Power System Stability

Power System Dynamics
Prof. M. L. Kothari
Department of Electrical Engineering
Indian Institute of Technology, Delhi
Lecture - 01
Introduction to Power System Stability Problem
Friends, during this semester we will be studying a course on power system dynamics.
(Refer Slide Time: 01:07)
This power system dynamics is also popularly known as power system stability. The modern
power systems are very widely interconnected while this interconnections result in operating
economy and reliability through mutual assistance. They also contribute to the stability problem,
what I am trying to emphasize here is that the power systems are widely interconnected and the
stability problem has become a very important challenge to power system engineers because of
this large scale interconnections.
The powers for power system stability analysis the mathematical model of the system is required
to be developed. The mathematical model is a set of nonlinear differential equations and a set of
algebraic equations, for this nonlinear differential equations and set of algebraic equations there
is no formal solution available. These equations are to be solved using non-linear using
numerical techniques, I am sorry using numerical techniques and the numerical techniques take
lot of time to solve the equations.
Over the years the power system stability has posed a problem to the power system engineers.
The problem is posed in two respects; the one is the modeling of the system. To get the correct
assessment of power system stability a detail model of the power system need to be developed.
Once the mathematical model is correctly developed, one has to obtain the solution through
numerical techniques. Because of the large size of the power system, the numbers of differential
equations are large in number and therefore solution through numerical techniques takes
enormous time.
Historically, historically the stability problem has been attempted from 1920 onwards. Earlier we
were not having digital computers and therefore computations were mainly done using hand
calculations or in those days slide rules were available for calculating. Somewhere in 1950 or so
the analog computers were developed and these were used for simulating the power system
stability problem.
Then in late 1950s digital computers came in and the first digital computer program for power
system stability was developed in 1956. At that time the program was mainly to analyze the
transient stability of the system. Over the years another development took place that is the
implementation or application of high response excitation systems. The high response excitation
systems were capable of improving transient stability of the system but with the application of
the high response excitation systems resulted into a problem of poor damping of the system
oscillations.
This problem of poor damping of system oscillations has been overcome by implementing what
is commonly known as the power system stabilizers. During this semester we will try to
understand, the development of mathematical model of the power system, the mathematical
model includes the mathematical model for synchronous machine, excitation systems, voltage
regulators, governors and loads. For having the correct assessment of the stability the models
need to be accurately developed.
Earlier due to the computational difficulties, many assumptions were made and therefore the, the
assessment through computation and the actual groundility there used to be lot of difference.
Now, this course we will cover in about 40 hours the text for, textbook for this course is power
system stability and control by Prabha Kundur. This is a very nicely written book. Over the years
a number of books have been develop, written a lot of research work or research papers are
available, from time to time i will refer to other books and research papers which are relevant to
our study.
The first lesson on power system dynamics is, introduction to the power system stability
problem. Today, we will address to these aspects, basic concepts and definitions classification to
the power system stability rotor dynamics and swing equation and swing curve.
Now let us first define what is power system stability? Power system stability may be broadly
defined as that property of a power system that enables it to remain in state of operating
equilibrium under normal condition and to regain an acceptable state of equilibrium after being
subjected to a disturbance.
If you carefully look into this definition, you will find that one has to emphasize on ability to
remain in operating equilibrium and second point we have to emphasize is the equilibrium
between opposing forces. A power system is subjected to a variety of disturbances, it is never in
steady state condition. Small disturbances in the form of load changes continuously come, while
large perturbations in the form of faults, tripping of lines, change in large load and dropping of
generators do come in the system.

A power system is designed and operated in such a fashion. So that it can withstand certain
probable contingencies. For the purpose of understanding this problem, the power system
stability problem is classified into 2 broad categories, one is called voltage stability and second is
called, I am sorry one is called angle stability and second is called voltage stability. Earlier years
the stability means angle stability. During the last one decade the another type of stability that
has come into picture is the voltage stability.

The angle stability is further classified into small signal stability, transient stability, mid-term
stability and long-term stability. We will briefly define these different types of stability problems
and try to understand what we mean by the different terminologies. Voltage stability is also
defined into two broad categories, one is called larger disturbance voltage stability and second is
small disturbance voltage stability.
Now, let us define what is rotor angle stability? The rotor angle stability is the ability of the
interconnected synchronous machines of a power system to remain in synchronism. Here the
emphasis on ability to maintain synchronism. This is the primary requirement for the operation
of a power system where, where all the machines of the system remain in synchronism.
Now for the system to remain in synchronism we have to study the torque balance of
synchronous machines, that is synchronous machine is the primary component here, where we
have to maintain equilibrium between the torque supplied by the prime over and the
electromagnetic torque developed by the synchronous generator. To analyze this power system
stability, we have to first understand the dynamics of the rotor and develop a mathematical
equation to describe the dynamics of the rotor.
For developing the basic equation, we make use of the principle of dynamics, elementary
principle of dynamics as for the rotational dynamics we all know that the accelerating torque is
the product of moment of inertia and angular acceleration that is for any rotating body the
accelerating torque is equal to the moment of inertia multiplied by angular acceleration and this
is the fundamental law on which actually the swing equation is best. Now as you know that
synchronous machine may operate as a synchronous generator or a synchronous motor.
When we look at the rotor of the synchronous generator there are two torques which act on the
synchronous generator rotor are, one is the mechanical torque which acts on the system and
another is the electrical torque or we can call it electromagnetic torque which acts on the, these
two torques operate or act on the rotor in opposite direction.
The mechanical torque is provided by the prime over and electrical torque is developed due to
interaction of magnetic field and a stator currents. The rotor rotates in the direction of
mechanical torque that is if you just look here, if I show the direction of rotation, it is in the same
direction as the mechanical torque is applied to the system.
Under steady operating condition, these two torques are equal and the rotor of the synchronous
machine rotates at synchronous speed. However, however when disturbance is occur, there exist
a unequilibrium between the two torques, these two torques are not equal and this this difference
is called accelerating torque.
When I look at the synchronous motor thus the driving torque is developed by the flow of
electric power from the supply and it meets the load torque, therefore load torque becomes the
breaking torque while the driving torque is the electrical torque and the rotor rotates in the
direction of in which the electrical torque is developed.
The swing equation or actually I will the swing equation is to be derived, a differential equation
can be written relating the accelerating torque moment of inertia and acceleration. This can be
written as J into d square theta m by dt square equal to Ta, where Ta is difference of Tm minus of
Te. Tm is positive for generator operation and T is also positive for generator operation, while for
motor operation they take negative signs. When they use mks system of units, J the total moment
of inertia expressed is in kilogram meter square, theta m the angular displacement of rotor with
respect to the stationary axis in mechanical radians. Now I am here emphasizing that in this
equation theta m is measured with respect to a stationary axis
Time is measured in seconds. Tm is the mechanical or shaft torque supplied by the prime over
less retarding torque due to rotational losses in Newton meters, that the unit for the torque is
Newton meters, this Tm is torque supplied with the prime over to less rotational losses. It is the
torque which is available for rotating the rotor. Similarly, Te is the net electrical torque or
electromagnetic torque in Newton meters. Ta is the net accelerating torque in Newton meters.

Now, if you see this equation then this theta m increases continuously even under steady state
conditions because theta m is measured with respect to stationary axis, okay. Now instead of
measuring the angle with respect to a stationary axis, the angle can be measured with respect to a
synchronously rotating axis, therefore theta m can be defined as omega sm into t plus delta m,
where omega sm is the synchronous speed of the machine. This is measured in radians per
second.

Delta m is angular displacement of the rotor in mechanical radians from the synchronously
rotating reference axis. Now this equation two, when you when you take the derivatives of this
equation with respect to time, the first derivative is written as d theta m by dt2 equal to omega
sm plus d delta m by dt, that is we can see the rotor speed is equal to synchronous speed plus this
additional torque.
Now we take the second derivative, d2 theta m divided by dt square is equal to d square delta m
by dt square, okay. Omega sm being constant its derivative is 0. Now we substitute the value of
rotor acceleration in the equation 4, we get expression as J times d square delta m by dt square
equal to Ta Tm minus Te. Now at this point I want to emphasize that all the terms in this equation
are torque terms in Newton meters. In power system studies, we are more comfortable with the
terms in power, power may be watts, kilowatts, megawatts and therefore, what we do is we
multiply this equation 5 by omega m that is d theta m by dt.
Now when we multiply this by this term omega m, our equation becomes J into omega m d
square delta m by d square equal to omega m Tm minus omega m Te. Now here, this speed into
torque, this is the power, the speed is measured in radians per second, power or torque is
measured in Newton meters this product is power in watts.
Now we will represent this term omega m Tm as Pm, omega m Te as Pe, where Pm is shaft power
input to the machine less rotational losses. We are emphasizing all the time, the term rotational
losses here. The meaning here is that this is not the torque supplied by the prime over but it is the
prime over torque minus rotational losses.
Pe is the electrical power crossing the air gap also called air gap power. I mean in this equation 7,
we substitute the expression for power and the equation become J into omega m d square delta m
by dt square equal to Pm minus Pe equal to Pa. Now this term J into omega m, J is the moment of
inertia in kilogram meter square, omega m is the speed in radians per second. Now this product
is called angular momentum. Now in practical situations, the rotor speed omega m is very nearly
equal to the synchronous speed.
The difference is very small, the difference may become large only when machine loses
synchronism right and therefore, what we do is for the purpose of simplicity, if it represents
omega m in this equation by this omega sm, then this product of omega sm and moment of
inertia J, it will be denoted as the inertia constant M and therefore my equation take the shape M
d square delta m by dt square equal to Pm minus Pe, where m is known as inertia constant.
Strictly speaking this coefficient should be J into omega m and since omega m is not constant
therefore this coefficient term, strictly is not constant but by assuming making this assumption
omega m is equal to omega sm, this coefficient become constant and it simplifies our analysis,
further the error incurred by making this simplification is negligible. Okay therefore we are
justified in assuming this coefficient of the acceleration term as constant and this is called inertia
constant M.
Now at this stage, I would like to emphasize that this term M, varies over a wide range
depending upon size of the machine, type of machine because we have two main types of
synchronous generators hydro machines, hydro generators and turbo generators and these two
machines have widely different value of inertia constant M and this inertia constant M will be
different depending upon the size of the machine.
Now to overcome this problem, we define another term or another inertia constant which is
denoted by the symbol H. You have to we have to very clearly understand the definition of H,
the inertia constant H or the constant H is defined as the stored kinetic energy in mega joules at
synchronous speed divided by machine rating in MVA. For any synchronous machine, when is
when the rotor is rotating at synchronous speed, we can find out what is the kinetic energy stored
in the rotor and we divide this kinetic energy stored in the rotor by machine rating. If we denote
the kinetic energy stored in mega joules and machine rating in MVA, then the unit of this term
becomes mega joules per MVA. We can also represent in kilo joules per kilo watt, per KVA not
kilo watt, per KVA but still the machine ratings are normally in MVA, we prefer to use the rating
in MVA and energy stored in mega joules, okay. Therefore, this can be written as 1 by 2J omega
sm square upon the machine rating in MVA as mach.
Now we will develop a relationship between the inertia constant M and constant H. So that the M
which is there in our differential equation will be replaced by constant H, from this equation 10,
we can write the inertia constant M as 2H upon omega sm into ah machine rating S mach.

Now we substitute for inertia constant M in our equation, we find here that this becomes 2times
H omega sm d square delta m upon dt square equal to Pm minus Pe upon of S mach. Now if we
assume, MVA based for the system as the machine rating in MVA, then Pm upon S mach
becomes the per unit mechanical power, similarly Pe divided by S mach becomes the per unit
electrical power okay. Therefore, this equation can now be written in terms of power expressed
in per unit, further as you all know that in power systems the per unit system of calculations are
very convenient and therefore the equation is manipulated and written in terms of per unit power.
Therefore, the equation 13 here is written as 2H upon omega sm d square delta m divided dt
square equal to Pm in per unit minus Pe in per unit. Now every time writing per unit is not very
convenient, therefore what we do with that, we drop this nomenclature here but we keep in mind
that power is expressed in per unit.
Therefore with this, we can write down equations describing the rotor dynamics of the
synchronous machine as 2H upon omega sm d square delta m by dt square equal Pm minus Pe. At
this stage we have to very careful and understand very clearly that this P m and Pe are expressed in
per unit while expressing Pm and Pe in per unit, the MVA base is the machine rating in MVA.
For any machine which I am writing this differential equation, if we assume the MVA base okay
then the MVA base is the machine rating in MVA and therefore Pm and Pe are expressed in per
unit in terms of MVA base of the machine.
Now here, this delta m is expressed in radians per second, I am sure radians not radians per
second, this is the correction omega sm is expressed in radians per second. The units of these two
terms should be consistent. We know a term known as angle in electrical radians, speed in
electrical radians per second.
Therefore, if we use instead of the angle in mechanical radians and speed in mechanical radians
per second, we can use delta in electrical radians and omega sm in electrical radians per second
and therefore what we do is that keeping in mind that these 2 terms would have the consistent
units, we drop the subscript m and we write the equation in the form 2 times H upon omega s
equal to d square delta by dt square equal to Pm minus Pe. In this is, this equation is known as the
swing equation of the synchronous machine.

This is applicable to generator as well as to motor, only difference is when we write this equation
for motor Pm becomes negative and Pe also become negative. So, that this term becomes Pe
minus Pm. Now here, we can express delta in radians, omega s in radians per second. If we do in
this manner, we can write the equation in the form H upon 2, H upon phi f that is omega s can be
replaced by 2 phi f and you will have the equation in the form H upon phi f d square delta by dt
square equal to Pm minus Pe okay, many time we express delta in electrical degrees and omega in
electrical degrees per second.
In that case omega s will be replaced by 360 into f. So that the resulting equation will be in the
form H upon 180 f d square delta by dt square equal to Pm minus Pe. We do use swing equation
either in this form or in the previous form but while using this equation, we have to very careful
in expressing the delta in proper units. You can in this case delta is in the electrical degrees while
in the equation 16 coefficient is H upon phi f delta is in electrical radians. As we will see here
actually that, this is the basic swing equation on which or around which the stability analysis of
the power system depends
Now we have 2 terms here Pm minus Pe, this Pe which is the electrical power output of the
machine is, is non linearly related to or it is non-linear function of delta and therefore, this
differential equation becomes a non-linear differential equation or we can say that the swing
equation is a second order non-linear differential equation. Suppose, you have any system a
number of machines number of synchronous machines then we have to write down swing
equation of each machine while in this, we have to write down the correct expression for
electrical power output.
Now when this swing equation is solved, the solution of the swing equation is known as swing
curve that is what we will get actually when we solve this equation, if you solve this equation
you will get the delta as a function of time delta as function of time and therefore we can define
this swing curve.
When the swing equation is solved we obtain the expression for delta as a function of time. A
graph of the solution is called swing curve of the machine and inspection of the swing curves of
all the machines of the system will show whether, whether the machines this is the wrong
mistake, I have whether the machines remains in synchronism after a disturbance.
I am writing here whether the machines remain in synchronism after a disturbance again this is a
big mistake here after a disturbance. For the purpose of solving the second order differential
equation, using numerical techniques or standard practice is to represent the second order
differential equation in terms of 2 first-order differential equations because when we apply the
numerical techniques for solving the differential equations using digital computers, it handles
first-order differential equations.
We can write down the second order differential equation as 2 first-order differential equations
by defining by defining a term omega but the omega is the actual speed of the machine and
therefore we can define these two differential equations in the form, 2H upon omega s, d omega
by dt that is you are having the d square delta by dt square therefore omega is defined as d delta
by dt okay. So that now here, we have one variable omega, then the second equation becomes d
delta by dt whose expression comes out to be omega minus omega s okay.
Now suppose you have a practical system where you have more than one machine that is called a
multi-machine system. In a multi-machine system, the output and hence the accelerating power
of each machine depend upon the angular position and, to the more rigorous also upon the
angular speeds of all the machines of the system. Thus for a 3- phase system, this is written for
mistake here again. Thus for a 3 machine system, for a 3 machine system, there are 3
simultaneous differential equations. The equations will look like this.

M1 d2 delta 1 by dt2 equal to Pm1 minus Pe1 which is function of delta 1, delta 2, delta 3and it is
also function of d delta 1 by dt, d delta 2 by dt, d delta 3 by dt. I will tell actually why this Pe1 is
function of this derivative terms also the, for the second machine we have to put inertia constant
of that machine and the electrical output of this machine is function of all these angles and these
derivatives.
Similarly, for the third machine, we have Pe3 function of all these derivatives. Now since the
system, we are considering here is in dynamic condition. Okay when system rotor is in dynamic
condition, it develops some damping torque and this damping torque is proportional to the speed
deviation. Then, let us correct it is proportional to the speed or speed deviation with respect to
the synchronous speed. Now this d delta 1 by dt, d delta 2 by dt, d delta 3 by dt, these are the
rotor speeds with respect to the synchronously rotating reference speed.
Now to simplify our stability analysis, we ignore this damping terms. Once we ignore this
damping terms, the electrical output will exclusively be function of these angles only okay. Now
I would like to tell something more about the electrical power output.
The important characteristic that has a strong bearing on power system stability is the
relationship between interchanges power and the position of rotors of the synchronous machines,
this relationship is highly non-linear. I have told you here the electrical power Pe right is function
of rotor angels and it is non-linear. Now if I take a simple example to illustrate this, let us
consider this case, we have 2 machines connected by a transmission line.
We can represent this machine 1 by a voltage source in series with a reactance. Similarly, the
machine 2 can be represented by a voltage source in series with a reactance. The machine 1 is
supplying power to machine 2. This can be understood as machine 1 is a synchronous generator
machine 2 is asynchronous motor. For this simple system, if we develop an expression for
electrical power output from machine 1, then this can be derived by writing or by first drawing a
phasor diagram.

The phasor diagram can be simply drawn in this fashion. We can start with the terminal voltage
of the machine, synchronous machine 2. Let I is the current, the internal voltage of the
synchronous motor is represented by this phasor Em, to this Em we add this voltage drops, where
the voltage drops which take place in the internal reactance of the motor 9 reactance and
generator reactance to get the internal voltage.
Therefore, this phasogram diagram shows that Em to this Em, if we add this voltage drop I into
XM, we get the voltage ET 2, to this we add the voltage drop I into XL, we get the voltage ET 1
and to this we add the voltage drop I into XG, we get the internal voltage of the synchronous
generator. Delta which is the phase difference between EM and EG is called power angle and
this delta is sum of all these three angles delta G delta L and delta M.
If we plot, a graph relating power angle delta and power P or electrical power Pe, we call this as
Pe, then this graph comes out to be a sine curve, the expression for Pe is the internal voltage of the
generator, the magnitude of this voltage, internal voltage of the motor EM divided by the total
reactance X into sine. Now, if we see here as the delta increases, the power output increases and
becomes maximum at delta equal to 90 degrees, on the same diagram if I draw the mechanical
input line, the mechanical input is not function of delta and therefore it comes out to be a line
parallel to delta axis okay.
In this diagram, we will write this as Pm mechanical power, now this mechanical power line and
this power angle characteristic intersect at 2 points, this is 1 point I call it a, this is another
point we call it b, a is stable equilibrium point while b is unstable equilibrium point that is
we will denote this a as stable equilibrium point normally called SEP and b is unstable
equilibrium point.
The system, if it is operating at this point a and if it is perturb then it develops the forces so as
so that it returns back to this operating point a. However if the system is made to operate
which is at the point b which is also the equilibrium point but if suppose some perturbation is
given then the system will lose stability. It cannot come back it cannot rest develop restoring
forces to come back to the positing point b and therefore, our stable operating point is a and we
shall represent this operating angle as delta that is when the system is under steady conditions the
mechanical power is equal to electrical power and operating angle delta naught.
Now I conclude here, what we have learn in this lecture. I have tried to give you the basic
classification of the power system stability. We have defined the stability involved terms, we
have developed swing equation of the machine and we have also defined a very important term
inertia constant H. We will continue further in the next lecture.
Prof. M. L. Kothari
Lecture - 02
Introduction to Power System Stability Problem (contd)
Friends, today we continue our discussion on introduction of the power system stability problem.
Today, I will discuss something more about inertia constant H, we introduced last time. I will
also discuss certain assumptions that commonly made in stability studies. We will discuss small
signal stability and voltage stability concepts; we shall be dealing with these topics in depth at
later stage. Today our intention is just to give you basic concepts related to a small signal
stability and voltage stability.
Last time, we have derived this swing equation, the swing equation is very important for analysis
of the stability of a power system, this equation is used for transitivity analysis and this is the
basis for small signal stability analysis also. This equation has the form the coefficient is 2H
upon omega s ah d square delta by dt square equal to Pm minus Pe. Initially when the equation
was written, the coefficient was M, the inertia constant. However, we define a another constant
known as the inertia constant H, let us see why we prefer inertia constant H for analyzing the
stability of the system.
The inertia constant H has the desirable property, that its value unlike that of M does not vary
greatly with the rated MVA and speed of the machine but instead has a characteristic value or set
of values for each class of machines. I had very clearly, very in particularly told you last time
that the coefficient M varies over a wide range depending upon the size of the machine, speed of
the machine, type of the machine.
However, the inertia constant H has some characteristic values and in that respect H is similar to
per unit reactance of machines as you all know that the per unit reactance of the synchronous
machines have some characteristic values. Similarly, the inertia constant H also has certain
characteristic values and sometimes in the absence of the specific knowledge about inertia
constant H, we can assume some characteristic value and do our calculations, manufacturers
specify or provide the value of H on the MVA rating of the machine.

This diagram shows the typical value of the inertia constant H for turbo generators. On this axis,
we have marked the rating of the machine in MVA and on this axis we have marked the inertia
constant H in mega joules per MVA. Now machines, where turbo generators are of different type
and different speeds, if you take a turbo generator whose lateral speed is 800 rpm a powerful
machine then the variation of the inertia constant is shown by this graph, that is when the rating
varies as something like about 10 mega, 10 MVA to say 100 MVA, the variation of H is in the
range starting from about 9.5, it goes to 6.5.
Similarly, for a turbo generator whose speed is 3600 rpm , the variation is seen in this graph, it
varies from 4 to as much as 7 or 7.5. Another graph which is shown here is the known
condensing type of steam turbines, the inertia constant again varies but over a certain range.
Now this graphs shows the characteristic value of inertia constant H for a hydro generator. Again
on the x axis, we have marked the rating of the machine in MVA and y axis, the inertia constant
H in mega joules per MVA. Here, you see actually that again you find that the machines having
different speeds do have different value of inertia constant H, it increases with increase in the
MVA rating of the machine but still you can see actually it lies in a certain range.
Now I would like to know from you what this inertia constant H is whether it is the inertia
constant of the prime over or whether it is inertia constant of the rotor of the synchronous
generator or it is inertia constant of the prime over and rotor grouped together, who will reply
this? The question I imposed is, that I am talking of inertia constant H for the synchronous
machine. This inertia constant whether it corresponds to the total inertia constant of the turbine
plus generator rotor or only generator rotor or only turbine.
Yes you are correct, the inertia constant is the total inertia constant of the system that is for the
rotating body okay. Now, here I will just give you a small example suppose in a power plant we
have two machines okay of different ratings and the information which is provided to us is about
the inertia constant on the MVA rating, if this two machines are tightly coupled. So that they
form a coherent group, coherent group means the machines will swing together then we can the
place this machine by equivalent machine.
Now just take the example, let us say that unit 1, there are two units are there in the plant, unit 1
is 500 MVA and its is a constant H1 is specified as 5 mega joules per MVA, unit 2 7000 MVA its
inertia constant say H2 is 3mega joules per MVA because the manufacturer provide the
information about this H constant on its on MVA rating okay.
These two machines are in the same plant, we want to find out the equivalent inertia constant of
the2 machines, how do we compute the equivalent inertia constant for that, we need one more
information that on both MVA basis. We want to find out the equivalent inertia constant. Let us
assume that we want to find out the equivalent inertia constant H on base MVA equal to 100
okay. Then the calculations proceed in a straight forward manner, first we find out the total
kinetic energy stored in the two machines, that is total kinetic energy will be equal to, take the
basic definition for the first machine it is going to be H that is 5, MVA rating is 500 plus second
machine H is 3 and MVA rating is 1000. This comes out to be 5500 mega joules.
Therefore, H I can say equivalent on base MVA equal to 100 is 5500 divided by 100 because
base MVA is 100, it comes out to be 55 mega joules per, this is the way one can find out the
equivalent inertia constant of a number of machines which are running in the same plant and are
coherent. Now before we talk further about the different types of stability, I would like to tell the
the commonly made assumptions for stability analysis.

The first assumption which is very commonly made is the mechanical power input remains
constant during the period of the transients, if you look into the swing equation, our swing
equation is 2H divided by omega s d square delta by dt square equal to Pm minus Pe okay. But
this assumption which I am telling you here is that this Pm will be assumed to remain constant
during the transient okay.
The logic for making this assumption is that in any any prime over whether is a stream turbine or
is a hydro turbine right, the mechanical power input is governed by governor and by
supplementary control action which you normally call a set point control. Now the governors
sense the speed deviation and actuate the turbine walls. So that more steam enters when the
speed drops and when speed increases the walls will be closed to decrease the speed decrease the
speed by decreasing the steam input to the machine.
In practice, the speed deviation is very small during the transient stability analysis until unless
the machine looses synchronism, the machine deviation is very the speed deviation is very very
small and therefore governors do not act okay and therefore, this assumption is a very very valid
assumption in all stability studies. We will assume this Pm to remain constant during the transient
period.
The second very important assumption is, we neglect damping or asynchronous power in the
stability studies and you can say the damping or asynchronous power is negligible. In fact when
the machine is in dynamic condition, it develops a synchronous power as well as asynchronous
power. This asynchronous power results into damping, now the swing equation which I have
derived, in this swing equation we have not accounted for damping.
Suppose, I want to modify this swing equation including damping aspect then I have to add one
more term here and that becomes minus D times d delta by dt. The damping torque will always
act in a direction opposite to the dragging torque, okay. So that is the accelerating torque at any
instant of time will now becomes Pm minus Pe minus D times d delta by dt but to simplify the
analysis we ignore this damping term and therefore, the second assumption is the damping or
asynchronous power is negligible.
The third assumption which is made the synchronous machine can be represented by constant
voltage source behind a transient reactance. This assumption is, is not very truly applicable
however, for classical stability studies when you make this assumption the results obtained are
not very far away from the actual results. As we will see when we talk about, the detail model of
the synchronous generator right, we can represent this synchronous generator model in more
detail while for performing classical stability analysis, this assumption can be made and this
results into a very simplified model or simplified representation of synchronous generator.
However actually in our studies, we will develop detail mathematical model of the synchronous
generator and this assumption will need not be made at that time. Another very important
assumption which is made is the mechanical angle of the synchronous machine rotor coincides
with the electrical phase angle of the voltage behind transient reactance, okay as here we have
assumed that a synchronous machine can be represented by a constant voltage behind transient
reactance okay and further assumption which is made is the phase angle of this voltage coincides
with the rotor angle
The last and the most important assumption here is the synchronous power may be calculated
from a steady state solution of the network to which the machines are connected because in our
swing equation, we have to compute the Pe , that is the electrical power output from the machine
this is the synchronous power okay and for computing the synchronous power, we assume that
the network is in steady state okay, although when the rotors of the synchronous machines are
oscillating right, the network is not in steady state in the real sense that is the voltage and the
frequency also is deviating slightly depending upon the but since this quantity is very small as
compared to the power frequency and the computations can be conveniently done by assuming
that the network is in steady state condition.

Now while classifying the power system stability we have classified into 2 categories angle
stability and voltage stability, angle stability was further classified into different categories.
Now, we will just look quickly, how we define the small signal stability the small signal stability
is the ability of the power system to maintain synchronism under small perturbations, this is a
important point, that is the machine or synchronous machine has the ability to maintain
synchronism, this is primary requirement for stability under small perturbations or small
disturbances.
The small disturbances are small variation in loads and generations, any power system loads
keep on changing continuously therefore this is one example of small disturbance, the second
important aspect here is the disturbances are considered to be sufficiently small for linearization
of the system equations to be permissible for purpose of analysis. As I have told you earlier the
swing equation is a non-linear differential equation, okay but when we consider small
perturbations this equation can be linearized around a nominal operating condition.
Once the equation is a linear equation, we can make use of linear control theory as well as we
can get the closed form solution of the differential equation once it is a non-linear differential
equation I have to solve by applying numerical technique however, if the differential equation is
a linear differential equation, we know the closed form solution for the equation and once the
differential equations are linear we can use a linear control theory to design controllers for the
system. Now here, I would like to take one example how do we linearize a non-linear differential
equation.
Our non-linear differential equations which we will consider is, same as the swing equation
derived okay. Now, when I say that we linearize this equation around a nominal operating
condition or point. The nominal operating condition here can be characterized by the initial value
of the angle delta. Let us say, machine is operating where delta is equal to delta naught and we
will assume here that Pe is the electrical power is given by this equation Pmax sin delta okay.

Now under steady state operating condition Pm is equal to Pe that is Pmax sin delta O, that is I
substitute the value of delta equal to delta O, that will give me the electrical power output and
under steady state condition these 2 powers are equal. Now what we do is that we give an
incremental change or we make an incremental change in delta, let us say that delta is equal to
delta naught plus delta okay.
Let us substitute, this in our equation, swing equation it becomes 2H upon omega s d square delta
naught plus delta divided by dt square, Pm we assume to be constant it remains Pm, Pmax sin I put
here delta naught plus delta, okay. Now you expand this equation and write down Pm minus Pmax
sin delta O cos delta, okay plus plus cos delta O, sin delta okay. Now let us look the at this
equation and since we know that, delta delta is a incremental change, small one.
So that cos delta delta is very nearly equal to 1, sin delta delta is very nearly equal to delta delta
good okay. We will make this substitution in this equation now let us write down what is the
value of this derivative. The derivative of delta O is 0 and therefore this derivative term becomes
d square delta delta divided by dt two equal to Pm minus Pmax into sin delta O cos delta delta we
have made it one, okay and I put plus here plus cos delta O delta delta
Now Pm minus sine P max sin delta O, Pmax sin delta O is equal to Pm therefore this becomes now
equal to minus Pm cos delta O delta delta okay or now I can write down my differential equation
as 2H upon omega s, d square delta delta by dt square plus Pm. Pmax right, we will write in Pmax
here cos delta O, this is also Pmax delta delta. This differential equation is a second order linear
differential equation, do you agree, why it is second order linear differential equation?
Because, here we do not have delta delta or any term, it is trigonometrical term or non-linear
term. So far this term is concerned, this coefficient is concerned this is a constant because Pmax is
known, delta O is known Pmax cos delta O is known. Pmax cos delta naught is known as the
synchronizing power coefficient, this term is known as synchronizing power coefficient and may
be denoted by the symbol Sp, that is Sp is equal to P max cos delta naught. So that we can write
down the swing equation as 2H divided by omega s d2 delta delta by dt2 equal plus Sp delta delta.

Further this is generally written in the form, sorry is written in the form d square delta delta by dt
square plus omega s, Sp divided by 2H delta delta equal to 0, in case Sp is positive the solution of
this equation will be similar to, similar to the solution of a simple harmonic motion. It is going to
be a sinusoidal, in case Sp is negative then it gives you a exponentially increasing value of delta
as a function of time. Therefore primary requirement for the system to be stable here stable under
small perturbations is that this synchronizing power coefficient should be positive.
Now, when we write the equation for simple harmonic motion the equation is written as d square
x by dt square plus omega n square x equal to 0, where omega n is the natural frequency of
oscillation. Now here by comparing the coefficients, we can write down that omega n is equal to
square root of omega s, Sp divided by 2H. This will be in radians electrical radians, we call it
electrical radians per second.
Now further, you can see that the synchronizing power coefficient is the slope of the power angle
characteristic at the operating condition, that is if you find out dPe divided by d delta it comes out
to be Pmax cos delta O and therefore dP by d delta is the slope of the power angle characteristic
and our primary requirement is that the synchronizing power coefficient should be positive for
the system to be stable under small disturbances, and that is why when I talk to you about the
stable and non-stable operating points, where this was our power angle characteristic P versus
delta.
This was your mechanical input line, these are the operating condition and this is the stable
operating condition because at this point the slope of the power angle curve is positive. I shall
take one small example to find out what is the frequency of oscillations, which we normally
come across.
Let us take a case where a synchronous machine as H equal to 5 mega joules per MVA. Further
we assume that Pe equal to 2 sin delta expressed in per unit 2 is your Pmax now. Let delta equal
to delta naught equal to 60 degree, okay. Then Sp comes out to be equal to 2 cos delta naught
which is equal to 1 okay. Let us take system frequency as 50 hertz, f equal to 50 hertz, omega n
natural frequency of oscillation of the rotor can be computed by substituting the values of small
s, omega s is 314 for 50 hertz system, Sp has come out to be equal to one synchronizing torque
coefficient 2 into 5H is 5.
This quantity is square root of 31.4, if you calculate it comes out to be equal to 5.6 electrical
radians per second. Frequency of oscillation in hertz can be obtained as omega n by 2 phi which
comes out to be in this case 0.89 hertz. This example is given here to give you the frequency of
oscillations we come across, system frequency is 50 hertz but when the rotors oscillate the
frequency of oscillation depends upon initial operating condition which determines the
synchronizing power coefficient, it depends upon the initial constant H of the machine and it also
depends upon the system frequency okay and it is of the order of 1 hertz in this case but it does
vary depending upon the value of this parameters but the variation is in the range of point 5 to 2
hertz normally.
Now we have defined the small signal stability and small signal stability is the ability of the
synchronous machines to remain in synchronism under small perturbations. Now when the
synchronous machines rotors are oscillating it develops an electrical power delta T, when the
rotors are oscillating it produces a electrical torque we call delta T. This torque can be
decomposed into two components, one is called synchronizing torque, another is called damping
torque.
The synchronizing torque is in phase with speed deviation, I am sorry there is a correction it is in
phase with phase with angle deviation delta delta and the damping torque is produced in phase
with speed deviation. when the rotor is oscillating and if the oscillations are sinusoidal in nature
then we can represent this torque as a phaser in a plane where, delta delta is the x axis and delta
omega is the y axis
Look here, actually in this diagram the plane is x axis is marked as delta delta y axis is delta
omega and in case, the delta T the change in electrical torque lies in the second quadrant then
you have 2 components, 1 component is damping torque component, another is the synchronized
torque component. Damping torque component is positive while synchronizing torque
component comes out to be negative

In case delta T lies in the first quadrant of the delta delta delta omega plane, then both
synchronizing torque and damping torques are positive for stability of the system both
synchronizing and damping torque should be positive. Okay I will discuss these aspects slight in
more detail when we talk about the small signal stability problem.
The small signal stability can be classified into 2 categories, one is called non oscillatory
instability not a stability instability, it can be non oscillatory instability this occurs because of
insufficient synchronizing torque, another is oscillatory instability this is due to insufficient
damping torque or by due to unstable control action.
Now we can see the response of the system under small perturbations, under 2 conditions, one is
that synchronous machine has constant field voltage there is no voltage regulator acting, another
case is where the synchronous machine is provided with a voltage regulator where we call this as
one as a constant field voltage another is with excitation control . Now suppose a situation is like
this where you have a negative synchronizing torque while positive damping torque the response
of the rotor to small perturbation will be of this form, starts from 0 and these oscillations are
oscillations are like damp because damping is positive but since synchronizing torque coefficient
is negative it increases monotonically. Therefore, this represents a case of insufficient
synchronizing torque.
Here is another case where synchronizing torque coefficient is positive but damping torque
coefficient is negative, the response of the rotor or del to perturbation starts from 0, you can see
actually that the oscillations are growing. Okay, therefore this is the case of oscillatory instability
and this is primarily due to lack of damping torque or here I can say that the system has negative
damping. In any power system, we will come across different modes of oscillations, a power
system is a large interconnected system okay.

We come across a local modes, inter area modes, control modes and torsional modes. The local
modes are the one where a synchronous machines located in a power plant oscillate with respect
to rest of the system and the frequency is in the range of .7 to 2 hertz. We have been taken the
example we found that frequency came out to .89 okay. Similarly, another case comes where a
group of machines of a power system oscillate with another group of machines connected by a
weak tie line and these group of machines will form one area another group forms, another area
and therefore we say this is inter area oscillation and when the group of machines are considered
the inertia constant H is large and the frequency of oscillation is in the range of point 2 to 1 hertz.
So for control modes are concerned the frequency of the different modes depend upon the
actually control phenomena therefore, no specific frequency is specified. The torsional modes is
another phenomena which occurs in the steam turbine and the different frequencies which we
come across are for typicals mass system it comes out be 16.3, 24.1, 30.3 and 44 hertz.
We will devote about two lectures to deal with about the torsional modes in detail at that time it
will be clear, how the different masses which are mounted on the same shaft will result into
different modes of sub synchronous oscillations and I will conclude here, my today is
presentation by summarizing what we have done.
We have addressed to 3pecific aspects, one is inertia constant H its importance, second aspect we
have considered here is the basic assumptions which are made instability analysis and the third
aspect we have studied is the small signal, stability, its definitions ,we have also talked about
how to linearize a non-linear differential equations and how to compute the natural frequency of
oscillations through an example. We will cover rest of the aspects of the introduction to power
system stability in the next lecture.
Prof. M. L. Kothari
Lecture - 03
Introduction to Power System Stability Problem (contd)
Friends we shall discuss today, 3 important aspects of stability analysis, One is the transient
stability voltage stability and long term and mid-term stability. All these topics will be covered in
detail in subsequent lectures.
Today I will simply give you the basic concepts related to these three stability problems. First we
consider the transient stability. As I told you earlier that originally, the transient stability was the
only stability which was of concern and which was analyzed in detail.
The transient stability is the ability of the power system to maintain synchronism when subjected
to severe transient disturbances. Let me emphasize here, that transient stability is the ability of
the power system to maintain synchronism subjected to severe transient disturbance, the word
severe transient disturbance is to be carefully understood the disturbances occur suddenly they
are severe in nature.
Now the examples of the disturbances are transmission line faults sudden large load changes is
loss of generation, generating units and line tripping. Now these are the examples which can be
consider which are considered for studying the behavior of the system. Now some of the
important features we have to understand about the transient stability are, the resulting system
response involves large excursions of rotor angles and is influenced by non-linear power angle
relationship.

Now here I am again trying to emphasize that whenever large disturbances occur then the, then
the resulting response involves large excursion of rotor angles, the delta varies over a wide range
and therefore the power output is highly non-linear with respect to delta, this is one important
aspect and therefore transient stability analysis require the solution of non-linear swing
equations.
The stability depends on both initial operating state of the system and the severity of the
disturbance that is for each operating condition the behaviour of the system is going to be
different. Similarly, the severity of the disturbance, now severity can be depends upon the type of
fault, location of fault and duration of fault.
Similarly, when we talk about the large load change what is the quantum of load which has come
on the system. Similarly, when we talk about the generator dropping what is the amount of
generation that has been taken out from the system and therefore these two important points we
have to carefully understand are that it depends upon the initial operating condition and severity
of the disturbance.
Usually the system is altered, so as so that the post disturbance steady state operation differs
from that prior to the disturbance. Here, suppose the system is initially operating in a bulk steady
state operate condition, the moment disturbance occurs there will be a change in the topology of
the system and therefore if the system is stable, the new steady state operating condition will be
different from the previous one.
The disturbances of usually varying degrees of severity and probability of occurrence can take
place in the system. Now when you talk about a practical system, the disturbances which occur
have different degree of severity and different probability of occurrence. When you talk in terms
of what you mean by the probability of occurrence in any power system faults occur in the
system but the probability of occurrence of line to ground fault is the maximum and the
probability of occurrence of 3 phase fault is minimum right and therefore one has to look into
this aspect while analyzing the transient stability of the system and also while designing the
system for operation and planning.

The system is designed and operated so as to be stable for a set of contingencies for example,
single phase to ground fault, phase ground fault or 3 phase fault. Now here when we talk about
the contingencies, we have to contingency stand for these all abnormal conditions okay and
whenever we design the system we have to see that what are the most probable contingencies
and the system should withstand those contingencies and it should not loose synchronism when
such contingencies occur.
Now when we analyze the transient stability of the system usually we assume the faults
occurring on the transmission line, sometimes we consider the faults at the buses or on the
transformers. In any system, when the faults occur the faulty system or faulty element is
disconnected by the operation of protective system and circuit breakers therefore, we have to be
very careful when we consider the analysis of transient stability, the operation of circuit breakers
and protective relays. In some cases, we consider the high speed reclosure also that is circuit
breakers have the capability to reclose after sometime delay.
Now this reclosure may be successful reclosure or it may be unsuccessful reclosure, in case the
fault is transitory in nature and it clears by itself then when you reclose the circuit, you will
regain the original operating condition. However, if suppose there is a permanent fault right then
the reclosure is going to be unsuccessful we will study the affect of auto reclosure on the
transient stability in our late, later lectures. With this basic points related to the transient stability
I will just show you a typical response of delta as a function of time for a particular for a given
disturbance.
This figure shows the behaviour of a synchronous machine under stable and unstable operating
conditions. Now I have shown in this diagram 3cases: case 1, case 2, and case 3. Case 1 shows
that following the disturbance, the power angle increases attains maximum value then further
decreases and oscillates with decreasing amplitude, this is a stable case. Case 2 is 1 where the
response is shown in curve 2, in this case the delta increases continuously till system becomes
unstable or till machines loose synchronism.
This type of instability is generally called first swing instability and this happens particularly due
to insufficient synchronizing power, the third case is a case where the power angle delta
increases attains maximum but subsequently the oscillations grow in magnitude This is a case
where system is stable from the point of view of first swing stability but the system is becoming
unstable due to, due to lack of damping torque that is the new operating condition is not small
signal stable.
Now these curves can be plotted for any operating condition, for any given severity or given
disturbance and then we can examine whether the system is stable or not by examining the swing
curves, they are all swing curves. These swing curves are obtained by solving swing equation of
the system, the method of solving this swing equations we will learn subsequently, okay.
We shall be devoting significant amount of time for analyzing the transient stability of a system
in our subsequent lectures. Now we address the problem of voltage stability, the voltage stability
has become the subject of studies during the last one decade. Now voltage stability is the ability
of a power system to maintain steady state voltages at all buses in the system under normal
operating conditions and after being subjected to a disturbance. You have to very carefully
understand the definition that is again it is the ability, stability is the ability of the system such
that it maintains steady state voltage at all the buses under normal operating condition and also
after disturbance.
Now here, we will see that actually when we talk about normal operating voltages that means
that acceptable voltage profile of the system sometimes the voltage may become very low it will
not be acceptable, we call this as a unstable system.
A system enter, enters a state of voltage instability when the disturbance increase in load demand
or change in system condition causes a progressive and uncontrollable drop in voltage. Now here
we have emphasizing the occurrence of voltage instability that is whenever a disturbance occurs,
the voltage drops and it is not controllable because any system, we have certain controllers which
control the system voltage but disturbance may be of such in nature. So that the controllers are
not in a position to control the system voltage and that is the condition for voltage instability.
The main factor causing instability is the ability of the power system inability of the power
system to meet the demand of reactive power. When you talk about the angle stability it is a
balance between reactive active power that is mechanical input of the turbine or mechanical
input from the turbine must be equal to electrical output from the generator, it is a real power
balance, here in voltage stability we require the balance between the reactive power that is
reactive power generated and the reactive power consumed in a system and in if in case the
system is unable to maintain this balance then the problem of voltage instability occurs.
The heart of the problem is usually the voltage drop that occurs when active power and reactive
power flow through inductive reactances associated with the transmission network. When you
again look, what are the main causes of voltage drop the, the voltage drop takes place in
transmission system in transformers or any part of the ah system due to the flow of real and
reactive power and the system has its impedance resistance and reactance and because of this
there is a voltage drop in the system and this drop is the primary, primary cause of voltage
instability. We shall be studying this voltage stability problem in detail in our later or subsequent
lectures however, I will just give you one basic ideas that the criteria of voltage stability.
When do we say that system is voltage stable or system is voltage instable? A simple criteria is a
criteria for voltage is stability is that a given operating condition for every bus in the system, the
bus voltage magnitude increases as the reactive power injection at that at the same bus is
increased this point is very important suppose, you have a large system you have number of
buses. Okay, if we inject reactive power at a particular bus then the voltage of that bus would
increase, if the operating condition is such that by injecting voltage at any of the buses and the
just correct me by injecting reactive power at any of the system buses, the voltage should
increase there should be no case where where I inject the reactive power and the voltage
decreases. This concept can be further stated in the form when we talk about voltage stable to say
voltage unstable.

A system is voltage instable, unstable if for at least one bus in the system. The bus voltage
magnitude V decreases as the reactive power injection Q at the same bus is increased because
when I talked about the voltage stability the condition posed was that at all buses, the voltage
should increase when reactive power is injected if at any one of the buses the situation is
different that is when you inject reactive power, the voltage decreases right then if any at any of
the one buses this happens the system is voltage instable or voltage unstable.
Another way of stating this problem is a system is voltage stable, if V-Q sensitivity is positive
for every bus and voltage unstable, if V-Q sensitivity is negative for at least one bus, it may be
more than one bus. Okay but if at any one bus, any one bus the voltage sensitivity becomes V-Q
sensitivity, we know what is in V-Q sensitivity that is change in voltage divided by change in
reactive power injected if you find this ratio right this is called the voltage V-Q insensitivity and
if the V-Q sensitivity becomes negative at any one of the buses, we say the system is voltage
unstable, voltage instability is essentially a local phenomenon however, its consequences may
have wide spread impact.
Another next one which I wanted to mention is, the concept of voltage collapse sometimes we
say that voltage collapse taken place in the system. Now this voltage collapse is a very complex
phenomena which occurs due to occurrence of a sequence of events but when, whenever the
system voltage becomes unacceptable, the system voltage profile becomes unacceptable to us we
say that voltage collapse has taken place.
Okay for analyzing the voltage stability problem it is divided again into 2 categories, one is
called large disturbance voltage stability and another is called small disturbance voltage stability,
just like as we had in the angle stability we divided this problem into 2 categories, small signal
stability and transient stability, transient stability is particularly related to large disturbance is a
large disturbance stability right.
So the power system stability was classified as angle stability and voltage stability. We have
studied basic concepts of angle stability, now we shall study the basic concepts of voltage
stability. The voltage stability is also classified like angle stability into 2 categories, a large
disturbance voltage stability and small disturbance voltage stability. This voltage stability a
phenomena has been studied during the last one decade in depth a lot of research work has also
been done in the area of voltage stability.
We will just study today the basic definitions of large disturbance voltage stability and small
disturbance voltage stability and the factors which determine these 2 stability problems. The
large disturbance voltage stability is concerned with systems ability to control voltages following
large disturbances such as system faults, loss of generation and circuit contingencies. Here I
emphasize that systems ability to control voltages, now in any power system, the power system
voltage is controlled by controlling the controllable devices, the controllable devices are the, the
automatic voltage regulators on the synchronous machine.
We have online tap changers on the transformers, we have capacitors a shunt capacitors, shunt
reactors and other devices are static wire systems and these devices when they are controlled to
maintain the voltage on the system and if the system can control the voltage following the
disturbance then such system is said to be voltage stable.
The voltage stability is determined by system load characteristics and the interaction of both
discrete and continuous controls and protections when we talk about the system load
characteristics, you have to understand very carefully that the power consumed by load depends
upon the magnitude of the voltage and frequency. There are some loads whose power consumed
depends upon the voltage only, there are some loads whose power consumed it depends upon
voltage and frequency both, an example is a heating load, where the power consumed is directly
proportional to voltage square.
However, it does not depend upon the system frequency but if I consider the induction motor
load, then the power consumed depends upon voltage and frequency both and therefore, when
we talk about the power system stability then the power system voltage stability that it is
primarily determined by the load characteristic and the interaction of both discrete and
continuous control protections.
When I talk about discrete controls these are the on off controls and continuous controls the
example is automatic voltage regulators. Similarly, the static wire systems are continuous
controls the criteria for large disturbance voltage stability is that following a given disturbance
and following system control actions, the voltages at all buses reach acceptable limits. In any
system, we can operate the system in case the voltages are within certain acceptable limits, in
case the voltages can be controlled within the acceptable limit the system is said to stable. Now
we discuss the small disturbance voltage stability.
Small signal voltage stability is concerned with a systems ability to control voltages following
small perturbations, such as incremental changes in system load. When we talked about the large
disturbance voltage stability the disturbances were large in magnitude while here when we talk
about the small disturbance voltage stability the disturbances are small in magnitude, the
example of small disturbance is incremental changes in system load.

A criteria for small signal disturbance voltage stability is that for a given operating condition a
system is voltage stable, if V-Q sensitivity is positive for every bus and unstable if V-Q
sensitivity is negative for at least one bus. I will again repeat here the meaning of V-Q sensitivity
that is at any bus if I inject a reactive power then the change in voltage divided by change in
reactive power injected this gives the V-Q sensitivity and for the system to be stable at all buses
in the system V-Q sensitivity should be positive and if V-Q sensitivity is negative at anyone of
the buses system is said to be unstable voltage unstable.
The voltage instability does not always occur in pure form, often the angle and voltage stability
go hand in hand, here I want to emphasize that system is one and whenever disturbances occur it
affect the system angle as well as the voltage and therefore both these phenomenon go together
and whenever we analyze the total system stability, one has to take care of both the aspects we
shall study the voltage stability phenomena in detail in subsequent lectures.
Now we will define another very important term known as the long-term and mid-term stability
this is also a new phenomena earlier, the designers and operators of power system did not
consider this problem this long-term and mid-term stability has been introduced as a result of a
need to deal with problems associated with the dynamic response of the power system to severe
upsets. The severe upsets are those disturbances that result in excursions of frequency and
voltage and power flow is either so great or so long lasting that they invoke the actions of short
ah slow processes protection systems and controls which are not modelled in the conventional
transient stability analysis because when we analyze the transient stability problem we do make
certain assumptions but however in long-term and mid-term stability problem, the disturbances
are so severe in nature, so it results into severe upsets in the system, the frequency and voltages
fluctuate over a wide range and this will invoke the action of certain controls which otherwise
were not modeled.
Long-term stability analysis assumes that, inter machine synchronizing power oscillations have
damped out and the result being uniform frequency and here I want to emphasize that whenever
disturbance occur system is in dynamic condition but the movement sufficient time has elapsed,
then the system will settle to a low frequency, low voltage and frequency can be assumed to be
reformed all through the system.

The focus is now on the slower and longer duration phenomena that accompany large scale
upsets and on the resulting large, sustained mismatches between generation and consumption of
active and reactive power.
These phenomenon include boiler dynamics of thermal units, penstock and conduit dynamics of
hydro units, automatic generation control, power plant and transmission system protection and
controls transformer saturation and off nominal frequency effect of loads and the network. This I
will just discuss these thing one by one, when we talked about the transient stability analysis, we
assumed that mechanical power remains constant. We did not model the boiler dynamics, turbine
dynamics for hydro turbines; we did not model the penstock and conduit dynamics. We also do
not model the automatic generation control which is a secondary control to regulate the real
power output of the generator. We have also not modelled the protection system in the
transmission transformer situation and the effect of off nominal frequency effect on loads and
network now here, when the system voltage and frequency deviate widely from the nominal
value, the transformer core may get saturated and it has effect on the system performance.
Similarly, when the voltage and frequency are low all these controls get activated or invoked
similarly, when frequency is quite away from the nominal voltage is away from the nominal it
effect the system load and therefore when we analyze the long-term stability, these phenomenon
need to be modelled and dynamics need to be analyzed considering the detail models of these
devices.
The next point is about the mid-term stability, the mid-term stability is between the transient
stability and long-term stability. Now we can say that mid-term response represents the
transmission between short-term and long-term responses that is it is basically transition,
transition between short-term and long-term responses.
In mid-term stability studies the focus is on synchronizing power oscillations between the
machines including the effect of slower phenomena and possible large voltage and frequency
excursions. As I mentioned that when we talk about the long-term stability, we do not consider
the synchronizing power oscillations while in mid-term we consider the synchronizing power
oscillations
Now severe system upsets result in large excursions of voltage frequency and power flows that
thereby invoke the actions of processes, controls and portions of and portions not modelled in the
transient stability studies.
The typical ranges of time responses, the transient stability is studied for a period of 0 to 10
seconds generally it is over a period of 0 to, 3 to 5 seconds only on those cases where we have a
large system and there are interior modes of low frequency exists in those cases the time period
of oscillations may be large and therefore the transient stability is analyzed approximately in this
range of 0 to 10 seconds. It is not a very a very rigid boundary between a transient stability mid-
term and long-term stability.
Now mid-term stability is from 10 seconds to few minutes while long-term stability is from few
minutes to 10 of minutes. Now with this, I will conclude that today we have, we have addressed
transient stability problem, voltage stability problem and and mid-term and long-term stability
problem. We have seen actually that the transient stability is the ability of the system to maintain
synchronism following large disturbance, it depends upon the system operating condition
severity of the disturbance.
Similarly, the small disturbance stability which we studied earlier that also depends upon the
operating condition of the system and type of disturbance. The voltage stability we have
addressed the basic concepts related to large disturbance voltage stability and small disturbance
voltage stability. We have seen actually the primary criteria for analyzing the voltage stability of
the system while talking about mid-term and long-term stability which are new in the area of
power systems, we have seen actually that whenever there are severe upsets in the system one
has to model the dynamics of slow processes controls and the effect of frequency and voltage
excursions on the system loads with this one is in a position to analyze the behaviour over a long
duration.
Further I would like to emphasize that these 3 stability problems mid-term stability, mid-term
and long-term stability, transient stability and voltage stability they are interrelated the modeling
is similar in nature with this I conclude my todays presentation. Thank you!
Prof. M. L. Kothari
Lecture - 04
Solution of Swing Equation
Method 1
Method 2
Friends we start today, the next topic that is on solution of swing equation. We have
derived swing equation or a synchronous machine we have also derived swing equation
for a multi-machine system. We have seen that the swing equation is function of or is a
non-linear function of the power angles, the there is no formal solution available or
possible because the swing equation is a non-linear differential equation and therefore
numerical techniques have been developed to solve the swing equation. The 2 methods
which is called method 1 and method 2 will discussed today. Before I tell you about the
method 1 and method 2 for solving the swing equation, let us consider a simple case.
A synchronous machine connected through a transformer to a double circuit transmission

line and hence and then connected to an infinite bus, a large system on this transmission
lines we provide circuit breakers at these locations. The purpose of providing circuit
breakers it is location is that as and when any fault occurs on the transmission line by
operating the circuit breaker this faulting line can be taken out.
We will consider a disturbance where due to some unknown reason this line trips, okay.
Now here before before the occurrence of this disturbance these 2 lines are in service.
Now if I draw the power angle characteristic I will denote this as Pe on this axis, we put
delta, okay for this case the power angle characteristic will come out to be a sin curve on
this path okay.
Now in case the system is operating under steady state condition I will represent the
mechanical input line by a straight line divided by Pm then this is our operating, now the
moment a disturbance occurs where out of the 2 lines, 1 lines trips then the post fault
system will have only 1 line in operation and the new power angle characteristic will be
different from the power angle characteristic before the occurrence of fault or we call it
pre-fault disturbance or pre-fault power angle characteristic or we can call it pre-
disturbance power angle characteristic.
Now let us represent the power angle characteristic after the disturbance we will this
characteristic can be represented as Pe1 equal to Pmax sin delta and this characteristic may
be represented as Pe2 equal to, let us say r into P max sin delta okay. Our initial operating
point is here, this denoted by this angle delta naught. Now what we see here is that the
moment fault has occurred or a disturbance has occurred the operating point will shift
from this position a to position b because the moment disturbance occurs right the angle
cannot change instantaneously, angle will remains same.
However, the power angle characteristic becomes Pe2 equal to r Pmax sin delta r is, r is a
quantity or a fraction which is less than 1 and it depends upon what where the values of
the transformer reactance, synchronous machine, transient reactance and transmission
line reactance and so on. Now what we see here is that the moment this disturbance takes
place there is difference in the mechanical input and electrical output, electrical output
determined by this point, this difference become the acceleration power Pe, we call it
accelerating power Pa.

Now this as we will see actually that because of the accelerating power the rotor will
accelerate okay and delta will increase, the increase of delta right as a the function of
time right or we call it actually the delta versus time this plot is our swing curve and they
are interested in plotting the swing curve of the system okay. Our primary objective is to
obtain the solution of the swing equation and that solution is your swing curve. Now what
I do here is that I plot the accelerating power Pa as a function of time.
We start with time t equal to say 0 at time t equal to 0, the accelerating power is equal to
ab and it is a positive as time passes the accelerating power is going to decrease okay.
Therefore, let us represent the variation of accelerating power by this graph I am not
showing the complete one because as far as this problem is concerned in this problem the
accelerating power is going to be positive all through and the resulting system will be
unstable system.
Now what we do is that for is solving the swing equation, we will divide the time into
small time steps. Let us consider 2 consecutive time intervals we will represent the nth
time interval, the let us say this is nth time interval. Now this nth time interval at the
beginning of this interval, let us call it time is t1 at the end of the interval, let time is say
ah t2. Okay then if this is the nth interval right then t1 is equal to n minus 1 delta t and t2
will be equal to n delta t, right this is the notation which we will be adapting. Now in the
step by step solution or we normally call point by point solution because these 2 terms are
used interchangeably in the method 1.
we assume that the accelerating power remains constant during the time interval and
equal to equal to its value at the beginning of the interval that is we calculate the
accelerating power at the time t1 which is the beginning of n minus 1th interval and we
presume at, we that this accelerating power remains constant during this interval it means
the accelerating power remains constant, okay and we shall denote this accelerating
power by the symbol Pa n minus 1 okay, with this introduction we can proceed to discuss
the method 1 for solving the swing equation.
As I stated in the beginning this, this method is called point-by-point solution therefore,
the point-by-point solution of the swing equation consists of two processes which are
carried out alternatively which are those two processes that is what we have to
understand.
The first process is the computation of the angular positions and angular speeds, I am
putting the word angular positions and angular speeds. At the end of each interval from
the knowledge of from the knowledge of angular positions and speeds at the beginning of
the time interval that is the, that is and the accelerating power is assumed for the interval.
Now this is general statement what will be the assumed value of the accelerating power
during that interval right is a subject of concern and we will see actually that the method
1, there is one way of choosing the accelerating power during that interval method 2, we
have another way of choosing the accelerating power. However, the accelerating power
during the time interval remains constant right and therefore the fist step we understand is
that we calculate the value of angular positions, angular speeds at the end of time interval
with the knowledge of angular positions, angular speeds at the beginning of the time
interval okay, this is the first step.
Now once we have obtained the angular positions at the end of the time interval then the
next step we want to second step will be or the second process we call it the second
process is the computation of accelerating power of each machine from the angular
positions of all the machines of the system because when we go from one step to the next
step right, we have to again compute the accelerating power of each machine, okay and
this accelerating power can be computed from the knowledge of network solution.
In general for a multi-machine system, one has to solve the network to find out the
accelerating power of each machine at the beginning of interval or we can say at the
beginning of next interval. Now this uh these 2 processes are carried out alternatively that
you can understand now that suppose I know the angular positions, angular speeds at the
beginning of time interval and I start with the knowledge of accelerating power at the
beginning of the time intervals or accelerating power during that interval, with this
knowledge we compute the angular positions, angular speeds at the end of the interval.
Then using this information we go to second step where we solve the network and we
find out what will be the accelerating power to be used for the next interval right and this
this these 2 steps are carried out alternatively that is I discuss in method 1.
The accelerating power is constant throughout the time interval delta t and has the value
computed for the beginning of the interval that value we have computed at the beginning
of the interval to presume this to remain constant during that interval, okay.
Now here in this diagram, instead of showing the plot of accelerating power versus time
what we are showing here is the acceleration that is d2 delta by dt2 which is equal to
accelerating power Pa by that is M is constant you divide this by M. So that instead of
plotting accelerating power if I plot acceleration versus the number of time steps. Now
instead of plotting here ah time in seconds I plot this as t by delta t. Okay and this plot is
similar to the plot for accelerating power except the units will be different because the
accelerating power has been divided by M.
Now if you see here actually that suppose I compute for time interval n minus 1 then this
is the assumed value of the accelerating power that is I compute the accelerating power at
beginning of the time interval and this will remain constant. Then when I come here, we
again compute what is the acceleration and this acceleration that assumed to remains
constant here like this okay. Please note down actually this graph here what we observe
in this graph. In this method 1, when the accelerating acceleration is decreasing the
assumed acceleration is always more than the actual acceleration and the we will see that
the error will depend upon what is the time step which we choose.

Suppose if I take very small time step then the assumed value of accelerating power will
very close to the actual value of acceleration or accelerating power. We will derive now a
simple algorithm to implement this method one. We start with our swing equation. Now
to illustrate this problem I am considering a simple system where one machine connected
to infinite bus and there is only one swing equation. Now this is our swing equation dt
delta by dt2 equal to Pa by M, okay.
Instead of using this coefficient here H by phi here, we prefer to use this M because it is
easy to what I told the time then you integrate this equation assuming that Pa is constant
accelerating power is constant then this integration will give you d delta by dt equal to
omega, this is the definition of dt delta by d, dt represented as omega this is going to be
equal to omega O plus Pa into t by M. Omega O is the initial value of angular speed.
Now here d delta by dt is not the actual speed of the rotor it is, it is excess speed of the
rotor over synchronous speed. Initially when the system is operating under steady state
condition then what happens Pa will be 0 and d delta d delta by dt is also 0 because there
is no acceleration, there is no difference in the in the rotor actual speed and the
synchronous speed, the rotor rotates at synchronous speed.
Now here with finite value of Pa the accelerating power right d delta by dt is omega and
this is the excess speed over the synchronous speed, now if you further integrate this
equation 4.2 with respect to time, we will get delta equal to delta naught omega O into t
Pa into t square by 2M. Now you have to very clearly understand the meaning of all the
terms which are involved in the these 3 equations 4.1, 4.2 and 4.3. Is there any doubt, this
is state forward okay. Now what we do is we will use these equations to write down the
speed and angular positions at the end of interval in terms of information at the beginning
of the interval.
(Refer Slide Time: 24: 44)

In the equation 4.2, I will substitute the value of the speed at the end of the nth interval.
The speed at the end of nth interval is omega n, the speed of the rotor at the beginning of
the nth interval is omega n minus 1, t delta t is the time, time state and Pa n minus 1 is the
accelerating power at the beginning of the interval. Okay, therefore from the equations
which we have obtained from swing equation by integrating.
Now what we are trying to do is that we are implementing or applying this equations to a
particular time interval. Now if we substitute the values of angular positions at the
beginning and at the end of time interval, we get the equation 4.5, delta n equal to delta n
minus 1,omega n minus 1 into delta t delta t square by 2M Pa n minus 1 okay and these
equations are valid for any value of n, they can start with n equal to 0 that is time t equal
to 0 okay and then we go from n equal to 1, 2, 3 like this, that is from one step to the next
step next step to next step and so on.
From the equation 4.5, we can write down the change in speed during the nth interval as
omega n minus omega n minus 1 as delta omega n which is equal to delta t by M Pa n
minus 1.This can be easily understood because this is the accelerating power accelerating
power by M is the acceleration and we are assuming this acceleration to remain constant
during the time interval therefore deviation in the speed during the time interval is called
delta omega n equal to the assumed value of acceleration multiplied by time period.
Similarly, we can write down the change in angular position during the nth time interval
as omega n minus 1 delta t equal plus delta t square by 2M Pa n minus 1. Now with the
information of the quantity at the beginning of the interval what are the quantity at the
beginning of the interval omega n minus 1, delta n minus 1 and Pa n minus 1.
We can solve this equation 4.6 and 4.7 okay and obtain this swing curve okay what we
have to we have to do is that we change, this angular change during the interval and we
can find out the actual angular position by adding this change to the angular position at
the n minus 1th at the beginning of n minus 1th interval okay and you can find out the
swing curve. Now when I said that there are 2 steps involved, this is the first step, the
second step will be that you obtain the value of this angle delta and substitute in the
power angle characteristic applicable during that transient period to find out the new
value of accelerating power that is your second step.
Once you obtain a new value of accelerating power we will substitute here in this
equation and obtain the new values of angle deviation and speed deviation. Now do we
require the information about speed deviation as function of time for assessing the
stability of the system, the answer is no. If I know the plot of delta as function of time
then by examining the swing curve I can say whether system is stable or not as we have
seen in last time and therefore if I am interested only in plotting the swing curve then
what we do is that we eliminate this speed term and obtain an expression which is
independent of or which does not contain speed term.
Now this can be easily done easily done, here equation 4.8 is same as equation 4.5, I will
just written for the safe of convince. Now what we do is that we write a similar equation
for the preceding time interval that is instead of n we put it n minus 1, n minus instead of
n minus 1 you put n minus 2.
So that we can write down the equation of this form delta n minus 1 equal to delta n
minus 2 delta t omega n minus 2 delta t square by 2 M Pa n minus 2. This equation is for
nth interval this equation is for n minus 1th interval, okay our next step will be that you
subtract these 2 equations, you will get delta n minus delta n minus 1 equal to delta n
minus 1 minus delta n minus 2 plus delta t into omega n minus 1 minus omega n minus 2
plus delta t square by 2 M Pa n minus 1 minus Pa n minus 2.
Now what we do is that we make some substitutions, we denote the speed change I am
sorry, correction we denote the angle change or angle deviation during nth interval as
delta delta n. Similarly, for n minus 1th interval delta delta n minus 1 and so on now if we
make these substitutions here in this equation we can write down expression in the form
delta delta n equal to delta delta n minus 1.
Now we had the term delta t into omega n minus 1 minus omega n minus 2, now this
term is nothing but delta omega n minus 1 this is the speed deviation in n minus 1th
interval and using this equation 4.4, we can write down that the speed deviation during
the n minus 1th interval can be written as or nth interval can be written as delta t by M Pa
n minus 1 that is equation 4.4 okay. Now you make that substitution here.
So that now what we find that in this equation we do not have speed term speed has been
eliminated okay. This equation these 2 terms can be combined and you get the resulting
expression for delta delta n equal to delta delta n minus 1e plus delta t square by 2 M Pa n
minus 1 plus Pa n minus 2. Now this is the algorithm for obtaining the swing curve or a
machine connected to infinite machine.
Similar equations can be derived if you have a multi machine system. A process goes like
this when I want to find out the change in total angle position during nth interval I know
what was the change in angle rotor angle position during the n minus 1th interval but that
was that computation has already been completed, we know this information, we also
know the uh assumed value of accelerating power at the beginning of n minus 1th
interval and at the beginning of nth interval, this is the accelerating power at the
beginning of nth interval.
The accelerating power at the beginning of n minus 1th interval is also known because
we have already done the calculation for that interval and therefore this algorithm goes in
a iterative manner you know the complete information on the which is required to
compute the expression on the right hand side of this equation, you get the value of delta
n.
Then the moment you get the value of delta delta n, you obtain the new value of delta n
as delta n minus 1plus delta right and you continue to do it because as I told you that the
process has 2 steps, the step one is to compute the new angular positions and new angular
speeds. However in the method one if you are interested only in swing equation then we
do not require the information about the angular speeds right.
Then once we get this new value of angle we refer back to the power angle characteristic
compute the electrical power output, mechanical power input is assumed to be constant
we compute what is acceleration power and use that accelerating power for the next
interval okay and therefore this process is to continue alternatively.
We have discussed actually this technique now how good this technique is in giving you
the correct solution and as we will see actually they are numerical techniques are not the
exact solutions they will give you always approximate solutions because we make some
assumptions. Only this is that you would like to get the solution which is very close to the
exact solution and to evaluate the method 1 step by step method 1 there we assume swing
equation of this form.
Let us presume that swing equation is given by this equation, now this is the swing
equation which we want to solve. Let me know whether this swing equation is a linear
differential equation or non-linear differential equation. This is a linear differential
equation, I have chosen the linear differential equation deliberately to illustrate the we
will say illustrate the effect of, effect of time step on the accuracy of the solution.
Now so far this swing equation is concern the initial conditions are given as delta naught
equal to phi by 4 and the initial speed is 0 d delta by dt initial is 0, with these initial
condition the formal solution of the swing equation is delta equal to phi by 2 minus phi
by 4 cosine square root of 2 f by H, t. This you can obtain yourself please do the it is an
exercise find out the formal solution of this linear second order equation. Now to
compare the accuracy of step by step method one we solve this equation by step by step
method because step by step method which is applicable are suitable for solving non-
linear equation can also be applied for solving a linear differential equation, for solving
this H is equal to 2.7 mega joules per MVA is assumed.

This graph shows ah swing curve obtained using the formal solution and by using this
method 1, this firm curve shows this graph 1, a firm curve it shows the solution obtained
using the formal solution. The graphs which are shown here in this diagram are obtained
for different values of time step delta T equal to 0.15 second. Then next graph is for delta
T 0.1 second then the third graph is for 0.05 second and the last one which is obtained
using the method 1 point by point solution is 0.0167 delta T equal to 0.0167 is one third
of the delta T equal to .05, originally a delta T equal to 0.05 second is chosen and in order
to understand how much improvement one gets in the solution delta T is reduce by a
factor of 3 and graph is obtained for delta T equal to 0.0 second.
This axis shows the time in seconds and y axis shows delta in electrical degrees, it can be
very clearly seen that as the time step increases the accuracy of the solution of the swing
equation decreases or deteriorates and hence one concludes that for getting solution
closer to the accurate solution one has to use very short time step. However, if we use a
short time step it will require more time for computation and this is the major drawback
of method 1.
In order to overcome this shortcoming of this method 1, a method 2 has been developed
and this method two is different from method 1 in terms of the accelerating power which
is assumed to remain constant during the time step. Now to understand this method 2 let
us look at this graph which shows the acceleration versus t by delta T, now here on this x
axis we have put t by delta T.
So that it shows the u the time step count, now this graph shows the variation of
acceleration with t by delta T or with t, now in this method two instead of assuming the
accelerating power remaining constant at the value equal to the 1 computed at the
beginning of the time interval.
We compute the value of the accelerating power or acceleration at the beginning of the
time interval for example, say time interval starts at n minus one and ends at n that is the
nth time interval. Now at this time step n minus 1, we compute the acceleration alpha n
minus 1 and assume that this acceleration remains constant from half the preceding
interval to the next half interval that we if we see in this diagram the accelerating power
computed at the beginning of nth interval remains constant from n minus 3 by 2 to n
minus half it can be very clearly seen actually that by assuming this accelerating power in
this fashion, the assumed value of the accelerating power is equal to the average value
over the time step.
Now we can write down with this assumption the change in speed during the time step n
minus half this can be written as delta t into alpha n minus 1 where alpha n minus 1 is the
acceleration computed at at step n minus 1. Now alpha n minus 1can be replaced by this
Pa n minus 1 by M. So that we can say that speed deviation delta omega n minus 1 by 2 is
equal to delta t Pa n minus 1 by M.
Now we can write down the rotor speed rotor speed at the end of n minus 1 by 2 interval
as the speed at the beginning of this interval that is omega n minus 3 by 2 plus delta
omega n minus 1 by 2. This quantity is the change in speed during the time step delta t.
Further, since the acceleration assumed to remain constant during this time step speed is
obviously going to vary during this time interval.
However, the total change in the time and total change in the speed will be accounted by
considering a step change in the speed that is the total change in the speed which is equal
to delta omega n minus half which is equal to delta t into alpha n minus 1, this changes in
the speed is assume to take place in a step manner at the instant at which we compute the
accelerations of the rotor, with this assumption with this assumption the speed remains
constant during the nth interval and it is its value is equal to omega n minus half, with
this speed, with this speed during this interval the change in rotor angle is equal to delta t
into omega n minus half.

Since omega n ,omega n minus half is constant during this interval therefore change in
angular position is obtained as the constant speed a omega n minus half into delta t
therefore, angular position of the rotor at the end of nth interval is equal to angular
position at the beginning of the nth interval plus change during the nth interval therefore,
we can write down delta n equal to delta n minus 1plus delta delta n.
Now when we solve this equations four point 1, 2 to 4.14 we can get the swing curve as
well as we can get the speed of the rotor as a function of time. However if we want to
obtain swing curve only we may use a formula for delta delta n from which omega has
been eliminated from the equations 4.12, 4.13 and 4.14.
We get an expression for delta delta n as delta t omega n minus 3 by 2 plus delta t square
by M Pa n minus 1. By analogy with equation 4.14, one can write down the change in
angular position during the n minus 1th interval as delta delta n minus 1 equal delta t into
omega n minus 3 by 2. So, that we can substitute the value of delta delta n minus 1 that is
the change in the rotor angular position in the equation 4.16 using equation 4.17. We get
the expression for delta delta n equal to delta delta n minus 1 plus delta t square by M Pa
n minus 1.
This is a equation which can be used for computing the swing curve the left hand side
term delta delta n, shows the change in rotor position rate rotor angular position during
the nth interval, this is expressed in terms of the change in angular position during n
minus 1th interval plus delta t square by M Pa n minus 1 and this can be easily
programmed and using this expression the problem which was solved using method 1 is
again solved here.
Now this graph shows the swing curve obtained using formal solution of the second order
differential equation and using method 2 considering the several values of time step. This
graph shows the solution obtained using the this graph shows the swing curve obtained
from the formal solution, you can say formal solution. Now I say formal solution means
it is the closed form solution of the second order differential equation. Now here this
swing curve is for delta t equal to 0.25 second, well this swing curve is for delta T equal
to 0.05 second.
Now if you, if you carry if we carefully examine this swing curves we can notice that,
that the amplitude of the swing curve or the maximum deviation of delta obtained using
the formal solution and that obtained using method 2, have hardly any difference, only
difference which we observe is the time, time period of the solution obtained within the
formal solution and that obtained using the method 2 have some difference and therefore,
I can conclude here that the method 2 provides the accurate swing curve and we can use
reasonable value of time step as we see that for time step equal to delta T equal to 0.05,
the solution obtained by method 2 that is the step by step method 2 is very close to the
formal solution.
Now we consider how do we account for the discontinuities in the accelerating power,
discontinuities in the accelerating powers occurs at the occurrence of disturbance or at the
instance of switching. In case the discontinuities at the beginning of the time interval that
is the first time interval then this can be computed delta delta 1 can be computed by using
this expression delta t square by M, Pa O plus by 2, where Pa O is the accelerating power
just after the disturbance and by putting 2 we are taking the average value of the
accelerating power half the accelerating power.
In case the discontinuity occurs at the beginning of mth interval then this can be the
accelerating power during the m th interval will be computed by this formula Pa m minus
1 minus plus Pa m minus 1 plus by 2, this minus indicates the accelerating power just
before the occurrence of disturbance and Pa m minus 1 plus indicates the plus sign
indicates just after the occurrence of incidence.
With this, we conclude the computation of the swing curve using 2 different methods
method one and method 2, method 2 is more accurate and this can be used. Thank you!
Prof. M. L. Kothari
Lecture - 05
The Equal Area Criterion for Stability
Friends today, we shall discuss ah the equal area criteria for stability.

We have studied that to analyze the stability of a power system, we have to solve swing
equations and solution of swing equations is through a numerical technique and it is a
time consuming process. The equal area criteria of stability is very powerful tool to
understand the basic concepts of stability. However, as we will see that this criteria has
its limited applications.
Now, today in our study we will establish this basic concepts pertaining to equal area
criteria and we will analyze considering a one machine swinging with respect to the
infinite bus, then we will develop the equivalent of one machine infinite bus system of a
2 machine system.
Then we will study the applicability of equal area criteria under what circumstances the
provoke systems this criteria is applicable and what are its limitations. Then we will
illustrate the applications of this criteria considering two examples, a sustained line fault
and second a line fault with subsequent clearing.
The equal area criteria of stability is a graphical method for determining whether the
system is in stable or not, that is in any stability studies our primary requirement is that
for given operating condition and for a given disturbance whether the system is stable or
not, many times we are interested in knowing if it is stable how much stability is and
what is the stability margin.
This information we can get by plotting the swing curve but as I told you that the
computation of swing curve is time consuming and for simple system for a two machine
system or a one machine connected to infinite bus, we can obtain this information by
applying a graphical method and that method is the equal area criteria.
Now when this criteria is applicable, its use wholly or partially eliminates the need of
computing swing curve and thus saves considerable amount of time. I am emphasizing
here, that it eliminates the computation of swing curve wholly or partially. Okay that we
will see actually when we attempt some example.
This criteria is applicable to any two machine system for which commonly made
assumptions are applicable right. We have studied the commonly made assumptions for
analyzing the transient stability problem and when these assumptions are applicable, this
can be applied to any two machine system.
Now first we will mathematically establish the equal area criteria of stability to establish
we start with the swing equation of a machine, here we are considering a machine
connected to the infinite bus, infinite bus is one which can be represented by a constant
voltage source its internal impedance is 0 and its inertia is infinite. Now you multiply this
both sides of this equation by a term 2 delta d delta by dt.
When we multiply both the sides by this term 2 times d delta by dt, we get equation in
this form 2 times d square delta by dt square into d delta by dt equal to 2 times Pa by M, d
delta by dt. Now this left hand side of this equation can be identified as derivative of d
delta by dt square, you look it very carefully that is if you take this term and find out its
derivative you will get the term 2 times d square delta by dt square d delta by dt, okay
and right hand side we are writing as it is 2 times Pa by M d delta by dt okay.
The next step is you multiply both sides of this equation by dt and we are getting
differentials instead of derivatives that is when you multiply both sides of this equation
5.3 by dt it becomes a differential that is d of d delta by dt square equal to 2 times Pa by
M, d delta. Okay, which is written here actually as d of d delta by dt square 2 times Pa by
M d delta okay. Now you integrate this equation 5.4, integrate it.

When you integrate, you will get d delta by dt square equal to 2 by M integral Pa d delta.
This integration we are doing over certain range of delta that is start with initial value of
delta naught delta equal to delta naught and go to some value of delta. I am just putting it
general that is delta naught to delta, I am not specifying what should be the value of delta
here.
Now from this equation 5.5, we can write d delta by dt equal to square root of 2 by M
integral Pa d delta, the limits are implied okay. Now we look at this equation and we
know that the condition which indicates the stability of the system is that d delta by dt
should be 0 that is a system starts where it is perturbed right with the initial value delta
naught and when it is, when delta increases it will attain a maximum value then start
decreasing that is the condition of stability, it means that when it goes to maximum and
start decreasing that maximum point, at that maximum point d delta by dt is 0 therefore,
the condition to indicate the stability is that d delta dt should become 0.
Now for this d delta by dt2 become 0 which is the condition to indicate the stability of the
system, our requirement is that the integral of Pa d delta from a initial value of delta equal
to delta naught to some value of delta m should be 0 that is if I, if I plot the area under the
curve Pa that is accelerating power starting from initial value of delta equal to delta
naught to some maximum value of delta m and in case the in the area under this curve is
0 right, then at the value of delta equal to delta m the d delta by dt will become 0 right.
Now this I will show here in this diagram, this is the power angle characteristic of the
system, this was the mechanical input line. Now in this diagram I am simply showing that
suppose initially the system was operating at delta naught and the power angle
characteristic applicable for the initial operating condition was not this but something
different. The moment fault occurs on the system the operating point has shifted from this
point to this point okay then delta will increase from initial operating from the point a on
the power angle curve Pe.
The initial acceleration is given by this or initial accelerating power is given by this line,
this is the initial accelerating power, okay therefore rotor accelerates when rotor
accelerates the delta will increase, speed increases then delta increases and now when it
reaches the point b right, the accelerating power becomes 0. Now at this point what is the
rotor speed?
The rotor speed is more than the synchronous speed because, because from when it is
travels from this point a to b, the rotor is accelerated gain some kinetic energy it has
gained some extra speed over and above the synchronous speed therefore, at this point
the rotor will not stop it will continue to, the angle will continue to increase. Now the
moment it crosses this point b we see that the accelerating power becomes negative that
electrical power output becomes more than the mechanical input.
Therefore, now the rotor will be subjected to retardation and when it reaches the point c
point c, if suppose whatsoever the kinetic energy it has gained if it is return back or is ah
lost then at point c the rotor will again attain a speed equal to synchronous speed and
therefore for the system to be stable, the rotor will swing from delta naught to delta m and
the area under the accelerating power.
Now here the accelerating power is obtained by separating therefore here I can say that
accelerating power curve is given by this difference, okay and therefore this area should
be 0. Now for this area under the accelerating power curve to become 0 requires that the
this this positive area, we call this as a positive area because accelerating power is
positive this is called negative area, the accelerating power is negative therefore these 2
areas should be equal and this is what is known as the equal area criteria of stability. Now
this equal area criteria of stability can also be interpreted in terms of the kinetic energy
gain and kinetic energy lost right.

Now we know actually that suppose a rotor is subjected to a torque T, torque T and when
it swings from delta naught to delta the, the work done on the rotor is equal to or work
done by the rotating body is equal to T into d delta. This is actually when the delta is
small and you integrate this expression from delta naught to sum value delta then this
becomes the work done when the rotating body accelerates, okay or decelerates it
depends upon the situation and hence we can say that the area a1 represents kinetic
energy gained by this machine or I can say this area a1 is directly proportional to kinetic
energy gained there has to be some proportionality constant I cannot say that a1 is in in
joules or mega joules or so it will depend upon what is the unit which we have attached to
it.
Now the derivation is which we have seen just now is considering a machine connected
to infinite bus. Now suppose you have 2 finite machines, 2 finite machines a problem that
of a synchronous generator supplying power to a synchronous motor through a
transmission line. Then this 2 finite machine system can be replaced by an equivalent
machine infinite bus system, this derivation is very straight forward we take the swing
equation of machine 1, d square delta 1 by dt square equal to Pa1 by M1 that is Pm1 minus
P1 by M1. Then we take a swing equation of second machine ah d2 delta 2 by dt equal to
Pa2 by M2 that is Pa2 is Pm2 minus P21 okay.
Now suppose we define here the relative angle between the two machines as delta equal
to delta 1 minus delta 2. Now whether these 2 machines are going to remain in stable
condition or become unstable is going to determine not by delta 1 and delta 2 individually
but the difference between this 2 angles delta 1 and delta 2, in case this difference
difference right remains within certain limits system will continue to be stable and
therefore the this 2 differential equations which we have written right, if we can be
written in this form that is you can subtract these 2 equations you will find that d square
delta 1 delta d square minus d square delta 2 by dt square. Okay this difference can be
written as d square delta by dt square because delta 1 minus delta 2 is delta and on this
right hand side you have Pa1 by M1 minus Pa2 by M2.
We can multiply both sides of the equation by this product M1 M2 or M1 plus M2 and we
may get the expression in this form. M1 M2 or M1 plus M2 d square delta by dt square
equal to M2 Pa1 minus M1 Pa2 or M1plus M2 .Okay now if we replace Pa1 and Pa2 by Pm1
minus P1 and Pm2 minus P2, we get this expression that is I have put here in this
expression wherever you had Pa1 Pm1 minus P1, Pa2, Pm2minus P2 and then simplify it in
this we find here that we can write this as M1 M2 or M1 plus M2 d square delta by dt
square equal to M2 Pm1 minus M1 Pm2 divided by M1plus M 2,that is in this term we have
the inertia constants and mechanical powers while here in this term we have inertia
constants and electrical powers P1 minus P1 and P2 okay.
Now this equation may be considered to be the swing equation of a machine connected to
infinite bus, where where the equivalent inertia constant is M1M 2 or M 1 plus M2, the
equivalent mechanical input is given by this expression and equivalent electrical output is
given by this expression.

Therefore, we can say that this expression written as M times d2 delta by dt2 equal to Pa
Pm minus Pe, where Pm is given by this expression, Pe is given by this expression, we can
say that Pm is the equivalent mechanical input Pe is the equivalent electrical output and
the equivalent inertia constant is M1 M2 over M1 plus M2. Now at this point you can
understand that the equivalent inertia constant is as if we are connecting 2 resistances in
parallel to find out the equivalent resistance.
Suppose you have 2 resistances r1 and r2 and you put them in parallel the equivalent
resistance is r1, r2 or r 1 plus r2 therefore this is the equivalent inertia constant is primarily
determined by or is going to be closer to the smaller one that is suppose I put M1 M2 M
plus M2 values, if M2 is large as compared to M1. Okay then we will find out the value let
us say I will just take example take just values actually let us say M1 is 5 and M2 is say 50
what will be the resultant? It is going to be less than 5 but it is closer to 5 not closer to a
50 therefore, the equivalent machine will have the inertia which is less than, less than the
the inertia of a smaller machine or inertia of a machine which has a smaller inertia
constant.
Now here at this stage, to illustrate the application of equal area criteria for analyzing
stability of a system, I will consider an example, the example we consider is a simple
example. Let us consider this example, we have a synchronous generator connected to a
double circuit transmission line to an infinite bus, we put infinity here to show that it is
infinite bus its inertia constant is infinite. To illustrate the application of equal area
criteria of stability what we will consider is that yes there is a machine infinite bus
system. We will consider a fault at the middle of the transmission line, one of the
transmission lines at the point P.
We will consider a three phase fault, a balance 3 phase fault okay. I will consider the
unsymmetrical faults in my next lecture because in any system we do come across
symmetrical, unsymmetrical faults and unsymmetrical faults are more frequent in
occurrence okay. Now let us assume some parameters okay, let us say Xd prime for this
system is 0.2, reactance of this transformer is say 0.1 that is impedance is j times 0.1.
The transmission line is loss free and its reactance is j times 0.4 and further we assume
that the infinite bus voltage V is 1.0 and we consider this infinite bus voltage as
reference, so that V 1.0 angle 0this is the terminal of the synchronous generator. We
denote the terminal voltage by the symbol Vt and let us assume that Vt magnitude is
given to be equal to 1.0 per unit and also it is assumed actually that this machine delivers
electrical power Pe equal to 1.0 per unit okay.
Now for this system, for this system we will obtain the power angle characteristic for 3
conditions, one is pre-fault operating condition, second is during fault or fault on
operating condition. Now once the fault is there in the system the fault will be cleared by
operating these circuit breakers at the two ends of the transmission line and therefore the
third operate, third power angle characteristic which we will obtain is post-fault power
angle characteristic, although we are considering a simple system but this exercise is
required to be performed even for a multi machine system.
We have to obtain the expressions for power outputs of machines under pre-fault
condition, during fault condition and post-fault condition and the approach we are
following here will also be applicable for a classical multi machine stability problem.
I am using about classical means here, we will be making those basic assumptions that is
synchronous generator can be represented by a constant voltage behind direct axis
transient reactance.
Now with this information given how to find out the pre-fault power angle characteristic
Therefore, here our primary requirement is that to get the pre-fault power angle
characteristic what do we need is the voltage behind transient reactance which is not
given, what is given is the voltage at the terminal of the machine. Now we can use this
information to compute, to compute ah pre-fault power angle what is first step is you
draw from the one line diagram a reactance diagram.
The reactance diagram can be drawn as a voltage source, we will denote this voltage
source as E prime, reactance Xd prime the value of this impedance is given to be is given
as j times 0.1, the transformer reactance is .1 per unit that its impedance is I am sorry, the
generator reactance is 0.2 not point .2 transformer is 0.1. These 2 transmission lines and
infinite bus voltage, we denote this as V at this point which is the terminal of the
synchronous generator the voltage is Vt and this magnitude of this voltage Vt is 1.0 while
this magnitude of V is 1.0 and delta 1 is taken as our delta is 0 the delta 1 is 0, okay. I can
call it delta, now the with this information or with this equivalent circuit what do we do is
we find out what is the phase angle of this Vt with respect to infinite bus voltage.
This can be obtained simply by using the relationship that the power output P can be
written as Vt magnitude into V divided by the reactance connecting the terminal voltage
or terminal of the generator to the infinite bus and let us say that the, the phase angle
between Vt and infinite bus voltage is alpha. Now we substitute the values of power
because it is we are supplying 1 per unit power Vt magnitude is given as 1 this is one and
the reactance between the infinite bus voltage and the generator terminal is how much
0.3 sin delta because here we are substituting only the reactance value there is it is not
impedance it is the magnitude okay.
Therefore with this, you can calculate the value of I am sorry there is a mistake here this
we will call alpha comes out to be equal to you calculate it and alpha in this case comes
out to be equal to 17. 458 degrees okay. Therefore I can say here that the terminal voltage
Vt is equal 1.0 angle 17 .458 per unit. Now the using this information we find out the
current supplied by the generator or I can be written as Vt minus V they are all phases
divided by the total reactance between the 2 machine or impedance, the impedance here
is how much j times 0.3.
Okay now if you substitute the value of Vt and V, the current comes out to be equal to
1.012 angle 8.729 degrees, okay. Now once you get the current I supplied by the
generator we can find out now the internal voltage E prime that is E prime can be written
as Vt plus j times Xd into I okay and you substitute the value of Vt Xd and I the calculated
value of E prime comes out to be equal to that is magnitude of E prime because at this
stage I am not interested in phase angle right.
Now E prime comes out to be equal to 1.05 per unit and the pre-fault power angle
characteristic Pe1 becomes now 1.05 into 1divided by total reactance between infinite bus
voltage and internal voltage that is coming out to be how much .5 sin delta here now
because when you are talking in terms of the relations with the power considering the
terminal voltage of the machine, we have written alpha.

Now delta is the angle and therefore the pre-fault power angle characteristic is now 2.10
sin, okay this these steps are extremely important. We will follow we may have to follow
the similar steps for a multi machine system also. The next step is we want to find out the
power angle curve or power angle characteristic when fault is on.
When the fault is on, on the system we can write down or we can again draw the
reactance diagram, since the I have consider the fault at the middle of the line therefore I
will divide this line reactance into 2 parts and show as .2 .2 on both the sides of the
faulted point or voltage at the point P.
Now since, we have considered a balanced three phase fault this P will be connected to
the reference bus directly connected there is no impedance involved however when we
consider the unsymmetrical fault we will see that to analyze the or to obtain the power
angle characteristic during fault conditions there will be some impedance connected
between the faulted point P and reference bus. This impedance will depend upon the type
of fault but for a 3 phase fault the impedance to be connected is 0.
Now here we are considering a three phase metallic fault okay there is no fault
impedance. Now in this the total reactance of the this these two components that is the
Xd prime and the transformer can be combined and this can be written as impedance is
0.3, this is j times 0.4, this is j times 0.2, this is j times 0.2 and this voltage is this voltage
magnitude is this is V voltage, this is E prime okay. Now what we do is we will try to
simply this network, so that we can find out the power angle characteristic during fault on
period.

This diagram can be redrawn as, now we will denote this node as 1 denote this node as 2,
okay this node as 0 and this as now here, if you examine this network then these
3reactances which are connected in star, j times between 1 and 3, 3 and 2 and 3 and 0.
Okay this star can be replaced by an equivalent delta, the equivalent using the standard
star delta transformation technique this exercise I have obtained and I find the equivalent
network after this star delta transformation as, E prime this reactance comes out to be
when you calculate actually comes out to be or this impedance comes out to be j times
0.65, this comes out to be j times 0. I am sorry it is j times 1.3, this is j times 0.266 and
what is this value? j times 0.2 and this is your V and this is E prime.
Now these terminals are this is your 1, this terminal is 2 and the node 3 has been
eliminated by star delta transformation this is our reference. Now the power angle
characteristic will depend upon the reactance connecting these 2nodes 1 and 2, one is the
node of at which E prime is connected two is the node at which the infinite bus voltage is
connected this characteristic is not going to be affected by the shunt branches because
whatsoever is the, you now current which is flowing through the certain branches, okay is
not going to affect what is the power which is going to be transferred and therefore the
the power angle characteristic, the power angle characteristic during fault condition can
be obtained as now I will call it as Pe2 equal to 1.05 which is the E prime into 1 divided
by the reactance connecting node 1 and 2 and that is 1.3 sin delta and this comes out to be
there.
We computed and its value comes out to be 0.808 sin delta or what you observe here is
that this Pe2 which is the equal to Pe max into sin delta right, the the maximum value of this
power angle characteristic will depend primarily upon what is the reactance connecting
the internal voltage of the synchronous machine to the infinite bus voltage more is this
reactance less will be the voltage okay.
Now we can very easily obtain the post fault power angle characteristic under the post
fault condition one of the transmission line is cleared or it is removed and therefore the
reactance connecting the two nodes, internal voltage of the synchronous generator and
infiltrate bus voltage that comes out to be how much .7 therefore our characteristic is now
this divided by .7 sin delta. Okay now these are the 3 important power angle
characteristics which one has to compute or obtain for analyzing the stability of a
machine infinite bus system either we apply equal criteria or you directly solve the some
equation, it is the a material. The these 3 characteristics if you draw it can be shown like
this.
I will call this is a Pe1 this is pre-fault power angle characteristic in this particular case
this is 2.1, P maximum value under faulted condition or the power angle curve under
fault condition was found to have the maximum value equal to around .808 and therefore
the characteristic can be shown to be like this and this is 0.808 after the fault is cleared,
the amplitude of this power angle characteristic how much what is the value of this1.5,
1.5 that is very good and therefore the third characteristic can be plotted here, then this is
1.5 this is your Pe3, this is your Pe Pe2.
In some of our representations what we will do is that the Pe1 is written as Pmax sin delta
Pe2 will be written as r1 times Pmax sin delta, okay and Pe3 will be written as r2 times Pmax
sin delta. Okay now this r1 and r2 these are the multiplying factors constants they will
always be less than 1 and now I shall take these 2 examples to illustrate how we apply the
equal area criterion of stability.
Now let us consider the first case where, we assume that there is sustain fault, fault is not
cleared. Okay now in this particular case the diagram shows that this is the output on
normal conditions or can say pre-fault power angle curve this is the output when fault is
on this is the during the fault power angle characteristic, the mechanical input line is
shown here Pm the initial operating angle is operating angle is delta naught where the pre-
fault power angle characteristic intersect.
Now the movement fault occurs the operating points shifts from A to B then it moves
from B on this during the fault power angle curve and as you have seen that when its
comes to this point C right the accelerating power becomes 0 and it swings beyond this
point and its swings to the point up to say D, this is corresponding to angle delta M if
these two areas become equal the rotor will swing up to an angle equal to delta M and
then from this point it will start returning back why it starts returning back because it is
now subjected to retardation electrical power is more than the and when it comes back to
this A what will happen will it stop here, no it will continue to move and ultimately it is
going to settle to this point C because because system has some damping which we have
not considered while writing the swing equation.
Now here depending upon actually the power angle curve, in case actually the height of
this power angle curve is less you will find actually that you may have to swing to to
larger angle to make this 2 areas equal. Now maximum swing up to which rotor is going
to be subject to retardation is where where the power angle characteristic intersects with
the mechanical input line at this point. In case suppose the energy or kinetic energy
gained is not returned back when it comes to this point then suppose the movement it
crosses this point, you will find that the machine is again subjected to acceleration and
therefore rotor is going to lose synchronism.
Now this is shown in this diagram, where the situation is slightly shown to be different
where even when the rotor has come to the point e these 2areas are not equal and
system may lose synchronism. Now the third case and the last case which I am discussing
here is that we have a situation where we, we switch off the faulted line or fault is
cleared.
Now you start looking at this there are 3 power angle curves this is the mechanical input
line initially we are operating at this point A, the movement fault occurs you shift to the
point B on the during the fault power angle curve. Now when you are moving on this
curve at this point C at on the angle equal to delta C the fault is cleared. We call this as a
fault clearing angle, the angle at which the fault is cleared then you will shift from this
point now to the post fault power angle curve then that is this is the, this curve is the post
fault output again as you know that this is this this is the area a1 that is this is bound by
these 2 angles and this power angle characteristic and mechanical input line this becomes
the accelerating area from e it will continue to swing it come to the point f and the
maximum angle becomes delta M at this point we find actually that these 2 areas equal
and therefore the maximum swing is up to delta M and the system will return back.
Now a new new stable operating point is now where the post fault power angle
characteristic intersects the mechanical input line that is in this diagram this is the new
stable operating point therefore, the rotor is going to swing around this point okay
oscillate around this point and because it has some damping, it will settle to this new
condition. Now with this I will just summarize what we have discussed today.
We have established the equal area criterion of stability. We have also obtained for a two
finite machine system an equivalent a machine infinite bus system. We have also
obtained for a given particular system the pre-fault, during fault and post fault power
angle characteristic a simple method is you can say discussed here and at the end I have
considered the 2 cases, one is considering a sustained fault and another is the fault cleared
after small amount of time, small time. Okay I conclude my presentation here and thank
you very much.
Power System Dynamic
Prof. M. L. Kothari
Lecture - 06
The Equal Area Criterion for Stability (Contd)
Friends, we shall continue our discussion on equal area criteria of stability.
Today, we will cover I will address the following problems. The basic definitions of a transient
stability limit, the meaning of critical clearing angle and critical clearing time then we shall study
the effect of fault clearing time on transient stability limit, next we shall study the effect of type
of fault on stability when we talk about the type of fault on stability we, I shall introduce the
concept of fault shunts then we will study the effect of grounding on stability.
Now, before I address these issues I would like to mention that this subject has been very well
addressed in power system stability written by Edward Wilson Kimbark there are 3 volumes,
volume 1, volume 2 and third volume is now known as synchronous machines. These basic
concepts are very well addressed in volume 1.

Now, let us define what the transient stability limit. We have defined earlier the transient
stability, now today we will emphasize on this word transient stability limit. The transient
stability limit refers to the amount of power that can be transmitted through some point in the
system with stability when the system is subjected to severe aperiodic disturbance that here the
stability limit is referred in terms of the amount of power that is what is the power in mega watts
that can be transmitted and this is referred to some point in the system.
Now the emphasis here is on some point in the system. Now if you take a machine infinite bus
system then there is only one transmission line and therefore whatsoever power that can be
transmitted on that line with stability becomes the transient stability limit. However when I
consider a multi machine system okay, in that system this limit refers to a point in the power
system that is there may be number of transmission lines and you may refer that in this particular
transmission line how much power can be transmitted for a given operating condition with given
type of severe disturbance, given type of severe disturbance here.
Now here we emphasize the word aperiodic the disturbance is disturbance comes not in
periodical manner but it comes in aperiodic manner it comes and goes it is not actually that after
every 5 second the disturbance keeps on coming. Okay, this is what is the meaning of transient
stability limit therefore, any power system when we operate, we have to operate below transient
stability limit.
Okay so that it will withstand the the particular type of disturbance for which the system is
designed. Now next term is the critical clearing angle. This critical clearing angle is specifically
referred to a machine infinite bus system because in the multi-machine system, you will have
number of angles right and therefore the definition to a multi-machine system is not that easily
available.
Now for a given system and for a given initial load, there is a critical clearing angle, if the actual
clearing angle is smaller than the critical value a system is stable and if larger the system is
unstable. I will explain this point in detail okay in our further discussion that here the meaning is
that there is some critical clearing angle and actual clearing angle, if it is less than this value
system is stable in case the actual clearing angle is more than the critical clearing angle system
will become unstable.

Similarly, we define another term critical fault clearing time again for a given system and for
given initial loading there is a critical fault clearing time, if actual fault clearing time is smaller
than the critical value the system is stable, if larger then the system is unstable. Now this
definition is applicable to machine infinite bus system or a multi-machine system because
whenever a system is operating okay and if fault occurs on a particular element of the system
then this fault is cleared by removing the faulted element by operating the circuit breaker at the
two ends and therefore there is certain time required to clear this fault.
In case this actual fault clearing time is less than the critical clearing time system is stable
otherwise, it is unstable further if suppose the critical clearing time for a system comes out to be
say .2 second and actual fault clearing time is say .1 second then the difference .1 second is
called the stability margin. Okay therefore, we have been discussing in these days in terms of
what is the stability margin and stability margin can be quantified in terms of difference in the
critical clearing time for the system and actual fault clearing time because actual fault clearing
time depends upon the operating time of the protection system and circuit breaker fault
interruption time.
Now let us understand what is the effect of fault clearing time on transient stability limit, now
here when I say it is a transient stability limit it means it is a certain amount of power that can be
transmitted without loss of stability. The transient stability limit depends on type of fault and the
duration of fault, a very we shall state these aspects type of fault and duration of fault. The power
limit can be determined as a function of clearing angle suppose it is a machine infinite bus
system we can find out the transient stability limit or power limit as a function of clearing angle
and clearing time can be found by solving the swing equation up to the time of fault clearing that
if I want to know know the transient stability limit as a function of fault clearing time.
The approach we will discuss here but for a machine infinite bus system the simple approach is
that you first apply the equal area criteria, find out for a given power what is the critical clearing
angle and then once we know the angle, we solve the swing equation up to the up to the fault
clearing angle and corresponding to that we read the value of time and that becomes our critical
clearing time and therefore when I say here, when I discussed earlier that when you apply this
graphical method that is equal area criteria of stability.
We do not completely diverse the need for solving swing equation but partially we do it wholly
or partially this is what we are the they are partially means partly you have to solve it and
suppose a swing curve is required to be solved for say 2 seconds for normal stability steady
analysis but in this particular case the time for which it is to be solved is very small suppose the
for critical fault clearing time comes out to be it is a .2 second okay, then I solve it from 0 to .2
second not from 0 to 2 second and therefore this saves my time or computation time.
As I have stated that fault clearing time is sum of the time that the protective relays take to close
the circuit breaker trip circuit and the time required for the circuit breaker to interrupt the fault
current.
The general conclusion that decrease in fault clearing time improves stability and increases
transient stability limit is just as valid for a multi machine system as for a 2 machine system.
This point I stated earlier also again we reiterate that general conclusion is the decrease in fault
clearing time improves stability and it improves the transient stability limit okay.
Now this conclusion is valid for machine infinite bus system 2 machine system and even for a
multi machine system and that is why the efforts have been made all through to reduce the fault
clearing time this has been possible by applying fast acting protective relays and fast acting
circuit breakers.
Now to illustrate this that how do we calculate the transient stability limit and obtain a curve
relating the transient stability limit and the fault clearing time, we will consider these 2 machine
system and a generator infinite bus and consider that the fault occurred at the middle of the line.
Now for this system we can find out the 3 power angle curves, pre fault, during the fault and post
fault, once we know this power angle curves we can apply the equal area criteria and determine
the certain points on the transient stability limit versus the fault clearing time.
Now we will consider the 2 extreme cases, first case we will consider that the fault is
instantaneously clear. Suppose fault occurs in the system and the time which it takes to clear the
fault is instantaneous it does not have any time practically it does not happen okay. Now this can
also be considered similar to that one transmission line is stripped, okay by operating the circuit
breakers.
Now in that case if you want to find out what is the transient stability limit then the approach
would be you make use of these two power angle characteristic, pre fault output characteristic
and the post fault output where during fault we do not require it because the system has not
operated with fault on the system.
Now in this case what we will do is, this mechanical input line, the mechanical input line Pm is
moved up and down, Pm is moved up and down till these two areas are equal that is a1 and a2 is
the area bounded by mechanical input line, the post fault power angle curve that is from delta
naught to delta 1 and a2 is again the area bounded by the post fault power angle curve and the
mechanical input line but they have opposite signs.
Now when these two areas are equal that will give us the stability limit that is this Pm becomes
the stability limit you have, what a what is to be done is that you have to move this mechanical
input line up and down you have to do 1 or 2 you know iterations and the movement these two
areas becomes equal that becomes the stability limit because here the fault has been cleared
instantaneously.
Now we will take another extreme case, where the the sustained fault on the system, that is fault
is not cleared in this situation we require the two power angle curves, one is the pre fault another
is the during the fault power angle curve. Now these 2 curves are plotted here again the approach
will be that you move the mechanical input line up and down till these 2 areas a1a2 are equal that
is when you do this computations you assume some value of Pm.
You know the value of delta naught initial operating angle, you can find out what is the value of
this angle delta 1 that will depend upon the intersection of mechanical input and fault on power
angle curve. Okay similarly, you can find out delta m which is equal to phi minus delta 1, okay
now you find out this area by process of integration and you equate this with the area a2, in case
these 2 areas are equal then this Pm becomes the transient stability limit with sustained fault.
Now the third situation will be that the fault is clear infinite time.
Now under this situation we require all the 3 power angle curves, okay. Now one way is that you
move this mechanical input line Pm up and down and see that these 2 areas are equal and but this
there actually we required the information and what is the fault clearing angle there were 2
parameters involved, one is the fault clearing angle another is the mechanical input line. The
easiest process will be we assume some value of Pm and instead of moving this mechanical input
line up and down you adjust this clearing angle, you will you move this line either on left or on
right it means you assume some value of fault clearing angle and see whether these 2 areas are
equal.
In case you find actually that for a assumed value of angle a1 is greater than a2 move this line on
left so that a1 decreases and a2 increases right and you do this exercise still these 2 areas are
equal it means what we have done now here is for given input Pm we have obtained the critical
clearing angle and then we integrate the swing equation from this point delta naught up to this
critical clearing angle and find out the corresponding value of critical clearing time.
Now this is the way you can find out a number of points assume some value, suppose the value
of Pm is small you will find that delta c will be very large and a stage may come when delta c
will coincide with delta m right that is the case for sustained fault that is when you are moving
here right and if this fault you find actually that Pm is such that system is stable when delta c
equal to delta m that is the condition for sustained fault then when you, if you are moving if the
Pm is moved up and when you are you have to move this line the moment you find actually the
delta critical clearing angle delta c same as delta naught that become the instantaneous occurring
time and therefore by this approach we can plot the curve relating the transient stability limit
versus the fault clearing time in seconds.
For example, this graph shows for a typical system on this x axis by plotting the fault clearing
time in seconds, y axis we are plotting the stability limit in per unit and the stability limit of the
system when the fault is cleared instantaneously is denoted as 1 per unit and for all other fault
clearing time it is going to be less than this okay.
Now here actually this .2 shows the sustained condition because there is a break here in the graph
right because when the fault is sustained the stability limit is going to very small, sustained fault
condition there is a fault is there and system is not losing stability it means that is the situation
which occurs only when the fault, I am sorry a a the power output is small right. Further as we
have seen that the post fault, I am sorry the post fault power angle characteristic depends upon
the element which has been removed and the remaining system.
Similarly, the during the fault power angle curve is concerned it depends upon the location of
fault on the transmission line. Suppose you consider the 2 machine system and for the whole
analysis what we have done is we have assumed the fault in the middle of the line suppose I shift
the fault location from the middle toward the sending end of the line or you shift this from
middle to the receiving end of the line in that case you will find that the stability limit will be
different the curve relating the stability limit versus the fault clearing time will be different.
Now these 2 curves show the fault at sending end of the line and this show the fault at middle of
the line now you can easily see that where the fault occurs at the sending end of the line okay
what will happen to be power angle curve the during the fault condition during the fault
condition no power will be can be transmitted on this when you consider the machine infinite bus
system let us say the system right, if suppose the fault occurs right at the sending end of the line
then this fault is as good as a fault on this bus.

Okay and therefore the voltage at this bus becomes 0 or it collapses because I am considering
here a balanced 3-phase fault and therefore no power can be transmitted from generator to the
infinite bus right, therefore you will find actually the r2 r2 which is multiplying r2 Pmax sin delta r2
will become 0. This is a very special case and you will find that the transient stability limit will
be less as compared to where the fault occurs in any other point on the transmission line that is
why these 2 curves have been plotted to illustrate the effect of location of fault on the
transmission line.

Now by applying the equal area criteria of stability, we can find out for a given value of
mechanical input and for given fault location we can compute the value of the angle fault
clearing angle delta at which the system just stable that is we can compute the critical clearing
angle for computing the critical clearing angle what is to be done is that you find out this area a1
that is you integrate, you integrate that is a1 can be written as integral of delta naught to delta c
Pm mechanical input minus Pe1 not P1 Pe2 d delta where Pe2 is equal to r2 times Pmax sin delta
okay.

Then the area a2 can be computed integral delta c to delta M, here we will be writing this as P e3
minus Pm d delta where Pe3 is equal to r2 Pmax sin. Okay, if you equate these 2 areas that is you
equate a1 with a2 and you can find the expression for critical clearing angle or the equation for
computing critical clearing angle. This equation has been obtained and it is a cos delta c equal to
delta m minus delta o, sin delta o minus r1 cos delta o plus r2 cos delta m divided by r2 minus r1
this, this expression can be derived without any difficulty by equating those 2 expressions.
Now, when you apply this formula, you have ensure that these angles are the delta m delta
naught they are they are yes substituted in radians sometimes people committed mistake they put
directly in degrees and therefore the result will be observed.
Now this angle delta m has to be calculated by using this formula delta m is equal to phi minus
sin inverse Pm divided by r2Pmax because delta m is delta m is obtained where the where the post
fault power angle characteristic intersects with mechanical input line like it intersect at 2 points
one is the angle which is given by this this equation and another will be phi minus this, therefore
delta m is equal to phi minus this therefore another sometime people commit mistake in
computing the value of delta m correctly and where this r1 is the x12 before fault and x12 during
fault r2 is x12 before fault x12 after fault, that is x12 is the reactance connecting the nodes that is
internal voltage of the generator to the infinite bus voltage and this formula is applicable to a 2
machine system only we do not have such formula for a multi machine system.
Now next point we have to address is how to determine power angle curve for unsymmetrical
fault, till now till now we have assumed a balanced 3-phase fault for our analysis and we also
assume that this 3-phase fault is a metallic short circuit there is no fault impedance involved,
under this situation the, the faulted point is directly connected to the reference bus in the
equivalent network and we analyze it but the moment you have unbalanced fault then things
cannot be as simple as in a 3-phase system because as you know actually that unbalanced faults
can be analyzed by by using the method of symmetrical components.
Okay and when we apply the method of symmetrical components we will come across positive
sequence network negative sequence network zero sequence network and we can compute
depending upon the type of fault the positive sequence currents, negative sequence currents, zero
sequence currents.
We also know that how to draw for a given system the positive sequence network, negative
sequence network and zero sequence networks before I tell you how to account for the
unsymmetrical fault for determining power angle curve during fault condition. We have to
understand some basic concepts one basic concept is which I will introduce and explain in
subsequent discussion.
The concept of fault shunt now here to ah before we understand this fault shunt, let us understand
a since the internal electromotive forces of 3-phase synchronous machine are of positive
sequence that is so for the three phase synchronous generators are concerned we always generate
positive sequence voltages. We do not generate negative sequence or zero sequence voltages, no
power results from interaction of positive sequence voltages with negative or zero sequence
currents. Although, the stator may be carrying negative sequence current, zero sequence current
but when this negative sequence current interacts with the positive sequence voltages no power is
generated no average power is generated.
Similarly, no average power is generated when positive sequence voltages interact with 0
sequence current, okay and therefore to compute the power angle characteristic which basically
relates with the power transfer from machine to the infinite bus okay. We are we have to, we
have to compute primarily the positive sequence currents okay and to compute the positive
sequence currents as we know actually that during fault condition the positive sequence, negative
sequence zero sequence networks are connected in a particular fashion if suppose there is a line
to ground fault these three networks will be connected in series if it is a line to line fault then
these two networks will be connected in parallel looking into the looking into the faulted
terminals.
We have to look where do we connect in parallel where the fault occurs in two faulted points
similarly, we have it is a double line to ground fault then the three networks will be connected in
parallel.
The simplest approach to account for the unbalanced or unsymmetrical fault is by connecting
shunt impedance ZF at the point of fault that is in the positive sequence network we return the
positive sequence network as it is earlier between the fault point and the reference we were
connecting zero impedance that is directly connected.
However, when unbalanced or unsymmetrical fault is there we have to connect an impedance of

value ZF that is called fault shunt ZF. The value of ZF depends upon the impedance Z2 and Zo
of the negative and zero sequence networks viewed from the point of fault that is ZF is function
of positive sequence I am sorry, negative sequence and zero sequence impedances. Here, I will
without a derivation right now I am just giving the result the this table shows the type of short
circuit and the fault shunt ZF is the impedance of the fault shunt, if it is a line to ground fault the
value of ZF is Z0 plus Z2, if it is a line to line fault the fault shunt impedance is Z2, if it is a
double line to ground fault if the fault shunt is the parallel combination of Z0 and Z2 and if it is 3
phase fault ZF is 0 okay.
Now this table is very important and I will show you a list through illustration, how do we get
this in time, now a typical statistics of the occurrence of type of faults or frequency of occurrence
type of fault.
A typical 132 K. V system the data were obtained and out of the total 72 faults which occurred
on the system, 58 were lined to ground faults double line to ground faults were 8 and 3-phase
faults were 6 in fact line to line faults generally gets converted into line to line to ground fault.
You can see very easily here that the frequency of occurrence of 3- phase fault is lowest and the
the frequency of occurrence of line to ground fault is the highest and therefore, in case you
design your system or operating condition of the system is design, considering 3- phase fault it
means we are very very pessimistic in our approach, it may be assumed actually that faults are
would be very severe and we are taking a very safe margin.
However, if you apply only considering line to ground fault definitely you are optimistic where
you feel that these faults may not occur because if you design considering line to ground fault
and if 3- phase fault occurs, system is going to lose stability similarly double line to ground fault
occurs it is going to loose stability in case you do not have any margin right and therefore the
practices I will tell you what are the practices which are as followed ah for designing the system
because we have to make a balance.
We have, we should not be very optimistic we should not be very pessimistic in our approach
before I tell you about the effect of grounding, let us just see how do we account and compute
the value of fault shunt impedance ZF okay. I have told you actually that fault shunt impedance
ZF is different for different fault and the the table had been shown to show you the expressions.
Okay let us take this the simple machine infinite bus problem where you have generator neutral
of the generate is grounded double circuit transmission line. In this case I have taken only one
transformer there is no transformer shown here but if that is there is a transformer here that can
also be considered this delta star connected transformer start point grounded infinite bus system.
Now when you solve the problem considering unsymmetrical fault we need information about
positive negative and zero sequence reactances of the system components for generator Xd
prime is 0.35, the negative sequence reactance of the generator is less than Xd prime is 0.24, X
naught is the lowest that is zero sequence reactance is always low 0.06, for the transmission line
is concerned its positive and negative sequence reactances are equal that is Z 1 is 0.4 and j times
0.4 the and the negative sequence reactance is also 0.4 the zero sequence reactance of the
transmission line is always more than the positive or negative sequence reactance, in this case it
is 0.65 the typical values it may be even more it may be sometimes 2 to 2.5 times even 2 to 3
three times it all depends upon this system. For a transformer we can assume this X1 X2 X0 equal
to .1 that is they are equal, with this the datas which we have assumed let us first obtain the
positive sequence, negative sequence and zero sequence networks.
The positive sequence network is simple is same as what you do actually for analyzing for
balance 3-phase fault condition, this is the internal voltage E prime the direct axis transient
reactance I am putting the reactance value only 0.35 transmission line reactance 0.4, 0.4 and the
transformer reactance here 0.1 and the infinite bus voltage okay this is the voltage V. I am
showing this as a voltage V, now V we shall consider that the fault occurs at the sending end of
one of the transmission lines right at this point.

This is the sending end of the line and therefore I will added this is as good as the fault occurring
at the bus therefore let us call this point as the point P on the series and this is our reference. Now
so far the negative sequence network is concerned it is it will have the same structure except that
the sources will not be present there are no sources because we do not generate the negative
sequence voltages and therefore the negative sequence network can be drawn.
This is my reference bus, okay we call this o this point continues to be P now what we will do
here is that this point we will call as P2 that is the fault point in the negative sequence network
and we will call this point as P1 in the positive sequence network, the points are same P1 and P2
are same on the physical system. The values here are now because in a zero sequence, a negative
sequence network, the generator reactance is 0.24 per unit, the transmission line is same 0.4, 0.4
transformer is also 0.1 okay.
Now we can find out the equivalent reactance of the negative sequence network looking into this
points P2 and o, this exercise when you do you will find actually that equivalent comes out to be
a reactance whose value is 0.133 for this problem you can say this is P2 and you can even call
this as a o2 the reference bus for the negative sequence network, therefore these 2 terminals are
important for us. The next step is to draw zero sequence I mean to draw the zero sequence
network for the system.
When we draw the zero sequence network, we have to consider the connection of transformers
because transformers may be connected in different modes and we have to also consider the
neutral impedance. In case you have put a impedance, in the neutral circuit in the equivalent
circuit that impedance will appear as 3 times the actual value in this particular case, the neutral
have been solidly grounded therefore they do not appear.

In this particular case if you draw this zero sequence network it will come out to be this value is
0.65, this is also 0.65, this is 0.1 this is 0.06, this is the zero sequence reactance of the generator.
This is our reference bus incidentally in this connection, the generator neutrally is grounded
therefore we can connect a neutral point to the reference. Okay if you had it not been grounded
this would have been open the transformer is a start delta transformer with start point grounded
and therefore again this point will be connected.
Now those who do not have the practice of drawing the zero sequence network, I will advice
them to refresh their knowledge about drawing the positive negative in zero sequence networks
particularly zero sequence networks, considering the different types of transformer connections
now here this is the fault point I will call this as a P000 and the equivalent impedance looking into
these two terminals has been computed it comes out to be equal to 0.053 these points are Po and
okay this so far we have obtained the positive negative in zero sequence networks and we have
also obtained the equivalent value of the positive, negative sequence impedances for the network
considering the fault location.
Now in order to consider the effect of different type of fault in the system right, these 3 networks
have to be connected in a proper fashion. They for line to ground fault which we have considered
these 3 networks can be connected in this fashion.
This diagram shows the positive sequence network and here we have not simplified this network,
the fault occurs at the point P1 in the negative sequence network is shown in the terminals P2 and
O2, well while zero sequence impedance is on between the terminals Po and Oo.
Now for simulating line to ground fault the 3 networks positive, negative and zero sequence
networks are connected in series that how we can connect these 3 networks in series that is O1 is
connected to P2, O2 is connected to Po and Oo is connected to P1 this becomes a series
connection. It can be very clearly seen here that the positive sequence network is modified and
here we have now and the impedance connected between P1 and O1 for line to ground fault, the
value of the impedance connected is Z2 plus Z0 and this becomes the fault shunt.

Now this network can be simplified, I will put in this simple form here between the node 1 and 3
the reactance is the transient direct axis transient reactance of the generator, this is the equivalent
reactance of transmission line and the transformer and between this node 3 and O, reference node
we have connected the fault shunt. If we have considered the lossless system then the fault shunt
impedance becomes a pure reactance that is j times XF. This network can be further transformed
that is the start connected three impedances can be replaced by an equivalent delta connected
impedances.
This figure shows the equivalent delta the nodes 1 and 2 are retained, the reference node is also
retained as it is now the reactance connecting the node 1and 2 is X12 and the reactance which is
directly coming across the E prime source E prime is shown here. Similarly, the reactance
coming directly across the infinite bus voltage is also coming and is shown here now know that
these 2 reactances do not affect the power transfer capability or power transmission capability of
the system therefore we concentrate only on the reactance connecting the nodes 1and 2. Now for
line to ground fault X12 is equal to 0.35 plus 0.3 plus 0.35 into 0.3 divided by XF where XF will
be the sum of the 2 impedances that is zero sequence and negative sequence impedance.
Now the value of XF will be different for different type of fault. The value of X12 is computed for
three different types of unsymmetrical faults for line to ground fault condition the X 12 is 1.22 per
unit, for line to line fault X12 comes out to be 1.44 per unit and for double line to ground fault it
comes out to be 3.41 per unit.
We can see here that the reactance connecting the nodes 1 and 2 which is primarily affects the
power flow on the transmission line increases as we go from line to ground fault to double line to
ground fault, if we determine the power angle characteristic during fault considering different
type of faults for line to ground fault, for this particular system consider the power angle
characteristic comes out to be P2 equal to 0.82 sin delta, for line to line fault the power angle
curve is P2 equal to 0.695 sin delta and double line to ground fault P2 is 0.293 sin delta that is if
you examine these 3 power angle curves we find that the power angle curve with line to ground
fault has the highest amplitude while the power angle curve corresponding to line to line to
ground fault or double line to ground fault is having the smallest amplitude or and hence the,
from the consideration of the transient stability limit or the power which can be transferred
without loss of synchronism, the line to ground fault will provide more transient stability limit as
compared to double line to ground fault.
This figure shows the plot of transient stability limit or power limit as a function of fault duration
in seconds. This curve shows the relationship between transient stability limit and fault duration
for line to ground fault, the second curve is for line to line fault, third curve is for 2 line to
ground fault and the last curve is plotted for a 3- phase fault.
Now we can easily see here that in case the fault is clear instantaneously that is when the fault
duration is 0, then the transient stability limit is same in all the 4 cases and therefore we can
conclude that the transient stability limit is not affected by the type of fault, if the fault is cleared
instantaneously.
However, if the fault is cleared the time delay then it is very clear actually that the transient
stability limit is lowest when 3- phase fault occurs and transient stability limit is highest for line
to ground fault therefore, we can see that from the point of view of severity the line to ground
fault is the least severe as compared to 3- phase fault or we can say that 3- phase fault is the
severest fault from the stability consideration.
Now we study the effect of grounding on stability the methods of grounding of a power system
modify the 0 sequence impedance this affects the impedance of the fault shunts for representing
the ground faults and thereby affect the severity of such faults.
A typical to a system has been examined and transient stability limit is computed for different
values of ZS and ZR for a 2 machine system.
Here the value of ZS and ZR are varied from resistive to the reactance value and this stable
shows that as the as the value of the impedance connected in the neutral of the receiving end side
and in the sending end side are varied from ohmic value to the reactance value, the stability limit
varies thus we can say that is grounding affects the stability. Now I can say conclude my
presentation that today we have examined the affect of fault duration type of fault location of
fault and the grounding on the stability of the system, thank you.
Prof. M. L. Kothari
Lecture - 07
Transient Stability Analysis of a Multi Machine System
Friends, today we will study the transient stability analysis of a multi machine system
using a classical approach.

We shall address to some of the important issues related to the transient stability analysis
of a multi machine system, I shall discuss the node elimination by matrix algebra,
classical model of the system using that model we will develop the power output
equations of generators in the multi machine system, then we will write down the swing
equations in a multi machine system and then for transient stability analysis whatsoever
preliminary computations are required to be made then we will discuss those.
One of the requirements while pursuing the multi machine stability problem is the
elimination of some nodes in the system where there is no current injection that is if you
have a system right and at some nodes in the system if there is no injection then these
nodes can be eliminated okay.
Now to understand how these nodes can be eliminated we start with the general system
and the mathematical model of the system is expressed as I is equal to Y bus into V
where I is the current vector and V is the voltage vector that is I are the currents which
are injected at the buses and V are the bus voltages okay, Y bus is a matrix which is a
square matrix okay and this is symmetric matrix you all know about it.
Now to understand how the nodes can be eliminated what we will do is that we will
partition this equation that is the current vector I can be partitioned into two sub vectors
similarly, the voltage vector will be appropriately partition and then Y bus matrix is also
partitioned, after partitioning this matrices we can write down IA IX this is the I vector and
IA is a vector, IX is also a vector.
It can be partitioned as the sub matrix Y matrix is partitioned as K, L, L transpose M,

VA, VX that is we are when we are doing the partitioning right of the system then when I
partition this current vector then the Y bus matrix is also appropriately partitioned and
bus voltage that is also appropriately partitioned. After this partitioning we can write
down the equations, the two basic equation in the form, IA equal to K times VA plus L
times VX and IX can be written as L transpose VA plus M times VX that is this is simply
obtained by multiplying the uh this two matrices on the right hand side of this equation
okay therefore, I get IA is equal to K times VA plus L times VX, IX is equal to L transpose
VA plus M times VX.
Now in case in case this IX is a null vector in the sense all elements are 0, okay in that
case what we can do is that we can write down VX in terms of VA and other elements of
this equation, what we do is that you subtract both the sides of the equation by L
transpose VA IX is anyway 0 and multiply by M inverse. Now when you do this two
operations that is you are subtracting both sides by this L transpose V A right and then you
multiply by M inverse, you will get an equation in the form minus M inverse L transpose
VA it is wrong it is L transpose VA.

Okay this is equal to VX, now we have obtained the expression or equation for VX in
terms of other non-quantities okay. So that what we do is that we substitute the
expression for VX in equation 7.3 right. So that this VX is eliminated and the moment we
eliminate this VX you can write down the equation in the form IA is equal to K times VA
minus LM inverse L times VA. Okay and therefore this equation can be put in the form IA
equal to K minus LM inverse L transpose into VA right that is if I write the equation in
this form, IA equal to K minus LM inverse L transpose this is L transpose into VA.
Here, I and V have the same dimensions and this is a matrix is going to be square matrix
have the same dimensions as that of the K and therefore now I can say that the my model
becomes IA equal to Y bus reduced Y bus reduced into VA. Therefore, this reduced Y bus
this you can say this is Y bus reduced okay and this reduced Y bus can be written in the
form as K minus LM inverse L transpose. Therefore, here the network which initially had
certain number of nodes and those nodes where there was no current injection okay.
These nodes have been eliminated and therefore this is now a reduced order model and
there are no nodes where there is no injection. Actually I will just illustrate this with the
help of a example
Now, let us consider that we have a simple system is the third order system let us say
okay and the Y bus matrix as original system is minus 3.330, 3.330 minus 7.5, 2.5, 3.333
and in, I can put 3 here also 2.5 and minus 10.833. This is let us say is Y bus of a system
which has 3 nodes okay. Now let us assume that the node number 3 is 1where there is no
current injection and I want to eliminate that node, what we do is that we partition this
matrix because there are 3 nodes 1 and 2 and 3, okay therefore we partition it in the
fashion.
So that K becomes now matrix minus 3.3300 minus 7.5 okay. L is a vector in this
particular case 3.33, 2.5, j will be there in all the cases okay and M is minus j times
10.833. Okay now you substitute the value of KL and M in the expression, K minus L M
inverse L transpose.
Now we make the substitution here this is a column column matrix this is simply a scalar
quantity this is a row where when you multiply this 2 it will be a 2 by 2 matrix and then
subtract this form the elements of K matrix and the resultant matrix which comes out to
be j times 2.308, 0.769, 0.769 minus 6.923.
Now I have illustrated through a very a you know small example but this is applicable to
any system which may have ah any number of nodes to be eliminated and this simplifies
the process to a great extent okay. Now our next step will be in solving the multi machine
problem, we have to develop the classical model.
When I say it is a they are using the classical approach we will have to develop the
model. Now in the classical approach for transient stability analysis we make all those
assumptions which are made earlier. These 5 assumptions which we made earlier I will
just reiterate them, first was that mechanical input power remains constant, second was
all asynchronous powers or dumping torques were assumed to be 0, third assumption was
asynchronous generator can be represented by a constant voltage behind direct axis
transient reactance, fourth was that synchronous power can be computed from the steady
state solution of network and the fifth was the phase angle of the voltage behind transient
reactance coincides with the total position with respect to the synchronously rotating
reference plane.
These were the 5 assumptions which we had made, the next assumption which we will
make here is all loads because in any power system we will have loads at various buses
all loads may be considered as shunt impedances to ground with values determined with
values determined by conditions prevailing immediately prior to the transient conditions
that is in the system we have loads connected and we will convert the loads loads by
equivalent shunt impedances okay and while calculating the shunt impedances, we will
consider the condition prevailing prior to the transient.
Okay whatsoever are the condition before the occurrence of transient or steady state
condition, okay using that condition we will find out the equivalent load admittances or
load impedances. Okay this is very important next assumption which is made
Now for analyzing the transient stability or for any a stability problem we always require
the initial operating condition or the steady state operating condition prior to the
occurrence of disturbance.
Now for a multi machine system the steady state pre fault conditions for the for the
system are computed using load flow, there is a mistake by mistake I did not write. The
steady state pre fault conditions for the system are computed using a load flow.
Therefore, a load flow for the system becomes a pre condition for starting the solution of
any stability problem in a multi machine system. Okay now the model for the system.

Okay now the model of the system actually with this all the assumptions which we have
made it looks like this. The transmission network is shown here, to this transmission
network we have generators connected at certain nodes. We can, so that these are the
points at which the generators are connected. There are certain points in the network
where the loads are connected. Okay now these loads are represented by constant
impedances connected to the reference bus.
Now this generators are replaced by ah voltage behind transient reactance direct axis
transient reactance. Now here, here we will not neglect the armature resistance of the
synchronous machine okay. Once we account for the armature resistance then we are
representing the generator by a impedance r1 plus j times xd1 prime r2 plus j times xd2
prime and we find out the voltage behind this impedance, if you neglect this resistance
then it becomes simply the voltage behind direct axis transient reactance but here when
we do the transient stability analysis there is no need to ignore the resistance and we can
account for it.
Once you account for this then these are the voltages behind the transient reactance.
Therefore, now if you see here the network which is enclosed in this rectangle is a
passive network it does not have any source now, okay all the sources have been taken
out that is at this nodes we have connected the voltage sources and all the loads have
been replaced by constant impedances and therefore what we exactly do when solving
this problem is that initially you will perform a load flow analysis okay and once you
perform the load flow analysis you will find out the voltages at all these nodes and all the
nodes where the all the nodes where the loads are connected.
Okay then using using this information the loads will be converted into equivalent
impedances and we will replace the generators by equivalent voltage in series with
impedance. Okay then then this becomes my model which is going to be used for
transient stability analysis.
Now in this case what we have done is that I have shown n generators okay and there
may be ah r nodes where the loads are connected. Okay therefore the system may have
originally n plus r nodes. Now when you perform this load flow analysis you will, you
will find out that the Y bus matrix considering only the line parameters. Okay therefore
now when you perform the transient stability analysis the Y bus matrix will be modified
and it will be obtained considering the machine impedances and this load impedances.
Okay once you obtain this Y matrix accounting for this machine impedances and the load
impedances, the load buses can be eliminated okay and therefore you will be left with
only these n buses where the generators are connected okay all the load buses you
eliminated and the the buses where this generators are connected they are also eliminated
because there is no injection here also but ultimately you will need or you will get a
model which will have the n nodes, where the generators are connected and these are the
constant voltages that is the voltages of constant magnitude their phase angles will very
during the dynamic conditions okay.
Now for this network we can write down the equations in the from I equal to Y bus E,
where I is the current vector are or current injected at all the n buses and Y bus is the
admittance matrix obtained after eliminating all the nodes. Okay therefore using this we
shall be in a position to obtain the equations for for electrical power output from all the
machines because we know actually that where we have to obtain or we have to analyze
the transient stability analysis we need the expressions for electrical power output from
the generators, okay. Therefore using this model which is having only the n nodes where
the generators or internal voltage of the generators are connected and this exercise is a
important exercise for which I have already the explained the methodology.

Now here, we will represent the uh diagonal elements of the Y bus matrix which are
called Yii will be equal to magnitude Yii angle theta ii, which will have real and
imaginary parts Gii plus j times Bii. Similarly, the diagonal elements will be shown as Yij
magnitude Yij angle theta ij, Gij plus j times Bij and these voltages, Ei will be written as Ei
angle delta i.
Now here I am not writing this prime right because once we understand that this is the
voltage behind the direct axis transient reactance and the state of resistance included.
Okay then there because writing every time E prime is little what inconvenient okay
therefore we consider that these are the internal voltages and the magnitudes of these
voltages will remain constant. Now we can find out the electrical power output Phi by
this basic equation that is the real part of Ei Ii star real part of Ei Ii star this is very
important.
Now here you look like this that this is the bus at which the voltage source is connected
say Ei this is i th bus, this Ii is the current which is injected by this voltage source. Okay
and we know actually the the complex power injected is equal to Pi Si equal to Pi plus j
times Qi and this can be obtained by the formula formula always and in this particular
system Ei into Ii conjugate, this star stands for conjugate. Okay and therefore when I take
the real part of this Ei Ii star that becomes the electrical power output from the I th
generator.
Now you we make the we substitute the expression for Ii, the expression for Ii is we have
the vector Ii can be obtained as Yi1 Yi2 Yin multiplied by this vector E1 up to that is you
have this is this is the I th row of the Y bus matrix and when you multiply this with the
voltage vector, bus voltage vector you will get actually this Ii as Yi1 plus Yi2 E2 plus Yin
En that is we can write down this Ii as summation of Yij Ej where j varies from 1 to n, this
is the way we will be writing the expression for Ii .
Now when you make this substitution here for the expressions for Ii that is you have to
the conjugate of this current and then perform simplifications you will get the expression
Pei that is electrical power output of I th machine as Ei square Gii plus j equal to 1 to n Ei
Ej cosine of theta ij minus delta i plus delta j and j will not take the value equal to i and
this equation is valid for i varying from 1 to n for all the machines that is you put i equal
to 1, 2, 3, 4. You will find that you will be in a position to get the expression for electrical
power.
This is a very important expression and for n machine system we will get n such
equations. Now you can if you just look this look at this expressions carefully, you will
find actually that this depends upon the magnitudes of all the voltages which are known
to us, they remain constant during the transient stability analysis that is during the
transient period. These are the Gii there is one term which is missing here actually just
please correct it. It is Yij it is there will be Ei Ej, Yij will also be there.
Okay cosine of theta ij minus delta i plus delta j therefore this, Yij is also known to us
only thing which is not known will be the quantity which vary the delta i delta ij
therefore, what happens is that these angles will be computed by using the numerical
technique and at the end of each step of computation these angles will be known.
Therefore, you substitute the value of these angles which are computed when you solve
the swing equations right and you get the expression for Pei at each step. Now, once we
know the expressions for electrical power output of the machines for all the n machines
then we can write down the n swing equations okay.
The swing equations for each machine will be 2 times Hi upon omega s, d omega i by dt
equal to Pmi, that is the mechanical input to i th machine minus the electrical output of the
i th machine again here. Yij term is missing. Okay that is what we have done is that this
expression is same as I have computed this is the expression for electrical power output
from i th machine okay then this is the main equation and we write down this d delta i by
dt equal to omega I, which is equal to the actual speed of the rotor minus synchronous
speed because d delta i by dt is the access speed of the rotor right and this it is access over
synchronous speed. Therefore, I am writing d delta i by dt equal to omega i which is
equal to d theta i by dt minus omega s.
Okay therefore what we see here is that we have written 2 equations, 2 first order
equations for one machine and therefore, if you have n machines in the system then you
will get total 2 n such first order equations. Now we need the information of what is the
mechanical power input to the machines Pmi and we assume that this Pmi is constant over
the transient period that is mechanical power input to all the machines are assumed to be
constant right. Therefore, there is a necessity to find out what will be the mechanical
power input initially okay.
Now how to do it because under steady state conditions the electrical power output is
equal to the mechanical power input and the system is in steady state condition that is the
derivative terms on the right left hand side of the equations are all 0 okay.
Therefore, with this condition we can write down the expression for mechanical power
input as Pmio equal to Ei square Giio, I am putting o here means this is the this is the
element of the Y bus matrix before the occurrence of transient again here you have Yij
okay, that is the complete expression of wherever actually ah angles were delta i, I put
delta io, delta j, delta jo, now theta ijo here you will find actually that when the system is
subjected to disturbance.
Okay therefore during fault conditions the Y bus matrix of the system will be different,
during post fault condition it is again going to be different right and therefore we will
have 3 different ah Y bus matrices one pre fault condition, one during fault condition,
third post fault condition. Now here to compute the mechanical power input initial
mechanical power input Pmio right. We use the conditions which are prevailing the in the
system before the occurrence of disturbance and this is, I am denoting by putting a further
subscript o, okay.
Now the equations which we have written were 2n equations, these 2n equations can be
written in the general form as X dot equal to a function a non-linear function X, Xo, t
where, X is a vector of dimensions 2n into 1X is a vector of dimension 2n into 1 where, n
is the number of machines and these are all first order nonlinear differential equations or
we can say a set of 2n first order couple non-linear differential equations they are all
couple.
Now X transpose is actually the X is the state vector the state variables are omega 1, delta
1, omega 2, delta 2, omega n, delta n right and let me just summarize here, how we obtain
the mathematical model of the system. The first step is therefore a given system you
perform a load flow study and find out the initial steady state operating condition okay,
once you have done this the next step will be to replace all the loads by constant
impedances.
Then next step will be to knowing the information which you have obtained in the steady
state ah condition you replace all the generators by constant voltage behind a impedance r
plus j times Xd prime okay. Once you have done this thing, you can illuminate all those
nodes where they are no current injections or no source is connected okay using the
matrix algebra.
Once you have done this thing you can find out the set of non-linear coupled differential
equations for the machine okay. Once these equations are there our task is to solve these
equations using numerical technique and plot the graph relating delta versus time that is
the, we have to plot swing curves for all the machines and then we examine the swing
curves to see whether the system is stable or not okay this is what is step.
Now here before we do all the transient steady analysis one has to perform some
preliminary computations. The preliminary preliminary computations to be computed
done are, all system data are converted to a common base this is a pre important
requirement that when you assemble the system data you will find actually that data are
given on the basis of there name plate ratings. Suppose you have a generator of 5hundred
MVA, okay then the generator parameters will be given on the basis of name plate rating
of the machine.
Okay but once you are consoling a system you have to convert all the data on a common
base and the common base generally chosen is hundred MVA, okay 100 MVA is
common because it looks to be little convenient for computations okay. Otherwise there
is no such hard and fast rule you choose the common base and when you say the one is
MVA base another is the voltage voltage in a particular circuit you will choose and then
you will compute the per unit values of all the system parameters on common base.
The second step is the loads are converted into equivalent impedances or admittances the
needed for this is obtained from the load flow study, okay this is what I have already told
you. Now how do we convert actually the loads into constant admittances these steps are
also important this we can write down here, that at suppose the load is represented as PL
plus j times QL at any bus as load is a PL plus j times QL.
This load can be written as VL into IL star because this is the complex complex load right
and with this can be written as the bus voltage , of the load bus, the voltage of the load
bus into the conjugate of the current drawn by the load IL star. Okay using this expression
we will compute the value of elements of the admittance which is going to replace the
load by a constant admittance. Now the admittance will be written as YL plus j times YL
equal to GL plus j times BL, okay.
This can be simply obtained that in this expression in this expression you can substitute
the value of IL, we know that the current drawn by an admittance is equal to admittance
into the voltage and admittance load admittance multiplied by the voltage that will be the
current and admittance is given by this formula, load voltage is V L therefore IL star is
going to be VL star into YL star IL star is VL star into therefore VL into VL star becomes
VL square and IL then YL will become GL minus times j times BL, YL star.
So that we can write down this load as VL square into GL minus j times BL right and
therefore, using this expression equation 7.8 right, we can say that the load admittance is
equal to PL minus VL square minus j times QL by VL square that is you can say that the ah
real part is PL divided by VL square and reactive part is BL divided QL minus VL square
okay and the proper sign is to be consider.
Therefore, this is one important step and this voltage is VL which are substituting in this
expression are the obtained by load flow analysis. Okay another computation which is
required to perform is the voltage behind the transient reactance. Okay now here you can
use this expression Ei prime equal to Vti plus j times Xdi prime into Ii.
Now here I have not considered the armature resistance right but you can modify this by
putting here instead of j times Xdi prime, we can put here Ri plus j times Xdi prime okay
and since the all information is are known, current is known, terminal voltage is known,
you can find out the while finding out this internal voltage we will use this expression,
the current injected by the generator is Pi minus j times Qi divided by Vti star.
This is very standard equation which can be used to compute the value of current and this
current is required in the equation 7.10 to compute the internal voltage. The internal
voltage which you compute, the internal voltage which you compute right will have its
magnitude and phase angle right and this phase angle which you get become the initial
value of deltas, okay and the magnitudes will remain constant during the transient.
Now once you have done these preliminary calculations okay we obtain using this
information, the swing equations. Now you know that we require swing equations under
3 different conditions, one is the pre fault operating condition, second is the during fault
and third is the post fault and therefore when you solve the swing equations suppose I
start solving the equations at time t equal to 0, okay now if the perturbation or
disturbances occur at time t equal to 0 right then immediately from that initial steady state
condition I will be using the swing equations, I will be solving the swing equations which
are applicable during the during the fault on period.
Then once I know that okay after this much time the fault is cleared right it means the
movement I reach that particular point i will change over from the fault on period swing
equations to post fault swing equations and continue to solve it right. Now one should
know actually that suppose the fault occur in the system, how to obtain the ah the swing
equations which will be applicable during fault conditions, pre fault conditions we have
already obtained.
Now the faults can be ah symmetrical faults or unsymmetrical faults. Now in case it is a
symmetrical fault then the problem is bit simple, if it is unsymmetrical then also it is not
difficult because we have already seen how to account for the unsymmetrical fault. Now
if it is suppose I take fault at a particular bus, let us say that fault occurs at a particular
bus then voltage of that bus will collapse it will become 0. Therefore, the occurrence of
fault at the particular bus is accounted by considering or by is by setting that voltage of
that bus to 0.
Okay then once the voltage of that bus becomes 0 we have to modify the Y bus matrix
which will be applicable to this period, this particular fault on condition. This
modification are not difficult one can, once you see that a particular bus is grounded okay
looking into that condition you can modify the Y bus matrix. Now simplest way is that if
you have Y bus matrix for the pre fault condition then if suppose at a particular bus let us
say ah bus number five bus number five if suppose there is a 3-phase short circuit then
the row and column corresponding to the bus number 5 can be deleted from the pre fault
Y bus matrix and the remaining, the remaining Y bus matrix right will become the Y bus
matrix applicable to post fault condition.
In fact actually suppose one particular bus is grounded or since it is sorted it gets get
grounded it they the number of buses which remain now will become less by one right
therefore Y bus matrix also reduce in dimension. Similarly, once you consider the post
fault condition right then, you can obtain the Y bus matrix during the post fault condition
because under post fault condition what happens one line may be out therefore for the
remaining system you have to find out the Y bus matrix.
Now this Y bus matrix will not be assembled fresh one can be modify the existing Y bus
matrix by modifying the elements of Y bus matrix some elements will be affected, for
example actually if suppose there is a Y bus matrix and if the line connecting the node
one and two that is tripped.
Okay then what is that is going to affect the or which elements of the Y bus matrix will
be affected that is Y11 will be affected because that line is out, Y22 will be affected and
Y12 will be affected others will not be affected, right. Therefore once we know that
particular type of line element has been removed right you can easily find out the Y bus
matrix applicable during the post fault period and once we know the Y bus matrix which
are applicable during post fault, during fault and during pre-fault conditions that you
already known because our equations that is the equations for electrical power output are
written using the reduced Y bus matrix therefore this reduction process has to be done for
all the 3 conditions pre-fault fault on condition and post fault condition.
Now I will suggest you to refer to these problems or these two examples which are given
in here, one example is given in the book written by P. M. Anderson and Fouad, power
system control and stability volume one, example 2.6 and 2.7, these two examples this is
example which is very very commonly used for illustrating the transient stability
analysis. It is a 9 bus system, it has total 3 generators, 9 buses okay and in these two
examples all the step by step computations have been carried out to illustrate the steps
involved in solving the transient stability problem.
One more simple example which is given in the ah basic book on power system analysis
that is elements of power system analysis by Stevenson, everybody is aware of this book
Stevenson here he has solved 5 bus system. Okay and in this 5bus system he has
illustrated all the steps which are required to obtain the complete transient stability
solution. Now once the model is obtained and you solve the swing equations, couple a set
of the coupled swing equations, you will get the swing curves.
Okay the swing curves will plot actually as delta one delta two delta three as function of
time. Now to examine whether the system remains stable or not we have to see the
relative variation of the swing curves that is not the absolute values of delta 1, delta 2
delta n that is important what is the relative value that is more important.
Suppose the two machines may accelerate simultaneously okay but the angles may
remain very close to each other or angle difference may remain very close then system is
not losing synchronism therefore, when you examine the stability of the multi machine
system we have to examine the ah the plots of relative ah angles that is we plot actually
delta 12 that is delta 1 minus delta 2.
You plot delta 1 minus delta three suppose there is a three machine system they are only
a 3angles involved and once this curves are plotted in case actually the curves or we can
say delta 12 reaches the maximum value and decreases it shows the stability condition,
in case delta 12 increases continuously it shows unstable condition. Okay with this let me
conclude what we have done today.
We have an we have given or we have studied the basic steps involved in analyzing the
transient stability of a multi machine system. Here we have used a classical model for
analyzing the stability I have mentioned that how do we obtain the required required
models for pre-fault, fault on and post fault operating conditions. Okay and the swing
equations which we obtained for the system are solved by using numerical techniques
and the stability of the system is understood by examining the swing curves which we
plot and particularly the is the swing curves are relating the relative relative angles of the
machines, okay thank you very much.
Prof. M. L. Kothari
Lecture - 08
Modelling of Synchronous Machine
Friends, we should study today the modeling of synchronous machine.
Till now the synchronous machine was model as a constant voltage behind direct axis transient
reactance. The synchronous machine modeling has been a challenge all through and lot of work
has been done over the years to develop more accurate models of the synchronous machine.
Today in our study we will develop the basic equations of synchronous machine and then we will
go to dqO transformation which is also commonly known as parks transformation.
Now this synchronous machine has two major parts, stator and rotor. We shall represent stator
has provided with 3 windings and we assume that these windings are sinusoidally distributed. On
the rotor, we have a field winding on the direct axis and we have amortisseur or damper
windings. In a synchronous generator we provide dampers and these dampers can be represented
by considering considering the amortisseurs located on the d axis and on the quadrature axis.
Now here in my presentation we will presume or we will assume one amortisseur on the d axis
and another amortisseur on the q axis. The convention which we will follow here is that the q
axis leads d axis by 90 degrees, although there are some some you know cases where the q axis
has been taken as lagging the d axis but in IEEE standards consider the q axis leading the d axis
by 90 degrees.
Now here the d axis is along the axis of north pole, it coincides with the axis of north pole then
we will measure the angular position of the, angular position of the direct axis with respect to the
axis of phase A of the stator that is here this straight lines shows the axis of phase A and the the
angular position of the rotor is measured with respect to the axis of phase A and we call this
angle as theta. Further we will be following the generator convention there is the stator currents
are leaving the terminals of the machine that is ia, ib and ic are leaving the machine terminals. The
rotor is rotating in the anticlockwise direction this is direction of rotation of the rotor which we
are presuming.
Now the currents in the rotor circuits are entering the rotor circuit, if you just see here this field
winding the current is entering the field winding and the applied voltage is efd the damper
windings are closed circuits amortisseurs are closed circuits. The current flowing is again into
the amortisseur windings closed circuit.
Some of the important nomenclature are ah will be use here, a, b, c stands for the stator phase
windings, fd stands for field winding, kd stands for d axis amortisseur circuit, kq stands for q
axis amortisseur circuit, this K stands for 1, 2, 3, n, the number of amortisseur circuits that is if I
put one amortisseur circuit on the d axis, k becomes 1 I can say 1 d if there is one amortisseur on
the q axis it is 1q. Okay therefore in general the amortisseurs are represented by putting
substitute kd or kq, theta is the angle by which the d axis leads the magnetic axis of the phase a
winding in the electrical radians and omega r is rotor angular velocity is electrical radians.
The ea, eb and ec are the instantaneous stator phase to neutral voltages that is the voltages which
are shown here these are the instantaneous values and they are with respect to phase to neutral
there is a raise from neutral to phase, instantaneous stator currents are shown as ia, ib and ic.
The field voltage is efd, the field and amortisseur circuit currents are denoted as ifd, ikd and ikq the
rotor circuit resistances will be denoted by Rfd, Rkd, Rkq with Rfd is the resistance of the field
winding, Rkd is the resistance of direct axis amortisseur circuit and Rkq is the resistance of the
quadrature axis amortisseur circuit.
Now here, we will see that we have stator windings, we have windings on the rotor and rotor is
rotating and because because of this we will find actually that the, we come across various types
of inductances in the synchronous machine, the inductances are the self- inductances of the stator
windings, the mutual inductance between the windings of the stator then mutual inductances
between the stator winding in the rotor circuits and self-inductances of the rotor circuits and
mutual inductances between the rotor circuits therefore we come across different types of
inductances in the stator in the synchronous machine.
The, we represent this by double circuit same circuit laa to denote that it is a self laa, lbb and lcc
stand for self-inductances of stator windings that is we will use double circuit notation to denote
the self-inductances or mutual inductances if there are self-inductances the two circuits will be
same if they are mutual inductances the two circuits will be different, like say lab lbc and lca stands
for mutual inductances between stator winding that is lab is the mutual inductance between stator
a phase and stator b phase, so on.
Then lafd, lakd and lakq represents the mutual inductances between the stator a phase and rotor
windings that is lafd is the mutual inductance between the stator a phase and field winding lakd is
the mutual inductance between the stator a phase and amortisseur on the d axis and similarly, lakd
then lffd, lkkd and lkkq represents the self-inductances of rotor circuit. Ra is armature resistance per
phase and we will represent this differential operator P which is your d by dt by the symbol P, P
is the differential operator.
Now in the case of synchronous machine, the self-inductances of the stator winding and the
mutual inductances between the stator windings and they they they are affected because of the
the non-uniform air gap. As we know that the magnetic field produced by the stator winding it
passes through passes through the stator core, through the air gap, through the rotor iron then air
gap and again return backs through the stator core right and therefore the flux produced by the
stator winding will be affected by the position of the rotor.
Now here in this diagram we saw the variation of permeance with rotor position means you
know that permeance is the reciprocal of reluctance. Okay now here I am considering a salient
pole machine and these are the pole location and we are just showing the expanded version. Now
the permeance is maximum when the, when the permeance is maximum along the d axis or we
can say the reluctance is minimum. This graph shows the variation of permeance as with respect
to the position that is angle alpha which is measured with respect to the d axis which coincides
with the North Pole axis okay.
We can easily see that this is maximum position, when it coincide with the Q axis it is minimum
and it again coincides with the d axis it is maximum and this variation is of the form P equal to
Po plus P2 cos 2 alpha that is when alpha is 0, alpha is 0 its value is Po plus P2 and when alpha is
ninety degrees its value is Po minus P2 right that is cos2 alpha becomes minus 1 and it is this
variation of this permeance right is having a strong bearing on the variation of self-inductances
mutual inductances and so on.
Now to understand the whole thing what we start with this we first write down the stator circuit
equations. The basic stator circuit equations are ea is equal to p, psi a minus Ra ia, eb equal to p
psi b minus Ra ib and ec equal to p psi c minus Ra ic, ia ib and ic are the instantaneous value of the
phase currents and p psi a stands for d by dt of psi a, psi a, psi b and psi c are the flux linking
phase a phase b and phase c respectively. Okay that means straight forward that the induced emf
is d by dt of psi a and this will be equal to the terminal voltage plus the resistance drop or now
this equation is drawn considering the generator action okay.

Now here, let us see actually that what determines the flux linkage in the stator phase winding
the flux linkage in the stator phase winding can be written as psi a equal to minus laa ia, now here
I will explain this minus terms but laa is the self-inductance of phase a la into ia minus mutual
inductance between a and b and multiplied by ib minus iac ic plus lafd ifd where lafd is the mutual
inductance between a phase and field winding ifd is the field current similarly lakd, ikd, lakq, ikq.
Now since we have assumed in the basic model here that the flux linkages are shown in the
direction opposite to the current and that is why actually the negative signs are appearing here
that is in these terms you can just see these are the negative signs while the currents are entering
the other three rotor windings therefore they are the positive signs. Now we will see that that
these self-inductances mutual inductances these are not constant these depend upon the position
of the rotor with respect to the windings, the stator windings and we will show that these depend
upon the angular position of the rotor and since the rotor is rotating the angular position of rotor
keeps on changing and therefore these inductances are going to be a function of angular position
theta. Okay now to understand this let us first start with the stator self-inductances, the stator
self-inductances.
Now here the stator self-inductance is denoted by the symbol laa okay and how when we define
this stator self-inductance, the basic definition is the flux linking the phase a winding divided by
the current that is the self-inductance of phase a winding with no currents on other windings that
is when only current ia is flowing and we find out what is the total flux linking the stator winding
a that is the self-inductance of stator winding a laa.
Now when the current ia is flowing okay then the MMF, MMF which is produced due to the flow
of current is Na into ia and this MMF is sinusoidally distributed along the surface of the stator or
along the air gap okay because the stator is suppose to produce a sinusoidally distributed MMF
okay and this MMF has the maximum value along the d axis right it is it is peak is along the d
axis and when you go away from the d axis both sides this is going to decrease.

Now here this diagram shows, this diagram shows the MMF produced by the stator phase a that
is the I am just showing this is the this is the axis of phase a okay and the MMF produced by the
axis of phase a MMF produced by the phase a or stator phase a is having its peak along the d
axis, this is the d axis. I am sorry, this is the not d axis, I am sorry this is the axis of the phase a,
the axis of the phase a. Okay it is a little correction this is the axis of phase a.
Now what we do is we split this MMF into two components both are having the sinusoidal
distribution, one having its peak along d axis another having its peak along q axis therefore this
this graph red graph which I have shown here, this shows the sinusoidal distribution having its
peak coinciding the d axis, this is the again sinusoidal distribution its peak is coinciding with q
axis and the q axis is leading d axis by 90 degrees okay.
Now this can be seen here in this diagram that the MMF produced by the stator right along its
own axis that is the MMF produced by a stator winding of phase a right is having its maximum
value along its own axis that is axis of phase a. Now we have assumed likely that the rotor is
rotating in the anti clockwise direction therefore axis of, now the d axis is shown here and q axis
is leading points and what we do is that this MMF is resolved into two components one along d
axis another along q axis. The d axis component is Na ia cos theta and q axis component is minus
Na ia sin theta, okay.
Now with this MMFs then we can find out what will be the flux produced at the air gap along
this d and q axis. Okay, now here we are showing that the MMFad that is MMF due to due to
current flowing in the stator a phase and it is component along d axis ad is equal to Na ia cos theta
and these are the peak values therefore when ia attains its peak value this will also become this
varying this is varying along as the ia is varying, then the along the q is minus ia Na sin theta
okay.
Now the flux produced long these two axis because of these MMF can be written as MMF into
Pd that is phi gad, g stand for air gap or gap flux okay the phi gad is equal to Na ia cos theta into
Pd and phi gaq is equal to minus Na ia sin theta into Pq. Now here this is the MMF and to relate
this MMF to the flux we are using this term Pd therefore, Pd is in general a permeance coefficient
we call it it is not only the absolute value of permeance but all other parameters which relate
flux to MMF because this is the MMF only this is the flux okay.

Now what is done is we again make use of this phasor diagram, the flux which is produced along
d axis in the air gap, flux which is produced along q axis in the air gap, we resolve them back
along the axis of phase a. that is when you resolve this right then this component will come out
to be equal to phi gad cos theta and the second component comes out to be phi gaq sin theta with
negative sign because there was negative sign already attached with it and therefore we can say
that the air gap flux due to current flowing in the stator winding a only comes out to be equal to
Na ia substituting these values on the previous equations in this form.
The Na ia Pd cos square theta plus Pq sin square theta and this expression when simplified it can
be put in the form Na ia into Pd plus Pq by 2 plus Pd minus Pq by 2 cos 2 theta.
Now here this is very important point to understand that the air gap flux produced, air gap flux
produced by current flowing in the stator winding of phase a is equal to is proportional to a term
Pd plus Pq by 2 and another term Pd minus Pq by 2, cos 2 theta that is this term does not depend
upon angular position, while this term depends upon the angular position. Now we define the
inductance.
The inductance, the self-inductance of the stator phase a due to gap flux only the flux which is
produced in the air gap, lgaa is equal to Na affective number of terms into the gap air gap flux
divided by ia and this comes out to be we substitute the value of phi gaa it comes out to be Na
square Pd plus Pq by 2 plus Pd minus Pq by 2 cos 2 theta okay therefore this can be put in the form
that is this is your Lgaa is the self-inductance of phase a due to gap flux only which can be put as
a constant terms Lg0 plus another term Laa2 cos 2 theta right because as I have seen the I have told
you that the permanence of the air gap varies as a with the position of the rotor and there we
found actually that it has a second harmonic variation. Here, also you find there is a constant
term plus a quantity varying as a function of cosine 2 theta okay.
Now to make the whole thing more complete there is some a leakage flux which does not cross
the air gap. Okay and this leakage flux also contributes the self-inductance of the stator phase
and therefore, when you account for the leakage flux then we can say that the self-inductance Laa
of the stator phase is equal to self-inductance due to leakage flux plus lgaa which I have obtained
in the previous equation that is due to the gap flux and then when you combine these two terms.
We can see here that this mutual inductance or the this second term will not be affected by the
leakage and therefore this leakage term is combined and you find here that the self-inductance
can be written as Laao plus Laa2 cos 2 theta.
Now this is the most important equation to understand that how the self-inductance of stator
phase varies as the position of the rotor varies the angular position of the rotor. Now this angular
position is measured with respect to axis of phase a.
Now this graph shows the plot for the variation of the self-inductance of stator phase a as
function of theta okay and you can identify here that this is the term Laa2 which varies this Laa2 is
constant and this is another term which we call Laao and the total inductance of the stator phase is
now written as Laa equal to Laao plus Laa2. These two terms are constant these constants these are
constant they do not depend upon the angular position that mean the total self-inductance
depends upon the angular position but these two coefficients are constant. Now when we
perform the similar exercise for phase b and phase c, since the the axis of the phase b and phase c
are displaced by 120 degrees with respect to axis of phase a right.

Therefore, the expressions which we have written here for self-inductance of phase b right will
be of the same form except theta is replaced by theta minus 2 phi by 3 and since these the
everything remains same therefore these terms are also same therefore it is not Lbbo but Lbbo is
same as Laao okay similarly, we write down lcc as Lao plus Laa2 cos 2 times theta plus 2 phi by 3
okay very straight forward.
Now next very important point we have to understand is the stator mutual inductances, the stator
winding mutual inductances again we will see that the stator winding mutual inductances also are
function of rotor position that will be function of theta. Now here, here when the when the axis
of the rotor is in the middle of the axis of stator phase a and stator phase b then at that position
the mutual inductance between a and b will be maximum for example the mutual inductance
between phase b and c when you try to see it will be maximum when theta is theta is 30 degree
minus 30 degrees and 150 degrees these they are the positions which we have to see. Okay, using
this information the flux linkage is, flux linkage is of phase b.
When current is flowing in phase a is are obtained that is we want to to find out the flux mutual
flux right that the flux linking flux, linking phase b due to current flowing in phase a okay and
then once you find out this flux. Okay, we can find out the mutual inductance because the the
inductance is the flux linkage by the current mutual inductance will be the flux linking phase b
due to current in phase a and then you divide by the current you will get the mutual inductance.
Here here following the same approach as we have done for done for obtaining the self-
inductance the, the air gap flux flux again the gap flux linking phase b with when current is
flowing in phase a is obtained in this form that is this is obtained in terms of these 2 components
phi gad and phi gaq that is this is the air gap flux along d axis this is the air gap flux along q axis
and after making the substitutions we find actually that mutual flux comes out to be equal to Na ia
minus Pd plus Pq by 4 plus Pd minus Pq by 2 cos 2 theta minus 2 phi by 3, okay.

Now you can easily see here actually that if you substitute here to make this quantity one to
make this quantity one. You can find it out actually ah what should be the value of theta right
and since this term is minus here Pd is always greater than Pq permanence along the d axis is
more then the permanence along q axis and therefore this is minus to have this also minus so that
the total quantity is added up. Okay you can find out the value of this angle theta and you will
find actually that when theta occupies either 30 degrees or 150 degrees it will be maximum. Now
this, this mutual inductance can be obtained as lgba divided by after dividing the the expression
for phi gba by ia okay.
Therefore, the expression for Lgba comes out to be in this form. Okay, now again it can be written
as minus 1 by 2 Lgo, Lg0 plus Lab2, now if you very carefully examine then this Lab2, Lab2 will be
of the same amplitude as Laa2, Laa2 right.
Similarly, you can find out the mutual inductance between b and a and we this ba mutual
inductance between the phase a and b that is equal ba or ab, they are always equal okay and the
expressions are written in the form minus Labo minus, now here actually when you have written
in this form what we have done is that we have accounted for some some leakage flux which also
leaks to windings right because there are, there is a air gap flux and there is some flux which
does not cross the air gap and once you account that we can write down these mutual inductances
in this form, okay again you can see that this depends upon theta. Similarly, you can write down
for bc and cb it comes out to be in the similar form and similarly lac and lca can be written like this
okay.

This diagram shows the variation of mutual inductance as a function of theta between the 2 stator
phases that is here we have shown the Lab and you can easily see that first thing which we see
here is that the the mutual inductances all through negative. Okay and its variation is shown in
this form therefore, this this quantity a constant quantity is Labo and over this is you superimpose
this sinusoidally varying quantity and variation is as a function of 2theta. Therefore what we
have seen till now that the self-inductances of the stator phases or stator winding and mutual
inductances between the stator winding.

Now we will consider the mutual inductances between stator and rotor windings, stator and rotor
winding. Now so far the mutual inductances between stator rotor windings are concerned that
they are function of angular position but they are not because of the variation in permanence here
because so far the rotor is concerned, rotor will always see the same permanence because the
stator stator is having the ah a uniform shape right and therefore, so far the so far actually the
rotor is concerned, rotor windings are concerned right the the there will be no variation in
permanence.
Now here the mutual inductances between stator and rotor windings vary because of angular
position. Now for example, if you take take the stator phase a and field winding in case the axis
of these windings coincide they will have maximum mutual inductance in case the axis of stator
winding of phase a and the field winding they are in quadrature, the mutual inductance will be 0
right and since the rotor is having rotating it occupies different positions therefore, when it
coincides where the direct axis of the rotor coincide with the stator phase a axis or b axis or c
axis they will have maximum mutual inductance and when the quadrature axis of the rotor
coincides with the stator phase axis, okay phase a axis or phase b axis or phase c axis then the
mutual inductance will be 0. Okay therefore, we can write down this mutual inductance Lafd
equal to L, Lafd into cos theta.
When suppose as we know that the theta is measured right considering the axis of phase a as
reference and theta is the angle between the d axis and axis of phase a. Okay therefore when
theta is 0, the mutual inductance between stator stator winding and the field winding is
maximum. Okay and 90 degrees it is now, so the the amortisseur on the direct axis is is also
going to have the inductance, mutual inductance in the form Lakd cos theta right because this the
the direct axis amortisseur is having axis coinciding with the field winding right and therefore
the variation is going to be similar.
Now the mutual inductance in the quadrature axis amortisseur and the stator winding will be
written by the formula Lakq cos of theta plus phi by 2, why this theta is replaced by theta plus phi
2 by 2 because q axis is leading d axis by 90 degrees therefore, this can be written as minus Lakq
sin theta.
Now what we have seen here is till now, we have obtained the expression for the self-
inductances of the stator windings, mutual inductances between stator windings and we have also
obtained the mutual inductances between the rotor windings and stator winding and we have
seen that all these are function of angular position. Okay now we again come to our fundamental
equations that is the stator voltage equations stator circuit equation ea equal to p psi a minus Ra ia
and we have seen that the flux linkage of phase a is now written as minus laa ia minus lab ib minus
lac ic plus lafd ifd plus lakd ikd plus lakd ikq therefore, now in this equation we substitute the
expression for laa, lab, lac, lafd, lakd, lakq.
Okay, then we get the expression for flux linkage of phase a as minus i a into Laa plus ib into Labo
Lab2 cos 2 theta plus phi by 3 this plus sign has come because there was negative sign here
earlier. Okay when you see this mutual inductance there was a negative sign therefore it becomes
plus here. Similarly, you have ic into Labo plus Lao cos 2 theta minus phi by 3 and so on, that is
what we have done is that in this in this basic equation, we have substituted value of all the
inductances okay, which were all found to be function of theta. Similarly we can write down the
flux linkage of phase b and flux linkage of phase c, they are exactly similar except you will find
that theta is replaced by theta plus 2 phi by 3 or theta minus 2 phi by okay.

Now after having written the equations for the stator windings, voltage equation for the stator
windings we can write down the rotor circuit voltage equations. The, in the rotors on the rotor we
have considered 3windings, 1 filed winding and 2 amortisseurs. Okay therefore efd the voltage
applied to the field winding is equal to P psi fd plus Rfd ifd. Now here, since we have assumed
that the current is entering the field winding and therefore the term here is P psi fd plus this is a
simple Rl circuit.
Suppose you have Rl circuit, then the applied voltage is equal to the rate of change of flux
linkages plus voltage drop in the resistance. Okay then the other 2 equations these relate to the
direct axis amortisseur, winding amortisseur circuit and quadrature at the amortisseur circuit here
since there is no external applied voltage therefore we have 0 term here. Okay therefore, there
are 3 basic rotor circuit voltage equations we have 3 basic stator circuit voltages equations. Now
let us write down the expression for these flux linkages psi fd psi kd and psi kq.
Now psi fd can be written as Lffd that is the self-inductance of field winding into ifd plus mutual
inductance of the mutual inductance between between field winding and amortisseur that is Lakd
into ikd okay and there will be no there will be no flux linking the field winding due to the
quadrature axis amortisseur because the the 2 axis are that that the the displacement of 90
degrees between the 2 axis okay and therefore there is no flux linkage contributed by by
amortisseur on the quadrature axis to field axis flux linkage.
Then these 3 terms are here Lafd into ia cos theta Lafd into ib cos theta minus 2 phi by 3and Lafd
plus ic into cos theta plus 2 phi by 3 that is when the 3 stator currents are carrying the values ia, ib
and ic and depending upon their mutual inductances this will be the flux linkage in the stator
winding. Now one point which I wanted to mention here is the so far the self-inductances of the
rotor circuits are concerned that is self-inductance of field winding, self inductance of
amortisseurs they do not depend upon the angular position because because so far actually the
the the magnetic circuit is concerned for computing the self-inductances of rotor circuits are
concerned, these the self-inductances are constant and the since that we have assumed like with
uniform internal surface of the stator okay and therefore no variation of reluctance so far actually
the rotor circuits are concerned.
Similarly, similarly the mutual inductance between the rotor circuits there is mutual inductance
between the field winding and amortisseur on the d axis they will be fixed, they do not depend
upon the rotor position right. Therefore for example, Lfkd is a mutual inductance between field
winding and direct axis amortisseur this is a constant quantity they will not depend upon the
rotor position.

Now this similarly you can write down the flux linkage of amortisseur on d axis and flux linkage
of amortisseur in the q axis okay.
Now with this with this we have developed the complete mathematical model that we have
written three stator circuit equations, we have written the rotor circuit equations we have
expressed all the inductances as function of currents and I am sorry ,not all the flux linkages as
function of currents and the self and mutual inductances.
Now this that is ah this is what is called complete model of the system however the basic
problems which arise are due to due to the variation of these inductances with the variation of
rotor angular position and to overcome this problem and seeing very carefully the expression for,
you see the expression for psi fd, we find here a term Lafd that is the along with this term where
ia cos theta plus ib cos theta minus 2 phi by 3 therefore, this has prompted us to obtain a
transformation and once we go we use this transformation, we will find that the equations can be
simplified and we can make these equations equations where they do not exclusively depend
upon or the inductances do not depend upon the angular position.
Okay the the transformation is of this form that is we define, we define this term ia cos theta plus
ib cos theta minus 2 phi 3 plus ic cos theta 2 phi by 3 multiplied by some constant kd as id that is
these 3 terms ia cos theta ib cos theta minus 2 phi by 3, ic cos theta plus 2 by 3 this complete term
multiplied with some constant kd is denoted by a term id.
Similarly we denote another term iq as minus ikq multiplied by ia sin theta plus ib sin theta minus
2 phi by 3, now this with this with this assumption or this transformation if we consider balance
three phase currents that is ia equal to Im sin omega s t, ib equal to Im sin omega s t minus 2 phi
by 3 ic equal to Im sin omega st plus 2 phi by 3. Okay that is, we are assuming that the 3 stator
currents are balanced with this 3 stator currents to be balanced, okay what we do is if you
substitute and find out the expression for id, the id will come out to be as kd into 3 by 2 Im sin
omega st minus theta. This is very important term that is with this transformation, this id current
id is equal to kd into 3 by 2 Im sin omega st minus theta.

Now if you assume kd equal to 2 by 3, if you assume kd equal to 2 by 3 then the the peak value of
id will be same as Im okay and therefore in the Parks transformation, Parks transformation kd
and kq are taken equal to 2 by 3, that is iq will also be taking the same minus iq into 3 by 2 Im cos
omega st minus theta.
Okay and therefore, if I take kq equal to 2 by 3 then the peak value of iq will be same as Im.
Okay now to make this model complete complete and assuming that suppose the 3 currents are
not symmetrical right then we can define one, 0 sequence current io as 1 by 3 ia plus ib plus ic and
with this definition the transformation looks like this it is very interesting thing this
transformation looks likes this that we can say id, iq, io, a vector consisting of d axis current, q
axis current and i0.
These 3 currents can be written in terms of the phase currents ia, ib and ic in terms of this matrix
and this is called transformation matrix that is transformation matrix is 2 by 3 the first row is cos
theta cos theta minus 2 phi by 3, cos theta plus 2 phi by 3. Okay similarly similarly the second
term and third term.
With this now inverse transformation that is if you write down the expression for phase currents
in terms of the dqO currents then this can be written in this form cos theta minus sin theta 1like
this this is called inverse that is sometimes if you know the value of id, iq, io we can find out the
phase currents.
Now interesting thing which happens is that if I substitute the values of the phase currents in
terms of dqO components right then I get the expression for flux linkage in the d axis called psi d
that is if the the all these fluxes flux linking phase a phase b and phase c right they are also
transformed currents are also transformed that is by applying this transformation. We find that
the flux linkage is can be written in terms of ah the constant coefficients that is psi d is equal to
minus Lao plus Labo plus 3 by 2 Lao Laa2 okay then Lafd ifd plus Lakd ikd that is here the coefficient
of id is a constant term it does not depend upon angular position theta similarly, for psi q and psi
o right we define we define these terms Ld equal to Lao plus Labo plus 3 by 2 Lao similarly Lq and
L0 are defined.

When you make this substitution we can write down the flux linkage psi d as minus Ld id plus
Lafd ifd plus like this. Similarly, when you apply the dqO transformation the flux linking the
rotor circuits are also expressed in terms of the rotor currents rotor currents rotor circuit currents
and the dqO components of currents and again you find actually that these flux linkages as well
as these 3flux linkages with the transform quantities are are independent of rotor angular position
and this is what it helps the whole thing and once you substitute these expressions in our stator
circuit equations.
We get these equation in the form ed equal to P psi d minus psi q p theta minus Ra ideq is equal to
P psi q minus psi d p theta minus Ra iq and eo equal to P psi o psi 0 minus Ra io. Now these are
the 3 basic equations which are written in terms of transform quantities or dqO terms or
sometimes in case in dq of frame of presentation. Okay today I have discussed the basic circuit
equations of the synchronous machine and discuss the dqO transformation and its importance.
Ultimately, we have obtained the stator circuit equations in terms of the transform quantities.
Thank you!
Prof. M. L. Kothari
Lecture - 09
Modelling of Synchronous Machine (Contd)
Friends, today we shall continue to study about the modelling of synchronous machine.

We have developed in the previous lectures, the basic model for the synchronous
machine and we found actually that if we use the model in the original form then the
coefficient that is the inductances were all function of rotor angular position. In order to
simplify the model the dqo transformation was introduced and with this dqo
transformation we found actually that the, the in the mathematical model synchronous
machine the all inductances become constant.
Now we will further discuss now the stator voltage equation in dqo components. We will
develop the equations for electrical power and torque. We shall discuss the dqO
transformation further and its physical interpretation. Any synchronous machine model,
we always prefer to use per unit system of representation and therefore, we will address
how do we convert the synchronous machine quantities into per unit quantities then I will
also discuss an alternative transformation which has all through been discussed in the
literature on synchronous machine modelling. Let us again look at the stator voltage
equations in terms of, in terms of the dqO components.
The stator voltage equations were written in the form of ed equal to p psi d minus psi q p
theta minus Ra id eq equal to pq p psi q plus psi d p theta minus Ra iq and eo equal to p psi
naught minus Ra i naught. Now here, if you look at this first equation we can see here
actually the means identify these terms, the first term p psi d is the rate of change of flux
linkages. Okay therefore similarly, in the second equation p psi q and in the third
equation we have the term p psi naught these are, these terms are known as the the
transformer voltages because they are taking place due to the rate of change flux linking
the circuit. Then we have next term as psi q p theta p theta is d theta by dt that is the
angular speed of the rotor that is omega R, we call it now here this term. Similarly, this
term psi d p theta these two terms are denoted as speed voltages.
We will further see that in in this equation similarly, in this equation the speed voltage
terms dominate over the transformer voltage terms and in many simplifications we may
neglect the transformer voltage terms as compared to the speed voltage terms the moment
you neglect this then this equation these equations that is the stator voltage equations will
become algebraic equations because if you just look here in this 3 equation then we have
this as a derivative term and the moment you can neglect this derivative term right the
remaining expression that is ed equal to minus psi q p theta minus Ra id Ra id this becomes
a algebraic equation right.

Now we will develop the expression for electrical power and torque in terms of dqO
components. The electrical power output we denote it by the symbol Pt here, as can be
obtained as product, some of the products of instantaneous voltage and instantaneous
currents that is for phase a it is ea into ia, phase b eb into ib, phase e ec into ic right. These
are the instantaneous powers you add these 3instantaneous powers you get the total
instantaneous power.
We all know actually that in case the system is having the stator currents and stator
voltage very sinusoidally right, then the some of the instantaneous powers come out to be
constant. Okay this is denoted by the symbol Pt, now what we do is that we apply dqO
transformation. We can apply the dqo transformation on this voltage as well as the
current currents and when you apply this dqo transformation then the power output P t can
be written as 3 by 2 times ed id plus eq iq plus 2 times eo io that is here the terminal power
is expressed in terms of dq components of voltage and currents okay.
Now we will see here actually in this that instead of having the term e naught, i naught
we have a term 2 times e naught, i naught when we talk about another type of
transformation that is called alternative transformation. We will find actually that the the
Pt will be expressed simply as ed into id plus eq into iq plus eo i naught it is something like
this the that becomes a power invariant transformation. Here this transformation is not
power invariant because you can see there these are 3 by 2 terms then this term is having
the coefficient 2 anyway this is resulting because of the type of transformation.
However when we convert these equations into per unit quantities, the per unit quantities
will be so chosen, so that this becomes a simple expression and this term 3 by 2 and 3 can
be eliminated we will find this thing okay. Now if the system is balanced we do not have
zero sequence quantities a balance system this e naught, i naught will be 0 and therefore
we can write down the total power Pt as 3 by 2 times ed id plus eq iq okay.

Now to give further insight into the power expression what we do is that we substitute the
expression for ed and eq, ed and eq in terms of the flux, flux linkages okay that is I
substitute ed equal to p psi d minus psi q p theta minus Ra id that is in this expression for
this power wherever ed and eq are present what we do is that we substitute the expression
for ed okay.
Now if we make this substitution and simplify then we can write down the output power
Pt as 3 by 2 term remains as it is id p psi d plus iq p psi q plus 2 times io p psi naught that
is in this expression we have this derivative terms. Okay the second term is psi d iq minus
psi q id into omega r that the omega r is your d theta by dt that is the angular speed of the
rotor minus id square plus iq square plus 2 times io square into Ra.
Now this 3 terms which we have in the expression for Pt can be identified as the first term
can be identified as rate of change of armature magnetic energy, rate of change armature
magnetic energy, second term is identified as power transferred across the air gap and
third term is armature resistance losses that is we have this 3 terms which can be
identified as the rate of change of armature magnetic energy that is in these in the
armature there is some magnetic energy stored because of flux linkages okay and this
magnetic energy which is stored right is varying at a certain rate right, that magnetic
energy is that are increases or decreases it dependents upon whether p psi d p psi q and p
psi o are positive or negative right and from that point of view this is considered as the
rate of change of armature magnetic energy and therefore when we are looking for
stability analysis right.
The the power which is transferred across the air gap is of concern that is in our swing
equation which we had written right the electrical power pe which we write that is the
power transferred across the air gap because that is responsible for producing the torque
and therefore now we write down Te the electrical torque as three by 2 psi d iq minus psi
q id omega r divided by omega mechanical now this torque is written in Newton meters
okay therefore we had this expression 3 by 2 psi d iq minus psi q id omega r, we are
dividing by this omega mechanical.
We know that this this term was the electromagnetic power that is the power transferred
across the air gap. You divide this power by speed that will give you the torque okay and
therefore, the expression for torque electrical torque is given by this expression now we
know the relationship that the omega r, the rotor speed in electrical radiance per second
divided by the mechanical speed in radiance per second this ratio is equal to Pf by 2,
where Pf is the number of filed faults.
Okay therefore, we can write down the expression for Te the electrical torque as 3 by 2
psi d iq minus psi q id Pf by 2 this is the most important expression which we have to keep
in our mind. Now here the next step which we will be studying will be the how do we
interpret the dqO quantities that this transformation what is the physical interpretation
associated with this transformation.
Now we all know that ah the stator of the synchronous generator when it carries balance
three phase currents it produces resultant mmf wave, this mmf wave is sinusoidally
distributed in the space that is the stator of a 3 phase synchronous generator produces
sinusoidally distributed mmf wave in the air gap okay. Now any quantity which is
sinusoidally distributed can be resolved into 2 sign terms.
Okay now here what is done is that this resultant mmf which are sinusoidally distributed
is resolved into 2 quantities one along the d axis another along the q axis okay and
therefore what is done is that the armature of the synchronous generator is replaced by 2
fictitious windings, one is along the d axis of the generator synchronous generator
another is across the q axis of the synchronous generator and when these windings are
made to carry the current id and iq right and they rotate at the same speed as the rotor it
means it means the these two fictitious windings are stationary with respect to the rotor
and therefore the mmf produced by this fictitious windings act on constant permeance
path and therefore the inductances which we come across are constant.
We will see that under steady state conditions, the currents which are flowing in these
two fictitious winding id that is direct axis fictitious winding and quadrature axis fictitious
winding these two current that id and iq will come out to be constant in magnitude they
will be just DC currents, okay this we will just now establish.
To establish this thing we will start with that let us assume that the armature of the
synchronous generator is carrying balance steady 3 phase currents that is system is
operating under steady state condition, okay no dynamics is involved let us assume that
ia is equal to im sin omega s t plus phi this phi is some initial phase angle of the current ia
this is arbitrary it can be 0 also ib is equal to im sin omega s t plus phi minus 2 phi by 3
and ic equal to im sin omega st plus phi plus 2 phi by 3 that is these 3 currents ia, ib and ic
form balanced balanced set of currents okay.
Now what we do is we apply dqo transformation and when you apply the dqo
transformation we will be in a position to write down id, iq and this io will be 0 because
we are considering the balanced steady state conditions that is in balance system the zero
sequence quantities are absent. Now here I will just add one thing that when you talk
about the zero sequence currents you always talk in terms of zero sequence current to be
the phasor right here the zero sequence term we are putting is as it is the instantaneous
component, instantaneous that is we are writing zero sequence as ia plus ib plus ic by 3 is
the instantaneous quantity but this is applicable here.
Now after applying this transformation this exercise has to be done that is in between you
have lot of trigonometric simplification to be done after you do this trigonometric
simplification the id the current in the du axis, d axis fictitious winding comes out to be
equal to im sin omega st plus phi minus theta iq equal to im cos omega st plus phi minus
theta and i naught is equal to 0. Okay now here what is this theta, theta is the { angular
position of the rotor and angular position at any time is equal to omega r into t, omega r is
the rotor speed under under steady state operating condition the value of omega r comes
out to be same as the synchronous speed.

Therefore, now we substitute here omega rt equal to omega s into t right the moment you
make this substitution here in this expression that is theta is to be put as omega s into t
this term vanishes and now we will have the id equal to im sin phi similarly, iq equal to im
cos phi right and therefore you can easily see here that for a given stator currents im is
constant it is a peak or it is the amplitude of the phase current as if id can be written as im
sin phi iq equal to im cos therefore what we see here that is id and iq these are constant they
are not varying as a function of time therefore they become DC current right.
However, under dynamic conditions when the system is dynamic condition rotor speed is
not exactly equal to the synchronous speed right in that case the id and iq will vary but the
variation will be a low frequency variation, it is not going to be the 50 hertz are 60 hertz
variation.
Okay now some of the advantages which we get by this transformation can be
summarized although we have seen these advantages but I will just summarize the
advantages of dqO transformation. The first advantage is that the dynamic performance
equations have constant inductances we have seen we have seen the mathematical model
which we have written in terms of dqo transformations okay and in this dqo
transformation by after applying this dqo transformation, the equations have constant
inductance this is the one of the biggest merit of applying the dqO transformation, second
is that under balance conditions zero sequence quantities disappear that I am not present
the id and iq under steady state conditions are constant in magnitude.
However, when we perform stability studies then stability studies involves slow
variations of id and iq the frequencies will be below 2 to 3 hertz that is the frequency of
oscillation of the rotor which will come across will be of the order of, order of point 5 to
say 3 hertz or we can say that the frequency of oscillation of these currents will be less
than 2 to 3 hertz is very low frequency variation this is most important point we have to
keep in mind and another very interesting advantage which comes is that that d and q axis
quantities particularly which will inductance ld, mutual inductance lq these can be
measured by performing experiments or by making some measurements on the terminals
of the machine.
Since all of you must have done experiment in your laboratory to measure the direct and
quadrature axis reactance of a synchronous machine. One important test which you
perform in on is the slip test right. Now we come to another important aspect so for what
we have done is we have developed the expressions for the stator voltage equations, rotor
voltage equations. We have also developed expressions for the stator and rotor flux
linkage equations we have also developed the expressions for electrical torque and given
some meaning to the transformation, physical meaning to represent transformation.
Now in any power system studies per unit system of calculation is the most convenient
one and merits of per unit calculations are well known to all of us and therefore now we
will study how do we transform, transform our system equations using per unit quantities.
Now whenever you transform the actual quantities into per unit quantities we have to
choose base quantities.
For example, in any circuit we may choose MVA base and the voltage base then we can
find out the base quantities in different parts of the circuit MVA base remains same while
the KV base depends upon the circuit voltage that is the transformers which are involved.
Now here, we have two parts, one is the stator another is the rotor. Okay therefore first
we will define how we obtain obtained the per unit representation of stator quantities.
Now to start with here we will define the base quantities, now in any system we can
define some base quantities and other quantities can be derived from those chosen base
quantities like we start with stator quantity we will use this subscript s to denote that
these are the stator quantities or the base quantities for the stator. We will define esbase as
peak value of the rated line to neutral voltage a slight variation is here that is instead of
defining the RMS value as the base value here we are defining the peak value okay.
Similarly, the isbase that is the current in the stator circuit or the base current in the stator
circuit is equal to peak value of the rated current in amperes. The additional quantity
which we have to refer here is the base frequency. Generally, we do not require actually
referring the base frequency when we talk about the power system network calculations
but here this is also an important quantity, therefore these 3 terms are primarily decided
therefore decision comes like this that rated frequency becomes my base frequency.
Now with this 3 quantities defined other quantities are defined in terms of the basic base
quantities like say omega base becomes 2 phi f base okay, omega m base that is the the
mechanical speed right. The base value of the mechanical speed is written as the base
value omega base into 2 by pf further the impedance base Zsbase is equal to es base upon is
base but we can define. The base value of the inductance in terms of stator base quantities
that is Lsbase is equal to Zsbase upon omega base. We can also define the flux linkage or
base value of the flux linkage that is psi s base equal to Lsbase into isbase. We know that
flux linkage can be written as L into i Lsbase can be replaced by this quantity the Lsbase into
is base can be replaced as esbase upon omega s base that is what we do here is this Lsbase
can be written as Zsbase upon omega base okay and then Zsbase into is base can be written
as esbase. So that the flux linkage the base value of the flux linkage can be written as esbase
upon omega s Weber turns.
Now let us see actually that how we define the three phase VAbase volt ampere base
because as I told you that your power system computations. We start with the 3 phase
MVAbase, since we started with the voltage and current bases. Okay therefore we write
down the three phase VAbase is defined as three times Ermsbase into Irmsbase that is the phase
voltage its rms value multiplied by stator current and its rms value this can be written as
now 3 rms value is equal to base quantity by root 2 because base quantity is chosen as the
peak quantity here.
Similarly, isbase by root 2. So that what we see here is very very important relationship
that is 3 phases VAbase comes out to be equal to 3 by 2 times the voltage base into current
base. Okay this relationship is very important because we always make use of the three
phases VAbase for our computations right but the relationship between the quantities
which you have chosen as the base quantities and the 3 phase VA comes out to be like
this.
Then the torque base the torque base can be always written as three phase VAbase divided
by omega m base. Now when you talk about the base quantities or base power we always
say base power is equal base MVA if you recollect in your power system calculations or
computations then whatsoever base MVA is there that is equal to base power in
megawatts and therefore power divided by the base speed will give you the torque base
therefore here the torque base is equal to 3 phase VA base that is omega m okay.
Now when you make the substitutions here in this expression for 3 phase VAbase equal to
3 by 2 esbase, isbase and this esbase and isbase, these quantities can be further expressed in
terms of flux linkage base. Therefore, torque is now expressed as 3 by 2 Pf by 2 psi s base
into isbase because we have seen earlier actually the torque is written in terms of flux
linkages and currents right therefore we prefer to specify the base or torque base in terms
of flux linkages and current base quantities. Okay this is in Newton meters.
Now what till now what we have done is that started initially the definition of base
quantities in the stator we have chosen the stator voltage and stator current and the
system frequency as the base quantities okay and then we have derived that what will be
the base quantities for other variables, other variables like say impedance, base
impedance, base torque base flux linkages, base flux base in base inductance and so on.
Now we start with transforming our stator voltage equations into per unit terms or per in
terms of per unit quantities is it not okay now again let us start or stator voltage equation
is ed equal p psi d psi q omega r minus ia, id into Ra okay this is the equation which we
have derived in terms of dqO components we want to transform this equation in terms of
per unit quantities how do we do it you divide this whole expression by voltage base
because these all these terms are the voltage terms therefore what you divide you divide
this by esbase.

Now when you divide this by esbase we make use of this relationship that esbase is equal to
is base Zsbase it is also equal to omega s base into psi s base this relationships we have
already derived that is this 3 the voltage can be written as the product of impedance and
current can also be written as product of omega and flux linkages okay. Therefore, what
we are doing here is that ed is divided by esbase p remains as it is we are not touching p
right now psi d since it is a flux term what we do is that we divide this flux term instead
of es base i put omega s base, phi s base here because esbase is same as this quantum.
Similarly, here we put omega s base, omega s base okay and in the third term where I
have Ra into id I put here not in terms of the flux linkages and speed by put errors Zsbase
into isbase because this the denominator terms which I have put here are same as esbase this
can be done okay. Now by doing this what happens is that this ed becomes now per unit
value. Okay that is the direct axis voltage ed can now it becomes a per unit quantity.
Similarly, the flux linkage becomes per unit, the resistance becomes per unit resistance,
now the current becomes per unit current, okay now to distinguish the per unit quantities
from the quantities a real quantities in amperes or ohms or watts right what we do is that
we will put a put a super script bar, okay a super bar on the top that is I will denote this
as ed bar equal to 1 upon omega base p psi d bar minus psi q bar omega r bar minus Ra
bar id bar as we will see subsequently the moment we develop a mathematical model of
the system in terms of per unit quantities right.
We will drop this super bar, drop it because every time writing this bar is not very
convenient but when you are trying to derive it first we will maintain this bar and then at
the end we will drop it for further further use of this model. Now here in this equation p is
d by dt and time is in seconds, okay therefore we can also define define the base value of
time and the definition of base value of time is the time in seconds required for the rotor
to rotor to rotate by 1 radian, 1electrical radian per second electrical radian turn per
second electrical radian that when rotor makes one electrical radian and the time which is
required to make 1 radian rotation that is called tbase that is 1 radial divided by omega
base that is when rotor rotates through a angular displacement of one radian the speed is
omega base base speed then the that particular time is called tbase base time okay.
Therefore, this base time can be written as 1 upon 2 phi fbase and if I now use this base
time then this term can be replaced by p bar, psi d bar minus psi q bar, omega r bar minus
Ra bar that is this expression is now in terms of per unit quantities that is p bar is also
expressed in per unit and all other terms are in per unit.

Similarly, you can write down for the q axis voltage and zero sequence voltage. We have
just now mentioned that p bar was defined as d by dt bar where dt bar is a per unit current
which comes out to be p bar comes out to be one upon omega base into p okay and we
have seen here actually that one upon omega base into p is the it comes out to be equal to
p bar that is why we have come to now this model, this model where ed bar is equal to p
bar, psi d bar minus psi q bar, omega Ra bar, id bar and similarly for eq and eo that is we
have developed the stator circuit equations in per unit terms okay.
Now the next step will be to develop the rotor circuit equations also in per unit quantities.
Now here to start with what I will do is we will assume assume the rotor circuit voltage,
rotor circuit currents as the base quantities, I am not I am not addressing at present how
do we choose that we will see later on but assuming that the for the rotor circuit, the field
circuit because in the rotor we have 3 circuits one is the field circuit another is the direct
axis amortisseur circuit, third is the quadrature axis amortisseur circuit and therefore for
the field circuit a base voltage will be efd base which will be written as omega base into
psi fd base that is psi fd is the flux linkage in the field circuit a base quantity in the field
circuit.
Similarly, we will have Zfdbase into ifdbase okay that is as we have done in the case of stator
similarly we specify the quantities in the rotor also. Okay but how do we choose this base
quantities in the rotor that we will establish in few minutes. Therefore, now following the
same approach as we have done for the stator we can write down efd bar equal to p bar
psi fd bar, Rfd bar, ifd bar that is in this equation where actually we have field circuit
applied voltage is efd okay the current flowing is ifd flux linkage is psi fd therefore we
can establish the equation in the per unit quantity they are very simple as state forward
the approach is exactly same you divide by efd base the actual quantities and then
wherever you have this Rfd, ifd divided Zsbase Zfdbase ifdbase.

Okay and you will get this equation other two equations also in the same form because
here we do not have term because they are all close circuits amortisseur closed circuit
therefore 0 equal to okay.
Now we come to very ah crucial thing actually that is how do we express the stator flux
linkages in per unit. Okay because so far what we have done is the stator voltage circuit
equation, stator circuit voltage equations, rotor circuit voltage equations. We have
expressed in per unit terms therefore now our next step is to express the stator flux
linkage equations in per unit.

Now to understand this let me ah rewrite here I have written again psi d is equal to minus
Ld id plus La fd ifd Lakd, ikd this equation was derived this is the direct axis flux linkages
psi d equal to minus Ld id plus Lafd what is Lf this is the mutual inductance between the d
axis and field winding. This is actually amortisseur and d axis winding that is why Lakd.
Okay now what you do is that you divide by this whole equation by base, when you
divide this by base psi d will be divided by psi s base, psi s base is equal to Lsbase into
isbase. Okay therefore you can you can the divide the whole quantities by this base
quantity right. Therefore you will have Ld divided by id divided by Ld into id divided by
Lsbase is base then Lafd into ifd divided by Lsbase into isbase.
Now here the problem comes that the per unit value, per unit value of the field current is
expressed in terms of field current base not in terms of the stator current base okay
therefore what you do is that you multiply this term is okay you multiply this term by
ifdbase into ifdbase. Okay then in this term what we can do here is that this ifd divided by
ifdbase can be written as per unit current in the field circuit then remaining quantities Lafd
divided by Lsbase into ifdbase divided by Lsbase this quantity will be denoted by a term Lafd
bar that is you can see very interesting thing that we have the now this 3 ratios. Okay we
can identity this ratio ifd upon ifdbase as ifd bar right then we have the term Lafd by Lsbase
then ifd base upon isbase okay.
Now this these are when you divide these two base currents it comes to be a multiplying
factor only and therefore what we do is that this Lafd will be denoted by a symbol Lafd
bar to ifd bar. Similarly, you will have Lakd bar ikd bar okay is it understood. Now once
you understand this thing actually the I have done it for one direct axis plus linkage
equation. Similarly. We have to do it for other flux linkage equations that is by doing this
exercise we have written the expression for stator winding flux linkage equations that is
the mutual flux linkage this psi d is the mutual flux linkage psi d bar equal to minus Ld
bar id bar plus Lafd bar, ifd bar plus Lakd bar ikd bar.
Now these terms are defined as Lafd bar is defined Lafd bar is defined as now we have just
now seen that Lafd bar is defined as Lafd upon Lsbase into ifd base upon is base right.
Similarly, Lakd bar Lakq bar right these terms were defined in terms of the mutual
inductances, base quantities, the stator and base quantities in the field circuit. This this is
very important actually speaking here here actually it is not state forward where we have
a base actual quantity divided by base gives you a per unit quantity here that is let me just
summarize here that the direct axis flux linkages are written in this form where these
quantities are to be defined okay.
Similarly, psi q is written in this form and psi naught is written in this form, okay they are
the three flux linkage equations in the stator circuit okay and where these terms are
defined as Lafd bar by this expression Lakd by this expression Lakq by this expression do
you understood. Now after we have expressed express the stator flux linkages psi d, psi q
and psi naught in terms of or or converted the stator flux linkages equations into per unit
quantities, next step is to convert the rotor flux linkage equations.
Now when you convert the rotor flux linkage equations, we will come across the psi bar
fd that is actually this is the flux linkage of the field circuit you can you again start with
the original equations, original equations and then divide by the base quantities right and
then then these terms may have to be again defined. These terms will not be required to
be defined that is Lffd will be simply equal to Lffd is the self-inductance of the field
winding right therefore, you divide this by the Lfdbase right where this is the rotor quantity
divided by the rotor base quantity it becomes a per unit quantity therefore there is no
problem here.
Similarly, for the current also there will be no problem but when we talk about this ikd the
La, Lfkd divided by ikd. Okay you will find this similar problem will come right and
therefore these quantities have to be expressed in terms of for example ikd is to be
expressed in terms of ikdbase. Similarly, ikq has to be expressed in terms of its own base
right.

Therefore, when you do this exercise you will find that we can express the rotor circuit to
flux linkages in terms of per unit quantities where these terms Lfda bar, Lfkd bar are to be
defined as all of the all this 5 terms I mentioned this 5 terms will appear in those three
equations are to be defined in this base.
Therefore, now what we see here is that that we have established, establish the rotor
circuit equations that is rotor circuit voltage equations, rotor circuit flux linkage
equations, okay in per unit terms while the basic problem here remains is here remains is
that how do we choose the rotor circuit base quantities because synchronous machine is
having stator and rotor parts, what is the relationship of base quantities in the rotor circuit
with the stator circuit base quantities right and for that we have to establish some basic
rules and I will just state what rules are to be established is that when we choose the base
quantities for the rotor and we want our our intention is our our intention or the objective
is we want to simplify the equations make this equations as simple as possible.
Therefore we have 2 major assumptions or requirements but in the per unit mutual
inductances between different winding are to reciprocal. This is one requirement that we
would like to choose the per unit quantities in such a fashion, so that the mutual
inductances become reciprocal right that is mutual inductance between the d axis field
winding d axis winding that is end the the field winding there should be equal therefore if
I write down that Lafd should be equal to Lfda.
Okay another thing which would like to do is that all per unit mutual inductances
between stator and rotor circuit in each axis are to be equal that that is another way of
simplifying that we would like to make this inductances mutual inductances on each axis
equal right with this I conclude my presentation today and let me summarize what we
have done today. We have discussed the stator circuit, voltage circuit equations in terms
of dqO components. We have also transformed the stator and rotor circuit voltage and
flux linkage equations in per unit quantities. Thank you!
Prof. M. L. Kothari
Lecture -10
Modeling of Synchronous Machine (Contd.)
Okay, friends today we will continue with the study of modeling of synchronous machine.
Today we will address some of these issues per unit system for rotor circuits, per unit power and
torque and alternative transformations. We have discussed dqO transformation and we will also
talk about the alternative transformation which has also been proposed in the literature and then
we then, we will talk about the steady state analysis of the synchronous machine. Now we have
developed the per unit equations for the stator circuits. Okay we have also developed that to the
per unit equations for rotor circuits.

Now these terms were defined and the definitions were given like this Lafd bar equal to Lafd
divided by Lsbase, ifdbase divided by s base that is these three terms were defined that they are
basically in the stator circuit equations you see this that the stator circuit flux linkage equations
psi d, psi q and psi o.

Therefore in these three equations we had defined these per unit inductances now these per unit
inductances are the mutual inductances okay.
Similarly, when we talked about the per unit rotor flux linkage equations, we had define these
terms these mutual terms Lfkd, Lfda these are all per unit quantities which were defined.
Now, now we want to establish the per unit system for the rotor. Now when I am trying to say
per unit system for the rotor in the sense that we have chosen the base quantities in the stator
circuits. Now how should we choose the base quantities for the rotor circuits, the rotor circuits
are rotor field winding and rotor amortisseurs. Okay now the primary consideration for choosing
the base quantities in the rotor are to simplify the the flux linkage equations to make this flux
linkage equation as simple as possible.
Now one requirement which we have put is or one way of simplifying is we make mutual
inductances between different windings reciprocal. We make mutual inductances between
different windings reciprocal. Now when I say that we make these mutual inductances reciprocal
in the sense that you there is a mutual inductance between the stator winding and the field
winding okay.
Now the if I write down the mutual inductance between the stator winding and field winding I
can write down Lafd. When I write down between field winding and stator winding it can written
as Lfda we want to make these two inductances equal. So that that is what is the meaning of
reciprocal okay another thing which we would do to simplify the equations will be that the
mutual inductances between stator and rotor circuits on each axis are to be equal that is when I
say each axis means, we have two axis d axis and q axis. The on the d axis we have the stator
direct axis winding, field winding and amortisseurs on the d axis therefore, there exists mutual
inductances between these windings we would like to make them equal.
Similarly, there exists mutual inductance between the q axis winding on the stator and q axis
amortisseur we want to make these mutual inductance also equal therefore we are trying to do
two things, one is that we will try to make or we will make the mutual inductances reciprocal and
the mutual inductances on each axis equal okay.
Now this will simplify the equations to a great extent. Now how do how do we make the mutual
inductances reciprocal, now to make the mutual inductance reciprocal you just look at the
suppose you look actually this equation I just put this, this is slightly I have to make that I think
Lakd. Okay, now this Lakd and another is Lkda we want to make them equal okay. Now these
quantities were defined now when to make them equal reciprocal, okay so that what is to be done
is that we equate the expression for this with the expression for Lkda, okay.
Now we will start first that is let us say that we want to make the per unit mutual inductance Lfkd
equal to per unit mutual inductance Lkdf right that is the amortisseur and field winding, okay that
is these are the two, you know mutual inductances one between between field winding and
amortisseur another is between amortisseur and field winding and want to make these reciprocal.
Now these quantities were defined now as per the definition this quantity L bar fkd that is the per
unit mutual inductance between field and amortisseur windings were defined like this that is Lfkd
upon Lfdbase, ikdbase upon ifdbase that is in the denominator here is Lfdbase, ifdbase here you have
Lkdbase, ikdbase. Now this is by definition which we have defined actually when we wrote the
equations for flux linkage equations in the rotor circuit. Now when we equate these two, okay we
will get this expression that is you just cross multiply
You will find that it becomes Lkdbase into ikdbase square equal to Lfdbase, ifdbase square you can just
check it it comes out to be like this that is ikdbase comes from this side, so okay and this these two
cancel out they are equal okay and we get this equation. Now here what we do is we multiply
both sides of this equation by omega base, once you multiply by omega base it becomes L omega
base into Lkdbase, ikdbase square similarly, omega base equal to Lfdbase into ifdbase square.
Now you can identify this quantity omega L into i this quantity this is actually the voltage this is
a voltage therefore I take this base frequency in radiance per second the base inductance and
multiply this by current, base current this will come out to be base voltage okay and now we are
taking these quantities for amortisseur the quantities are for amortisseur Lkd ikd therefore the base
voltage will become the amortisseur circuit base voltage this d axis amortisseur circuit base
voltage similarly, here.

Now when you do this you will find this is very important relationship that is the base voltage in
the direct axis amortisseur circuit multiplied by the base current in the direct axis amortisseur
circuit this product must be equal to the the base voltage in the field circuit into base current in
the field circuit, it is very simple that is the the base vi in the amortisseur circuit should be equal
to base vi in the field circuit
Now if you do this you will find actually that the mutual mutual inductances right between the
field winding and the amortisseur winding will become reciprocal. Okay therefore this is one
important requirement that base VA in the field circuit should be same as the base VA in the
amortisseur circuit now we do one similar exercise for mutual inductance between the stator
winding that is the stator direct axis winding and the field winding that is stator and field that is
Lafd, we want to make this as Lfda that we want to make these two per unit mutual inductances
equal right now these quantities were again defined when we wrote the equations for flux
linkages okay.
Now when we equate this two terms and follow the same procedure as we have done for the for
the mutual inductance between field and amortisseur. Okay therefore here again we come to very
important result and that result is that vibase, v is that is the efd base into ifd base that is the
vibase in the field circuit comes out to be equal to 3 phase VAbase for stator, 3 phase VAbase for
the stator that is in the stator circuit when you assume that the the base voltage as peak value of
the phase voltage, the the base current in the stator circuit is the peak value of the phase current
and we have established that this quantity is equal to 3 times VA, 3 times VAbase for the stator
that is 3 phase VAbase for the stator the total volt ampere rating right.
Therefore, this is a very important relationship because in the we have also established the
relationship that the base VA in the kd circuit direct axis amortisseur should be same as that in
the field circuit it means now we can conclude actually that these 3 windings one in the field
winding, another stator winding, third amortisseur winding, direct axis amortisseur winding.
Now if you want to make this mutual inductances reciprocal then the volt ampere rating or volt
ampere base in the stator circuit must be equal to the volt ampere base in the field circuit and volt
ampere base in the amortisseur circuit because in the stator with the the volt ampere base is equal
to the 3 phase VAbase 3 phase VA total MVA rating.

Okay this is exactly similar to what you normally do in any system we take total 3 phase MVA
as the base quantity and then we talk about the base for the rotor circuits we find actually that but
rotor circuit now if the field circuit if you consider it it has a applied voltage Vefd and the current
flowing is ifd therefore efdbase into base ifdbase becomes the this product becomes the VAbase
for the field circuit therefore this one condition will be established that if we want to make, we
want to make the mutual inductances between the circuits reciprocal then this is the criteria
required for choosing the VAbase for these circuits, okay.
Now this the similar exercises have to be done for for quadrature axis, stator winding and the
quadrature axis amortisseur because so for we have talked about the d axis and here also we
established the relationship and the relationship is that ekqbase into ikqbase is equal to 3 by 2
esbase into isbase not now this quantity is nothing but the 3 phase VAbase for the stator
therefore ultimately what is our conclusion that all the circuits state all the all the circuits that is
all which are in the rotor field circuit, amortisseur circuits right these would have their VAbase
equal to the 3 times or 3 phase VAbase of the stator circuit this is the and this will make all the
mutual inductances reciprocal.
Therefore, now I conclude here that in in order to satisfy the requirement that per unit mutual
inductances between again the emphasis is that we are trying to make the per until mutual
inductances reciprocal actual mutual inductances which exist, we have seen actually that they are
not reciprocal when you see the actual equations which we have written right they are they
appear in the equations in a different fashion between different windings the reciprocal okay and
the requirement is that VAbase must be same and equal to the stator 3 phase VAbase, this is the
main conclusion okay.
Now the next requirement which we have posed here is posed here is that the mutual inductances
on the same axis we want to make them equal first we make them reciprocal and then
requirement was there want to make them equal.
Now to make them equal we we define the per unit self inductance of the d axis winding, d axis
stator winding as leakage inductance per unit leakage inductance plus per unit mutual inductance
that is when you look at the d axis winding on the stator. Okay the total per unit inductance is
defined as Ld bar you know this bar stands for rotor super bar stands for per unit.
Now the total flux which is produced right which links the d axis winding does not pass through
the field circuit or the amortisseur circuits d axis amortisseur circuit there is some flux which is
the local leakage flux therefore this Ll is accounts for the inductance of the d axis stator winding
due to leakage. Okay therefore this is this becomes our mutual inductance Lad bar similarly, for q
axis.
Now here our requirement is that in order to make all per unit mutual inductances between stator
and rotor circuits in the d axis equal that is we have in the stator circuit a fictitious d axis winding
okay and the rotor circuit we have on the d axis two windings okay and we want to make them
equal for making them equal or this by definition this is actual value of mutual inductance Lad
divided by the Lsbase okay and this this is equal to Lafd which has which was earlier defined by
this expression when we wrote the wrote the mutual or we can say when we write the
expressions for the psi d psi q and psi naught right, the Lafd was defined like this per unit value of
Lafd was defined by this expression that you can refer to the pervious equations okay.
Similarly, now we want to make Lad to equal to the per until Lakd okay while Lakd was defined by
this expression. Okay now what we want that we equate these 2 as we did in the previous case
and when we equates these 2 we come to one very important relationship that the ifdbase should be
equal to Lad upon Lafd, isbase. Similarly, ikdbase comes out to be Lad upon Lakd divided by Lsbase while
these quantities are the actual inductances, actual mutual inductances right while what we trying
to make them equal are the per unit mutual inductances. Okay, now this this gives a relationship
that suppose I have chosen the stator base current then the base current in the field circuit should
be given by this equation.
Similarly, if I choose in the base current in the stator as is base then the base current in the
amortisseur circuit, direct axis amortisseur circuits will be given by this expression. Okay
similarly you can do for q axis also that is q axis will give you that ikq base should be equal to Laq
divided by Lakq, isbase I hope you understand this one is this is this is a straight forward okay, what
we are trying to do is that what we want to make them equal. We want to make the per unit per
unit the mutual inductances on the same axis d axis equal that is right, we want to make Lad per
unit equal to Lafd which is the definition of Lad same as Lfd we want to make this Lakd and Lafd
equal to make them equal you equate these quantities which was the basic definitions for this per
unit Lafd and per unit Lakd.
Okay, and we finally get the relationship that the field currents base value of field current base
value of the amortisseur circuit current and the base value of the quadrature axis amortisseur
therefore current are expressed in terms of the stator, the stator circuit base current okay
Similarly, we had also established that VA base for all the rotor circuits in terms of the 3phase
VA base of the stator and therefore since actually when I talk about the VA that the field voltage
base into ah field voltage current base okay.
Now this product is equal to the three phase VA of the stator current right but the individually
currents will be defined by this expression now once the current is current base is given for these
circuits the , we compute the voltage base for the rotor circuits using the previous relationship I
hope it is clear to all of you now once we have done this and define the the model for the
synchronous machine is complete, the synchronous machine model is complete we for all for all
further discussions what we will do is that this super bar which we had put to identify these
quantities as the per unit quantities we will drop it but we will always keep in our mind that these
are per unit quantities once these are per unit quantity because every time writing super bar is not
very convenient okay.
Now further we can express the per unit power and torque per unit power and torque the in our
previous derivations the terminal power, total terminal real power was expressed as 3 by 2 ed, id,
eq, iq plus 2 times eo, io. Okay that you can recollect now here what you do is that you replace all
these quantities by that is you divide by base 3 phase base VA of the stator here that is you
divide this by base volt ampere, you divide all these quantities by the base voltage and currents
when you do this exercise the power in per unit can be written as ed bar id bar plus eq bar iq bar
plus 2 times eo bar io bar.
Okay that is only thing which happens is that 3 by 2 term which appears in this expression
disappears in the expression for per unit generator power, okay other expressions are same
except that the per unit quantities this exercise I will just suggest you to verify, you substitute the
values and then you verify that you get these 3 by 2 is then we obtained the per unit electrical
torque or air gap torque which will come out to be equal to per unit direct axis flux linkage psi d
into iq minus psi q into id. In fact this this is the most important expression when we talk about
the synchronous machine model why is this expression is, so important can you tell me.
Other, other stator circuit rotor circuit equations are also ah equally important right because that
gives the model complete model we have to write down the stator circuit equations, rotor circuit
equations, stator circuit flux linkages, rotor circuit flux linkages. Okay and all these things have
to be written in per unit by establishing proper per unit base but when I say that the per unit air
gap torque is given by this expression and this is quite important, can you tell me why? yes
instead of calculating the torque in terms of three phase we are getting the torque just by 2
values. No, because the torque is computed in per unit using the per unit flux linkages and per
unit currents this expression is going to be used in our swing equation.
When you have written the swing equation in the swing equation, we have to write down
mechanical torque minus electrical torque and this this is the electrical torque that will come in
the swing equation okay.
I will just mention briefly about alternative per unit systems and transformation that is the dqO
transformation which we discussed till now and another form of dqo transformation which has
been discussed in the literature and also used by some people. Here, in this the the dqo quantities
are expressed in terms of this transformation matrix into the phase currents that is the phase
variables ia, ib and ic.

Now if you see this transformation matrix the main difference here is that the dqO
transformation which we used right this quantity or this term kd this is a kd was a kd or kq they
have both taken equal to 2 by 3. Here they are taken as square root of 2 by 3 and correspondingly
the the last row that is the third row the terms become instead of 1 by 2, it becomes square root
of 1 by 2 and so on.
Now with this transformation if you write the inverse transformation that is you write down the
phase currents ia, ib and ic in terms of dqo quantities right. Then this this this becomes the inverse
of our transformation matrix, this is the transformation matrix we use therefore if you invert this
matrix right then we can write down using this inverted matrix the relationship between the
phase currents ia, ib and ic in terms of dqO components. This inverse if you just see here then this
inverse is nothing but the transpose transpose of the transformation matrix.
You just see here actually that this row is the column here right while this row appears as column
in the inverse okay and therefore this is this inverse this is the inverse of the transformation
matrix but is also the the transpose of the transformation matrix. Now whenever you have this
type of situation we call this transformation as orthogonal transformation now when you resort to
this transformation one very important result which we get is like this.
The terminal power Pt right which when we write down as ea into ia plus eb into ib plus ec into ic
this is a terminal power okay. This Pt is average or instantaneous, instantaneous but what is
average?
The average and instantaneous become equal in a 3 phase system, 3 phase system the total power
is total or the average power is constant, instant we say instant is power we calculate this sum
will come out to be equal to average power right therefore now if you substitute here that is you
write this equation, that is this equation can be written as Pt can be written as ea, eb, ec multiplied
by ia, ib, ic is it not this expression is written you can put in the matrix form this is row vector
multiplied by this column vector this will give you ea, ia plus eb, ib plus ec, ic. Now you replace
this row vector by dqO components you replace this also by dqO components and multiply these
2 matrices when you do this multiplication the end result comes out to be of this form that is the
Pt will come out to be equal to ed, id plus eq, iq plus eo, io it is something like this that when I had
3 voltages and 3 currents okay I got the terminal power by multiplying this quantity.
Similarly, I have now the direct axis voltage quadrature axis voltage and zero sequence voltage
similarly the direct axis current quadrature axis current and zero sequence current I multiply and
add this comes out to be equal to right now for this this type of therefore this transformation is
called power invariant transformation, power invariant transformation and it has some merits
which has been discussed in literature but it has some demerits also and that is why in most of
the models which have been developed the dqo transformation which was discussed earlier is
used where we use kd and kq equal to 2 by 3. Okay having developed the synchronous machine
model you have developed is synchronous machine model completely in using per unit quantities
and the model is developed in dqo frame of reference okay.
Now this model is the complete model and now what we will study from this point onwards will
be first we will study the steady state aspect okay and then we will go how this model can be
simplified for our stability studies. Okay that is this is the complete model actually right but
sometimes we can make some assumptions and simplify the model okay and how do we simplify
this model which will be suitable for stability studies that is what we will study in the subsequent
lectures.
Now let us quickly do the steady state analysis because steady state analysis is one which is
known to everybody all of you know this analysis but now we will develop the steady state
equations for the machine starting from our dynamic model of the synchronous machine. Okay
now when we talk about steady state conditions we are presuming that the synchronous machine
is operating under steady state condition and the stator currents and voltages are balanced 3phase
currents and voltages that is this is the assumption that these steady state operating condition and
the stator and stator quantity stator currents as well as the stator voltages are balanced.

With this assumption once it is steady state this derivative terms will be absent, in the steady
state condition we do not have anything like d by dt of psi suppose, I take this psi will be
constant therefore this derivative term derivative term will be absent, second is that all all zero
sequence terms will also be absent because it is a balanced 3 phase system okay.
With this assumption the stator circuit equations can be written now in per unit again remember
that these are all per unit stator circuit equations the bar is not written here while the stator circuit
equations are written in the form ed equal to minus omega r psi q minus Ra id eq equal to omega r
psi d minus Raid and efd equal to Rfd, ifd it is very simple actually, you can just in all the 3
equations the p psi d terms have been p psi terms that is the derivative terms have been set equal
to 0 okay.
Now under a steady state operating condition the rotor speed is equal to synchronous speed
therefore, we substitute here omega r equal to omega s and in a per unit system we will see that
omega r is equal to omega s equal to 1. Okay now these are the 3 voltage equations these 2
voltage equations for the stator circuit this voltage equation for the field winding and these are
the flux linkage equations that is psi d and psi q in per unit quantities they are all per unit terms
we have not done anything except that except that in these equations the the derivative terms
have been set equal to 0 right. Another thing while writing here we have also set the mutual
inductances per unit mutual inductances on the d axis windings are equal right therefore, you will
find here actually when you write down this psi d is Lad ifd.
We will write down psi fd again I have put Lad because the mutual inductance is mutual
inductance between the stator d axis winding and the field winding have been made equal in per
unit in per unit system okay. Therefore, this simply appears as Lad okay. Further actually when
we have written these equations suppose there is one amortisseur on the d axis k becomes one
here, suppose you take 1 or 2 amortisseurs on the q axis one will, one equation will be for psi 1q
there is a second amortisseur will be psi 2q right and they will be written in the by the same
equation minus Laq iq why the same Laq will come because mutual inductances between the 2
amortisseur on the q axis, when expression per unit are equal.
Therefore, this is okay now here we can just derive one simple expression for field current what
we do is that in these equations just start with this. Suppose you take my interest is to write down
what will be the value of field current required or field current in per unit required to produce the
required steady state operating condition. So that using this expression you can write down the
expression for ifd that is ifd is written first here as psi d plus Ld id divided by Lad and then this psi
d that is the per unit flux linkages in the d axis will be replaced by the this equation psi d equal to
minus Ld id not not here actually this psi d because we will resort back to our stator winding
equations that is psi d can be written using this equation. Okay that is eq equal to omega r psi d
minus Raiq okay and when we make these substitutions here we get very important relationship
is this relationship that is the field current which is required to produce certain value of eq the iq,
id with these quantities known they are all the circuit machine parameters.
Now stator per unit stator resistance per unit direct axis inductance, per unit mutual inductance
between direct axis and the field winding these are all now circuit parameters or synchronous
machine parameters they are all known. Okay and therefore when system is operating with some
load the that load will determine iq and id, okay and to have then for a particular value of eq this is
the value of ifd required therefore for all our studies actually as we will see further when we talk
about the models of the synchronous machine for our stability studies this information is required
but this is simply obtained using those equations that is the stator circuit equations flux linkage
equations and making some appropriate substitutions.
Now we use this omega r is equal to omega s and this omega s into Ld we will represent by xd
omega s into Ld will be represented by xd right. Similarly, you can see omega s into Lad will be
represented by xad reactance omega L is equal to reactance always okay.
Therefore if this equation is written again exactly the same but instead of writing in terms of the
per unit per unit d axis inductance and mutual inductance which is written in terms of Xd and Xad
Xd is the direct axis direct axis reactance while Xad is the mutual reactance between between the
field and the d axis winding.
Now we will quickly see, now the phasor representation before we talk about the phasor
representation let us reinstate the some of the facts. Under steady state conditions the id, iq these
two currents these two currents are constant in magnitude therefore basically the current flowing
in the fictitious d axis winding and the fictitious q axis winding are dc current direct currents.
Similarly, the ed and eq they came out to be constant voltages that we have already established
under steady state operating conditions. Now the question arises is that the when we, when we
draw a phasor diagram right all the phasor are corresponding to a sinusoidally varying quantity
that is all the quantities which vary sinusoidally and having the same frequency right can be
represented in the phasor diagram. Okay but here also although these quantities are constant in
magnitude because of certain trigonometrical relations which exist we will be in a position to
represent this d and q axis voltage and currents as phasor but with the with the clear clear
understanding in our mind that d and q axis quantities are scalars.
\The ed, eq, id, iq they are scalars right but because there exists trigonometrical relationships
between the certain phasors and the d and q axis components therefore we can represent these
quantities in the phasor diagram. Now to start with let us say that we have a balanced steady state
operation the phase voltages are given by these equations. Okay that is ea is equal to Em cos
omega s t plus alpha alpha is some arbitrary angle a reference angle or may be to start with and
this eb is lagging by 2 by 3, 2 phi by 3,120 degrees ec by leading by 2 phi by 3. Now you apply
dqO transformation to this equation.
When you apply the dqo transformation you will get the d axis component ed and q axis
component eq as Em cos omega st plus alpha minus theta Em sin omega st plus that is in fact when
you apply this transformation right to arrive at this result we have to make use of many
trigonometrical identities that is when you when you when you put these equations right ea eb and
ec this this column vector in that transformation matrix.

You will find actually that this trigonometrical terms will have to be multiplied with the elements
of that matrix transformation matrix and finally to come to this result you may have to use
number of trigonometrical identities. Okay and the final result is going to be in this form okay.
Now we can represent the angular position theta as omega r into t plus theta o and omega r is
same as omega s because it is a steady state operating condition. Now when you make this
substitution ed comes out to be equal to em cos alpha minus theta o eq comes out to be equal to em
sin alpha minus theta. Okay and therefore in this equation when we find here in this equation the
there is no term like omega st right and therefore ed is constant it depends upon the peak value of
the stator phase voltage and the alpha and theta naught, okay that is why I said that this ed and eq
come out to be scalar quantities under steady state operating conditions.
Now ed and em the ed is d axis voltage and is in per unit. Okay now we are more more
comfortable when we use the RMS value of the voltage rather than using the peak value of the
voltage. Now suppose, I assume the peak value of the stator voltage as my stator base voltage
then RMS value RMS value of the stator base voltage when it is expressed per unit will come out
be same as the peak value. It is something like this that I have chosen the base voltage in the
stator circuit and that base voltage is peak value of the per phase per unit of the phase to neutral
voltage.
Okay now if I want to express ah any quantity in terms of this then they are expressed in terms of
the peak value. Now if suppose I have chosen the stator base voltage as peak voltage I want to
chose the the base voltage in terms of RMS value, then it is automatically going to be the base
voltage which is the peak value divided by root two and therefore if you express these quantities
that is ed and eq not in terms of peak values but in terms of RMS value the equation will remain
same because the base values of the stator voltage, when it is expressed as a in terms of peak or
where it is expressed in terms of RMS value they will have a relationship of one by root 2.
Suppose I take the stator base as 100 volts then this is a peak value then RMS value, RMS base
value will be equal to 100 by root 2 and therefore this equation can be written in terms of the
RMS value now I put et, et as the RMS value and it remains same there is absolutely no change
this is also per unit, this is also per unit okay.

Now when we represent these quantities in the phasor diagram it will look like this. Let us say
this is the terminal voltage phasor I represent this Et tilde putting this as a phasor quantity
because we use bar earlier super bar earlier for representing the bar per unit quantities because
that is why we are putting Et tilde and let us show that this is my d axis, this is the q axis, q axis
lead d axis by 90 degrees and here the angle is alpha minus theta because you have seen here
actually that the angles comes out to be alpha minus theta o right. Then, then ed is represented by
Et cosine alpha minus theta o and eq is equal to et sin alpha minus theta o therefore, this phasor
diagram, phasor phasor diagram is expressing these quantities ed, eq in terms of Et.
Similarly, you can draw a phasor diagram showing the currents. Let us say this is my terminal
voltage phasor so this is the current phasor and let us say that phase difference between the
voltages and current is phi right.
Then we can express express the d and q axis components of current as Et, I am sorry current
can be expressed as id equal to It sin delta i plus phi iq is equal to It cos delta i plus phi. This
exercise is done similarly to what we did for voltages right therefore in the phasor diagram I can
simply show It equal to id plus j times iq okay and the diagram I have just now shown to you,
okay.
Now here the angle between the q axis and the terminal voltage we will denote by delta i. Okay
now one small exercise which is important exercise to be done is done is to establish the position
of d and q axis with respect to the terminal voltage that is I know the terminal voltage and with
this terminal voltage as phasor I have to establish what is the position of q axis what is the
position that is if I establish the position of q axis I can automatically get the position of d axis,
now to establish this, now the some of these are equations are all repeated that you know the ed is
given by this expression, we have already established.
To establish the relationship what we do here is we will define a voltage eq, a phasor eq as
terminal voltage plus Ra plus j times Xq into It. This is an this relationship is a very well known
to all of us right that is just start like this that my my intension is that to establish relationship
between the terminal voltage and the d and q axis, we define a voltage eq as Et plus this
impedance Ra plus j times Xq into It.
With this definition and we substitute the expression for ed Et as ed plus j times eq and It as id
plus j times iq and we make simplifications that is you multiply and simplify this expression. You
will find that eq comes out to be j times Xad, ifd minus Xd minus Xq id what this what does we get
from this equation, what this equation conveys to us. This equation conveys that this voltage eq
which I have assumed right is in quadrature with d axis that is the axis this is along q axis not
part of it it is of course quadrature of d axis but it is along q axis because our reference with
reference to the reference this quantity is I had by 90 degrees that is j term is existing because the
other terms are all scalars here actually the terms are all scalar.
Okay and therefore the phasor diagram which I draw here showing this Et plus Ra It plus Xq It
this gives a voltage Eq that is with respect to this terminal voltage Et the position of Eq comes out
to be along q axis right and therefore we have now established a relationship between between
the terminal voltage and q axis and once we establish relationship d axis position is known and
all further computations can be done the moment we know the dq and other quantities. Friends I
will conclude my presentation here by by summarizing what we have done in this lecture. First,
thing that we have established the base volt amperes for the rotor circuit.
We have also established the how to choose the base current for the rotor circuits in order to
make the mutual inductances reciprocal and also to make the mutual inductances on d axis and
mutual inductances on q axis equal. Okay then we have also established the steady state relations
and developed a simple phasor diagram to show that, to show that the the q axis can be obtained
by obtaining the position of a phasor which is equal to terminal voltage plus plus an impedance
equal to Ra plus j times Xq multiplied by the current that is if you compute this quantity eq equal
to terminal voltage plus impedance that is Ra plus j times Xq multiplied by the current this gives
me the position of q axis, okay. Thank you!
Prof. M. L. Kothari
Lecture - 11
Modeling of Synchronous Machine (Contd.)

Friends, today we will continue our discussion on modeling of synchronous machine,
okay. We will continue on the steady state analysis and then we will talk about the
magnetic saturation. Now while discussing the steady state analysis we had discussed
earlier that a voltage, a voltage behind voltage behind the impedance Ra plus j times Xq
lies along the q axis and that gives us the location of q axis with respect to terminal
voltage.
Once we establish the location of q axis we are in a position to obtain the d and q
components of voltage and currents which are required, okay.

Now here this is the phasor diagram which shows the that voltage Eq bar or Eq is along
the q axis what we have done is that this is the terminal voltage, this is the load current or
terminal current, the phase difference between the terminal voltage and the current is phi
to this we are adding this voltage drop Ra It and adding the voltage drop Xq It this gives
you Eq.
Now this voltage Eq is along the q axis and the angle between Et terminal voltage and this
Eq that is the q axis right this is denoted by the angle delta i. Now we will establish a term
which is a rotor angle or the power angle, although we have mentioned right now that the
power angle is the angle between the q axis and the terminal voltage.
Now we will establish actually that in case there is no load on the machine that is
machine is under no load condition right then the terminal voltage, terminal voltage Et is
along the q axis and therefore when the machine is loaded right Et falls back with respect
to the q axis right now to establish this we start with our original equations they are the
stator voltage equations that is ed equal to minus omega r psi q minus Ra id eq equal to
omega r psi d minus Ra iq and the field voltage efd is equal to Rfd ifd while the flux
linkages are minus Ld id plus Lad ifd psi q equal to minus Lq iq and the flux linkages in the
field winding is psi fd is equal to Lffd ifd minus Lad id.
In the amortisseur on the d axis psi 1d equal to minus Lf1d, now we are putting
amortisseurs is identified as 1, K equal to 1 ifd minus Lad id psi 1q equal to psi 2q. I have
we have assumed here that they are 2 amortisseurs on the q axis minus Laq iq. Now when
you consider the system under no load condition right under no load condition there is the
id and iq will be 0. Okay now once id and iq are 0, you can see here psi q will be 0. Okay
and once psi q is 0 and id is 0 therefore what you see here is that ed become 0.
Okay now with this substitution we write here the under no load condition the d axis flux
linkage is equal to Lad ifd that is in these equations you substitute id, iq and ed equal to 0
and psi q is also 0 therefore, we when we make the substitution we find here that psi d is
equal to Lad ifd that is the direct axis flux linkage is directly proportional to field current
this is very important relationship under no load conditions psi q is 0 ed is 0 eq is equal to
Xad ifd that is the q axis component of the terminal voltage is directly proportional to the
field current. Okay now we know the terminal voltage Et is equal to ed plus j times eq
right that is what we established earlier. Now since ed is 0 ed is 0 and eq equal to Xad ifd.
So that Et comes out to be equal to j times Xad ifd that is in the no load condition the
terminal voltage is in quadrature with the direct axis that is j is here and is proportional to
the field current ifd right.
Now this is all known to all of you because under no load condition the terminal voltage
is directly proportional to the field current excitation and this voltage is called excitation
voltage under no load condition the terminal voltage is the excitation voltage. Now we
can establish a very simple equivalent circuit under steady state operating conditions for a
round rotor synchronous generator that if you consider the round rotor synchronous
generator then Xd and Xq are equal and this we will denote as Xs synchronous reactance,
Xd equal to Xq equal to Xs for a round rotor right.
Now when you substitute the value of substitute the value of Xd equal to Xq equal to Xs in
the equation for Eq because Eq is now becomes Et plus Ra plus j times Xs It, we have seen
earlier that Eq is equal to j times Xad ifd minus Xd minus Xq id now the moment the
moment the id is 0 id is 0 Eq is equal to j times Xad ifd right. Now this is what we are
trying to establish here that if you do not neglect the armature resistance R a will come
right and Xd Xq neglecting the leakage reactance right Xad becomes Xs right. Otherwise,
Xd is equal to Xq equal to Xs while Xad is the mutual reactance okay.
Therefore, we can write draw a simple a simple equivalent circuit of the synchronous
generator where terminal voltage is denoted as Et angle 0 and the excitation voltage as Eq
angle delta i and we put an impedance Ra plus Ra and Xs that is impedance is Ra plus j
times Xs that is a very simple equivalent circuit which is obtained and where this we have
establish that this Eq the magnitude of the Eq is equal to Xad Ifd that is called excitation
voltage, a very simple expression. Now we can derive the expression for the complex
power in terms of d and q axis components okay, complex power S is equal to Et into It
star It star that is I star stands for conjugate now when we substitute the expression for Et
and It right in this expression, we get S equal to ed id plus j times eq iq plus j times eq id
minus ed iq. Therefore, this term is identified as the real power and this term is the
reactive power.
So that we can write down here that the real power at the terminal of the machine is e d id
plus eq iq that is you know the dx is component of voltage the dx is component of the
current qs is component of the voltage qs is component of the current right you multiply
the corresponding voltage and currents and add them you get Pt. Similarly, you obtain the
expression for Qt as eq id minus ed iq. Okay now we will establish one very interesting
expression for torque because we have seen that the expression for torque is psi d iq
minus psi q id, this expression is the very general expression which we have established.
Now in these expressions we substitute the expression for psi d and psi q from the
previous relationships established relationships and then simplifies this equation. You
will find that this torque Te that is a air gap torque will come out to be equal to the
terminal power plus Ra into It square or we can say that I square R loss what we what do
we understand here.
Now since in per unit quantities power and torque are equal when we use the per unit
quantities then the torque is same as the power. Now this torque is the air gap torque or
we can and this air gap torque can be obtained as the terminal power plus the armature
resistance loss right and therefore whenever you performed stability steady analysis this
is one computation which is required to be performed that first we find out what is the
terminal power then to that terminal power we add the armature resistance loss okay and
that gives us the electro magnetic torque and that is what we will be substituting in the
expression for swing equations that is in the swing equation.
We have to substitute Te the Te is computed using these expressions and and under
steady state conditions right the mechanical torque is equal to the electrical torque
therefore mechanical torque is computed by this expression right while electrical torque
will be will continue to be computed by this expression because under dynamic
conditions this is the expression for electromagnetic torque under dynamic conditions for
under steady state condition we can say that the steady state torque is equal to terminal
power plus losses. This is a very important step in case you neglect the armature
resistance then terminal power is same as the electromagnetic torque but today with the
availability of computing facility right and the capability of the digital computers there is
no need to neglect the armature resistance because armature resistance losses are
substantial okay and they have bearing on the stability results.
Now, we come to the very important aspect of modeling of the system. So for, so for we
have neglected the saturation in the magnetic circuit. This was neglected to simplify the
mathematical model. However, however the magnetic saturation need to be accounted
and if we neglect it then the errors in the results are sometimes substantial and therefore
there is a necessity to understand how do we account for the saturation in the magnetic
circuit. Now to understand how do we account for a saturation in the magnetic circuit, we
will look into the open and short circuit characteristic of the synchronous generator all of
you are aware of open circuit characteristic of the synchronous generator short circuit
characteristic of the synchronous generator and we will see that the open circuit
characteristic is very useful in determining the saturation characteristic of the machine
okay.
Now we write under no load conditions we established earlier also id iq psi q and ed these
four terms where 0 when the machine is not loaded right the id and iq components of the
current are 0 psi q is 0 and ed is also 0 this is what we have established in our previous
discussion. Then the terminal voltage Et came out to be equal to Lad ifd that is this
terminal voltage is directly proportional to the field current. Okay and this terminal
voltage is also called excitation voltage under no load conditions under no load condition
this is also the excitation voltage now let us look at the at this characteristic.
On this axis I am I have shown the field current ifd and on this axis we are showing the
open circuit terminal voltage in per unit. Okay when you perform the open circuit
characteristic or perform the experiment to obtain the open circuit characteristic what we
do is we run the machine at rated speed okay and note down the terminal voltage for
different values of the field current.
Now here while plotting this characteristic I have neglected the residual flux, okay
otherwise there is some residual flux and you will get that even if the field current is 0
there will be some that is that has been neglected here. Now the the open circuit
characteristic will look like this. Okay that is it remains straight line and then starts
getting saturation, now if you draw a tangent to the open circuit characteristic in the
initial portion right then we get a straight line, we can draw a straight line which is
tangent to the initial portion of the OCC, we call this air gap line.
Now here here we will define two field currents that is if the terminal voltage is one per
unit, the field current required when you consider the saturation is IfNL. In case we
ignore the saturation then the field current required is IfNL (ag), ag stands for that or the
air gap line considering the air gap line. Okay now if we plot short circuit characteristic
that is on this axis I am putting short circuit armature current in per unit there is again the
same field current. This short circuit characteristic comes out to be a straight line while
you perform the short circuit characteristic the machine is run at rated speed and the
terminals of the machine are put short shorted okay.
Now there is no saturation when we are performing short circuit characteristic that is you
consider actually the short circuit current right for different values of the field current
even beyond the rated current, you will find there is hardly any saturation. The basic
reason for for the short circuit characteristic comes coming out to be straight line can be
attributed that under short circuit condition right the armature current is having the
demagnetizing affect okay and because for the demagnetizing affect the resultant
magnetic field which is produced in the air gap or in the iron core is very low and
therefore saturation does not take place and this is very important point.
Now here we can now define unsaturated reactance of the machine and then we define
saturated reactance of the machine. Now to define this that is what we do here is that
when the synchronous machine, when the synchronous machine terminals are shorted and
let us say it carries one per unit current right at that time whatsoever the induced emf due
to this field current is used in overcoming the drop in the short circuited path.
If we neglect the armature resistance armature resistance then we can say that the induced
emf is a some constant times Ifsc equal to 1 into Xs unsaturated okay that is there under
short circuit condition there is no saturation in the system right therefore the reactance
which we talk about the unsaturated reactance and the voltage induced is some constant
time in the field current okay, then if we see here the open circuit characteristic then one
per unit voltage is produced when the current is equal to IfNL okay.
Now being the proportionality constant same that is K IfNL ag equal to one right
therefore I using these 2 equations we can establish that the unsaturated value of
synchronous reactance is equal to IfSC divided by IfNL (ag), this is a very important term
that is what is the value of unsaturated synchronous reactance or unsaturated value of
synchronous reactance that is Xs unsaturated is obtained using the open circuit
characteristic open circuit characteristic and short circuit characteristic that is on this
short circuit characteristic we draw we find out what is the field current required to
produced one per unit current in the armature under short circuit condition.
Similarly, we find out how much is the how much is the current which is required to
produce one per unit voltage under on air gap line again this is I am putting a air gap line
therefore this ratio of IfSC upon IfNL the ratio is of IfSC and IfNL (ag) gives you a
saturated unsaturated value of synchronous reactance. Okay now when we look at the
saturated value of the synchronous reactance.
Now we can find out the value of saturated synchronous reactance that to produce one
per unit voltage under open circuit condition the field current required is IfNL, okay this
IfNL ag was on air gap line while IfNL is on OCC right therefore, since open circuit
characteristic there is some saturation taking place therefore the field current required is
IfNL and following the same approach as we have done earlier now we can write down
that saturated value of the synchronous reactance is IfSC divided by IfNL. Normally, we
perform the open circuit and short circuit test to find out the synchronous reactance of the
machine the synchronous reactance which we compute by doing open circuit and short
circuit test are the will give you the saturated value of the synchronous reactance okay.
Now our task is that to relate to relate the unsaturated value of synchronous reactance to
the saturated value of the synchronous reactance. Okay now to relate this before I relate
this one more term which is very commonly used I am just introducing that is the short
circuit ratio, the short circuit ratio is defined as IfNL divided by IfSC and it comes out to
be the reciprocal of reciprocal of saturated synchronous reactance right the if the if the
since short circuit ratio is low that is what that synchronous reactance is high and in fact
in industry we very very frequently used this term short circuit ratio and it also gives the
information about the quality of the machine.
Now for representing the magnetic saturation for stability studies we make some realistic
assumptions the assumptions made are that the leakage inductances are independent of
saturation when we talked about a various inductances, we found mutual inductances as
well as the leakage inductances right. For example, Ld direct axis inductance Ld is equal
to Lad plus LL, LL is the leakage component okay and Lad is La ed is due to the flux which
which crosses air gap goes through the field winding and again return backs to the
armature okay.
Now since this leakage flux is mostly takes path through the air right and it crosses the
iron but but it is but it is it does not contribute toward the saturation of the iron right
because the reluctance to the leakage flux path is mostly due to the air and therefore we
make this realistic assumption that is the leakage inductances are independent of
saturation, second is the leakage fluxes do not contribute to saturation of iron cores that is
the same thing because leakage fluxes do pass through the iron core but we assume here
that do not contribute to the saturation. This is an assumption the realistic assumption but
not very very strictly they given but acceptable for stability studies.
This is another very important is that saturation relationship between the resultant air gap
flux and the mmf under loaded conditions the same as under no load conditions. Now this
is very important assumption here that we will establish the saturation relationship using
the open circuit characteristic right while while we assume here that that this relationship
is also applicable under loaded conditions.
Okay that is the saturation relationship between the resultant air gap flux and mmf under
loaded conditions is same as under no load conditions, okay that is why the open circuit
characteristic is very useful and used for determining the saturation characteristic okay
then another is the there is no magnetic coupling between d-axis and q-axis as a result of
nonlinearities between the nonlinearities introduced by saturation because here here
under linear conditions we assume that the d axis and q axis are magnetic circuit do not
have any coupling okay and even if there is nonlinearities now in the magnetic circuit, we
assume that that coupling does not exist even under nonlinearities this is what is the
meaning of this. Now with this assumption we will now concentrate upon the these terms
mutual inductance that is Lad and Laq, this now this Lad is the mutual inductance
considering saturation and we express Lad their mutual inductance with saturation is equal
to Ksd times mutual inductance unsaturated, u stands for unsaturated Lad u and this term
Kad is called direct axis saturation coefficient Ksd.
Okay similarly for quadrature axis we can write down Laq is equal to Ksq Laq Laqu that u
again stands for unsaturated value of mutual inductance and this Ksd and Kaq are the
saturation coefficients but our basic task is to obtain this coefficients. Okay now here
again we look at the open circuit characteristic.
Now in this open circuit characteristic on this axis that is on y axis, we are putting the
voltage open circuit voltage or flux linkages because we have established that the open
circuit voltage is proportional to the flux linkage okay and on this axis you can put field
current or the mmf. Okay, now we see two important things here that for a given field
current I right, if I consider the saturated characteristic that is open circuit characteristic I
find that the total air gap flux produced is psi at, okay if you neglect the saturation then
the for the same field current the flux produced is psi ato is not because this is a line this
is air gap line okay and we are operating at this point the field current is say I, okay for
the same for this field current the flux produce is psi at when you consider the OCC and
the flux produced is equal to psi at O if I consider the air gap line right.
The saturation coefficient Ksd is defined as psi at divided psi at O this this saturation
coefficient on direct axis or direct axis saturation coefficient Ksd is equal to psi at that is
the flux linkage corresponding to current I on open circuit characteristic and psi at O is
the flux linkage on air gap line for the same current and you can see here very carefully
that this ratio is same as psi O, Io by I and it is going to be always less than 1.

The saturated value of the inductance is going to be less than the unsaturated value of the
inductance therefore we can see here actually that whether I take the ratio of psi atO to
psi at or psi at to psi ato or I take ratio of Io to I they will give the same result okay. Now
we define here difference that is psi atO minus psi at this difference flux linkage is
designed by psi I and this psi I is very important to determine the saturation characteristic
of the machine or we define psi I as psi atO minus psi at and Ksd can now be written as
psi at divided by psi at plus psi I that is i replace this psi at O by psi I plus psi at right
therefore the major exercise will be that for a given value of psi at, we have to find out
what is the value of this psi I.
So that we can compute the saturation coefficient Ksd now here, now we have to
basically model this OCC open circuit characteristics to be modeled that is we have to
develop a model for the open circuit characteristic once we I know this this is the model
of the open circuit characteristic I can always find out find out this psi I for a given value
of psi at okay.
Now for modeling if you look at this characteristic the initially this characteristic is a
straight line. Okay then we will find actually that from then this characteristic is
practically following exponential shape and then once it gets fully saturated again it will
it is going to become a straight line, if you carefully examine the open circuit
characteristic right then you can divide this open circuit characteristic into 3 segments
that is the first segment is one where the flux linkage is proportional to the field current
that is in the unsaturated condition, second segment yes the the iron core starts getting
saturated and the characteristic is a non-linear one.
Okay, then in the third segment this characteristic becomes a straight line under saturated
condition what happens is that when you increase the mmf the air gap flux is going to
increase right but this increase is going to be linear relationship again therefore, for
modeling the open circuit characteristic or magnetization characteristic of the machine
we divide this characteristic into 3segments and these 3 segments are unsaturated
segment, non-linear segment and third is fully saturated linear segment.
This is the new thing which you have to very carefully understand okay now these 3
segments are shown here and psi T1 is the boundary value of flux linkage for the segment
one psi T2 is the boundary value of the flux linkage for second segment psi T2 and
beyond this it is the linear linear characteristic okay.
Therefore now what we do is that whenever you to try model right we this characteristic
is available to us experimentally you obtain the characteristic whichever way you obtain
it, this characteristic is known to you. Okay, now I can define here in this linear portion if
I take one per unit value on the on the x axis that is I take this as a x axis and I call this
vertical height as L i n e r incr incremental L incremental right then I can write down the
slope equal to the slope of this characteristic can be written as iLncr that is this is the
straight line characteristic with slope you have to obtain. Okay our the these 3 segments
can be model very easily.
So for actually the segment one is concerned segment one is concerned the psi at is less
than equal to psi T1 which is the boundary for segment one right and therefore the this
psi I term is 0, now this the actual OCC and the air gap line they are coinciding right and
therefore the psi I which is difference between psi at O and psi at is 0. The second
segment is a non-linear segment okay.

The boundary values are psi at should be greater than psi T1 and less than equal to psi
T2, okay this is the range we have already established for the second segment that is the
second segment is the psi at anywhere on this line right this is going to be in this range
psi T1 and psi T2. Now this can be modeled by a the equation of this form the
exponential form of equation the psi I can be written as psi I can be written as Asat that is
A saturation this is a coefficient this the subscript sat is used to indicate that we are trying
to model the saturation that is psi I equal to A saturation e to the power B saturation
multiplied by psi at minus psi a T1 what is this psi I, this is the difference between the psi
atO and psi at and this psi I can be model by this expression where you for any value of
psi at psi at you find out this difference psi at1 psi T1 is known to you and these
coefficients have to be obtained, these coefficients have to be obtained actually for a
given machine okay.
Once these coefficients are known psi I can be computed. Now here if you see that if I
put psi at equal to psi T1 right that is actually at the beginning of the segment psi at is
equal to psi T1, you can again look here at this point right at this point psi I is 0 right but
when I use this equation psi will come out to be equal to Asat, this is a slight discrepancy
at this boundary. However, in practice this term psi at is small right and therefore this
discrepancy is not very significant because when I use this equation and try to put
actually the initial value that is psi at equal to psi T1 right then psi I should come out to
be 0 but it is not coming out to be 0 by this equation right.
However since A sat this coefficient is generally very small and therefore this
discrepancy is a negligible or can be ignored because again let me emphasize here that
we trying to develop the model okay and for this developing the model we have to very
carefully obtain these coefficients because when I said is exponential I have to do curve
fitting okay and once these coefficients are obtained we obtain actually the model for psi
I. Now the last segment which is again a straight line fully saturated segment third
segment the fully saturated one okay.
(Refer Slide Time: :42:45)
Now here we have to find out find out what is the value of psi I when psi at is beyond psi
T2 right now to obtain this what we really do here is we may have to write down the
equation for this segment is a straight line, you can write down the equation of the
segment in the form of y equal to mx plus c a straight line okay and then you know this
equation for this line what is the equation for this air gap line, you simply write down psi
at equal to Ladu into ifd because this is a path line passing through origin while this line is
not going to pass through origin therefore, you have to make use of the initial conditions
that is actually that our initial conditions are that when the flux is equal to psi T2 right psi
T2 the field current is ifdo something, you can note it down that is because here this is the
segment which we are trying to model right.
Now you do this exercise look this exercise you can do and obtain the expression for the
psi I which is which is the difference between actual psi at and psi atO right. Now psi I
has been obtained in this case as psi g2 plus L ratio into psi at minus psi T2 minus psi at.
Now this I will suggest all of you to derive the expressions psi I equal to psi g2 plus L
ratio psi at minus psi T2 minus psi at, I have already checked it, it comes out to be okay
there is no discrepancy.
Now here what is done here is that when the flux linkage is psi T2 right we find out
actually the value of flux linkage on air gap that is for the psi T2, you note down this
current and for the same current you can find out a flux linkage which is on air gap line
corresponding to the air gap line that is called psi g2 and using this information that for
this particular current this point is lying on the third segment characteristic right and
using this value that the flux linkage is psi g2 therefore basically at this point at this point
psi I is equal to psi g2 right.
Therefore if I substitute here in this equation in this equation right if psi if I take this psi
at if I take this psi at equal to psi T2 and this term become 0 right and psi I becomes psi
g2 minus psi at, no no psi psi T2 is nothing but psi at right therefore this equation is
satisfied. Okay and this can be derived now once you are in a position to obtain this
quantity psi I the coefficient Ksd is known okay now next point is how to obtain the
saturation coefficient for q axis okay.
Now, so far actually d axis is concerned we have OCC for q axis we do not have similar
open circuit characteristic we cannot plot it. Now here generally we made one assumption
that in the q axis the saturation is 0, a very realistic assumption that you can make this
assumption that on q axis that is the saturation coefficient Ksq is equal to 1 because the
magnetic the flux takes its path, so larger air gap when you talk about the q axis of the
machine, okay.
Now when we are talking about this the flux psi at is equal to psi ad square psi aq square
that is the flux which we have been talking about right is the total flux in the air gap psi at
equal to psi ad square plus psi aq square.
Now here we will just establish one very is simple relationship using our expressions for
psi ad and psi aq, psi ad and psi aq these are the expressions which were known to you
the relationships are that is you write down the expressions for psi at that is you write
down the expressions for psi ad which is then psi plus Llid and psi aq as psi q plus Lliq.
Okay and psi d can be written by this expression psi d psi ad is equal to eq plus Ra iq psi q
is written as minus ed minus Raid these are the equations which we have already
developed earlier and now if you substitute this in the equation for psi at right, you will
find actually that psi at when you express in per unit because we are all doing per unit
calculations this psi at will come out to be simply as a voltage Ea Et plus Ra plus j times
Xl into It that is this psi at is a voltage behind psi at is a voltage behind or is equal to the
voltage behind behind this impedance that is Ra plus j times Xl that is Eq when we are
talking about that was the voltage behind Ra plus j times Xq. Okay now here it is only
talking about the leakage reactance this is this is established by substituting these
expressions. Now with this we have established how do we obtain the saturated value of
the reactance is required for calculations.
Now if you see it very carefully here then for each operating condition you know for each
operating condition, you have to find out what is the saturated value it is not same for all
operating condition this coefficient Ksd is different for different operating condition you
can easily see here that psi at when I say psi at is obtained corresponding to one particular
operating condition. Okay and for that we have to find out psi atO, okay and this ratio psi
at upon psi atO gives you saturation coefficient and this psi at is different and therefore
we get the different value of saturation coefficient and that is why, when you perform the
stability studies stability studies right the saturation has to be computed as the loading
condition keeps on changes for different loading condition different value of saturation
coefficient is to be used. With this I am concluding here the modeling of synchronous
machines for stability studies. Today, we have talked about the method of modeling,
modeling the saturation right.
How how it has been done basically we started with the basic definition of what we mean
by saturation coefficient and then the OCC open circuit characteristic is divided into 3
segments and for each segment we have written the expression for the flux linkage psi I
which is the difference between psi atO and psi at that is psi atO is the flux linkage for a
given field current on air gap line right and then we compute the value of saturation
coefficient. Further, we have assumed here the saturation coefficient on q axis as 1that is
saturation is neglected and this can be considered with this.
Let me say that we have completed the synchronous machine modeling this modeling is
complete in the next chapter we will study the synchronous machines model for stability
studies. Okay when we talk about the actual model of the synchronous machine we are
considering here now the stator circuit, rotor circuit transient while in the stator circuit we
have we have made the assumption that the derivative terms are negligible as compared
to the speed voltage terms therefore, this assumption will be carried over for a stability
studies also that is when we have written the stator circuit equations we have ignored the
transformer voltages and therefore this assumption will be carried over beyond for
stability studies also.
However, there will be some more assumptions that can be made and simplified to obtain
are made to obtain the simple synchronous machine model for stability studies. Now
when we talk about the modeling the next step will be to model the excitation system and
the models for turbines and governors that makes the complete model for the energy
system that is turbine, governor, excitation system and synchronous generator that is the
complete thing, we will devote few more terms to develop the models for the excitation
system and this okay. Thank you very much.
Prof. M. L. Kothari
Lecture - 12
Synchronous Machine Representation for Stability Studies

Friends, today we shall study about the synchronous machine representation for stability studies.
In this we will study the simplified model with amortisseurs neglected and a constant flux
linkage model. Now here, before I discuss these simplified models let us see the per unit system
of equations which we have developed so far. I will just summarize the per unit system of
equations which we have developed for the synchronous machine.

The these 3 equations that is the these represents the stator circuit equations in dqo frame and all
these terms are expressed in per unit all the currents, flux linkages, the voltages are expressed in
per unit, the time is also expressed in per unit. The second set of equations are the per unit rotor
voltage equations. Now here while writing the per unit rotor voltage equations I have considered
one amortisseur on d axis and 2 amortisseurs on the q axis. The field winding circuit voltage
equation is efd equal to p psi d plus Rfd ifd. Similarly, the equations are written for amortisseurs
right 0 equal to p psi 1d R1d, I1d. Now here I put this substitute one to indicate that this is the
there is 1, 1 amortisseur on the d axis or amortisseur number 1on the d axis right if there is a 2
that will become second winding also on d axis okay 0 equal to p psi 1d, psi1q plus R1q, i1d and
similarly the last equations.
Then in these equations which we have written, we need the expressions for stator flux linkages
therefore another set of equations are the stator per unit stator flux linkages, linkage equations
per unit stator flux linkage equations. These are written as psi d equal to minus Lad plus Ll id Lad
ifd plus Lad i1d. Now here if you see here this Lad is the mutual inductance between the stator and
field winding, okay. Now in the per unit system of representation, we have made all mutual
inductances on the d axis equal so that you will find that this Lad appears here.
Similarly, this term in the this term represents the mutual inductance between direct axis
amortisseur and d axis of the stator winding right therefore, again this comes out to be same
similarly, similarly for quadrature axis windings, we have flux linkage psi q equal to minus Laq
plus Ll iq plus Laq i1q plus Laq i1, i2q okay if again these are all mutual inductances are made are
equal the last equation is the flux linkage due to the zero sequence currents.

Then we have equations for rotor flux linkage equations which which are expressed in per unit
rotor flux linkage equations okay now again if you see here again we find actually that in the
rotor circuit also we get the these terms Lad Lad okay.
Now we have to see the last equation is this per unit air gap torque Te is equal to psi d iq minus
psi q id this is the per unit torque in the synchronous generator. Now this is the complete set of
equations for representing a synchronous machine okay and if you want to study the dynamics of
the synchronous machine then this set of equations plus the swing equation have to be solved and
you have seen that these equations are non-linear differential equations, okay and these equations
can be solved by applying numerical technique. Now in case actually the system is small right
then we can easily solve these equations and obtain the complete solution but when we talk about
the power system stability problem where we may come across hundreds of generators thousands
of buses a realistic power system which we have today even in our country right has large
number of generators and large number of buses.
Therefore the therefore the dimension of the system is very large and for analyzing the stability
of the system right if we use this detailed model which we have developed while developing this
model we have not done any approximations, no simplifications were made further actually I
would just emphasize that in these equations which we have written right the the all inductances
are are the saturated values of these inductances that is you account for the magnetic saturation
also while computing the inductances okay.
Now if you want to use this equations for stability studies it becomes very very difficult because
the system becomes very large and for large system the number of equations become very large
therefore, when we when we look in terms of simplifying the system model we always look from
the point of view of reducing the dimension of the model that is we want to reduce the number of
the differential equations, second is that we would like to have numerical technique which can be
used to solve our these equation quickly that is the fast numerical techniques and low order
model of the system. Okay these are the main thrust looking from that point of view we shall be
studying actually today that how we can make some approximations, simplifications.
Now what we will do that we will consider some various degrees of approximations okay that
can be made to simplify the model minimizing the data requirement and the computational effort
many times we have a problem also that ah certain datas which are required to use the model are
not available, one example is that normally the synchronous machine data related to the
amortisseurs are very are not easily available right and therefore we have to see that what
approximation can be made, so that computations can be fast or computational efforts are
reduced as well as the datas which are not available right can be or we can do the studies
without using those datas or getting some approximate datas.
Now while looking at these simplifications, the first simplification which has been looked into is
that we neglect this p psi d and p psi q terms in the stator circuit equations this is the first
simplification. Now let us see what is the justification for making this simplification or
neglecting the p psi d and p psi q terms, these is terms are the transformer voltages. Okay earlier
I talked about these equations and mentioned that these transformer voltages terms are smaller as
compared to the speed voltage terms right but if you see this equation this stator circuit equation
only because of this transformer voltage terms these equations are differential equations right
and if we neglect this transformer voltage term or the derivative terms right then the stator
voltage equations become algebraic equations in fact these terms represent the stator circuit
transients, okay.
Now I will just discuss what is the justification for neglecting these terms? Now the stator circuit
is connected to the transmission network right. Now for so far the transmission side or you say
the transmission network transients are concerned. Okay there is a transmission network
transients the time constants of this transient is generally very low and whenever whenever some
transient phenomena occurs in transmission system it quickly dies time constant is very small
right and therefore for system stability studies, we ignore the transmission system transients. We
assume that the transmission system is in steady state.
We consider only the sinusoidally varying voltages and currents in the transmission system
although whenever disturbances occur in the system like say fault we will find actually that when
the fault occurs it will have decaying DC components of the fault current it will have a high
frequency currents which are also adjusting in the during the fault system okay and but because
of the the low time constants, we always ignore the transmission network transients. Okay now if
we ignore the transmission network transients can we consider the stator transients because they
are part of the same circuit.
Therefore, if you consider the stator transients it becomes inconsistent and therefore therefore for
all system stability studies this assumption is made then the question arises whether this
assumption is a realistic assumption whether the results which we will get by ignoring these
terms will be acceptable, people have done lot of research on this and after performing a detail
studies the conclusions drawn are that yes we can ignore these terms that is the the stator circuit
transients can be neglected.
I will suggest you to look into the discussion given by P. Kundur in his book on power system
stability and control, power system stability and control. In this he has discussed the results
reported in some research papers and compared the the dynamic performance of the system
including including the stator circuit transients and excluding the stator circuit transients and
compared, okay and once you read this you will, you will come to a very important influence
that yes for stability studies one can, one can comfortably neglect the stator circuit transients
right and therefore the moment you neglect the stator circuit transient we get very big
simplification that the stator circuit voltage equations become algebraic equations right that is ed
equal to minus psi q omega r minus Ra id eq equal to psi d omega r minus Ra iq and eo equal to p
psi o minus Ra io okay they are becoming the algebraic equations, okay the another important
assumption which is made or approximation which is made is in this equation this rotor speed
omega r is assumed to be equal to be synchronous speed that is in per unit system, we make
omega r equal to one right this is another assumption.
Now in practice by making this omega r equal to 1 there is not much simplification in the system
model however, the very interesting thing which is there is that if you assume omega r equal to
one then this assumption to some extent counter balances this assumption that is where you
neglect the stator circuit transients that is if you analyze the system performance neglecting only
p psi d term p psi q term and p psi o term. Okay second time you do the studies neglecting both
these terms right then not neglecting both the terms sorry by making the second assumption
correction that is we are neglecting this transformer voltage terms or stator circuit transients and
assuming this rotor speed to be equal to the synchronous speed right, if you make both these
assumptions right then the results which we get are more closer to the to the actual results right
and therefore this assumption is also a very desired assumption.
Now in the same book, Kundur in his example 5.1 in example, 5.1 he has established established
considering small perturbations that effect of neglecting the speed variation, counter balances the
effect of neglecting the stator circuit transients I will suggest that you should carefully study this
example 5.1, okay now these are the two major assumptions which have been made.

Now with these assumptions, let us see develop the model the simplified model which we will
get after okay that is if you make this assumption then the equations will becomes like this for a
balanced three phase system ed will become equal to minus psi q, where omega r is 1 minus id Ra
eq equal to psi d minus iq Ra that is these are the two stator circuit equations which we shall be
using for further system studies and no more simplifications are done so far the stator circuit
equations are concerned.

Now with these simplifications, we can derive the expression for the electrical power output and
electromagnetic torque or air gap torque. We will see that the power output Pt is ed id plus eq iq,
now you substitute the expressions for Ed and Eq from these equations, okay these algebraic
equations you substitute this.
With this substitution, we will find actually that the terminal power Pt will come out to be equal
to psi d iq minus psi q id. Okay psi d iq minus psi q id this is the expression minus Ra times id
square plus iq square, this term can be identified as stator circuit loss or stator circuit resistance
loss. Therefore, we say that terminal power is equal to the air gap torque minus armature loss of
the stator that is why I said that when you have to calculate actually the torque Te right or air gap
power we first calculate the terminal power we know it to that we add the armature resistance
drop I am I am sorry armature resistance loss not drop and this is very important equation which
we shall be using.
Okay, now after making these 2 major simplifications we can make some further simplifications
that is if you look into the model in the stator circuit equations become the algebraic equations,
okay but when you look at the rotor circuit equations then the rotor circuit equations are
differential equations. Okay now next next set of simplifications which can be made are that we
neglect the, neglect the amortisseurs.
We have amortisseurs located on the ah rotor , okay now here I would like to just question you
that manier times manier times people have considered the amortisseurs in the mathematical
model right but we can make some simplification here by neglecting the amortisseurs. Now the
justification for neglecting amortisseurs are like this one is that the synchronous machine
parameters related to amortisseurs or synchronous machine parameters related to amortisseurs
are generally not available or not easily available okay this is one and therefore if the data is not
available we have to take some approximate data and use it, this is one second is that this
transients associated with the amortisseurs are also fast transients their time constants are also
low and generally they decay fast.
Therefore, the moment you neglect the amortisseurs in the system or amortisseurs on the rotor
circuit what we gain, what do we gain just can you tell me just tell me, what will you what do
you gain actually by this because, I had mentioned that these are the 4 equations related to the
voltage, related to the rotor circuits. Okay this is the equation related to the field voltage and the
field circuit sorry this is related to the field circuit, this is related to the amortisseur on the d axis
and these are the two related to the amortisseurs on the q axis.
Now if I neglect this amortisseurs then these 3 equations disappear and these are the 3differential
equations okay and therefore, the immediate advantage which I get is the ah the model becomes
of lower order number of differential equations which I have to write down for the system will
reduce this is one advantage which I get immediately that is the model order gets reduced and
you can easily understood that if the number of differential equations are less right then the
solution becomes faster.
The second advantage which we will get here is that while performing the simulation studies or
dynamic simulation studies, we can use higher or the larger time step for integration because the
time step which we use for integration of the differential equations it depends upon, it depends
upon the time constants of the system right and in case you want to capture the complete
dynamics of the system right, our time step should be small enough to take care of all the time
constants
In case the moment we neglect these 3 differential equations, we get very important advantage
and that advantage is that the remaining equations have larger time constant and we can use
larger time step for simulation and once you use larger time step then our computational efforts
are reduced therefore we get two major advantages both one in terms of reduction in the order of
the system, second is reduction in the computational time or simulation time by using higher or
larger value of time step for integration. Okay now with this 2 assumptions which we have made
here. The stator circuit you know equations are now algebraic equations.
Okay the rotor circuit the the equations related to amortisseurs we have neglected then we are
left only with one differential equation that is efd equal to p psi fd plus Rfd ifd. This equation is
corresponding to the field circuit of the synchronous generator. Now this can be written in this
form p psi fd equal to efd minus ifd Rfd this is a differential equation, we write down the
derivative term on the side and therefore one has to note down very carefully that this is the only
differential equation associated with the electrical characteristic of the machine right.
The the swing equation are always there right therefore whenever you study the dynamics of the
system then we have to consider the differential equation associated with the synchronous
machine the stator and rotor circuits but after the simplifications we find that we have only one
differential equations which is associated with the electrical characteristic of the system and
therefore, my model becomes very very much simplified. Now here I will just present how the
alternatives form of machine equations.
So far the equations which we have written right these equations were written in terms of the
field current okay and flux linkages. Now another form of equations this there is only the change
in the form actually there is nothing new is that we define the terms like this. Here, we define the
term EI this is slight a mistake here EI a voltage proportional to ifd that is Ei equal to Lad ifd a
voltage proportional to ifd that is you know field current right, field current is ifd and a term when
you multiply this by Lad mutual inductance right then all these things are all expressed in per unit
quantity right. Therefore this EI term is proportional to ifd because this Lad becomes a
proportionality constant therefore this is one term which we define here, second term which we
define here is a voltage Eq prime this is proportional to the field flux linkage that is psi fd that is
instead of talking about the field flux linkages.

We talk about a voltage term and this voltage term is proportional to its flux linkages and this is
the proportionality constant and the third term which we define is Efd. Now this Efd is the voltage
applied to the field circuit right we transform this by multiplying with a multiplying factor and
again these are per unit quantities and therefore this is a constant and Efd a voltage term
proportional to the field voltage okay therefore instead of calling this term Efd, I will now express
in terms of this term Efd which is related to this voltage by this multiplying term that is the
coefficient.

Okay now when I use this when I use these terms now or the newly defined terms and and use
our equations. Okay the equations which we have got are these are the basic equations therefore,
let us write down the equation for psi d, okay the equation for psi d can be written now after
making all the simplifications here psi d equal to minus Ld id plus EI.

You can just look into the simplified equations which we have developed I will just show you
again anyway okay. Similarly, you can write down another equation for Eq prime using the
previous equations that is equation number 12.9. Okay you multiply these both sides by Lad and
divided by Lffd, you will get the equation in the form Eq prime equal to minus Lad square
divided by Lffd id into Ei. This term which is the coefficient of id has been identified as
difference of difference of the direct axis inductance minus direct axis transient inductance. In
fact, this I shall establish in my next lecture that is Ld minus Ld prime comes out to be equal to
Lad square upon Lffd. This derivation is to be done separately therefore, if you make this
substitution in the equation 12.1 2, 12.12, we will get the result in the form that is Eq prime is
equal to Ei minus Ld minus Ld prime id right. Now here this Id is normally this is small id not
capital id, we use this direct axis component of the stator current okay.
Similarly, similarly we can use this equation 12.10 to obtain an equation of the form. This pEq
prime equal to 1 upon Tdo prime Efd minus EI. Now here I think I will just show that equation,
if you take this equation this is the equation where p psi fd equal to efd minus ifd Rfd. Okay you
multiply both sides of this equation by Lad over Lffd, okay when you multiply both sides we will
have the term here p multiplied by Lad divided by Lffd into psi fd and this term is identified as Eq
prime as per our new definition. Okay similarly, this efd when it is multiplied by Lad upon Lffd
you multiply both side by Rfd and divide by Rfd that is you multiply and divide this term by this
term Rfd.
So that what happens is that we can identify this term as one upon Tdo prime this is called the
open circuit field time constant, open circuit field time constant because what it this Lffd it is the
this is the inductance of the field circuit and Rfd is the resistance therefore, the L by R is the time
constant okay and this time constant term is a very important term for our stability studies this
time constant term is generally of the order of 5 to 10 seconds. Okay, therefore this is the
equation now which is put in the from pEq prime equal one upon Tdo prime Efd minus EI that is
the equation which we have earlier which was in this form the differential equation p psi fd efd
minus ifd Rfd this has been transformed in terms of the newly defined variables that is Efd is the
voltage proportional to field voltage, EI is the voltage proportional to field current and Eq prime
is the voltage proportional to flux linkages, field flux linkages. Okay therefore now you have in
this equation you find that we have only voltage terms not the current or flux linkage term this is
the only difference.
Now I am here summarizing the model of the synchronous machine which is most commonly
used for system stability studies that is we write down psi d the flux linkage of the d axis equal to
minus Ld id plus EI okay, psi q is equal to minus Lq iq Eq prime expression for Eq prime is EI
minus Ld minus Ld prime id and pEq prime dot that is the d by dT of Eq prime is equal to one
upon Tdo prime Efd minus EI. Therefore in these four equations what we see here is that these
three are the algebraic equations and one is the differential equation therefore, the synchronous
machine is very commonly represented by these equations only one differential equation and
what this differential equation is if you look it very carefully this differential equation is related
to the field circuit of the synchronous generator the field circuit.
You know we started with the this was the equation you know Efd equal to p psi fd plus Rfd ifd this
is the equation for the field rotor field circuit okay and this has been transformed in this
particular form and the terms like EI is the related to id Ld like this similarly Eq prime which is in
this equation is related to other terms. Okay therefore these are the inter related terms this is the
differential equation but these algebraic equations are also required to be solved along with the
differential equations.
Okay now next very important thing is that we can we can draw a phasor diagram, phasor
diagram under transient conditions. Sometimes you even feels like actually that how can we
draw a phasor under transient conditions but since we have neglected the stator circuit transients
and therefore, so far the stator circuits are concerned it carries the balanced 3 phase sinusoidally
varying currents right and therefore you can represent or you can draw a phasor diagram under
transient conditions.
Now to draw the phasor diagram, we again start with our basic model there is original equations
we start with original equations note there is a lot of papers here. We start with these equations
because we have developed these equation ,okay ed equal to minus psi q minus Ra into id or id
into Ra. Okay, similarly eq equal to psi d minus ia iq into Ra okay therefore we will make use of
these equations to develop a phasor diagram .
Now let us take this term eq equal to psi d minus Ra iq, Ra iq there is a mistake here okay now this
psi d term this is the flux linkage term can be replaced by psi d can be replaced by minus Ld id
plus EI and in per unit system the inductance is equal to the reactance that is Ld can be written as
Xd in per unit system right because Xd is nothing but omega into Ld and omega is one therefore I
can write down this equation now eq is equal to minus Xd id plus Xad ifd minus Ra id that is this
term is identified as EI and therefore now i can write an equation very interesting equation for EI.
You can see this started with eq this is the fundamental stator circuit voltage equation okay or
stator circuit equation. Now in this equation we are substituting the expression for psi d which
we have already developed and we are also, we are also replacing this Ld terms or Led terms by
Xd or Xad terms because in per unit system they are same the reactance is the same as the
inductance per unit. You get the equation in the form EI equal to eq plus Xd id plus Ra iq this is
one very one important equation what we do is that this is actually the algebraic equation that
these are all magnitudes of these quantities we have got.
This can be transformed into phasor equation what we do is we multiply both the sides by j, you
multiply both the sides by j then you will get j times EI prime EI equal to j times eq j time Xd id
plus j times Ra iq, you note that anything accept that this equation is multiplied by j all through
once you multiply this by j then this term j EI can be identified as a phasor right therefore I can
represent this as EI delta what will be the location of this phasor it is going to be along be
quadrature axis because there is a term j associated with it j times eq again it is located along the
quadrature axis then here in this term j times Xd id, I put j Xd id tilde.
Now this j is not replaced because id is a reference therefore id when it is a reference I can call
this as a id phasor ah straight away okay since it is a reference there is no but in the case of eq j eq
is the phasor because it lies along the q axis. Similarly, like this right therefore I get a equation in
this form now when you see this equation here. You will find actually that this this is eq which is
along the q axis this term is also along the q axis that is id when it is multiplied by j times Xd
becomes comes across the q axis this is anyway along the q axis therefore, the this term EI which
we defined as as voltage proportional to field current is is lying along the q axis. Okay then we
make use of this equation Eq prime okay.
Now in this equation, we have Eq prime equal to EI minus Ld minus Xd prime into id okay here
we replace this EI by this term eq plus Xd id plus Ra iq after this replacement and replacing Xd Ld
and Ld prime by Xd and Xd prime we will get the equation Eq prime in the phasor form Eq prime
in phasor form as Eq prime equal to eq plus j times Xd prime id plus Ra iq that is this again these 3
terms you can see here they are this this sum is lying along the q axis that is this Eq prime also
lies along the q axis EI also lies along the q axis okay.
Then earlier we had defined one term Eq that is the voltage behind this impedance Ra plus j times
Xq that is terminal voltage to this terminal voltage we added this term that is R a plus j times Xq
into It this term also came out to be along the q axis. Okay therefore, this Eq in the phasor form is
represented as EI minus j times Xd minus Xq id that is we can express this Eq which was defined
earlier as a voltage behind this impedance okay this is also along the q axis and we have seen
actually that it is this voltage phasor which define the location of q axis with respect to the
terminal voltage therefore whenever we start drawing the phasor diagram, the first requirement is
that we have to establish the location of d and q axis with respect to the terminal voltage and that
now we have established the relationship between this Eq and EI okay. Now this can be written in
this final form.
This term define not deline, define we can define one more term here. Now let us see what this
um can you just interpret this equation, for example I had the equation in this form Eq equal to Et
plus Ra plus j times Xq into It okay therefore we have another term if we divide E prime, E prime
equal to Et plus j times Et plus this impedance into It, this impedance is Ra plus j times Xd prime a
sub transient direct axis reactance.
Okay it means this is the voltage behind this impedance terminal voltage plus voltage drop in this
impedance therefore, these voltages and these voltages can be simplified and expressed in the
form you just make the substitutions right Et can be written as ed plus j times eq It can be written
as id plus j times iq and therefore this E prime can be written as as ed plus Ra id plus j times Xd
prime iq plus eq plus j times Xd prime id plus Ra iq, this this whole thing will come that is
substitute the value of ed eq from the expressions for ed and eq which we have already developed
that is in the previous expressions I think okay.
Now here one thing which we observe is that if you take this voltage term, okay then this has two
components one will be along the d axis another will be along the q axis. The q axis component
of this quantity comes out to be Eq prime the Eq prime which we define as a voltage proportional
to field flux linkages right and the q axis component of E prime comes out to be quantity Eq
prime right and therefore now I have drawn a simple phasor diagram to represent all these
quantities.
This is not a very unknown phasor diagram to you right but to make the things complete what we
do is that you start with terminal voltage with Et. Okay let us say this is the terminal current now
this and the phase difference between the terminal voltage and terminal current is always known
to you then to this Et we add this voltage drop Ra It plus plus plus j times Xq It Ra It plus j times
Xq It you will get this term Eq and this Eq will define the position of q axis.
Once this q axis is defined you resolve the terminal voltage into two components Et as Eq and ed.
Similarly, the terminal current can be resolved into two components Iq and id right then in this
phasor diagram we can show this term E prime a voltage behind direct axis transient reactance
that is what you do is that you take Et to this you add Ra It plus j times Xd prime It therefore, this
is the location of E prime and and the when you resolve this E prime into 2 components the q
axis component of E prime is Eq prime right therefore, most of the ah relations which we have
derived can be easily remembered with the help of this phasor diagram.
Okay, now any doubt I will suggest that you start with our basic equations and make the
simplifications step by step, okay and establish establish the new equations or equations in the in
terms of new variables. Once they are established in terms of new variables right you can express
these variables as phasors that is EI is a phasor Eq prime is a phasor and Eq is a phasor right and
then also the term which you defined is E prime this as a phasor then the whole everything is
expressed in the form of a simple phasor diagram okay.
Now let me summarize ah what we have studied today. We started with the complete model of
the synchronous machine then we have made the simplifications, we have discussed the
simplifications, simplifications discussed are that the stator circuit transients are ignored another
simplification we have considered is the the omega r is assumed to be constant then third
simplification which we have incorporated is we have neglected the amortisseur circuits.
With this the resulting resulting model of the synchronous machine to describe the electrical
characteristic has one differential equation and that differential equation is ultimately written in
terms of in terms of the the variables which we have defined in terms of Eq prime in terms of EI
and efd, okay and this model is very commonly used although we do make some further
simplification to obtained a classical model right but this is the model which is very commonly
used and any simplification beyond this will not give you the desired results although it will give
you the information right but sometimes you can get approximate information about the stability
and then go for detail information. Okay with this I conclude my todays presentation. Thank
you very much.
Prof. M. L. Kothari
Lecture - 13
Synchronous Machine Representation (Contd.)
Friends, today we will discuss about the synchronous machine representation. Now, so far we
have developed the synchronous machine model, considering the rotor circuit and stator circuit
resistances and inductances. Now with this stator circuit and resistor circuit, rotor circuit
inductances the complete model of the synchronous machine is developed.
Now however there is 1 practical problem that the the the various inductances which are used
actually in modeling the synchronous machine are in general not available or cannot be measured
directly from the certain standard tests and the parameters that is the the inductances and
resistances of the stator and rotor circuits are which are used in modeling right have been called
fundamental parameters or basic parameters.
Now the standard ah the the practical practice is now to use derived parameters which can be
measured by performing test at the terminals of the synchronous machine. Now these derived
parameters are very important and the machine data are given in terms of the derived parameters.
Okay therefore, today we will try to relate the the derived parameters with the fundamental
parameters of the synchronous machine.

Now to understand this concept of derived parameters, let us first look at the direct axis network
of the synchronous generator. Let us say that this black box represents the direct axis network
and we have 2 pairs of terminals shown here. This this pair of terminals represents the stator side
of the synchronous generator, this pair of terminals represent the rotor side of the synchronous
generator that is at these terminals we we we inject the field voltage. Okay and at these terminals
the the stator circuit current is coming out okay.
Now we are we will be representing this machine by 2, 2 networks 1 direct axis another
quadrature axis right. Now in the direct axis network the the direct axis current is the output from
this network and we are looking for what is the flux linkage in the direct axis due to this
particular excitation at the field winding and the load which is supplied by the synchronous
generator terminals.
Now here the convention is that we have used the synchronous generator convention where the
terminal the the currents are leaving the stator terminals that is the generator convention. Now
with the generator convention the direct axis current is leaving the terminal of the d axis network
okay.
This network, this diagram represents the q axis network in this q axis network we represents the
current leaving the network as iq and the at the terminals we look for the quadrature axis flux
linkage psi q. Now here in this 2 diagrams which I have shown here instead of showing the
current id flux linkage psi id and applied voltage efd, we are we will be representing this in terms
of small perturbation. So that input quantity shown are delta efd and the output quantity shown
here are delta id and delta psi d, similar is the situation in the q axis network.
Now here we will be representing the mathematical model relating the flux linkage delta psi d,
now when we develop this model we will be developing first the model in terms of what we call
as a operational parameters, operational parameters that is we will represent delta psi d as equal
to minus Lds into ids plus a transfer function we will call as Gs delta efds.
This is a very standard way to start with that is we are trying to relate the the current flux linkage
and the voltage applied to the terminals of the synchronous at the field winding of the
synchronous generator in terms of what we call as the operational parameters that is this Lds is
the operational direct axis inductance, operational we understand what is the meaning of
operational here right but this parameter s which I have shown here the s is the familiar Laplace
transform operator. Now if we recollect actually, your circuit theory then in circuit theory. We
represent the operational impedance of a RL circuit as R plus sL is it not, this is the operational
impedence we can call it Zs that the same approach is used here.
Now the our our problem basically will be to obtain the expression for Lds and Gs in terms of
fundamental fundamental parameters or we call it the basic parameters of the synchronous
generator.
Now these this equation I have written here because when I consider the small perturbations the
same equation is written here okay, the minus Lds into delta ids plus Gs into delta efds okay equals
psi ds. Now similarly we can write an expression for quadrature axis network the delta psi qs
equal to minus Lqs delta iqs because in the case of quadrature circuit there is no applied voltage
and there is no field circuit on the q axis okay, on the q axis we have always the damper circuits.
Now in this expression this transfer function Gs is the stator to field transfer function, stator to
field transfer function you can easily see here that Gs suppose delta id is 0 that is this machine is
under no load condition, okay then you can easily see that delta psi ds will be equal to Gs times
delta efds therefore, this transfer function relates the stator to field transfer function Lds is the
direct axis operational inductance. Okay Lqs quadrature axis operational inductance okay.
Now our interest here is now to develop develop a relationship between the operational
inductances and this transfer function Gs with the fundamental circuit parameters or fundamental
parameters of the synchronous generator. Now for doing this, we again start with our basic
equations of the synchronous generator okay. Now in this case of synchronous generator, we had
developed the flux linkage equations these flux linkage equations now are written in the
operational form that is psi ds can be written as minus Ld ids the current is now operational
current Lad ifds plus Lad i1ds.
Now here while writing the equations I am considering 1 damper on the direct axis right
therefore, we were denoting the damper winding by the symbol kd, k used to stand for the
number associated with the damper winding therefore, since k is 1 in this case it becomes L1d
further we are assuming that the all mutual inductances are same, so that the this Lad appears in
both the terms okay.
similarly we write down the flux linkage for equation psi fds as minus Lad ids plus Lad i1ds plus lffd
ifds right and the flux linkage of the damper winding in the per operational form is written here.
Now these 3 equations are were derived earlier and what we have simply d1 is that we have now
put in the operational form okay.
Similarly, the efds efds when it is written that is the field winding voltage right, we are writing
now expression for efds in operational form again. Now when you write this expression for the
operational form efds the we will write the equation right then it comes out to be s times psi fds
minus psi fd0 plus Rfd ifds that is you take again the standard equation for the field circuit right
because in the case of field circuit, we have the term d by dt of psi fd efd has a term d by dt of psi
fd and therefore when you take the derivative and then when you take the Laplace transfer of the
derivative, you will have the initial condition that will appear here because in the previous
equation these were the algebraic equations.
Now equation for efd it has a derivative term and therefore this appears similarly, actually we
write down an equation for an equation for the damper winding there is no applied voltage right
but still we have d by dt of psi 1d therefore, again you will find that initial term come therefore
now what we do is that in order to avoid the initial values we express these equations in terms of
incremental values, okay so that under steady state conditions the incremental values are 0.
Okay therefore, when you use this incremental values the equation will have straight away the
form delta efds equal to s times delta psi fds plus Rfd ifds this is very important equation is very
straight forward actually if you look in terms of our basic equation right then we have applied
voltage is equal to Ri plus Ldi by dt. Okay therefore you can write that Ri plus d psi by dt and this
is what what we have already derived.
Now in this expression you substitute the expression for delta psi fds okay delta psi fds, we have
already written earlier that is use this expression for this delta psi fd and here these are all
constants you can convert this delta id delta i1d delta ifd. Okay and you substitute the expression
for delta psi fds here in this expression and then simplify when you simplify this this delta efd will
appear as minus s times delta Lad delta ids plus Rfd plus s times Lffd delta ifds plus s times Lad delta
i1ds that is we have written now incremental field voltage right in terms of the 3 currents. Okay
similarly you write down this expression for damper winding also because in case of damper
winding the applied voltage is 0 right therefore, you have this is minus sLad delta ids plus sLad
delta ifds plus R1 plus sL11d delta i1ds.
Now at this stage, at this stage again me let me reiterate actually that we want to relate the stator
circuit quantities to the rotor circuit quantities right in terms of the stator variables and rotor
applied field voltage that is that is the expression which I have if you look at this expression delta
psi ds is equal to Gs delta efds minus Lds delta ids in this expression or in the second expression the
rotor currents do not appear rotor circuit currents that is i1d or ifd these 2 currents do not appear
therefore, what we do is that we eliminate this currents that is from this using these equations
what we do is that you eliminate this currents.

Now to eliminate this what you do is that that you use this equations and write the expression for
delta ifds and delta i1ds in terms of delta efd and delta id that is these 2 equations associated with the
other equations which are ah derived earlier right what we do is that we we write the expression
for delta ifds and delta i1ds in terms of this quantity and this known quantities that is we are trying
to eliminate the field current and the damper circuit current.

Now this this step is little what involved step but you can obtain it. Once you obtain this
expression next step will be to to substitute this in the expression for delta psi ds because if you
look actually the expression for delta psi ds, delta psi d this is that equation for delta psi d or psi
d equation equation for psi d is here. In this equation we have id, ifd and i1d, okay what we have
now d1 is that ifds the this will be expressed in incremental form then once you express in
incremental form in this expression you replace this delta ifds and delta i1ds using those
expressions.

Once you express you will find that you can write the expression for delta psi ds as Gs delta efds
minus Ld ids do you appreciate this these steps. Now once you do this thing what will happen is
that this expression for Gs which you get will be written as that is 2 expressions Lds and Gs, the
Lds will be written as Ld multiplied by an expression of this form 1 plus T4 plus T5 into S plus T4,
T6 S square divided by 1 plus T1 plus T2 S plus T1, T3 S square that is when you do this
simplification and express delta psi ds in terms of delta efds and delta ids right.
Then the coefficient of delta efd term will come out to be, coefficient of delta efd term will come
out to be in this form, that is Go equal to ah Gs equal to Go into 1 plus divided by 1 plus T1 plus
T2 S, T1 3S square while the expression for Ld will come out to be equal to this the the yes, then
these constants which are shown in the this equation are related to the fundamental parameters of
the synchronous generator by this expression that is Go equal to Lad by Rfd, Tkd is Lfd by R1d, T1
equal to Lffd by Rfd, T2 equal to L11d by R1d, T3 1 upon R1d multiplied by L1d plus Lad Lfd divided
by Lad plus Lfd T4 1 upon Rfd Lfd plus Lad into Ll divided by Lad plus Ll and T5 as 1 upon R1d
multiplied by L1d plus Lad L1, Ll divided by Lad plus Ll and similarly, T6.
You can easily see that when you perform this simplification right right, the you will be in a
position to express the Lds that express this Gs as a ratio of 2 polynomials right. The denominator
here is of the second order and numerator is 1 order less than the denominator similarly, Lds is
expressed as ratio of 2 polynomials where the order is same further the denominator of Lds and
Gs are same that is this denominator for Lds is same as that of the okay, this is these derivations
can be carried out.
Now at this stage what is d1 is we write the expression for Lds that is the operational direct axis
inductance as Ld multiplied by by numerator that is by this expression which is the numerator as
1 plus s time Td prime into 1 plus s time T double prime 1 plus s times Tdo prime plus into 1 plus
s time Tdo double prime that is what we do is now this this operational inductance will express in
terms of time constants and these time constants as we will see are are the the standard
parameters of the synchronous generator because these are the parameters that is going to
determine the the the transient performance of the system okay.
Similarly since Gs has the same denominator right therefore, we when you put in this formula
where Gs can be written as 1 plus s time Tdo prime and T double prime okay now our interest is
that we know we know the expression for this time constant T1 to T6 in terms of basic circuit
parameters is it not these are all known. We want to relate this these standard parameters that is
this time constant Tdo prime Tdo double prime Td prime Td double prime to our basic parameters
or fundamental parameters. Now to correlate this before I correlate I will just write 1 more
expression that if we follow the same approach as we have d1 for d axis we can write the
expression for Lqs as that is operational quadrature axis inductance as Lq multiplied by 1 plus s
times Tq prime into 1 plus s times Tq double prime divided by 1 plus s times Tqo prime 1 plus Tqo
double prime. Now again these constants Tq prime Tq double prime Tqo prime Tqo double prime
these 4 time constants right are again the standard parameters of the synchronous generator right.
Now to to obtain an expression for these parameters in terms of T1 to T6 what we have to do is

that we compare this denominator and numerator terms after all actually we had written earlier,
let us say Lds Lds is written in this form see this Lds is written as Ld multiplied by this expression.
Okay and Lds we are now writing in the form of in this form in terms of this time constants, okay
while these are our standard parameters and as we will see that they can be measured from the
certain tests which can be conducted at the tests on the synchronous generator right. Now at this
stage we make to we see here actually this expression look at this expression.
First I will just look at the denominator 1 plus T1 plus T2 s plus T1 T3 s square. Now in case
suppose you have instead of T2 here as T3 then this can be broken into fractions, you know
suppose actually you have the expression of this form 1 plus T1 plus T3 into s plus T1 T3 s square
the the partial, you can find out the factors of this expression the factors will be 1 plus T1 s into 1
plus T3 s. Okay now if if it is in this form then directly you can identify that T 1 is equal to Tdo
prime and T3 equal to Tdo double prime but incidentally the expression is here 1 plus T1 plus T2 s
plus T1 T3 s square.
Now here we make use of 1 simplification or approximation we can say, the approximation is
that in practice in practice R1d is very very large as compared to Rfd, R1d which is the resistance
of the damper winding on the d axis or amortisseur on the d axis is very very large as compared
to the field winding resistance. When you make this assumption T3 and T2, T2 and T3 come out to
be very small as compared to T1 and T5 and T 6 will come out to be very small as compared to T4
right therefore in the expression for the the expression which we have is in the denominator we
have 1 plus T1 plus T2 therefore this T2 is negligible as compared to T1 okay.
Now T3 is also very small as compared to T1 therefore first step suppose I neglect T2 I will not
create much error second actually step is that suppose I add T3 to this term then also I am not
going to create big error therefore what we do is that this denominator this is actually 1 plus T1
plus T2 into s plus T1T3 s square right will be replaced by this term 1 plus T1 plus T3 s plus T1 T3
s square, the justification I have just now told you that T2 was is very small as compared to T1
similarly, T3 is also very small as compared to T1 right therefore what we do is that we neglect T2
but we add T3. We do not create much error but we simplify our results and expression.

With this you can see here 1 plus s times Tdo prime becomes 1 plus s times Tdo double prime
which is of course is standard. Now when you make this assumptions I just told you the
approximations not assumption the approximations then you can write down this this. You can
equate the 2 denominators and you get this type of expressions, in the case of numerator the
approximation is that T5 and T6 are very small as compared to T4 right therefore, we are now
getting a simple expression of this form. Remember that, remember that when I am making this
assumption that T3 is very small as compared to T1 right in the product term T1 T3 s square I am
not making any approximation T1T3 remain there.
Now when you do this you will find actually that Tdo prime can be identified as T1 Tdo double
prime can be identified as T3, Td prime can be identified as T 4 and Td double prime can be
identified as T6 while the while we know that T1 is equal to Lffd divided by Rfd. Similarly, Tdo
double prime is equal to 1 upon R1d multiplied by this expression and expression for Td, T4 and
T6 were also derived therefore they are all time constants.
You can easily see that the this this term T1 is equal to ratio of inductance to resistance that
means it has the characteristic of a time constant. Okay, here also in this bracket you have
inductance term divided by resistance therefore, this also is a time constant right and now we are
relating actually these 2 time constants that is the we call actually the standard time constants
right are related to the the basic parameters of the synchronous generator we will see actually
that this time constant can be measured by performing tests okay
Now at this stage let me just sum up what we have d1 till now is that we have started with with
the ah operational model for d axis and q axis winding for the synchronous generator okay and
then using the the standard equations. Okay for the synchronous generator stator winding and
field winding right we have established the transfer functions for Gs and Ld and similarly, for Lq
then assuming that assuming that this can be expressed in terms of some time constants or the the
standard parameters. We have established actually that this time constant can be related to the
basic parameters okay.
Now till now actually in my all discussion I have not talked about what is the transient
reactance? what is the transient inductance? what is the sub transient reactance, sub transient
inductance but we all know actually that whenever there are sudden changes in the synchronous
machine operation, let us say a fault okay then the currents are induced in the rotor circuits and
this currents will decay at certain rate and the rate of decay of these currents are determined by
the time constants. A certain time constant which determine the rate of decay of the currents I
will just take the case that whenever there is a short circuit currents will be induced in the
damper windings they will induced in the field winding right similarly, in the damper winding of
the q axis and so on and they decay with certain time constants okay.
Now first we will try to establish that what will be the inductances offered offered by the
synchronous machine when sudden changes take place or the transient take place. Now to
understand this what we do is we start with, let us look at this expression for Lds this is the
expression for Lds okay.
Now suppose I want to know the behaviour of the machine under sustained disturbance condition
that is the when the machine is carrying the steady state currents at that time I can substitute s
equal to 0, s equal to 0. When I substitute s equal to 0 then this Lds will become equal to Ld that
is this operational inductance for s equal to 0 will become Ld that is Ld is your synchronous direct
axis synchronous inductance because at that time the currents are under steady state conditions in
sustained conditions.
Therefore, now I can say that under steady state condition s is 0 and therefore, Ldo that is the
operational impedance for s equal to 0 is same as the synchronous inductance or d axis
synchronous inductance. Now you consider another case let us say s is approaching infinity, a
very high frequency transient when s approaches infinity it means actually the transient is of
short duration right then the frequency associated will be very large.

If you make this situation that is the rapid transient condition then this Ld double prime, we will
represent this operational impedance for s equal to infinity as Ld double prime right and this will
come out to be equal to Ld multiplied by this this expression. Now what we are trying to say is
that the inductance offered by the direct axis under rapid transients is denoted as sub transient
inductance and this sub transient inductance is related to the synchronous inductance by this
expression these time constants further further if you assume assume actually that the there is no
damper winding in the circuit, let us say that there is no damper winding in the circuit direct axis
only there is a field winding right.
Now at that time, at that time when there is a rapid transient right the inductance is denoted as Ld
prime and again s is approaching infinity right, this will come out to be equal to Ld into Td prime
upon Tdo, Tdo prime therefore now I can say here that we have we have identified the 3 different
type of reactances offered by the synchronous machine that is synchronous inductance, sub
transient inductance Ld double prime and transient inductance Ld starting from the operational
inductance point and the expressions for this if you just look into the complete expression
because when you substitute here the expression for Td prime, Tdo prime, okay in terms of the
fundamental parameters you will get the full expression for Ld double prime okay and this these
are expressed in terms of the basic parameters the expressions are quite lengthy you can write it
down.
Similar exercise can be performed for q axis that is you can perform the same exercise for q axis
and you will find actually that the time constants are Tqo prime equal to this and Tqo double prime
equal to this again expressed in terms of fundamental circuit parameters right and this transient
and sub transient inductances of the quadrature axis winding damper, winding and their
parameters are also given. Now we have to just look slightly in more detail about these time
constants.

Now here if suppose I perform a test on the synchronous generator with with stator terminals
open that is under no load condition. When we perform a test right, where we we we consider the
synchronous terminals open then this term will absent, this terminal will be absent under this
case actually we have relation with the field direct axis field flux linkages, direct axis linkages
with respect to delta efds therefore this transfer function Gs affects the affects the change in d axis
flux linkages as the change in the applied voltage of the in the field circuit. Okay and this change
is determined by the 2 time constants, 1 time constant is Tdo prime another is Tdo double prime.
Now since this test is performed under open circuit condition, okay therefore this time constant
Tdo prime is called direct axis open circuit time constant or we always call the open circuit direct
axis time constant Tdo prime. Okay and the Tdo double prime is called called the open circuit sub
transient direct axis time constant here actually the Tdo prime is called basically full full name is
it is open circuit direct axis transient time constant, the transient word is also there Tdo prime.
Okay and this time constant is very important actually when you model our system in fact 1 can
perform 1 test very simple test actually on the synchronous generator and that is normally called
the steady state frequency response no not steady state h what you can call another term is stand
still frequency response test the stand stand still frequency response test what is d1 is that you
keep the machine stationary not rotating, okay you apply field voltage and vary the frequency of
the applied voltage.
Okay and measure the output voltage of the machine and plot plot actually the phase plot the
gain versus that is the applied voltage at the field winding and the output voltage at the terminal
of the synchronous generator for for a range of frequency right once you make this plot and then
this plot can be that is this can be this curve can be fitted with the characteristic of this Gs. You
can do the curve fitting right and by doing the curve fitting you can identify the time constant Tdo
prime and Tdo double prime and Tkd also right.
Now over the years in the literature actually there is lot of lot of emphasis has been laid down on
measuring the synchronous machine standard parameters therefore, when I talk about standard
parameters the standard parameters are the the synchronous inductance sub transient inductance
transient inductance then these time constants right.
Now these are parameters called as the standard parameters and these parameters can be
measured by by performing tests at the synchronous machine terminals and lot of work has been
going on actually to measure accurately these standard parameters of the synchronous generator
particularly those who are actually working in the field actually right. Sometimes you will come
across where you have to measure these parameters by performing the standard tests and a lot of
literature is available on performing this test even IEEE has given some standards, standard test
to be conducted for measuring these standard parameters.
Now in the case of salient pole machine, salient pole machine there is 1 small additional point to
be learnt and it is something like this in the case of salient pole machine we provide the damper
winding in the phase of the pole there is a the pole is a laminated pole, you know while actually
if you talk the round rotor machine the rotor is as solid it is not laminated 1 right and therefore
you can always identify or you can represent actually quadrature axis by 2 windings, 2 damper
windings right while in the case of salient pole machine there is 1 damper winding provided
there is only 1 damper winding.
Now once there is only 1 damper winding, you cannot have any distinction between the transient
and sub transient inductances Ld prime and Lq prime and Lq double prime this distinction is not
possible, they are same okay and therefore this for salient pole machine the sub transient and
transient they are same that is the transient there is no mention of transient we always call of sub
transient okay because there is they are considering the damper winding 1 damper winding only
these are the expressions for salient pole machine.
Further you will see actually in the literature actually after this discussion will show you that
these are the relationship between the different types of reactance because when we talk in per
unit system the inductance is same as the reactance okay these are the standard relations which
are given.
Now, I will just quickly tell you 1 more point actually before I close my discussion here that
when we started the study of transient stability I talked about the classical model and in the
classical model I said that the synchronous generator can be represented by a constant voltage
behind transient reactance and this representation is valid, when we assume the field flux linkage
psi fd constant that is if you assume this psi fd constant right then the then the voltage behind
transient reactance or direct axis transient reactance becomes constant.
Okay now this we will quickly derive actually and show you that yes how it happens now the
steps involved are like this, you write down the per unit flux linkage in the d axis psi ad, you can
just write down psi ad is equal to minus Lad id plus Lad ifd. Here, we are assuming the mutual
inductance is same, okay psi d is equal to psi ad minus L1 id. Okay this is the difference between
psi ad and psi d psi fd is psi ad plus Lfd ifd these equations are standard. We all know it about now
using this last expression what we do is that is we find out the expression for ifd.
We find the expression for ifd, the ifd comes out to be equal to psi fd minus psi ad upon Lfd. You
eliminate we eliminate the term ifd from the first equation that is psi ad which is equal to minus
Lad id plus Lad ifd, you substitute the expression for ifd then psi ad will be written in the form
minus Lad id plus Lad upon ifd psi fd minus psi ad. Okay when you rearrange the whole thing you
rearrange the term that is psi ad you take it out on this side and express psi ad in terms of other
parameters it will come out to be equal to psi ad is equal to Lad prime into this term where Lad
prime is defined like this okay. Similarly, for psi aq 1 can write down the psi aq as equal to Lq
prime minus L1. Now here we will make 1 more small assumption here that the transient
saliency, saliency we will neglect we assume that Xd prime is same as Xq prime right.

With this assumption, now we can write down our d axis voltage ed equal to Ra id minus omega
psi q which is written in terms of psi aq id and iq this is ed that is in this expression for ed ed, you
replace this psi aq now once you replace the psi aq ed can be written in terms of id iq and ed prime
that is this expression for ed is written as minus Ra id plus Xd prime into iq plus Ed prime this Ed
prime is a important term here which is expressed here as Ed prime equal to minus omega Laq
prime into this. If we assume this psi 1q constant that is the flux linkage in the q circuit also
constant right then this term is a constant term.

Similarly when you you can write down for eq where you it will be written in this form minus Ra
id iq minus Xd prime into id plus Eq prime, where this Eq prime is written here as omega Lad prime
into psi fd upon Lfd. Okay therefore this Eq prime is written in terms of psi fd and Ed prime is
written in terms of psi ad therefore, this Eq prime and Ed prime if you assume these 2 fluxes to be
constant they will remain constant right.
Once these 2 voltages are constant I have established here I have establish here that this voltage
E prime that is the the voltage E prime that is the voltage behind transient reactance can be
written as Ed prime plus j times Eq prime and when these 2 voltages are constant this E prime is
also constant.
Now here I you you have to look like this that if suppose consider that this is my q axis, okay this
will become d axis let us say this is our E prime. This quantity is going to be Eq prime this is
going to be your Ed prime. Now if Eq prime is constant Ed prime is constant E prime is constant
if suppose these are our q axis let us say that the reference axis is shown somewhere here this is
R and I these are our reference axiss and we measure the angle with respect to the reference axis
that is E prime this angle is measured as delta.
Now you can easily see here actually is that so far this E prime is concerned right the magnitude
of E prime is constant and its position with respect to q axis is fixed right and therefore for
studying the stability I can measure the angle of E prime with respect to some arbitrary reference
it may be synchronous rotating reference prime or so on therefore here you will find actually that
the component of E prime call it say ER and EI this is our actually I and R they will be varying as
the delta varies therefore there are 2 important points to sum up here 1 is that this E prime
remains constant when we make these assumption second point is that position of this E prime
with respect to q axis is fixed and therefore, I can measure the angle with respect to angle
between E prime and any reference frame for performing my stability studies right. Let me sum
up here that today, we have developed the relations between the standard synchronous machine
parameters with respect to the basic parameters and also we have established that that the
classical model in a classical model the the if you assume this flux linkage is constant then the
voltage behind transient reactance is constant right. Thank you!
Power System Dynamic
Prof. M. L. Kothari
Lecture - 14
Excitation Systems
Friends, today we will study excitation systems.
In this topic we will cover what are the excitation system requirements then what are the basic
elements of excitation systems then the dynamic by performance measures dynamic performance
measures are categorized into 2 categories large scale performance measures and small signal
performance measures. Then we will discuss about the various control and protective functions
which are provided in excitation systems the some of the control and protective functions which
we will discuss are AC and DC regulators, load compensation, under excitation limiter, over
excitation limiter and field shorting. Now we will briefly address to all these points.
So far the excitation system is concerned it plays a very significant role in the stability analysis
of a power system and it is also very powerful control which can be used to improve the stability
of the system. Here I would like to tell that ah over the years, over the years the different types of
excitation systems for the synchronous generator have been developed. You know that we have
in today our power plants, the DC excitation systems, AC excitation systems and static excitation
systems and in AC you have brushless excitation systems and the static or stationary rectifier
excitation system right. A variety of excitation system have been developed and these
development has been have been possible with the development of the power electronic devices.
Okay we will read into go into these details as we proceed across or proceed with, now first let
us see the basic function of excitation system.
These are all it is known to all of you that the basic function of an excitation system is to provide
direct current is to provide direct current to the synchronous machine field winding. This is a
very simplest thing that in any synchronous machine right, we have to provide DC current or
direct current right it operates. This is the main purpose of excitation system then the excitation
system also performs various control and protective functions. We will see control and protective
functions, these functions are essential for the satisfactory performance of the power system that
is the modern excitation systems are provided with a variety of controls right and variety of
protection functions, the very propose of providing these control and protection functions are, so
that this excitation system can give you the satisfactory performance for the operation of the
power system as a whole. The as I told you that there are two functions control functions and
protection functions okay.
Now what are these control functions let us see, control functions are basically you can put at
control of voltage and reactive power flow. In excitation system right we provide various
controls so that we can maintain the terminal voltage constant under varying load conditions. I
hope all of you aware of the capability chart of a synchronous generator where a synchronous
generator operates operates in the given region in the pq plane, okay and this ah capability
charge is decided from the consideration of over heating of the armature, over heating of the
rotor circuits or field system.
Similarly, the small signal stability or stability considerations right therefore, any synchronous
generator is required to operate within the capability chart. In addition to this the synchronous
generator has ah certain short time capability also. So that under for a short time you can operate
the synchronous generator to enhance the system stability right and we take advantage of short
time capability of the synchronous generator, the short time capability is generally of the order of
30 to 60 seconds right and therefore, when I talk about the excitation system it has large number
of controls right and these controls are provided primarily to the achieve these two goals, one is
the control of voltage and reactive power flow, okay second is enhancement of system stability
okay that is the excitation system can be appropriately controlled so that you can operate the
system with larger transient stability limits. Okay that is called the enhancement of the stability
of the system.
Then the protective functions, why do you require the protective functions? okay the protective
functions ensure that capability limits of synchronous machine excitation system and other
equipment are not exceeded that is you operate the system within the capability limit of the
excitation, limit of the synchronous generator, limits of the excitation system right and any other
associated equipment as I have just now mentioned that a synchronous generator has to operate
within the capability limits and that limits are obtained in the form of a capability chart okay and
this synchronous generator has a short time extra extra capability it can supply extra reactive
power, extra real power but for a short time okay and therefore, our controls are so provided so
that we can use the or we can take the advantage of this short time capability for improving the
transient stability, long term stability, dynamic stability of the system.
Okay but we provide in addition to this control functions protective functions. So that in case any
of the capability limit is violated right the protection functions will operate right and and keep
this or operate in such a fashion. So that you operate within the capability limit we should not go
outside the capability limit because any attempt of the system to go outside the capability limit
means system is likely to be or generator is life is going to be affected. Okay therefore, these are
2 important functions which you have to look into all the time control functions and protective
functions.
Now we we can look into the excitation system requirements from 2 considerations, one is the
generator considerations from the consideration of synchronous generator and another is from
the consideration of power system where synchronous generator is part of the large power
system right and therefore, we will look into the excitation system requirements from the 2
considerations, one is generator considerations another power system considerations.
When you look into the generator considerations the main function of the excitation system is to
maintain terminal voltage as output varies it has to automatically adjust the excitation current as
the output varies and keeps the terminal voltage constant. This is one of the important
requirements from the generator point of view second is that the excitation system must be able
to respond to transient disturbances with field forcing.
Now this terminology is very important that is our excitation system should be such that such
that whenever there is some transient disturbance okay and that particular following that
disturbance if you want to reserve to field forcing field forcing means we apply a large voltage to
the field winding of the synchronous generator so that more current can flow through field
winding that is excitation current can be increased temporarily or for a short duration.
Now this field forcing capability is one of the important requirement of excitation system as you
will see that that to have good field forcing capacity the ceiling voltage of the excitation system
right the ceiling voltage when you design the excitation system it will have a high ceiling
voltage.
Now when you look from the power system power system considerations then the excitation
systems which are provided on synchronous generator should be should have effective control of
voltage and enhancement of system stability that is we should be in a position to control the
system voltage by controlling the excitation system okay and we shall also be in a position to
enhance the system stability. Okay, second is to achieve these goals its response should be high,
rapid response, rapid response, so that you can improve the transient stability. Okay that we will
talk subsequent in our subsequent discussion that how the rapid response or high response of the
excitation system improves the stability of the system.
Okay but first we will understand what we mean by rapid response how we quantify the dynamic
performance of the system okay. Then we should have a provision to modulate the field current
to enhance small signal stability that is when the system is oscillating you know whenever there
are some perturbations taking place right the oscillations are created in the system the system is
never in steady state condition it is always in the quasi steady state condition because load
changes are taking place continuously okay and whenever the load changes take place the
generation changes the load change load changes and therefore in this process of generation
changing the load right the system is in the dynamic condition okay and there are always small
oscillations in the system when I say small oscillations magnitude of oscillations is small and the
frequency is the low frequency oscillations right and to there is a necessity to improve the
damping of these oscillations by providing power system stabilizers right.
Therefore the power system stabilizers are provided in the excitation system to enhance the small
signal stability therefore, the from power system consideration I can say it should have the
capability to enhance transient stability for improving or for achieving this goal its response
should be high right and then we should have the facility or capability to modulate the excitation
system. So that we can improve the damping of the low frequency oscillations, okay now these
are the requirements from the power system considerations. Now let us see what are the main
elements of an excitation system?
The elements of an excitation system can be listed as exciter, regulator, terminal voltage
transducer and load compensator, power system stabilizer, limiters and protective circuits that is
if broadly if I look into the complete excitation system then the various building blocks of the
excitation system can be put as the exciter there is a regulator there is a terminal voltage
transducer and load compensator, a power system stabilizer and limiters and protective circuits
okay.
Now this is the block diagram which shows how these elements are interconnected or form the
complete excitation system. Now this is the main block where you can say this is our generator, a
synchronous generator this synchronous generator is connected to the power system, okay the
field winding of the synchronous generator gets its current from exciter that is the exciter
supplies the field current this is your exciter then the field the the output of the exciter is
controlled by providing a regulator this is the regulator.
You can see here that regulator has number of inputs here. I will explain but one of the biggest or
very important important building block of the excitation system is a regulator because here we
are not talking about the exciters only we are talking about excitation system okay. As I have
told you that the primary function of the excitation system is to is to provide DC current to the
field winding of synchronous generator, okay and we provide various controls and protection
system so that we achieve the requirements of the system from generator considerations and
from power system considerations.
Now first, let us look at actually at we maintain the terminal voltage How do we do it? we sense
the terminal voltage from the synchronous generator terminals. Okay now this is the block which
shows the terminal voltage transducer and load compensator I shall separately talk about load
compensator but you can understand that when I want to regulate the terminal voltage of
synchronous generator right we sense the terminal voltage okay.
Now since the synchronous generators terminal voltage is AC terminal AC voltage it may be 11
kv may be 16 kv depending upon the design of the machine therefore, at this point the voltage is
stepped down okay and then it is rectified and we get a DC signal which represents the terminal
voltage of the synchronous generator. Okay that is why what I used is the terminal voltage
transducer that is you step down the voltage you rectify and bring down to a level which can be
used for the purpose of control okay and this signal which is the which represents the terminal
voltage or a function of terminal voltage and load current because we will be showing about the
load compensator also this is fed to the regulator, this is one of the input signals that goes to the
regulator. Okay the second block which I have shown here is the limiters and protective circuits
okay.
Now the limiters will also sense the terminal voltage load current okay and looking and it is
going to process these signals current and voltage which are sensed from the terminal of the
machine okay and then we will decide whether the synchronous generator the operating
condition is within the limits or not in case it is going outside the limit it will take appropriate
action, again the action is taken through regulator. Therefore, this this signal which is coming
from the limiters and protective devices also is fed to the regulator.
Then you have one feedback loop around the exciter this this loop is generally we call this the
excitation system stabilizer that is here you sense the output of the exciter okay and in case you
do not have this feedback loop then the exciter itself may become unstable because once you
have a close loop control system right the stability is a very important requirement and therefore,
you will find that there is one feedback loop right that loop is called excitation system
stabilization okay and that again operates through regulator and the last feedback loop is through
power system stabilizer. You can see this power system stabilizer again this power system
stabilizer takes its input signal from the generator terminal, the input signal to the power system
stabilizer is generally the speed deviation.
When a synchronous generator speed deviates from synchronous speed this division is sensed
another signals which have been tried are the deviation in the power output instead of sensing the
speed deviation you can sense the deviation in the power output, the delta pe right or deviation in
frequency also has been looked into it or a combination of these quantity, these signals okay.
Now this power system stabilizer is basically an auxiliary controller it will operate only during
the dynamic conditions not during the steady state conditions okay and this also operates through
the regulator and any regulator basically basic function of the regulator is to is to perform all the
required functions and maintain terminal voltage constant and therefore, we have a reference
therefore input signals which come to the regulator okay they come through the terminal voltage
transducer and load compensator, a voltage signal which is function of terminal voltage and also
the load current, a properly processed we have to discuss in detail about this block then limiters I
have told you that we , we have to see that synchronous generator always operates within the
operating limits whether actually you want to use the short term capability then still the limits are
decided.
Okay we have to operate within the short term and long term operating limits of the synchronous
generator. Then this power system stabilizer is to damp the low frequency oscillations which are
normally provided or which happens because or which exists low frequency oscillators exist
because of because of the perturbations which are taking place continuously in the system and
this oscillations need to be damped they should not increase they should not grow, otherwise
system will become unstable. Okay now these are the basic building block of a complete
excitation system okay is it okay or any doubt.
See see these these functions are provided in a complex manner, complex manner but limiters
have they just they just monitor magnitude of the voltage okay terminal voltage magnitude of the
current which is flowing through a synchronous generator, magnitude of the power which is
output, real power, reactive power not only one quantity they will monitor so many parameters
because when I talk about the capability limit of the synchronous generator then we look into the
pq, pq plane okay for a given value of p there is certain value of q possible.
Okay therefore pq plane we have to we have to maintain this is there therefore this is not only
sensing the terminal voltage but this limiter will sense all other parameters also and not this
block diagram is just to give you basic features of the excitation system but it senses not only the
the generator parameters but excitation parameters also exciters will also not go beyond its
operating limits right.
Therefore this is a it limiters means you just look at the movement limits are crossed it will take
action the action may be that it may even trip the synchronous generator from the system
suppose it feels that now the synchronous generator is going outside the capability limit then
these limiters can even trip the synchronous generator. Okay they are very very I say actually
that ah a very complex control protective functions are provided and our main consideration is
that we take the best out of the synchronous generator right from the system point of view and at
the same time its its safety is safe guard is ensured okay, is it okay. Now we will look into some
basic performance measures.
Here as I have stated that the excitation system control should be such that we are in a position to
improve the stability of the system. Okay and to improve the stability of the system, we have to
look into what are the desired characteristics okay and and therefore we look into the dynamic
performance measures, when the system is under dynamic condition what are the performance
measures okay.
Now this performance measures are further classified into 2 categories, the large signal
performance measures and small signal performance measure that is when the system is
subjected to large disturbances okay we have to achieve certain performance measures related to
large disturbances then whenever the system is subjected to small perturbations, okay then we
have to basically look into the the stability of the system okay.
Now large signal performance measures in the sense that suppose a 3 phase fault occurs in the
system or a important transmission line trips right then the system is subjected to large
perturbations and at that time how the system should behave and how are our excitation system
and excitation control should behave right therefore, we want to quantify that what should be the
dynamic performance measures of such a system. Now related to the dynamic performance
measures there are some basic definitions, we have to clearly understand.
The excitation system ceiling voltage is the maximum direct voltage that the excitation system is
able to supply from its terminals under specified conditions. It is the indicative of the field
forcing capability of the excitation system; higher ceiling voltages tend to improve transient
stability. This is a very important factor that the excitation systems which have higher ceiling
voltages right will have a higher field forcing capability and resulting transient stability limit will
be high, the next term in the excitation system ceiling current.
The maximum direct current the maximum direct current that the excitation system is able to
supply from its terminals for specified time right therefore, time is a constant here for a short
time it can supply current more than the rated value for prolonged disturbances this value
depends on thermal duty of the excitation system. In typical excitation systems the ceiling
current is of the order of 1.6 times the rated current. The next important term is the excitation
system voltage time response.
The excitation system output voltage expressed as a function of time under specified conditions
that is if I create certain control conditions so that the excitation system system causes increase in
the terminal voltage and this increase is when it is plotted as a function of time right then that
graph represents the voltage time response of the excitation system right. Suppose I give given
example actually that suppose you make a step change in the reference voltage right, the terminal
voltage of the synchronous generator is going to increase and if you plot voltage as the function
of time right then that curve is the voltage response curve. Now to quantify the voltage response,
we define a voltage response time.
The time in seconds for excitation system voltage to attain 95 percent of the difference between
the ceiling voltage and the rated load field voltage under specified conditions just to illustrate this
suppose we this is the rated voltages let us say it is 1.0 if I create terminal conditions or the
voltage regulator operate in such a fashion, so that voltage is made to increase in this fashion and
let us say that this is the ceiling voltage right, this is the ceiling voltage.
Then you find out the difference between the ceiling voltage and the rated voltage and the time
which is required to attain 95 percent of this value. Let us say this time that is this is the 95
percent of the difference between the ceiling voltage and the rated voltage and this time t right,
the time t in which actually the the voltage will rise from no rated value and it will attain 95
percent of the difference between ceiling voltage and the rated voltage this is the voltage
response time.
Now at the modern excitation systems have very high response and we do specifically
specifically characterize the excitation system as high excitation systems whose response time is
.1 second or less than .1 second right that is you have variety of excitation systems okay and
those excitation systems whose response time is .1 second or less than .1 second are
characterized as high response excitation systems.
Then we also define a term which is called nominal response of the excitation system or
excitation system nominal response. Now to understand this excitation system nominal
response, let us look at the voltage response of an excitation system that is here in this graph we
are operating so that the excitation system output voltage is represented by oa and the terminal
conditions are so created so that the voltage increases and follows this graph of this curve right.
Then on this time axis we choose a time which is equal to .5 second or this is represented by the
line ad okay then we draw a line a straight line passing through this point a in such a fashion. So
that the area of this space that is this area the area abd is equal to the area of this triangle this acd
that is you are this is the actual response curve.
Okay you draw a straight line such that such that this area abd is equal to the area acd that is the
acd is the area of the triangle then the slope of this graph represents the nominal response and it
is expressed as the slope here, the slope of the um curve is you can see the slope the slope is cd
divided by ad. You get this slope and you divide this slope by the rated voltage okay. So that you
get nominal response now there is a reason for choosing this time .5 second for defining the
nominal response the whenever whenever actually the system is subjected to the large
disturbances right and the the the swing curve which we plot the peak value of the swing curve
normally occurs in the range in the time range of around .4 second to .75 seconds and of course
this these figures are not very rigid but when we subject a system to a large perturbation or large
disturbance like a 3 phase fault.
Okay then the swing curve when you examine right the for a stable system the swing curve will
attain its peak value in a time something like .4 to .75 second and in case we the excitation
system is to come to rescue the system and improve the transient stability then it must act within
this time. In case suppose it is slow and it acts after this peak is out right then you cannot do
anything right therefore, the the .5 second is the time which is being chosen and used.
Now we will talk about the very important performance measure that is small signal performance
measures till now what we discussed was the certain definitions which are related to large
disturbance performance measures or large signal performance measures. The small signal
performance measures provide a means of evaluating the response of the closed loop excitation
system control to incremental changes in the system conditions that is here these measures which
we will discuss right these measures help us in understanding the behavior of the excitation
system when small perturbations take place right.
Now when we want to examine the performance of the excitation system when small
perturbations take place there we use a linear model that is the system is linearized around the
operating condition. Once we use a linear model we can use the linear control theory to examine
the behavior of the closed loop control system or the excitation control systems okay.
The performance measures for small signal behavior or the small signal performance measures
can be categorized or can be put in two ways one is the making use of the time response right
that is the indices associated with time response, another is the indices associated with the
frequency response that is we talk about the performance of the excitation system following
small perturbations either making use of the time response of the system or frequency response
of the system okay.
The indices associated with the time response are rise time over shoot and settling time, rise time
over shoot and settling time these are the indices indices with which we can we can quantify the
performance of or small signal performance of the excitation system. Now to understand these
terms, let us look at the at the step response of the excitation system.

Now this step response is plotted under no load condition for a closed loop control system that is
when the generator is not loaded under no load condition right. We obtain the step response, now
this this graph or this curve represents the step response of any closed loop system right in the
series this is step response okay now if you have given a step input equal to 1 per unit that is unit
step input then this is the steady state value that is under steady state condition right the output of
the systems will attain a value equal to 1 because we have given step input or unit step input
okay.
Now the method of defining this rise time, rise time is that you look at this 10 percent value that
is .1 per unit value look at this on the graph the time corresponding to the .9 per unit value. Okay
and then the time required for the for the response to rise from .1 corresponding to .1 to .9 right,
this time is called the rise time. The lower the lower the value of this rise time the response is
fast. Okay the another measure is the overshoot, you can just see here this is the overshoot
actually this is the steady state value the this is the positive peak and therefore overshoot is
defined by this value. Okay therefore, this is another you know measure for for characterizing
the dynamic response or small signal response and third is the settling time that is the time in
which the time in which the response settled to the steady state value within certain specified
limits because you will find that this response is going to have some oscillations.
Okay and the moment it comes within these limits right then this time is called the settling time
that we are not to wait till it attains the value equal to one the moment it comes in this band
which is the permissible band right that is the specified band for settling time right. Now this
band is again ah different actually but normally we can have 5percent plus minus 5 percent band
therefore once it reaches within plus minus 5 percent.

Okay then this is the therefore these are the three important parameters which characterize the
small signal performance of the excitation system the another indices which are used making use
of the frequency response of the excitation system. The indices associated with the frequency
response are low frequency gain, low frequency gain cross over frequency and phase and gain
margin. These are the three important parameters before I talk about these three parameters I will
like to highlight here that since it is a closed loop control system. Therefore, stability and the
degree of stability these are of importance okay and another thing which we need is actually in
the closed loop control system we want the low steady state error steady, state error should also
be low because suppose I want to regulate the terminal voltage right then I want to regulate the
terminal voltage within a small band and that that the steady state error will depend upon again
the characteristic of the excitation system and as you will also see actually that these
requirements are conflicting.
Suppose you want to achieve achieve a very low steady state error then the stability may have to
be compromised right therefore these are the conflicting requirements and in the frequency
response these are the 3 important indices the low frequency gain crossover frequency gain and
phase and gain margin. Now here here that need to be listed this let me see.
See the any control system any control system can be put in this form. Okay that is the G1and G2
are the transfer functions right and now what happens is that when we are interested actually
noting the frequency response of this system, we plot the frequency response of this 2 transfer
function G1into G2 right G1 into G2 or you can call it actually the open loop transfer function
open loop transfer function.
Now this graph shows shows the frequency response of open loop transfer function of a typical
system. Okay now when we plot this graph, we plot 2 curves one is the gain another is the phase
angle we plot phase angle and gain, the gain is plotted in decibel that is gain in dB. Now the gain
in dB is defined that is gain in dB is equal to twenty times the log of G1 G2 that is this G1, G2, G1
G2 is the transfer functions of the open loop system okay and G1 and G2 both are function of
Laplace transform s right and when we plot the plot the frequency response characteristic of the
system then on x axis we mark frequency frequency and on y axis we mark the gain in dB okay
this is the very standard terminology used and since G1 and G2 both are function of frequency
right therefore magnitude and phase angle of this open loop transfer function vary as the
frequency changes .
Now this graph you can just see this is a typical graph this is a typical graph I hope it is visible, a
typical graph is shown like this that is this is the gain this is plotted in dB. Okay and this is the
frequency and then you will see actually that these are plotted actually in the logarithmic scales
the frequency is also marked in log of omega n. Okay now the low frequency gain is shown here
actually when the frequency is low this is the gain right. The higher the value of this low
frequency gain right the steady state error is going to be low right because when the system is
under steady state condition right then at that time the system is not the omega becomes 0, we
can say right therefore, at very low frequencies are actually when omega is 0 steady state
condition this gain will represent the steady state error higher this value, lower will be the steady
state error. Then this is the there is a frequency here at which this gain becomes 0, this is called
crossover frequency, this is the crossover frequency.
Now similar to similar to this graph on the same diagram, we plot the phase angle characteristic
now typical phase angle characteristic is shown here like this. Now on this axis we have marked
the phase angle in degrees okay. The frequency at which the gain becomes 0 is called crossover
frequency and higher is the crossover frequency better will be the response. Okay now when we
talk about the gain margin and phase margin now to understand what the meaning of this gain
margin is, you look at the phase characteristic or phase angle characteristic.
Now you can see here actually that when this phase angle for this particular frequency the phase
angle has become minus 180 degrees. Okay now corresponding to this phase angle that is at this
frequency I find out what is the gain, for example for this phase angle is minus 180 degrees the
corresponding gain is this and this gain is called gain margin, gain margin this this quantity is
called gain margin.
Now then we also specify what is the phase margin to specify the phase margin what we do is
that we look at the crossover frequency what is the phase angle of the system at the crossover
frequency, what is the phase angle of the system and let us say phi c is the phase angle at the
crossover frequency and this difference this 180 minus phi c this represents the phase margin,
this quantity that is the phase angle corresponding to crossover frequency and the phase and this
one eighty degrees phase angle right, this difference is called phase margin.
Now for any control system higher the value of gain margin and higher the value of phase
margin higher is the stability of the system okay now we can see one very interesting thing here
that suppose suppose we want to increase the low frequency gain. Now for increasing low
frequency gain means the I increase the gain setting of the this block G1 or G2, okay now when
you increase this gain setting means this graph will go as a whole up that this graph to make this
quantity more this graph has to be shifted up.
(Refer Slide Time: 54:30

Now when you are doing this shifting in this graph up both this term this Gm will reduce that is
the gain margin will reduce and phase margin will also reduce and this very clearly shows
actually that when you are trying to tune one parameter right the other it has a detrimental effect
of other parameter, other aspects. Okay and in field actually the AVR tuning is important ah
exercise to be performed where the gain setting and other parameters are to be set so that we
obtain a certain amount of phase margin, gain margin and actually the desired steady state error.
One more important parameter they are not one but they are two more parameters which are also
important which are obtained obtained from the frequency response characteristic and here we
plot we plot the frequency response of the closed loop system that is G1, G2 right this loop is
closed that is the feedback loop is closed and the the the generator is disconnected from the
system that is generator is not loaded under unloaded condition. Okay now the graph which is
plotted here is one is this graph that is the gain gain, you can see this is graph for gain like this
another is the phase angle. Okay now the peak value of this graph that is generated by Mp, let us
say Mp is the peak value which occurs at a frequency omega M.
Now this peak value also represents the overshoot in the time response it has a relationship
between the it is a relationship respect to the time response if this peak value is high then the
overshoot is also going to be high, the another term which is used is the bandwidth bandwidth.
Now to understand this bandwidth if we plot a graph that is omega versus the gain in dB right.
Let us say for a typical I am not showing this overshoot anyway that is like this then in case
suppose I instead of plotting gain in dB let us say the actual gain, let us say this is 1.0 and the
frequency at which this drops to 0.707 right then this frequency is called bandwidth and if I plot
the graph in terms of gain in dB, gain in dB then this point will become 0 because if I if I take the
log of 1 to the base 10 it becomes 0 right and this will become minus 3that is on this in the same
graph if I plot gain in dB then this is 0 this is minus 3 and therefore this the bandwidth is
determined by plotting the okay by plotting the frequency response of the closed loop system
okay.
For any particular system to have good characteristic, the features are the gain margin should be
6 dB, phase margin should be about greater than 40 degrees, greater than equal to 40 degrees and
overshoot should not be more than 5 to 15 percent that is overshoot is around 5 to 156. Now let
me conclude here that we have discussed the excitation system requirements performance
characteristics, the performance characteristic related to large disturbances and small
disturbances. For small disturbances, we have discussed the time response as well as the
frequency response. Thank you!
Prof. M. L. Kothari
Lecture - 15
Excitation Systems (Contd.)
Okay gentleman, we start with the study of excitation systems. Today we shall cover the control
and protective functions.
The control and protective functions provided in excitation systems are AC and DC regulators
load compensation, under excitation limiter, over excitation limiter and field shorting circuits.
We shall also study today the types of excitation systems, types of excitation systems briefly.
As you all know the control and protective functions are required to achieve the desired
performance of the generator and power system we have excitation control, excitation limiting
and excitation protection that is there are 3 different functions are control, protection and
limiting. Here this is the most general block diagram of the excitation system which shows,
which shows the different control protection and limiting functions. Now let us just look at here
in this diagram this is our exciter just as usual the the exciter is a field shorting circuit, we will
discuss the need for the field shorting circuit and the output of the exciter is given to the field
winding of the generator right and it is connected to the system.
Now the different controls of protection functions are here, one is the excitation system stabilizer
circuits and just see it here. This is excitation systems stabilizer circuit here what we do is that
we take the signal from the terminal of the exciter and this is fed to AC regulator then there are a
3 limiting circuits, one is the over excitation over excitation limiter, under excitation limiter and
volts per hertz limiter and protection.
Now this say over excitation limiter takes it input signal from the output of the exciter and again
feeds this to the AC regulator, under excitation limiter gets its signal from the terminal of the
generator and again operates through the AC regulator, voltage per hertz limiter as well as the
protection right this also gets its signal from the generator terminals okay and it feeds to the AC
regulator. Okay earlier we have seen that the control functions start like the voltage sensing and
load compensation that also acts through the regulator, AC regulator. We have power system
stabilizer that also acts through the AC regulator right and in addition to the AC regulator
because in the previous block diagram, I had sawn only one regulator.
Now here we have shown two regulator one is called DC regulator another is called AC
regulator. The the DC regulator has a very limited role most of the time only the AC regulator
which functions okay but when the AC regulator is taken out because of some or the other
reasons in that case the DC regulator comes in or is put into service, DC regulator instead of
controlling the terminal voltage in generator it basically regulates the output of the exciter. The
output of the exciter is a DC voltage and in a DC voltage which is regulated by the DC regulator
while the AC regulator is concerned it has all these multiple functions control protection limiter
all these functions are through AC regulator.
Now we will briefly discuss these components before I talk about the component separately, let
me mention here that in any excitation system you may have all these functions or some of these
functions it is not necessary that any given excitation system will have all these functions
associated with it depending upon the requirements some functions may be present some
functions may not be incorporated. Now, let us just discuss briefly AC and DC regulators the
basic function of the ac regulator is to maintain stator voltage right that is the output of the
synchronous generator. The dc regulator holds the holds constant the generator field voltage and
is also referred to as manual control okay.
Now in any excitation system, any excitation system we require the excitation system
stabilization right and the the simple block diagram of the excitation system stabilizer is shown
here. This this block represents the exciter and AVR okay and the output of this exciter is the DC
voltage which is represented by the Efd the voltage which is applied to the field winding of the
synchronous generator then in order to have stable performance of this exciter and AVR, the
simplest stabilizer circuit which is used is derivative feedback, derivative feedback and this
derivative feedback is achieved through this transfer function.
We will discuss the details about this derivative feedback when I did talk about the the models of
components of the excitation system because after we discuss the basic functions, we will go to
the modeling and while talking about the modeling we will first talk about the models of various
components and then the excitation system model as a whole right. Now this feedback block you
can see here has the transfer function of this form sKF divided by one plus s times TF right.
Now this denominator represents the time delay which is provided and in the numerator, we have
actually this function sKF, the s stands for d by DT and therefore it provides a derivative
feedback signal and by properly adjusting the values of these parameters KF and TF, we can
achieve the desired dynamic performance of the exciter okay.
The next component is load compensation. This load compensation here, we derive a voltage Vc
which is obtained by adding a voltage drop Rc plus j times Xc into It 2 terminal voltage that is to
the terminal voltage of the synchronous generator, you add a voltage drop which is represented
by this term that is a complex number Rc plus j times Xc is a complex number multiply this by
the terminal current right and you perform this phasor addition and then obtained at quantity
which is proportional to the magnitude of this quantity that is Vt Et plus j Et plus Rc plus j times
Xc into It this is a phasor is a complex number find the magnitude of this and therefore, Vc is the
a scalar quantity right and this quantity this Vc represent the terminal voltage plus voltage drop in
the complex impedance or in the impedance Rc plus j times Xc.
Now in a practical system practical system the the voltage Vc is derived by this basic schematic
diagram using this schematic diagram from the terminal voltage you put a potential transformer
and get a voltage signal which is proportional to the terminal voltage in the output of the
synchronous generator right, we put current transformer and extract a current proportional
proportional to the load current right.
Now these two quantities that is this is a quantity proportional to the terminal voltage here is the
quantity which is proportional to the terminal current. Okay these are fed to a block diagram
which is called load compensator right it has a real term Rc and a imaginary term which is
represented by Xc that is this can be visualized as if you have a impedence of the value of the Rc
plus j times Xc right, practically it will not be an impedence it is only actually a a complex
number which can be used actually to derive a voltage signal Vc.
Now this voltage signal is fed to the regulator right and ultimately the field current of the
synchronous generator is regulated like this. In case in case we do not want any load
compensation then the Rc and Xc will be set to 0, okay so that what you regulate will be the
terminal voltage of the synchronous generator okay.
Now suppose if I take the Rc and Xc positive, positive then the voltage which we regulate is not
the terminal voltage but what we regulate is the voltage which is which is somewhere inside the
synchronous generator because here this that is from the what is all some voltage is generator
inside the synchronous generator right that voltage drops and we get a terminal voltage here.
Therefore, what we get by a setting Rc and Xc positive the voltage drop across the compensator is
added to the terminal voltage. Okay the compensator regulates the voltage at a point within the
generator and thus provide the voltage drop. Now this is particularly required when you have 2
generators right which are connected in parallel and we want to do the reactive load sharing
right. Now the reactive load sharing by the two generators will depend upon the internal voltage
which is generated and therefore, here we are trying to regulate the internal voltage.
In case you put Rc and Xc negative right because as I told you that you can make the real term
negative reactive term also can be made negative the compensator regulates voltage at a point
beyond the machine terminals that is terminal voltage is et right from this et we had subtracting
some voltage drop and therefore basically you are trying to regulate a voltage somewhere beyond
the terminal voltage of the machine right.

Now this type of ah compensator is required where you parallel the two generators having
individual transformers there is a terminal voltage, a generator it is on generator transformer, a
generator it is on generator transformer and then they are parallel on the high voltage side. Next
component is the under excitation limiter. Now the under excitation limiter means actually that
in case the excitation current in the field winding of the synchronous generator goes below a
certain value then this type of operation is not permissible that is we should not reduce the field
winding current below a certain set value. The basic reason is here, the under excitation limiter is
intended to prevent reduction of generator excitation, reduction of generator excitation to a level
where the small signal stability limit or the stator core end-region heating limit is exceeded.
Now these details actually one has to look into the capability chart of the synchronous generator
where if the excitation becomes very low right then the small signal stability limit is evaluated
okay and there are different names which are given to this under excitation limiter.
When we look the operation of the synchronous generator in the PQ plane right. Now this is the
P axis that is the real power output this is the reactive power output when the then it is over
excited it supplies positive Q and where the under excited it supplies negative Q and the
operating limits this is the under excitation limiter which is shown here the under excitation
limiter setting is done. So as that it is coordinated with respect to the small signal stability limit
that is this curve so the small scale of the small signal stability limit that is in case you operate in
this portion of the PQ plane the system is stable for small perturbations. If you operate
somewhere in this region that is on this side of this characteristic system becomes unstable right,
the small signal stability we will be studying separately okay.
Now keeping some margin the under excitation limit is set like this there is also one
characteristic which is also provided it is called the loss of excitation relay that is whenever the
synchronous machine looses its excitation then there is a relay which is used to protect the
synchronous generator right. The characteristic of the loss of excitation relay is also shown here
in this diagram, this is the characteristic of the loss of excitation relay right. Therefore, these
three characteristics actually that when you set the under excitation limiter settings. Okay these
have to be coordinated with respect to the small signal stability limit and loss of excitation relay
characteristic, okay and we have to see at the same time that both these characteristics are
outside the setting of the under excitation limiter. This margin whatsoever we provide that is a
safety margin, the next limiter is the over excitation limiter.
The purpose of over excitation limiter is to protect the system from overheating due to prolonged
field over current. In case, in case the field current is more than the permissible or the the the
rating of the synchronous generator right and if that current is more than the rated value for a
prolonged time then naturally because of overheating of the field winding the field winding may
get damaged right. Therefore the purpose of the over excitation limiter is to protect the system
from overheating due to prolonged field over current. This limiter is also referred to as the
maximum excitation limiter whether you call it over excitation limiter or maximum excitation
limiter they are the two different names which are used in the literature.
The over excitation limiting function typically detects the high field current condition there is a,
it has to detect the high field current condition and after a certain time delay acts through the ac
regulator to ramp down the excitation to the present value. In case, in case the current in the field
circuit is more it is detected then since actually the over excitation is permitted for some small
amount of time, okay then this limiter what it will do is it will ramp down it will bring down the
field current and bring it to the rated value.
Now this graph shows the characteristic of the over excitation limiting that is on this axis, we
have marked axis what is marked is generator field voltage per unit of rated value that is you
start from a rated value is 1, 1.1, 1.2, 1.3, 1.4, 1.5 like this and on this y axis we are putting the
time in seconds right, the times are 30 second, 60 second, 90 second and 120y second. This
gives you the, you know range over which actually the excitation system, over loading is
permitted or over excitation is permitted.
Now this characteristic A which you can see it here right this characteristic shows the field
thermal capability. You can easily understand actually that suppose I the generator field voltage
per unit of the rated value, now here the whatsoever voltage you apply right the corresponding
current will flow because under steady state, steady state conditions the field current is equal to
the field voltage divided by the field resistance right and therefore, this generator field voltage
per unit of rated value also represents the corresponding field current and you can easily see that
as this value increases the time for which it can be permitted is decreasing. This is practically this
type of characteristic inverse characteristic and the characteristic of the over excitation limiter is
to be coordinated with respect to the field thermal capability.
You have to always keep some margin and therefore, the characteristic of the over excitation
limiting is given by giving some margin, in the sense that suppose I take this example here for
example if you let us say field field voltage is 1.3 per unit okay then it can withstand this over
voltage for this much time right but keeping some margin, our limiter will operate in a time
which is less than the upper limit right. Therefore, this this is the margin which is provided a
safety margin you can call it.
Now this table shows the time in seconds verses the field voltage of the current in percent of
rated value right that is from this graph we have just picked up some values and that shows
actually that if the applied voltage is 2.08 per unit or 208 percent of the rated value it can
withstand for 10 seconds. If it is 146 percent right it can withstand for 30seconds, if it is 25
percent more than the rated value it can withstand for 60seconds and 12 percent more than the
rated value withstand for about 2 minutes is 120 seconds.
Now this is the this is the thermal characteristic of round rotor generator, turbo generator, round
rotor turbo generator because these characteristics are going to be different for different
machines but typical machines this has to be provided by the manufacturer then the next
important limiter and the protection function is the voltage-per-hertz limiter, voltage-per-hertz
limiter. Now here we all know that the the flux produced in the core is function of the terminal
voltage that is we we have a standard emf equation right therefore, the voltage is related to the
flux and the frequency okay.
Now suppose we are operating and at a situation where frequency drops and voltage rises,
voltage becomes more frequency becomes less. Okay it means v by f becomes more under that
situation the the operating flux will become more it will increase and when the flux increases it
will cause additional core losses right and therefore there is a necessity to provide the protections
against such eventuality, one simple example is that when standard generator is you know
disconnected from the load or actually the load is thrown off.
At that time the there is tendency for the machine to speed up and the frequency rises voltage
also rises right and in case the v by f ratio is beyond a certain permissible value then this limiter
will come into action and take necessary, you know action right. Therefore, basically this is
provided to safeguard the generator from the excessive magnetic flux. Now not only the
generator actually the transformer which is connected to the generator is also effected when
when a v by f ratio is more than certain value.
Now the ratio of per unit voltage and per unit frequency referred to as volts per hertz is readily
measurable quantity that is proportional to the magnetic flux that is v by f ratio but here the ratio
is taken the voltage is expressed in per unit frequency is also expressed in per unit right and that
ratio v by f not the absolute value, this per unit ratio reflects the over fluxing of the synchronous
generator this is also called over fluxing.
Here this table shows the voltage per hertz in per unit you can see it is a this is 1.051. I think this
this is slightly mistaken, no this is 1.25, 1.2, 1.15 this may be 1.1, okay this is 1.1 correction 1.1,
1.05. Now here damage time of the machines right damage times are given in minutes here. The
time is expressed in minutes instead of putting in seconds here write is in minutes.
Now damage time for generator and transformer are separately mentioned here that is in case this
ratio V by hertz or voltage per hertz is 1.25, the the generator can withstand this for .2 minutes if
it is 20 percent more than for 1 minute, it is if 15 percent it is about 6 minutes, it is 10 percent it
can be withstand for 20 minutes and if it is 5 percent more than infinite, this is the capability that
is if this V by f ratio right in per unit is 1.05 then there is nothing to worry. For the transformer
the permissible times are more, we can easily see here that instead of .2 here it is ones minute for
20 percent ah V by f right ah 1.2 times V by f is 5 minutes 1.15 it is 20 minutes and 10 percent is
infinite right. Therefore, based upon this permissible values right the setting of the limiter or the
voltage per hertz limiter settings are decided.
Now this ah field shorting is very special provision which is provided on the synchronous
generator. Especially when when we have the the excitation systems where we use the rectifiers
to supply the field current because you will you will see that we have two categories of
excitation systems ah one is AC excitation system where the output of the AC excited is rectified
and given to the field winding.
Similarly, the static excitation system right. We take the terminal voltage terminal current rectify
and give the field current to the synchronous generator right. In these 2 types of excitation
system we cannot make the field current to reverse because rectifiers are provided therefore, the
field current cannot reverse and under certain situation, certain situation it is desirable to allow
the current to flow in the reverse direction particularly during pole slipping or during short
circuit conditions where in the stator carries the stator carries the dc component whenever the
fault occurs at the terminal of synchronous generator right. The the voltages which are produced
in the field winding of the synchronous generator right if they try to drive the current in the
reverse direction.
In case we do not have this provision then the voltage produced will be very high which may
damage the field winding of the synchronous generator and therefore there are two different
circuits which are provided, one is that across the field winding you put a thyristor control field
field discharge resistance right that is this resistance is always provided actually along with the
excitation system right and the moment actually the voltage across the field winding becomes
more than a certain set value then this thyristor is switched on right and it will allow the current
to flow in the reverse direction right.
Otherwise actually the the the field current is flowing say in this direction okay and this thyristor
is off therefore, this there is no current flowing here and this field current cannot reverse it
cannot flow in the reverse direction it will flow only is in this direction. Now suppose the due to
some ah faults on the terminals of synchronous generator right in the field winding some
voltages are produced which have this type of polarity and in case this the in case actually the
this we provide a path for the reverse current to flow then this voltage will not go beyond a
particular limit. If this path is not provided then this voltage becomes high and it is dangerous to
the life of the synchronous generator field winding.
Another arrangement is that use make use of varistor, varistor which is a non linear resistance
when the voltage across this circuit is below set value or some less value then it offers infinite
resistance and when voltage goes beyond certain value it offers low resistance and therefore it
allows the field current to flow in the negative direction right. They are the two arrangements
which are provided as a special ah protection for the field winding of the synchronous generator.
Therefore with this I can say that we have discussed the various control and protection functions
provided on the synchronous generator excitation system it is not it is synchronous generator
excitation system.
Now we will briefly talk about the different types of excitation systems. Excitation systems are
broadly classified into 3 categories DC excitation system, AC excitation system and static
excitation systems. The excitation power required for a synchronous generator is of the order of
2 to 3.5 kilo watt per mega watt rating of the machine or per MVA rating of the machine.
The DC excitation systems were the oldest of all the excitation systems and the main limitation
of DC excitation system is the amount of power which can be delivered and the commutation is
the main problem in the excitation system, the more the current to be delivered, more is the
commutation problem. The in nineteen sixties the DC excitation system has been superseded by
AC excitation systems in AC excitation systems we have 2 different type of excitation systems
ah stationary rectified systems and rotating rectifier systems while we study excitation systems
we have different types of arrangements, one is the potential source controlled-rectifier system,
compound-source rectifier system and third is the compound-controlled rectifier excitation
system. Now this figure shows the schematic diagram of a DC excitation system.
The this the DC generator is the source of power for supplying DC current to the field winding of
the synchronous generator. If you see here then this represents the DC exciter this is the armature
this is the field winding the output of the DC generator is feeds the field winding of the
synchronous generator through slip rings. The DC exciter may be driven by the turbine shaft or
by separate motor also and this may be self excited or a separately excited machine. The the
voltage regulators may be a rheostatic type or they magnetic amplifier or amplidyne time and
this type of excitation systems are still existing in many of the power plants and therefore,
modeling of this DC excitation system is also important. We shall discuss the modeling of DC
excitation system separately. Now this diagram shows the AC excitation system.
In AC excitation system the the field power is obtained from AC exciter, the AC exciter is a AC
generator and this AC generator may be driven again by a separate motor or by a shaft of the
turbine. Generally in this AC exciter is driven by the turbine shaft itself. Now here in this
arrangement we are using the stationary diode arrangement that is the output of the AC exciter is
rectified with the help of stationary diodes and the DC current is fed to the field winding of the
synchronous generator through slip rings.
Now here when we use the non control rectifiers then the output of the exciter is regulated by
controlling the field current of the AC exciter therefore, if you see here in this diagram then the
AC regulator adds to control the firing angle of the control rectifier and it regulates the output of
the AC exciter. The another arrangement which is ah possible is that we use stationary controlled
rectifier that is instead of using the stationary uncontrolled rectifiers as is shown here. We can
use the stationary controlled rectifiers other arrangement is exactly the same as in the previous
case.
Now once we use this stationary controlled rectifier then the output of the excitation is adjusted
or controlled by controlling the firing angle of the controlled rectifier here itself that is this AC
regulator will directly control the firing angle of the stationary controlled rectifier and this AC
exciter provides out AC output, AC output and the control output is obtained by rectifying the ah
this DC ah AC output to DC but in a controlled fashion.
The other arrangement is exactly similar to what we discussed in the earlier case. Now in these
two arrangements which we discussed for the AC exciters, AC exciters the armature of the AC
exciter is stationary and field system is rotating while the field of the synchronous generator is
getting it supply through slip rings. Therefore, in this AC exciter we have still the presence of
slip rings and the another ah alternative arrangement which has been used is where we do not
make use of slip rings and that is called the brushless excitation system.
In brushless excitation system we have AC exciter uncontrolled rectifier and the output of the
AC exciter is fed to the field winding. Now the armature of the alternator or the armature of the
AC exciter is made rotating right the rectifiers are also rotating the field system is rotating and
therefore, we can see here that armature of the AC exciter, the rectifiers and the phase field
system they are rotating and they are connected rigidly without having any slipping contact. The
excitation of the AC exciter, how do you provide the excitation for the AC exciter? The AC
exciter excitation is provided by using a pilot exciter, this pilot exciter has the a permanent
magnet field system right and the armature is stationary.
Now in this arrangement the rectifiers are not accessible to us because they are rotating and
therefore, we cannot have controlled rectifiers in the brushless excitation system the control of
the excitation system is achieved by controlling the field current of the AC exciter that is the AC
exciter gets its field current field current through a controlled rectifier that is the AC regulator
this is the AC regulator.
This AC regulator acts on the controlled rectifier which are provided here and we supply the
controlled field current in the AC exciter and in this arrangement. We will be in a position to
control the output of the exciter by controlling the field current of the AC exciter. However,
another problem is that the the field voltage and field current cannot be measured directly that is
since this field system is rotating rectifiers are rotating and therefore, there is no way directly to
measure the field current directly and therefore if you want to get the information about how
much field current is flowing this can be obtained indirectly by measuring the field current of the
field current of the pilot exciter in the not in the AC field field current of the AC exciter that is
the output of the pilot exciter.
The major advantage of this type of excitation system is that we do not have slip rings and once
we do not have slip rings right the problem of maintenance of the slip rings. The carbon brushes
flashovers all these things are eliminated the excitation systems whether we use a stationary
rectifier excitation system or brushless excitation system the dynamic performance of these two
systems have been found to be practically identical therefore from the point of view of
performance. There is no special choice whether we choose a stationary rectifier excitation
system or brushless excitation system both these arrangements provide the similar dynamic
performance. Now we briefly talk about the static excitation system.
The static excitation system now this figure shows the schematic diagram of a static excitation
system. In a static excitation system, the excitation power is derived derived from the terminal of
the synchronous generator or from the auxiliary bus of the power plant. Now in this arrangement
the the we make use of this exciter transformer the input to the exciter transformer comes from
the terminal of the synchronous generator itself and output of this exciter transformer is rectified
using the controlled rectifier and this output is fed to the field winding of the main generator or
main synchronous generator through slip rings. Therefore, in this static excitation system, we do
require the slip rings the control is the output of the excitation system is controlled by controlling
the firing angle of the controlled rectifiers that is the AC regulator directly acts this is the AC
regulator. This AC regulator directly acts on controlled rectifiers and we get the regulated current
in the field circuit of the synchronous generator.
Now one major advantage of the static excitation system is that we do not have any rotating part
so far the excitation, excitation or the excitation system is concerned another thing is that the the
the response time of this static excitation system is very low that is they are very fast there is one
more arrangement of static excitation system.
Now here this is the main generator to get the required field current instead of making use of
only potential or power potential transformer to get the required field current. We make use of an
arrangement which is called compound source rectifier excitation system here here. The field
current in the main generator is made function of the terminal voltage of the synchronous
generator and the load current therefore, if you see here in this arrangement that we make use of
a transformer which is called saturable current transformer this saturable current transformer is
connected in the neutral circuit of the main alternator then we can obtain a quantity proportional
to the load current. We can obtain a quantity proportional to the terminal voltage these two
quantities are these terminal voltages and load current, they are appropriately combined and we
get an output which is fed to the fed to the power rectifier and the the output of the power
rectifier field gives the field current to the synchronous machine field winding.
Now the major advantage of this is that whenever suppose a fault which occurs at the terminal of
the synchronous generator. If we use a simple static excitation system using the exciter transform
only then when the 3 phase fault occurs or some sort of fault occurs at its synchronous machine
terminals right then the voltage available to this transformer becomes 0 or it dips then this will
affect the field current of the excitation system now in this arrangement. When the voltage at the
terminal synchronous generator dips but simultaneously current becomes high right.
So that, we will be in a position to maintain the field current even under the disturbance
conditions. With this I will conclude the that we have discussed today is that the various control
protection and limiting functions of the excitation system and different type of excitation
systems. We have also seen the limitations of DC excitation system, the merits of AC excitation
system and particularly the merits of brushless excitation system, the today in major power
plants; major power plants the excitation systems are either AC excitation systems or static
excitation systems. Thank you!
Prof. M L Kothari
Lecture - 16
Modelling of Excitation Systems
Friends, in this lecture we shall discuss about modeling of excitation systems.
While talking about the modeling we will study non reciprocal per unit system then we will talk
about the modeling of excitation system components because any excitation system has several
components therefore first we will discuss about modeling of the components of the excitation
system one by one. The components which we will be model, we will start with the separately
excited dc exciter, self-exciter, dc exciter, ac exciters and rectifiers as we know that in the ac
excitation system the output of the ac exciter is rectified and supplied to the field winding of
synchronous generator and therefore there is necessary to model the three phase rectifiers which
are commonly used.
The other components which require to be model are the amplifiers, excitation system stabilizing
circuit, wind up and non-wind up limits. As we will see that there exist certain limits that the
output of say AVR cannot actually certain limit like this therefore, we will have certain limits
which are imposed and they are two different categories of limits wind up and non-wind up
limits. Then to implement the under excitation limit and the over excitation limit you will have
gaiting functions and terminal voltage traducer. Over the years a variety of excitation system
models have been developed. We will confine our discussion to a few typical excitation system
models which have been documented by IEEE in the standards published by IEEE.
We will discuss the excitation models type DC1A exciter model, type AC1A exciter model, type
ST1A exciter model, type ST A exciter model. These are the 4 models which we will discuss
however, number of models are available in the IEEE standards. The over the years the IEEE has
developed the excitation system models for power system stability studies. The 4 important
references references which cover the excitation system models are the first report that is first
IEEE committee report.
This was published in the year 1968 and the IEEE transaction on power apparatus systems then
subsequently subsequently the the models were upgraded taking care of the new developments
which have taken place then the next IEEE committee report was published in the year 1981 and
when you look into the this IEEE committee report, you will see that they have 3 different
models for DC excitation systems, 4 different models for ac excitation systems and 3 different
models for static excitation systems. Therefore, in all they have discussed 10 different models
excitation systems. Now these different models were needed to cope up with the variety of
excitation systems manufactured by different manufacturers and existing in the power system
today.
Recently in 1996 another, you know a committee report came in where they have discussed
specifically the digital based excitation systems because today the excitation systems have micro
process based controls and with this digital controls right there was a necessity to model the
excitation systems and give standard excitation system models. Then one very comprehensive
standard which is IEEE standard 425.5 published in 1992, it discusses in detail the different
models which have been developed the material which we will be presenting here is drawn from
these 4 references.
Now here when we talk about the excitation system models, we have to develop the per unit
system for excitation systems. We have developed earlier the per unit system for synchronous
generator models and while developing the synchronous generator model in order to simplify the
synchronous generator equations, we had made we had made certain assumptions and the model
developed is normally known as the reciprocal model right.
Now if we use this reciprocal model there are some problems particularly if we use the reciprocal
model then then the per unit values of the exciter output voltage becomes very low. The typical
value is which is mentioned is that it may be as low as .001 per unit, if we use the same same per
unit system and therefore the another per unit system is proposed for the modeling the excitation
systems and that is now called as non-reciprocal per unit system.

In this non-reciprocal per unit system the the definition goes like this 1 per unit exciter output
voltage is required to produce rated synchronous machine armature terminal voltage on air gap
line as the definition goes that one per unit exciter output voltage is required to produce produce
the rated terminal voltages of the synchronous machine on the air gap line not on that saturation
curve but on the air gap line. Now here we define the one per unit exciter output current is the
corresponding current corresponding current which flows in the field winding of the synchronous
generator that is when you apply certain voltage at the terminal of the synchronous generator
which produces the the rated voltage at the terminal of the synchronous generator right.
Then at that time whatsoever is the current which flows in the field winding, we call that current
as 1 per unit I will just take the example suppose you have a synchronous generator whose
terminal voltage is say 11KV okay now to produce the 11 KV voltage on the air gap line if you
require say 600 volts to be applied to the synchronous generator field winding right then 600
volts will be considered as 1 per unit volts per unit voltage of the excitation system.
Similarly, suppose at that 600 volts applied the field current comes out to be say 200 amperes
right then this 200 ampere will be considered as 1 per unit field current or 1 per unit excitation
current excitation system current. Okay now with this definition, now there is a necessity to
establish the relationship between the reciprocal per unit system with the non-reciprocal per unit
system. We know that the excitation system is is interfaced with the synchronous generator both
at the field winding as well as at the synchronous machine terminals that is for controlling the
output of the excitation system, the input control signals are derived from the terminals of the
synchronous generator that is you change the terminal voltage you use actually the load
compensator and obtain a voltage VC, okay this is obtained from the terminally synchronous
generator.
Similarly you may sense the speed of the synchronous generator, okay or you sense the power
output of the synchronous generator all these things right real power, reactive power all these
quantities are fed to the AVR of the excitation system right. Therefore, so far the input to the
excitation system is concerned is obtained from terminal of synchronous generator while the
output of the excitation system is fed to the field winding of the synchronous generator therefore,
excitation system is interfaced interfaced at if with the field winding of the synchronous
generator and the terminal of the synchronous generator right and therefore we have to have the
the arrangements. So that these 2 different forms of per unit systems are properly interfaced.
Now in order to achieve this interfacing we first look at the basic open circuit equations that is
even the synchronous generator is open circuited right. We have the equations that ed, ed is equal
to 0 and eq equal to psi d which is equal to Lad into ifd. Now this equation these two equations are
the basic stator winding equations in terms of dq axis components okay.
Now we we know also that under open circuit condition this voltage or this quadrature axis
component is also equal to the terminal voltage of the synchronous generator. This was already
established, okay therefore now if I plot a characteristic relating eq or et as a function of ifd that is
the field current right then the characteristic will look like this. This is open circuit characteristic
of the synchronous generator. On this axis we have marked the field current in per unit i fd. Now
this ifd when I use this symbol this is actually the symbol used in the reciprocal system of per
units or per unit reciprocal system and we represent the per unit terminal voltage on the y axis.
Now this is the open circuit characteristic if you draw a transient to the characteristic passing
through origin, okay that is the initial portion then this is your air gap line right.
The slope of this air gap line slope of this air gap line is the mutual inductance unsaturated Ladu
that is the mutual inductance unsaturated that is the slope of this characteristic right. Therefore
now to produce on this axis I have put this ah current in per unit of i fd that is reciprocal system
and for producing 1 per unit voltage here, the current required will be current required will be
that is ifd required will be equal to that is to produce et equal to 1 that is seem same as eq right
therefore, ifd required will be 1 upon Lad and since we are considering here the air gap line we are
considering the unsaturated value of mutual inductance and therefore the the current required to
produce 1 per unit terminal voltage in reciprocal system of units right is one upon Ladu.

Now this current will be denoted as 1 per unit in the non-reciprocal system of units that is same
quantity. Okay will be said as 1 per unit okay therefore, using this information we can establish
relationship between the between the ah non-reciprocal per unit system and reciprocal per unit
system because we have seen actually that the ifd required to produce one per unit terminal
voltage is one upon Ladu right and since this should be equal to 1 per unit in the non-reciprocal
system and therefore the relationship which is established between established between the non-
reciprocal system of unit.

We call the current in non-reciprocal we will denote by ifd capital. Okay in the reciprocal system
we have been using this symbol ifd therefore, the ifd will be equal to Ladu ifd this you can check
yourself that the relationship between between the reciprocal per unit system of field of the field
current and non- reciprocal per unit system of the field current they are related by this quantity
Ladu.
Okay, that is similarly when we talk about the field voltage okay the efd the field voltage efd is
equal to Rafd ifd okay and I replace this ifd by one upon Ladu. So that the efd efd if applied it
becomes Rfd upon Ladu in per unit right and this is denoted as 1 per unit, so far actually the non-
reciprocal per unit system is concerned and hence the relationships which we establish are like
this that a non-reciprocal per unit system, the field voltage efd is equal to Ladu divided by Rfd efd in
the reciprocal system of unit therefore, these two equations that is equation 16.5 and 16.6 relate
relate the field current and field voltage in reciprocal system of per unit reciprocal system of
units to non-reciprocal per unit systems.
Okay now here I have shown in the form of block diagram, exciter model non-reciprocal per unit
system right. The we have a exciter model now the quantities will be Efd and Ifd. Okay
synchronous machine model reciprocal per unit system the quantities at the terminals of the field
winding are efd and ifd , okay the relationship between these 2 quantities that is efd is equal to Rfd
upon Ladu Efd and ifd is equal to Ifd upon Ladu right therefore whenever we develop the complete
model of the system right considering the synchronous generator and the excitation system right
the excitation system first will be modelled considering the non-reciprocal per unit system and
then it will be interfaced with the synchronous machine reciprocal per unit system using these
relations right.
Now now, we will devote our time for discussing the models of 3 basic component of the
excitation system that is separately excited dc exciter, self excited dc exciter, ac exciters and
rectifiers. Now as I have told you earlier that although the dc excitation systems right have have
been superseded by ac excitation systems or static excitation systems but still in the power
systems existing today there are number of dc excitation systems operating number one, second
point is that the when we develop the basic model for separately excited dc exciter the model for
ac exciters is also similar to that similar to that with some modifications. Okay therefore, we let
us start with developing the model for separately excited dc exciter.
Okay, and in order to develop the model for the separately excited dc exciter. You look at the dc
exciter this is the this is the armature of the dc exciter. Okay it is a dc exciter is a its field
winding okay, this is the field winding. We represent the field winding by a resistance Ref and
inductance Lef that is this field winding of the dc exciter right has resistance and inductance right.
Now the field winding is generally very highly inductive right and therefore, the inductance
plays very significant role in the modeling of excitation system.
Okay therefore we will start like this. Let us say that the voltage applied to the applied to the
terminal of the field winding is Eef and the current flowing is Ief okay. Now here the output
voltage of the exciter right we call this is as a Ex this is denoted by the symbol Ex okay. Now this
Ex is related to Iex okay and the relationship is a non-linear relationship. We all know actually
that I you plot the open circuit characteristic of a dc generator or separately excited dc generator
it comes out to be a non-linear characteristic saturation exist in the system. Further when this dc
exciter is loaded loaded right due to armature reaction the voltage at the excited terminals will
further drop right therefore, while modeling the dc exciter we may have to account for two
things, one is the loading effect another is the saturation effect and in fact while modeling the dc
exciter, we combine both the effects okay.
Now the basic equation for the field circuit is written as the Efd this will be slight a mistake is
there Efe. We will denote the symbol by Eef a mistake here. You make it Eef Eef is equal to Ref
into Ief plus d psi by dt where psi is the flux linkage which can be written as Lef into Ief this is the
basic circuit equation of the field winding of dc exciter.
The output voltage of the exciter is proportional to field flux linkage psi and Kx is the
proportionality constant that is output voltage at the terminal of the synchronous generator output
of the exciter dc exciter not synchronous generator is directly related to the flux linkage of the
field winding and is directly proportional okay.
Now this constant Kx depends upon the speed of the exciter it also depends upon the design of
the exciter and other parameters. Now here this is a most important point to understand about
modeling of dc exciter or separately excited dc exciter
If you plot the open circuit characteristic relating the Ex to Ief right then this open circuit
characteristic is shown here. Okay and the when we draw the transient to this we get air gap line
right and the slope of this air gap line, if you find out the slope of this air gap line right then it is
because this is a steady state characteristic, this is a steady state characteristic what will be the
slope of this? Resistance resistance of the field winding okay and since the we are considering
the air gap line we will consider this resistance as Rg shown here. Now if suppose for a certain
output Exo right, if I try to find out what will be the field current required under the open circuit
condition then the field current will be equal to the field current required corresponding to the air
gap line plus some additional current to account for the saturation.
Now in order to account for the loading effect right what we do is that we plot another saturation
characteristic and that characteristic is shown here. This saturation characteristic is the constant
resistance load saturation curve. You have to understand very carefully what we mean by the
constant resistance load saturation curve. Now this can be explained like this you have your dc
generator armature across this you put a resistance, okay and this is the field winding call this
current as Ief.
Okay and run this machine as rated speed and now if you plot the characteristic relating the
terminal voltage Ex with respect to Ief. This characteristic will take care of two aspects saturation
as well as the ah armature reaction effect and this characteristic is called, this characteristic is
called constant resistance load saturation curve.
Okay now for a given value of Ex the field current which is required will be now written as the
field current corresponding to the air gap line plus this additional term delta Ief this delta Ief this is
additional quantity right this takes care of saturation as well as the loading effect and therefore,
now we can say that for any operating condition any operating condition the Ief will be equal to
Ex divided by Rg plus delta Ief delta Ief.
This additional current which is required to account for the saturation and the loading effect right
is function of is function of the voltage Ex which is produced and a non-linear function of Ex that
is if you, if I try to quantify what is the value of Ief required then we can easily see here actually
in this graph that this Ief is different for different values of Ex right. Therefore, we can say that
the Ief required is is equal to the voltage Ex into a non-linear function Se Ex and this Se Ex is
called the saturation function which is dependent on the voltage Ex.
In fact this has to be obtained graphically this value whenever you want you can say develop the
model for a given dc exciter this has to be obtained experimentally. Now what we do is we have
developed basic equation starting from the field circuit equation that is Eef is equal to Ref into Ief
plus plus d psi by dt, where where psi is the flux linkage.
Now what we do is that in this expression in this the first basic expression, in this equation that is
Eef equal to Ref Ief d psi by dt, you substitute the value of Ief substitute the value of psi okay and
we will get after making this substitution an equation of this form that is E ef is equal to Ref upon
Rg Ex plus Ref into Se Ex into Ex plus 1 by Kx dEx by dt that is when we have obtained this
equation, we have made use of the subsequent relations which were derived earlier.
Now in this equation, we can see is here that the field current is eliminated there is no field
current term does not appear directly. Okay and there is no flux linkages also everything is
expressed in terms of the saturation function the field resistance of field winding and this
constant Kx. Okay now this equation is converted into per unit system of equations therefore,
what we need is the base voltage for this that is Ief, we require base voltage and the base voltage
is chosen like this that is Ex base is same as the Efd base that is the field voltage which we apply
right that base value is the base for Ex and Ief base is same as Efd base divided by Rg that is the
base value of the field current as we already seen that this voltage divided by the by the air by the
air gap by the slope of the air gap line that is resistance Rg and the resistance base is equal to Rg
itself. Okay now when you when you convert this equation into per unit system of equations that
is you divide that you divide this equation by Efd base the throughout and then you can write
down the equation in the form Eef bar bar is stands for per unit and this conversion is right. Now
it is actually the non-reciprocal per unit system.

Okay now when writing this equation, okay some new terms are introduced here which are
defined as follows that is Se bar Ex bar is defined here as Se bar Ex bar is defined as delta Ief bar
divided by Ex bar is a straight forward you know definitions which you can derive and I will
suggest you to derive these things yourself okay.
Now here here again I am showing the constant resistance load saturation curve of the dc exciter
now on this axis now instead putting putting the field current in reciprocal per unit system I am
put this is field current in non-reciprocal per unit system right PU Ief and this is per unit Ex. Okay
therefore, now you can see in this equation that for any value of Ex right the field current
required will be so much right. Now I call this quantity as A and the field current required
corresponding to the air gap line that is the, if we neglect the saturation and loading effect field
current required is denoted by B okay then this saturation function which we have defined can be
written as simply A minus B divided by B.
Okay and by the basic definition of Kx the Kx can be expressed again in terms of the per unit
output voltage of the exciter and the per unit field current. Okay now here we have a term that is
Kx is equal to Rg divided by Lef Ex bar Ief bar. Okay therefore what we do is that we define
another ah term which we call as Lfu is equal to Lef Iefo Exo bar that is for a given operating
condition right you can denote this term as Lfu that is this Lef Ief divided by Ex, you can call this
term as Lfu.
Now if you denote this term by Lfu then our equation will reduce in a standard form that is Eef bar
equal to Ke Ex bar Se Ex bar Ex plus Te d Ex bar upon dt that is you can see this equation here that
you have all per unit voltages, saturation function which we as I have told you that can be
computed from the constant load resistance ah saturation curve okay.
Now once we have come to this level we can write this expression in the form of a transfer
function model because when I say that I want to develop the excitation system model then we
can express this model this is now the excitation system model of a dc exciter. Okay now this
can be put in a compact form in the transfer function model that is what you do is that you take
the Laplace transform of all these quantities right and when you express this in the transfer
function model form it looks like this.
We have seen you just um look at this equation if you look at this equation what is to be done is
at one summing point you have the term Ke Ex to this Ke Ex, you add this term these are the two
algebraic terms, okay you add these two terms subtract from Ief that is Eef okay and when this
whole quantity is integrated you will get Ex right. Therefore this model come something like this
in the form here is Ex therefore you this is the gain KE therefore output from this block will be KE
into Ex output from this block will be a saturation quantity that is Vx which is Ex into Se Ex okay
these two are added and subtracted from Eef.

Now whatsoever comes is multiplied by this transfer function you get the Ex therefore the
transfer function model of a separately excited dc exciter is given here these quantities all these
quantities are expressed KE of TE KE and the saturation function right they can be easily
computed from the relations which have been derived. For example Ke is equal to Ref by Rg TE is
equal to Lfu by Rg and SE Ex bar is equal to Se bar Ex bar Ref by Rg right. Therefore these
quantities are all known to us and these are obtained from the parameters of the dc exciter and
the constant resistance saturation curve.
Now this block diagram which I have just shown can be further simplified and put in the form of
a first order transfer function because here we have seen actually it is a first order transfer
function therefore, the whole thing can be simplified and you can make this model as a small
perturbation model is only small change in the field voltage applied to the exciter and small
change in the output voltage delta Ex right can be written in the form of K upon 1 plus s times T
because we are whenever we represent any system right we always try to develop a model with
input output quantities okay and the the transfer function models are always always linear
models non-linear models actually we do not have anything like transfer functions okay.
Now the model which I have developed here will be applicable for one operating condition for
once you take one particular operating condition and take small perturbations around that then a
small perturbation model is put in the form of a transfer function model right these quantities K
and ST are given by these equations.
Here, K is equal to 1 upon BEX SE EFDo plus KE, I will just explain this term Bx separately and T
is expressed as TE divided by BEX SE EFDo plus KE, where the SE Ex of the saturation curve you
know this is the non-linear saturation ah function this is modelled by this exponential function a
constant AEX e to the power BEX EFDo at any operating condition at any operating condition right
we can find out the value of SE E Exo either graphically or if you have modelled this saturation
function by this this exponential function where these terms will be known to you right then that
is why actually the the time the constant K and this T are expressed in terms of the the
parameters of saturation function and all other parameters like KE is there TE is there right. Now
next point is that we will develop the model for a self-excited dc exciter okay.

Now the self-excited dc exciter will like this you have synchronous I am sorry, the armature of
the dc exciter, this is the field winding of the dc exciter. We will have some regulating rheostat in
series with this field winding right and the output of the AVR, output of the AVR will be
connected in series with the field circuit and put across the armature that is in case suppose if it is
not a regulated one right then we will connect this field winding directly across the armature but
here what we do is that in series with the field winding we connect the output of the regulator
right. So that here here we can write the Eef equal to VR plus ExVR plus Ex that is in case it was a
separately excited exciter Eef will be as same as VR but now Eef is going to VR plus Exn. Now with
this change this is the only change with this change one can develop the complete model of the
self-excited dc exciter the that is in the equation which we had derived earlier that is 16.22 we
substituted substituting equation 16.22 in equation 16.13 right that is wherever we had this Eef
we are putting VR plus Ex other things are same because that is at the only change actually in the
circuit.

Okay when this model is simplified simplified you will find actually that the same model which
we develop for separately excited dc exciter is applicable except the definition of KE now
becomes Ref divided by Rg minus one this is only difference other things are exactly same that is
this model see this model this was the model developed for a separately separately excited dc
exciter right therefore, the KE which was defined is now modified for a self-excited dc exciter
and the value of KE comes out to be Ref by Rg minus 1 and other definitions are same.
Okay now with these definitions, okay we have completed the model for dc excitation system
that is self-excited as well as and what we see here is that it comes out to be a first order model
the model is a first order model. Okay now when we go for ac excitation system, ac excitation
system the in ac excitation system we have the main source of field power is ac generator, an
alternator. The output of this alternator is rectified with the help of 3 phase full wave bridge
rectifier now this 3 phase full wave bridge rectifier may be controlled or uncontrolled.
Okay now to in order to develop this model model ah we account for the demagnetizing effect or
the armature reaction effect separately in the case dc exciter. We have accounted for the armature
reaction and the saturation together while in the ac exciter the practice is to separately account
for armature reaction and and the characteristic of the ac exciter which is plotted under no load
condition there is open circuit characteristic plotted under no load condition and the open circuit
characteristic right because open circuit characteristic always under no load condition right. The
open circuit characteristic is used here and the saturation is defined making use of the open
circuit characteristic that is you plot open circuit characteristic on this axis I am putting the per
unit exciter field current and it is the per unit VE. Now this VE is is not the voltage which is
applied to the field winding of the synchronous generator but VE is the output of the exciter okay
and the saturation function which is required here in this model the saturation function is
obtained from the open circuit characteristic. Okay now to account for the armature reaction
effect we have one more block here where this the actual field current Ifd is multiplied with the
multiplying factor KD which is called demagnetizing factor and that is KD into Ifd is added at
this point therefore.
If we see the complete model the difference between the dc exciter and ac exciter is one is this
term that is we are putting additional term to account for the demagnetizing effect, second is that
this saturation function is obtained from OCC right and third thing is that the output here is the
terminal voltage of the exciter not the voltage applied to the field winding because this voltage
which is the output of the exciter is rectified and fed to the field winding therefore there is a
rectifier in between, this is all main difference between the dc exciter and ac exciter.
Now as I have told you that the this is saturation is obtained by this formula AB minus B. The
output of the rectifier is denoted by Efd that is a function Fex a multiplying factor Fex into VE, this
is very important this Fex is very complex function and as we will see actually that we have to
model the rectifier characteristic. The 3 phase rectifiers 3 phase rectifiers when they see the
system there will be the the input impedance is purely purely inductive in nature input
impedance is purely reactive in nature and this impedance has the effect of delaying the
commutation commutation and when the there is some delay in the commutation commutation
means the change of current from one valve to the another valve right and this affects the output
voltage now for modeling the 3phase rectifiers we always define 3 different modes, mode 1
mode 2, mode 3.These 3 modes are depending upon what is the load current, how much current
it is supplying right, it is very non-linear characteristic.
Therefore to obtain the characteristic of the rectifier or this is rectifier regulation model what is
done is the output of this ac exciter is multiplied with the multiplying factor which is obtained
through this loop that is you have VE multiplied some term you get Efd, this multiplying factor as
I have just now told you is Fe Fex, Fex is the multiplying factor. Now this Fex is obtained obtained
from Ief Ifd there is a field current flowing in this field winding of the synchronous generator and
the output voltage that is you obtained a non-linear obtain function IN equal to Kc Ifd by VE that is
uh using this terminal voltage which is available what is the field current which is supplying we
obtain a current IN and this Fex function is a non-linear function of IN that is Fex is equal to a
function of IN.
Now here here in mode one mode one f IN is equal to 1.0 minus .577 IN where, IN is less than
.433, this is in the mode 1 whether operating in mode 2, f IN is equal to square root of .75 minus
IN square where IN is greater than .433 and less than .75 and we are operating in mode 3 this
function f IN is equal to 1.732 into 1 minus IN where, IN is free than greater than .75 and less than
one that is this non-linear function, okay is different for different modes and they are they can be
computed using these expressions.

While IN is computed knowing knowing this Ifd and VE and this factor Kc which is the
commutation commutation reactance Kc is the, this Kc stands for a constant which depends
commutation reactance right. Now with this I conclude my presentation saying actually that we
have developed the non-reciprocal per unit system for excitation system and we have developed
the models of the dc excitations, dc exciter that is self-excited and separately excited and ac
exciter right. We will continue the discussion on modeling in our next lecture. Thank you!
Prof. M. L. Kothari
Lecture - 17
Modeling of Excitation System (Contd.)
Friends, we shall study modeling of excitation systems.
Last time we have discussed the modeling of DC exciters, AC exciters and today we will discuss
the modeling of some more important comp1nts of the excitation system and then we will
discuss a few typical excitation system models at the end. Now in any excitation system we have
a regulator AC regulator and the output signal of the regulator is amplified right and for that we
use amplifiers. The amplifiers may be electronic amplifiers may be magnetic amplifiers or
rotating amplifiers right. Now these amplifiers, we will just see how these amplifiers are model a
basic model of a amplifier whether it is rotating or whether it is electronic can be put in this form
that is a amplifier can be presented by the gain constant KA and a time constant TA that is the
transform function of the amplifier is KA divided by 1 plus s times TA.
Now this amplifier will have the maximum and minimum limits that it respective what is the
input signal coming the output will be limited to the maximum value because this VRmax and
VRmin. For example, if you take the electronic amplifier okay there is always saturation and
therefore the maximum value which is available will be determined by the characteristic of
electronic amplifier.
Now these 2 time constants KA and these 2 constants, the gain constant KA and this time constant
TA play very significant roll in the stability of the excitation system and therefore these
parameters are required to be tuned. The next, the next important comp1nt of the excitation
system is the excitation system stabilizer, the excitation system stabilizer is a closed is a closed
loop system to stabilize the the performance of the exciter itself okay.
Therefore here, if I consider in general and exciter having its armature and field winding right
the output of the exciter we represent as Ex. Now the this 1 simplest arrangement for stabilization
which I have shown here is using a ah transformer that is you take a transformer and connect the
primary of the transformer across the output of the exciter and the output of this transformer it is
the secondary of this transformer is used to give a signal for stabilization.
Now here in this arrangement you can see here that this V error is the signal or error signal
coming from the regulator that is from the summing point. In any regulator, we will have a
summing point right and we get the error signal. Now to this error signal we separate this signal
which is the output of the this transformer, okay you consider this transformer and this signal
which you see here actually is Verr minus V2 right. This signal is amplified and fed to the field
winding of the exciter therefore here is a feedback loop.
Now this feedback loop is used for stabilizing the exciter this is called the exciter in system
stabilizer right. Now we will just see something more about how this exciter system stabilizer is
model that is we will develop a transformer function model of the exciter system stabilizer. Now
to develop the transformer function model what we require is that what the transfer function is
relating V1 to V2 that is V1 is the input signal and V2 is the output signal. We will relate this V1 to
V2, now this can be d1 by writing the basic circuit equations that is you can assume actually that
this is the primary of the transformer has resistance R1 self-inductance L1 and mutual inductance
M.
Okay, then the primary circuit equation can be written as V1 equal R1 i1 plus sL1 i1 plus s times
M i2. Okay therefore here we are using this Laplace transform variable s, okay instead of d by dt
okay therefore the basic equation is V1 is equal to R1 i1 plus L1 times d i1 by dt plus M times d i2
by dt right therefore, this derivative term we have replaced by Laplace transform operator s. The
output voltage V2 can be written as R2 i2, L2 times d i2 by dt this is again written as s times L2 i2
plus s times M into i1. Okay they are the 2 basic equations which can be written for the next
stabilizer transformer.
Now here 1 important thing which we have to observe is that the current i2 which flows in the
input of amplifier is very low any electronic amplifier or what is your type of amplifier we look
for right their input impedances is very high and therefore the current i2 is practically equal to 0
right. Now when we make this assumption that i2 is 0, we substitute i2 equal to 0 in these 2
equations right then our equations will reduce to this form, V1 equal to R1 plus s L1 into i1 V2
equal to s M times i1 that is in these 2 equations I am substituting i2 equal to 0 right and therefore
the transfer function relating V2 to V1 that is V2 by V1, V2 by V1 can be written as s M divided by
R1 plus times L1 right now you divide both numerator and denominator by R1 right then you will
get the transfer function in the form sKf plus divided by 1 plus s times Tf right where this
constant Kf becomes M divided by R1 this is M divided by R1 and this time constant Tf will
become L1 divided by R1 and therefore, we have a transfer function relating the output voltage
this will be mistake I think there is a mistake here it is V2 by V1, V2 by V1 output to input
transfer function to input V2 by V1 is equal to s times Kf divided 1 plus s times Tf right.
Now these 2 parameters again that is the gain constant Kf and this time constant Tf they are the
important parameters, these parameters are also required to be tuned. Okay further you can see
here actually that this is this represents derivative feedback, this represents a derivative feedback.
In fact in this arrangement, in this arrangement this transformer primary winding is designed to
have high resistance so that when there is no variation of this output voltage, there is no variation
of this output voltage right then there will be no output coming on the secondary of the
transformer right. Therefore since this is the this is the DC voltage here right therefore as as long
as there is fluctuation or variations are taking place in this output voltage there will be a
corresponding feedback signal coming otherwise, the moment actually the we have steady state
operation right then then there is some DC current flowing in this circuit and there will no output
from this stabilizing transformer.

Now this is 1 arrangement which is used for realizing the derivative feedback although there may
be alternative arrangements to achieve this derivative feedback. Now here, we will just briefly
talk about the limits in the literature documented by IEEE standards or the IEEE standards. We
define the 2 types of limits, 1 is called windup limits, another is called non windup limits. Now I
will just distinguish between what we mean by the windup limits and non- windup limits in
windup limits to illustrate this what we have taken is that we have consider an integrator a
transfer function of an integrator. Now in the windup limit this is the transfer function of the
integrator 1 by s, okay input signal is u output is v. Okay then we are putting actually the limiting
functions LX and LN and the resulting output is y that is the output form this complete limiter is y
while input is u right.
Now here in this arrangement when we look actually this transfer function model of this
integrator then the we can write down equation dv by dt equal to u that is this transfer function
basically represent this first order differential equation. Okay dv by dt equal to u right, now here
in this arrangement in case this v that is the signal v which is coming here that is output of the
integrator if v is greater than LN and less than LX that is it is in this range it is lying in the range
between maximum and minimum then y is equal to v, this y is same as v.
In case this v the signal that is output of the integrator is greater than equal LX then y is set equal
to LX. Okay that is this derivative this derivative of dv by dt equal to u this when it is solved okay
if you get actually the signal v if it exceeds this upper limit LX then the output will be set equal to
LX. This is very simple actually any system any system right the basic concept of the limiter is
that if the signal is exceeding then we set it to the upper or lower limit.
Similarly, if v is less than equal to LN we set these 2 the lower limit LN right. So that so that the
they the device which is used actually to ah give output v or basically the output y right right, We
will have this ah output confined to the upper and lower limit and when this output is less than
this and greater than this the v is y is equal to v this is very simple this is called actually windup
type limits. There are certain devices when they are used for limiting right then the integrator
function of those devices can be represented by this model.
The another limit or another type of limit is called non windup limit, a non windup limit the the
symbol which is used to represent a non- windup limit is we have this integrator transfer function
1 by s here. The input is u, output is y and instead of putting it outside this LX and LN that is we
had put earlier here. Now they are put along with the block the idea here is basically that this
output y, output y which is going to be produced produced will be confined between LX and LN
right and this out output will never be equal to the what we have given actually in the previous
equation that v, where the v was shown here right therefore there is a limit which is imposed by
the device itself right.
Now the equations which describe the this type of limit is that if y is greater than LN or less than
LX then dy by dt is put to be equal to u that is in this function, if input signal that is the the you
have the output signal y which is within the limits then dy by dt is put to be equal to u right in
case this output is greater than LX then dy by dt becomes 0 and and y equal to LX.
Similarly if y is less than equal LN and dy by dt is less than 0, okay that is it is it is less as well as
this quantity is less than negative then dy by dt again set to 0 and y is set equal LN. This is the
basic difference between the 2 types of limits that is the windup and non windup type limits
therefore, when you lead the read this literature on excitation systems you will find that these 2
terms are frequently referred.
Now another next component of the excitation system model will be discussed now that is the
low voltage gate and high voltage gate that is HV gate and LV gate. Now we have seen actually in
our discussion on excitation systems, we have to incorporate the under excitation limit and over
excitation limit. Therefore, basically this is a gate where we get we put 2 inputs u and v right the
for for LV gate, LV gate the condition is that if v if v is less than equal to u right then y is equal to
u that is out of the 2, suppose u is less than v then output will be equal to u.
Similarly, if u is greater than v output will be equal to v that is ah a very simple ah gate where we
have 2 inputs 1 output the this gate will always ensure actually that output is equal to the smaller
of the 2 inputs and the next type of gate is called HV gate, HV gate where input signals are again
u and v output is y right. Now here if u is greater equal to v right, y equal to u it means the output
is equal to the larger of the 2 signals, if u is less than v, less than v then output is y is equal to v
because here v is greater v is bigger than u right. Therefore, these 2 gates are also required
actually to model the complete excitation systems because we have to we have to implement the
under excitation limit and over excitation limit and these limits are ah realized by using these H V
and LV gates.
The next important component of excitation system is the terminal voltage transducer and load
compensator, terminal voltage transducer and load compensator. We have discussed is our
previous discussion the need for load compensator and the function performed by the load
compensator. Now the this block is used to compute a voltage VCI, VCI which is derived by by
adding a voltage drop Rc plus j time Xc into It to the terminal voltage Et right and therefore, so far
this block is concern input will be the terminal voltage a phasor quantity, It is the terminal current
of the machine and these 2 quantities are processed here in this block and we get a signal VCI
which is a magnitude of this this signal Et plus plus Rc plus j times Xc into It okay.
Now this signal VCI okay is when it is processed right there will be some time delay in the
processor and therefore this time delay is represented by this transfer function 1 they are by 1
plus s times TR. Now this time constant TR is generally very small in many cases of a many
studies or many simulation studies, this time constant can be even neglected but its value is
generally very small therefore we can say that the terminal voltage transducer and load
compensator is having 2 blocks, the first processing block where we get a signal which is
obtained by computing this expression and taking the magnitude of this expression and then it
passes through the first order transfer function which represent the time delay of the transducer
okay.
Now we have come to a stage where we will be in a position to discuss the complete model of
the excitation systems. Now for discussing this complete model, let us let us again look into the
ah models which were developed for DC exciter and AC exciters because I will be discussing
only 3 models, 1 is known as DC1A model IEEE DC1A model another is IEEE AC1A model
and third is IEEE ST1A model. In the IEEE, IEEE standard 1992 right it has large number of
models, now the the substitutes which are put here I will just mention why these different
substitutes are left that is DC1A that is initially when the first publication came they were called
as DC exciters or DC IEEE DC model or IEEE DC exciter model, IEEE AC exciter model, IEEE
ST or static excitation system model.
Then in 1981the IEEE gave another report where to distinguish this first generation model from
the second generation model they put as subscript 1 they called DC1, AC1, AC2 like AC2 like
that therefore 10 different models were given then in 1992 when they further revised right these
models were added 1 more subscript that is DC1A that is A was added then somewhere in 1996
when another we can say report came where DCA was replaced by B. Therefore you will find
actually in the last or the recent report they have used the symbols say DC1B or AC1B, B stands
for that somewhere in 1990, we distinguish from the different models.
Now here since it is not possible to draw this model in totality on 1 sheet right therefore, what I
have d1 here to illustrate, we will take we will represent actually the various building blocks
separately and these blocks will be assembled. Okay now if you look actually the DC exciter
model then in this DC exciter model we can say that output is EFD or EX and input is the output
of the regulator that is VR that is Eef equal to VR therefore I will be representing this DC exciter
model with the help of a block okay, by saying that input is VR or output is EFD. Okay therefore
to draw all these things in the same sheet becomes very complicated and it is not possible to
show in a small sheet.
Now when you look at the complete model for the IEEE type DC1A excitation system that is the
IEEE type DC1A excitation system is shown here. This block represent the DC exciter model
which is now we have developed right which has saturation function which has 1 upon s time T
and also the constant K right all these are part of this model, output is EFD and input is VR. Then
we have 1 model is called amplifier model that is we require the amplifier right therefore I am
just showing the amplifier with non-windup limits that is here in this block diagram the amplifier
is transfer function is KA divided by 1 plus s times TA with those VRmax and VRmin limits then
input to this amplifier comes through an HV gate, HV gate and here here this HV gate will get a
signal from a transfer function this is normally called compensator that is that is for obtaining the
desired performance of the DC excitation system, we may add a compensator in the forward loop
okay.
Therefore this is basically a compensator, okay. The output of this compensator is compared with
the with the under excitation limit and we know actually that where this under excitation limit
okay the output will be consider this this is realized using actual gate that is output will be
corresponding to the higher of the 2 signals. So that we will realize actually that the the
excitation system will not go below a certain limit prescribed that is the under excitation limit.
Now here I have put the star here to show that this under excitation limit can be realized by
feeding the signal at the summing point also therefore there are 2 alternative ways of feeding this
under excitation limit that is this signal if it is not required to be fed if it is fed somewhere right
summing point. Okay then we can get rid of this HV gate now the first summing point which I
am shown here we have reference voltage the output Vc that is Vc is coming from terminal
voltage transducer and load compensator and from the reference we are subtracting this Vc that is
the output from the terminal voltage transducer and load compensator.
Okay then this Vs is the auxiliary signal which comes from power system stabilizer right
therefore in this summing point. We have these 3 signals shown then the next summing point.
We have put the derivative feedback is VF is the output of the excitation system stabilizer okay
and this VF is subtracted from the signal which is the V reference minus Vc plus Vs and the under
excision under excitation limit can be injected here itself right therefore, there are 2 alternative
ways of realizing the under excitation limit Now this this read feedback transfer function is what
we have derived earlier the s times K upon 1 plus s times TF okay.
Now this transfer function model represents the both both self-excited, exciter self-excited DC
exciter and separately excited DC exciter. The only difference is in the in the parameters of the
model particular the Ke right. Now here one interesting thing which is d1 here is that when we
have a self-excited DC exciter, we set the parameters in such a fashion so that we get the desired
value of this field voltage with VR equal to 0, under steady state condition, steady state condition
right the exciter ah field rheostat is set is such a fashion. So that the desired value of field voltage
is obtained with with output from the regulator equal to 0 that is VR is equal to 0 with VR equal
to 0. We should be in the position into get the required value of EFD.
The typical parameters of the DC1A exciter model are given below that is the parameters KA is
187 TA is .89, TE is 1.15, AEX equal .014, BEX is 1.15, KF 0.058, TF 0.62, TB .06, TC .173, TR
0.05 and VRmax 1.7 and VRmin 1. minus 1.7, the KE is computed so that initially VR equal to 0 and
load compensator is not used that is RC and XC will be set equal to 0 when they are not using the
load compensator RC and XC will be set equal to 0. Now if you see here actually then for this
excitation system model we have large number of time constants gain settings right and the
dynamic performance of this model is depends upon this time constants and when you design the
excitation system right then these parameters have to be properly optimized or tuned.
So that we get a required dynamic performance this is a quite an important exercise which is to
be done right at the design stage and when you install this excitation system the excitation system
parameters are required to be tuned to achieve the required dynamic performance, I have talked
about what we mean by the required dynamic performance of the excitation system, could you
just tell me what are the required dynamic performance suppose I look from the point of view of
say frequency response of the system what are the parameters which are set for right that is that
is let me just mention here that the 3 parameters which we call is the gain margin, phase margin,
bandwidth and overshoot right and the MP which was the overshoot that was to be 1.0 5 to 1. 15,
phase margin should be greater than equal 40 degrees, similarly the gain margin should be
greater than equal to 6 DB and it should have a high bandwidth, higher the bandwidth better will
be the time response right. The next model we will discuss is the model for AC exciter.

Now here, let us again look into the model of the AC exciter which we had developed earlier.
The AC exciter model is similar to the DC exciter model except it has additional features there in
this AC exciter model the demagnetizing effect of the load current is accounted by putting a
constant KD and the field current IFD right this is 1 point, second is here output of this exciter is
1VE VE which will be further processed through the rectifier or will be rectified and we fed to the
field winding of the synchronous machine right therefore, while modeling the complete AC
excitation system I have represented this portion in the form of a block diagram I will say this
portion when I just put from this portion this whole thing I {represin} ((00:36:54 min)) I am
representing in the form of a block diagram that is you can see here the input is VR 1 output is
VFE, I am showing this VFE separately in the model VE and input is IFD. Okay now this to about
the congestion on the same diagram, the another very important component of this AC excitation
system model is the rectifier regulation model.
In this rectifier regulation model again input is VE and field current IFD the output is EFD
therefore, again I will be showing this as input output model showing this VE as the input, IFD as
the input and EFD as the output. This we have already seen actually that to to account for this
rectifier voltage regulation the process is that we take the field voltage which is there which is
flowing in the field winding of the synchronous generator. We take the output voltage of the AC
exciter these 2 quantities are fed into this block to compute a current IN this IN equal to KC IFD by
VE.
Now this constant KC, KC is function of commutating reactance, KC is the function of

commutating reactance. Okay then after computing this IN we fed this to a non-linear transfer
function that is FEX equal to f of IN and the output of this block that is FEX, FEX will be the out of
this block right it is multiplied with VE to get EFD that is ultimately we can say that the as the
loading of the exciter increases right or depending upon the loading right the the voltage which is
the applied to the field wining of synchronous generator and the voltage which is produced at the
terminal of AC exciter they vary and this variation is determined by this multiplying factor FEX
that is FEX into VE equal to EFD. Okay therefore so far this model is concerned I have shown as
input as VE and IFD and output EFD.

We had also seen that then when the rectifier operates in different modes, different modes the
function fn is defined as I function of IN that is this function is the function of IN is defined as
mode 1 operation, 1 minus 0.557 into IN if IN is less than equal to 0.433 that is this 3 phase
bridge type rectifier right is model model right. So that we identify 3 different modes of
operation mode 1 mode 2 mode 3 right and in mode 1 the IFD is such that or the field current
IFD is such that the IN which is computed quantity is less than equal to 0.433 right and this is the
mode 1 operation. Similarly in the mode 2 operation the this non-linear function is square root of
0.75 minus IN square where, IN is in the limit .433 and .75 okay.
In the mode 3 operation this this function f IN equal to 1.732 into 1 minus IN where IN is greater
than .75 greater than equal to .75 and less than equal to 1 right. Now I have given you the
expression for computing this quantity IN earlier okay. Now the as I told you that in AC
excitation systems we have different possible arrangements the model which I will be discussing
here that is AC1A model that is IEEE type AC1A model right.
This model represents the the brushless excitation system it represents the brushless excitation
system. Now in the brushless excitation system we have rotating AC exciter armature, rotating
non control rectifiers and the field system right. This is the rotating part and the the control is
excised here by controlling controlling the field current of the AC exciter you will just see here
actually that this is the AC exciter the field system is stationary okay and by having a control
current in the field winding of the AC exciter we control the output of the brushless excitation
system therefore the model which I am discussing here that is AC1A model right, it pertains to it
pertains to the brushless excitation system.
Similarly, there are other AC excitation system and the other models represent the other type of
excitation systems because this variety of excitation system models have been given to take care
of the different types of excitation systems right because from 1 excitation system to another
excitation systems variations are there is it clear.
Now this block diagram gives you the complete model for the IEEE type AC1A excitation
system, this is the IEEE type AC1A excitation system model. Now here you can start from the
output side. Now this block diagram represents the rectifier regulation model rectifier regulation
model. We have seen actually that rectifier regulation model input quantity was VE and field
current IFD that is you have to sense the field current of the synchronous generator and this field
current is fed to this and output is EFD. Okay therefore when you want to draw the complete
diagram if this block diagram can be replaced by the AC replaced by the rectifier regulation
model, okay which had which had actually various components.
The components again let us see the comp1nts here were the computation of IN and this non
linear function right therefore basically you compute the value of IN there will be logic function
here. Okay and this will give you the output FEX and ultimately you get the output EFD right this
is the rectifier regulation model. Then the next block is the AC exciter model I have shown
actually that the output of the AC exciter model is VE which is the terminal voltage of the exciter
VE the input were IFD, IFD and the the output from the regulator that is the VR, VR is the therefore
here this summing block I have shown separately. Okay therefore in the summing block VR is
compared with VFE, VFE which is taken as output from AC exciter okay.
Okay, therefore the these 2 blocks can be represented by the detailed model to show the
complete excitation system, this AC excitation system, AC1A model of the excitation system.
Now the rest of these things we can start from the summing point V reference is always required
any excitation system model the V reference is with the positive polarity VC is with negative
polarity that is the output of the terminal voltage transducer and load compensator. We add the
stabilizing signal that is the auxiliary signal Vs the power systems stabilizer output signal Vs that
we will show separately when we talk about power systems stabilizers and the excitation system
stabilizer that is VF this will come as a negative signal therefore here I have shown in the same
summing block all the signals these are the reference voltage, the terminal voltage, the auxiliary
signal that is the power systems stabilizer signal or there may be some other auxiliary signals
also right and the excitation system stabilizer.
The excitation system stabilizer input here VFE, VFE while when you looked into the DC
excitation system right the input to the excitation system ah stabilizer was EFD that is the voltage
is applied to the field winding state here but here since this is a non-linear function when there is
nothing to be stabilized in this portion right because this relationship is a fixed relationship
between the VE and EFD we have a fixed relationship right and therefore input signal is taken
from the AC exciter that is VFE which we had shown earlier.
This is again a compensator which may or may not be required right this is again for the purpose
of proper stabilization of the whole thing then the amplifier model. We have HV gate for
realizing the under excitation limit UEL HV gate then the LV gate to realize the over excitation
limit. Okay and therefore this becomes the complete model of the field controlled rotating
rectifier excitation system right. Here as I have told you actually that the control is exercised by
controlling the field current of the exciter right. In case you have the stationary rectifier system
stationary rectifier system then control is transferred from field circuit to the rectifier.
You can use control rectifier with different models have different different excitation systems
will bee controlled in a different manner because in the in the case of a brushless excitation
system since the rectifiers are rotating they are not accessible therefore control cannot be put on
the rectifier therefore, they are non controlled rectifiers, the control is on the field circuit of the
exciter.

The last model, we will discuss will be the static excitation system but before I talk about the
static excitation system the typical parameters of AC1A exciter model are given here KA 400, TA
0. 0 2, this in this particular case TB and TC are set to be 0 it means compensatory is not required.
KF is 0. 03, TF is 1.0, KE is 1, TE is .8, KC .2, what is this KC? A parameter which is function of
commutating reactance and it is required to take care of the rectifier regulation characteristic, KD
equal to 0.38 what it this KD, KD we have discussed number of time.
This is to take care of the armature reaction effect of the alternator that is AC exciter is an
alternator right therefore KD is 0.38 maximum minimum limits are 0. 7 .33 and 6.6 that is minus
6.6 6 that is minimum. VAmax the output of the output of the amplifier can be 15 or minus 15 and
the saturation function is modelled by AEX equal to .1 and BEX equal to .03 right again these
parameters are required to be tuned actually and these are the typical tuned parameters of a
particular system therefore this they are representative this should not be considered to be the
optimized parameter that just representative to give some idea about the typical parameters of the
excitation system.
The last excitation system which we will talk about the static excitation system model, the static
excitation system model here I will discuss the excitation system model with exciter transformer
we have a controlled rectifier that is the excitation system power is derived through a exciter
transformer this is a 3 phase transformer right and the output of this is rectified using controlled
rectifier and fed to the field winding of the machine, of the synchronous machine through slip
rings right and therefore our regulator acts on controlled rectifier that is AC regulator which gets
inputs from current transformer po10tial transformer this will directly act on the control rectifier
this is the very simple excitation system model time is over actually. I will just complete here of
course I may have to complete afterwards some more portion time is little bit over actually right.
The the model looks like this the model ah looks like this it has all the functions similar the
summing point is same your V reference, VC under excitation limit stabilizing signal, the
excitation system stabilizer right. This signal is called VI this is the whole summation or VI it is
fed to HV gate this is fed to HV gate to realize the under excitation limit then there is a
compensator which is put here again. Okay then amplifier and the output signal which comes
here is VS.
Now another one which is shown here is that the stabilizing signal can be put at this point as or
add the summing point itself there are 2 alternative ways you can put the stabilizing signal here
right and the this is fed to under excitation limiter then we have over excitation limiter and we
get the output voltage from this. Now here this EFD max is function of terminal voltage that is Et
VRmax and the KC the constant depending upon the commutating reactance and IFD that is this
maximum voltage is function of terminal voltage because because the potential transformer takes
its input from the terminal voltage itself right and minimum is Et into VRmin. Okay it also has a
field current under limiter field current limiter also is provided here right. This field current
limiter it relays that you take this limiting value uh compared with the field current then you
have some gain setting and fed here.
Now this completes the presentation of the different type of excitation system models and I
suggest you to look into the IEEE standard which is published in 1992 right for further details
about the excitation system models. Thank you!
Prof. M. L. Kothari
Lecture -18
Small Signal Stability of a Single Machine Infinite Bus System
Friends, today we start with the study of small signal stability. We shall devote our discussion
towards the small signal stability of a single machine infinite bus system that is for
understanding the concepts related to the small signal stability, we will consider a simple system
and that is a single machine connected to an infinite bus. Okay the small signal stability is the
ability of the system to maintain synchronism under small perturbations. The small perturbations
continuously occur in any power system due to changes in loads and generations.
Now for analyzing the small signal stability of any system, the system model can be linearized
around an operating point that is the disturbances are considered to be so small or incremental in
nature. So that we can develop a linear model of the system around the operating point, once we
develop the linear model of the system we can understand the behaviour of the system under
small perturbations, various parameters of the system which affect the stability of the system
further the moment we have a linear model we can apply the linear control system theory for
designing the controllers.
Now here when I talk about the controller particularly we are interested in designing the
excitation system control that is the voltage regulator and the power system stabilizers as we will
see that in any power system right, the actual system is some what complex it is not as simple as
a machine connected to infinite bus is always a multi machine system right. Now in any multi
machine system as you know that the system will have different modes of oscillations, the modes
of oscillations are classified as as local modes of oscillations, inter area modes of oscillations and
the control modes okay.
Now, primary requirement of the system is that this system should have for stability this will
have positive synchronizing torque coefficient and positive damping torque coefficient. The
stability of the system will be affected if any of these two torques or any of these two torque
coefficients become negative. Okay now to start with we will study first a simple model
considering constant flux linkages in the field winding that is we will first start with the constant
flux linkage model, next we will include the the field winding dynamics that is we will consider
next step the constant field voltage and must we assume these voltages applied to the field
winding as a constant okay we the model is developed so that we take care of the the dynamics
of the field winding that is changes in the field flux linkages.
This particular stage when we say that the applied voltage is fixed means it is a practically a
manual control after studying this model with constant field voltage, we will extend it to the
model which will include the automatic voltage regulator and we will study the affect of the
parameters of the excitation system and gain setting of the automatic voltage regulator on the
stability of the system. Then the next step will be we include auxiliary controllers that is the
power system stabilizer that is this the stage by step development will give us the complete
insight into the small signal stability problem.
The system which we consider for small signal stability to start with is a machine infinite bus
system where a synchronous generator is connected to a large system which can be characterized
by a infinite bus, the equivalent circuit can be represented as generator the we represent the
terminal voltage of the generator, the line is represented by an equivalent impedance Z
equivalent equal to RE plus j times XE and connected to an infinite bus right.
Now here here even in a multi machine system, multi machine system right. Now if you if you
take out one generator and the connected line right then just will be system if you represent by a
infinite bus then it becomes a machine infinite bus system. Now as we continue our studies to
simplify our model and understanding the resistance of the line or the equivalent resistance R E
we may may be ignored.
Now when we consider a classical model that is we assume the field flux linkage is to be
constant. Okay now when we consider the classical model of the synchronous generator then the
classic the synchronous generator is represented by a constant voltage behind direct axis
transient reactance right. Now here we will represent the voltage behind transient reactance by
the symbol E prime and the voltage of the infinite bus as EB, okay and the phase angle, phase
angle between E prime and EB is denoted by delta right and that is the power angle of the system.
Now for this machine infinite bus system we have derived earlier the the power angle
characteristic okay the power angle characteristic to derive we can start with the the terminal
current of the synchronous generator as the E prime minus EB. Now here we are representing the
ah voltage behind transient reactance if it is represented as the as E prime angle 0 that is you you
consider this as a reference voltage generally we do other way around we consider the infinite
bus voltage as reference and the internal voltage leads the reference voltage by some angle delta
okay but it does not matter very much.
We can write down the expression for current It this is the first step for solving this problem by
this what we know will be the the you know whenever we start solving the problem the terminal
voltage of the machine will be known to us right and therefore, once the terminal voltage is
known we can find out the internal voltage by this formula E prime equal to Eto plus j times Xd
prime into Ito, where here now we denote this total reactance Xd, XT as Xd prime plus XE earlier
the power angle characteristic which we derived was relating the terminal voltage with the
internal voltage but now here we will be deriving the power angle characteristic relating the
infinite bus voltage with respect to the internal voltage. Okay therefore the total reactance which
comes will be Xd prime plus XE where XE is the reactance of the transmission line okay.
Then the complex power behind Xd prime is given by S pi, S prime is equal to P plus j times Q
prime you can call it P prime. Okay this is E prime into It complex I am sorry It conjugate, It
conjugate that is the complex power is P prime plus j times Q prime and this is obtained by
multiplying this voltage E prime by conjugate of It, okay when you make the simplifications.
We can write down the expression for air gap torque. Now we all know actually that in case we
neglect losses right then the air gap torque is equal to the the power output impairing your
system that when you consider the per unit system then the power output at the terminal of the
machine is same as the air gap torque. Okay and this is our standard power angle characteristic,
now here at this stage I would like to mention that in case instead of having a non-salient pole
machine, if we have a salient pole machine right then the power angle characteristic will be
different from what is given here, it will have a another term which is known as the reluctance
torque. Okay and one can write down the expression for the power angle but this is the primary
requirement to start with. The next step is you will linearize you linearize the expression for air
gap torque around an operating point.
Our initial operating condition is characterized by the angle delta equal to delta naught that when
we are operating and transferring certain amount of power right the initial value of delta is delta
equal to delta naught, okay. The linearized equation is written in the form delta Te is equal to
partial derivative of T with respect to delta into delta delta where this expression is very simple
there is only a Te is function of function of one variable only here. We are assuming this E prime
and EB constant, okay. The infinite bus voltage assumed to be constant the voltage behind
transient reactance is assumed to be constant the total line reactance is constant and therefore the
torque is function of delta only one variable and therefore when you obtain the small change in
the air gap torque delta T is equal to partial derivative of delta T divided delta delta into delta
delta.
Therefore, here basically this partial derivative becomes derivative term itself right there is no
other you can say parameter on which the Te depends okay but in case actually these are also
variables right then when I talk about this term it is the partial derivative with respect to delta not
delta when you take other terms also then you have to write the partial derivative of this quantity
with respect to not delta say when you take this partial derivatives there there so many variables
will be there.
You have to keep on taking all the variables then you will have delta delta delta E prime delta E
prime like this it is basically it is actually the Taylor series exponents time to expand the
equation around the operating condition right. Now we come to the basic equations of motion,
we have developed earlier the swing equation of the synchronous generator and the swing
equation is a second order non-linear differential equation. This equation can be represented by
two first order linear I am sorry 2 first order non linear differential equations it { it will continue
to be non linear.
These two equations are written here as p delta omega r equal to 1 upon 2H Tm minus Te minus
KD delta omega r and second equation is p delta equal to omega naught delta omega r, where
delta omega r is the deviation of the rotor speed with respect to synchronous speed right. The KD
is the damping torque coefficient to for the sake of completeness right, we include damping term
also well earlier when we discussed actually the KD was not explicitly included in the equation
but the complete equation is in this form.
Now when the equation is written in this form the speed deviation the delta omega r x is is
expressed in per unit delta omega r is expressed in per unit the mechanical torque, electrical
torque are also expressed in per unit. We know how to define the initial constant H the time in
this equation is the time instead of putting in per unit we always prefer to express time in second
is right and therefore, the second equation has p delta equal omega naught into delta omega.
Now these are the 2 basic equations around which the the we will be you can say developing the
small signal model using these equations, small signal model and studying the basic concepts of
small signal stability. Now you what we do is that this this these two equations that is 18.5 and
18.6, we will linearize around the operating condition okay.
Now to linearize this around the operating condition one assumption here which we are making
is that there is a mechanical power input remains constant that is Tm is constant right and Tm is
equal to the initial power output that is Teo. Okay and therefore when we talk about the small
perturbations our equation will become p delta omega r equal to 1 upon 2 H delta Tm minus Ks
delta delta minus KD delta omega r this delta Tm delta Tm is considered to be a change in
mechanical torque.
Okay but generally this change is included in the model included in the model to obtain the
dynamic performance by giving a small change in the mechanical torque right. Otherwise the
mechanical torque remains constant this delta T term has been replaced by Ks delta delta and
this term Ks is synchronizing torque coefficient right and what will the unit of this coefficient
unit per unit torque. Now in this equation actually be the delta delta is expressed in radians
electrical radians. Okay similarly, the unit for KD the torque in per unit per, sir in previous
equation T delta equals to omega r naught into delta omega r, so all this equation comes because
T delta equal to I think omega naught plus delta omega delta r. Now if you see the basic equation
it is like this, d delta by dt is equals to omega minus omega naught that omega naught is the
synchronous speed okay. Now what we do here is that this difference we do not by in fact
actually this speed actual speed is called omega r is the speed of the rotor. Okay this will do not
this difference is noted by delta omega r. Now what we do is that divide this by omega naught,
so that this quantity becomes in per unit and multiplied by omega naught.
So that I can represent d delta by dt as as delta omega r into omega naught while this delta omega
r is expressed in per unit. Okay while in this equation in this whole equation delta is expressed in
electrical radians because when you solve these problems, you have to be very clear about the
units. Okay see originally when you see the our swing equation right, it is of the form d omega
divided dt d omega r divided by dt 1 upon 2, I am sorry 2H divided by s 2H divided by omega r
omega r, okay equal to accelerating power okay.
Now here when I express this omega r in per unit this is this is omega naught not omega r this is
omega naught this is omega r right then this term is combined here. Okay so that I have this
coefficient equal to 2H only that is that is when you write the equation in this form, 2H d omega
r by dt equal to Pa. Okay here the this omega r is a per unit further omega r can be represented as
as omega naught plus delta omega r that the omega naught is in this particular case when it is per
unit omega naught will become 1 that is you can put there is a 1plus delta omega r therefore,
when you substitute this expression here I can write down the equation as 2 times H d by dt of
delta omega r equal to Pa, is it okay that is that is since we know actually the derivative of this 1
is 0 therefore d by dt of delta omega r is same as d omega r by dt and that is why when you look
at the swing equation which we have written here.
Okay original swing equation this is not a linearized equation we can write down this delta
omega r. Okay is it clear actually ah these 2 equations are very important actually because when
you solve a given problem right the it is very important to understand that how the quantities are
expressed what quantities are in per unit? what the unit of time? what is the unit of delta? what
is the unit of speed deviation? all these terms are very important. Okay now when you look at
this equation.

This is a equation in terms of small variations of speed angle and mechanical torque, okay and
therefore we can take the Laplace transform of this equation right. When you take the Laplace
transform it will become something like this s times delta omega r, s equal to 1 upon 2 H delta
Tm (s), okay minus Ks times delta delta s minus KD times delta omega r s that is in this equation
all these variables are the Laplace transform. Okay and therefore now what we can do is that we
can write down we can write down delta omega rs as 1 upon 2H times this whole expression
okay.
Now this equation is represented in the block diagram form here at this summasing at this
summation point we have delta Tm Ks times delta delta that is your delta Te that is delta delta is
here delta T and KD times delta omega r. Now when I represent this in the block diagram form
all these variables are the Laplace transform but for the sake of convenience, we may not write
ah delta delta as delta delta with s. Okay but we understand that in the in the block diagram or
the transfer function model all the variables which we represent they are the Laplace transform
of the original variables. Okay delta Tm also delta Tm (s) therefore this is the summation point
where we derive a quantity which is given in this bracket delta Tm minus Ks delta delta minus KD
delta omega.
Now this quantity when it is multiplied, now here it is 2 times Hs when you multiply this by this
block is one upon 2 times Hs right, we get delta omega r here okay. Now the next step here is to
linearize the the second equation. Our second equation was p delta equal to omega naught into
delta omega r but this delta delta is our our actually the delta can be written as delta o plus delta
delta and therefore p delta is same as p delta delta because delta o is constant that is if you take
the derivative of this term that is d delta by dt right then this comes out to be same as d delta
delta by dt.
Okay now if you take the Laplace transform of this equation then it will be s times, s times delta
delta s equal to delta omega r s into omega naught therefore, I can now write down delta delta s
equal to omega naught by s into delta omega r s. Okay therefore the the speed deviation right is
related to the angle deviation by this transfer function omega naught by s. Okay and therefore in
this block diagram the this transfer function omega naught by s relates the angle deviation that is
delta delta to delta omega r. Okay and therefore this block diagram is extremely important to
understand the small signal stability of a power system because we will be developing the
complete model including including additional blocks to this basic block diagram right. Now the
next step to understand the whole thing is that we develop a characteristic equation characteristic
equation of this system right.
Now to develop the characteristic equation, we can make use of this block diagram and relate
this output delta delta this output delta delta is to be related to change in mechanical this is the
input right therefore, if you look at this block diagram input is changed in mechanical torque
output is delta delta therefore this is a single input, single output system. Okay they normally
they call it SISO single input single output system and you can simplify this model and obtain a
simple transfer function relating the delta delta to delta Tm that is first we write down what is
delta delta as omega naught by s 1 upon 2 time Hs into this whole quantity which is delta Tm
minus Ks delta delta minus KD delta omega, okay.
Now in this equation what we do is that we replace this delta omega r, delta omega r by delta
delta upon omega naught into s right because this relationship is there in the block diagram. You
can see this block diagram relationship delta delta is equal to omega naught by s into delta
omega r right. So that when you look at this equation we have now delta Tm delta delta and other
constant terms.
Okay and therefore when you rearrange the whole thing, you can write the equation in this form
s square delta delta KD by 2H s delta delta plus KS by 2H omega naught delta delta equal to
omega naught by 2H delta Tm right. Now when you equate this expression to 0, we get the
characteristic equation for the system that this is this is the input input term right.
Therefore to to obtain the characteristic equation we can use this expression therefore the
characteristic equation is in the form s square plus KD by 2 Hs plus Ks omega naught by 2H
equal to 0. Since the system which we have considered is a second order system therefore
characteristic equation is also a second order characteristic equation. Now you can see here the
the coefficients are depending upon the damping term KD the system inertia constant H the
synchronizing torque coefficient Ks. Now these are the system parameters further you can easily
see that Ks depends upon the operating condition. Okay it also depends upon the line reactance
whether the expression for Ks is for this single machine infinite bus system the expression for
synchronizing torque coefficient is E prime into EB divided by Xt into cos delta o right and
therefore we can easily see here actually that in this characteristic equation the the coefficients
are function of loading that is system operating condition and system parameters also the Xt
depends upon the line reactance.
Similarly, it depends upon direct access transient reactance right, now when this equation is
similar to our standard equation in this form s square plus to times zeta omega n into s plus
omega n square and you can identify identify that this coefficient 2 times zeta omega n, 2 times
zeta omega n is equal to KD by 2H. Similarly, this omega n square term that is this term is equal
to KD by 2H while this omega n square term is equal to Ks omega naught divided 2H.
Okay now when you when you obtain the roots of this characteristic equation right the the roots
can be written as s1, s2 the small s s2 equal to minus zeta omega n plus minus j times j times
omega n square root of 1 minus zeta square that is in this case the roots will depend upon
whether they are complex conjugate or real on the value of zeta, if zeta is if zeta is greater than 0
and less than 1 right. We will get these 2 roots as complex conjugate roots right.
Now a system whose zeta is greater than 0 and less than equal to one right we say that the system
is under damped under damped. A system whose zeta is equal to 1 is critically damped is
critically damped and a system whose zeta is greater than 1 is over damped right and and system
whose zeta is less than 0 is unstable it is in negative damping system is unstable right in any
system, we have to ensure that the damping is adequate.
So that the oscillations which are generated are damped, whenever we design the controller for
the system we have to achieve certain minimum damping for all the modes which are present in
the system right. Now here we can also give some more interpretation to the damping term zeta
that is if you plot these roots in the s plane then assuming that damping zeta is positive the roots
will appear in the s plane because here we, so the real part we call it sigma.

We call this as omega and this axis and any any root is written as sigma plus minus j omega
right. Now if the system is a stable system like having its complex conjugate or the roots lying in
the left half of the s plane then this term that is this can be this can be is written equal to how
much zeta omega n while this term will be equal to how much omega n square root of 1 minus
zeta square what will be this quantity omega n therefore, this omega n is the distance of the root
from the origin okay.
Now if you find out this if you look at this angle theta right then cosine theta is equal to zeta
right therefore whenever we perform this Eigen value analysis of the system. We we can draw in
the s plane a line which will represent certain value of zeta that is you can that is suppose this is
your s plane right and if I say that this line represents zeta equal to say .1, this line will represent
zeta equal to say something more than .1, .2 like this.
Now in case all the roots lie on this side of this line it means each mode or there is no mode
whose damping is less than .2 right therefore whenever we design the control for this for
improving the system stability particularly the AVR tuning or power system stabilizer tuning
right, we perform the root locus analysis okay and vary the parameters like gain setting and see
actually that you get the optimum performance but at the same time you ensure actually that no
root is having its damping less than the desired value that therefore this this is a very important
way to to ensure that we do not go to a stage where any do any of the roots is less than or any of
the roots is having its damping less than desired value okay.
Now for this system which we are considering here the omega n is the natural frequency of
oscillation it is given by this formula omega n is equal to square root of Ks omega naught by 2H
and damping ratio zeta is one upon 2 times KD divided by square root of Ks 2H into omega
naught. Okay these 2o terms you can derive and see that you get the value of omega n and zeta
like this. Now the next point which we will study is, see this the two differential equations which
we have derived the linearized differential equations one differential equation was written in this
form, okay the.
Similarly, we are written data written the second differential equation in the linearized form,
okay. Now you can arrange these two differential equations right and write the model in the
vector matrix form that is finalizing the stability of the system right the most convenient form of
representing the model is vector matrix form that is we write the model in the form of X dot
equal to Ax plus Bu right.
Now these two equations which we have derived can be arranged in this form that is d by dt of
delta omega r equal to minus KD by 2H delta omega r minus Ks by 2H delta delta plus1by 2H
delta Tm second equation is d by dt of delta delta equal to omega naught into delta omega r that is
all. Okay therefore, these two equations which we have derived can be written in the vector
matrix form and when we make use of the the say mat lab for analyzing the system performance
right then we can find out the Eigen value of the system matrix A right for studying the stability
characteristic of the system right.
Now you have to understand very clearly here that this matrix A is function of function of
operating condition and system parameters this I will be you know emphasizing again and again
here that whenever you write down the model in the form X dot equal to Ax plus Bu right then
the basic stability characteristics are determined from the system A matrix. Okay therefore live
let me summarize here that what we have done till now is we have developed the small signal
model for the machine infinite bus system and we have represented this model in the form that is
vector matrix form X dot equal to Ax plus Bu.
We can write down the A matrix we can write down the B matrix and the state variables in this
model are delta omega r and delta delta, these are the state variables while delta e delta Tm in this
case is a input, okay. Now I will just give one more important information that the importance of
this term 2H okay, now to understand this importance of this term 2H, let us look at our swing
equation again.

Our swing equation is d by dt omega r equal to 1 upon 2H Ta okay, where omega r is expressed
in per unit, okay. Now now suppose you have a synchronous generator and which is at stand
state 0 speed okay and if you apply the rated torque rated torque then in how much time the
synchronous generator will attain its the rated speed right. Now to obtain the time in which this
synchronous generator will attain its rated speed can be obtained by integrating this equation that
is you integrate this equation assume Ta equal to 1 per unit okay, Ta equal to 1 per unit and just
tell me in uh quickly in how much time the synchronous generator will attain its rated speed.
Okay the synchronous generator will attain its rated speed starting from 0 towards rated value
when it is applied with a rated rated accelerating torque that is Ta is equal to 1 per unit and
therefore this time 2H is also denoted as the mechanical starting time of the synchronous
generator that is TM is equal to 2H therefore, if you see in the model we have a term 1upon 2H
therefore this 2H is similar to a time constant and is the mechanical starting time of the
synchronous generator in literature you will find actually that many times we represent this term
this TM by a symbol capital M that is you will find actually when you read the literature that this
block 1 upon 2 times Hs is also represented as one upon Ms.

Now here you should not get confused with the term M, the angular momentum of the rotor right
the M here is actually the in the small del will be equal to 2H and which is the mechanical
starting time of the synchronous generator rotor when it is subjected to a rated torque. Similarly,
if the machine is running at constant speed a rated speed and if you apply a retarding torque
equal to the rated torque then its speed will come down to 0 in a time equal to 2H right.
Therefore, that also gives the information that how much time the machine will take to retard
from its lateral speed to 0speed therefore with this today I will conclude my presentation by
saying that we have developed a linear model of machine infinite bus system considering
considering constant field flux linkages. Okay we have also studied the ah the the the importance
of damping ratio zeta right and whenever we design the system right the zeta is going to be one
of our design parameters, okay. Thank you!
Prof. M. L. Kothari
Lecture - 19
Small Signal Stability of a Single Machine Infinite Bus System (Contd.)
Today, we continue our study on small signal stability of a single machine infinite bus system.
Now today we will include the effect of synchronous machine field circuit dynamics, in the small
perturbation model. Earlier the the transfer function model which we developed earlier in this
transfer function model. We assumed the constant field flux model, okay or the voltage behind
direct axis transient reactance were assumed to be constant that is the classical model now we go
one step further and include the dynamics of the synchronous synchronous machine field
circuits.
Okay now when we developed the synchronous machine model, okay at that time I emphasis
actually that if we neglect the amortisseurs then the synchronous generator can be represented by
a first order model that is the synchronous machine equations there was only one differential
equation. The differential equation was corresponding to the field flux linkages and the equation
when it is written written considering time in seconds is put as P psi fd equal to omega naught
into efd minus Rfd ifd.
Now this term omega naught comes because we are now considering the time in seconds rather
than in per unit. Okay when we developed this equation in per unit it was simply d by dt of psi fd
equal to efd minus Rfd ifd. Okay now here we will do once that transformation because this this
equation is written considering this efd and ifd in reciprocal per unit system. Okay we replace this
efd by this term using this equation Efd equal to Lad by Rfd efd that is where we derived the
synchronous machine model.

We introduced this term efd and basically this introduction was to to link the non reciprocal per
unit system to reciprocal per unit system and here when I replace this efd by the field flux, field
voltage that is Efd is the field voltage or output of the exciter right. The equation becomes p psi fd
equal to omega naught Rfd divided by Ladu Efd minus omega naught Rfd ifd right. Therefore, this is
the additional equation that has to be considered while accounting for the field flux dynamics or
we can say the field circuit dynamics.
The complete model considering the rotor dynamics and the field flux dynamics can be written
in this form that is the these two first two equations represent the dynamics of the synchronous
machine rotor, okay these two equations are well known to us. Now to this these two equations
when I add this third equation that is d by dt of psi fd equal to omega naught Rfd Ladu into Efd
minus omega naught Rfd ifd. Okay now these three equations equations right describe the the
complete behavior of the synchronous machine under dynamic conditions including the field flux
dynamics while while writing these equations we have made one assumption and that
assumption is that the the amortisseurs are ignored , another assumption which we have made is
that the transformer the transformer voltage terms are also neglected and the speed deviation that
is the omega r term which appears in the stator equations, stator voltage equations that omega r is
also assumed to be equal to 1 okay.
Now when we look at these equations, we will clearly see that the state variables are delta omega
r delta and psi fd that is these are the 3state variables. Okay but when you see on the right hand
side of these equations we have a term T that is the air gap torque right. Now this air gap torque
is neither a state variable nor it is a input to the system similarly, this ifd that is the field current is
not a state variable or nor it is a input to the system because when we write the equation this Tm
is the input to the system, okay Efd is the input to the system that is the mechanical torque applied
to the system and the field voltage applied.
Therefore these are the two input quantities and these are three state variables while this Te and
ifd these are not the state variables as well as they are not the input quantities and therefore there
is a necessity to to express these two quantities Te and ifd in terms of state variables. Okay the
moment we we succeed in expressing the air gap torque and the field current in terms of state
variables we can write down the complete model of the system because these equations are
concerned in these equations right. We have these quantities which are to be expressed in terms
of state variables.
Now to express this quantities in terms of state variables we we make use of the the synchronous
machine equations and we will also we make use of the network equations because here we are
considering a machine connected to infinite bus right and therefore the transmission line right,
we represent the network that we will be in a position to write down the network equation. We
will be in position to write down the synchronous machine equations and since this these two
components are connected right we will be in a position to correlate them. Okay that is we will
find actually the network constant equation okay.
Now these are the three basic equations which represent the d axis field flux linkage q axis field
flux linkage and the field winding flux linkage psi fd, okay psi fd is a state variable while this psi
d and psi q are not the state variable but we can see here that the psi d can be expressed in terms
of these inductances id id that is the direct axis component of the synchronous machine stator
current and ifd. Now these equations are known to you further actually this Lffd this is the this is
the self inductance of field winding this will be expressed as Lad plus Lfd, Lad is the mutual
inductance and in the reciprocal per unit system right this mutual inductances are same for direct
axis stator winding, the field winding. Okay these are the same we have made equations same
then we will make use of this equation that is the air gap torque is equal to psi d iq minus psi q id.
Okay therefore with this we will be in a position to simplify the model, now these steps are
clearly given in the text book which is prescribed for your course okay.
To start with what we did write that this psi d this the direct axis flux linkage can be expressed as
minus Ll id plus Lads into minus id plus ifd that is if we make use of this equation right then what
we do here is that this Lad Lad these two terms are combined and further further, we will make
use of this the saturated value of the mutual inductance instead of using the unsaturated value,
we will be using the saturated value and the saturated value and unsaturated value are related by
the saturation coefficient okay and therefore the equations now which I am writing here for psi d
and psi q this Lad is replaced by Lads that is we are using the saturated value. Okay therefore this
this term we can see this term that is Lads into minus id plus ifd this term represent the flux linkage
the inductance into current and this we will denote as psi ad psi ad.
Similarly for psi q, we will represent this as minus Ll into iq plus psi aq. Now here we are using
these terms which are normally called the intermediate variables which will ultimately be
eliminated. Okay similarly this equation that is the field flux linkage equation psi fd right. Now
for for this Lffd you substitute Lad plus Lfd. Okay and then these Lad terms will be combined the
terms which have the coefficient Lad.
You combine and we can write the equation for field flux linkage in this form psi ad plus Lfd ifd
right. Now when we write this field flux linkage in terms of psi ad which is which is also
appearing in our other two equation that is psi d equal to minus Ll id plus psi ad. Okay therefore
the first step here is that now I express this ifd that is the field current in terms of psi fd which is
my state variable psi ad and Lfd Lfd is a parameter of the machine is a machine parameter, okay
synchronous machine parameter it is a known quantity.

Therefore, this is this is one step where we are now expressing the field current in terms of the
one of the state variables and psi ad and Lf that is this term is additional term which again we
have to eliminate. Now these steps you can perform yourself that is the psi ad was defined as
minus Lads id plus Lads ifd that is this term when we put in the expanded form right it was minus
Lads into id plus Lads into ifd that is it was written as Lads into minus id plus ifd. Okay this can be
simplified and now expressed in terms of id that is the direct axis current id is with this derivation
one can perform and obtain the expression for psi ad in terms of direct axis current id, direct axis
component of the synchronous machine current id and psi fd therefore, if you see this equation
that this psi ad has been expressed in terms of the terminal current a component of the terminal
current and the psi fd which is the state variable okay.
Similarly, we can express psi aq in terms of iq that is what we are doing here is that these two
terms which are the intermediate variables psi ad and psi aq have now been expressed in terms of
the state variable and the the direct axis and quadrature axis component of stator current okay.
Now these two steps are these steps are very important then we will be further making use of the
stator circuit equations now because the stator circuit is connected to the infinite bus through a
transmission line that is it is a machine infinite bus system right and we have the standard
equation that is the ed equal to minus psi q into Ra id, eq equal to psi d minus Ra iq these are the
stator circuit equations these are the algebraic equation, a stator circuit equations. Okay where
this psi d psi d and psi q can be written in terms of currents that is iq and there is a little mistake
here id okay.
Now in these equations we see here very interesting thing that psi q and and psi d are expressed
in terms of the our intermediate variables psi ad and psi aq and iq and id. Okay now what we do is
we write down the network equations, okay and then the once we write down the network
equation we will be in a position to write we will be a position to eliminate this ed and eq.
Okay now to do this thing you, we write down the equation for ed and eq which where earlier
written in this form that is this psi q and psi d terms will be replaced by these terms. Okay so that
ed and eq are written in terms of psi ad and psi and aq. Okay therefore the direct axis and
quadrature axis component of the stator voltage is now written in terms of the currents and these
intermediate variables. Okay Ra Ll these are the circuit or the synchronous machine constants.
Now let us see what are the what are the network constant equations. Okay now to understand
this network constant equations what we have to do is we we make use of our reference axis
right because in a machine infinite bus system right. The infinite bus voltage can be represented
as a reference and the rotor position can be described in terms of position of the q axis right.
The position of the rotor with respect to the infinite bus voltage phasor can be expressed in terms
of q axis that is here if I represent this as my infinite bus voltage okay and this is the terminal
voltage of the machine synchronous machine right, if this is my q axis this is ed that is the d axis
that is this is your q axis. Okay this all known to you okay and this is your d axis okay. Then the
definition of this power angle delta is the angle by which q axis leads EB the delta is positive
when q axis leads the infinite bus voltage by an angle delta while the internal angle is the angle
by which by which q axis leads terminal voltage the delta i is the internal angle and delta is the
power angle, total power angle okay.
Now here you can write down the components of infinite bus voltage that is if you resolve this
infinite bus voltage EB into 2 components one along q axis another along d axis then the
component along q axis is EBq which is going to be EB cos delta, component along d axis will be
EBd equal to EB sin delta okay and when you express this terminal voltage and infinite bus
voltage as phasor we can write down terminal voltage as ed plus j times eq and infinite bus
voltage as EBd plus j times EBq, okay this is very straight forward nothing to be explain.
Now let us write down the ah network constant equation this your synchronous machine,
terminal voltage is Et phaser, the line impedance is RE plus j times XE and here is our infinite bus
the phase the voltage here is EB phasor right and It the current here is written and the phaser
formula right. Therefore our network constant equation is is Et is equal to EB plus RE plus j times
XE into It this is the network constant equation a very simple equation which relates the terminal
voltage of the synchronous generator to the infinite bus voltage in terms of the line impedence
and current okay. Now what we will do here is that we will replace now Et by ed plus j times Eq
EB will be replaced by EBd plus j times Ebq, It will be replaced by Id plus j times Iq and then we
will, we will equate real part with the real part and imaginary part with the imaginary part to get
two two algebraic equations, okay in terms of the real variables.
Now this exercise has been done here. This is our network constant equation okay, I am just
introducing here you substitute in this equation ah replace this Et It and EB by the components
and you perform simple algebraic simplifications and we will get these two equations that is ed as
ed equal to EBd plus id into RE minus Iq into XE eq equal to EBq plus id into XE plus Iq into RE
okay this is very there is nothing very difficult, okay anybody can do it.
Now once we have done this, we can we can now express this current Id and Iq, Id and Iq in terms
of our state variables that is you make use of see this this is one these are the two equations
where ed and eq are expressed in terms of these our intermediate variables id and iq. Okay now
this ed and eq will be replaced by ed and eq will be replaced by this ed and eq right you replace
these by the expressions given in the right hand side of these equations right.
Therefore, once you do this thing you will get equation right where you have only id and iq and
they will be written in terms of these terms these are intermediate variables psi aq psi ad and other
machine parameters and ultimately when you do all the simplifications algebraic simplifications,
you can write down the expressions for id, id in terms of the state variable this is our state
variable this is also a state variable and other synchronous machine constants that is all the all the
previous steps which we have made are basically to come to a stage where we can replace we
can express the obtain the expression for the id and iq in terms of these state variables okay.
Similarly, the expression for iq is given like this which you can derive yourself you have to sit
and derive these equations. The definitions of the terms which appear in those equations are also
expressed here. Okay and further when you write these terms in a per unit system, per unit
system the line inductance or synchronous machine inductances are equal to the synchronous
machine reactances and line reactance. Okay therefore we do not make any difference between
the inductance and reactance because in per unit system our omega o is 1 okay.
Once we have obtained the expression for id and iq in terms of the state variables okay the model
is complete it is a it becomes out to be non-linear model if you look it very carefully carefully
then you will find actually that the psi ad and psi aq terms are expressed in terms of these
currents. Okay and this currents are now expressed in terms of state variables we have also seen
actually that the the air gap torque is expressed in terms of in terms of the flux linkages and
currents right therefore ultimately what happens is that we are in a position to express the the
torque term which comes actually in the equation for the synchronous machine that is the swing
equation in terms of the state variables, torque is expressed in terms of the state variables right.
Now these are the non-linear equations we are not done any linearization. Now this we can
obtain now the linear equations considering small perturbations the first step which is considered
for linearization is to obtain expression for incremental changes in id and iq to obtain the
expression for incremental changes in id and iq this id which is function of these state variables.
Okay you can obtain the expression for incremental changes in terms of the state variables that is
delta id can be written as m one times delta delta plus m 2 times delta psi fd, these m1 and m2 right
these are the quantities which depend upon the synchronous machine parameters, the line
parameters, the initial operating angle of the system that is delta o that is operating condition of
system that is that is this these two equations which are the incremental equations for currents
direct axis current and quadrature axis current are now written in terms of the state variables and
these coefficients are function of system operating condition and system parameters which
includes the synchronous machine parameters and the line parameters, okay.
Now here actually I have derived the expressions for m1, n1, m2, n2 and it can be clearly seen that
these constants are function of function of the machine parameters and initial operating condition
that is delta is delta naught here sin delta o where you find cos delta o term all these line
reactances the infinite bus voltage okay.
Now to complete our ah hum model we also obtain now the incremental changes in these
intermediate state variable that is delta psi ad and delta psi aq in terms of our state variables for
example delta psi ad can be written in terms of the incremental changes in the state variables that
is delta psi fd and delta delta delta psi aq can be written in terms of again in terms of these state
variables and the coefficients of these terms are again function of the machine parameters,
system parameters and operating condition.
We had written the expression for ifd, if you just see this equation you see the equation here this
equation ifd was express as psi fd minus psi ad Lfd. Okay now since psi ad is expressed in terms of
the state variables I can now express incremental change in the field current in terms of the state
variables.
The expression comes out to be as delta ifd is equal to this coefficient into delta psi fd and this
term into delta delta that is the incremental change in field current is expressed in terms of the
state variables incremental changes in the intermediate variables that is psi ad and psi aq are
expressed in terms of the these variables because psi ad and psi aq have to be ultimately
eliminated. Now we come to our main equation that is the expression for for air gap torque.
We have to write down the incremental change in air gap torque that is delta Te, delta Te as we
have seen earlier delta Te is written as I am sorry not do this later it is confusing first we will
write down del Te not delta Te, Te is written as psi d iq minus psi q id. Okay now this psi d right if
you consider the leakage inductance then to obtain the expression for air gap torque this psi d has
to be replaced by psi ads iq minus psi aqs id we will be considering the saturated value of these flux
linkages right and difference between this and this term is the leakage flux which is not
contributing towards the air gap torque. Okay now we want to linearize this equation to linearize
this equation the simplest way will be you write down this Te as Teo plus delta Te express this
term as psi adso plus delta psi ads iq will be expressed as iqo plus delta iq minus this is also perturb
psi aqso plus delta psi aq right into id is expressed as ido id naught plus delta id.
Now under steady state condition Teo is equal to psi adso iqo minus psi aqo so into ido okay and we
will neglect the the terms terms which are the product of two incremental terms like when I
multiply this you know expressions or simplify I will come across the term incremental change
in the psi ed psi ad multiplied with incremental change in current this two incremental changes
when they are multiplied will give you very small quantity which can neglected. Now when you
do this exercise the ultimately you can write down delta Te equal to psi ado delta iq plus psi iqo
delta psi ad minus psi aqo delta id minus ido delta psi aq, here actually the we are considering the
saturated values s is not written just for the simplification okay.
Now we have the expressions for delta id and delta iq, we have also obtained the expression for
incremental changes in the air gap flux linkages psi ad and psi aq right. Therefore, when you
when you substitute these incremental quantities that is delta iq delta psi ad delta id delta psi aq
right and these are the initial values. Okay we will we can write down the expression for
incremental change in air gap torque as K1 time delta delta plus K2 times delta psi fd that is the
the all the derivations which we have done right from beginning right were basically aimed to
obtain an expression for change in air gap torque in terms of state variables and we get the
expression for incremental change in air gap torque this delta Te right equal to K1 time delta delta
plus K2 time delta psi fd.

Okay this is very a important step that what we find here is that here that the the delta Te that is
change in air gap torque depends upon the change in rotor angle that is the power angle and
change in field flux linkages that is delta psi fd okay. The the model which I had derived earlier
where we assume assumed the field flux linkage is constant right.

The the incremental change in torque that is air gap torque was simply equal to Ks into delta
delta while now here the incremental change depends upon the change in flux linkages in the
field winding, field circuit and delta delta. Okay therefore now instead of representing the
coefficient of delta delta as Ks, I am representing this as a K1. Okay this is now going to be the
difference in what we are writing. The expressions for these constants can now be written in the
desired form which can be computed from the knowledge of initial operating condition
synchronous machine parameters and line parameters.
Okay these expressions have been derived because they are only the algebraic simplifications
which you have to do. Okay just, now let us look again at our mathematical model. This was the
starting point for our further derivations. Okay now this equation right has been linearized in the
previous case we have linearized it where mechanic change in mechanical torque we have put as
a delta Tm change in electrical order is delta Te and this is our damping torque term right and
therefore this equation which is going to be a non-linear differential equation the moment you
linearize this this becomes a linear differential equation right.
Now we have now to linearize this equation that is the field circuit equation, we have not done
the linearization of the equation this can be done very easily because here we will consider the
change in field voltage as delta Efd. Okay and change in field current as delta ifd, okay but this
delta ifd see this is not a state variable is replaced by the incremental changes in terms of the
other state variables.
Let us now examine the complete model of the system, these 3equations represents the complete
model of the system, first two equations represents the the dynamics of the synchronous machine
rotor or rotor dynamics. The third equation represents the the dynamics of the synchronous
machine field winding. This these two equations were linearized earlier and we obtain a simple
stats pay simple model in the transfer function form. This third equation is now linialize
linearized and when we linearize this equation we will have the incremental change in field flux
linkage, incremental change in field winding voltage and the incremental change in field current.
We have earlier obtained the expression for delta ifd in terms of state variables.
Now when we when we substitute the expression for delta ifd in this equation the the linearized
equation can be now added to the first two state space equations and the linear equations appear
in the form delta psi fd dot equal to a32 delta omega r a3 plus a33 delta I am sorry this has
correction here. The delta psi fd dot appears as a32 delta delta plus a33 delta psi fd plus b32 delta
Efd okay. This can be written in this form here as delta psi fd dot equal to a32 delta delta plus a33
delta psi fd plus b32 delta Efd. Now we we take the Laplace transform of this equation and obtain
the transfer function model to represent this equation in the transfer function form. Once we take
the Laplace transform the equation can be written in the in terms of the Laplace transform
variables that is delta psi fd is equal to K3 divided by 1 plus pT3 into delta Efd minus K4 delta
delta. Now p is equal to s, okay now we will incorporate the transfer function relating, we will
incorporate the the transfer function relating the change in field winding flux linkage is to change
in rotor angle and field winding voltage.
Earlier, we have developed the transfer function model relating the variables variable
considering the first two equations in the of the system. Now to this model we will add the
equation relating the field winding variables, now while deriving this model we can do small
simplification that is this portion of the model that is this portion of the model can be simplified
by using the simple block diagram simplification and this portion can be represented by an
equivalent model of the form 1 upon 2 times Hs plus Kd where output of this will be delta omega
r and input to this will be delta Tm minus delta Te.
With this simplification, with this simplification incorporated the complete transfer function
model incorporating the field circuit dynamics is shown here this portion of the model is the one
which was developed earlier. Now since we have to incorporate the affect of field circuit
dynamics since, delta Te is equal to K1 delta delta plus K2 delta psi fd right therefore at this
summation point at this summation point, we to obtain the expression for delta Te we obtain a
quantity equal to K1 into delta delta and to this we add a quantity corresponding to K2 into delta
delta delta psi fd.
Now this sum is realized by using the summation block okay therefore in this transfer function
delta Te is obtained by summing the 2 terms K1 delta delta and K2 delta psi fd. Okay then we
have the transfer function to relate delta psi fd to delta efd and delta delta. Now this is shown in
this portion of the block diagram where input to this transfer function K3 over 1 plus s times T3 is
delta Efd minus K4 delta delta right.
Now this k4 and K4, K3 and T3 are again constants and they depend upon the system operating
condition and parameters. Now therefore this is the model which accounts for the the
synchronous machine field winding dynamics. Now we will try to understand the affect of affect
of the including the synchronous machine field dynamics on damping and synchronizing torque
coefficients. To to examine this aspect let us write down, let us write down the delta Te that is the
change in change in air gap torque produced due to the change in field flux linkages only right.
Therefore delta Te due to the change field flux linkages can be related to the change in delta delta
by this expression where delta Te divided by delta delta due to delta psi fd equal to minus K3, K2
K3, K4 divided by 1 plus s times T3, okay. Now here we can examine the affect of frequency of
oscillation of the rotor, in case the in case the frequency oscillation is 0 right then we can
substitute here in this equation or in this expression s equal to j omega equal to 0. Then delta Te
can be written as minus K2, K 3, K4 delta delta that is the change in air gap torque due to change
in delta psi fd is expressed in the form given by this expression okay that is delta Te is equal to
minus K2, K3, K4 delta delta.
Further, we know that the change in change in the air gap torque due to change in delta delta
which operates directly is written as delta Te equal to K1 delta delta right. Therefore therefore we
can say here the the delta Te is written as K1 minus K2, K3, K4 into delta delta. When we consider
the affect of change in field flux linkages also now for any practical system the constant K2, K3
and K4 are positive and we can see here that this term represents the net synchronizing torque
coefficient right and when the field flux linkages were assumed to be constant the synchronizing
torque coefficient was equal to K1 right but the movement we account for the the the change in
field flux linkages due to change in rotor angle delta right. The net synchronizing torque
coefficient decreases it becomes K1 minus K2, K3, K4 therefore for steady state stability limit the
condition is K1 equal to K2, K3, K4 okay.
Now let me conclude my presentation today here that we have developed the small pertition
small perturbation model of the machine infinemet infinite bus system including the synchronous
machine field winding dynamics. This model has four constants K1, K2, K3, K4 and we have also
seen that when we account for the the synchronous machine field dynamics then the
synchronizing torque coefficient under steady state operating condition decreases that is we can
say that the demagnetizing affect has a has an adverse affect on the synchronizing torque
coefficient. Thank you!
Prof. M. L. Kothari
Lecture - 20
Small Signal Stability of a Single Machine Infinite Bus System (Contd.)
Friends, today we will continue with the study of small signal stability. Last time we have
included the field flux dynamics in the small perturbation model which we developed. We have
also seen that due to change in change in field flux that is delta psi fd the electrical torque
produced delta Te the delta of Te that is the change in electrical torque or air gap torque that is if
you find out this transfer function relating the incremental change in air gap torque to
incremental change in angle then this transfer function comes out to be equal to minus K2, K3, K4
divided by 1 plus ST3. Now we have also seen last time that in case, we analyze the performance
under steady state conditions that is when S tends to 0 right then the incremental change in air
gap torque is given by this equation delta Te is equal to minus K2, K3, K4 delta delta, okay.
Now if you look carefully here in this transfer function block diagram which we had developed
last time that delta Te this is the net net deviation in the electrical torque is now some of the 2
components, 1 produced due to this constant K1 that is K1 delta delta and another produced due
to change in field flux linkages that is K2 delta psi fd. Okay now under steady state condition we
can see her actually that the total change in torque or air gap torque will be equal to K1 minus
minus K2, K3, K4 minus K2, K3, K4 into delta delta right. Now the net affect is that the the
synchronizing torque coefficient when we assume the field flux linkage is constant right is is
more than when you consider the change in field flux linkages right that is the change in field
flux linkages takes place basically due to the demagnetization affect.

Okay and the steady state stability limit is reached when we have this condition satisfied that is
K1 is equal to K2, K3, K4 that is when we when this product K2, K3, K4 is equal to K1 right then
the then the net synchronizing torque coefficient becomes 0, okay and that is the steady state
stability limit. Now we shall examine the case when the rotor of the synchronous machine is
oscillating with some frequency that is we examine a case where rotor of the synchronous
generator is oscillating at some frequency at rotor oscillation frequency.
Now to analyze this we again take the same transfer function delta Te divided by delta delta
right. This is equal to minus K3, K2, K3, K4 or 1 plus S time T3 okay. Now we will not set S equal
to 0 okay you multiply this numerator and denominator by a term 1 minus ST3, okay this is to
rationalize you will get the expression for this transfer function as minus K2, K3, K4 into 1 minus
STs, T3 divided by 1 minus S square T3 square.

Now this expression can be written in this form that is delta Te due to change in field flux
linkages only. Okay that is equal to minus K2, K3, K4 divided by 1 minus S square T3 square into
delta delta. Now when you multiply this this this the second term becomes plus again plus K2,
K3, K4, T3, K2, K4, K2, K3, K4,T3 divided by the same term that is 1 minus S square T square into
S delta delta into S delta delta okay.
Now in this expression we will set S equal to j omega that is you substitute S equal to j omega
where omega is the the frequency of rotor oscillations, omega is the frequency of rotor
oscillations and we also substitute here delta delta equal to omega naught by S delta omega r.
Okay where delta omega r is the deviation in rotor speed with respect to synchronous speed
therefore if you make these 2 substitutions in this equation right you can easily see that S delta
delta is equal to omega naught into delta omega r right and this term is when you put S equal to j
omega this comes to be real term right it means K2, K3, K4 are constants 1 minus S square T3
square will become 1 plus omega square T3 square. Okay S equal to j omega square of this is
equal to minus omega square and therefore it becomes plus, okay.
Now here when I am putting here S equal to j omega j omega here but 1 thing we can observe
very carefully here is that before I substitute the value of delta delta right, if you see in this
equation if I substitute S equal to j omega then this term this term can be characterized as
synchronizing torque component because this is because of this coefficient multiplied by delta
delta therefore this term is in phase with speed in change is in is in is in phase with change in
rotor angle okay, while when you see this term it will come out to be in quadrature with rotor
angle.
Okay and when we make this substitution in this equation the change in electrical torque or air
gap torque due to change in flux linkages only is written as Ks is a function of delta psi fd this I
am writing this Ks it is not a product it is not a multiplication that Ks is a synchronizing torque
coefficient due to delta psi fd only into delta delta plus Kd due to delta psi fd that is the damping
torque coefficient due to variation in field flux linkages only right multiplied by omega naught
into delta omega r.
Now whens when we can see very carefully here that this term represents the synchronizing
torque component because when I multiply this and take the complete term then the can take the
when I take the complete term it becomes the synchronizing torque component. When I take this
term completely it becomes damping torque component. Now if I take the coefficient of delta
delta then this portion becomes synchronizing torque coefficient and this portion of the
expression will become damping torque coefficient right.
Now to understand this thing more carefully or more clearly, we take an example where for a
particular operating condition the K1 is 0.7642, K2 is 0.8649, K3 is equal to 0.323, K4 is 1.4817
T3 is 2.365. Okay we take a specific rotor oscillation frequency the that is rotor oscillation
frequency is taken as 6.41 radius per second electrical radius per second. Now with this
parameters okay assumed with this constant assumed right, when we substitute in this expression
for delta T equal as a function of delta psi fd delta T produced by delta psi fd only, if you
substitute in this expression then the synchronizing torque coefficient due to variation in field
flux linkages comes out to be minus .00172 and damping torque coefficient comes out to be 1.53
right.
Now here we can we have to very carefully understand that these 2 terms which you have
computed here they are function of these K1 to K4 parameters or K1 to K4 constants T3 and the
rotor oscillation frequency okay therefore as the operating condition changes, you will find
actually that this K1 to K4 constants will vary T3 will also vary right, this rotor oscillation
frequency also varies and therefore the synchronizing and damping torque coefficients
contributed by demagnetizing affect will also vary but 1 conclusion which we draw is that when
the rotor is not in steady state condition or when the rotor is oscillating with some frequency
right it contributes to negative synchronizing torque coefficient and it contributes to positive
damping torque. Okay this is a very important conclusion which we should always keep in our
mind.
Now further further we want to see the what what is the total synchronizing torque coefficient
then this total will come out to be equal to K1 plus Ks produced or contributed by variation in
field flux linkage that is Ks is the net synchronizing torque coefficient, this will be equal to K1
plus the synchronizing torque coefficient contributed due to field flux variations or field flux
dynamics. Okay now particularly in this case the value of K1 is 07642 and the second component
is minus 0. 00172, okay and 1 can find out the uh resultant that will be 8 how much 6 okay1, 2,
3, 4, 5 okay. This is 88 goes it will be 4 here 4 here 267 that is 0.76248. In fact the decrease in
synchronizing torque coefficient is not very significant it is a very small decrease okay, this is
what the message is or which is the most the conclusion which we draw from this okay.
Now in all these studies which we have reported till now, we have assumed assumed the you can
see this in this diagram in this diagram this is our transfer function model which we have
developed. This delta Efd was set equal to 0 the meaning of setting delta Efd equal to 0 is the
constant field flux constant field voltage condition that is the field voltage applied is constant del
delta Efd is 0 it means Efd is constant okay.
Now now we will include the effect of excitation system we will include the effect of excitation
system in our model. Now to include the effect of excitation system in our model. Let us start
with the terminal voltage condition because to the exciter or excitation system right at the
summing point. We give input signal that is corresponding to the terminal voltage of the
synchronous generator. Let us start from the terminal voltage the terminal voltage is written in
terms of d and q axis component as ed plus j times eq. Okay then we can write down Et square
equal to ed square plus eq square.
Now what we do is that we we give small perturbation to these variables that is Et ed and eq. Let
us assume that Et is perturbed and the perturbed value is or the or the increment is delta Et
similarly ed increment is delta ed and increment of eq is delta eq, when you make this perturbation
we can write the equation in this form.
Eto is the is the steady state voltage okay and Eto plus delta Et is the terminal voltage okay
therefore the perturbation is around this voltage Eto therefore, Eto plus delta Et square equal to edo
plus delta ed square okay plus eqo plus delta eq square this is a very straight forward and simple
expression. Now what we do is that we expand this equation we just expand it and neglect the
second order terms in terms of perturbed values second order terms which are involving the
perturbed values that is we will ignore delta Et square similarly, delta ed square delta eq square
because they are small and the square will be negligible. If you make that assumption then we
can write down this this equation as Eto delta Et equal to edo delta ed plus eqo delta eq. Okay
because Eto square is equal to edo square plus eqo square right therefore this this terms will vanish
from the resulting equation okay.
So that now we can write down delta Et delta Et the change in terminal voltage as edo divided by
Eto delta ed plus eqo divided by Eto delta eq okay. Now at this stage our problem is that this delta
ed and delta eq these are not the state variables, these are not the state variables right and
therefore we have to eliminate these incremental terms from the expression for delta Et. Now to
to eliminate this we make use of our generator, stator circuit voltage equations ed equal to minus
Raid plus Lliq minus psi q, psi aq, eq was equal to eq is equal to minus Raiq minus Llid minus psi ad.

Now these 2 equations were derived earlier and these equations represent the stator circuit direct
and quadrature axis circuit equations. Okay therefore what we do in this equation is that we we
give small perturbations okay and since this coefficients are constant R a Ll these are constant
quantities, okay you just perturb them.
Now when you give this perturbation you can write down delta ed equal to minus Ra delta id plus
L1 delta iq minus delta psi aq, okay similarly delta eq is equal to minus Ra delta iq minus Ll delta id
plus delta psi ad that is we get the equation for delta ed and delta eq in terms of delta id, delta iq
delta psi aq and psi delta psi ad further further the expressions were derived or the equations were
derived for deviation in delta id, delta iq delta psi aq delta psi ad and they were all expressed in
terms of state variables. If you look into the equations derived earlier right then these 4 terms that
is the deviation in direct axis component of the current quadrature axis current then psi aq psi ad
they were all derived therefore, if we if I now substitute here in these expressions expressions for
these quantities in terms of state variables and then you substitute delta ed and delta eq in this
equation that is 18.61.
Now when you make these substitutions then ultimately we can write down the expression for
delta Et equal to constant K5 delta delta plus K6 delta psi fd. Now this this equation is the most
important equation that is this change in terminal voltage of the generator can be expressed as K5
delta delta plus K6 delta psi fd okay.
The expressions for K5 and K6 are derived and they are written in the form K5 is equal to edo
divided by Eto into minus Ra into m1 plus L1 n1 plus Laqs n1 plus eqo divided by Eto minus Ra n1
minus L1 m1 minus Lads prime m1 that is if you see here the expression for K5 right then then all
the all the terms in this expression are known they are function of the initial operating condition
and the circuit parameters or the system parameters.

Similarly, the expression for K6 is written here and K6 is also function of operating condition and
system parameters therefore for any given operating condition 1 can compute the constant K5
and K6. Now after having completed completed the inclusion of inclusion of excitation system
model in the small perturbation model before I talk about how this model is completed, let us
first look here.
If you see this this portion of the model this is the transfer function model this portion of the
model it was derived earlier where this was the delta Efd input signal, okay. This portion here the
state variable delta psi fd delta Te delta omega r delta delta this was already derived right. Now
we add this equation that is delta Et equal to K5 delta delta plus K6 delta psi fd right. Therefore
we add this summation point here in this block diagram where 1, 1 signal is coming as K5 delta
delta another is K6 delta psi fd right and what we get here is the delta Et delta Et the change in
terminal voltage of the generator.
Now we have also seen actually that the terminal voltage of the synchronous generator is sensed
by putting a transducer that is we call actually the voltage sensor or voltage transducer. Okay,
now we can represent this voltage sensor by a first order transfer function that is 1 over 1 plus S
times TR, this time constant TR is generally very small right small and in some studies, we will
ignore this time constant or in some studies this can be included there is obsoletely no problem
only thing is the order of the model will increase when you include actually this transfer function
okay.
Now you can see here that delta Et is the input to this transfer function which which is the
transfer function of voltage sensor or voltage transducer. We will denote the output of this block
by a another state variable you call it delta V. Okay this delta V is now now compared with delta
V reference with the small perturbation model okay, delta V reference that is this error delta V
reference minus delta V is fed to the exciter Gex s represents the transfer function of the exciter
and output of these is exciter is delta Efd okay therefore this transfer function model right is a
complete transfer function model which accounts for change in field flux linkages which
accounts for change in terminal voltage of the synchronous generator and which also includes the
exciter transfer function model. Okay therefore if you I look at this transfer function model the
we have 2 inputs here 1 is the change in mechanical torque delta Tm is a input signal then delta V
reference is another input signal I can change the reference voltage right, if I want to regulate the
terminal voltage okay.
Similarly, if I want to change the mechanical output you can change the position of the governor
to change the mechanical input to the system right now for this system you can develop a
complete state space model in the form x dot equal to Ax plus Bu and the state variables will be
delta delta delta omega r delta psi fd delta V and delta Efd. Now in case this transfer function of
this exciter is simplified is simplified and sometimes we represent the exciter by a first order
transfer function, sometimes we represent this exciter by a by a ah constant gain only right if this
is represented by a constant gain only then delta Efd will not be a state variable. Okay therefore I
will suggest all of you to develop yourself a state space model of this system okay with input
variables delta Tm and delta V reference and the state variables delta omega I mean delta delta
delta omega r delta psi fd.
Okay and delta V here delta Et is not a state variable okay and the the the transfer function model
which has been developed in the text book is assuming assuming that the exciter is is a stator
excitation system and its time constant is negligible, so that you can represent the exciter by a
gain KA only right. Therefore I suggest you to assume the exciter model as a a constant KA okay
and then you develop the complete transfer function model of the system in the form of x dot
equal to Ax plus Bu and write down the A and B matrices. Now here in the text book the that is
text book we are following right the model is developed assuming delta V reference equal to 0
delta V reference is assumed to be equal to 0 right that is there is no change in the reference
voltage right. Now in case this delta V reference is assumed to be 0 then there is only 1 input
variable right and the model is developed here in this form that is delta Efd is equal to KA into
minus delta V1, you just see it here this will come out to be straight away, this is delta we call it
delta V1 delta V1.

Okay this is KA, this is 0 therefore there is a minus sign attached here right and because of this
there is minus sign coming and therefore delta Efd is equal to minus delta V 1 into KA. Okay and
once you have this quantity delta Efd you can write down the equation for P delta psi fd because
we have to write down the equation for the derivative of or equation expressing the derivative of
delta psi fd right and this can now be written as you can easily see here, when you write this
expression, when you write this expression.
So delta psi fd can be written as K3 divided by 1 plus S times T3, okay what is the input to this
input to this is delta Efd minus K4 times delta delta. Okay and this delta Efd we have just now
seen is equal to minus KA minus KA times delta V1 right and when you make the simplifications
and eliminate all those terms all those terms where which are not the state variables. We can
right down the equation for delta psi fd dot or d by dt of delta psi fd in the form a 31 delta omega r
plus a32 delta delta plus a33 delta psi fd plus a34 delta V1, okay this is the equation which we can
write down and in this case when you have assumed actually the exciter or represented the
exciter by a constant KA or gain KA right the the model which you get will be 4th order model
and the state variables will be delta delta, delta omega r delta psi fd and delta V1 these are the
state variables and you can write down the model in the complete form.
Further you can study the Eigen value of the a matrix and analyze the stability of the system
because ah for any system the stability can be analyzed by analyzing the Eigen value of the
system matrix A right.
Now I will just tell you the ah little historical development, this model which has a 6 constants it
is popularly known as K1 K6 model and it was originally developed in 1952 by Heffron-Phillips
that is it is called as or it is known as Heffron-Phillips model.
The that is K1 K6 model also called Heffron-Phillips model. These were developed the the this
this was reported in a research paper that is at that time it is to be AIEE, American Institution of
Electrical Engineers AIEE in 1950. Then subsequently in 1969 the detailed studies were carried
out by by the 2 authors popularly known as DeMello and Charles Concordia, C. Concordia and
the results of their studies were reported in a IEEE transaction on power power applicative
systems in the year 1969 and therefore many a times people refer to this model K1 K6 model as
DeMello Concordia model right but originally it was developed by Heffron-Phillips and still
actually it is known as the Heffron-Phillips model and this is the very very important model
which is used extensively for analyzing the small signal stability of a machine infinite bus
system because this model is basically for single machine infinite bus system.
Similar models have been developed for a multi machine system also. Now I will just skip our
some of the transparencies. Now we will analyze 1 very important aspect that is what is the
effect of AVR on synchronizing and damping torque components? we we have now included the
excitation system right therefore for once we have included the excitation system right we can
either refer to this as effect of excitation system control or we can refer that the effect of AVR
automatic voltage regulator.
Okay now we will study the effect of this AVR by analyzing analyzing the synchronizing and
damping torque components okay now to obtain the to obtain the synchronizing and damping
torque components which are produced when the when the rotor is under steady state condition
or when the rotor is oscillating with some frequency that is we will analyze again again under 2
conditions 1 is the system is under steady state operating condition, another is when the rotor is
oscillating with some frequency okay.
Now to analyze this again we come to the expression for delta psi fd right. Now the delta psi fd
can be written as K1 divided by 1 plus S times T3 into minus K4 delta delta because our
expression here was delta Efd minus K4 delta delta therefore minus K4 delta delta is written here
and and we are writing. Now here the expression for delta Efd that is minus Gex S or 1 plus S
times TR multiplied by delta et this is delta et the K5 delta delta plus K6 delta psi fd.
Okay that is if you again look at this expression then what we are trying to do is that now we are
trying to find out an expression for delta psi fd. Okay now this delta psi fd we can write first as
delta psi fd equal to this transfer function into delta Efd minus K4 delta delta then this delta Efd is
replaced by this transfer function Gex Gex into delta V1. Okay and then this delta V1 can be
replaced by this expression this transfer function into delta et okay.
Now with this things represented we have represented all the things here that is delta psi f d is
equal to 1 upon 1 plus S times T3 into minus K4 delta delta minus Gex S divided by 1 plus S times
TR because we are not assuming that Gex equal to K in the beginning but for for understanding
and for simplicity we will fa replace Gex by KA later on. We are also taking the time constant of
the voltage transducer. Okay therefore this is the complete expression which is now written in
terms of in terms of the state variables because delta delta is a state variable delta psi fd is the
state variable okay.
Now you you rearrange the terms, so that delta psi fd terms come on 1 side and other terms come
on the right hand side of the equation right now by rearranging the expression looks like this
delta psi fd equal to minus K3 into K41 plus S times TR into ah closed plus K5 Gex S divided by S
square T3 TR plus S times T3 plus TR plus 1 plus K3 K6 Gex S delta delta right whereas I have told
you, just now just now that we are transmitting all the terms right all the terms which are having
delta psi fd on left hand side and remaining term on the right hand side, right hand side terms will
have delta delta.
Okay and then we write down the expression for delta psi fd as this transfer function into delta
delta this is exactly similar to what we did actually when we assume the delta Efd equal to 0 right
but since now we are including the exciter or excitation system the expression for delta psi fd
becomes a slightly more complicated. Okay but here this delta psi fd which we are writing down
here right is is is now function of these constants.
You can see the all the the constants which are coming in this picture except K1 you have K3, K4,
K5, K6 right K2 does not come in picture so far actually this delta psi fd is concerned right the
function of these 4 parameters the function of the transfer function of the exciter is also the
function of this constants like TR and so on therefore I can say that delta Te due to delta psi fd is
equal to K2 delta psi fd.
Okay that is the electromagnetic torque produced or air gap change in air gap torque change in
air gap torque due to change in delta psi fd is equal to K2 into delta psi fd right because we have
seen the actually the total change in air gap torque is K1 delta delta plus K2 delta psi fd right
therefore the air gap torque or change in the air gap torque due to delta psi fd is equal to K2 delta
psi fd.
Okay now to illustrate what is the effect of AVR gain AVR gain we take this typical example
where the constants of the model are given as K1 equal to 1.591, K2 equal to 1.5, K3 equal to
.333, K4 equal to 1.8, T3 1.91, K5 is minus 0.12, K6 is 0.3 the time constant TR is taken as 0.02 a
small time constant the exciter is represented by a gain KA, okay inertia constant H of the system
is 3 and KD equal to 0 in fact these are the parameters of the complete system right although
when we are trying to analyze the synchronizing torque coefficient produced due to delta psi f d
or due to the AVR action right the H will not be required.
Now let here I am you know trying to emphasize that knowingly we are taking K5 equal to a
negative term for any system for any system as the system loading changes changes the K5 varies
from positive to negative that is if you per if you if you compute the value of K5 considering
wide variation in loading condition right then the K5 is positive for light loading condition and
K5 becomes negative when loading becomes high that is heavy loading condition K5 is negative
okay.
Now we have taken deliberately K5 equal to minus .12 a negative value. Now with this
parameters which we have assumed first we will study the steady state synchronizing coefficient,
the steady state synchronizing coefficient that is for obtaining the steady state synchronizing
coefficient what should we do in this expression we substitute S equal 0, S equal to 0 when I
substitute x, S equal zero in this expression right. You will find actually that delta Te due to delta
psi fd is equal to 0.06 KA minus 0.9 divided by 1 plus 0.1 KA that is I have not substituted the
value of KA I have expressed this torque in terms of excite AVR gain. Okay KA that is delta Te
due to field flux ah changes or deviation in field flux linkage is written as 0.06 K A minus 0.9
divided by 1 plus 0.1 KA into delta delta okay.
Now if I take KA equal to 0, K equal to 0 is a condition similar to constant field voltage that is
when you apply constant field voltage, okay it means we are not assuming the AVR action that is
KA equal to 0. If you take KA equal to 0 this quantity that is the synchronizing torque coefficient
due to delta psi fd is equal to minus 0.9. Okay that is what we had seen earlier that the for
constant field voltage condition right the effect of armature reaction or demagnetization is to
contribute negative synchronizing torque coefficient therefore you get a negative 1.
Now if you take KA equal 15 then this numerator will become 0 for KA equal to 15, you can
easily see that 15 into .06 equal to .9. Therefore if you take KA equal to 15 then the contribution
of delta psi fd to synchronizing torque coefficient becomes 0 now if I take KA equal to 200 which
is this case the synchronizing torque coefficient becomes positive that is 0. 529 because the the
movement the movement you put KA is equal 200 and compute this expression to this this term
is the synchronizing torque coefficient.
Okay therefore the now the total synchronizing torque coefficient is equal to Ks is equal to K1
plus Ks delta psi fd right because K1 is we have assumed the value of K1 equal to how much 1.
591 therefore, if I substitute actually K1 equal to 1. 591 and plus .529 you will get 2.12 right that
is when K5 is negative K5 is negative K is taken equal to 200 you find that the net synchronizing
torque coefficient increases it becomes 2.12 as compared to 1.59, okay. Now you and there is
under steady state condition it did not contribute anything to damping torque coefficient because
the in this expression right what we are getting is that the torque is directly proportional to delta
delta. Now you consider the the damping and synchronizing torque comp1nts ah rotor oscillation
frequency okay.
Now we again start with this expression right in this expression we substitute the values of all the
constants and we will substitute the value of substitute the value of S equal to S as j times 10, a
typical value is put here S equal to j times 10 perhaps this value is chosen to simplify the
calculations okay. Now if you make this substitutions delta Te due to delta psi fd again comes out
to be 11.1 minus j times 0.18 divided by 17.18 plus j time 19. 3 delta delta okay.
Therefore what you see here that this comes out to be a this coefficient of delta delta is the
complex number you rationalize this expression find out the separate the real and imaginary parts
then you can write this equation in the form 0.2804 delta delta minus 0. 3255 j delta delta. Okay
therefore, now you can easily see that this is the synchronizing torque coefficient right and this is
positive while the damping torque coefficient in this particular case is to be computed as yet
what is to be done here is that in this expression in the expression what you will do is we have
this relationship.
We have the important relationship that is delta delta equal to omega naught divided by s delta
omega r. Okay therefore now here in this equation if I put S equal to j omega right then I will
have the term j times delta delta equal to omega naught, omega delta omega r right and this
expression that is I can write down delta Te due to delta psi f psi fd equal to 0.2804 delta delta
okay minus 0.3255 omega naught divided by omega delta omega r okay omega naught is in this
particular case 37 it is 60 hertz system and omega is taken to be equal to 10.
Okay therefore this coefficient the this coefficient is the coefficient of delta omega r right
therefore the this this is the total damping torque component and this term 0.3255 into omega
naught by omega becomes the damping torque coefficient right the damping torque coefficient is
computed here and it comes out to be equal to minus 12. 27.
Okay, the important conclusion which we draw from this discussion is discussion is that when K5
is negative right the if K is taken around 200 right. It contributes to negative synchronizing
torque coefficient, I am sorry it contributes to positive synchronizing torque coefficient and
negative damping torque. Okay with this let me conclude that we have developed the complete
transfer function model that is the Heffron-Phillips transfer function model for a single machine
infinite bus system and we have also analyzed the effect of AVR on on synchronizing and
damping torque coefficients. Thank you!
Prof. M. L. Kothari
Lecture - 21
Small Signal Stability of a Single Machine Infinite Bus System-Power System Stabilizers
(PSS)
Friends, today we will continue our study on small signal stability of a single machine infinite
bus system. In the previous lecture, we have studied the effect of automatic voltage regulator on
the damping and synchronizing torque coefficients. We have seen that in case the operating
condition is such that such that the constant K5 is negative and if we increase the AVR gain then
we find actually that the damping torque coefficient becomes negative while the synchronizing
torque coefficient is positive that means we have a very conflicting type of effect on the damping
and synchronizing torque coefficients, when the AVR is considered. Okay that is now we have
also computed through else simple example that if we increase the AVR gain right then the
damping torque coefficient becomes more and more negative.
Now to illustrate this thing further for a typical problem, we have computed the K1cases
constants over a wide range of operating condition and we will just examine how these constants
vary as the system loading varies.
In this figure in this figure we have plotted the variation of constant K1 as a function of real
power output P for 3 different values of reactive power output that is this graph shows this graph
shows the variation of K1 variation of K1 for Q equal to 0 that is the reactive power output of the
generator is 0 and we are varying the real power only okay.
We can see here actually that constant K1 increases up to some value of real power output and
then starts decreasing okay then for reactive power Q equal to 0.5. Okay, now here when I say Q
is positive means the we are supplying lagging lagging power factor load, okay therefore for Q
equal to .5 right the constant K1 is varying in this fashion again its variation is similar to what we
got for Q equal to 0 it increases it increases and it becomes maximum and then starts decreasing
you can see here actually that the value of K1 is less than is less for Q equal to .5 as compared to
what we got for Q equal to 0. Okay another graph which is shown here is that Q is equal to
minus .5 that is it supplies leading power factor load okay.
Now we can see here that again K1 first increases and then starts decreasing, okay therefore the
trend is similar and 1 thing we observe that in this complete range where P is varied from .2 to 1
per unit right the K1 remains constant I am y K1remains positive not constant K1 remains positive
and it is its magnitude is affected by the reactive power which it supplies.
This graph shows the plot of constant K2 as function of P that is the real power output of the
generator for 3 different values of Q right. Now we find here actually that in all the 3 cases
where Q equal to 0, Q equal to 0.5 or Q equal to minus 0.5 right. The K2 is increasing over the
range which we have considered from 0.2 to 1 that is this computations have been done or the
graphs are plotted for varying real power output from very low value that is .2 per unit to 1 per
unit right and further we can see here again that as the reactive power output becomes negative
the value of K2 is more that is K2 is more for leading power factors load and it decreases when
the power factor is lagging right.
This constant K3 if you see that it is also affected by the reactive power output Q equal to 0 this
is the graph Q equal to minus .5 this is the graph and Q equal to 0.5 this is the graph but for a
given value of Q the variation in K3 as the value the value of P varies is very insignificant, okay
this practically it is a flat curve.
Now this graph shows the variation of constant K4 as function of real power output P for 3
different values of Q again that is Q equal to 0, Q equal to .5 and Q equal to minus .5 again you
can see here actually that the constant K4 varies as the real power output varies and is also a
heavily depended upon the value of reactive power output. But still we can see actually that this
constants K1 to K4 right over the completely over operating range which we have considered that
is ah real power output varying from very low value like point 2 per unit to 1 per unit and the
reactive power varying from .5 to minus .5 right the ah constant K1to K4 remain positive all
through.
Now here is the plot for constant K5 again as function of real power P for 3 different values of Q
now this line represents the reference axis that is K5 is 0 right. Now you can see here actually
that if I consider Q equal to 0 right in this particular case the the value of K5 is negative all
through from .2 to 1 per unit value of P that is the real power P is varying from .2 to 1 or Q equal
to 0, it is all through then when it is plotted for Q equal to .5 right it is seen that for P varying
from .2 to some value around .35 or so right the K5 is positive and then it becomes negative.
Similarly for Q equal to minus .5 K5 remains positive for about for P equal to .5 nearly and then
it becomes again negative therefore this K5 is the most important parameter right which affects
the small signal stability of the system and sometimes we {u} ((00:10:31 min)) we call this K5 as
the culprit. Now we can very clearly see actually in this plot that K5 is positive for light loading
condition right while it becomes negative for heavy loading conditions but even actually in this
particular case we find actually that the for a moderate loading also it is negative and once this
K5 is negative right it produces produces negative damping torque coefficient when the AVR
gain is increased, okay.
The last plot is for the constant K6 which shows actually the variation of constant K6 as function
of P for 3 different values of Q, Q equal to 0 Q equal to minus .5 and Q equal to .5 again we can
see here actually that yes the constant K6 is positive all through right and it decreases as the real
power output P increases for any value of Q that you can see that Q equal to .5 or Q equal to 0 or
Q equal to minus .5 it decreases but we can clearly see actually that the value of K6 remains
positive okay.

Now to understand more clearly the effect of K5 on damping and synchronizing torque
coefficients, the computations were carried out for synchronizing torque coefficient as a function
of KA that is the AVR gain. Now we have considered 2 loading conditions, 1 loading condition is
corresponding to ah loading condition 1 loading condition is such that K5 comes out to be
positive therefore this that is K5 equal to 0.0309 where the real power output is .4 per unit and
reactive power output is 0 that means it is a light loading condition K5 positive.
Now when we increase this AVR gain then synchronizing torque coefficient this is the
synchronizing torque coefficient it decreases right although it is positive all through it decreases,
its magnitude decreases.
For the for the same operating condition the damping torque coefficient is plotted here and it is
seen actually that the damping torque coefficient increases with increase in K A and it becomes
maximum for come value of KA then it starts further decreasing. Now over this wide range of KA
that is um in this graph the KA is varied from 0 to 500 right.
Now over this complete range we can clearly see that that the KS is positive but it decreases
however the KD that is the damping torque coefficient it increases and then starts decreasing but
still it is positive all through over the complete operating range that is the K varying from 0 to
500. Then we have consider another loading condition a heavy loading condition the loading
condition is that a real power output P is .9 per unit and the reactive power output is .3 per unit
for this operating condition and other system parameters okay, the K5 is computed it comes out
to be minus 0.1475.
Now considering this K5 equal to minus 0.1475 okay and all other constants which where
computed for this operating condition right the the value of KS synchronizing torque coefficient
and damping torque coefficient are computed. The synchronizing torque coefficient you can see
here it increases with the increase in AVR gain KA right it you know the over a small range over
a small range practically it is constant but afterwards it increases and it is positive all through
okay. Now you is this graph shows the variation of damping torque coefficient KD okay, you can
easily see that when KA is 0 the value of KD is slightly positive.
We have seen actually in the previous discussion right if we do not include the AVR okay and
assume assume constant field voltage right and at that time the effect of demagnetization or
armature reaction is to give some positive damping torque. Therefore K equal to 0 a small
positive damping torque and as the value of K is increase the damping torque becomes 0 right
and then it remains negative it becomes it attains some maximum negative value and then it
decreases okay and we can see here actually that although it first it is it is positive initially K
equal to 0 then it it decreases in a fast rate, it becomes negative right and all the range 0 to 500
the KD is negative all through right.
It means we have find here that that the when you include the AVR AVR then the damping
torque becomes negative. Okay now in any practical system automatic voltage regulatory is
required to regulate the terminal voltage. Okay because we cannot say that we will go for a
manual control and keep this field voltage constant if you do that you will find actually that they
negative damping torque is not produced but AVR is required and further to regulate the terminal
voltage within the certain certain tolerance right. You have to have certain steady state gain
larger the value of AVR gain right less will be the steady state error.
Okay and therefore AVR is required to regulate the terminal voltage it has to have large gain
because larger the gain of the AVR right its response is going to be fast and the excitation system
whose response is fast is capable of improving the transient stability limit of the system.
However, when the when we consider a small signal stability or a dynamic stability of the
system right we find that it will add negative damping torque coefficient or negative damping
torque.
Now when the system has negative damping torque whenever the oscillations are created in the
system these oscillations will grow. Okay and when the oscillations grow you will find actually
that on any power transmission line you will find actually that due to the rotor oscillations right
the the there will be oscillations in the power flow on transmission lines right and the protective
relays are provided to to detect these oscillations and protective relays will disconnect the system
or trip the transmission line whenever such oscillations are produced, before the system losses
synchronism right. But the problem basically is that can be allow the situation to happen answer
is obviously that yes, we have to find a solution to overcome this problem of negative damping
produced due to the AVR action and the answer is the incorporation of power system stabilizer
right.
Now first let us see what is the function of the power system stabilizer? Here I have written basic
function of power system stabilizer you have to very clearly understand the basic function then
we we will try to see how we achieve this objective. Power system stabilizers add damping to
electromechanical oscillations this is a fundament thing that we are we are talking of electro-
mechanical oscillations that is the oscillations of the synchronous machine rotors right.
It is a electro-mechanical oscillations essentially essentially the power system stabilizer act

through the generators excitation system in such a way that a component of electrical torque
proportional to speed change is generated that is we can see here, the PSS act through the
generators excitation system that is when I discuss the different type of excitation system at the
AVR input I have repeatedly told that there is 1 signal coming from the power system stabilizer
okay and we that signal was denoted as VS a stabilizing signal right.
Therefore we can say here that this act through the generator excitation system and it acts in such
a way that a comp1nt of electrical torque proportional to speed change is generated. We know
that a torque which is produced in phase with the speed deviation right is the damping torque.
The damping torque is KD into delta omega R right therefore if we can produce a torque which is
in phase with speed deviation then it will be it then it acts as a or it acts to damp the oscillations.
Now we can further say that this this torque which is produced is the additional this is an
additional an addition to the damping torque because system has its own damping also there is
always some damping which is there inherent in the system therefore the PSS will provide a
damping torque which will be additional to the existing system damping. A power system
stabilizer is used to used to add a modulation signal to the generators AVR reference input that
is they they stabilizing signal which we which we obtained through the power system stabilizer
right that signal will be added at the at the summing point of the AVR or we can say actually that
it is added to the AVR reference input okay.
Now let us again come back to our K1 K6 model that is the Heffron-Phillips model now in this
model let us see how to incorporate the PSS. Now this is a summing point, this is a summing
point where the reference voltage is compared with the terminal voltage in the AVR we have this
this is the small signal or you can say the small perturbation model therefore all the quantities are
shown as deviations from the nominal values right delta V reference is the deviation from the
reference of voltage okay then this signal delta V1 is is corresponding to delta Et because this
transfer function was incorporated to take a account for voltage transducer. Okay therefore at the
summing point we want to add a signal this signal we will, we are representing as VS and this
will be added we are putting a plus sign here right.
Now this signal is derived from the speed deviation that is in this transformal transfer function
model this is the point where I can sense the speed deviation delta omega r. Okay therefore here I
put a transfer function we will denote this as Gpss that is the transfer function of the power system
stabilizer. Its input is derived from the speed deviation, okay the output of this transfer function
that is the will be the stabilizing signal and this stabilizing signal is added to the summing point
of the AVR. Okay that is you can see here actually in this block diagram what I have done is I
have taken taken input from this point which represents delta omega r. Okay it enters the transfer
function of the PSS and output of the transfer function of PSS is the stabilizing signal, this
stabilizing signal is put here at the summing point okay.
Now, now we have to see actually that in practice in practice how do we physically realize the
speed deviation signal because you can you can sense the speed by putting a speed transducer on
the shaft of the synchronous generator that is generally today the standard practice is to have a
toothed wheel mounted on the shaft to the synchronous generator and we put a magnetic pickup
right. Therefore, the number of pulses which are generated by the magnetic pickup are
proportional to the speed of the shaft right and ah we can very very precisely measure the actual
speed of the synchronous generator and this speed is compared with the reference speed and the
difference is the delta omega the speed deviation. Okay and now what we have to see is see is
that what should be the structure of this Gpss what should be the transfer function of the power
system stabilizer.
Now to understand this what we do is we look at we look at this transfer function model
carefully. Now if you see here actually that when this signal VS is injected here. Okay the let us
see that we want to obtain obtain air gap torque or variation air gap torque by varying this VS.
Okay suppose actually I am just putting signal stabilizing signal VS and I want to know how
much is the air gap torque produced due to this signal only that is I want to know how much is
the air gap torque produced by this stabilizing signal only.
Now you can easily say that this signal ah or the air gap torque produced can be easily obtained
through this loop can you see this this is the loop actually because here we are not going to
consider the deviation in angle right that we are looking only that if I am putting this signal how
much is the stabilizing produced therefore you have to see very careful here the the let us just
look actually how much is the delta psi fd produced due to VS. Okay then the the torque produced
due to the PSS will be equal to delta psi fd into K 2 that is this is the torque which is going to be
produced. Okay and therefore we can say that if we can write the expression for delta psi fd as a
function of VS right then we can make out actually that if I vary this VS how much is the
stabilizing torque produced, okay.
Now you can see that if this VS is coming here right now to find out this delta psi fd due to this
stabilizing signal only we have to consider only this portion of the transfer function, okay. Now
delta psi fd can be written, Now if I just see it here we can write down looking at this expression
delta psi fd equal to equal to this transfer function that is you put K3 over 1 plus S times T3.
Okay, now input to this is coming through this Gexs and we will be assuming Gx as as K right
therefore this is coming through KA. Okay and now input to this is VS minus delta V1 VS minus
delta V1 there are assuming delta V reference to be 0, okay therefore we can put KA you can take
this K into into VS minus further to simplify this understanding, we will assume this Tr equal to 0
this time constant of the voltage sensor or voltage transducer is very small and you can make this
Tr to be 0. So that we can say that here it is VS minus minus K6 times K6 times delta psi fd. Okay
that is we get once simple expression this multiplied by VS then this delta psi fd, I can write now
here delta psi fd produced due to stabilizing signal only can be written as K3 divided by 1 plus S
times T3 multiplied by KA into VS minus K6 delta psi fd, okay.
Now this expression this expression which I have written here is simplified and put in this form
delta psi fd equal to K3 KA, 1 plus S times T 3 in bracket minus K6 delta psi fd plus VS. Okay this
is a very important step once you very carefully derive and see that the delta psi fd produced due
to the stabilizing signal VS can be written in this form. Now here and if you simplify this and and
write the expression for delta psi fd divided by VS that the transfer function relating the change
in flux linkages due to stabilizing signal only the delta psi fd upon VS, K3 KA divided by 1 plus
K3 K6 KA plus s times T3. This is the simple transfer function you will get relating the change in
flux linkage to a stabilizing signal VS. Okay now to have clear understanding okay about this
effect of stabilizing signal on the on the air gap torque produced due to stabilizing signal.
We take this example for a typical example which we consider earlier these where the constants
K1, K2, K3, K4, T3, K5 while taken as negative K6 is taken as 0.3, this TR is neglected although
when the previous case we have taken TR equal to 0.02 but for the purpose of simplification we
are ignoring or we are neglecting this TR making TR equal to 0 the effect of neglecting TR is the
or or I can say that the error error caused due to this assumption is practically insignificant .We
will assume this Gex the transfer function of the exciter as KA and its value we will take as 200 a
typical value equal to 200. Okay now with this parameters assumed if we make the compute this
ratio delta psi fd divided by VS it comes out to be equal to 66.66 divided by 21plus 1.91s, okay.
Now you assume that the rotor oscillation frequency is 10 radians per second 10 electrical
radians per second, so that put as equal to j10 then this transfer function becomes 66.6 divided by
21 plus j times 19.1 it was 1.91 omega is 10 it becomes 19.1. Now we can write down the delta
TPSS that is the air gap torque produced due to the PSS action only is written as can be written
as delta Te due to PSS that is delta T PSS is delta T or air gap torque due to PSS which is going
to be equal to K2 times delta psi fd due to due to PSS okay. Now we have just now computed
actually this delta psi fd okay as the the transfer function delta psi fd by VS okay.
Therefore, now you substitute the expression for or the value for TK2 right we get the delta T due
to PSS as as 1.5 into 66.6 divided by 21 plus j times 19.1 into VS into VS. Now for this for this
complex you can say the function right we can find out its magnitude and phase angle. We can
find out the magnitude and phase angle that is what you have to do is you take the magnitude of
the numerator divided by magnitude of denominator and this ratio will give the magnitude then
find out the phase angle of the numerator in this case phase angle of the numerator is 0 because it
is a real number, find out the phase angle of the denominator. For example, in this case phase
angle of this denominator term is going to be say theta is the phase angle its transient inverse
19.1 divided by 21. Okay and therefore the net phase angle of this function will come out to be
negative right therefore delta TPSS is obtained as 3.522 the phase angle is minus 43.3 degrees into
VS okay.
Now suppose I want I want that this the torque produced due to PSS is a pure damping torque
pure damping torque. Okay in that sense that I must produce this VS that the stabilizing signal
stabilizing signal it should be such that it will it will compensate for this phase length. Okay and
therefore if I if I now assume the transfer function I am just putting very high very, you can say
a simple case that is if I assume GPSS equal to I am just putting GPSS equal to some constant I call
it KA, a K stabilizer KASTA stabilizer.
Okay and its phase angle I make it 43.3 degrees, okay then then delta P delta TPSS can be written
as 3.522 KA stabilizer I think we can put this STA as STAB stabilizer, sometimes we use the
word K stabilizer only that A can be deleted here into delta omega r because VS is equal to what
is VS, VS is equal to GPSS into delta omega r therefore, if I now the assume actually that there
GPSS is equal to a constant and having its phase angle equal to 43.3 right then the the air gap
torque produced due to PSS is 3.522 into K stabilizing delta omega okay.
Now if you if you we can write this same thing which I have written just now as delta TPSS is
equal to gain of PSS at omega equal to 10 into 3.522 into delta omega r this is what we have
written now we will write write down this damping torque coefficient due to PSS, KD due to PSS
will be equal to gain of PSS into 3.522 because this torque is purely a damping torque right and
we have compensated for the phase length which is produced by choosing the appropriate
transfer function of PSS.
So that we can say that damping torque produced due to PSS is equal to the gain of PSS what so
gain you put into 3.522 and the net damping torque which is going to be produced will be equal
to KD due to AVR action plus KD due to PSS. The net damping torque which is going to be
produced will be equal to KD due to AVR action plus KD due to PSS plus there will be some
inherent damping torque which may be present in the system because we have seen actually that
when I did not consider AVR, we did not PSS still there was some damping torque produced
positive damping torque due to demagnetization effect.
Okay now suppose I want actually that I want to simply nullify the effect of AVR action right
then the the KD PSS should be equal to the KD due to AVR action that is negative, for example in
the problem which we have taken problem which we had taken earlier we found that K D equal
was found to be minus 12.27 when we have considered only the the effect of effect of field flux
changes that is the demagnetization effect or armature reaction effect right. We found actually
that the damping torque contributed I am sorry not it was due to the AVR action with KA equal
200 the damping torque contributed is minus 12.27 right.
Therefore now if I take the the PSS gain if I take PSS gain equal to 12.27 divided by 3.522 right
then this PSS is going to produce positive synchronizing torque coefficient which will just
nullify the negative damping torque produced due to the AVR action right but now if I take gain
if I take gain which is more than this that is if you set the PSS. So its gain is more than this ratio
that is 12.27 divided by 3.522 right then it is going to add to additional damping, it will provide
additional damping torque right and therefore we come to 1 very important conclusion that the
that the PSS PSS which we have to provide should be in a position to to compensate for this
phase length and if you have certain gain.
So that it can provide additional damping okay because the the damping due to demagnetization
effect was very small although it was positive when we assume this manual control right there
was some positive damping but that was very small and therefore actually we want certain
damping which will quickly damp out the oscillation whenever there are produced right and
therefore there is a necessity to design the PSS properly.
Now another 1 point which you have to understand is that suppose I provide PSS which will not
fully compensate for this time ah for this phase length because here the phase angle is 43.3
lagging and if I make PSS to have phase length less than 43.3 that is the PSS compensates for an
angle which is less than this that is I can take instead of 43.3 degrees, let us take 30 degrees right.
Now this situation is known as the under compensation that is we are we are not fully
compensating the phase length which is there between the stabilizing signal and the air gap
torque. Okay now if you have this situation what will happen if you look at the let us look at the
let us look at a plain here. I denote this as a delta delta and delta omega r is along y axis okay if
suppose you fully compensate for the phase length by choosing the PSS parameters right then the
delta TPSS delta T due to PSS will be along this axis right and this is a pure damping torque,
okay.
Now suppose you instead of fully compensating if you do ah under compensation under
compensation right in that case this is your delta delta this is your delta omega r, delta TPSS will
lie somewhere here delta TPSS will be in this position. Okay you understand this point because
we have we have we have partially compensated for the system phase length. Now this is the
situation then this torque can be this delta TPSS can be resolved into 2 components 1 along delta
omega r another along delta delta right.
Now what we see is that this component this is the component of delta PSS along delta delta
what is this component know as synchronizing torque and this is this component is the damping
torque therefore, the moment you have a situation where you have under compensated right, we
get some additional synchronizing torque right and it is desirable because we know that the
whenever you design a system when you design a system we should have synchronizing as well
as the damping torque coefficient to be positive and here we add this line.
Now you take a third case where if you over compensate for this, suppose you over compensate
that is instead of having the phase angle of the GPSS exactly equal to the phase length of the
system. Then the delta TPSS is produced can be shown in this plain it will come in this second
quadrant that you can put delta TPSS and now you can see here actually that we resolved this into
2 components.
We get a component in phase with delta omega r which is the damping torque and we get another
component which is in phase with delta delta synchronous torque coefficient but now we get a
negative synchronizing torque coefficient right and therefore in practice in practice it is it is
necessary to design the PSS is in such a fashion. So that we do not over compensate okay under
compensation if it is there there is no harm the literature shows that 1 can go for under
compensation as much as 10 degrees to 30degrees under compensation of the order of 10 degrees
to 30 degrees.
Now what we have considered till now is a simple situation where the excitation system was
represented by a constant gain KA only right. Now this representation is applicable to static
excitation system right where the time constant of the exciter is very very low.

However, when we consider other excitation systems right then we have to account for the actual
transfer function of the excitation system and further we can see 2 important points that the the
the phase length which is produced depends upon the operating condition because it is function
of the system constants and the constants are function of operating condition. Okay and it also
depends upon the type of excitation system which we have considered right.
Therefore from that point of view that point of view the transfer function which we have to
assume for the for the GPSS are I am sorry transfer function of the power system stabilizer, okay
will not be a very simple 1 a general form of the transfer function block diagram of a of delta
omega PSS is shown here. Without emphasizing I have told you that the input signal to the PSS
is the speed deviation.
Okay and since we want to produce a torque in phase with the speed deviation therefore this is
most ah suitable input signal to PSS and this signal is delta omega therefore in the literature this
type of power system stabilize or the power system stabilizer with speed deviational input signal
is known as delta omega PSS, this is delta omega right therefore it is we use this terminology.
Now here the general transfer function of PSS is shown here that is input to this is delta omega
output is the stabilizing signal therefore here first block is showing the stabilizer gain, stabilizer
gain. The second block shows ah what is called washout? this is a washout block and these 2
transfer functions represent the phase compensator, phase compensator we have seen the
function of these gain how much gain is to be provided because the the gain will determine the
amount amount of damping which is going to be produced.
We have also seen the role of this phase compensator because this phase compensator is to
provide the required phase compensation that is phase length of the system is to be compensated
by putting phase lead therefore compensator is to be provide the desired phase lead for a given
frequency. Okay now the function of this washout we will just understand why we require the
washout block?
Now we want actually this stabilizing signal to be present only when the rotor is oscillating or
there are oscillations in the system. Suppose there is a steady state deviation in the speed from
the rated speed to some other value, the steady state deviation right during the steady state
deviation we do not want any stabilizing signal to be produced because our basic function is to
provide damping to the rotor oscillations. Okay and therefore we provide here a transfer function
transfer function whose gain is going to be 0, when their frequency of oscillation is 0.
Therefore, in example in case see sTw or 1 plus s time Tw that if I put s equal to 0 then this
transfer function provide the 0 gain and when there is a steady state change in delta omega no
stabilizing signal is produced. Now suppose you provide stabilizing signal or allow the
stabilizing signal to be there, stabilizing signal to be there under steady state deviation then it is
going to unnecessarily affect our terminal voltage because this signal is entering at the summing
point of the AVR. We do not want actually the stabilizing signal to affect the terminal voltage
also okay the this transfer function should be a high pass filter. Okay now I will just sum up what
we have discussed today. We have studied the effect of variation of loading condition at different
constants and we have also studied the how the phase is to be incorporated to obtain the rewire
required positive damping. Thank you!
Prof. M. L. Kothari
Lecture - 22
Small Signal Stability of a Single Machine Infinite Bus System-Power System
Stabilizers (PSS) (Contd)
Okay gentleman, we start today on the the study of power system stabilizers.

Last time I have introduce the function of power system stabilizer and we have seen
actually that the power system stabilizer is provided to damp electromechanical modes of
oscillations. Okay and in order to provide the required damping to the electromechanical
mode of oscillations, we provide power system stabilizer.
The structure of the power system stabilizer is shown here. This power system stabilizer
is given the name delta omega power system stabilizer. The reason for giving this name
as delta power system stabilizer is that the input signal is derived from the from the
deviation of the rotor speed right. Now it has primarily 3 blocks, the first block we denote
as K stabilizer it is the gain block, the second block is denoted as s times Tw or 1plus s
times Tw this is called wash out block. The third block in this particular case we have
shown 2 transfer functions right this 2 put together will form the phase compensator
because the the power system stabilizer is to compensate for the phase lag between
between the between the exciter input to the air gap torque okay and to produce a torque
in phase with the speed deviation right.
This power system stabilizer is to be design, so that we compensate for the phase lag of
the system that the phase lag of the system is the phase lag of the exciter and the
synchronous generator. Now, I will now discuss ah in detail the role of different blocks of
the PSS. Let us first start with the wash out block this wash out block is provided to allow
to allow the high frequency signal of interest to pass and end not to allow the steady state
deviation to pass and end change the terminal voltage because if this block is not
provided then if there is a steady state change in the rotor speed then this steady state
change will be reflected in the terminal voltage change.
Now if you look here at the transfer function of this block under steady state condition, s
is 0, under steady state condition s is 0 therefore when I substitute s equal to 0 in this
transfer function right then the this the gain of this block becomes 0 right and therefore
any signal which comes in to the is power system stabilizer there will be no out put
coming from this block. However, when the frequency s, frequency of oscillate rotary
such that if this term 1 plus s the T omega right.
Now this term will become or you can say that it becomes for s equal to j omega right.
The the transfer function of the wash out block can be return as j omega Tw divided by 1
plus j omega Tw for any frequency of oscillation of the rotor.
Now if magni magnitude of this term is j omega Tw is very large as compared to 1 or I

can say if omega Tw is very large as compared to 1 right. In that case this transfer
function a gain equal to 1 because this 1 can be 1 is 1 plus j times omega Tw can be
written as j times omega Tw and therefore this transfer function will behave as if its gain
is equal to 1 right and therefore the the speed deviation signal right will be allowed to
pass right and therefore this wash out block is also known as also known as high pass
filter. Okay and generally the characteristic of this wash out block will depend upon of
the value of time constant Tw right. In the literature lot of studies have been done on the
choice of Tw.
We have also studied the effect of variation of Tw and we found actually that this Tw is
not very critical and 1 can choose the value of Tw in the range of 1 to 20 in the wide range
that is Tw equal to 1 second right up to 20 seconds they have been chosen. Now generally
generally the a time constant Tw Tw equal to 10 second is consider to be quite suitable.
Further actually some times what happens is that the system gets islanded because of
suppose some the major for in the system and view to resort islanding, in that case what
will across the frequency of this islanded system may go up or go down and at that time
at that time this frequency is the steady state deviation right at that time also we do not
want actually this phases phases to create to any problem in terms of variation of terminal
voltage right.
Therefore when you Tw equal to 10 or around that right then the desired characteristic of
the wash out block is okay the next transfer function block which we have to consider is
the phase compensatory. Now here I have shown 2 stage phase compensator that 2 stages
which you have shown here is normally call 2 stage phase compensator okay.
Now I will discuss, how to design 1 stage phase compensator and in case you require to
provide 2 stages then the second stage can also be design accordingly the basic function
of this phase compensator is to provide phase lead right and therefore we normally call
this as a phase lead controller or phase lead compensator. Now in order to provide the
phase lead this time constant T1 should be greater than T2 okay.
Now I will further discus, how the how the phase lead controller is designed and how this
can be physically realized using operational amplifiers because when you want to realize
this phase lead compensator right for implementation, okay then we have to see that how
this transfer function is physically realized using operational amplifiers although there are
other ways to realize this transfer function right but the the modern practice is to make
use of operational amplifiers.
Now when we design the phase lead compensator, we have to very careful and design in
such phase so that it provides the damping to the oscillations right which are which are
local modes of oscillations and interior modes of oscillation because we we have seen
earlier that local modes of oscillations have a frequency range of around .72 hertz, okay
while interior modes of oscillations have frequency of the order of point 1 to .8 hertz and
therefore if you look the total spectrum right the frequency range over which we want to
produce damping is varying from .1 to 2 hertz and the and the phase shift or phase lead
controller is to be designed.
So that we get the required phase compensation right but as we know as we have seen
actually that you can design this compensator to compensate exactly the phase lag of the
system at 1 particular frequency okay.
Now further we have to understand very carefully that the phase lag of the system is not
constant it depends upon the system operating condition and system parameters. Okay
and therefore as the system operating condition changes the phase lag characteristic of
the system also changes and therefore actually the design of the phase compensator, so
that it meets the requirement over the over the wide range of frequency is a real
challenge, okay.
Now this this phase lead phase lead compensator the transfer function of phase lead
compensator can be written in this form. In general I am just putting a general any phase
lead compensator can be written as gain Kc in to 1 gain Kc multiplied with this transfer
function that is s plus z1 divided by s plus p1, this z1 is the 0 of this transfer function and
p1 is the pole of this transfer function and for this transfer function to provide the phase
lead p1 should be greater than z1, okay.
Now here to realize this phase lead controller or phase lead compensator, the simplest
arrangement for its implementation using operational amplifier see shown here that is in
this amplifier, in this portion of the operational amplifier circuit this is the feedback
impedance pass 1 and this is the input impedance right. Now in the feedback portion we
have taken a circuit comprising of resistance R2 in parallel with capacitor C2. Similarly,
in the input a circuit is comprising of a resistance R1 in parallel with C okay.
Now you all know actually that when use 1 operational amplifier right there is a phase
inversion. Okay therefore to correct for this phase inversion the make use 1 more stage of
operational amplifier. So that we correct this phase inversion because this will in this
particular case the gain is 1 or the the feedback resistance is taken same as the input
resistance.
So that the gain of this operational amplifier is 1 and the phase shift is 180 degree
therefore the phase shift created by this operational amplifier is corrected by second
operational amplifier. The transfer function relating the output voltage Eo to the input Ein
can be simply obtained by writing the writing the operational impedance of this R2 C2 in
parallel and R1 C1 in parallel.
Suppose, you have circuit in comprising of resistance a capacitance C1 and the resistance
R1 right the we I can call the operational impedance of this a Z1, this will be equal to
parallel combination of these 2 impedance that is R1 in parallel if 1 upon C1 s divided by
R1 plus 1 upon C1 right which can be simplified and put in the form R1 divided by 1 plus
R1 C1s, okay.
Similarly, you can write down for the of feedback impedance that is I call this a feedback
as Z2 as the operational impedance of the feedback impedance that can be written as R2
divided 1 plus R2 C2 s and therefore we can write down the transfer function Eo divided
by Ein that is Eos divided by Eins as it is Zf divided by Z1 Zf is here. This quantity and Z1
we have already therefore when you divide it will come out be 1 plus R1 C1s, 1 plus R2 C2
s this multiplied by R2 by R1. Okay therefore you can see here actually the transfer
function is similar to what we are looking for but you are interested in having a transfer
function of the form 1 plus T1 s divided by 1 plus T2s.
Therefore, this the simple operational amplifier with the with the of feedback comprising
of a resistance in parallel with capacitance, similarly input also a resistance in parallel
with capacitance. We can get a simple transfer function which and by choosing the value
of this R1 C1, R2 C2 right. We will in a position to obtain the require phase characteristic
of the transfer function, okay.
Now here I have written this expression, we just now we derived you call this as a Gcs
that is the transfer function of the compensator Gc , c stand for compensator as Eos by Eins.
Now this can be put in the form the 1 which we have just now derived right it can be now
put in a different form right that is you can put this in a form C1 by C2 s plus 1 by R1 C1 s
plus 1 by R2 C2 and therefore we can say that this is this transfer function is of the form
gain Kc s plus Z1 s plus p1 where we can identify the Z1 as 1 by R1 C1 and p1 as 1 by R2
C2, okay.
Now in order to in order to reduce the number of independent parameters of the circuit,
we can choose C1 equal to C2 equal to C that is this these 2 capacitors can be chosen to be
equal, once you choose the C1 equal to C2 equal to C this transfer function can be written
here in this form it is put C1 equal to C, C2 was also equal to C and you divide this whole
expression by R1 and denoted by R2.
Okay then finally it is written in this form at R2 by R1,1 plus R1 Cs,1 R2 Cs which can be
R2 by R1 can be written as 1 by a where a is equal to R1 by R2 and T can be written sR2
into C, okay therefore now I get the transfer function of the compensator in the form 1 by
a multiplied by 1 plus a times Ts, 1 plus Ts right therefore what we have done is that we
have now 2 parameters to be computed 1 is the a, another is the time constant T and I as
have told actually that that T1 is greater than T2 and therefore a is greater than 1, a is
always greater than 1 to realize R2 realize a phase lead controller right.
Now the problem arises is that how to choose the values of these 2 parameters a and T.
Now this this quantity 1 upon a will be consider as part the stabilizer gain, you have this
is a multiplying factor a constant therefore this can be club with the stabilizer gain. Okay
now we will now a study how to design the this phase lead compensator to to achieve the
required phase compensation. Now if you look here in this transfer function is, excuse me
sir, yes.
No no here here this, if you look at this transfer function this portion of transfer function
right that this represents T1, this represents T2 and we want phase lead compensator and
for this phase lead compensator this a in to T should be greater than T, therefore a has to
be greater than 1as if you want to get the phase lag then a will be less than 1 because
sometimes we do required phase lag compensators in the design of controllers right at
that time a will be less than 1 and this a equal to R1 by R2 the values of R1 and R2 can be
so chosen, so that we get the required ratio.

Now the simplest approach to design this compensator is that we identified 2 corner
frequencies of this transfer functions, this transfer function as 2 corner frequencies 1 is
omega equal to 1 upon aT quite omega 1 another corner frequency we can call it omega 2
equal to 1 upon T and the this particular transfer function will provide you the maximum
phase lead at a frequency which will be the geometric mean of these 2 corner frequency
right.
Therefore here we consider this omega m as a geometric mean of the 2 corner

frequencies which is obtained as square root of 1 upon 80 in to 1 upon T that is these 2
corner frequencies we identify we find out the geometric mean. Okay and the geometric
mean of these 2 corner frequencies is equal to 1 divided by square root a in to T, the idea
here is that when you plot plot for this transfer function right the phase angle versus
frequency right then the maximum phase angle of this transfer function will occur at a
frequency equal to omega m and this frequency is the geometric mean of these 2 corner
frequencies, okay.
For this transfer function for this transfer function Gcs for s equal to G omega. Let us
write down what is the phase angle of this transfer function right to obtain the phase
angle of this transfer function what we do is that we write down Gc j omega equal to 1
upon a remains 1 plus j times a omega T that is for s equal to j omega 1 plus j omega
okay, what is the phase angle of this transfer function. Obviously, the phase angle is that
phase angle of the numerator minus the phase angle of the denominator and the phase
angle of the numerator is tan inverse a omega T and of the denominator is tan inverse
omega T and this is actually the phase angle of Gc j omega that is for this function for this
transfer function phase angle is given by this expression.
A very simple formula to compute the now my interest is my interest is to express the this
phase difference right phase difference in a simple form that is in terms of a and T. Okay
now what we do here is that we make use of this identity that is phase angle of this
function is let us say phi then 10 phi is equal to omega aT minus omega T divided by 1
plus omega aT in to omega T that is, we know that its suppose your this function transient
theta 1 minus theta 2 right this is equal to tan theta 1 minus tan theta 2 divided by 1 plus
tan theta 1, tan theta 2.

Okay, now here tan theta 1 is equal to omega aT tan theta 2 is equal to omega a omega T
not a and the the phase phase angle of this transfer function is theta 1 minus theta 2, you
will call this quantity as phi. So that now I can write down tan phi equal to omega aT
minus omega T divided by 1 plus omega aT in to omega T, okay a very simple
expression of the phase angle of this transfer function.
Now we know that for omega equal to omega m equal to 1 upon square root of aT the
phase shift is maximum. Okay and therefore what we do here is that we write down the
expression for tan phi m in this form that is you substitute in this expression omega equal
to omega m, okay and when you substitute omega equal to omega m in this expression
and simplify you will find that you get an expression for tan phi m equal to a minus 1
divided by 2 times square root of a, this step you can just substitute and verify, okay.
Now this expression is further written in the form sin phi m equal to a minus or a plus 1
or a can be expressed as 1 plus sin phi m divided by 1 minus sin phi m then we have
obtained very simple expression for a in terms of the the required phase angle of the
transfer function or the phase compensator that is if I want actually that this phase
compensator should provide a phase shift equal to phi m right then what should the value
of a and the value of a is obtain by the simple formula a equal to 1 plus sin phi m divided
by 1 minus sin phi m.
This intermediate steps can be easily verified right then we have also now an expression
omega m equal to 1 divided by square root of at that is if I know that this is the frequency
at which I want to get the required phase shift there is omega m known to me I start with
this omega m the frequency at which I want to compensate for the phase lag of the
system.
Okay, therefore omega m is known to me, so that using this expression and using this
equation to I know both this things, I know what is the what is the phase shift of achieve
that is the how much should be the phase shift or phase lead this phase compensator is to
provide. I also know at what frequency these 2 provide okay. Therefore knowing these 2
information I have a very simple formula we are first I compute the expression for a
because phi m is known to me and then use this expression use the value of a to obtain
the expression or the value of T right. So that I can now I have now both the parameters a
and T both.
Now the question arises that what is the frequency at which I should design my power
system stabilizer this frequency is the rotor frequency, rotor oscillation frequency.
However this rotor oscillation frequency we have seen varies over wide range, over a
wide right and therefore this power system stabilizer should be design in such a fashion.
So that you design for 1 frequency but when the frequency varies from that value on or
increases or decreases it still it should be in the position to give you reasonable phase
compensation but I just now mens I told you very clearly that we would like to design
power system stabilizer in a such a fashion.
So that so that it is it may have some under compensation but it should not have over
compensation. The moment you have as a some degree of under compensation, we get
the advantage of additional synchronizing torque coefficient, positive synchronizing
torque coefficient and is helpful right but we should we should not target to over
compensate because the moment you try to over compensate it will produce a negative
synchronizing torque coefficient.
Okay a many times the required phase shift cannot be obtained using 1 stage of the phase
lead compensator and in that case we go for 2 stages that is called 2 stage phase lead
compensator, many times 1 can have 2 identical blocks that is here I shown this phase
compensator right.
I have discussed how to design 1 stage phase lead compensator and we we can have 2
stages identical in that case suppose I want to compensate for say 85 degree let us say
right therefore, instead of designing 1 phase compensator to compensate for this total 85
degrees I can design the so that 1 is for 85 by 2 another is for 85 by 2 generally the thumb
rule is that 1 phase lead compensator block should be design to compensate around 50
degrees, the to understand the basic concepts of designing the phase lead compensator or
for that purpose any linear compensator, I will advise you to read this chapter 10 which is
design of control systems that is a title of the book is design of control systems written B.
C. Kuo.
I am just sorry, let me correct it title of this chapter 10 is design of control systems the
title of the book is automatic control systems the 7th edition by B. C. Kuo, B. C. Kuo is a
very standard text book on automatic control systems and I am sure that you will be in a
position to very clearly understand how this to be design. Now I will just quickly tell you
how to obtain the state space model of the complete system that is when you have
incorporated in a power system stabilizer right automatic voltage regulator in the system
therefore, we can develop a complete state space model of the single machine infinite bus
system including AVR and PSS.
We had developed earlier the state space model model considering considering the input
signal from the exciter as delta Efd that is this transfer function model actually if you just
look here. This transfer function model I am sorry, the from this transfer function model
that is if you see this is this portion actually that is we are not considering PSS we are not
considering the excitation system and AVR therefore the state variables which are
required to model the system in the state space form delta omega r, delta delta and delta
psi fd. With this 3 state variables right 1 can write down the state space model of the
system and input to this model or input to this state space model will be delta Tm and
delta Efd right.
The the model is written in the form delta omega r dot delta delta dot delta psi fd dot at
this is my X dot and this is the A matrix A, this is the state vector this becomes the input
matrix B and this is your input vector Bu right therefore in this case the input is
comprising of delta Tm and delta efd.

Now you can obtain more detailed model including including the AVR and PSS. Now to
develop this model I will just tell you few salient steps, the additional portion of the
transfer function to be incorporated in the model is shown here is the additional portion
because we already developed actually the transfer function model of the third order
system. Now the exciter is represented by a gain K only you can represent the exciter
depending upon the what type of exciter you are using okay.
We know the detail models of the different type of excitation system therefore here for
the purpose of simplicity we are representing this by a constant Ka only and this
representing static excitation system IEEE sT1 a model. Okay this is the change in the
terminal voltage delta Et this is the transfer function of voltage transducer terminal
voltage transducer that is we get a delta V1 this is the transfer function model of the PSS,
this portion you can just see here I have shown in this model a single stage phase lead
compensator right. If if we necessary 1 can put 1 more stage 2 stage phase lead
compensator right and accordingly you can modify your study space model and the
question is that how do we develop the differential equations, linear differential equations
for this system.
So that they can be augmented with the existing or the existing system model is
augmented by adding these equations additional equations. Now the state variables which
we will have to define new state variables 1 is because earlier these are the only in put
signal here right use state variables which your define will be delta v1 which is the output
of this transfer functional block, Vs the output of power system stabilizer then here then
we call this another state variable you call this is a delta V2.
Okay and here there is no state variable required because this is only gain therefore at this
point it is the signal is K stab into delta omega, okay therefore so far this this block is
concern in this block the input is this quantity output is delta V2. So far this second block
or phase compensator is concern input is delta V2 output is Vs, this block 1 is concerned
input is delta Et output is delta V1. Therefore, these 3 state variables are required to to
develop the complete model of the machine infinite bus system including PSS and AVR
right.
Now the equations can be written very easily let us us write the equation for first a block
that is here transfer function is 1 over 1 plus s times TR output of this is delta V1 input of
this is delta Et, okay. Now this model which we are talking about is small perturbation
model and in small perturbation model all these state variables initially are 0. There is the
perturbed values like see delta V1 delta Et right a tan equal to 0, R0 and therefore we can
replace s by p okay and you can write down the equation in this form delta V 1 equal to 1
1 plus S times Tr into delta Et or you can write down here now, 1 plus S times TR into
delta V1 equal to delta Et a very simple right.
Now here if a replace S by d by dt S is replace by d by dt, then you will and divided by
TR all through you can write the equation in the form that delta V1 dot will be equal to 1
upon TR delta Et minus 1 upon TR delta V1. Okay, now to make this equation complete
equation because delta Et is not a state variable delta Et is not a state variable but delta Et
can be expressed in terms of the state variables.
Therefore delta V1 dot is written as 1 upon TR delta E t is equal to K5 delta delta plus K6
delta psi fd, okay minus 1 upon TR delta V1 okay therefore now I have written 1
additional equation which is a differential equation it is expressed in terms of the these 3
state variables delta delta delta psi fd and delta V1 okay. Similar approach is to be used for
deriving additional state variable equations right I will just illustrate for delta V2 also.
Now to understand how to develop the equation corresponding to this wash out block, we
can write down delta V2 equal to STw or 1 plus S times TW input is KSTAB into delta
omega r delta omega r is a state variable. Now you perform the similar operation you will
find that S times delta V2 equal to KSTAB S delta omega r minus 1 upon Tw delta omega
delta V2. Okay that is you cross multiply and then collect the terms that you put S delta
V2 on left hand side and all other terms on right.
Now in this equation you find here S delta omega r, S is d by dt. So that this is basically
your delta omega r dot term delta omega r dot term because whenever you write
deferential equation on right hand side we will have only the the expression which is the
linear function of state variables right therefore what we do is that this delta omega r dot
is replaced by by is replaced by this equation, we have the equation here delta omega r
dot is equal to a11 delta omega r plus a 12 delta delta plus a13 delta psi fd plus b11 delta Tm
we have it right.
So that when you now substitute the expression for d by dt of delta omega R in this
expression, you will get a equation for delta V2 dot and this delta V2 dot is now written in
terms of all the state variables like delta Omega r delta delta delta psi fd delta V2 and delta
Tm because when you write this right this differential equation it will be function of state
variables plus input variables.
Now here input is delta Tm right you follow the similar approach for writing the transfer
function of phase lead compensator that is in the phase lead compensator the input is
delta V2, output is Vs the transfer function is 1 plus S time T1, 1 plus time T2 right now Vs
is a one of the state variables. So that we can write down the equation in this form you
have to simplify the expressions fallowing the same approach, ultimately our requirement
is that the left right hand side of the expression should be function of all state variables
and input way the last equation is for this block a summation block here.

Now what we do is that delta Efd is written in terms of K that is delta Efd is equal to K
into delta V reference plus Vs minus delta V1. Okay, now this delta Efd appeared in our
earlier equation for delta psi fd that is if you see this equation for delta psi fd this delta psi
fd is a substitute actually the express what delta efd because when delta psi fd was written
delta psi fd dot equal to right function of delta delta delta omega R delta psi fd plus
function of delta efd therefore this efd is now not state variable is to be eliminated.
Okay, when all these things are done there is you can eliminate this delta efd and you will
get the expression in the form delta psi fd dot equal to a2, a32 delta delta a33 delta psi fd a3
for delta V1a36 Vs ab32 Ka delta V reference that is what we have done here is that we
have written 3 additional equations, okay and modified the the third equation the reverse
corresponding to delta psi fd dot.
Okay and the the total model of the system can be written in this form it comes out to be
a 6th order model there is there are 6 difference state variables delta omega r delta delta
delta psi fd delta V1 delta V2 Vs. Okay and we can write down this model in the form
again X dot equal to AX plus Bu, now this u is delta Tm and delta V reference that is
input is now the mechanical talk that is the the incremental change in mechanical dot
delta Tm and the change in reference voltage, delta V reference and delta efd has now
been eliminated from the model.
Now in this model once the model is developed we know actually that all these
parameters that is coefficient the the elements of this a matrix, they all function of
loading condition and system parameters right, one can find out the Eigen values of this a
metrics to understand the stability of the system. Okay, now I will just quickly tell you 1
or 2 more points before I close the discussion.
The basic limitations of delta omega PSS basic limitation the the the limitations are 1 is
that this delta omega PSS can excite the torsional modes, can has the has the potential of
exciting the torsional modes. In case you have large gain of the PSS, if the PSS gain is
kept large then it can excite the torsional modes. Many times to overcome the problem of
excitation of torsional modes, we provide filter in the PSS path the moment you provide
filter right then this is again restricts the PSS gain that is we cannot have the gain as much
as we want because if you if you want to can say have the desired dumping right.
Then the gain setting of the PSS will give the required dumping ratio right but if you try
to design the PSS so that you achieve the required dumping for electromechanical mode
right then there are other problems the other problems are the possibility of excitation of
torsional modes and the limitation on the PSS gain. Okay, let me some of here that we
have discussed today the details of the power system stabilizer to be applied we have
analyzed the analyze the system with PSS. Okay and we have also seen how this PSS can
be realize using operational amplifiers and at the end we have derived the complete state
space model of the machine infinite bus system including AVR and PSS.
I will suggest that all of you should derive this model yourself right and that you can
study the behavior of the system with PSS, you can make use of the mat lab software
package to find out the Eigen values of the system matrix and to simulate the dynamic
performance your system by giving some perturbations may be perturbations may be in
terms of perturbation in mechanical torque you can just make delta Tm equal to some .01
or change in the reference voltage and see the behavior of the system. Thank you!
Prof. M. L. Kothari
Lecture - 23
Dynamic Modelling of Steam Turbines and Governors
Friends, today we will discuss dynamic models of steam turbines and governors.
So far we have developed the model for synchronous generator and the associated excitation
systems while studying the stability, we have made one assumption that mechanical input power
remains constant right. Now while developing actually the model for transient stability analysis
right we our swing equation our swing equation is written in the form 2H divided by omega s, d 2
delta by dt2 equal to Pm minus Pe.
Now till now we have assumed that this mechanical power input is constant. Okay and the this is
very valid assumption however whenever you want to do study involving a severe upsets in the
power system and we want to know the long term midterm and voltage stability analysis of the
system then the the turbine and associated governor models also required to be included and
today, we will devote our studies to develop the models of steam turbines and associated
governors before we developed this let us look at this block diagram.
This block diagram shows the complete power system net power system. Now you can identify
here the different blocks suppose you start with actually the power system network. This is the
block which shows the power system networks which has which we represent the generators
transmission line network and loads, okay. Now from this network we get output as electrical
power and for the control of this power system as a whole we have different controls okay.
Now one control is actually the generation control and for controlling generation what is done is
that we change the interchange power on various tie lines and find out the net interchange which
has the system frequency this is processed in a controller which is called automatic generation
control and the output signal which comes to the automatic generation control right. We will
change the position of the speed changer. Okay, then when you look the major blocks you have
the speed governor, speed controller mechanism and governor control valves and gates.
Now here we are writing valves and gates, the valves basically pertain to steam turbines and
gates pertain to hydro turbines and the output or the position of the gates right control the input
to the turbine and therefore this was the turbine here. Now we are while for for presenting the
system dynamic model what we do is that we represent the turbine separately and then the output
from the turbine will be represented as a mechanical power output and this mechanical power
output will be used to write the swing equation that is we wild this, we will separate the turbine
generator inertia and write the swing equation so that input to this block which we basically used
to write the swing equation. We call it actually the turbine generator inertia input is electrical
power and mechanical power the different of this is the accelerating power and the output from
this block will give you the angle change in angular position and change in the speed.
Now the speed deviation signal is processed through a speed governor therefore the speed
governor is a most important component in the control mechanism. It has two inputs, one coming
through the coming from the terminal of the synchronous generator and that is the speed signal
another is the a control signal coming through the AGC, we call actually the speed changer
position, okay. Now these two signals when they enter through this governor governor will
process it and we have a speed control mechanism and then ultimately the valves okay.

Now this is the complete schematic diagram which represents the location of steam turbine the
governors with respect to the total system. Now our study will be confined to representing the
steam turbine and the speed governing system okay. The reference material for steam turbine
models and hydro turbine and associated governors is available in this IEEE committee report
because when we discuss the excitation system models I have referred to a number of papers of
published in IEEE transaction and IEEE ah standards. Okay similar to similar to the excitation
system models the IEEE task force on power system dynamics has developed the models for
steam turbines, governors and hydro turbines in associated governors.
This is published in the year 1973 and while developing these models, since this subject involves
the mechanical system right along with the electrical system right and therefore this is written by
by with the association of IEEE working group and American society of mechanical engineers
right. Therefore, these two groups have contributed in developing the models which are
published in IEEE transaction in the year 1973 this is the standard paper I advise all of you to
take a copy and read it. Okay while developing these models there some assumptions which have
been made because when we see the system as a whole right we have contingency controls for
example fast valving right.
Now these type of controls have not been shown here in the models we will ah develop these
models actually and discuss these models from beginning. Now to develop the model what we
need is basically the relationship between the input and output. Now since here we will be
dealing while developing the steam turbine model right where we know that the steam is
generated in the boiler and then it flows through the various stages of steam turbine and then
ultimately it is exhausted to the condenser. Okay now in order to develop the model let us start
with a simple steam vessel, let us say that there is a steam vessel and volume V is the volume of
this vessel. In this vessel the steam is entering and let us say that Q in is the steam in kilograms
which is entering per second and Q out is the steam in kilograms leaving this channel.
We will develop actually a transfer function model relating the Qout to Qin and this basic model
will be used to represent various stages of the steam turbine. Our basic objective is to develop a
transfer function model relating the the Qout that is the the steam coming out to the steam coming
in to the vessel.
Here we can we make use of a basic equation which is called continuity equation, the continuity
equation is written in the form d by dt of W, W is the weight of the steam W is the total weight
of the steam which is inside the vessel. Okay this rate of change the dW by dt is equal to V times
d rho by dt. The rho here is the the density of the steam, okay we can say that it is the weight in
kilograms per metre cube. Okay and V is the volume therefore V into rho is equal to total weight
V into rho that is the total weight of the steam therefore rate of change of total weight of the
steam content is equal to this difference Qin minus Qout a very simple mechanism that there is a
vessel where steam is entering at a certain rate steam is leaving at certain rate right and this
difference will represent the rate of change of way to the steam inside the vessel.
Now basic meaning of the terms here are W is the weight of the steam in the vessel in kilograms
which is equal to V into rho, V is the volume of the vessel in metre cube, rho is the density of
steam in kilogram per metre cube and Q is the steam mass flow rate, Q represent the steam mass
flow rate kilograms per second, T represent time in seconds a very simple equation. Now here
we will relate the the rate at which the steam is leaving the vessel to pressure right.
Now let us say that Qo is the steam leaving the vessel per second when pressure is Po right
therefore at any other pressure the Q out will be equal to Qo divided Po into P where, P is the
pressure which is different from the nominal pressure, okay P is the pressure of the steam in the
vessel Po is the rated pressure and Qo is the rated flow of the vessel it means at a particular
pressure the steam flow is Qo and at any other pressure the rate of steam flow or steam flow is
proportional to the pressure.
Now to develop the equation we need actually the expression for deer d rho by dt okay, Now to
obtain the expression for d rho by dt. We can write like this d rho by dt can be written as dP by dt
that is the derivative of pressure with respect to time or the rate of change pressure with respect
to time into d rho by dP that this chain rule type of relation we have written that d rho by dt is
equal to dP by dt into d rho by dP or you can put other way around d rho by dP into dP by dt that
will be much better if you write down d rho by dt as d rho dP into dP by dt. Then, we substitute
this expression in our continuity equation we get Qin minus Qout equal to volume into d rho by dP
into dP by dt.
Okay, now here dP by dt can be replaced dP by dt can be replaced from this expression that is
you can write down you can just take derivative of this equation. So we write down the dP by dt
is equal to Po by Qo into dQ by dQout by dt. Okay use this expression so that I can write now Q in
minus Q out equal to V into d rho by dP Po by Qo into dQout by dt. Okay here in this equation in
this equation we will represent this total term V d rho by dP into P o by Qo by a term a constant
called Tv dQout by dt.
Now here this term d rho by dP is a known quantity that is how the the density varies with the
pressure this is the rate of change of steam density with pressure this can be this is a known
quantity therefore all these terms are known Po is known Qo is known therefore this coefficient
will be represented by a constant Tv at at right and ultimately you will see that this becomes the
time constant of the vessel. Okay therefore, now I have a simple equation a first order differential
equation this Qin minus Qout equal to Tv into dQ out by dt you take the Laplace transform of this
equation. Okay and then you can express the the Qout in terms of Qin right and you can write
down the transfer function in this form Qout by Qin equal to 1over 1 plus s times Tv.
A simple first order transfer function which we have obtained here for this vessel, okay. Now
this this is the these are the Laplace transform variables output input and this Tv is the time
constant of the vessel. Okay now you understand actually the what are the parameters which
determine this time constant Tv of the vessel okay because if you see it here again T, Tv depends
upon the volume of the vessel these are the vessel right mode will be the time constant it will
also depend upon these parameters what is the, what is the ah Qo for a given value of Po and this
quantity that is the property of the steam okay therefore with with knowing these quantities or
these parameters this Tv can be computed. Now let us come to the actual turbine representation
or steam turbine representation. So far the steam turbines are concerned we have different
configurations it all depends upon the rating of the machine right.
Now this block diagram represents the simple tandem compound single reheat steam turbine.
Now here the basic blocks in this tandem compound single reheat steam turbine are that we will
have control valves and steam chest that is that is the control valves right and the position of the
control valve is adjusted by the governor, okay. The governor controls or speed control
mechanism right they will be adjusting the position of the valves, okay. Therefore first block
diagram which I can show here is h control valves and steam chest these valves are enclosed in
an enclosure right which also houses some substantial amount of steam right therefore we call
this is a steam chest.
The steam from this through this control valve enters the high pressure section of the turbine.
After developing some power in the high pressure section the steam temperature and pressure
reduces it gets expanded right therefore it is send to re heater after re heating the steam in the
superheated condition again and then it enters the next stage of the turbine is called intermediate
pressure turbine section, IP. The steam leaving from the intermediate pressure section enters the
low pressure sections through cross over pipings okay and then it is exhausted to the condenser.
Now in this schematic diagram we have not shown certain valves which are provided in the
system. For example I have not shown the interceptor valve there is a interceptor valve provided
actually here in the reheater circuit that is the steam leaving the high pressure section there will
be an interceptor valve we call IV. Similarly similarly we have here a control valve which we
will be showing but along with the control valve there will be stop valves also therefore there
will be always two valves stop valve which will be used to completely stop the flow of steam
similarly along with the interceptor valve also we have a stop valve right.
Now in this models which we will be developing we will not be representing this stop valves.
Similarly, I am not showing here the interceptor valve but they can also be represented okay
further depending upon the size of the machine we may have a single stage reheating or two
stage reheating. In case it is a two stage reheating we we will have one more section then the first
section is normally called VHP which is a very high pressure turbine section, next will be called
high pressure then intermediate pressure and so on that is the steam which leaves the very high
pressure section and after working and when it leaves the very high pressure section will enter
the reheater again it comes out enter the high pressure when it leaves again and enters into the re
heater and so on.
Therefore it is called two stage sometimes we have other configurations which are called
compound configuration there here you will find that all these sections are mounted on the same
shaft that is high pressure interference, intermediate pressure low pressure and then generator but
in the compound type of steam turbines they have two shafts right and but both this you know
turbine sections are controlled simultaneously right they are not independent if they are
independent then they become actually independent steam turbine sections okay.
Now our interest is to develop the transfer function model relating the power output from the
turbine and the turbine output is denoted as mechanical torque or mechanical power okay and
valve position that is what is the position of the valve right therefore, we will relate the relate the
mechanical torque to the valve position that is suppose a time T equal to 0 if I change the
position of valve right then how the output at the shaft will vary right.
Now you can easily see here now that we have control valve and steam chest therefore this is a
steam vessel you can call it, high pressure section is also a steam vessel because here we have
two types of blades stationary and moving right and the steam expands in the nozzles which are
formed in the each section right therefore this also behaves like a vessel this also behaves like a
vessel this also behaves like a vessel therefore the we can represent represent each of these by a
time constant and the time constant will depend upon what is the volume of this what is the
pressure and other parameters, okay. Further, suppose actually the total power is generated, the
total say mechanical torque developed is say Pm.
Now this total mechanical torque is developed in these 3 sections okay a part of it is produced
here in high pressure section half part in intermediate pressure section, part in the low pressure
section right and generally generally you will find that around 70 percent to the power is
produced due to the intermediate pressure and low pressure section and about 30 percent is
produced in the high pressure section therefore we have to account while modeling the section of
the power which is produced in each of the sections okay.
This is the simplest transfer function model of a single reheat tandem compound steam turbine.
The configuration of that particular tandem compound single reheat steam turbine we have just
now seen, the first block diagram this one this represents the steam chest and control valve the
time constant associated with this is TCH, CH is stand for the chest, okay. A certain amount of
power is produced that is PGV is stand for position of the governor valve PGV, P stands for
position PGV define the position of the governor valve even this governor valve position is
calibrated in terms of power right then whatsoever power is produced here in the, now this TCH
will also include the time constant associated with the high pressure section, high pressure
section right.
Therefore this block diagram represents the the vessel that is the chest along with the high
pressure therefore the total is represented by time constant TCH, then when the steam enters
through the reheater okay now the reheater pipings you can say the pipings are there and it has
capacity to store a large amount of steam right and therefore associated with the heater is the
time constant called TRH right. Therefore we represent this reheater separately as one upon one
plus S times TRH and at the end the steam which leaves the intermediate pressure because this
reheater along with the intermediate pressure is also that any time constant associated with the
intermediate pressure is included in the re heater then at the end we have cross over pipings and
the lower pressure section therefore we represent by one more time constant called TCO, CO
stand for the cross over pipings right.
Now at this point some power is generated in the high pressure section if I call this power as PHP.
Okay the power which is produced here in the intermediate pressure section I will call it as PIP
and power which is produced here in the low pressure section we will denote by PLP. Okay now
the total power which is going to be produced will be PHP into FHP plus PIP into FIP to PLP into FLP
and the total is Pm. Okay this is what is the therefore now we have a complete transfer function
relating the mechanical power Pm which is developed by the steam turbine and the governor
valve position.
In case you have another configuration instead of having single reheat if you have a 2 stages
reheat system right then we may have to add one more section right therefore depending upon
actually the what is the actual configuration one can develop a transfer function relating the
governor valve position to the mechanical power input and for each of these sections we can
write down the differential equations because when we write the model right the model is to be
written by writing a set of differential equations. For example in this particular case there are
three time constants and therefore three differential equations will be required to describe the
steam turbine.
Now the state variables state variables in this case will be PHP, PIP and PLP these are the 3 state
variables okay and in this particular arrangement if you want to study the small signal stability
then there will be the incremental changes like the delta PHP will be the state variable, delta PIP
will be state variable but when I am writing the complete non-linear model then these quantities
as it is will become the state variables and therefore we will require the initial values to solve the
problem. Therefore whenever I want to solve the stability problem I should know what is the
initial value of this PHP? initial value of PIP and initial value of PLP they are my state variable any
doubt.
Now the the terms associated were explained I will just reiterate PGV is the output power from
the control valve or we can even represent the output position of the control valve but that is also
expressed in terms of power calibrated in terms of power okay TCH is the steam chest time
constant TRH is the re heater time constant TCO is the steam storage or cross-over time constant.
Now here I wanted to just mention that out of all these time constants the TRH the re heater time
constant is the largest its value is around 10 seconds, 5 to 10seconds while this TCH and TCO the
time constants are of the order of .2, .3 like that fraction of a second a small right and once when
you see that the time constant TRH a large time constant right then whenever you are trying to
study the midterm or long term dynamics right then this model will have its effect on the
stability.
Now, similarly this FIP stands for IP turbine power fraction FLP is the LP turbine power fraction
Pm is the equivalent generator input mechanical power, this is the mechanical power input which
we have been talking about. Now the 3 differential equations which you can write can be written
in the simple form, 3 differential equations are because there were 3 single order a first order
transfer functions for each transfer functions you can write down one differential equation d by
dt of PHP as PGV minus PHP divided by TCH similarly, d by dt of PIP equal to PHP minus PIP divided
by TRH, d by dt of PLP equal to PIP minus PLP divided by TCO and the mechanical power Pm is
equal to PHP, FHP plus PIP, FIP plus PLP, FLP okay.

Now these 3 differential equations right we will will form the the model for the steam turbine
and if you go for slightly you can say more detail not more detail but for different configuration
you can write down the number of equations required to describe that particular configuration.
Now let us see what is the sum of these 3terms, now these 3 fractions that is the fraction of
power which is generated in the HP section plus fraction of power which is generated in the IP
section fashion ((00:34:47 min)) of power which is generated in the LP section, this sum should
be equal to 1 okay.
Now we will discuss the speed governor for steam turbine or we can say that speed turbine
control, steam turbine control. Now before I go into details I wanted to specify here mention here
that over the years there has been lot of development in the governing mechanism and the
associated controls ah originally the mechanical hydraulic controllers are used they are still
actually in operation but these days they have what we call electro hydraulic controllers or
electro hydraulic governors but remember that in all these the hydraulic servomotor plays very
significant role.
Now when we look into the control for the steam turbine the main function of this control is to
open and close the steam turbine valves this is the end objective that we want to adjust the steam
turbine valve position so that the required amount of steam flows this is what and these valves
have to open against the pressure of the steam and they are are their inertia and their weight is
very substantial NTPC engineers know actually that their weight may be in terms of few tons,
am I right and therefore to make this valves to move, you require huge amount of torque or
pressure right and therefore this pressure amplification is done in hydraulic servomotors,
required amount of pressure amplification is done in hydraulic servomotors therefore here the
basic components of speed control mechanism can be look like this.
We have a speed governor first block is the speed governor, input to the speed governor is the
speed actual speed of the machine then the output of this governor will be the speed governor
position. Okay depending upon the actually the speed there is some position of the governor this
enters into the first block here. We call speed relay right I will discuss these three these
components separately but there is a speed relay to this speed relay we have two inputs coming
one is the one is the governor speed changer position, another is the speed governor position
because as I have told in beginning in the first block diagram which had shown you therefore
there were two inputs coming, one coming through AGC which is the speed changer position
where you sometimes you can use that instead of speed changer position actually the the power
output position your reference what the set point you can call it.
Okay and this set point can be adjusted by the AGC automatically otherwise it is done manually
and another signal coming is through the governor these two signals enter the speed relay, speed
relay is comprising of a pilot valve and a servomotor then there is one more servomotor
hydraulic servomotor to amplify this power and then we have the governor control valves and
output will be the valve position therefore you can easily see here that input to this control
mechanism.
We have two signals one is a speed another is the speed changer position and the output is the
valve position because when talked about the steam turbine models right the input to the model is
the valve position right and now in this particular case the input will be speed and the set point
these are the two inputs.

The governor which senses the speed right. This is the mechanical speed governor a flyball type
of mechanism and although there is a lot of developments which have taken place in sensing the
speed still flytype, flyball type speed governors are in operation these NTPC people is it
operating in your system now or replaced therefore now the we use what is call actually the
electric sensor, the speed sensors we call it for the speed sensing instead of using this mechanical
flyball type of arrangement we use actually the electric speed sensors basic simple actually you
can use a permanent magnet generator PMG and the speed can be sensed by knowing the voltage
output of that okay.
Now here irrespective what type of you can say the speed sensing you use right, we have a signal
which is proportional to the speed okay. Now this diagram shows a simple speed relay pilot
valve. Now when I talk electro hydraulic governor right the mechanical hydraulic governor or
controller the M is replaced by E therefore basically the speed sensing mechanism is replaced but
so far processing is concerned still we have this speed relay and servomotor, hydraulic
servomotor right because the amount of power which is required is enormous the amplification is
very high level, okay. Now in this arrangement I will just show you here actually this
arrangement if you see this this rod which is coming here that represents the position of the
speed governor.
Now when this rod actually will is attached to a lever like this and this lever is attached to a pilot
valve and a spring loaded servomotor. Now here actually you will see a very interesting thing
that always whether I talk of a hydraulic servomotor or a pilot valve right there will be linkage
between the the valve piston and like the servomotor piston this linkage is basically to provide
what is called negative feedback it is a negative feedback always. Now the arrangement is very
simple here the high pressure oil is here you maintain right.
When the pilot valve suppose suppose actually the the pilot valve is made to move down okay
then this high pressure oil will move into the servomotor and push the piston up the the the
moment of this piston when it is moving up it is actually restricted by or controlled by the spring
pressure, okay. Now you can easily see that when this is trying to move this will also come down
while at the same time when this oil enters and tries to push it up through this same mechanism
this pilot valve is also pushed up therefore this acts as a negative feedback.
Now of course here in this diagram it is shown actually that as if this servomotor is controlling
the steam valve but in case we need one more stage of amplification then this rod or this piston is
going to control the hydraulic servomotor. In this hydraulic servomotor again the similar
arrangement is there you have this pilot valves high pressure oil comes here right and the this
linkage shows actually the connection to speed relay right and you can easily see actually that
when this lever moves up right the this pilot valve moves goes up it opens this and the oil will
can come down when it opens what happens is the oil will come down from this side and it enter
from the top right, in the sense that is moving down that is actually when you move it down the
oil will come from the from the the end, oil will leave from these two sides but enter through
high pressure section only and it will move from the top push it down and oil will come out like
this. Therefore through this mechanism the hydraulic servomotor actually you will be because of
the high pressure oil which is maintained right therefore this power amplification takes place.
Okay and the amount of amplification will depend upon what is the what is the cross section area
of the hydraulic servomotor or piston right and what is the pressure which is maintained in the
high pressure oil right these are the parameters which will determine. Now when we try to
develop the model for these devices you have to look into two aspects one is that the moment of
this piston piston for a certain particular opening will depend upon the flow of oil at a certain
pressure therefore this moment can be considered to be function of time as the time passes it
increases therefore basically this can be represented by an integrator.
It will represent because when the oil is flowing at certain rate the piston will keep on moving
right the position will keep on changing. Further further since in these all the devices there is
always a negative feedback. So that output whatsoever is the output position here is fed back
directly as a negative feedback, okay another thing which you have to see is the that there are
two limits which we will be talking about one is called actually the rate limit that is the rate at
which this valve is going to move up and down, another is the position limit, upper and lower
position limits that is when the this can go to the top it will come down to the bottom there are
two limits okay therefore there are two limits which will be shown one is called actually rate
limit another is called the position limit. While when you talk about actually the speed relay
these limits are normally not required to be shown.
Now this block diagram represents the transfer function model of simple steam for a transfer
function model of steam turbine governor. Now we can just start like this this is the speed
governor, input to this is the speed okay therefore output of this will be a K times speed. They
are basically it is a it is a speed sensor speed deviation is going to be sensed here. Now this is
going to compared with the reference the error comes here.
Now this block diagram which I have shown here this block diagram it represents actually the
transfer function of the speed relay, it is something like this. Speed relay we represent time
constant of this as TRS, TSR not RS, TSR I can put it we will call it TSR speed relay an integrator
this is basically integrator. The position is fed back there is a negative feedback loop here okay
and this is the plus this is minus. Now this can be replaced by an equivalent transfer function one
over 1 plus s times TSR that is this transfer function which has actually which represents the
integrator and a negative feedback is represented by a first order transfer function of this form
right.
Similarly, here actually is the transfer function of the servomotor, while in the servomotor we try
to represent the rate limit and the position limits and therefore these two this is this is the time
constant one upon TSM servomotor time constant, one upon TSM at this point we can sense what
is called the rate and after integrating we can sense the position right therefore these two are
provided. Then the here you will find actually the there is a the control valve moment and
actually the servomotor moment there is a non-linear relationship and therefore this is the valve
characteristic.
Now in order to linearize this we actually a which will have a characteristic opposite to this. So
that when we put together it will be linear characteristic now this is the complete transfer
function model of a steam turbine governor. Now here I have not shown actually in this diagram
but the speed relay output can be used to control other valves also because you may have for
example interceptor valve, interceptor valve can also be controlled through this speed relay that
is why it is shown other valves okay.
Now using because for stability studies, for stability studies we can model this non-linearity this
non-linearity can be modelled and we can write down the transfer function or the differential
equation relating the output of speed relay to the input. Similarly we can relate the output of the
servomotor to the input and these relations are available therefore we can write the complete
mathematical model of the and you can see really that this CV which I have shown here control
valve position is basically the PGV, PGV the governor valve position or control valve position they
are same.
Okay now with this ah let me summarize our presentation today. We have discussed today the
the transfer function model of steam turbine and steam governor a steam turbine governor. We
have also developed the differential equations which can be used to represent the steam turbine
and the differential equations to represent the the steam turbine governor okay. Thank you!
Prof. M. L. Kothari
Lecture - 24
Dynamic Modeling of Hydro Turbines and Governors
Friends, today we will study the dynamics modeling of hydro turbines and hydro
governors. We have studied the dynamics models for steam turbines and a turbo
governors, the characteristic of the hydraulic turbine are different from that of a steam
turbine and therefore different dynamic models are required for hydro turbines the the
physical arrangement for a hydro power plant is shown in this diagram.
This is the hydro turbine, this is the water storage tank, we will call it fore bay and this is
the penstock through which the water flows from storage tank to the hydro turbine. The
transient characteristic of the hydro turbine is determined by the dynamics of water flow
in the penstock and we shall determine the transfer function of the hydro turbine. Now
when we consider small perturbations about a steady state operating condition then we
can relate the various parameters by linear equations.
Now this equation that is q equal to a11 h plus a12 n plus a13 g and another equation m
equal to a21 h plus a22 n plus a23 g. Now this these two are the algebraic equations in this
equation q stands for per unit deviation in flow, h stands for per unit deviation in head, n
is per unit deviation in speed, g is the per unit deviation in gate position, m is the per unit
deviation in torque.
There we can say here therefore small perturbations we can relate the per unit deviation
in flow to per unit deviation in head, per unit deviation in the speed, per unit deviation in
gate position. Similarly the per unit deviation in torque is related to per unit deviation in
head, per unit deviation in n that is the speed and per unit deviation in gate position. The
now these are the two algebraic equations which relate the small change in the water flow
water flow following following change in head speed and gate position.
Similarly this is the change in torque produced following the small changes in head speed
and gate. Now using these equations and using the dynamics of the water flow in the
penstock a transfer function model for the hydro turbine is obtained. The transfer function
relating the mechanical power output of the hydro turbine to the gate position or position
of the gates that is PGV is the position of the gates is related by these constants which
were shown in those algebraic equations and an important term Tw.
Now this Tw is the water time constant or water starting time constant of the hydro
turbine. We will discuss about this time constant Tw in detail, further these coefficients a
one one a 11, a 13, a21, a22, a23. These can be interpreted as the partial derivatives of q with
respective to head partial derivative of q with respect to speed deviation and so on now
for ideal turbine which is normally considered to be a loss less turbine, the n for rated
speed ideal turbine at rated speed these coefficients are having their values equal to a11 is
.5, a12 is 0, a13 is 1, a21 is 1.5, a22 is minus 1 and a23 is 1. Therefore, if we make use of
these values and obtain a transfer function relating the mechanical power developed to
the valve position.
Now this transfer function comes out to be a very simple transfer function of the form
one minus STw divided 1 plus 0.5 STw. Now let us understand what are the factors on
which this water starting time or time constant Tw depends upon.
Now this time constant Tw is associated with in write Tw is associated with the
acceleration time for water in the penstock between turbine inlet and forebay but
basically represent the acceleration time. Now the equation for this time constant Tw is L
into V divided by HT into g, now here here L is the length of the penstock in feet, V is the
velocity of the water in feet per second, HT is the head in feet and g, g here is the
acceleration due to gravity, g here is the acceleration due to gravity its value is feet per
second square its units are like that that is if you put actually value of g it comes out to be
32.2 feet per second square 32.2.
Okay now what we do is that there exist a relationship between the power developed and
the various parameters of the system therefore in this equation the velocity will be
replaced by power and other terms that is the power generated is given by this equation V
into HT into A into e divided by 11.8, here P is the power in kilowatts, V is the velocity in
feet per second velocity of the water in the penstock, HT is the head, A is the cross
sectional area of the penstock in feet square or square feet and e is the combined
efficiency of the generator and turbine is the total efficiency divided by one therefore
what we do the from this equation we express the value of V in terms of power, head, A
and e and then substitute this value of the expression of V in this equation 24.2.
When we make this substitution we get the expression for water starting time constant T w
equal to 0.366 PL divided by HT square A into e. Here the value of g has been substituted
as 32.2. Now this time constant Tw is a very important parameter. So for the model of the
hydro turbine is concerned the hydro turbine the transfer function has some special
characteristic and we will discuss the special characteristic of this hydro turbine transfer
function.
The transfer function of the hydro turbine which we have just now derived is 1 plus TwS
1 plus 0.5 TwS this is minus here, correct thank you, this is minus 1 minus TwS 1plus 0.5
Tw okay is this, okay is minus. Now we can write down here mechanical power output Pm
and the gate position PGV.
Now this transfer function has one special characteristic that is the if I look at the poles in
0s of this transfer function then it has 10 in the right half of the S plane and a transfer
functions or the systems the systems which have at least 10 or one pole in the right half of
the S plane are known as the non minimum phase transfer functions or non-minimum
phase systems.
We shall analyze the special characteristic of this turbine hydro turbine by considering
unit step input to this turbine that is we change the valve position suddenly and once we
give a any step input to this hydro turbine transfer function right then we can write down
the expression for mechanical power developed.
Now here instead of considering a PGV, we will prefer to consider this deviations
deviations in the gate position okay and deviation in the mechanical power developed
therefore when I consider the deviations I can write down the model in this form that is
delta PGV is the input and output is delta P, okay.
Now when we consider the unit step change in the position of valve or gate of this hydro
turbine then delta PmS can be written as 1 by S 1minus TwS over 1 plus 0.5 TwS. Okay
because we have considered the unit step change in gate position. So that the transfer
function ah not the transfer function where the Laplace transform of unit step input is 1
by S.
We shall write down this function in this form 1 by S take this Tw out, so that we can
write here 1by Tw minus S divided by take the 0.5 Tw. So it write down 2 by Tw, okay
plus S that is we are writing in the in the form S plus A or here actually is the S minus A
this is the numerator directly it with not going to come actually in our expression for the
response. This can be written as 2 by S 1 upon Tw minus S divided by 2 by Tw plus S.
Okay, we obtain the time response by taking the inverse transform of this transfer
function that is we can obtain delta Pm(t) by taking the inverse transform that is Laplace
inverse of this whole function which I can write down here as 2 by 2 by S1 by Tw minus S
divided by 2 by Tw plus S, okay. Now we obtain the partial fraction of this and the partial
fractions will come out to be like this and this will come out to be equal to Laplace
transform of Laplace inverse of 1 by S minus 3over 2 by Tw plus S that is if you simplify
this expression you will get this expression you can just check it.
Okay therefore now we can say that delta Pm(t) is equal to 1 minus 3 to the power minus
2 by Tw into T this is the this is the response of the hydro turbine to step unit step change
in valve position, okay. Now if I put t equal to 0 that is at time equal to 0 that is delta
Pm(o) is how much is equal to minus 2 and if I put now t equal to infinity in this equation
I will get delta Pm infinity equal to 1. Now we can you can appreciate actually or the
difference in the normal response which we get for the transfer function for a system
suppose there is a system and we give a some input we except that the output should start
following the input.
Now here you find that when at time t equal to 0 when I give a step change in valve
position then immediately the change in mechanical power is minus 2. We have given a
unit step input but the output is negative and this is as high as minus 2 and under steady
state condition the output is same as input right. Now this is the main difference or many
special characteristic of the hydro turbine.
Now if I plot the response that is time t delta Pm, we start at minus 2 and t equal to 0 its
value is minus 2 and it settles to value equal to 1and it is very exponentially, so that
response is of this valve the time constant is time constant is Tw by 2 right that is if you if
you draw a transient to this response curve at t equal to 0 then the curve will be the this
transient will be like this and this time can be identified as Tw by 2, the meaning is that if
the if the change in mechanical power changes at the initial rate then it will reach its
steady state value in time equal to Tw by 2 right, otherwise it is very exponentially.
Now to understand what is the implication of this special characteristic of the hydraulic
turbine on the type of governor which will be required. Now let us first consider that we
use a simple governor no special features, a simple governor and let us see the what will
be the requirement for stable operation for this system.
To understand this the a special requirement let us start with the simple governor we can
represent the governor by a simple transfer function 1by R the transfer function of the
hydro turbine or hydraulic turbine is 1minus TwS, 1plus 0.5, TwS, output of this hydro
turbine will act on the inertia of the turbine generator system is 1 upon 2 HS, I am
neglecting the damping term this becomes my speed deviation delta omega R and here is
a negative feedback. Okay this is negative this is change in reference this this term is
delta omega reference I give as and is a plus.
Now to analyze the stability of this system what we will do is we will consider, we will
consider the H equal to so let us say H equal to 5, okay. Let us take Tw equal to what was
the taking 22. Okay now with this parameters we shall find out that what are the
requirements on speed regulation parameter? Our basic requirement is it is a closed loop
system and this system should have high degree of stability it should have a required
stability margin.
To analyze this requirement we write down the characteristic equation of this system, the
characteristic equation can be written as 1 plus 1by R, 1 minus TwS over 1 plus 0.5TwS
into 1 upon 2 HS. Now here here this this is the forward loop transfer function 1 by R, 1
by R into this transfer function into this this is the forward loop transfer function and the
the feedback is unity feedback and the of the characteristic equation is 1 plus gH which is
the standard formula and it comes out to be equal to that is 1 plus 1 by R.
Now will let us substitute the value of Tw equal to 2 and H equal to 10 and let us see 1
plus 1 by R,1 minus 2S, 1 plus S, 1by 10s okay. We can write this as 10 SR plus is it
okay or we can write this is in the form now 10 R S square plus 10 R minus 2 into S plus
1 equal to 0. Therefore this is a second order characteristic equation and for this second
order characteristic equation to have its roots to lie in the left half of the S plane. Our
requirement is that these coefficients should be positive therefore from this consideration
that these coefficients to be positive one requirement is now the 10 R should be greater
than 0 or we can say R should be greater than 0 the second requirement here is the 10 R
minus 2 should be greater than 0 this puts the requirement that R should be greater than
.2, it means from the consideration of stability the minimum value of R which we can
choose for this governor is 20 percent.
Generally the the speed drop which we choose out or which is the governor parameter or
2 parameter of the governor is taken around 5 percent. Now if I take the R equal to 5
percent then this system is going become unstable, okay. Now suppose I want actually
the response to be critically damped then for critically damped response our requirement
is 10 R minus 2 whole square minus 10, 4, 40 R equal to 0 okay that is b square minus 4
ac should be equal to 0 and if you solve this equation the value of R which is required
comes out to be .736 and another value of R is 0.05, what is the value 536 okay.
It means we get 2 roots of this equation one root says that the value of R or if I call it R1
we call this as R2 one root says that it should be R should be .736, the meaning of .736
means it is a 73.6 percent right and when I say R2 equal to .0536 it is 5.36 percent but if
you assume R2 equal to .0536 then it does not made this requirement let R is greater than
0.2 and if you take this value of R equal to .0536 this will come out or this will result into
a damping is so equal to minus 1 that is negative damping and with R1 equal to .736 the it
will have a critically damped response where will be equal to 1 right but this is too high
and therefore in order to overcome this problem the the hydro governors which are
provided have to have different characteristics.
Now we will discuss the characteristic of the hydro governor which is required for hydro
turbines. Now this is the block diagram of a hydro governor that is the governor for hydro
hydraulic turbine. The the building blocks of this governor or a speed control mechanism
or the hydro governor is similar to that we use in a steam turbine, the only difference
which we will see here is that this block dashpot okay.
Now let us just discuss what are the main blocks, the first block we can see here is the
speed governor, the speed governor sensor the speed and it gives you a position which is
which is proportional to the speed deviation or speed right there is the speed governor
position. This first block pilot valve and servomotor, the input signal to this pilot valve
and servomotor are the speed governor, speed changer position which is the signal which
comes from AGC of the system and another is the position of the governor speed
governor position. These two are the input signal to the pilot valve and servomotor, the
output of this pilot valve and servomotor goes to another servomotor hydraulic
servomotor which is named as distributor valves and gate servo motor.
Now these 2 servo motors are required to amplify the power which is required to move
the gates and the output of this hydraulic servomotor control the governor control valves
or gates of the hydro hydraulic turbine and therefore output is the gate position, okay.
Now this dashpot is to provide feedback signal and this feedback is a derivative feedback
this dashpot is to realize if derivative feedback and you can see actually that there are 2
input signals coming one is directly coming from the position of the servomotor another
is coming through dashpot right. Now these 2 these two feedback signals are required to
obtain the required characteristic for the hydro governor.
Now this shows or this block diagram shows the transfer function of various building
blocks of the hydro turbine governor or hydraulic turbine governor we start like this, this
is the this is the transfer function of the pilot valve and servomotor this is the time
constant of the distributor valve and servomotor TG stand for time constant of the
distributor valve and servomotor. Here we show the rate limits then this transfer function
one by S to represent the integration function and the another limits which we have is the
position limit right, therefore these two limits are shown in this transfer function diagram
and whenever we are studying the small perturbation dynamics or whenever the system is
subjected to small perturbation then these these limits particularly may not be touched
particularly the the position limit may not touch but the rate limits may have to
incorporated. Again, you can see here there exist a non-linear function to relate the gate
position to the hydraulic servomotor position that is when the hydraulic servomotor
piston moves right it moves the gates and the movement of gates and the movement of
the servomotor they are again related by a non-linear function. Therefore this show the
nonlinear function the transfer function of the dashpot is delta STR over 1plus STR.
Now this transfer function you can see here that it is the derivative feedback, this is a
derivative feedback because we have a a term S in the numerator that is STR or 1 plus
STR this becomes the derivative feedback then this term delta delta is called transient
drop delta is called transient drop then the another feedback which was shown in the
block diagram is through this sigma.
Now this sigma is same as R which we use actually in the models for the hydro for the
governors. Now this sigma is known as the permanent drop, permanent drop the meaning
here is that when the system is in dynamic condition right there will be output from the
dashpot and the movement the system attain the steady state condition right there will be
no output signal from the this transfer function and therefore the the net drop will be
determined by the permanent drop sigma only, okay. The typical, the typical parameters
of this hydro governor are this TR is 5 seconds its it ranges in the range of 2.5 to 25 a
wide range TG, TG is the time constant of the distributor distribution and valve and
servomotor this time constant is small and its value is 0.2 second and its range is .2 to .4.
The TP which is the time constant of the pilot valve and servomotor this time constant is
very small of the order of .04 second and it is in the range of .03 to .05 second. The delta
the temporary drop its value is .3 and its range is .2 to 1, the sigma or R its range is its
value is .05 and it varies from .03 to .06 percent. Normally it is 2 percent or one in drop is
considered and some places it may be lower than this or slightly higher than this but its
the range is very narrow or is .03 to .06.
Further further this time constant TR,TR is depends upon the water starting time constant
right, therefore the typical value is the TR should be equal to 5 times Tw this is the thumb
rule okay and suppose if Tw is 1then TR is 5, suppose Tw is 2 TR becomes 10 and so on
similarly this temporary drop is 2.5 Tw divided by 2H. Therefore you can say that the
temporary drop requirement depends upon the inertia constant of the turbine generator
system and the water starting time right because as we discussed actually that this special
feature that we have to provide a temporary drop compensation right particularly to take
care of special characteristic of the hydraulic turbine and there is a relationship which
relates a temporary drop to water starting time constant Tw and inertial constant H.
Now the question arises the how do we optimize the parameters of this hydro governor,
one way is that okay you take this thumb rules that is you set the value of delta as given
here in the range of .3 around .3, .221 or use this formula delta is 2.5 Tw divided by 2H.
For example in this case when I put Tw equal to 1 and H is taken as 5 how much it comes
out to be .25, 2.3 is one which we are talking about but however in any particular system
particular system the procedure is that you you model the complete system take the hydro
turbine, take the its associated governor right and you obtain the dynamic response of this
system and we can use, we can use different techniques to set the parameters of these
governors right.
Generally the governor parameters are set under no load conditions the the general model
for a speed governing system is developed and this is shown in a compact transfer
function like this where input is delta omega the speed deviation signal in the transfer
function is in the form K into 1plus S T2 divided by 1 plus S times T1 or 1 plus S times
T3. The output of this transfer function represents the change in power delta P, now this is
added to the reference setting that is Po which is the reference setting and this output
which is the sum of Po and delta P right is the gate valve or gate valve position okay.

Now here as you know actually that this Po is the position of the speed changer, in the
AGC when we talk about this is the speed changer position. Now this time constants T1,
T2, T3 right. These time constants are can be obtained from the typical values which are
given here for this TP, TR delta sigma and TG so on that is these time constants T1, T2, T1
and T1, T2 and T3 right are related to the parameters of this governor and this is put in the
simplified form for the purpose of simulation studies, okay.
Now another important point which we have to see here is that the there are different
manufacturers of hydraulic governors, in they are hydraulic governors have been
manufactured by the general electric company by a wasting house and so on right and
therefore this manufacturers give the parameters of the model of the governors and
developed by them right therefore the standard references are available where the
parameters of different manufactures are given, if you look into the reference material
which I have given to you last time that is the committee report right then in this report
the parameters of different types of hydro governors are are given.
Now these parameters are generally the generally the typical values however however for
a any utility for any utility has to set its own parameters by performing certain studies.
Now as we have studied actually that the excitation system and automatic voltage
regulator parameters need to be set, similarly the parameters of the governors need be set.
We are so far actually the turbine parameters are concerned they are constant only
flexibility which we have is actually in setting the parameters of the hydro, hydro
governors for hydro turbine and uh other governor for the steam turbine.
Now this constant K which I have shown here this K is the gain of the governor this gain
is basically if the reciprocal of the permanent drop that is R if R is taken as 5 percent the
gain will become 20 right and the standard techniques which are used for for setting the
parameters are by either by time domain simulation or by frequency response analysis
okay. Let me the sum up that we have studied the dynamic model for the hydro turbine.
We have also seen the special characteristic of the hydro turbine transfer function and to
meet the requirement of the hydro turbine, the hydro governor is required to have a
special feature and that feature is to have a temporary drop concept. Now when this hydro
governor with the temporary drop is provided and is the parameters are properly set it
will give the desired dynamic response for the system. Thank you!
Prof. M. L. Kothari
Lecture - 25
Load Modelling for Stability Studies
Friends, so far we have studied the modelling of synchronous generator and steam turbines,
steam turbine governors, hydro turbines and hydro turbine governors. We have also studied the
modeling of excitation systems. Now this completes the modelling on the on the generator side
right. We have also assumed actually that for power system stability studies the transmission line
network is modelled under steady state conditions.
Now we have to model 1 very important component that is the loads in the system. The
modelling of loads has has important influence on the stability of the system. Further the
modelling of load is the most complicated process, the reason being that whenever we whenever
we try to model the load in the power system stability studies right we identified some load buses
and at that load bus we have variety of loads which are supplied. The load supplied at the
lighting loads, heating loads, cooling loads, motor loads, furnaces and so on right and all these
loads have different characteristics.
Now in our classical stability studies, we represent load at a bus by constant impedance load.
However, the studies have shown actually that this representation may give some erroneous
results and therefore there has been, there has been over the years efforts being put to develop
develop the models of the system. Now the problems is because the the loads are of different
variety and this composition of load at a particular bus is also not fixed it varies from time to
time is a function of time it is also is a function of season, weather conditions and so on and
therefore whenever we develop the load models we do lot of simplifications right. Now today in
our study we will look into the modelling aspect of the loads for power system stability studies.
Now traditionally traditionally for the purpose of modelling or for the purpose of representing
loads in stability studies, the load models are classified into 2 categories, 1 is called static load
models another is dynamic load models. The the dynamic load models are normally required or
essentially required for long term stability study, voltage stability and and when you want to
study the inter area oscillations right. Now today we will briefly discuss how the loads are
represented as the static loads and dynamic loads.
When I talk about the dynamic loads, we have a number of loads which are thermostatically
controlled loads when I say thermostatically controlled loads for example most of the heating
loads are thermostatically controlled loads right, the temperature is controlled and the thermostat
is on and off. Similarly, there are number of examples like refrigerators, coolers right and so on.
Then we have another a very big component in our load that is the discharge lighting loads
where you have the fluorescent tubes, we have the sodier vapour, sodium vapour lamps and
mercury vapour lamps.
Now these are all under the category of a discharge lighting loads. The third most important
component in the dynamically dynamic loads is the induction motor loads in any system about
60 to 60 percent load, 60 to 70 percent load is due to motors and the the modelling of these
motors is a very important and lot of attention has been given to modelling these large size
induction of motors for stability studies right because in power plants or actually in some big
industries we have large size induction motors and their characteristic affect the system stability
as a whole right.
Now first we will address the modelling of static loads when I talk about the modelling of static
loads right then we represent the load load at any time a function of voltage magnitude and
frequency right. Now here the load will be expressed in terms in terms of voltage and frequency
ah using algebraic expressions because its not a dynamic load right therefore the equations which
are used to represent static loads will be algebraic equations okay and since the loads are
function of the magnitude of the system voltage and system frequency we will represent these
loads as using algebraic expressions use in terms of voltage magnitude and system frequency.
Now for representing these load models we will consider, we consider the real power and the
active power components of the load separately because the behaviour of the reactive power
consumed by a load and the real power consumed by the load following variation in voltage and
frequency they are somewhat different. Now let us see what different load models have been
used in the literature. A simplest load model considering the voltage variation is shown in this
form that is the P is the real power consumed by a load at any voltage V equal to Po into V bar to
the power a a is an index a is some index right.
Therefore we can say where this V bar is defined as V divided by Vo, where Vo is the nominal
voltage at which the power consumed is Po right and therefore this V bar is a fraction of the
actual voltage and the nominal voltage or we can say that this V bar is a per unit voltage
expressed in terms of Vo okay. Therefore P can be expressed as the the nominal value of the real
power consumed Po multiplied by this V bar to the power a, then similarly the reactive power Q
is represented as Qo into V to the power b, V bar to the power b, a and b are the 2 different
indices.
Now here suppose if I take a and b equal to 0 both let us say that these coefficients are 0 what
does it mean P equal to Po it means the load is independent of voltage right. Now suppose I take
a equal to 1 and b equal to 1 if I take these coefficients a equal to 1, b equal to 1 what does it
represent yes therefore here here the meaning is that the load load draws constant current,
constant current right and therefore the power is proportional to the voltage, power P and Q
which are consumed they are proportional to the a voltage.
Now if I take these indexes a and b, indices a and b equal to 2 that shows that the the power
consumed is proportional to V square and that is the characteristic of a constant impedance load.
For example in our transient stability studies classical transient stability studies which we
discussed we were representing the loads by constant impedance at that time the the index
becomes 2. Therefore, we have by putting the different values of a and the b we can have
different representation of loads whether the load is a constant power load, constant current load
or constant impedance load.
Now in practice the load behave different loads behave in a different way, different devices. I
just take the example of say incadance incandescent lamp okay. Now this lamp does not
consume reactive power right and the the power consumed by this will be proportion to V
square, V square although it is not very linear for it is for a so small variation it is proportional to
V square and therefore this load will behave like a constant impedance load. While there are
other loads they have different characteristics and in any at any bus, at any bus we have different
devices and therefore we we may divide this load into 3 different categories like say constant
power load, constant impedance load and constant current loads right.

Now if we in case we want to represent a composite load, composite load then these coefficients
a and b will be different for example for a composite load this this index a varies from .5 to 1.8
similarly b has a wide variation that is 1.5 to 6 for composite loads. Again it depends upon what
is the composition. Now here this the loads right which we represent by static loads they
generally have very high response in the sense that when the voltage and frequency changes right
the power consume also changes practically, instantaneously, there is no time delay right and
therefore at any instant of time they can be represented as static loads and the variation can be
represented using algebraic equations. Any composite load, any composite load can now be
represented as in terms of 3 components we normally call this as ZIP representation Z means
constant impedance I stands for constant {pa} ((00:14:08 min)) current and P for constant power.
Therefore we can have a load representation by this equation P equal to P0 multiplied by p1into V
square, V bar square plus p into V bar plus p3 now this p1, p2 and p3 these are the fractions of the
load which represent the constant impedance constant current and constant power fractions okay.
Suppose I take p1 equal to .5, p2 equal to say .3 and p3 equal to .2 that means the 20 percent load
is a constant power load, 30 percent load is a constant current load and 50 percent load is a
constant impedance load.
Similarly reactive power that is reactive power Q is represented as Q0 into V1 q1 into V bar
square plus q2 into V bar plus q plus q3 because it is multiplied by q0 okay. The basic problem
comes is actually is difficult to know this these fractions p1, p2 and p3.
The next important aspect of static load modelling is the frequency dependency of loads, this is a
very important component very important aspect of load modelling that is the loads depend on
the system frequency. In most of the cases the real power consumed by the load increases when
frequency increases and real power consumed by the load decreases when frequency decreases
and in our Indian Power System where we do not have enough generation, enough generation
right the load depending upon frequency right that is the these loads decide actually the system
frequency that is generations are particularly constant when you put additional load frequency
drops right and therefore at a lower frequency lower frequency we are in a position to meet the
additional loads which comes, although the total power which is fed is same right but the
whenever additional loads comes what happens.
Suppose your system is operating at a certain frequency and let us put that I switch on some
additional load of say 1 megawatt this additional load will be met from from where it comes
because the answer is answer is that under steady state condition under steady state condition this
additional load is met by met by 1 component is the release of the existing load plus the increase
in the generation, while in dynamic condition we have all the 3 points that release released by the
existing load, increase in generation and change in kinetic energy.
But the moment you attain the steady state condition right this change in kinetic energy
component become 0 because you attain the constant speed again right this is the way the system
you known operates right when you are trying to match the generation with the load. Okay
because whenever additional load comes the first thing which happens is the the frequency
drops. Okay and initially this additional load is met by kinetic energy mainly change in kinetic
energy then the regulators volt the the governors will come into action will increase the
generation and when the frequency drops the existing load will start consuming less power they
release some power right.
Therefore these are the 3 components and therefore here we are trying to address these thing that
if the frequency changes how the real power and reactive power consume vary. Now this
equation is shown here shows the variation of real power consumed by the load with the
variation of magnitude of voltage as well as frequency that is, if I consider this general equation
to represent the real power variation with magnitude of voltage and I multiply this expression by
this term 1 plus Kpf into delta f right.
Now this term Kpf is the this coefficient is an important it is the characteristic of the system okay.
Similarly the reactive power consumed can be represented as Qo V bar to the power b into 1 plus
Kqf into delta f. Okay in a sense that both P and Q are now represented as function of system
voltage magnitude and frequency deviation okay.
Now you you suppose if we represent the load, load by this expression that is ah ZIP
representation okay. So you can represent P equal to P0 into p1 V bar square plus p2 V bar plus p3
then this whole thing is multiplied by this fraction this factor that is 1 plus Kpf into delta f. Okay
similarly Q can be represented as Q0 into q1 V bar square plus q into V bar plus q3 into this
therefore this becomes a very general representation of a load which now accounts for change in
magnitude of the voltage and also change in frequency that is frequency deviation.
Now here 1 has to understand what are the values of this coefficient Kpf and Kp Kqf. Now so far
the real power consumed is concerned right as the frequency increases the real power consumed
increases and therefore this Kpf will always be positive, Kpf is always positive and this delta f,
delta f is expressed as delta f is defined as the actual frequency of the system minus the nominal
frequency f minus f0the delta f is positive when the actual system frequency is more than the
nominal frequency right and therefore this coefficient Kpf is is always positive but it varies, it is
different for different loads it different for different loads. There are some loads where the power
consumed does not depend upon frequency, lighting load flow particularly this is a you know the
these powers particularly the fluorescent tubes are you know all sodium vapour lamp or mercury
vapour they do depend upon frequency but particularly incandescent lamps.
Suppose I have actually a a room heater a rod type room heater then it will not depend upon
frequency so far magnitude of voltage is same right but since the loads are composite loads and
therefore this is a very important coefficient to be known and is difficult again the estimating this
coefficient is a very difficult task actually for power system those who are taking the course on
power system control and instrumentation we use the term load frequency constant d a parameter
load frequency constant d this d is same as this term Kpf which I call it.
Now what about Kqf that is this is also a positive all through or whether it is always negative or
what should be the value of this Kp. The the response of the audience here is that may be positive
or negative but answer is not correct, is always negative, always negative. We can easily
understand why why it is negative?
The typical values of Kpf is from 0 to 3 and Kqf is in the range of minus 2 to 0 minus 2 to 0.
These are the typical values of these coefficients this is a wide range depends upon the type of
load.
In the literature there is 1 more way of representing this representing the static loads right that is
the static load model can be represented as P equal to P0 multiplied by PZIP plus PEX1 plus PEX2
where PZIP is represented by this expression p1 V bar square plus p2 V bar plus p3 that what we
have already seen then PEX1 that is exponential component is represented as p4 V bar to the
power a1 multiplied by this term 1 plus Kpf1 delta f PEX2 as p5 V bar a2 into 1 plus Kpf 2 delta f.
In fact the you know the load models which have been used actually and used in analyzing the
stability of the system right the the results which are obtained and the actual actual thing which
happens in the ground right. Suppose actually for a given operating condition right my stability
study shows that I have this much stability margin while in practice you will not achieve that
much and that has been the concern actually for the power system engineers because generally
generally we do most of this studies of light using the system models.
Okay and the when you try to we can see what happens on the field in the field right for the
similar operating conditions right sometimes it is seen actually that the the results obtained by
offline studies and what happens on the ground there is a lot of variation and that is why actually
the all through the efforts are being made to have more detailed model of the system because
when I talk about the stability studies stability studies then the the crux of the whole problem is
the detailed model of the system right.
As you have seen actually that we have to model all the sub systems in detail synchronous
generator excitation system, governing system, turbine system and even we are not g1 to the
boiler model system, boiler models are also required, transmission line models are required, load
models are required right and in case there is a big approximation or error in this modelling the
results are going to be different or erroneous and that is why lot of efforts are being made to
develop um better and better model and to estimate this the these parameters by doing some
experiments in the field right this is because suppose I take this model then it has how many
coefficients will be estimated I require this coefficient Kpf1 and Kpf2 in this I have 2 different
coefficient which determine the frequency dependency of loads right. Similarly this p1, p2, p3, p4,
p5 right and these are to be estimated actually for a particular load bus and this is a quite a load
task.
Now we will discuss the dynamic load models, this is all about the static load models and
justification of static load models is that whenever the magnitude of frequency and voltage
changes right the the new value of power consumed changes quickly in a sense the response time
is practically load and therefore we can represent these loads by static models. But there are so
many loads actually which need to be modelled in a dynamic way and particularly this dynamic
modellings or load modelling is needed when we are going ah interior oscillation studies that is
you want study the gear of the system when it is subjected to inter area oscillations is a long term
phenomena.
Similarly we are studying the long term or midterm stability or voltage stability of the system
right during that period the load models or dynamic load models are also required now here the
these aspects should be considered will be the discharge lamps. I have talked about the
discharged lamp, the fluorescent lamps, the sodium vapour lamps and the no incandescent lamp
is not a discharged lamp, mercury arc or mercury vapour lamps.
Okay in these lamps there is a very important actually that the moment the voltage is around .7 to
a .8 per unit this these lamps extinguish and when the voltage recovers within 1 or 2 seconds
they are restored back. Understand this is a very interesting thing that in a power system when
you are doing the stability studies under disturbance condition the voltage at various buses vary
and in case the voltage at a particular point at a particular bus becomes less than about 70 to 80
percent then these discharged lamps extinguish and when the voltage is restored back right then
they they are restored back after 1 or 2 seconds because when I switch on a fluorescent lamp it
takes 1 or 2 seconds to we will put on right. This is 1 of the important characteristic to be model,
then there are many devices which have the over current and thermal relays these are also
required to be modelled. Then the thermostatic control of loads particularly these thermostatic
control loads are they are mostly heating and cooling loads air air heating, air cooling,
refrigerators and some other devices which we can then another very interesting thing which we
have to understand is that in a distribution system we have under load tap changing systems
ULTC sin distribution systems we have some voltage regulators which regulates the system
voltage ULTCs are also for regulating the voltage and then we also have I think distribution
system the switch capacitors, voltage control capacitors right.
Now whenever actually the system voltage, system voltage dips or become slow then this
ULTCs will try to operate and regulate the voltage. You know that is suppose actually at a
particular time the voltage has dipped right then the unload tap changers or under load tap
changers right they will operate to increase the voltage but the response time is quite high, the
the voltage is slow after about 1 minute they start operating and their operation is completed in
about 1 or 2 minutes time, they are slow devices. Then there are many voltage control capacitors
right if the voltage is high the capacitors will be switched off and when the voltage is low
capacitors are switched on right and further the reactive power which is consumed by these
capacitors also depends upon the system voltage and therefore these are all ah the the aspects
which have to be considered while doing the modelling.
In fact actually the 1 stability program which has been developed by the EPRI, Electric Power
Research Institute USA right they have modelled the thermostatically controlled loads and also
the induction motor load in their system they call it the program is called load seal. There are
many you know a companies they have purchase those programs although question is that
whether we are fully utilizing those software packages for operating our system that is a different
issue but detail detailed stability programs accounting for the load models are available globally.
Now here I would like to discuss actually the thermostatically controlled loads. Let us take the
example where where actually the load is to heat a particular space, space heating right and the
differential equation which governs the heating process or temperature control is written by this
equation that is K times d tau H by dt equal to PH by PL. This is very important equation because
once I say it is a a dynamic model then these derivative terms will come, we will have to have
differential equations to represent certain devices that is a constant coefficient K d by dt of tau H
PH minus PL.
Here this tau H is the temperature of the heated area, tau H temperature is the temperature of the
heated area which we want to regulate, we want to control, how do we control is by by
increasing or decreasing the heat generated by the heating device. Now here on this right hand
side we have the 2 terms PH and PL the PH stands for the power from the heater, this PH can be
represented as a function of constant KH, G is the conductance of the system and V square.
Now since actually we are trying to study the stability affect of these devices on stability because
the power consumed here is function of voltage V square right therefore the the power from the
heater is shown as a constant into conductance into V square. Then PL stand for the heat loss by
escape of escape to ambient area atmosphere, this is the heat dissipation which can be by which
can be represented by a constant K into tau H minus tau A the tau A is the temperature of the
ambient area or we can say ambient temperature where G is the load coefficient okay.

Now this equation, this equation has been put in a form where I can realize the transfer function
relating the actual temperature of the system and actual conductance of the system. Now to do
this thing what we do that in this equation we substitute for PH and PL these quantities. Now
when you make this substitution this K times tau H by dt for PH I am putting K H GV square and
for PL we are putting KA into tau H minus tau A.
Okay then you simplify this equation and arrange the terms we will get the term this d tau H by
dt equal to K1 by T1 GV square plus 1 by T1 tau A minus 1 upon T1 tau H, where these quantities
are defined like this T1 equal to K by KA and K1 is defined as KH by KA by just these definitions
okay. Now this thing this equation I have now if you take the Laplace transform of this equation
we can put in the form of a transfer function I will just show you how it is put in the form of a
transfer function.
In any system, any system we want to regulate the temperature right then we start with the
reference temperature let us call it tau reference and we want to know what is the actual
temperature at any time T of the space okay. Now this this equation that is see this equation, this
differential equation relates relates actually the time at a any time T with other parameters of the
system this is simple differential equation, now this can be put in this form.
Now if you see here actually this differential equation then we can represent the this differential
equation by this form that is G is the system conductance, you multiply this G by V square GV
square multiplied by K1 okay to this you add this ambient temperature tau A okay. Then when it
passes through this transfer function 1 upon 1 plus S and T 1 you will get a term tau H. Now in
practice what is d1 is that this error this is the error signal we call it tau error right this will be put
into a controller to regulate the conductance of the system. Generally, we use a PI controller if
you use a PI controller you can have a proportional gain setting KP here and we put integral
control I am just putting on this side is, it can be it is KI divided by some time constant Tc S and
here we have to put the upper maximum limit GMAX. You can add these 2 outputs again there
will be a a maximum limit which has to be put GMAX and this conductance is the 1 which bring
back.
Now this becomes a simple closed loop control system which we can show here to regulate the
temperature of a space which is to be heated. Now here tau H is the state variable state variable
this is only simple and this is the PI control which I have shown. Now depending upon this error
it is going to increase the or decrease the conductance right the more the conductance right. The
more is the heat which is going to be generated right but since we have using this integral also so
that the we want actually that the maximum conductance should be limiting therefore maximum
limit either it may be touched by integral integrator or by the sum because here it is plus okay,
therefore this is 1 way of we can representing a thermostatically controlled loads.
Now under steady state conditions under steady state conditions the tau reference equal to tau H
right. Now if you make this substitution in this equation in this differential equation put the
derivative term to be 0 and I make derivative term equal to 0 right then we can write down the
tau reference equal to K1 times V0square Go plus tau A that is the reference temperature setting
which is required to attain the required value of tau H, where the constant K1 is equal to tau
reference minus tau A divided by Vo square Go okay.
Now with this I will just give briefly introduced actually that how a dynamically controlled or
thermostatically control load can be represented. These differential equations are to be written
along with the system differential equations.
So far actually the discharge lamps are concerned their characteristic can also be shown as
function of voltage but these characteristics are quite non-linear in nature. I will shown in this
axis P and Q consumed by the discharged lamps and on this x axis you put voltage V okay. Now
let us first see the power P consumed by a discharged lamp. At a certain voltage up to certain
voltage say V1, V1 the discharged lamp should be off right therefore no power is consumed by
discharged lamp etcetera when the voltage becomes V1 the suddenly these devices are switched
on and when the voltage increases beyond this the power consumed that is at a when the voltage
is V1 the power consumed suddenly jumps from 0 value to this value then as the voltage further
increases the power consumed may vary linearly and by another segment.
In fact actually you can do ah this non-linear characteristic and represented by a piecewise
linearized model that is you can put at this voltage V1 from V1 to V2 it can be represented as
some KP1 into V and here it can be represented as KP2 into okay while this so far the reactive
power is consumed. The reactive power variation is a very non-linear in nature for reactive
power variation for these devices can be shown this is type of curve we can call this as a Q
which is equal to Q0 the value at this particular point multiplied by V to the power some
coefficient K a very non-linear type of variation right and therefore since the system voltage is
varying and we want to model these devices this type of model is to be shown that in case the
voltage is less than V1 they are off when it is more than V1 to V2, P and Q have to be shown like
this that is the model you have to have that facility where you represent the loads are present as a
function of voltage. Okay the next important component for modelling is the induction motor
load.
Now induction motor load or modelling of induction motor has has attracted the attention of
many researches and many practicing engineers in the past because over 70 percent of the system
load is induction motor load. Now for developing this dynamic model for the induction motor the
approach is similar to what we do for a synchronous motor and we represent the model using DQ
transformation, DQ transformation I would not go into details about all these transformations but
here we have 1 simple advantage actually for so far this induction motor is concerned, the rotor
of the induction motor is a cylindrical 1 right and therefore the the a variation in mutual
inductances and self-inductances right are different as compared to that of a synchronous motor.
Let us just see so far the stator windings are concerned they will they will not see any difference
in the in the reluctance of the magnetic circuit as the position of the rotor changes rotor is
cylindrical therefore depend whichever is the position of the rotor right. The stator windings will
not see any difference in the reluctance and therefore the self-inductance of the stator winding,
mutual inductance between the stator windings are not function of angle this is the advantage,
second point is that the mutual inductance between the stator and rotor windings vary vary
because because the axis of the rotor and stator windings are having different orientations.
When the axis of the rotor and stator windings are aligned they will have maximum mutual
inductance when they are in quadrature it will be 0 mutual inductance that is the only variation
and here the stator voltages are written in terms of the D and Q axis components just as we do in
the case of synchronous machine but here we also require the transformation for rotor quantities
this is the main another difference.
Further for studies the transformer voltages in the stator circuit will be neglected, the stator
circuit for example for stability studies in the synchronous machine we have ignored the
transformer voltages in the stator voltage equation. Similarly, here also we ignore it however
actually the the transformer voltage in the rotor circuit will not be ignored while this so so far
rotor circuit is concerned the rotor is at closed circuit therefore D and Q axis voltages in the rotor
circuit are 0 that is Vdr and Vqr were set equal to okay, these are the simple equations which can
be easily derived and they are available in the standard literature.
The stator of the synchronous generator can be simply represented by a model of this form. This
is simple model which 1 can arrive at for the stator equation because the stator equations we are
not considered the transform voltages right therefore this is a algebraic equation.

We can represent this voltage Vs bar as Rs plus j times X bar into Is plus V prime this is an
important voltage component okay when you represent these voltages in terms of D and Q axis
components we can write down Vs equal to Vds plus J times Vdqs V prime equal to Vd prime plus
j times Vq prime.
Then here the 2 components of this voltage V prime V prime at these 2 components Vd prime
and Vq prime right these these voltages are obtained from these 2 differential equations that is p
vd prime is given this equation and vp vq prime is given by this equation that is they are the 2
differential equations in fact basically these these equations represent the rotor dynamics, rotor
dynamics as I have told you that we do not neglect the rotor dynamics here if you neglect {lo}
rotor dynamics then this model also becomes a static model okay.
Now here there is a term very important term To prime this is the transient open circuit time
constant, transient open circuit time constant and this time constant this To prime characterizes
the decay of the rotor transients when the rotor is open circuited, I am sorry there is mistake here
this when the stator is open circuited when the stator is open circuited. The differential equation
for the induction motor just like actually the we have a swing equation for synchronous jetter.
Similarly, we have a swing equation for induction motor is d by dt of p omega r bar equal to 1
upon 2H Te minus Tm. Now here here electrical power right is the input and mechanical power is
the output therefore Te minus Tm is the accelerating torque right therefore this is the swing
equation for the induction motor and this electrical torque is obtained by this equation Te equal to
vd prime ids plus vq prime iqs. Further the load torque load torque to this Tm is the load torque this
Tm is the load torque right this load torque also depends upon the speed and therefore the
equation which is used to represent the load torque Tm is To omega r to the power m omega r bar
here omega r bar means it is actually the speed of the rotor expressed in terms of or we can say
synchronous speed of the machine. Okay that is you can say it is a per unit speed omega r bar is a
per unit speed another can express in for torque may be Tm equal to To into A times omega r
square plus B times omega r plus C.
Now let me conclude the todays presentation, today we have discussed the static and dynamic
models for the loads which are required to be modelled in the stability studies of the power
system we have modelled the various types of loads and the loads which we consider are actually
the static loads and the dynamic loads, in dynamic loads particularly particularly we have
determined the dynamic model for the induction model. Thank you!
Prof. M. L. Kothari
Lecture - 26
Numerical Integration Methods for Solving a Set of Ordinary Non-linear
Differential Equations
Friends, today we will study the numerical integration methods for solving a set of
ordinary non-linear differential equations.
We have developed the mathematical model of the system for analyzing the stability we
have developed the mathematical model for synchronous generator, excitation systems
turbines and governors and also loads, okay. Now when we want to analyze the transient
stability of the system we need to solve a set of non-linear ordinary differential equations
along with a set of non-linear algebraic equations and we can say that we have to solve
the differential algebraic equations for solving the algebraic equations we have standard
techniques available like say technique Newton-Raphson technique and so on.
Now for solving this non-linear ordinary differential equations we have to resort to
numerical techniques because there is no analytical solution available for non-linear
differential equations. Now the numerical techniques will provide you approximate
solution, I am emphasizing this word approximate solution. Here, when I say that we
apply numerical technique for solving this problem what we do is that the time duration
for which we want to get the solution is divided into small time segments.
We can call this time segments say staring from initial time t equal to t0 to t1, t2, t3, t4 and
say tn and these segments may be equidistant or of the equal length or may be unequal
length that is the time step from initial value to to t1 and from t1 to t2, these 2 time steps or
ah so on may be equal or unequal. Now today we will study these numerical techniques
and broadly these numerical techniques are divided into 2 categories known as the
explicit integration methods for solving a set of ordinary differential equations, we use
the what explicit integration methods and another technique is known as the implicit
integration methods.

We will be in a position to understand what is the difference between the implicit and
explicit integration methods and their merits and demerits? Right, as we will discuss the
different techniques. We will under the explicit integration methods we will discuss
briefly the Euler method, modified Euler method and 4th order Runge-Kutta method.
Although number of other techniques are available in the literature right but however we
will confine our discussion to these 3 techniques which come under the category of
explicit integration methods. We shall discuss 1 simplest technique under the heading of
implicit integration method that is called trapezoidal rule that is trapezoidal rule of
integration, okay.
Now to understand the solution of ordinary non-linear differential equation equation. We

start with start with the a general differential equation, general non-linear differential
equation of the form dx by dt equal to f of xt where, x is a vector x is a vector it will have
its dimension depending upon the size of the problem which we have to solve. For
example, in the classical stability analysis the x will comprise of the angles that is the
rotor angles deltas and rotor speeds like say delta 1, delta 2, delta n, if it is a n machine
system and the speeds like say omega 1, omega 2 up to omega n this will be the state
variables.
Then we will come across a set of differential equations which are put in the compact
form as dx by dt equal to f of xt. For solving these equations, we know the initial
condition that is as time t equal to 0, time t equal to 0, x is equal to x0 that is the initial
conditions are known to us and in knowing the initial condition we have to find out find
out how the x varies as a function of time that is what is the meaning of solving this
differential equation that is we want to get the solution for x as function of time.
The Euler method is the simplest of all the methods and let us consider considers scalar
system that is we have only 1 unknown to be solved to understand the Euler technique we
will start first with the scalar differential equation that is dx by dt equal to f of xt where x
is a scalar right. Now to solve this what we know initially is that x equal to x0 at t equal to
t0 this is the initial condition which is given to us for example, when you solve our swing
equation right at time t equal to 0 we know what is the value of delta that is delta equal to
delta 0 at t equal to t0 right, this is the initial condition.
Now to understand how we obtain the solution of this let us assume that the solution for x
as a function of time is known to us I will say that this is the exact solution of this that is
x varies as a function of time by this very non-linear way. This is the initial point this is
the this point is t equal to 0 and corresponding to this the x is equal to x0 this is the initial
operating condition, okay what we do is we we find the derivative of this function that is
this function that is obtain the derivative at x equal to x0 it means after basically finding
out slope of this curve at x equal to x0 and let us say that this line represents the slope at x
equal to x0 right.
Now let us take a small time step and call this time instant as t1 where, this time step can
be denoted by delta t that is t1 minus to equal to delta t okay. Now if this you take a small
value of this step right then value of x at t equal to t1 can be represented by this equation
that is first we will find out first we find out what is the changed delta x that is this
quantity delta x that is at t equal to t1 right we find out the value of x that is x is equal to x
naught plus delta x, this delta x in the change change when you increase the time by a
small step delta t right.
So that now we can write the solution in the form that is x1 equal to x0 the initial value
plus dx by dt at x equal to x0 into delta t right. This is this is what is the approach we
make use of in the Euler method that is you find out the slope of the curve at t equal to t0
that is x equal to x0 right and then find out find out actually the delta x that increment in x
when t is changed by time step delta t.
Now this increment is determined by this expression that is dx by dt at x equal to x0 into

delta t. Now here here here this Euler method can be considered to be considered to be
the first order method that is suppose you expand this function if you expand this
function around initial value x0 using Taylor series expansion then the Taylor series
expansion will give you x1 equal to x0 plus delta t into x0 dot plus delta t square by
factorial 2 x0 double dot and so on that is if we take any function and expand at the initial
operation condition using Taylor series expansion then x1 that is the new value at time t
equal to t1 can be obtained as by this series that is here effectively we are truncating this
Taylor series by by considering only the first order derivative okay these higher order
terms we are neglecting and therefore obviously obviously the the the accuracy of
solution will depend upon delta t, if I take delta t very small very small then naturally the
x1 will come out to be very close to the actual value right.
Now here when we use this Eulers method ,we assume we assume that the slope of this
curve is same all through the interval while if you see the actual solution right then during
this interval the slope is different at different points right and therefore therefore there is
some error that will be incurred actually when you solve using the Eulers method. For
example, if you get this point that is x equal to x1 I will call it as x1 corresponding to t1
now if you find want to find out the next step next step that is you want to find x2 using
x1 right.
Now what is going to be our approach x2 will be obtained as x1 plus dx by dt evaluated at

x equal to x1 into delta t okay. The next step will be that you obtain x2 as x1 that is value
at t equal to t1 that is they have becomes the initial value find the derivative at x equal to
x1 okay and multiplied by the time step delta t okay. Now graphically when you see here
let us say that the new slope comes out to be like this right then the the new value of x
that is at x at t2 will be obtained by this point, let us say this point this is your t2 and this is
your x2, this is the point which we get.
Similarly, if you keep on going you will find actually that the the solution which are
getting by this technique is is deviating from the actual solution and this solution is going
to be very close to the actual solution if you take delta t very very small, this is a very
important requirement. Now here this method is called explicit method because the new
value of x at anytime is obtained in using the previous value of x that is that is the when
you are solving this problem, solving this problem right we find out the derivative this
derivative is explicitly computed using the known value of x okay and that is why it is
called explicit integration method here in this method there is a problem that in case you
take delta t large right the error which is created will propagate and the solution may
blow off right and we call this as numerical instability right.
Therefore therefore numerical instability is a problem problem when we use explicit

techniques for solving non-linear differential equations therefore in general you can write
in algorithms using this Euler method as xn plus 1 is equal to xn plus dx by dt, x evaluated
at xn that is you evaluate this derivative at x equal to xn and this is to be multiplied by time
step delta t therefore this is the algorithm a simple algorithm for evaluating the the or for
obtaining the solution of first order differential equation.
Now in case we have instead of scalar equation we have a set of differential equations
right then the same procedure is adopted because these these equations are coupled these
equations are coupled right and therefore when I write the solution this x will be replaced
by vector that is you will write x, I can write down here.
The algorithm will now becomes xn plus 1 equal to xn plus dx by dt evaluated at x equal
to xn into delta t this is the algorithm when you solve a set of differential equations. Now
when I say dx by dt, x is a vector what is the what is the meaning of this now suppose I
say here x is equal to x1 x2 let us say I will use the word say xp, we do not want to mix up
with this xn okay let us say they are p components that dx by dt the meaning will be here
dx by dt this is transpose here. Okay x is a column vector dx by dt can be written as dx1
by dt, dx2 by dt dxp by dt whole transpose it means when I say the derivative of this x
means I am taking the derivative of each of the variable dx1 by dt, dx2 by dt, dxp this is the
way this is to be obtained because when you write down the set of differential equations
right, first order differential equations on left hand side we have the derivative terms delta
1 dot delta 2 dot delta omega 1 dot delta omega 2 dot like that.
Okay for our stability analysis right therefore they are all derivative terms this I
individually evaluated and put in the top now when I say the dx1 by dt is function of all
the state variables and initial condition right. Now in order to overcome the the the
problem of numerical instability this Euler method has been modified modified to reduce
the error which will accumulate and propagate right.
So that we can use a reasonably large value of delta t the time step because if you use
very very small delta t then to get the solution the number of steps will become very large
and it will take more time for getting the solution because at each step you have to do the
similar type of calculations.
In the modified Euler method the approach goes like this the first step is exactly the same
as we use in the Euler method that is we find out the value of x1 we call this is a predicted
value x1p that is the predictor step. The first step this step is called predictor step that is
x1p is called to x naught plus dx by dt evaluated at x equal to x0 into delta t that is you
evaluate x1p you call this as the predicted value initial value x0 plus derivative at x equal
to x0 into delta t.
Okay now, using this predicted value this predicted value x1 is the value at t equal to t1 is
it okay t1 therefore using this predicted value we will find out the derivative of this
function at t equal to t1 okay and using these 2 derivatives on predicted value. We will
find out the corrected value of x1 okay therefore the approach the algorithm will go like
this that x1c is called corrector step x1c is equal to x0 plus 1 by 2 dx by dt at x equal to x0
that is what we have already d1 find the derivative at x equal to x0 then you find out the
derivative of this function right dx by dt at x equal to x1p okay and then take the average
of these 2 derivatives where this is the slope of the solution or to the function at t equal to
t naught or x equal to x0 okay this is the slope at x equal to x1 predicted or we can call it
this is the slope at t equal to t1 okay.
We take the average of these 2 slopes and obtain the a value of x1 and this is going to be
naturally more accurate as compared to the 1 which you obtained by considering only this
term because now we are considering the average of the 2 slopes right. Now here 1 need
not to stop here you can use this corrected value that is x1c right consider this as a
predicted value and further you can find out the more accurate value of x1 that is what
can be d1 is that you can call this as x1 I will call this x1c2, I will just now suppose I use
actually 1 second more the second step then I will call this as this equal to x0 plus 1 by 2.
This initial derivative will remain same that is dx by dt at x equal to x0 plus we will
calculate this dx by dt at x equal to x1c, c1 I will call this as c1 in that case if I go for 1
more step I will call there this is c1 use this x equal to x1c okay and then multiply by delta
t.
Now in case actually these 2 values are very close you can stop here suppose when you
find out the corrected value in the first step then you will find out the corrected value in
second step also if suppose these 2 corrected values come out to be very close or 2
conjugative corrected values come out to be very close you can stop here. Now generally
generally this second step may not be required you can use a small time step and so that
you can stop at the first correction step itself right but if you want to you can say go for
further, you know further you can same a sophisticated value or accurate value you can
use this steps you know instead of terminating at first corrected value you can go for
second, third so and so and since this is can this has to be programmed therefore is
nothing very big actually so far uh computations are concerned.
We can always develop a program we will do it because this type of modified Euler
method is used very commonly because this is very simple a very simple and more
accurate also and the the we can use a slightly larger time step and the problem of
numerical instability also is less probability. Okay the next step which is or the next
method which is commonly used for solving set of non-linear differential equations is
known as the Runge-Kutta method Runge-Kutta method.
Here I will discuss only the 4th order Runge-Kutta method this is more popularly used
because in the literature the second order Runge-Kutta method, 4th order Runge-Kutta
method these are discussed right .Now here as we have seen actually that the Euler
method is a first order method where we are looking in terms of the Taylor series
expansion as if we are we are terminating at first derivative term. Here this 4th order
Runge-Kutta method is as if you are using up to the 4th derivative terms actually you will
not be evaluating the 4th derivative term 4 derivative terms but it it mimix the Taylor
series expansion as if you have taken the the terms up to the 4th derivative.

Now the algorithm for evaluating the uh value of x at n plus 1th step is xn that is the value
of the variable at previous step xn plus 1 by 6 times this term that is x1 k1 plus 2 times k2
plus 2 times k 3 plus k4 I will explain this k terms that is here what is d1 is you obtain the
value of xn plus 1 in terms of the previous value plus this term so this is basically delta x
okay now these terms k1 2 times k2, 2 times k3 and k4 are defined as follows here k1, k1,
k1 is the f of xn tn into delta t that is that is you evaluate this derivative term that is f of xn
tn at t equal to t.
Suppose that we are starting at t equal to t0 then this will become f of x0 t0 if it is going
next step it is becoming f of x1 t1 and so on therefore this is by this is a way and this is
then multiplied by delta t. Okay then we calculate k2 k2 as f of xn plus k1 by 2 xn plus k1
by 2 and this means actually that the value of this term xn plus k1 by 2 is at tn plus delta t
by 2 that is you know in all all the differential equations we come across right the time is
is not explicitly available in the equation that is the time is implicit function right.
Suppose if I solve this the swing equation right in my swing equation on the right hand
side when I evaluated this function it is function of delta and so on the time is not
explicitly visible but it is implicit because delta is corresponding to certain time right
therefore here when I say that this is the value of x right it means it is the at the beginning
of the interval xn plus this quantity can be considered to be at the mid of the interval.
So that I can say interpret this total term as at the mid of the and the time corresponding
to this will be tn plus delta t by 2, okay this again you multiplied this by delta t. So that
this becomes a increment in x, k3 we make use of this k2 and obtain this term derivate that
is f of obtain this function xn same xn remains same plus k2 by 2 tn plus delta t by 2. It
means again the this this increment is computed in the last 1 we obtain the function at xn
plus k3 full not half full that is tn plus delta t okay then what is d1 is to obtain the
increment delta x we we use this 4 terms and we the waiting given to this k1 is 1, k2 we
multiplied by 2, k3 multiplied by 2 and k4 by 1 and then this is added and divided by 6.
So that it becomes a weighted average right, this is the formula you can see here xn plus 1
is equal to xn plus 1 by 6, k1 plus 2 times k3 plus 2 times k, I am sorry 2 times k2 plus 2
times k3 plus k4 by 6 it is basically these are all small small increments in delta x and the
the increment in a small step delta t is obtained by taking the weighted sum and dividing
by 6 that is the weighted average you can call it and here as you can easily see here
actually that we have not computed computed the second order derivative or third order
derivative or 4th order derivative but this process process gives you the solution, similar
to what you will you get if you include up to the 4th order derivatives in the Taylor series
expression now I shall illustrate the application of this techniques by taking an example
after discussing these techniques.
When you physically interpret this 4 constants then k1 can be considered to be slope at
the beginning of the time uh step that is the slope at the beginning of the time step into
delta t then k2 can be interpreted as first approximation to slope at the mid step is a first
approximation, first approximation to this to slope at the mid step into delta t therefore
we are multiplying slope by delta slope with delta t.
So that this becomes a increment in x okay k3 will be interpreted as second

approximation to slope at the mid step into delta t and finally k4 the slope at the end of
the time step delta t right and then increment delta x is equal to 1 by 6 of k1 plus 2 times
k2 plus 2 time k3 plus k4.This is what is the 4th order Runge-Kutta method and simpler
form of this is called second order Runge-Kutta method right where where actually they
use less number of terms. Now I will just discuss something about the numerical stability
of explicit integration methods.
When we discuss this numerical instability the meaning is the meaning is that the the
errors propagate and accumulate and solution will blow off and the although the actual
solution is possible to get the correct solution but by this numerical technique we are
getting a solution which is which is incorrect and that is what is normally called actually
the numerical instability.
Now numerical instability can be avoided by choosing the appropriate value of time step.
Now in any system when you solve any system when you solve right the system has
number of time constants number of time constant. For example, when you consider our
uh stability problem then you will have the time constants which will be very short of the
order of you can say fraction of second right to few seconds.
Suppose actually I take the the excitation system right then if I take the if I consider the
time constants associated with the with the sensing and transducer circuit the voltage
sensing and transducer circuit it is of the order of 0.02 second is very small, if I take the
open circuit time constant of the field circuit it is of the order of 5 seconds, 10 seconds
and so on.
Similarly, if you take actually the times constant the mechanical starting time constant
that is tm one of that is 2H okay that is also a very large because H is of the order of 5 6.
So that tm becomes around 10 and therefore the we define what is called a term steepness
of the system which relates the ratio of largest time constant to the smallest time constant.
Suppose largest time constant is 10 second and smallest time constant is say .02 seconds
then this ratio determines represents a steepness of the system right. The steepness is
defined in terms of another term which is called actually the largest value of the largest
Eigen value of the linear system to the smallest Eigen value that is you take a system find
out the Eigen value and find out the magnitudes largest Eigen value divided by the
smallest Eigen value this will give you the measure of steepness of the system and for the
system to be for the the that a solution to be numerically ah stable right the the time step
should be related to the smallest time constant right and here we discuss one more
important technique where the numerical stability is not a problem, okay and that
technique is known as the implicit integration method.
We will first understand conceptually what do you mean by implicit integration method it
is a very simple approach. Again we start with the differential equation dx by dt equal to f
of xt x equal to x0 and t equal to let us say that this is a scalar equation to start with let us
consider a scalar equation now I want to solve this equation how do I solve this equation
let us say I want to solve this equation the 1 simplest approach would be you integrate
both sides.
Suppose I want to get the solution of this equation dx by dt which is a function of xt

you integrate this equation respect to time and I want to integrate this over a time step say
t0 to t1 dx by dt. This can be written written as integral of f of xt dt to to t1. Now what is
the suppose you integrate and obtain actually this value what is the value of this delta x or
if I call this as x1 minus x0 that is I am integrating dx by dt with respect to t, so that it
becomes a integration of dx only.
So that you can say that at t equal to t0 its value is x0 t equal to t1 its value is x1 what is the
integral of this function what is the value of this integral therefore, we can say that this is
this is area under this curve from t0 t1 okay. Now now I can because this area I cannot it
find I cannot find it out because I not I do not know how this this function is varying and
it is a non-linear function I do not know therefore what we can do is that suppose I take
this situation.
If I plot the function f of xt as if function of time as I plot this function f of xt is a function

of time and let us assume that this is the this is the actual solution okay. Now as you very
clearly told actually that if I am trying to evaluate this integral then it is the area under
this curve from t naught to t1 okay. Now this area under this curve can be approximated
by the area of this trapezoid only error which will cause is small error that is you can
approximate the area under this curve from t0 to t1 the small time step delta t as as f of x0
t0 plus f of x1 t1 multiplied by delta t by 2 is it not this is very important step actually
people sometimes actually fail to appreciate this step which is a very simple step this is
the most important step again you can see here actually that area actual area under this
curve and the area which is approximated using the area obtained by trapezoid right there
is a slight difference and this difference will can be reduced, if I take delta t as small as
possible you reduce this delta t this difference will also vanish okay.
Therefore, I can say now the x1 can be written as from that equation as because here x1
minus x0 let put x0 on this side you will get x1 equal to x0 plus delta t by 2f of x0 t0 plus f
of x1 t1 therefore this is the this is the value of x1 obtained obtained and next at time t
equal to t1 but the problem here is that this x1 is expressed in terms x1 itself. Here, we see
you see this step in this step what is happened is that x1 I have expressed in terms of x1
itself but this equation is now an algebraic equation this is an this is a non-linear algebraic
equation.
Okay and therefore this non-linear algebraic equation can be solved to obtain the value of
x1 using standard equation techniques standard techniques like Newton-Raphson
techniques which are used for solving the non-linear algebraic equations because this is a
the the algorithm can be written as xn will be equal xn plus 1 let us say xn plus 1 will be
equal to xn plus delta t by 2 f of xn tn plus f of xn plus 1, t of n plus 1. This is the algorithm
but you I think you appreciate that you have to solve this algebraic equation, non-linear
algerice algebraic equation using again numerical technique.
This is what actually I have just written here. Now to illustrate actually this trapezoidal
rule of integration, let me take the transient stability problem of a machine connected to
infinite bus we will take a classical problem.
The differential equations which we have to solve can be written as d delta omega dt
equal to 1 over 2H Pm minus Pmax sin theta, second differential equation is d delta by dt
equal to delta omega into omega naught. These are the 2 first order ordinary non-linear
differential equation.
Now these 2 equations are coupled equations that is you will find actually that in this
equation delta term comes right. While in this equation omega term comes delta omega
terms come therefore in general you will find that if you take any problem for a stability
analysis you will get a set of ordinary differential equations ordinary non-linear
differential equations but they are all coupled nonlinear differential equations okay. Now
let us say that how we apply the trapezoidal rule of integration and what equations you
will get okay
Let us assume actually we know that at t equal to tn let at t equal to tn delta equal to delta
n and delta omega equal to delta omega n okay that is at time t equal to tn we know the
value of delta delta n and omega as omega n okay. We write down now the 2 first order 2
non-linear algebraic equations by applying the trapezoidal rule of integration that is I
consider the first equation first equation I am to obtain this term delta omega n plus 1 this
will be equal to delta omega n plus delta t by 2 here I have to write down here f of xn tn f
of xn tn.
Now in this expression in this expression I will write down here 1 by 2H I am not
substituting the values 1 by 2H Pm minus Pmax sin delta delta n correct. You can close this
bracket here plus 1, 1 by 2H Pm remains same Pmax sin delta n plus 1, you close the
bracket here and this is the equation to obtain the value of delta omega n plus 1 in terms
of delta omega n and in terms of delta n plus 1 because now delta n plus 1 is coming here.
The second equation we write down here will be delta n plus 1 equal to delta n plus delta
t by 2 multiplied by f of xn tn. Now in this case it is delta omega n into omega naught plus
delta omega n plus 1 into omega naught. Okay now these are the 2 non-linear couple
algebraic equations, okay and initial values of these quantities are known delta omega n
and delta n initial values are known. Okay and you use the simple simplest technique like
technique to obtain the value of delta omega n plus 1 and delta n plus 1 at t equal to n
plus 1 the numerical stability of these techniques can be illustrated considering a set of
linear differential equations.
Let us examine these equations written by this in the matrix form X dot equal to AX. The
value of X is known say Xn is known at time t equal to tn we want to obtain the Xn plus 1
at time tn plus1. Now this represents the set of linear differential equations and applying
the uh Eulers method we can write down Xn plus 1 equal to Ax evaluated at t equal to tn
plus Xn. Now we can write down this equation in the form Xn plus 1 equal to I plus A
delta t into Xn which can be written as which can be written as F function F into Xn.
Now this represents this represents the discrete form of this differential equation X dot
equal to AX, now for the system to be stable the Eigen values of this matrix A matrix F
not A, matrix F which is equal to AI plus A delta t must be must lie in a in a unit circle
that is the Eigen values, all the Eigen values that is lambda j for j from 1 to n must lie in a
unit circle that is magnitude lambda j must be less than 1.
Now here we can see that the the matrix F is function of the time step delta t as the time
step delta t increases the the Eigen values of this matrix F may lie outside the unit circle
and the and the system response becomes unstable. Now here a small correction when we
write down xn plus 1 equal to Ax t equal to tn this should be multiplied with delta t. Now
with this I conclude my todays presentation by saying that we have discussed the various
numerical techniques which are suitable for solving a set of ordinary non-linear
differential equations. Thank you!
Prof. M. L. Kothari
Lecture - 27
Simulation of Power System Dynamic Response
Friends, so far we have studied the modeling of various sub-systems of the power system.
The sub-systems are synchronous generator, excitation system, turbine governors,
transmission network, loads. These are the different sub-systems and we have studied
how to develop or how to form the dynamic model of these systems and also the
algebraic models.
Now the basic problem is that system is a complete integrated system and therefore we
have to we have to integrate the models of individual sub-systems into a total system
model and then the equations have to be solved right. The techniques which are used for
solving the set of non-linear or any differential equations we have studied say we can use
explicit methods of integration maybe Eulers method, modified Eulers method, Runge-
Kutta method or some few more versions of this explicit techniques or we can use
implicit method of integration.
Now the first step of integrating the sub-system models into a complete system model is
that we have to start looking at the synchronous generator equations. Now if you see the
equations which you have derived for the synchronous generator, these are derived in
terms of dq axis components and and the dq axis locations will be different for for
different machines.
When the system is in dynamic condition right or even under steady state condition the
location of dq axis will be different for different machine depending upon the loading
condition or under dynamic conditions their positions will vary. Now when we solve the
load full equations that is the network equations right we take slate bus as our reference
bus and the voltage of all the buses are obtained with respect to the reference bus right
therefore considering this reference bus as our reference axis.
Now what is to be d1 is that we have to transform transform the the voltages and currents
which were earlier written in terms of dq axis components with respect to the reference
axis and that is what is the first step which is to be done that is called the axis
transformation this is axis transformation is to be done for both voltage and currents
right. Now to understand how this axis transformation is carried out, let us look at this
phasor diagram.
Now in this diagram I have simply shown that let us say this is my d axis and
corresponding q axis is ahead of d axis by 90 degrees. This is our convention which we
have been following that is the q axis is 90 degrees ahead of d axis, if you read some ah
other books or in some books actually they prefer they do take q axis legging the d axis
but our convention is that we have to all through follow that q axis lead the d axis okay.
Now if suppose the terminal voltage of the generator is Et, Et is the terminal voltage of
the generator then the d axis component is the component of Et along d axis that is Ed.
Similarly, the q axis component of Et is eq okay, now now if if the reference axis are
shown like this that is R is the reference axis this is the R reference and in quadrature
with this is the imaginary part of the reference axis, this is reference axis and quadrature
component of this right we call R and I, R and I that is the real and imaginary parts, real
axis, imaginary axis okay is a very standard thing.
Now if you if you take this voltage Et and find out its component along R and I axis then
these components can be shown as ER EI that is Et is same for this its components are ER
and EI, ER is the real part of Et along with the R axis and EI is the imaginary part of Et
along this axis. Okay and now if I represent this Et, Et as ed plus j times eq Et is ed plus j
times eq.
Okay then if I now use this my reference axis as ER plus j times EI right then this ER plus
j times EI can be obtained by by rotating the dq axis components or actually this dq axis
by an angle which is equal to the angle between q and I or angle between d and R that is
you just rotate this phasor through an angle and this angle can be written in this diagram
you can just physically it can be written as delta minus phi by 2 and the angle will be
always measured with respect to the real axis that is our reference axis and the q axis of
the synchronous machine, this is the measure of our rotor angle okay.
Therefore, the rotational rotation is to be given by how much amount that is delta minus
phi by 2. Okay and it is to be given in anticlockwise direction with reference to this
diagram and therefore what I have done is that ER plus j times EI can be written as ed plus
j times eq multiplied by e to the power minus delta minus phi by 2 right. Therefore, this is
actually if you multiply this by this term e to the power minus delta minus phi by 2 here
there is a term missing j it should be e to the power minus j times okay you put it j here e
to the power minus j times delta minus phi by 2 okay therefore this can be written as ed
plus j times eq e to the power j times phi by 2 minus delta phi by 2 minus delta.
Now you write write this term this this term e to the power minus minus disappear is now
only j, I think you can write down here e to the power j is now phi by 2 minus delta okay
this is going to be equal to cosine of phi by 2 minus delta plus j times sin of phi by 2
minus delta okay and if now perform this algebraic operation you will get ER as ed sin
delta plus eq cos delta and EI as eq sin delta minus ed cos delta that is what you are what
we are doing basically is that we are simply performing this algebraic operation.
You separate the real and original parts you will get this curve okay ah which can be
written as ER EI equal to this matrix sin delta cos delta minus cos delta sin delta ed eq
when we know ed eq you can transform ed eq into ER EI okay therefore this is the very
important step which is required to be followed when you integrate the all the equations
with the network equations.

Now here I have tried to show the various sub-systems which are required to be
integrated to form the complete system model. Now we start like this take the we are
considering 1 synchronous machine right. Similarly for other machines also 1 has to look
for we have stator equations and axis transformation this shows here this block diagram
shows that we have stator equations these equations are algebraic because we have
neglected the transformation voltage in the stator equations right therefore these are the
algebraic equations and we are performing the axis transformation you will get this terms
ER, EI right.
Similarly, this axis transformation has to be performed on currents also that is IR and II
that is this is the real part of the current this is the imaginary part of the current okay
this these these are the input to the network that is the transmission network equations
including static loads right that is you write down the equations of the transmission
network shall basically the load flow equations including the equations for the static
loads
Suppose the static loads are represented as constant impedance load and they will can be
included accordingly if we are representing as ZIP or any form that load has to be
included, load model has to be included. Then we have the generator rotor circuit
equations, generator rotor circuit equations. Suppose if I consider consider a situation
where we do not have amortisseurs or neglect the amortisseurs then we have 1
differential equation corresponding to the field winding okay.
If we consider the amortisseurs then we have additional differential equations

corresponding to the number of amortisseurs on d and q axis considered okay. Therefore
these are the differential equations then we have swing equation for each machine swing
equation for each machine is a second order non-linear differential equation therefore
corresponding to the swing equation we have 2 differential equation okay. Then we have
excitation system which is connected to the rotor wire circuit or field circuit of the
generator therefore excitation system equations will come.
Similarly, the prime mover and governor if we include the prime mover and governor in
our model the prime mover governor equations will also come and therefore, if I see here
in this enclosed lines okay then the generator rotor circuit equations, swing equation
excitation system equation, the prime over and governor equations all these equations are
are differential equations and these equations are written written in terms of d-q axis
terms that is when we write the equations for synchronous generator rotor windings okay
right therefore all these equations are in terms of d-q therefore here there is no need of
any transformation.
Now if you see here in addition to this we have to represent the motor loads that is to the
transmission network we have some motor loads which are required to be represented by
their dynamic model then this equations will also come. In case your network has HVDC
link suppose in some power system networks we have few HVDC links right and when
we are developing the complete model of the system, the high voltage DC link right is is
also to be represented by a set of algebraic and differential equations therefore the
mathematical model of HVDC link has to be evaluated.
Similarly, if you have FACTS devices in the system flexible AC transmission system that
is if the FACTS devices are incorporated maybe like say ah STATCOM or TCSC or SVC
right or actually the unified power flow controller then these FACTS devices are also
required to be model and incorporated in the model.
Now this the these models we have not discussed right but adequate the these devices
have to be adequately modeled and required to be incorporated therefore you can see here
actually that so far the common reference frame is concerned this is included in this
purple line if you can just see here that the stator equations and axis transformation they
are with respect to R-I axis right reference axis similarly, the all transmission line
equations are with respect to the common reference frame R-I okay.
Now the equations corresponding to this common R-I reference frame they are all
algebraic equations while the equations which are in corresponding to the generator rotor
circuit swing equation excitation system prime mover they are all differential equations
okay and therefore the necessity is to solve solve the a set of set of differential algebraic
equations for solving the stability problem okay.
Now there are various production grade transient stability programs are available which
have been developed by various you can say research organizations like say EPRI is 1
which has developed a large number of programs similarly, PTI is another organization
which has developed this programs and available, okay but they include all this aspects.
Now we will study study how how we solve the complete set of equations right. Now
when we solve the complete set of equations therefore first we have to see the individual
ah sub-system equations okay.

Now we can write down the individual sub-system differential equations in this form
general form Xd dot is equal to fd is a function of Xd Vd I will explain these terms the set
of differential equations these are all this a vector the Xd dot is equal to a function of Xd
Vd. The algebraic equations will be written as Id equal to function this gd is another
function non-linear function which is function of Xd and Vd again.
Now here Xd now I am considering d the subscript d is used to say that it is

corresponding to a device okay therefore suppose I am writing the differential equation
for a 1 synchronous generator okay for 1 synchronous generator I can write down a set of
differential equations right therefore Xd stands for the state vector of that sub-system, 1
generator which may which may include my all excitation turbine ,swing equation rotor
equations and so on right therefore Xd is a state vector of the sub-system right.
Now Vd, Vd stands for stand for the terminal voltage to which the machine is connected
to which terminal voltage of the machine and this Vd is expressed expressed with respect
to the common reference plane right. The Id Id is the current which is injected to the bus
and expressed in common reference plane right therefore {wic} ((00:20:07 min)) these
are the 2, 2 equations this Xd dot is a represents a set of differential equations and Id equal
to gd Xd Vd represents the set of algebraic non-linear equation or non-linear algebraic
equations right. These are basically that represent our load flow equations this represents
our a set of differential equations okay.
Therefore definition of these terms are Xd is a state vector of individual device Xd is the
state vector of individual device Id is R and I components of current injection from the
device into the network Id is R and I components of current injection from the device into
the network okay, Vd is R and I components of bus voltages and therefore, if I write
down then Id this, this is a current will be written as we can write down this as IdR, IdI that
is the real and imaginary component this forms a this is a small is a vector of 2 is a2
dimensional vector with real part and imaginary part.
Similarly, Vd is VdR and VdI okay now all these differential equations we can assemble
together that is you start with say Xd1, Xd2, Xd3 up to say Xd depending upon all these
dynamic devices all the devices which are representing by a set of differential equations,
you just put together, so that then you can write down this equation in the form X dot
equal to f of xV.
Now this x here is is the total state vector the which represents the state vectors for
individual devices put together right that is Xd1, Xd2, Xd3 they are all put together to form
the complete state vector V, V is a bus voltage vector bus voltage vector right that is if
you have n bus system right then these voltages are V1, V2, Vn and V1 will have 2
components real and imaginary like that set of algebraic equations.
Now we all know actually that the current which is injected at any particular bus current
injection at a particular bus is function of all the state variables and the network voltages
right that is you take any 1 particular bus it will depend upon the state vector of the
system and the voltage vector of the buses that is the bus voltage vector therefore we can
write down this I as x of V it means we take Id1, Id2 up to the last bus and then we form a
vector which is called as I okay, bus current vector and this can be this is a function of the
state variables and voltages and the bus injections can be written in the form of Y Y bus
matrix YN, I am putting this N here this is Y this is the this is the matrix of the network
admittance matrix of the network that is YN into V, YN into V equal to bus current vector
okay and now we are writing this specifically to emphasize that this bus current vector is
function of the state state vector and the bus voltage vector this is what is the meaning of
this terminology. Is it clear?
Therefore the movement we have written the model in this form that is X dot is equal to f
of x, V and I of x V equal to YN V right, this is the complete model representing the
differential algebraic equations okay. Now what is known when we start solving this
problem what is known to us is that at time t equal to 0, t t equal to 0 we know that state
vector x is so much similarly the voltage vector is so much that is we know at time t
equal to 0 the initial conditions so that xo and Vo are known to us. We start solving the
problem and that definition that x is the state vector of the system, V is the bus voltage
vector, I is the current injection vector they are all standard definitions.
Now for solving this differential algebraic equations, we will discuss 2 techniques, 1 is
called partitioned solution with explicit integration, partitioned solution with explicit
integration another is a simultaneously with implicit integration that is when
I say that we use explicit integration method to solve the differential equations right. The
meaning here is that you apply any of the explicit integration technique okay and solve
this differential equations solve this algebraic equations and you do it iteratively that is
you solve this differential equation, you solve the algebraic equation then you can keep
on repeating actually this process till you get the final solution.
Now here suppose the system is operating in the steady state, okay this is a steady state
condition and if say at a particular bus a fault occurs right then at t equal to 0, t equal to 0
that is at the just at the inception of fault, whether this state vector x will change
instantaneously or not just tell me that is at t equal to 0, a fault occurs that is we say that
the at the inception of fault at the instant of occurrence of fault m will the will the state
state vector or the state variables vary instantaneously, the answer is no.
You are correct that so far the state variables are concerned they cannot change
instantaneously now let us look at the network variables can this the bus voltages change
instantaneously therefore whenever whenever actually the fault occurs right, we know
before the occurrence of fault the pre-disturbance condition is xoVo and the movement the
fault occurs right at that particular instant of time the network variables that is the
network voltages will suddenly change from 1 value to the other value while the state
variables will remain at the same value that is I can say xo at x at t equal to 0 minus will
be same as x at t equal to 0 plus right because normally we represent this type of thing
that is the state vector at x at t equal to 0 minus equal to x equal to t equal to 0 plus
this 0 minus stands for just before the occurrence of fault and 0 plus represent the just
after the occurrence of fault, in fact there is a no difference between these 2 time stands
right and these 2 will remain equal so far the solution is concerned and therefore
whenever you start solving this equation the first step will be that the at the movement the
fault has occurred.
You solve this equation and find out find out at next time stand what are the voltages x
will remain same here also when you solve this equation right the x is not going to
change therefore considering the x what is known to you, you find out the next time step
what will be the voltages using this time step we can now solve this equation to find out
the next value for the state vector x at t equal to t1 that is delta t let us say.

Now depending upon what technique we use, now suppose we use actually the I am just
putting the 4th order of Runge-Kutta technique okay. Now in the 4th order Runge-Kutta
technique we calculate this 4 coefficients k1, k2, k3, k4 these constants and then we find
we find the weighted average to find the change in change in state this is what is the
approach therefore when you apply the 4th order Runge-Kutta technique then in first
instance you you you start with this find out the voltages voltages at time, your t equal t1
equal to say delta t when we start from initial condition using this information you solve
this equation and obtain k1 right. Then using this values of k1 which will which has the
information about this x you can again go back to the load flow equations and get the
new value of x and V using this new value of x and V you can find out k2 right. Then
using this new with this value of k2 and so 1 you can again solve these equations.
Therefore you may have to solve in each step right the the algebraic equations a number
of times therefore depending upon what technique we use actually these equations have
to be solved so that we get more accurate results this is what is the further when we use
the word partition the meaning here is that whenever you solve these equations this
quantity is known to you because x dot will be evaluated in terms of known quantities,
explicit method meaning meaning is that the the function f of xV right that is the basically
represent the derivatives will be evaluated in terms of the quantities at the previous step
that is if I want to know the value of x at t equal to tn plus 1 then x, the xn plus 1 is
obtained in terms of xn right that is this is what is the concept of xn and therefore basically
these equations are all decoupled equations. I can write down actually x1 dot can be
computed because everything is known on the right hand side.
Therefore one way of is that you decouple all the equations solve one by one right and
another way is that you you partition these equations ah device wise that is you have the
generator 1, generator 2, generator 3 and solve the equations device wise that is why they
use the word partitioning. Now the the drawback of this explicit integration method is the
possibility of numerical instability if you take a time step which is large enough then we
cannot rule out the possibility of numerical integration and to avoid avoid this numerical
instability we have to use a small time step for solution okay and therefore 1 can use
another technique that is we use simultaneous solution with implicit integration.
Now we know that when we use implicit integration, implicit implicit integration method
for solving these equations right then then the numerically this is a stable technique.
Although, the error will depend upon or accuracy will depend upon the length or or the
length of the time step but numerically it is going to be a stable right that is what we have
already established, let us see how we the simultaneous.
Now in simultaneous the first step will be you apply the trapezoidal rule of integration
apply trapezoidal rule of integration to the set of differential equation that is to this
equation x dot equal to f of XV I apply the trapezoidal rule of integration and let us write
down actually if I apply the trapezoidal rule of integration what will be this.
I can write down xn plus 1 minus xn this will be equal to delta t by 2 f of x n there is
function of V also x of xn Vn plus f of xn plus 1 minus not minus plus ,I am sorry not
minus plus x of n plus 1, V of n plus this is what the solution that is if you take this
equation x of n the x dot when equal to f of XV right then when I apply the trapezoidal
rule of integration then in the time step t equal to tn 2, t equal to tn plus 1 right this is what
is the my formula therefore now I can write down xn plus 1 equal to x n plus this quantity
right.
This is what we have written here as xn plus 1 equal to xn plus delta by delta t by 2 f of xn
Vn plus f of xn plus 1 Vn plus 1 right this is the this is the step or this is the equation
obtained by applying the trapezoidal rule of integration, people have to be very clearly
understand how to apply it. Now here in this on this right hand side right the these are all
non-linear function of xn and Vn okay.
We can write down our algebraic equation our algebraic equation we can write down I, I
as a function of xn plus 1, Vn plus 1 equal to YN V of n plus 1 that is what we are doing is
that corresponding to time tn plus 1 we are writing that current is function of xn plus 1, Vn
plus 1 and voltage is Vn plus 1 right therefore our equation I equal to Y and V are simply
substituting the values computed at time t equal to n plus tn plus 1 okay.
Therefore these we have this is 1 set of equations okay which are written from the
algebraic equation and this is another set of equations the algebraic equation obtained
from the differential equation right. Now what we can do here is in this expression, in this
expression suppose suppose actually the I have obtained a solution let us say I have
obtained a solution if I substitute here x equal to n plus 1 and V equal to Vn and V equal
to Vn plus 1 similarly x equal to xn and x equal to xn plus 1 that is you substitute right it
means this quantity should be equal to this 1 this will satisfy right but in case actually in
case our guess for the next step is not correct. You will find actually that this will not be
equal to this 1 right.
Therefore if I now write now if I now write down that xn plus 1 minus xn minus this term
and denote this by a function okay therefore when I get the solution that function should
be equal to 0 therefore here we are defining a function F of xn plus 1 Vn plus 1 as xn plus
1 minus xn minus delta t by 2, f of xn plus Vn plus f of xn Vn that is this whole term whole
term I defined by a function right.
Therefore we can say that in this function when you write down which is on the right
hand side if you see right xn is known to you, Vn is known to you right therefore the
unknowns are only xn plus 1 and Vn plus 1 similarly, if you see this equation if I take this
{qua} ((00:41:21 min)) this this term minus this term and if this difference I defined by
another function right when I obtain the solution that function should also become 0
right.
Therefore the the another function which you define is written here as G of x n plus 1 Vn
plus 1 as YN Vn minus 1 not Vn plus 1 minus minus I of xn plus 1 Vn plus 1 right therefore
you define these 2 functions, 1 function is obtained from the form the algebraic equations
obtained from by applying the trapezoidal rule of integration, another function obtained
from the network equation okay and at solution when I get the solution this function F, x
of n plus 1, V of n plus 1 right should become 0 G, x of n plus 1, V of n plus 1 should be
0 means these 2 functions should ultimately become 0 when I get the complete solution is
it clear these steps.
Now if you have understood this step then next step you can tell how to how to solve
these algebraic nonlinear algebraic equations just tell me how will you solve these non-
linear algebraic equations. We can apply any technique any technique, iterative technique
which is used for solving a set of non-linear algebraic equation. You can apply you can
apply Newton-Raphson technique or any other technique actually which is applicable
okay therefore we will just write down few equations to solve this a set of nonlinear
algebraic equation using Newton-Raphson techniques if you apply Newton-Raphson
technique, now here when you apply the Newton-Raphson technique our variables which
we want to compute will be x of n plus okay.
Now when you are applying the Newton-Raphson technique this is the iterative technique
and we will represent the {intera} ((00:43:59 min)) iteration count by a ah by a by a
number say k iteration count is denoted by k okay and these iteration counts actually
what we do is that that we start with k th iteration find out the value at k plus 1 th
iteration then using k plus 1 th iteration value go for k plus twelfth iteration and so on
right. Generally if you apply Newton-Raphson technique you will get the solution in 2, 3
iterations right.
So therefore when you apply this what we really do is we write down, we write down this
xn plus 1 at k plus 1 th iteration. Our value we want to find out this x at n plus 1, Vn plus 1
at k plus 1 th iteration will be written as the values corresponding to the previous iteration
that is at k th iteration, x of n plus 1 at k, V of n plus at k plus these changes the change is
called delta xn plus 1 k delta Vn plus 1 k right these are the increments which you will get
in each step okay. Now this is what is the way where we we use any iterative method.
Now these increments can be obtained by applying the Newton-Raphson technique
suppose when you apply the Newton-Raphson technique these increments can be
obtained by solving these equations.
Here, what exactly has given is the we have applied the Taylor series expansion and
considered only the first order terms because in the Newton-Raphson technique we do
consider only the first order terms therefore here we can write down minus F of xn plus 1
k, V of n plus 1 k similarly, minus G of xn plus 1 k, V of n plus k equal to Jacobian, this
is the Jacobian where you find out the partial derivative of F with respect to x partial
derivative of F with respect to V because our vector is now comprising x and V unknown
vector right is comprising x and V.
Therefore we are getting first the partial derivatives with respect to with respect to x and
partial derivative with respect to V that is you will get the terms will be suppose this F
will have number of terms F1, F2, Fn right and x depending upon the number right
therefore this will be a this is a matrix this is also a matrix right therefore similarly delta
G by delta x delta G delta V.
Now this Jacobian the size of the Jacobian will depend upon the number of the state
variables and number of network variables and so far the network variables are concerned
we are considering them as reals that is the real part and imaginary part to form vector V
okay then that this is what we normally do in the Newton-Raphson technique again we
consider the real part and imaginary part and using this you find out this correction vector
that is when you just invert this matrix multiply with this you will get the correction
vector and this correction vector when you add to the previous values you will get the
new values right.
Now in fact if you when when 1 carefully examines the uh Jacobian right then this
Jacobian can be written in the form of AD, BD, CD, YN plus YD that is when you obtain
this Jacobian of the system then you can look actually as it has it comprises of 4 matrices
AD, BD, CD and YN plus D the YN is the network matrix and this AB, AD, BD, CD and YD
are in this form.
This AD will be diagonal matrix where on this diagonal in each diagonal we have on each
diagonal we have the ah block, it is the block diagonal you can call it actually right it is a
matrix Ad1 is a matrix, Ad2 is a matrix, Adn is also a matrix BD is written as Bd1, Bd2, Bdm.
Similarly this YD is written as Yd1, Yd2, Ydm basically this YD when you write down right
this will be corresponding to the ah static load models. Okay and CD equal to Cd1, Cd2 up
to Cdm. This is the way 1 can solve the differential algebraic equations applying the
implicit rule of integration okay.
Now you can understand here understand here that when you are applying this implicit
rule of integration this the the non-linear algebraic equations are solved solved during
each time step once you have completed 1 time step and you go to the next step the same
thing has to be repeated because our time steps are say delta t this starting from to to t1, t1
to t2, t2, tn and each time step right the the non-linear equations which you obtained have
to be solved.
Now only 1 small which I wanted to tell is that since we are solving a transient stability
problem right and during this solution we will come across switching phenomena where a
fault will occur fault will be cleared and so on and therefore the problem is therefore how
to simulate the fault conditions. Now whenever you consider actually a 3 phase fault in
the system then the movement the 3 phase fault occurs the voltage of the bus at which the
3 phase occurs reduces to 0 and therefore to simulate this 0 condition, one way is that you
connect between the faulted bus and the reference bus a large value of conductance
between the faulted bus and the reference bus you connect a large value of conductance.
If you connect a large conductance then the the impedance connected is very small and
therefore the voltage of that bus will become 0 right. Therefore without disturbing our Y
bus matrix or anything the simplest way to incorporate the the fault will be to connect a
large admittance normally it is of the order of 10 to the power 6 right. So that voltage at
that bus becomes 0 that is simulated in case it is a unbalanced fault right we will find out
the fault shunt right depending upon the type of fault and that fault shunt will be
connected between the faulted bus and the network and accordingly accordingly the self
admittance of that particular bus will change that is what is to be changed.
When the fault is cleared cleared and if you have removed 1 particular line to clear the
fault right then the admittance matrix will be appropriately changed by modifying the
mutual admittance and the self admittance terms correspondingly. With this let me
conclude today that we have discussed how to simulate the dynamic response of the
system as a whole.
We have discussed the 2 different approaches, one we call actually the where we use the
explicit explicit integration methods for solving the differential equations or another
technique which we have discussed is where we use the implicit integration method,
where the differential algebraic equations are simultaneously solved. Thank you!
Prof. M. L. Kothari
Lecture - 28
Dynamic Equivalents for Large Scale Systems
Friends, today we will study dynamic equivalents for large scale systems. We know that the
stability studies is one of the most fundamental and important study to be performed for system
expansion planning for design of the system, for operation and control also.
Now to solve actually the stability problem what is to be done is that we have to develop the
model of the system and then solve these equations using numerical techniques. Further, we
require actually large number of stability studies to be performed and since there is no analytical
you can explicit solution when has to resort to the numerical techniques.
Now this numerical techniques take lot of time to obtain the solution. Further the size of the
system has become very large and therefore when the size of the system becomes very large we
will have a mathematical model of very large dimension or very high dimension and therefore
that two problems one is actually that the dimension of the mathematical model of the system
another is the time which is required by using the numerical techniques and in practical
situation, in practical situation it is practically impossible to represent the entire system in detail
and therefore there is a necessity to to resort to some equivalency.
Now here the when you talk about the equivalencing the basic philosophy is something like this
that whenever a disturbance occur in a particular portion of the system then the that portion of
the system which is very close to the disturbance is severely affected, while the system which is
away from the disturbance disturbance is is lesser affected right and the system which is very far
away may have very little affect on the system and therefore one can one can represent the part
of the system which is very close to the location of the disturbance in detail and the system
which is away can be represented by an equivalent.
However however, actually the system which is which is not that severely affected but affected
to some extent right it has an affect on the dynamics of the machines in the study portion right
and therefore there is a necessity to adequately capture the dynamics of the system which is
away from the study system right and therefore what we do here is that whenever the system
large system stability analysis do you perform right we divide the system into study system,
external system and third portion we call is remaining system.
Now to understand what we mean by study system or external system or the remaining system
let us just take this small system it has a number of generators you can see here actually in this
diagram there is number of buses I am not showing the loads but the number of buses the loads
are connected.
Now suppose I am interested in studying the behaviour of this this portion of the system, let us
say this is the portion in which I am interested right it was my interest is to know if there was
there is a fault in this portion of the system right what will be the dynamic behavior of the
system. Therefore, what we do is that we call this portion as the study system the study system.
Now to this study system we have system outside the study system.
Now some portion of this system which is outside the study system has significant influence on
the dynamics of the study system. We will different this by we call it is the external system,
external system then beyond this external system we have this portion, this can be given the
name as remaining system. We will call this as a remaining system, the philosophy here is that
we are interested in the behaviour of the of the system whenever there is a fault occurs in this
portion of the system dynamic behaviour of the system and to know the transient stability
solution of the system then what is to be done is that these external systems should be
represented by an equivalent okay, while representing the external system by the equivalent this
equivalent should be accurate enough. So that we capture adequately and accurately the the
affect of the system onto our study system okay.
Similarly, the remaining system is very far off from the study system and it has very little
influence on the dynamic of the study system then this can be represented by approximate
equivalent, very approximate equivalent or we can say very simplified model we can represent it
okay many times the system can be divided into two parts only study system external system
may not have remaining system.
Now here you can just see these two buses at which the external system is connected to the study
system right of these two buses these buses where study system is connected to the external
system, we denote these buses as boundary buses that is you can say that this bus this bus, these
are boundary buses, boundary buses. Similarly for this external system right the the buses for
example this bus or this bus where where actually the external system is connected to the study
system we call this as a terminal buses these buses will be called at the terminal buses okay.
Now over the years, over the years I can say at least during the last four decades efforts have
been made to develop dynamic equivalents okay. Now when we develop dynamic equivalents
we have to ensure two things one is that the equivalent is accurate enough so that so that we
capture the affect of this equivalent on the study system adequately, secondly that when we
obtain the equivalent the process of obtaining the equivalent should be very fast that is there is a
mathematical way or there is a methodology which is used for obtaining the equivalent for the
system and therefore the time which is required or computations which are required for obtaining
the equivalent of the external system should be should be reasonable okay.
Now many times we may require some more time to obtain the equivalent but since we may
have to perform perform many stability studies right and if we use this equivalent then
whatsoever time you spend in obtaining the equivalent is compensated because actually in
practice that you solve the the stability of the study system using the equivalent right and
equivalent is to be obtained only once once right and therefore when you do the multiple stability
studies that whatsoever time which is required is compensated. In fact the literature on this
dynamic equivalents is very worst and we will confine to only few techniques which are quite,
we will say popular and used in practice.
The techniques are classified under the these 3 categories that is the dynamic equivalencing
techniques we can say or we can classify broadly in these three categories number one dynamic
equivalents based on distribution factors, dynamic equivalents based on distribution factors, the
second technique is based on a model analysis and order reduction model analysis and order
reduction and the third technique is known as the coherency based dynamic equivalencing
technique there are 3 techniques on each technique, you will find lot of literature in the published
okay. Recently, people have developed techniques which are called actually the model coherency
based techniques that is they take the advantage of the the model equivalencing technique and
the coherency based equivalencing techniques right.
The dynamic equivalents which are obtained using the distribution factors right is a simplest of
all the techniques in this arrangement what is done is that this whole external system including
the remaining system is represented by an equivalent generator connected at the boundary buses
that is suppose they are two boundary buses right. At these two boundary buses you connect 2
equivalent generators right and if you have number of loads also then you represent by 2
equivalent loads right.
So that when you perform the stability study you have to perform the study study stability study
of this study system for example in this study system. Now you will have four two generators
which are there already and two equivalent generators right but but this technique is based upon
some heuristic approach it is not having very sound mathematical base right.
Now earlier when the when the when the you know computers are very limited capacities were
available then this was the best approach for obtaining the equivalents but now since we have
better computing facilities available right therefore another techniques we have also have been
considered and they are having a sound mathematical base okay, we will deal with this I will just
mention about this model analysis technique.
In this model analysis technique basically the first one has to obtain for the complete system
linear model, one is to obtain linear model obtain the with this linear model you obtain the Eigen
values of the system okay and then using this information about the Eigen values you eliminate
you eliminate certain modes okay and then retain only those modes which have a significant
influence on the system dynamic behavior okay.
There may be some modes which decay fast you can just ignore them those modes which are
very high you know very low time constant actually and those modes which have high time
constant you retain them okay because we are studying actually the stability if this is a slow
phenomenon okay because you always know actually in 2 in 2 seconds, 3 seconds and wherever
you have actually the time constant of the modes which decay very fast in say fraction of a
second you can ignore the the major drawback of this model equivalencing technique is where
one is to obtain the Eigen values of large system and then then eliminate or reduce actually the
dimension of the system, this is what is now one the very serious problem which comes here is
that this reduced order model you cannot you cannot actually correlate each mode with the
physical system.
You can appreciate this point actually that there is a flatten physical system right and with this
and you have obtained actually this sort of model using this model analysis then these reduced
order model right you cannot have one to one correlation ship between the model and the
physical system and in view of this we shall be discussing mainly this coherency based dynamic
equivalent okay and in this coherency based dynamic equivalents when we study, first let us see
what do we mean by coherency based dynamic equivalents. I think I can first tell you what is the
definition of coherent group of generators.
The definition is like this, a coherent group of generating units for a given perturbation is a group
of generators oscillating with the same angular speed and with terminal voltages in a constant
complex ratio, this definition you have to very clearly understand that is in a power system you
know large number of generators. Okay whenever a particular disturbance occurs.
Okay then then you will find actually that a group of machines swing together right they all
swing together it means where when the when you plot actually the delta versus time out the
swing curves then those machines which have the identical swing curves right form a coherent
group of machines okay and therefore if we can identify for a particular system the coherent
group of machines then we can replace this the coherent group by one equivalent machine right
and this is a very you know sound mathematical technique that is yes there is a system where
whenever you have perturbation the machines which swing together which have the identical
swing curves right they form a coherent group and this coherent group can be replaced by an
equivalent machine that is when you talk about the angular speeds they all have the same angular
speed okay. However, the terminal voltages of these machines right will maintain a constant
complex ratio okay.
Now when we want to obtain the dynamic equivalents using coherency technique then they are 3
important steps which are involved, the first step is identification of group of coherent generators
obviously that before I replace I before by replace the system by equivalent I should know that
okay these machines are coherent, these machines are coherent, these machines are coherent, the
groups have to be obtained, this is a first step and second step is called network reduction.
The terminal buses of each group of the coherent generators are replaced by a single equivalent
bus. You try to appreciate this point that suppose the first step is that we have done some study
and identified that okay these machines form a coherent group, okay then each machines has
connected to a bus okay, therefore what is to be done is that we have to we have to do network
reduction and and obtain an equivalent bus where all the machines are connected okay.
So that once the suppose take an example that if I find that machine number 1,3 and let us say 6
right they form a coherent group then I want to all these 3 machines will have their buses their
different buses at which they are connected okay but we want to replace this by one equivalent
bus where all these machines will be connected in parallel and then these machines will be
replaced by one equivalent, is it clear, this is a second step and the third step is called dynamic
aggregation. The dynamic aggregation means once you have obtained a common bus at which all
the machines which form a coherent group are connected right then we represent a equivalent
machine by what is called a dilute aggregation phenomenon which will have equivalent inertia
constant, equivalent excitation system, equivalent turbine governor model, equivalent transient
reactance and so on it means the the this is a very important step where where you obtain a
equivalent.
Now because in the sense that these are the three important steps which are involved that is you
have to first identify the coherent group of generators and these are the other two steps. Now
here to identify the coherent group of generators we have very important reference that reference
is the paper is written by Robin Padmore, the title of the paper is Identification of Coherent
Generators for dynamic equivalents that emphasis is for identification identification of coherent
generators for dynamic equivalents, this paper is appeared in IEEE transaction on power
apparatus and systems volume PAS 97 number 4 July August 1978. There is another paper that
is G. J. Berg and A. Ghafurian representation of coherency based equivalents in transient
stability studies. This has appeared in electric power system research volume 6 which has in
1983 pages are 235 to 241.
Today, we will concentrate on identification of coherent generators for dynamic equivalents

okay and then in the next lecture we will talk about the other two aspects that is network
reduction and dynamic aggregation and there is one more research paper which deals with
dynamic aggregation that I will tell you in the next term.
Now here when we talk about identification of coherent group of generators. Let us see what is
the simplest technique, how do we how to identify? can you tell me how to identify the coherent
group of generators because you know what we mean by coherent coherent group or what is the
meaning of coherency the the response of the class is that we first obtain the swing curves, swing
curves of all the machines right and those swing curves which are identical we group into one
coherent group and similarly other coherent group, it means we have to first perform a detail
transient stability analysis.
You have to take a system and you perform a detailed transient stability analysis okay and then
you can group them this is one technique this is a straight forward and accurate technique.
However however there are certain facts which have been obtained through lot of research work
the facts are like this and these facts can be used to obtain obtain the approximate swing curves
for identifying the coherent groups of machines instead of going for the detail transient stability
analysis using these facts we can obtain the approximate swing curves right and we can find out
the coherent group of machines and these can be used actually to finally obtain the equivalent the
the assumptions are the facts are like is their assumptions as well as actually the facts.

The coherent group of generators are independent of size of disturbance coherent groups of
generators are independent of size of disturbance this is a very important point here that is at a
particular bus right if a fault occurs whether the disturbance is say .1 second duration or whether
it is of .2 second duration because the size of disturbance will depend upon the duration of the
disturbance to which the system is subjected, for example if you take three phase fault at a
particular bus right if the fault persists for more time then it is a it is a bigger disturbance. If
suppose fault persist for very short time it is a disturbance of small magnitude because if you
suppose somebody performs the study that okay you consider disturbance and state in the
dynamic behaviour considering the different fault clearing time it means you are basically
studying the behaviour of the system system with varying magnitude of disturbance.
Now suppose you consider actually the three phase fault or say unsymmetrical fault right the
unsymmetrical fault is less severe in as compared to three phase fault right and therefore from
that point of view the one very important conclusion here is that it is it is independent of size of
disturbance and therefore coherency can be determined by considering a linearized system that is
you can obtain the coherency behaviour using a linearized system. Okay this is one very
important and on this basis we can get actually the coherency behaviour or coherent groups very
quickly.
The second assumption is or the fact is that the coherent groups are independent of the amount of
details in the generating unit models that is if you look here that is the coherent groups are
independent of the amount of details in the generating unit models that is whether I model a
generator by classical model or I model using the detailed model considering the amortisseurs
field winding, dynamics excitation system, turbine right. The coherency behaviour is same and
therefore what can we do, we can use actually for coherency identification as simplified model
classical model why should I go for all detail model right and therefore therefore the point is that
one can use a simplified model that is therefore a classical synchronous machine model is
considered and excitation system in turbine models are ignored right.
Therefore with these two important assumptions right but these assumptions have been verified
by doing number of studies okay and the the assumptions are valid to a large extent right okay
therefore what we will study here today is that how to identify identify the coherent group of
generators making use of these basic assumptions that is we will represent each machine by
classical model and we will use a linearized model to obtain the swing okay.
Therefore now if I write down the swing equation for a ith machine, for an ith machine
considering small perturbation okay this equation is written as Mi d by dt of delta omega i equal
to Pmi delta Pmi minus delta Pei minus Di delta omega i okay. Now this term is to take care of
damping which is present in the system that is Di delta omega i is a damping term. The second
equation is d by dt of delta delta i that is 2phi f o omega 0 into delta omega i. Now here Mi is
equal to 2 Hi, 2 Hi and the speed deviation that is delta omega i right they are represented in per
unit, okay that is the this is all what you know it.
The next step will be that we write down the the model of the network now here small
perturbation model for the network can be represented like this, I can because here you know we
will have number of buses okay and we will classify the buses into two categories one is internal
bus of the generator internal bus and load buses. Here, we will not resort to any any you can say
elimination of nodes nodes will remain where they are okay therefore you can write down the
equation here delta Pg delta PL delta Qg and delta QL as delta Pg divided delta delta delta Pg
divided by delta theta I will explain delta and theta delta Pg by delta E, delta Pg divided delta V.
Second term will be delta PL by delta delta delta PL by delta theta delta PL by delta E, delta PL
divided by delta V. The third row is delta Qg divided by delta delta delta Qg by delta theta delta
Qg divided delta E and delta Qg divided by delta V, last will be delta QL divided by delta delta
delta QL divided by delta theta, delta QL divided delta E delta QL divided by delta V okay,
multiply by a vector called correction vector that is delta delta, delta theta, delta E and delta V
that is for any network any network right we can write down the equation of this form that this is
basically the Jacobian of the network, this is a Jacobian of the network.
Now here delta because here we will talk about the two types of voltages, voltages these are the
internal voltage of the generator are representing by generator by classical model internal voltage
of generator I will denote this internal voltage by the symbol E right and this this E is a vector
which will have the terms E1 E2 up to En depending upon the number of the generators V is the
voltage of the load buses right delta is the phase angle of internal voltage and theta is the phase
angle of voltage of the load buses right delta Pg, delta Pg is the generator injected power okay
similarly delta PL and delta QL these are the load bus residuals and delta Qg is similarly the
deviation in generator power okay.
Now here in this model model once we represent once we represent the generated by classical
model okay and what about this delta E, if you represent the generator by a classical model then
the internal voltage of the generator is constant magnitude of the internal voltage is constant and
therefore for classical model delta E is is 0 okay, second point what we do here in this case is
that the affect of variation of load bus voltages on the delta Pg, delta Pg and delta PL is negligible
that is the power transfers mainly depend upon actually the angle delta that is we know actually
this decoupling there exist a decoupling right and therefore you can simplify this model and we
can represent the the delta Pg and delta PL in terms of these two quantities that is when we
simplify this when delta E is anyway 0 and this these generations these these quantities hardly
depend upon delta V okay.
Therefore with this assumption or assuming this decoupling right we can I have a simplified
model a simplified model can be written here as delta Pg and delta PL as as delta Pg divided by
delta delta delta Pg divided by delta theta, delta PL divided by delta delta and delta PL divided by
delta theta that is we will use in our study this simplified model. Now here I wanted to
emphasize that this delta PL that is the the the we call actually the load bus residual power right
this delta PL is also 0 in normal case situations. We will use this term delta PL to simulate certain
type of disturbance to simulate certain I will just explain to simulate certain type of disturbance
we will represent this the this term delta PL will come into picture okay.
Now in order to represent this disturbances because we will come across the different type of
disturbances right and we can represent these disturbances in a very simplified manner. Let us
first consider actually the a fault occurs in the system and how do we represent the fault. Again
since, we are trying to do some simplifications let us say that how do we represent a fault and
that is depend the system during fault condition let us say.
Now if you see very very carefully then during fault condition during fault condition since the
duration of fault is small may be 2 cycles, 3 cycles, 4 cycles okay the normal fault duration
which the system can withstand is of the order few cycles okay. Now during this period the
change in delta is small that is when you look on the system as a whole right when the fault
occurs the system is subjected to acceleration okay and the delta the the the rotor angles changes
but during this period of about .1 second is a short duration of within fault condition the variation
in delta is not very large do you agree with this.
If you just try to visualize the criteria of stability right then during the fault condition if you see
the see the change in delta it is there but it is not very large further since delta is very has varied
over as small range or small variation is there therefore the acceleration rotor accelerations at the
insertion of fault and the rotor acceleration at the fault clearing time that is the the accelerating
power at the inception of fault and at the instant of clearing the fault there is hardly any
significant difference, you appreciate this point that is let us say that I am just taking single
machine infinite bus system okay. Let us say when the fault occurs that before the occurrence of
fault the mechanical input say .8 per unit.
The moment the fault occurs the electrical output becomes from .8 to let us say .3 okay therefore
the accelerating power at the inception of the fault is .5 per unit is .8 minus .3 this is the
accelerating power now at the instant of clearing the fault at the instant of clearing the fault right
mechanical input is same let us say electrical output instead of being .3 let us it is .31 because
they are moving on the during the fault power angle characteristic. Therefore the acceleration
power what accelerating power is .8 minus .31 that is say .49.
Therefore, we can make an assumption that okay during this fault the acceleration power remains
at the level of 3.5 and will not create any view big errors small error will be there but still and
therefore this situation can be simulated by increasing the mechanical input by .5 for that
particular period, my mechanical input was initially say .8 I increase in my model mechanical
input by that much amount okay and then the system behaviour of dynamics is not going to be
affected it remains same therefore in this case I have I have not to I have not to consider obtain
the the mathematical model during the fault condition I have to simply increase the mechanical
input by that much amount right this is the one way or simply you can say simulating the fault
condition right.

Now suppose you have a situation where where I am just taking an example, simple example.
Let us say that you have a system a simple system with 2 generators and 3 buses like this. Let us
call this is bus number 1, bus 2 and say bus 3, a simple system. Now suppose I want to consider a
disturbance where this line connecting bus 2 and 3 is tripped the line line connecting bus 2 and 3
is tripped one simple way is that okay, I obtained actually the model model right considering this
post fault or post tripping situation another way of simulating will be simulating will be that okay
you look before the occurrence of fault what was the power which was flowing on this line.
Let us say I am just showing here that after all I am putting opposite air holes. Let us say the
power at this point is entering this bus I call as P2 and power entering at this bus is called P3
because of this line connection, for example if suppose you consider a lossless line right if P2 is
say 50 megawatt P3 will be minus 50 megawatt okay then this this particular situation can be
represented that is at this bus I inject a power equal to P3 I inject a power equal to P2 at this bus
therefore this two, these two this condition and this condition they are same okay. Now I can
simulate I can simulate the tripping of the line 2, 3 by setting P3 equal to 0 and P2 equal to 0
okay.
Now once I say that P three equal to 0 means what is the change in the load power it has changed
from P3 to 0 therefore delta PL will represent actually this this disturbance here that is at this bus
number 3 whatsoever is the change in the power right that I put in my model delta PL delta PL
equal to so much. Similarly, for P2 I represent actually as change in delta PL because this delta PL
which I am trying to say is a vector vector where all the load buses are considered therefore here
in this particular case this two entries right I will make P2 and P3 to represent the line tripping
right therefore when you are trying to take a particular fault condition or line tripping condition
or even you you consider actually the generator tripping.
You have not to change your model the network model remains same and by just you can say
changing the values of delta PL or delta Qg, I am sorry Qa Pg you can change the condition of
the or they simulate the condition of right. Now the question basically is that how do we obtain
the how do you obtain the swing curve considering this type of situation.

Now here easily see that in this model this delta P L and can be is is basically to simulate a
particular type of disturbance okay. Now to solve this problem again start with this these two set
of equations, first is our linearized differential equation, second is the network model okay with
this linearized differential network model what we do is that you apply the trapezoidal rule of
integration, to convert this linear differential equation into an algebraic equation okay. I will
write down this step that is I have the equation of the form which I will integrate both the sides
okay.
I will have this equation Mi integral d omega i by dt dt this is equal to equal to we have term
delta Pmi dt minus integral delta Pei dt minus Di is constant integral integral delta omega i okay.
We will integrate this equation over a time interval delta t okay, over time integral delta t
applying the trapezoidal rule of integration.
Now following the conventional given in the reference paper I will start like this that when you
integrate over the time interval t minus delta t to t1 can take t to t plus delta t but I am following
the same convention which is given in the reference paper okay t minus delta t to t. Let us write
down the value of this integral it will be Mi into delta omega i omega i, t minus delta omega i t
minus delta t okay so far this term is concerned and I integrate this over this time integral I will
get delta omega i t minus delta omega i, t minus delta t.
Now here this term delta Pmi this delta Pmi is the change in mechanical power okay. Now we use
this this term to simulate the type of fault suppose you are considering a fault and you are trying
to simulate the fault by increasing the mechanical power by some amount which will be equal to
the the accelerating power at the instant of occurrence of fault therefore delta P mi will remain
constant will remain constant therefore, what will the integral of this delta Pmi.
Of course it is a constant I am saying as a function of t why I am saying the function of t because
because when the fault is cleared this value has to be changed when the fault is cleared okay then
post fault system is going to be different therefore I am putting this a function of t okay but it
will remain whatsoever is there it is stable on particular time this into delta okay minus what will
be the integral of this by applying the trapezoidal rule of integration delta t by 2 which is the
multiplied by delta t by 2 multiplied by delta Pei Pei t minus delta t t minus delta t plus delta Pei t
okay.
Here you can use this right then one more term is there minus Di delta omega i, t minus delta t
plus delta omega I t multiplied by delta t by 2. Okay that is the first differential equation is
integrated over a time interval t to the minus t minus delta t to t and we get actually in the
equation applying the trapezoidal rule of integration, this is algebraic equation.
Similarly, you take the second equation when we integrate the second equation, we will get delta
delta i t minus delta delta i t minus delta t equal to 2 phi f0 that remains same you will have the
term delta omega i t minus delta t plus delta omega i t into delta t by 2, after this the both these
equations are now converted into algebraic equation.
Now these equations along with the the network equations will have to be solved we will do this
thing in the next turn afterwards let me sum up what we have studied is we have studied the
basic concepts of dynamic equivalents and we have also studied that the dynamic equivalents can
be obtained using a linear model and considering a classical representation of synchronous
generator okay. Thank you!
Prof. M. L. Kothari
Lecture - 29
Dynamic Equivalents for Large scale Systems (Contd)
Friends, today we shall continue continue our study on dynamic equivalents for large
scale systems

In the last time we have studied the basic concepts of dynamic equivalents, we have also
studied the different techniques which are used for obtaining the dynamic equivalents out
of these the 3 techniques which we discussed last time when based upon the distribution
factors, another based on the model analysis and order reduction and the third one was
the based upon the coherency equivalents.
Coherency equivalent is one which is very very simple and it has many merits as
compared to the other techniques. First step in the coherency equivalencing is that one
has to identify identify the group of generators which are coherent and for identifying the
group of generators which are coherent coherent one is straight away approach is that
perform a full transient stability analysis obtain the swing curve and and group those
machines which have identical swing curves in to an equivalent one.
However, however one can one can use a simplified approach to obtain the swing curves
quickly and then identify the coherent group of generators. Now here the two
assumptions on which the symplications, simplifications can be made are number 1that
the coherency behavior of the synchronous machines is independent of independent of
the magnitude of disturbance and second is that is independent of the degree of modeling
whether I represent the system by a simple model or I model the system the details the
coherency behavior does not change.
Therefore from these 2 points actually we can use a simplified approach for for
identifying the coherent group of generators the from the first consideration that the
coherency behavior does not depend upon the magnitude of disturbance. We can make
use of linear model, second is the since it does not upon the details in which the system is
represented we can use a classical representation of the system or classical model right
and then obtain the swing curves of the system.
We shall discuss the simple algorithm which is used for identifying the coherent group
of generators. The linear form of the swing equations is given here Mi d by dt of delta
omega i equal to delta Pmi minus delta pi minus Di delta omega I, this is the one
deferential equation out of the two differential equations which represent swing equation
of the system. Here, delta Pmi is the change in mechanical power of the i th machine delta
Pi is the deviation in mechanical power delta omega i is the deviation in speed of the
machine.
Now here in this equation the speed is is represented in per unit, second equation is d by
dt of delta delta i equal to 2, phi fo delta omega I. Now in this equation Mi is equal to 2
times Hi their network can be modeled model and we can find out actually the the small
changes in the network variables or the system variables using this Taylor series
expansion. Now here in this in this set of equations which I have written here, we have
considering the the generator internal voltages are represented by the symbol E and the
load bus voltage are represented by V, the phase angle of the generator internal voltage is
delta and phase angle of load bus voltage is theta.
Now when we represent this generator by classical model then delta E that is the change
in magnitude of the internal voltage of the generator is 0 and further assuming that the
that the change in power flow due to change load bus voltages is also negligible, we can
simplify this equations and represent in the simple formula like this delta Pe delta Pl
equal to delta P by delta delta delta P delta theta or del P L by del delta L del pl by delta del
theta that is these are all the partial derivatives partial derivatives actually with respect to
delta and theta. You can write down this equation in a compact form as HGG, HGL,
HLG, HLL this is the matrix this is the Jacobian of the system.
Now we will make use of the linear differential equations of the system and the
simplified network equations. Now in this model this change in load power that is delta
PL will be used to simulate a particular type of disturbance okay. Now the algorithm for
identifying the coherent group of generators is to be derived and the first step is that we
we integrate the both sides of the differential equations over the integral over the time
period delta t minus delta t to t, therefore if you integrate both sides of the differential
equation we can write down Mi is a integral of t minus integral d by dt of delta omega i dt
Similarly, integral of delta Pmi dt integral of delta Pi dt minus Di delta omega i dt.

Now the integrals are taken over a time step delta t that is t minus delta t to t. Now we
integrate and perform this integration function using the trapezoidal rule of integration,
when you apply this trapezoidal rule of integration we get equation in this form that is Mi
into delta omega i t minus delta omega i t minus delta t that is this integral right will be
simply integral t minus delta t to t of delta omega i and therefore this will be simply Mi
into delta omega i t minus delta omega i t minus delta t. In this so far the second term is
concerned or this term is concerned delta Pmi is constant this term is there to simulate a
particular type of fault. Therefore, since delta Pmi t is constant the the value of this
integral or may be this term is delta Pmi t into delta t.
Now performing this integration and this integration by trapezoidal rule of integration we
get this expressions. Now in this this this equation is a is now an algebraic equation.
Similarly, when you perform the perform the integrate similarly when you consider this
second equation and integrate this using the trapezoidal rule of integration then we will
get the algebraic equation in this form that is delta delta i t minus delta delta i t minus
delta t equal to 2 phi fo into delta omega i t minus delta t plus delta omega i t the whole
multiplied by delta t by 2.
Now we have now two two algebraic equations, now what what we do here is that we
arrange the equation so that all the terms which are corresponding to time t will be kept
on left hand side and all the terms which are non or corresponding to time t minus delta t
will be kept on left hand side. When you make this arrangement of the terms the first
equation is written in the form 1 plus Di divided by 2 Mi delta t into delta omega i t plus
delta t by 2 Mi delta Pi t equal to Ai t minus delta t.
Now this similarly similarly the second equation is written in this form that is on left
hand side we have the terms which are corresponding to time t that is delta omega i t
delta pei t delta omega i t delta delta delta i t while on this side the terms are known okay
therefore we represent this right hand side term as Ai t minus delta t and Bi t minus delta t
where, Ai t minus delta t is given by this expression which one can derive and satisfies
similarly, Bi t minus delta t is written in this form right.
Now the next step is that we eliminate delta omega i t delta omega i t using these
equations that is you substitute the expression or the value of delta omega i t from this
equation in the first equation.

When you make this substitution we get an equation delta Pi t equal to Ci t minus delta t
minus MTii delta delta i t that is we express the deviation in power output of i th generator
in terms of a term which is non that is Ci t minus delta t this is it quantity. The value of Ci
minus t minus delta t is given by this expression this is the total expression that is C i t
minus delta t is equal to 2 times Mi by delta t Ai t minus delta t plus 2 Mi divided by delta
t square pi fo multiplied by 1 plus Di divided by mi delta t it means all the terms on this
side of the expression are known to us therefore Ci t minus delta t is known similarly,
MTii is a quantity which is which is express in terms of non-quantity.
Now using this expression this expression and the and the network model that is this
model the delta Pe delta PL Jacobian multiplied by this correction vector it is in this form.
Now using this model or this equations along with this equation that is delta Pi t equal to
Ci t minus delta t minus MTii delta delta i t what we do is that we we replace delta Pi t in
terms of Ci t minus delta t and delta delta i t it means they are eliminating delta Pi t. Now
when you perform this operation then you can write down the equation in this form
C t minus delta t equal to HGG prime delta delta t plus HGL delta theta t, where HGG
prime is equal to HGG plus MT. Now here this MT, MT is a diagonal matrix diagonal
matrix with entries MT11, MT22 and MTnn depending upon the number of generators.
Now this equation along with this equation these two equations the whole we can arrange
the two equations in the matrix form you write down C of t minus delta t and delta P L t as
HGG, prime HGL, HLG, HLL delta delta t and delta theta t. Now this is the equation
which can be solved because this term is known because this is expressed in terms of the
quantities which are corresponding to the previous time step and this term simulates the
type of fault and therefore when we solve this equation we get the value of deltas and
thetas.
Now this this matrix this matrix actually is basically a Jacobian right the elements of this
matrix are constant elements of this matrix are constant is a real constant matrix and
therefore for solving this equation one simple approach is that you perform
triangularization of this matrix that is you perform the triangularization operation of this
matrix and once you performance this triangularization operation this is to be perform
only once because this matrix is constant then we can find out the value of the change in
angles right for in each time step that is first I will start with time as t equal to 0 then I
can get the value corresponding to t equal to delta t right and I can iteratively keep on
updating the value of these angles delta and theta the let us now summarize the complete
algorithm which is used for for obtaining the approximate swing curves, I am using the
word approximate because we have made some assumptions.
The algorithm is the first step of the algorithm is the even you initialize initialize the
deltas, omegas and delta Pe 0 that is the initial values of all the angles, speeds and delta
P. Now here the angles speeds and power they are all deviations from the steady state
values therefore initial values of these variables will be 0 is it okay then next step is
increment the time from t minus delta t to t that is the first the third step is set set delta Pm
t delta Pl t according to the disturbance being model.

Now we have we can consider different type of disturbances, disturbance may be a 3
phase fault, may be tripping of a transmission line, may be load sudden change in load at
particular bus or tripping of generator all these disturbances can be modeled model by
appropriately choosing the value of delta Pm t and delta Pl and making some minor
changes in the model. Next step is to compute Ai t minus delta t, Bi t minus delta t and Ci
t minus delta t for all values of I. Suppose you have say n generators, n generators right
then we have to we have how many equations will be there for each generator there are
two equations and therefore I will be paring from 1 to n not 2n, 1 to n because when I
take the first differential equation that is one equation the constant term is Ai d by i 1 to n
then we solve a set of algebraic equations written here that is solve the following matrix
equation for new bus voltage angles delta theta t, delta delta t, delta theta t theta t is the
change in angle of the bus voltage while delta delta t is the change in the angle of the
internal bus voltage or the voltage behind transient reactance.
This operation can be performed easily easily by the process of triangular resistance of
this matrix where we may not have to invert this matrix otherwise this a large has to be
inverted the process of triangularization will avoid the of inversion of this matrix. Now
one small important point is that when you solve these equations the time step which we
have to choose is important right.
Now for solving these equations the time step delta t equal to 0.1 second is considered to
be quite adequate even some times you you can choose a time step of the order of .2
seconds but however the steady so that if you take delta t equal to 0.1 second right the
results will be quite acceptable. However actually the since the we are solving these
equations by using the implicit integration technique and therefore the problem of
numerical instability would not come the the delta t should be kept small to obtain the
accurate results from the consideration of accuracy.

Now once we have obtain once we have obtain the the deviation in angles right we can
compute the deviation in generator power delta P ei t can be computed using this equation
which was derived earlier even compute that expression for change deviation in generator
power output and this is no this is expressed in terms of the known quantities because Ai
ts Ai t minus delta t is known, Bi t minus delta t is known that is the all the terms which
are there on the right hand of this expression are known.
Then in case we are interested in knowing the variation of angular speed as a function of
time then we can obtain the value of delta omega i t in terms of the calculated value delta
delta i t and the known quantity like Bi t minus delta t right that is when you come up to
the step 7 we have calculated calculated the values of this parameters delta delta delta
omega delta Pei at time equal to t right and since we will be solving this equation for a
certain time duration.
Suppose you want to get the solution for two seconds in the moment the time is equal to
the time for which you want to get the solution you stop otherwise you go to step two
where to increment the time step and continue the integration process that is step two is
increment the time form t minus delta t to t and you continue this process till you get the
final solution okay.
Therefore, the algorithm which we have discuss just know is simple algorithm which can
be use to get the approximate swing curves quickly right and then the next step will be
that we have to cluster the swing curves together or you obtain actually the group of
generators which we have identical swing curves.
Now here we can we can cluster the generators considering the the swing curves
corresponding to the bus voltages or considering the the swing curves corresponding
internal internal generator voltages that is I can cluster or I can examine the swing curves
corresponding to bus voltage that is the delta theta versus T for all the buses or I can
cluster cluster the swing curves using the internal bus voltage angles.
Now in practice the first step is that instead of clustering with respect to the internal angle
we prefer to first find out first find out the swing curves corresponding to the bus voltages
right and then the moment you find out that okay these are the generators where the bus
voltages, swing curves are identical okay then you form equivalent group. Now this
approach is helpful in network reduction and once we have done this second step of that
is called network reduction where we put the all the generators on the same bus and then
we check whether these generators form the coherent group or not. Generally it is found
actually that when you when you cluster the cluster or the the the swing curves
corresponding bus voltages then obtain the coherent groups using the using the delta theta
versus T curves okay. Then it is found actually that the machines machines are coherent
with respect to the angle also okay.
Therefore the clustering algorithm is an important step and the algorithm goes like this
what we do here is that suppose you have a group of machines or a set of swing curves
which are to be clustered therefore you start actually with the one let us start with one
swing curve and consider that this is the reference curve, then take the swing curve 2 and
compare the swing curve of the second generator with this reference in case these two
swing curves are identical then the generator 2 belongs to that particular cluster
otherwise, you put the generator in to second cluster then you go to the third generator
you compare with the compare with cluster one and cluster two if it belongs to either of
the two you put it there otherwise you form a third cluster.
The algorithm goes simply like this that delta theta i t minus delta theta r t should be less
than equal to when where is the i is the index, r is the index of the reference reference
swing curve right. In case this this difference is is less than prescribed value over the
complete time interval that is that is for the complete time interval of set to seconds over
with the swing curves are obtain then the then the the generator whose swing curve is
corresponding to theta i qualifies to be put in along with reference generator otherwise,
you form like another cluster.
Now when you perform this step you will be in a position to put the the swing curves in
different clusters and that will give us that which these are the groups of generators which
form one cluster this is second cluster is a third cluster and once we have identified these
groups we have to go for the next step to find out the equivalent machine.
Now at this stage I would like to give some results of some studies which have been
conducted and reported in this important reference that is in this reference which is very
important reference paper written by A.J. Germond and R. Podmore spell is R. Podmore
dynamic aggregation of generating unit models I EEE transactions on power operators
and system volume PAS 97 number four July August 1978 pages 1960 and 1969.
Now this is one reference paper which tells us how to how to aggregate aggregate the the
generator models and obtain an equivalent 1 okay. Now before that naturally one has to
first perform a exercise to identify the coherent group of generators. The earlier reference
paper which I had given which was written by Podmore in that reference paper the two
case studies have been presented in the one case study is done for power system is called
Machigan system. In this system the system considered is having 1027 buses, 1027 buses
and the number of generators in this system were 295, were 295 derivatives.
Now for this system for the system the the algorithm which we just now described was
applied to find out the coherent group of generators for identifying this a 3 phase fault, a
3 phase fault on a bus on a bus was considered that is the bus which is considered in this
typical system is a 345 KV bus okay. The the duration the duration of the fault
considered in this case is 12 cycles there is no fault consider here for this study 12cycles
on 60 hertz bases how much time it comes out to be 12 by 60 seconds which comes out to
be .2, .2 second, its a quite large time.
Now up and the the swing curves approximate swing curves are obtain applying the
algorithm which we have discussed here and the result show that these 295 generators
can be grouped in to total 50 equivalent generators that is when the equivalence is
applied then these 295 generators can be replaced by 50 equivalent generator further out
of these 50 it was seen that 25 groups having only one generator that is when you put the
all the machines in to different clusters what is found was that you can put all the swing
curves of the 295 generators in to total 50 clusters and in in in 24 different clusters was
seen that there is only one generator that belongs while the other generators were
grouped in to number of group together or you can easily say 295 minus 25 that is
minus 24 not 25 then this remaining is 271 they were definitely but put in to only 26
cluster and therefore each cluster must be having large number of generators.
Therefore, consider it that the once you reduce from 295 to 50 then naturally the the
model dimensions is going to reduce significantly. Now further the time which is
required because the result the result which are reported in this research paper so that the
time which is required to obtain the equivalent is 180 seconds just to get the equivalent
and then once you solve the problem transient stability analyses solve using the
equivalent the time required is 55 seconds while if you solve the complete transient
stability problem then the time required is something like 341 seconds.
Now we can see actually that the the time which required to get the solution has
drastically reduced from 341 second to 55 second. Now these figures are corresponding
to a typical computing system which was used right it but relative timings will still
remains, the second case study which has been reported in this research paper the system
belong to this WSCC, is western system coordination committee we call it where the
number of generators considered were or number of buses in the system were 1308 buses
you see this dimension of the problem 1308 buses and it had 397 generators, 337
generators for this particular system again a 3 phase fault was considered, 3 phase fault.
The duration of fault considered in this particular case is 83 cycles on 60 hertz system
instead 12 cycle in the previous case it is 8 cycles and the time, integration time delta t
which is chosen is a again equal to 8 cycles fault duration of the early 8 cycles
integrations stable also up 8 cycles and for clustering for clustering the swing curves, the
tolerance used here is in the previous case it was 5 degrees in the present case for the
WSCC system, CC system it is a 4 degrees and when this exercise was perform right it
could be replaced by 87 equivalent generator but the original system had 337 generators,
while the equivalent system has now got eighty seven generators and out of these 87
generators you 44 generators or groups 44 groups are having only one generator.
Further yes further it was seen actually that these 44 generators which form group of
single generator they were the generators closed to the disturbance, close to the
disturbance and the same thing was observed for Machigan system also in the since
actually that the machines which are very close to the disturbance right may have to be
may retain their independent swing curves while you machines are away they will behave
generally more in a coherent fashion okay.
Now the next step for our study will be that okay we have identify now identify the
coherent group of generators. Now our requirement is that you want to find out the
parameters of the equivalent generator when I say the equivalent generator means there
what is the equivalent inertia constant, what is the equivalent model for the turbine and
governors right what is the equivalent excitation system for this because each machine
has its own excitation system we want to find out what the equivalent of this. Similarly,
each generator synchronous machine right you have to find out equivalent synchronous
generator similarly if it power system stabilizer we want to find out the equivalent power
system stabilizer right.
Now this exercise is quite a quite a complex and involve exercise and we I will just tell
you today that how do we obtain the equivalent equivalent rotor dynamics, first rotor
dynamic let us see it now the equation or the swing equation for each machine written
like this 2 times Hi di a delta omega here also we can put delta omega i right or we can put
omega i is equal to omega 0 plus delta omega i.
So that this equation now the when you just look at the definition of coherence right the
all the machines which are coherent right the the the speeds of the rotors that is delta
omega i the speed deviation will be same for all the rotors right and therefore if I have
something like n machines then I can write these n such equations okay then I add all
these equations.
Now since the speed deviation is same for all the machines right I can write to the
equation of the form summation to Hi for all is multiplied by d delta omega by dt equal
to summation or all i Pmi minus summation Pei minus summation di or i delta omega.
Now I am dropping this subscript I here right because all the machines which are forming
a coherent group they have the same speed deviation and therefore you can see here that
equivalent machine will have inertia constant which is equal to sum of the inertia
constants of all the machines.
Similarly the equivalent damping constant will be equal to the sum of the damping
constants of all the machines and the equivalent mechanical input will be sum of all the
mechanical input that is total mechanical input. Similarly, the electrical output is sum of
all the electrical output therefore this is one thing which we see that when I try to find out
the rotor dynamics of the equivalent machine then that equivalent machine has total
inertia which is the the sum of all the inertias and similarly the damping damping that is
damping term di this the total damping is the sum of damping of individual machines.
Now we will discuss one more important aspect suppose we want to find out the
equivalent equivalent model of the turbine governors that is each machine has its own
turbine and governor right. Suppose you have a group of machines which form a coherent
group I want to find out the equivalent turbine model similarly, the equivalent governor
model.
Now we can write down the the transfer function of of of governor turbine to the let us
say let us say the governor and turbine together, output of the turbine is going to the
change in mechanical power input is speed deviation right.
Therefore, if I represent the transfer function of i th machine turbine governor Gi s this

can be written as change in mechanical power output delta Pmi s divided by delta omega i
s but this is the transfer function, transfer function of so i th turbine governor system
output is mechanical power input is speed deviation.
Now since since all these machines have the same speed deviation, same speed deviation.
So that I can write down equivalent mechanical Pm delta Pm s as summation summation
Gi s sum over all is multiplied by delta omega i s delta omega s or we can drop this I
that is you can say that the total mechanical power the equivalent mechanical power can
be written as as the sum of the transfer functions of the individual generator, transfer
generator I am sorry individual turbine transfer functions that is we can see here
ultimately that I can just write down here summation over i Gi s multiplied by delta
omega s.
Therefore now question arises is that how do I find out the equivalent equivalent transfer
function which will represent the equivalent of all the individual individual governor
turbine transfer functions put together. Obviously you can seen here actually that that the
equivalent transfer function is the sum of the transfer functions of individual turbine
governor model okay. Now here generally what is done is that we come across two type
of machines one is the steam turbine and steam turbine governor another is hydro turbine
and hydro turbine governors.
Suppose in your study you find that in one cluster or in group of coherent machine there
are few machines which are turbo generators, there are few machines with hydro
generators right. In that case when you find out the equivalent we will find out equivalent
of all those turbo turbo generators and put by one equivalent machine we find out the
equivalent for hydro generators and hydro generator governor in put as one machine, we
will not combine the the hydro machines with the steam turbines are driven generators
okay.
Now just to tell you actually that how do we obtain the equivalent model for this hydro
hydro governor hydro governor to this this model which I have considered here that is
governor turbine model it can the speed in to two parts, one is the governor model
another is the turbine model and when these two models are combined together will give
the total model about the hydro governor model can be put in the general form as K times
I can use the word GG s, G is stands for governor equal to K times one plus s times T2
over 1 plus s times T1, 1 plus s times T3 square.
This is the genera form of the hydro governor model and the hydro turbine model if we
put it actually GT s it can be put in the form form 1 plus s Tw, 1minus 0.5 Tw s this is the
model okay. Therefore, what is done is that you put these two together form this is form
equivalent model of the hydro governor and turbine then what we are interested is
actually that what should be this constants K T1, T2, T3 and this Tw right of the equivalent
model. When we know the transfer functions of individual models right the approach will
be the approach will be that what you do is that.
Suppose you want to you can say study our certain frequency range then you you take
discrete frequency range frequencies that is you put s equal to G omega i takes in discrete
values for each value of frequency, you find out the transfer function of individual
individual models and add them that is a suppose I want to find out what is the transfer
function of say model number one it will be G1 s, s equal G omega a substitute that value
of omega i you can some discrete value then this quantity will come out to have actually
its magnitude and phase angle.
Similarly second will also have it magnitude phase angle therefore you can add all these
together to form in equivalent magnitude and phase angle at particular frequency okay
and therefore for all discrete frequencies you can perform this exercise.
Now let me define an error function equal to summation over all is G j omega i minus G
star j omega i magnitude divided by G of j omega i. Now this error function represents
the numerator of this error function is G of j omega i that is this is the transfer function
evaluated for discrete frequency omega i and this is the value of the transfer function of
the equivalent turbine generator the difference will come out to be complex number we
take the magnitude of this divide by divide by the magnitude of this transfer function G
of j omega i for the actual system at the discrete frequency then we do this exercise for
over the complete range of frequencies and sum up this function this is define as error
function.
Now our objective is to obtain the parameters of the equivalent turbine governor model
the equivalent parameters are the gain constant K, the time constants T1, T2, T3 and Tw
and now in order to obtain the the time the parameters of the equivalent generator what
has to be done is that we apply some some technique normally called numerical green
technique.
So that this error function is reduced and when this error function is reduced and comes
to a minimum value at that will give us the the equivalent turbine governor time
constants and density. Now if I plot plot the frequency response of the equivalent
equivalent turbine generator model and the frequency response of the actual system right
then the these two frequency response should be identical right then we can or I can say
that you you plot or it is because of border plot for magnitude and phase angle right and
the border plot for the actual system and for the equivalent system they should match
closely right that means we have obtain the parameters of the hydro governor and turbine
very accurately, this in case now instead of having the hydro governor and hydro turbine
if you have steam governor and steam turbine right. We can we we can choose the
equivalent transfer function of the turbine governor model right and and performing the
same exercise that is we define error function and minimize this error function using
gradient numerical gradient techniques and hence obtain the parameters of the equivalent
governor and turbine okay.
Now let me sum up that today we have discussed an algorithm for identifying a coherent
group of generators then we have also discussed how to aggregate aggregate the the rotor
dynamics and obtain the equivalent inertia constant of the machine equivalent damping
constant of the machine right and we also obtain the the or we also discussed the method
which can be used to obtain the the equivalent turbine governor model of the system.
Thank you!
Prof. M. L. Kothari
Lecture - 30
Dynamic Equivalents for Large Scale System (Contd)
Friends, we shall continue with the study of dynamic equivalents for large scale systems. So far
we have studied the basic concepts of dynamic equivalencing and also we have learnt about the
algorithm which can be used for identifying the coherent group of generators. Now the next step
will be that we have to do network reduction right and then aggregate the models of of the
generators which form which form a coherent group that is when I say that if you have once we
have identified the group of generators which belong to 1 coherent group then these generators
have to be replaced by an equivalent generator okay. Therefore, the first step is that we have to
first do network reduction okay. Now for performing this network reduction let us again look at
the definition of the coherent group.
Our definition shows that a coherent group of generating units for a given perturbation I am
emphasizing this word given perturbation is a group of generators oscillating with the same
angular speed and with terminal voltages in a constant complex ratio, therefore we will make use
of use of the definition of coherency to do network reduction okay.
Now here there are 2 aspects, 1 is that the machines in the coherent group will have same angular
speed and second is that the terminal voltages in a voltages of this machines will be in a constant
complex ratio okay. Therefore, for network reduction we will be using the second, the second
important component of this definition okay what we really do is that you have say group of
machines which are coherent okay and they are connected to their own individual buses therefore
now what we have to do is that we have to reduce the number of buses and retain only 1 bus. So
that all the machines which are coherent are connected to 1 bus this is what is the what is the
main thrust of network reduction.
Now to illustrate this concept what I will do is that we will consider that a system which has say
N generators and N buses okay and let us say that the generator N minus 1 and N they are they
form a coherent group okay. We have just start actually we will take that there are 2 generators
which are numbered as N minus 1 and N they form a coherent group and we will replace replace
the the 2 buses, bus number N minus 1 and bus number N. So that we will reduce the number of
buses okay.
Now in this process what is normally done is that we retain retain 1 bus of this group with the
same voltage because different buses will have different voltages right therefore let us say that
we we retain bus number N minus 1 and the voltage of that bus will be retained as it is right
therefore when we do this simplification or network reduction. Let us see that what will be the
changes in the network model okay.
Now to perform this network reduction we will use this condition that is the voltage of N th bus
and voltage of N minus 1 th bus right the ratio of these 2 voltages is a constant equal to b and this
b is a complex number that the voltages maintain maintain a constant complex ratio okay
therefore, we will make use of this condition. However, we all know that for identifying the
coherent group of generators we have used this first condition first condition that is delta N
minus 1 t minus delta N that is the for all time instants over the complete period right the 2 swing
curves were identical if the this difference is a constant and it is it is less than some tolerance
literature survey shows actually that that the tolerance of 4 degrees or 5 degrees has been
considered therefore now we will use this condition to obtain a reduced network okay.
Now we when you obtain the reduced network the process is very simple we start with the actual
network model. Let us say that this is the network model that is Ii is equal to summation of j
equal to 1 to N yijVj right. This forms our steady state network model, where Ii is the current
injection at bus I, Vj is the voltage of j th bus okay and yij are the elements of by bus matrix okay
this is a very straight forward this has to be no problem. Now in this when you write actually you
will have say N equations where i is varying from 1 to N equations you can write down. Now in
these N equations wherever you have VN we can replace this VN by b into VN minus 1 okay.
Now if you perform this exercise let us see what type of equations we will get we will let us
write down I1, I1 can be written as y11 V1 plus y12V2 y1 N minus 1 we write up to N minus 1 term
first N minus 1 VN minus 1 plus y1 N voltage is now replaced VN is replaced by b times VN minus
1, okay. Therefore this equation now can be written as I1 can be written as y11V1 plus y12V2 last
term will now be equal to y1 N minus 1 plus plus b times b is a complex number into y1 N into VN
minus 1 okay. Therefore, when I replace the voltage of n th bus by b into VN minus 1 I get the
equation for I 1 as y11 V1, y12 V2 and the last term will now be the y1 N minus 1 plus b times y1 N
into VN minus 1.
Therefore what I see here is that now in this in the in the by bus of the network right the the last
column first thing is that this y bus is going to N minus 1 into N now and the last column is
going to be different otherwise y11, y12 these elements up to y1 N minus 2 they will be same
okay.
Now similarly, you can write down for all the buses okay. Now suppose now I write down for
say corresponding to N minus 1 th bus. Let us write down the equation for N minus 1 th bus that
is you can write down here first let us write down the current injection will be IN minus 1 okay
this will be now yN minus 1, V1 plus I am not adding these terms lets us write down this last term
here plus what will be this thing now you will have here yN minus 1 what is the value of now
this 1? N minus 1 plus plus y that is what is plus plus I am just putting the alphabet first plus yN
minus 1 is going to be yN minus 1 N into b and this sum multiplied by VN minus 1 okay. Now
now you write down the last term that is IN IN will be now yN1 V1 that is on the last term here yN
yN, N minus 1 okay plus plus yN, N into b into VN minus 1 okay because VN is replaced it is very
simple. Now the number of equations still are equal to N although although the number of
elements in the row have been reduced from N to N minus 1.
Now what we do is that we we add these 2 equations but before addition what we do is that we
multiply this IN by b star I will tell you why we do this but you multiply this IN by b star and add
these 2 equations and we will call this a very symbol Ie as IN minus 1 plus b star that is b star
stands for complex conjugate of b into IN that is that is on this left hand side when I perform this
addition that is I multiply this last equation by b star add it to IN minus 1 right then I get on left
hand side the current term which is IN minus 1 plus b star IN.
Okay then then you can write down Ie equal to I will write down only 2 terms let us say, Ie will
be equal to yN minus 1 1 plus plus b star yN, 1 this multiplied by V1 okay. The last term it can be
written here as it is a then the term that is you can write down here as yN minus 1, N minus 1 plus
yN minus 1 N into b plus b star into yN N minus 1 plus b star yN, N okay into VN minus 1 okay, b
star yN, N into b yes very correct therefore it is going to be I will put b here b into b star yN, N
okay. Therefore now, what we have obtained is the N equations not I am sorry N minus 1
equations that is number of equations which we were having earlier were N now we have we
have reduced them to N minus 1 equations okay and the in the last equation that is the N minus 1
th equation the current injection Ie, Ie is equal to IN minus 1 plus b star IN okay. Now to
understand why we have done this particular operation multiplied by b star and added to the term
IN minus 1.
Now let us see what is the complex power injection at N minus 1th bus by doing this equivalency
that is the voltage is VN minus 1 if you multiply this by Ie star what does it show, this product
shows the complex power injected at the equivalent bus okay. Now this can be written as VN
minus 1 into into IN minus 1 plus b star IN whole star if you write you expand this term then it
can be written as VN minus 1 IN minus 1 star plus VN minus 1 b IN okay, what does this
expression represents what does this expression represent.
It represents the some of the some of the complex power injected in the original network at N
minus 1 th bus and at N th bus therefore by doing this operation we are retaining the total
complex power injected at N minus 1 th bus equal to the total complex power injected at N
minus 1 and N th bus because this is the requirement actually because when we are trying to to
reduce the network, we have to ensure actually that this condition of power invariance is
maintained okay and that is why we have performed this operation here where, we multiply this
last equation by b star and added to the equation corresponding to N minus 1 th bus, okay.
Therefore with this operation performed we get we get now the reduced network, it has now N
minus 1 buses N minus 1 buses okay and therefore I can write down the equation of the reduced
network or the reduced network model as Ii equal to Yij, Vj, j varying from 1 to N minus 1, j
equal to 1 to N minus 1 okay.
Now this Yij this this bus matrix which is a N minus N, the dimension of this Y bus is now N
minus 1 into to N minus 1 and its elements are modified with respect to the elements of the
original network okay. This is the first step which is called the network reduction.
Now in case suppose you have instead of 2 generators which form a coherent group, suppose
there are there is 1 more generator which forms the coherent group you can perform additional
similar operation and you can keep on reducing the number of buses right and while doing this
network reductions process what is to be done is that we retain the voltage of 1 of the buses at
the original level that is when I have done this reduction process the voltage of voltage of N
minus 1 th bus was retained right but suppose you have instead of 2 machines you have suppose
say 4, 5 machines which form a coherent group then you can select 1 of the machines machine
bus admission retain that bus at its voltage right.
So that now what we can say is that the the we have a network that equivalent machine is
connected to the bus which is retained and the equivalents machine we have to find out what will
be the parameters of that equivalent machine okay. The network reduction part we have seen is a
very simple process it can be done.
Now in order to obtain equivalent model of the machine okay first we will consider the classical
model of the machine that is let us say that we represent the generator by a classical model and
then we will find out actually that what will be the equivalent machine which will represent the
classical model of the system I think this point is already covered. Now to obtain the equivalents
machine we again look at the at the dynamic model of the system.

The dynamic model of the system is written in this form that is for Ith machine d delta I by dt is
equal to omega I minus omega 0 which can be written as if you use in per unit system or the
speed is represented in per unit this can be written as delta omega into omega 0 okay, d omega I
by dt equal to 1 upon Mi Pmi minus P okay where these terms are appropriately you know
dimension d that is dimension of all these terms have to appropriate that we all know it there is
nothing now here is the very important point that if I represent a generator by equivalent.
Okay therefore, we know that the the generator model considering the classical representation is
of this form that terminal voltage Vi is equal to internal voltage Ei minus zii Ii this is what is the
classical model for a synchronous generator that is the terminal voltage is equal to the internal
voltage minus its own impedance that is the impedance behind I am sorry the impedence which
is equal to zii in case you neglect the resistance then this becomes the sub-transient direct axis of
sub-transient reactors okay and the the power injected is the real part of Ei star into Ii this is just
to repeat actually there is nothing new this we all know it okay where our definition for this zii is
Ri plus j times xdi prime Ei is magnitude Ei angle delta I, Pmi is a constant and the magnitude of Ei
is also constant because it is a classical model right we represent the generator by constant
voltage behind direct axis transient reactance.
Now here to obtain the equivalent equivalent let us first write down the 2 equations using the
original voltages for the machines N minus 1 and N that is if I write down for machine number N
minus 1 right I can write down EN minus 1 that is the voltage right. This can be written to be
equal to VN minus 1 plus impedence of this machine right.
Now let me call this as a there are different symbols are used but I will use I to show that it is the
internal impedance Zi N minus 1 into IN minus 1 that is let us look actually that system is not
reduced we have actually the N machines okay and then the the equation of the classical model
looks like this that the internal voltage is equal to the terminal voltage plus this drop. Similarly, I
can write down for N th machine as VN plus Zi, N IN.
Now what is what should be the next step because we have these 2 equations and the and 1
condition which we know is the relationship between VN minus 1 and VN we can use that
relationship so that in this equation what you do is you first let us first replace VN by b times VN
minus 1. So that you can write down this EN equal to b times VN minus 1 this term remains same
okay. Now you divide this equation I will call this equations say for reference say A and I will
denote this equation as B okay, you divide this equation A by Zi N minus 1 all through so that
you can write down now EN minus 1, Zi N minus 1 equal to VN minus 1 divided by Zi N minus 1
plus I of N minus 1 okay.
For the second equation we can write down EN divided Zi N equal to b times VN minus 1 divided
by Zi N plus I okay. We shall number this equation as C okay and number this equation as D
okay the next step will be that you multiply this equation D by b star and add to equation C okay
then what will you get now, when you do this operation you will get EN minus 1 divided by Zi N
minus 1 this remains same to this we are adding a term multiplying by this by b square the
whole equation that is b star not b square b star put EN divided by Zi N okay, on right hand side
we will get the that is this second equation I am multiplying by b star and I am adding it to
equation C therefore this comes out to be equal to VN minus 1 divided by Zi N minus 1 okay plus
b star b VN minus 1 divided by Zi N plus IN minus 1 plus b star b star IN okay. Therefore next
what should we do this is okay IN is multiplied b star this term here see here, the this is our
equation which remains as it is okay to this I am multiplying this the whole equation by b star so
that the first term is okay b star EN by Zi N1 here here I have not multiplied here VN minus this
VN minus 1 upon Zi N minus 1 his remains same here b star into b VN minus 1 upon Zi N right
here IN minus 1 as it is plus this is multiplied by b star into IN what is the problem there is a slight
mistake. You this is to be called as b star okay this is just a slip okay. Now now what we can do
is here we can easily see 1 very interesting thing here that this is the equivalent current that is the
equivalent current that is to be injected at the particular bus VN minus 1.
Therefore, now we can write down this whole thing this whole thing can be written as you, we
write down this first this term VN minus 1 I will take it out okay then coefficient of VN minus 1
we can write down here as 1 upon Zi N minus 1 plus b star b divided by Zi N okay plus Ie plus Ie
equal to equal to this term EN minus 1 divided by Zi N minus 1 plus b star EN Zi N okay.
Now the equivalent generator equation if you write down the generator equation of the
equivalent generator then the equivalent generator equation is written in the form VN minus 1
okay plus let me write down Z equivalent Z is Ze into Ie must be equal to E equivalent that is if I
try to write down the equivalent generator. The equivalent generator internal voltage is Ee
equivalent impedence is Ze and equivalent current is Ie therefore equivalent generator model
when I put in the classical form is VN minus 1 plus Ze into Ie equal Ee okay. Therefore now you
can identify identify from this equation what is this Ze and what will be the Ee, Ze can be simply
obtained because you divide divide this whole equation by this coefficient.
So you write down VN minus 1 is equal to this reciprocal of this plus Ie plus equal to reciprocal
of this term into this whole thing and therefore the Ze can be identified that is the equivalent
impedence Ze, equivalent impedence Ze can be identified as as 1 upon Zi N minus 1 plus b star b
divided by Zi N inverse right and equivalent voltage, internal voltage Ee is identified as 1 upon Zi
N minus 1, N minus 1 plus b star b divided by Z i, N inverse multiplied by EN minus 1 divided by
Zi, N minus 1 plus b star EN divided by ZiN.
Okay that is you can see that all the terms on this side of this equation are known to you because
the generator equivalent generator is obtained obtained by replacing the 2 generators 1
corresponding to the bus N minus 1 another corresponding to bus N and the equivalent generator
what we have done is that we have retain the voltage of N minus 1 th bus intact right what we
have done is that we have obtained, what should be the equivalent impedence and what should
be the equivalent internal impedance.
Now suppose suppose now instead of you can say having 1 machine, 2 machines forming a
coherent group, if you have more than 2 machines then this process can be repeated and once
you repeat this process actually you can get the equivalent generator representing the coherent
group therefore this exercise has to be done.
Now, so far what we have done is that we have considered considered the classical model but
suppose, if I do not want to go for classical model if it is a classical model then in this classical
model actually I am not representing the excitation system I am not representing actually the
synchronous generator dynamics or stator equations in that form right. We are assuming that
the flux linkage flux linkage is constant okay that is what is the classical model but in case you
want to have a more detailed representation right then the procedure is that you replace each
component of the synchronous machine by equivalent that is what is to be done is you replace
the turbine governor models that is each machine has a turbine governor and suppose there are
few machines which form a part of the coherent group then you have to obtain an equivalent
turbine governor model okay.
Now for obtaining the equivalent turbine governor model the the procedure is very simple
because here here the machines which are coherent all the machines have the same same rotor
speed okay. Therefore, if you I look actually the turbine governor model then input is speed
deviation and output is the mechanical power output is it not therefore each each of the turbine
governor model input is delta omega because speed deviation is same for all these machines and
therefore the total mechanical output I can get simply by adding adding all the transfer functions
because total mechanical power will be equal to the transfer function of machine 1 plus machine
2 plus machine 3 multiplied by delta omega right and therefore these transfer functions can be
aggregated to obtain an equivalent model.
Now suppose you consider the rotor dynamics okay. Now since this group of machines are
coherent coherent right therefore all the swing equations you can add together because omega is
same d omega 1 by dt is same as d omega 2 by dt. Therefore, all these swing equation can be
added together and when you add this swing equations what will be the equivalent inertia
constant it will be some of the all the machines inertia constants of all this machines therefore
this is a equivalent inertia constant. Now when I want to replace say the excitation system
models by an equivalent, now here what is the input to the excitation system can you tell me,
suppose I try to write the transfer function model of excitation system what are the inputs.
We have V reference and terminal voltage, we have V reference and terminal voltage now all
these machine have the same terminal voltage because they are connected to the same bus do you
understand this point all the machines have the same terminal voltage right therefore the input to
the excitation system of each of the machines will be same terminal voltage and therefore now
you can obtain an equivalent excitation system right because because the input is same and you
can aggregate all these models. Similarly, if you take the power system stabilizer input to the
power system stabilizer is going to be delta omega speed deviation and since the speed deviation
is same for all the machines therefore you can find out the total stabilizing signal by adding by
adding the transfer functions of all these individual machines and considering the same input as
delta omega.
Now this exercises this exercise where you replace the turbine governors by equivalent turbine
governor excitation system by equivalent excitation system the power system stabilizer by
equivalent power system stabilizer right. For this for this what we have do is that we first
assume the model of the equivalent component and then try to fit the transfer function
characteristic with the transfer function of the total system right.
Now the best reference for for studying how how the individual models can be aggregated
together is given in this reference paper that is by A. J. Germond not this one, yes A. J. Germond
Germond and R. Podmore dynamic aggregation of generating unit generator unit models title is
very straight forward, dynamic aggregation of generating unit models IEEE transaction on power
apparatus and systems volume, PAS- 97, number, 4 July/ August 1978, then pages are 1060 to
1069 that is when dealing with the dynamic equivalencing I started with these 2 references.
The first reference was Robin Podmore as the author identification of coherent generators for
dynamic equivalents therefore this paper addresses how to identify the coherent group of
generators. The second paper written by the Berg and Ghafurian representation of coherency
based equivalents in transient stability studies here he is dealing mainly with the network
reduction and and representation of the equivalent generator by a classical model, while in this
paper in this paper actually which is which is actually considered to be companion paper of this
these 2 papers are put together practically they have appeared in the same year and in the same
issue of the IEEE transaction right they dealt with the 2 aspects 1 aspect is the how to identify
the coherent group of generators and then second aspect is how to aggregate them.
Now just to inform you that this these papers were written out of out of the EPRI project report
that is EPRI has sponsored some research project and at the end when the research project was
completed these 2 papers gave the information about the techniques which they have developed.
Now just to give you 1 slight information that the in these research papers to substantiate
substantiate the algorithm which they have developed they have applied these to 2 large systems,
1 large system is the system of Michigan system another is called WSCC system in Michigan
system the number of buses is 1027 and the number of generators were 295 okay and the for
identifying the coherent group of generators a 3 phase fault at a particular bus was considered of
12 cycle duration in the first case and the equivalent model which they obtained at the end right
had 50 generators.
Original system had 290, 95 generators and the equivalent had only 50 generators and out of this
50, it was seen actually that the 24generators or 24 coherent groups had only 1 generator and
remaining coherent groups were 26 do you understand what is the meaning of a group having
only 1 generator that means that machine has its own independent dynamics, it is not similar to
other machines right and the the studies were done actually that if you perform now the transient
stability analysis using the equivalent the time required was 55 seconds and if you do for the full
complete system the time reported is 341 seconds.
Therefore this is the reduction which 1 gets in in computing the transient stability of the system
using dynamic equivalent similar results are given for WSCC system. Now I want to 1 point out
1 more point here that is a last point that is the coherency behavior depend upon the location of
the disturbance because when I discussed the definition of the coherency we have written
actually a particular location fault or the disturbance at a particular location and generally what is
done is that suppose in a particular system I am interested in knowing the dynamics that becomes
my location for the fault and I find out the equivalent for the whole thing okay.
Now some times you come across a situation where you find machines say hydro-machines and
turbo generators may be identical or may form a coherent group in that case generally what is
done is that machines the hydro-machines are represent by equivalent hydro machine and the
turbo machines or turbo generator represent by equivalent turbo generators they do not we do not
mix up the model of hydro and steam turbine generators.
Okay, with this let me conclude today that we have studied the an an an algorithm for network
reduction and also for obtaining the equivalent generator, first considering the classical approach
and then broadly we have considered that how do we obtained the equivalent model of excitation
system, turbine governors, equivalent power system stabilizer and equivalent rotor dynamics and
similarly, for equivalent synchronous machine. Thank you!
Prof. M L Kothari
Lecture - 31
Direct Method of Transient Stability Analysis
Friends, we shall study today direct method of transient stability analysis.
Over the years efforts have been made actually to assess the transient stability of the system by
applying some direct methods. As you know I that to assess the transient stability of the system
one has to solve a set of differential algebraic equations and for a large system large system the
number of equations are large in number and it is a quite a time consuming process, even if we
resort to dynamic equivalence right the dimension of the model is quite large. Here, we will
concentrate upon one approach which is which is known as the transient energy function
approach transient energy function approach.
In direct methods when we tried to determine the stability one can assess the stability of the
system without explicitly solving the system of differential equations or we can say that we can
partly partly avoid or partly avoid the solution of differential equations we may have to solve the
differential equations for a short interval.
For example, as you know that if you take a single machine infinite bus system then for this we
can apply the equal area criteria of stability to assess the stability of the system. However to find
out the critical clearing time we do have to integrate the equations from initial operating
condition up to the fault clearing time because because we know that if we if you obtain if you
obtain for a given system given system the critical clearing angle then we have to find out
corresponding critical clearing time okay and therefore we may have to integrate this differential
equation from initial operating condition up to the critical clearing time okay.
It means actually we may not have to integrate the differential equation or solve the equations for
the complete time period and in that case we have partly partly solving the differential equation
for the part of the period okay. Now in direct methods in fact they are mostly at the reach level
and and as yet actually the application in practical large system is is not reported because in
practical large system as we will see that when we try to apply the direct methods there are some
problems.
Now to understand the basic concept or basic philosophy of applying the direct methods we
make use of a simple rolling ball analogy. In this rolling ball analogy let us say a physical
system where you have a bowl, this is a bowl this bowl has uneven or non-uniform rim okay here
is a ball a small ball which is lying here in this bowl okay.
Now the the the ball occupies this lowest position in the bowl and this position is called the
stable equilibrium point okay. Now when this bowl is perturbed let us say that we we apply some
force for a short time so that this bowl is perturbed and when you perturb right you are injecting
some kinetic energy to this bowl okay. Therefore, this bowl will roll inside this surface of this
bowl it will roll along some path the path along which it rolls will depend upon the direction in
which the disturbance is applied right.
Now now if suppose the kinetic energy injected is large enough so that the ball comes right up to
the rim right and it crosses over this rim right then we say that the system is unstable because
once it crosses it enters into the unstable region and it cannot return back to the stable operating
condition okay but suppose if the kinetic energy injected is such that such that it climbs a certain
height in the a bowl in the bowl and and returns back right then the system is stable okay
therefore, here if you see here that for the stability of this bowl and stably of stability of this ball
in this bowl right.
There are two important points to be noted, one is the kinetic energy injected one is the initial
kinetic energy which is injected to the ball okay, second is the height of the rim at the crossing
point height of the rim at the crossing point then the location of the crossing point depends on the
direction of the initial motion, the location of the crossing point depends on the direction of the
initial initial motion, it means when you when you give some disturbance right then the ball is
going to move over a certain trajectory inside the bowl okay and depending upon the direction in
which it is made to move right the height is going to be different because I have assumed that the
rim is non-uniform uneven and therefore if I want to find out whether this this ball which is lying
inside the bowl right forms a stable system right I should know what is the kinetic energy which
is injected to the system and what is the height of the rim?
Now as you know till that suppose when the ball is perturbed right as it moves along along then
it acquires potential energy potential energy which is which depends upon the position of the ball
or actually height of the ball with respect to the stable equilibrium point, this is the potential
energy and and when it is moving it looses kinetic energy.
Now suppose suppose when it just comes to the particular point on the rim rim and the speed of
this rim speed of the ball become 0 right it means at this particular point the ball has acquired the
potential energy only it has the total potential energy, it does not have kinetic energy.
Now in case the ball comes up to this particular point and it does not have any kinetic energy
then the forces will act on this ball to bring back to the stable equilibrium point and therefore if I
want to assess the stability I must find out what is the energy or what is the what is the potential
energy with the system, with the system can provide or we can say another way that what is the
capability of the system to observe the energy imparted to the ball and the energy imparted to the
ball depends upon what is the what is the maximum potential energy which it can it can acquire
therefore from the stability point of view, we can consider like this that I can define a energy
function.
I can define a energy function and let us say the energy which this system acquires energy which
the system acquires is denoted by a symbol V. We will denote this energy by a symbol V or V is
a energy function right now let us denote the Xcl as the state of the system at fault clearing when I
say state of the system means for a single machine infinite bus system, the state can be described
by the two two state variables one is delta another in omega speed speed and angular change
therefore this delta and omega these two are the state variables for a single machine infinite bus
system. Now if I know what is the state of the system at at the fault clearing time right then I call
this Xcl as the state of the system at fault clearing time and let us add V which is the function of
the state of the system as Vcl, V of V at Xcl okay.
I will denote this as a transient energy function is called Xcl. Now we find out a critical energy
Vcr denote the critical energy and the the difference that is the critical energy minus minus the
energy at at fault clearing time that is V which is the function of Xcl right this difference is the
transient energy margin a very simple concept right
Now if I again look back at this rolling ball analogy right the critical energy is the energy
correspond corresponding to the potential energy potential energy which this system can provide
right at the crossing point okay and since, since actually the the disturbance can make this ball to
move in different directions right therefore critical energy is different for different directions of
motion right.
Now when I look actually, now a practical power system then we can apply this analogy to a
practical power system also how that suppose the system is operating in the steady state
condition and a fault occurs in the system okay. Now the moment the fault occurs during this
fault on period right. The rotors of the synchronous machines accelerate okay and acquire some
kinetic energy.
Similarly, the angular position of the rotor is also change therefore we can say that the the
machines in the system acquire some kinetic energy, the total kinetic energy plus some potential
energy. This potential energy it depends upon the angular position of this rotor with respect to
the stable operating angular positions okay now if the system is capable of observing the kinetic
energy which you have injected into it right then the system will be stable if the stable is cannot
absorb that kinetic energy completely right then the system becomes unstable this is a very
simple concept.
Now here we will first concentrate upon the application of this transient energy function
approach to a single machine infinite bus system and we will also show that the equal area
criteria which we very much familiar right can also be interpreted in terms of, in terms of energy
functions. Therefore, now our first step is to define a transient energy function how we define the
transient energy function.
Now to define the transient energy function I will start with with a simple system and that is the
single machine infinite bus system that is we will find out the energy function for a single
machine infinite bus okay once you understand the basic concepts, we will extend this concept to
a multi-machine system. The basic swing equation of a single machine infinite bus system I shall
write down in this form M d omega by dt equal to Pm minus Pe. Another equation I am writing
here is d delta by dt equal to omega I am writing in this form which is slightly its it appears
slightly different from what we have been using.
Now here this omega is the deviation of the speed from the synchronous speed therefore instead
of writing delta omega I can write down simply omega for the sake of simplicity that this the
omega is the speed deviation okay. Now here here I am not stating that the speed deviation is in
per unit or speed deviation actually in radians per second okay it will the value of this if you
want to be in per unit then appropriately the M has to be chosen right. Now the energy functions
energy functions are always computed for post fault system that whenever we compute the
energy function or define energy function right the energy functions are always constituted for
post fault system.
Now for the post fault system, let us define that Pe is equal to Pmax3 sin delta here actually the the
three power angle characteristic we will define that P1 equal to Pmax1 sin delta P2 equal to P max2
sin delta and for post fault condition it is Pmax3 sin delta therefore the in this equation Pe is equal
to Pmax sin delta okay.
Now what we do here is that using this equation these two equations this equation and this
equation okay we can write down here dt dt can be written has M d omega divided by P m minus
Pe that is suppose I write down here dt I just consider actually this small pertubation and let us
say that this d omega and dt can be separated right. So I can write dt equal to Md omega by Pm
minus let us write down this denominator in the form of a function I can write down this as Md
omega by a function this is basically a function of delta of delta, for the second equation for the
second equation we can write down dt equal to d delta by omega.

Now if I take this second equation dt equal d delta by omega now we can write down here M d
omega divided by f delta equal to d delta by omega because dt equal to this dt equal to this
therefore they are made equal okay or we can say here M omega d omega equal to f delta d delta
okay. Now you integrate both sides of this equation then we can write down here as M integral
of omega d omega equal to integral of f delta, d delta which can be written as 1 by 2 M omega
square equal to I will just say integral of f delta, d delta. Here here at any instant of time at any
instant of time the the 1 by 2 M omega square is equal to integral of f delta d delta because these
are equal.
Now so far omega is finite this quantity is finite right and therefore this is also going to be a
finite. The energy function is defined we use this definition V omega delta is defined as 1 by 2 M
omega square minus integral of f delta d delta this is what is the basic definition of energy
function here. Now we will identify that this this part of this energy function represents the
kinetic energy therefore we call this as VKE while this portion or this part of the energy function
right it depends upon the angular portion delta is a function of delta only and therefore this is
called, this part of the energy function is called VPE and we denote especially or we put delta
along with it so so that this is this is a potential energy and it depends upon delta.
Now here if you clearly you can say try to write down the expression for V PE delta right then this
VPE delta can be considered to be because in this equation suppose you see VPE delta is basically
basically the negative of integral of f delta d delta right therefore I can write down for this single
machine infinite bus system the value of VPE delta VPE delta can be written as as negative of
integral of f delta d delta okay and what is your f delta f delta we have denoted as Pm minus Pmax3
sin delta therefore now here you can put here is minus Pm minus Pmax3 sin delta d delta, is it
okay? or its value you can write down as you integrate with respect to we are integrating with
respect to delta it can written as minus Pm delta. Now when you integrate this quantity sin delta
will have a minus cos delta that is integral of sin delta is minus cos delta okay and therefore this
continues to be minus minus Pmax3 cos delta okay.
Now here we do one interesting coordinate transformation that is we would like to we would like
to you can say express this potential energy such that for the post fault stable equilibrium point
its value is 0 that is we want to we want to express this VPE delta that is the potential energy in
such a way so that its value is 0 for stable post fault stable equilibrium point.
Now suppose in this expression if I integrate this equation if I integrate and you put the limits let
us say that I will put the limits here from delta s to some angle delta and you may say that is the
integration is performed. So that we do it from some angle delta s to delta okay then we can write
down this as VPE is now function of delta and delta s okay this can be written as minus P m delta
minus delta s minus P max3 cos of delta minus cos of delta s that is this expression, this expression
we have written while in general I have not put any initial and final limits. Here, what we are
doing is that we want to we want to you can say have the coordinate system in such a fashion, so
that this VPE is 0 for post fault stabler stable equilibrium point.
Now for example in this equation if I put delta equal to delta s if I substitute in this equation delta
equal to delta s then this quantity is 0 right, yes if delta equal to delta s then VPE delta delta is
equal to minus Pm minus Pe because there is a sorry, see I putting delta equal to delta s so that
this term is 0 and this term is also 0 cos delta is minus cos delta has become 0 that is instead of
making the the our coordinate system such that the potential energy is measured for delta equal
to delta naught or delta equal to 0 we even like to have to coordinate system in such a fashion. So
that at the post fault stabler stable equilibrium point the potential energy is 0 right.
Further, we will show that show that this the the the total energy function the because when I talk
about the energy function this is also called the transient energy function which is sum of kinetic
energy and potential energy right. For the post fault condition we will show that this quantity is
constant for the post fault system rights the transient energy function is a constant term. Now to
show this thing we can proceed in a simple fashion like this that you you have your basic
equation M times d2 delta by dt2 equal to equal to Pm minus Pe okay.
Now because the the way we have define the way we have defined actually the potential energy
the we can write down here that d negative of negative of VPE d delta is equal to Pm not Pmax Pm
minus Pe because when you you have seen here actually, when we define this VPE you just look
at this equation the definition of VPE is VPE delta is equal to minus integral of V f delta d delta.
Okay therefore, if I take the negative negative and take the d d by d delta of this quantity then it
is equal to f delta okay that is I can say now here that in this equation in this equation we can
write down M d2 delta by dt2 can be written as minus d by d delta of VPE okay.
This is this is by definition we multiply both sides of this equation by d delta by dt that is you
write down this as M times M times d2 delta by dt2 d delta by dt equal to minus d by d delta VPE
into d delta by dt okay. Okay, now this term can be identified you can then take this side or take
on this on this side this can be identified as M by 2 this is identified as d by dt of M by 2d delta
by dt whole square plus plus what will be the quantity here VPE d by dt that delta you can take it
out and therefore d by dt is here it is plus VPE and this quantity is 0 that is you take this equation
okay and you write in this form and therefore if you if you take the derivative this expression you
will get the upper equation right and therefore this term is 0 the meaning of this is that the
expression within the bracket is constant right derivative of this complete expression becoming 0
means this this term this is within the brackets is a constant term or we can say here that 1 by 2
M what is d delta by dt omega omega square plus VPE is constant okay or I can say here that the
energy function which has two components VKE plus VPE equal to the constant.
Now this is very important thing actually physically I want to just explain that when the system
is disturbed okay and let us say that the fault is cleared you are looking for the post fault system,
post fault system energy function is is constituted for a post fault system right. Then in the post
fault system whatsoever the energy which is injected right when the system is in motion what
happens is that the the total energy is a portion between VKE and VPE right.
Therefore suppose a situation comes where VKE become 0 then this total energy will become
potential energy only right but this total remains same now here I want to just tell you one more
point here that is any system any system you will find stable equilibrium point another is the
unstable equilibrium point stable unstable equilibrium points for a single machine infinite bus
system right under the post fault condition post fault condition we can find out the stable
equilibrium point by equating this term Pm minus Pmax3 sin delta equal to 0 right. So that the
stable equilibrium point that is delta s is equal to sin inverse Pm divided by Pmax3 this is the stable
equilibrium point.
Now the we have unstable equilibrium point there will be two unstable equilibrium points which
will satisfy this equation, one unstable equilibrium point I will denote as delta u this will be phi
minus delta s, another will be we will denote as delta u say cap as minus phi minus delta s that is
you substitute in this equation delta u equal to phi minus delta s this equation is going to be
satisfied. Similarly, you substitute delta u hat as minus phi minus delta s then also this equation
is going to be satisfied.
Now these two points are called unstable equilibrium point okay and the critical energy, critical
energy that is VCR which we call right is the potential energy at this unstable equilibrium point
right. Now just to demonstrate how we can plot actually the we plot a graph for potential energy
and show actually that how the the potential energy varies as you vary the angle delta.
Now to plot this graph let us start like this on this we will on this axis we will put delta and here I
am putting VPE is a potential energy as per our as per our definition of the energy function the
energy function the potential energy is 0 at delta equal to delta s. Let us say this point is delta s
okay now and let us say that this point is phi minus delta s that is delta u when I say that there is
another unstable equilibrium point it is on this side that is minus phi minus delta s.
If we now plot the potential energy as a function of delta then this plot will come something like
this I will use a different color. Let me use of this form therefore this is the value of VPE this is
the value for VPE at delta u of this quantity is VPE at delta u at okay this is the value of VPE at
delta u and at delta s which is the stable equilibrium point this quantity is 0.
Now if the total energy if the total energy is equal to some quantity lets say E some quantity say
E and let us say that this is the delta that is if angle at fault clearing that is the fault clearing angle
then at this point this will represent this represents the potential energy we can call it VPE and if
the total energy is say E, I am using the word say total energy E E is the total energy then this
portion will be your VKE at this stable equilibrium point the the total energy will become now
how much 0 VPE is 0 and the the the total, if E is constant therefore at this point it is going to be
only kinetic energy that the stable equilibrium point when the system comes actually it has a
large maximum speed delta omega the speed derivation is large at the stable equilibrium point
that is why it keeps on oscillating right.
Now suppose the total energy which is which is associated in the post fault system which I have
told you that this is a constant and if this is less than this VPE corresponding to delta u my system
is stable, okay conceptually the simple the the subject is very simple here that if we can find out
the critical energy and the critical energy in this particular system is corresponding to is the is the
potential energy corresponding to the unstable equilibrium point and if I know what is the total
energy which is injected into the system right and this total energy remains constant this total
energy has two parts the potential and kinetic and and actually as the system moves right you
will find actually that the the total energy is a portion between the potential energy and kinetic
energy right and therefore condition for stability is that the potential energy at unstable
equilibrium point which is the critical energy is more than the total kinetic total energy injected
or it is the total kinetic energy which is injected which at any other angle is actually a portion
between the two parts.
Therefore, whenever if you want to find out whether the system is stable or not right what is to
be done is that one has to find out what is the critical energy and one has to also find out what is
the total energy injected into the system right and if we can find out actually critical energy
without actually solving the set of differential equations right then to find out the total kinetic
energy injected one can one may have to integrate the equations from T equal to 0 up to the time
T when the fault is clear only for a short time during fault condition you may have to integrate to
find out what is total kinetic energy injected.
Now we will establish here the relationship between the equal area criteria of stability and the
and the transient energy function approach for assessing the transient stability. We will establish
that the equal area criteria of stability is same as the transient energy function approach right this
TE as I have just told you that transient energy function is is function of delta and omega and it
has two components, the kinetic energy and potential energy.
Now to illustrate this that we have we consider the same machine infinite bus system and this is
the diagram which is drawn and it shows the three power angle characteristics the pre fault this is
the post fault this is the faulted one okay. The the peak value of this post fault power angle
characteristic is one is Pmax3 the peak value of during the fault is Pmax2 okay and the mechanical
input is Pm okay this is I think we are all very well familiar with this.
Now, let us say that we are operating at initially at an angle delta naught on the pre fault power
angle characteristic and the fault occurs and which drive the system on this post fault power
angle curve and let us say that fault is cleared at delta equal to delta cl so that now you shift your
operating point from from faulted curve to post fault curve and you move on this. In this diagram
you can see that the unstable equilibrium point is shown here where where the post fault power
angle characteristic intersect with Pm that is this is your unstable equilibrium point its value is
equal to phi minus delta s and delta s the stable equilibrium point is the intersection of Pm with
post fault power post fault power angle characteristic okay.
Now to establish this what we do is we have to look for the 3 areas the area this is area A1 okay,
this is the second area A2, this area A1 is is actually the area which represents the kinetic energy I
will explain this represents the kinetic energy, we will show that it represent kinetic energy this
area A2 right is a part of the potential energy okay and our our stability criteria is that A2 should
be greater than A1 right a simply criteria is A2 should be greater than A1.

Now here we when try to apply the transient energy function approach, we look at this third area
also this area A3 this area A3 is bounded by the post fault power angle curve, mechanical input
line and the angle delta s and delta cl that is this area is A3 okay therefore if we our stability
criteria is that A2 A2 should be greater than A1 this is our equal area criteria of stability.
Now what you do is that you add you add this area A3 to both sides of this equation. So that I can
say A2 plus A3, A3 there is any change because I am adding the area A3 to both side of this
equation therefore A2 plus A3 if it is greater than A1 plus A3 right then my system is stable. Now
this A1 plus A3, A1 plus A3 if you look here this area when you are at this point delta cl right this
area which is A1 it represents the kinetic energy that is 1 by 2 m omega square where omega is
the speed corresponding to delta cl while this A2 represents the potential energy VPE
corresponding to delta cl okay and this this sum must be equal to this total okay, in fact this area
that is A2 plus A3 this represents the the critical energy.
We can establish now by obtaining the or by by deriving the expression for A1, A2 and A3 and
then we will show that A2, A3 represents represents the critical energy that is the potential energy
corresponding to delta u and A1 plus A3 represent the transient energy or transient energy at delta
cl which has 2 components one is the critical energy one is the potential energy right. Now this
can be easily proved, I will just do partly because you can just write the expression for this 3
areas A1.
Now A1 you can write down the let us just write down what is A1, A1 is A1is you can find out A1
by integrating this expression that is Pm minus Pmax2 sin delta from delta naught to delta cl, d
delta okay. Now here what we do is this quantity this quantity is equal to M, this quantity is
basically your M times d omega by dt. During the this is this is what we are doing is we are
trying to find out the area and this represents the the integration over the angle delta cl to or delta
naught to delta cl right that is when the faulted condition and this expression is always equal to
M times d omega by dt okay and you put this as d this is what we are replacing here, this is
replaced by this is d delta.
Now now d delta by this d delta by dt is equal to omega as per our definition so that I can write
down this expression as as M omega d omega, where d delta is replaced by omega into dt and
this quantity is your 1 by 2 M omega square. Now here your are integrating from 0 to omega cl
right because at delta equal to delta naught the speed deviation delta omega cl or omega is 0 right
therefore this quantity can be written as therefore A1 represents the kinetic energy that is you can
say that this is the kinetic energy imparted to the system.
Similarly, you write down A2 and A3 and establish establish that the the A2, A3 is actually the
potential energy corresponding to the unstable equilibrium point delta u right and and the the
transient energy transient energy corresponding to this fault clearing angle delta cl is this kinetic
energy plus the potential energy which is equal to delta which is equal to or which is
proportional to your area A3, if I suggest you to write the expression for A2, A3 and establish that
A1 plus A3 represents the transient energy function right which is corresponding to corresponding
to the clearing angle delta cl and the remaining the expression that is A2 plus A3 represents the
critical energy.
Okay with this let me summarize that today what we have done is that we have studied the basic
concepts of direct method of transient stability analysis, we have devoted our time to transient
energy function approach and the studies are confined to a single machine infinite bus system.
Thank you!
Prof. M. L. Kothari
Lecture - 32
Direct Method of Transient Stability Analysis (Contd)
Friends, today we shall continue our study on direct method for transient stability analysis.
We have already studied how to apply the transient energy function approach for analyzing the
stability of a single machine infinite bus system. Today we will address this transient stability
assessment of a multi-machine system using transient energy function approach. Before I discuss
this let me emphasize here that this transient energy function approach is still in the researching
stage and so far, so far this transient energy function approach for transient stability assessment
has been applied to multi-machine system considering considering the classical model
considering classical model and some efforts have also been made to apply this approach
considering the detailed mathematical model of the system.
When I say that we make use of classical model for transient stability assessment using transient
energy function approach, the synchronous generator is represented by a constant voltage behind
transient reactance and the all the loads are represented by constant impedances. Further further,
because of this type of representation classical representation the the system model is reduced
retaining only the internal generator nodes, okay and therefore whatsoever I discuss here will be
will be considering the system model which is a reduced system model where only the internal
generator buses are retained okay.
Now to start with we we look into the we start with the basic swing equation of the machine. The
basic swing equation of the machine as is very well known to you is two H divided by omega o
d2 delta by dt2 is equal to Pm minus Pe. Now when I consider the ith machine, ith machine then
this equation will take the shape 2 times Hi divided by omega o, second derivative of delta i input
becomes Pmi and electromagnetic power is Pei.

Now here we will represent this by this symbol Mi okay. This electrical power output that is Pei
Pei, we have already derived the expression for Pi but generally for this derivation for this
transient analysis function approach it is put in a slightly different form that we will just establish
that is Pei is equal to a real part of Ei Ii conjugate okay which can be written as a real part of Ei
into sigma j equal to 1 to m yij Ej star okay.
Now here we will denote this yij as as Gij plus j times Bij that is the element of the y bus matrix
we will represent as Gij plus j times Bij. Here, m stand for number of machines right because we
have written only the internal nodes of the synchronous generators with this we can simplify and
write down that real part of Ei okay E, Ej star we put here as Gij yij star we will put as Gij minus j
times Bij and this is summation over this total m equal to j equal to j equal to 1 to m that is j
equal to 1 to m okay this expression well let us simplify we continue this real part for time being.
We will represent Ei as Ei angle delta i okay therefore let us write down this as Ei angle delta i
summation for Ej we will write down Ej Ej star or Ej conjugate is Ej angle minus delta j okay j
equal to 1 to m, this is multiplied by Gij minus j times Bij. This can be now further simplified as
real part of summation in j equal to 1 to m Ei, Ej, Ei, Ej, Gij minus j times Bij here the angle we
had delta i minus delta j okay.
So that now we can write down this as cosine of delta i minus delta j plus j times sin delta i
minus delta j okay this whole thing closed. We will denote here this delta i minus delta j is delta
ij okay and in this operation operation we can further if I denote this term Ei, Ej, Bij as Cij and Ei,
Ej, Gij as Dij okay.
Now when I take the real part the real terms will be corresponding to to Ei, Ej, Bij sin delta i
minus delta j okay plus Ei, Ej, Gij cos delta i minus delta j and therefore if we make these
substitutions. With this substitutions we can write down write down Pei equal to equal to
summation of j equal to 1 to m Cij sin delta ij plus Dij cos delta ij simple therefore the expression
for this real power output Pi has been written in a very compact form as Cij sin delta ij plus Dij
cos delta ij and is summed over j equal to 1 to m.

Now further further even j equal to i right then sin delta ij equal to 0 okay and cos delta ij will be
1 cos delta ij will be 1 and this term Dij this Dij will reduce to this form Dij if I write down here
Dij will become Dii will become Ei square Gii right that is if I take a value of j equal to i that in
this expression the first term vanishes and the second term, second term becomes simply equal to
Ei square Gii right therefore the the standard way of representing or writing the expression for
electrical power output Pei is equal to is equal to Ei square Gii plus summation j equal to 1 to m, j
naught equal to i i Cij sin delta ij plus Dij cos delta ij okay. This is the expression which we will
be using when we write down the the expression for the transient energy function okay that we
should know what we what we mean by this term Cij and this terms Dij.
If you look carefully for the expression for Cij right the Cij is equal to magnitude of Ei magnitude
of Ej into Bij the susceptance part while this Dij again these two voltage term and the conductance
part right therefore this we will we will actually keep in mind that this C ij Cij is having the
voltage terms as well as the susceptance part while Dij has the voltage terms and the conductance
part but since we are using the classical model right therefore these magnitudes are constant and
therefore Cij and Dij are constant.
Now at this stage let me explain one very important concept as when we analyze the transient
stability of the system right what we do is that we we plot swing curves right say for machines 1
to n or machine 1 to m right and these angles are measured with respect to the synchronously
rotating reference frame that is the common reference frame that is the synchronously rotating
reference frame and therefore to assess transient stability what we have to do is that we have we
have to look at the relative variation of the angles delta 1 2, delta 13, delta 14 like that the
another very standard formulation is known as the center of inertia formulation and then center
of inertia formulation we define a mean angle and we will see that all other angles will be
measured with respect to a center of inertia reference.
Therefore, first let us define what we how we define center of inertia which is also sometimes
called center of angle the when we define the center of inertia we have to define two terms, one
is delta COI that is the center of angle this center of angle is defined as 1 over MT summation of
Mi delta i that is if delta i is the angular position of the rotor with respect to synchronously
rotating reference frame Mi is your inertia constant where MT is the sum of all the inertia
constants right.
Therefore we define a term center of angle that is delta COI right and all other all the angles will
be measured with respect to the delta COI for assessing the stability of the system that is the
relative position of the rotors with respect to COI. We define here the center of speed also that is
omega COI is also defined similar to the center of angle as one upon MT the total inertia
constant, summation of Mi omega i, there is are 2 important terms now. Therefore, we will
formulate the problem problem where for assessing the transient stability the actual angles will
be measured with respect to the COI that is center of angle.
Now if you can see conceptually which we visualize then the center of angle is basically a some
sort of mean waited mean of these angles you know the waiting is Mi because waiting parameter
is Mi delta i, this M1 delta 1 plus M2 delta 2 plus like this and all divided by total inertia right.
Now with these 2 terms defined our next step is to write the swing equation in terms of relative
position of the angles or rotator angles with respect to COI, COA right. Now to transform this
equations in that COI reference frame we always use the word COI reference frame what we
define that we define this angle theta I, we define an angle theta i as as the actual angle delta i
minus delta COI.
Okay theta i dot can be written as theta i dot can be written as delta i dot minus delta COI dot
okay. Further delta i dot is omega i dot I am sorry it is only omega i, I will just cut it here we will
write down this as omega I this can be written as omega COI and therefore, this relative speed
this is this now omega is the speed of the rotator with respect to the synchronously rotating
reference frame. Therefore, now when we find out the speed with respect to COI then this we
will denote as omega i bar okay. Our next step is to now write the equation of motion of a ith
machine in COI reference frame right.
Now to start with we can do is something like this that Mi our basic equation start with our swing
equation Mi d2 delta i by dt2 equal equal to Pmi minus minus minus, here we can put here Ei E
square Gii minus summation of summation of j equal to 1 to m, Cij sin delta ij plus Dij cos delta
ij, we can put it this way completely okay. Now with transformation we have just now seen the
transformation we just now seen this delta ij can be written as can be written as theta i minus
theta j, you just check it delta ij which is equal to delta i minus delta j right and with this
definition of theta i is equal to delta i minus delta COI right one can write down this and this will
be denoted by a symbol theta ij that is wherever delta ij occurs we can put theta ij.
Our interest is that this whole equation should be expressed now in terms of in terms of COI
reference frame and these angles are now relative angles are measured with measured with
respect to the COI reference frame and they are denoted as thetas okay. Further further this term
this term we will denote this as Pi instead of writing this thing whole every time we will denote
this as Pi equal to Pmi minus Ei square Gi okay. Then in this equation for delta i delta i you write
down delta Ei can be written as as theta i plus delta COI this is by definition and therefore our
equation our basic equation Mi the swing equation d2 of theta i plus delta COI, Mi d2 of this angle
divided by dt2 right this is written as Pi minus in the previous equation j is not equal to i okay, Pi
minus summation j equal to 1 to m j not equal to i okay in this bracket you have Cij. Now I can
write down here sin theta ij plus Dij cos theta ij okay.
Now you simplify this left hand side I can write down now as Mi d2 theta i by dt2 this is second
derivative of theta i with respect to time plus Mi, Mi d2 delta COI dt2 equal to this whole term
this you can write down, is it okay.
Therefore, this is Mi where the the equation for the complete system can be written as in the COI
reference frame for ith machine it is written as Mi d2 theta i by dt that is second derivative of
theta i with respect to time is equal to Pi minus summation of j equal to 1 to m, j naught equal to i
Cij sin theta ij plus Dij cos theta ij, j minus Mi d2 delta COI dt2 okay, therefore if you look at this
equation in this equation when I have changed over from from synchronously rotating reference
frame to COI reference frame what change we observe here or what change has come in the
equation.
Now this this this left hand side is exactly similar this portion is exactly similar what we have
seen except that delta ij is replaced by theta ij but we have this additional term this additional
term. Now this term this additional term which which have come it can be shown that is this term
Mi d2 dt2 of delta COI this can be this can be written as this can be written as Mi divided by Mt
Mi divided by Mt into PCOI, PCOI this can be established actually and I will just skip over this step
here there is no problem only thing that you resort to the basic definition of delta COI okay and
when you resort to the basic definition of delta COI right you will be in the position to write this
as this term as Mi upon MT PCOI, where where PCOI is PCOI is written as PCOI is written as
summation of Pi, i equal to m minus 2 times summation of i equal to 1 to m and further
summation j equal to i plus 1to m Dij cos theta ij that is this term PCOI which I have just now
written here right is can be established can be established to be equal to summation of P i, i
varying from 1 to m.
Okay this Pi we have already defined as Pi is equal to Pmi minus Ei square Gii okay. Then this
second term which finally comes after simplification is very interesting term this 2 times i equal
to one to m and j take the value i plus 1 to m that is when you expand this expression you start
with say i equal to 1 okay and then you take j equal to i plus 1 to m okay then therefore when
you expand this expression the completely right the PCOI comes out to be like this.
I suggest all of you to derive this and verify that the PCOI comes out to be like this this is a very
important expression. Now with the establishment of with the establishment of expression for
PCOI we can see very clearly here that this PCOI is function of angles thetas okay and other things
are constant okay and therefore now, I can write down the equation of motion or the swing
equation in terms of COI reference frame as Mi d2 theta i by dt2 equal to a function fi of theta, see
in this expression we have written the equation in this form Mi d 2 theta i by dt2 equal to Pi minus
this term okay minus Mi this term okay.
Now when this whole thing is simplified simplified right you will you know we have just now
established this term okay. This term is equal to Mi upon MT PCOI okay therefore on this side on
this side this term is function of deltas only other things are constant I am sorry they are not
deltas but thetas okay and therefore this equation this is the swing equation of i th machine is
written in the form Mi d2 theta i by dt2 equal to fi theta where this theta is now theta is now a
vector theta 1, theta 2, theta m is it okay.
Now when we try to analyze the stability of the system we will come across across model of the
system for the faulted condition and model of the system under after the fault is cleared okay
therefore, if I write down the model of the system during the fault condition as Mi d2 theta i by
dt2 equal to fi F I put this superscript F to denote that it is for faulted condition right. Now this
equation will be valid for time t, time t is greater than 0 and less than equal to tcl, where tcl is the
fault clearing time okay similarly, you can write down another equation for post <a_side> post
fault condition and that is Mi it will be exactly similar Mi d2 theta i by dt2 equal to I will just
write fi theta right where where in this particular case t is greater than tcl is the post fault
condition okay.
Now the problem is that we want to we want to assess the transient stability, we have formulated
formulated actually the model of the system in the COI reference frame okay, this part is clear or
any doubt is a straight straight forward we have defined what is the meaning of COI delta COI
that is center of angle, center of speed we have defined and using these two basic definition and
the swing equation of the machine we have now formulated the model in terms of COI reference
frame okay.
Now our problem is to write down write down the energy function and energy function we
know is written for the post fault system I have emphasized last time also the energy function is
written considering the system model for post fault conditions okay. Now to in order to write
down the model for the energy function or energy function, expression for the energy function
right what is what can be done is is like this actually you can write down the equations of the
system.
Now this can be written now in a slightly more compact form is Mi d omega i bar by dt equal to
fi theta and we can write down here that d theta i by dt equal to omega i bar that is this swing
equation I have written as set of two first order differential equations, using this equation you
know our basic approach for writing the energy function is we integrate both the sides of this
equation okay and we integrate both the sides but what we need the information is what is the
post fault stable equilibrium point okay.
Therefore, here energy function will be expressed expressed in terms of in terms of post fault
stable equilibrium point and any angle theta okay therefore now the question is that I want to
know what is the post fault stable equilibrium point, how do we get the post fault stable
equilibrium point say this model is ready okay this is the model which is applicable to the post
fault operating condition therefore how do I obtain the post fault stable equilibrium point see
when I say stable equilibrium point, when the system has attained the stable operating condition
then naturally this term should be 0 and therefore I can obtain the stable equilibrium point by
solving this equation, you solve this equation fi theta equal to 0 because because what will be this
quantity constant constant, under stable equilibrium point this is the relative speed with respect
to COI COI and therefore relative speed will become 0. Therefore, this omega i bar will become
0 that is how we can say the stable equilibrium point can be obtained by solving this equation
further what is this theta theta is basically in general a vector we can call it say theta 1, theta 2 up
to theta i.
Further one can establish that this term this expression Mi theta i is equal to 0, i equal to 1 to m
one can establish that this summation Mi theta I, i varying from 1 to M is 0. Now to establish this
thing this is a very straight forward you replace this theta i by its basic definition that is this can
be written as this summation i equal 1 to m Mi theta is your definition is theta i delta i minus
delta COI, delta COI okay. This delta COI is does not depend upon i okay the is not a function of
i therefore this can be written as this can be written as summation i equal to 1to m Mi delta i
minus minus Mi delta COI okay, just I have expanded it.
Further this portion this is by definition by definition what is this quantity, MT delta COI is it not
minus delta COI summation i equal to 1 to m, Mi the summation of Mi is MT and therefore this
quantity comes out be 0 therefore we have established that this identity okay.
Now what is the advantage of establishing this identity what is the advantage which we get by
establishing this identity? that because of this equation I can express theta m interms of theta 1 to
theta m minus 1 right and therefore instead of having theta 1 to theta m as my independent angles
right therefore when I solved for obtaining the stable equilibrium point right I have to solve only
for theta 1 to theta m minus 1 both are independent, this identity is there right therefore we will
solve this equation and obtain the stable equilibrium point once you obtain the stable equilibrium
point we we perform the integration of our swing equation swing equation and define the energy
function.
Now let us write down the now to write down this the simplest approach is that you write down
the equations of each machine right. Suppose I write down the equation for machine 1, now my
equation for machine 1 is Mi d omega i by dt equal to fi theta okay.
Now I can write down this number of such equations right for each machine where, I is going to
vary from 1 to m. So that you can we write down the another equation is your another equation is
dt2 d theta d theta i by dt is equal to omega bar. Now from these two equations, from these 2
equations we can write down dt equal to dt equal to Mi d omega i bar fi theta equal to equal to d
theta i divided by omega i bar okay and when I take this equation now, okay then when you
integrate this equation you will find that you will get 1 by two Mi Mi omega i bar square equal to
integral of fi theta d theta i okay.
Now we will integrate it with respect to with respect to the stable equilibrium point that is our
base starting point will be the stable equilibrium point therefore so far actually the for the stable
equilibrium point is concerned right the omega i bar is 0 therefore for any other omega i right
this is the integration that is starting from 0 to omega i, now here we will start from theta i s to
some angle theta i and my energy function.
Now my energy function now becomes I can say that energy function for ith machine right Vi
which is now the function of function of omega i bar theta i is equal to 1by 2 Mi omega i bar
square minus integral theta i s to some angle theta i fi theta d theta i. This is the energy function
this is by definition by definition of this energy function we integrate both the sides right and
then you the this quantity is actually the integration of the left hand side term, this is the
integration of right hand side term okay and you transfer the right hand side term to the left hand
side it will be the negative sign here then this whole term is called the energy function of ith
machine right and therefore now you can write down the energy function for the complete
system by by adding all the such equations that is we write down the energy function machine 1,
2, 3, 4 like this and when you add all these equations you will get the energy function for the
complete system.
When you perform this exercise where you add all the energy functions energy functions, for all
the machines and and explicitly write down the expression for theta i fi theta i, fi theta it was
defined okay therefore now you write down the expression for fi theta in the complete form right.
Then the total energy function will be written in this form, this is the expression which I have
written here this is the most important expression that is V this is the energy function V theta
omega bar as equal to 1 by 2 i equal to 1 to m Mi omega i bar square okay because you can see
that for one machine it is 1 by 2Mi omega i bar square therefore for all the machines it is
summation of such terms from i varying from 1 to m then since fi theta is now is now replaced
by it is original expression and we known that fi theta bus function of Pi therefore when you
considering this Pi constant and when it is integrated you will get theta i minus theta is you again
add from i equal to 1to m then you had these terms Cij cos theta ij and Dij cos theta ij okay and
then you also had the term PCOI.
Therefore when you make this total summation the expression comes out to be in a very compact
form as this is the first term, the second, third term and fourth term, there are four terms here.
Therefore in fact here this first term is your kinetic energy we call it VKE and all these other
terms second third and fourth put together forms the potential energy. Okay now this the each
term can be explained, the first represent the kinetic energy is stored in the rotators.
Okay the second term is the potential energy due to position of the rotators, this third term third
term represents the magnetic energy stored in the elements of the system because I have told you
that Cij has a term Bij that is the susceptance term right. Then this can be interpreted as the
magnetic energy stored in the elements of the system. The last term represents the energy
dissipated in the elements of the system because Dij has the term Gij conductance part right
therefore the potential energy is concerned it has the 3 components which represents the the
energy due to the position of the rotators minus these 2 components right this is called.
Now once this expression is obtained, now here I wanted to emphasize that for this derivation
one has to one has to do certain algebraic simplifications that is you write down the sum up the
total sum of the energy function of individual machines right and then you have to simplify this
to arrive at this function okay. Therefore, now let me just tell you that if you want to find out the
find out the critical energy because we require actually the expression for the critical energy Vcr.
Now for finding out the critical energy Vcr what we need is the unstable equilibrium point
because as you seen in the single machine infinite bus system right, the critical energy right is
depends upon the unstable equilibrium point and since this system has number of machines right
and therefore there will it will have number of unstable equilibrium points right and therefore the
definitely the problem basically is to find out actually the critical energy, critical energy which is
actually a realistic one for a particular system for defining the transient stability of the system
right.
Therefore the the exercise will be to find out an unstable equilibrium point and then find out the
critical energy for the system. Then second is that you have to find out the energy energy
corresponding to the fault clearing time that is you find out the energy energy we call it say V cl
that is the energy corresponding to fault clearing time. You find out the energy there is a critical
clearing I am sorry the critical energy okay.
Then then in case Vcr stands for critical energy and Vcl stands for the energy of the system total
energy kinetic energy plus potential energy right at the fault clearing time then this difference
shows the a stability margin this is our stability margin, okay.
Now generally to find out the critical clearing time what is done is that we integrate the
equations we integrate the equations of the system till these two energies are equal that is you
integrate the equations and calculate this Vcl, Vcr is to be calculated separately the moment these
2 are equal we find out the time which is t l or this tcl will now become equal to tcr the critical
clearing time okay.
Now at this stage, let me tell you that the work of the problem is to find out the critical energy
and lot of research work has been done, how to evaluate the critical energy and this critical
energy depends upon the unstable equilibrium point and in any system we will find that for this
multi-machine system right you can find out there will be number of unstable equilibrium points
right and I will suggest you to read chapter 9 of this book power system dynamics and stability
by Peter. W. Sauer and Professor. M. A. Pai, chapter 9 of this book deals with the transient
energy function approach for transient stability analysis.
There large number of references are given at the end of this book which deal with the with the
problems for associated with computing the critical clearance critical energy computing the
critical energy and for assessing the transient stability. Those who are interested they should read
the reference papers to have more knowledge about the transient anality function approach for
stability assessment. Thank you very much!
Prof. M. L. Kothari
Lecture - 33
Subsynchronous oscillations
Friends, we shall study today sub synchronous oscillations in the power system. In the
analysis of power system dynamics so far we have considered the rotor of the turbine
generator system as a single mass system such a representation accounts for the system
mode of oscillations, the frequency of this mode of oscillation is in the range .2 to 2 hertz
in the system, in reality the the turbine generator rotor is a very complex in the structure
mechanically it has several masses mounted on the same shaft. It has high pressure
turbine section, intermediate pressure turbine section, low pressure turbine sections,
generator and exciter mounted on the same shaft and in such turbine generator rotor when
it is disturbed or perturbed, torsional oscillations are produced and these torsional
oscillations in the sub synchronous frequency range in certain conditions may interact
with the electrical system and such interaction may be an adverse interaction.
In view of this it is necessary to study the sub synchronous oscillations in detail. The
problems related to torsional oscillations are torsional interaction with power system
controls, sub synchronous resonance and torsional fatigue due to network switching. We
shall study all these aspects while discussing the torsional oscillations or sub synchronous
oscillations in the system. Before, we address these specific problems the first step in the
study of sub synchronous oscillations is to develop the detailed model of the turbine
generator shaft system.
Now in order to develop the dynamic model of the turbine generator shaft system, we
consider a tandem compound of steam turbine which consists of high pressure section,
intermediate pressure section and low pressure section B and low pressure section A. All
these 4 steam turbine sections are mounted on the same shaft in addition to this the
generator is also mounted on the same generator rotor. Here, we will consider the static
excitation system and in case in case the system comprises of a rotating exciter then 1
more mass we have to be represented on the same shaft.
Now, here in order to develop the dynamic model of the shaft system or turbine
generator shaft system, we will consider that the torques which are developed in the
various turbine sections are represented as THP represents the torque developed in the
high pressure section, TIP is the torque developed in the intermediate pressure section,
TLPB is the torque developed in the low pressure B section and TLP is the torque
developed in the low pressure A section and the air gap torque developed in the generator
is represented by TE.
Further, further the the shaft connecting the high pressure and intermediate section
similarly, the intermediate pressure and low pressure sections the the shafts have finite
stiffness coefficient the stiffness coefficients will be represented by the symbol K45
joining the high pressure and intermediate pressure section. Similarly, K34 is the stiffness
coefficient of the shaft connecting a intermediate pressure and LPB and similarly the K23
and K12 are the stiffness coefficient of the shaft sections connecting LPB and LPA and LPA
and generator sections. While we develop the dynamic model system, we will assume ,we
will assume that the that the torque developed in this various turbine sections are
constant.
Now this assumption is particularly used for developing the small perturbation dynamic
model of the system. Now here each section that is the generator LPA, LPB, intermediate
pressure and high pressure section they have their inertia constants. We will represent the
inertia constant of generator by the symbol H1, inertia constant of the generator that is the
not generator but steam turbine section LPA by H2 and similarly, the inertia constant of
LPB, IP and HP sections. Further, we will assume that there is some damping associated
with the oscillation of these masses. Now here in this model the the the various various
turbine sections and generator are represented by lump mass. The dampings associated
with the these 5 different masses are represented as D1, D2, D3, D4 and D5. Now let us
understand what causes damping in the turbine generator shaft system.
Now when when the the when the blades of the steam turbine are rotating in the steam
which is flowing right therefore during in this process there exists some viscous damping
right and similarly the damping is also produced when the generator rotor is rotating
therefore the generator rotor also produce some damping because of the damper windings
and other phenomena which is exist in the system. Similarly, the when this this system
connected to the electrical network then this electrical network also produces some
damping.
Now damping is also produced by the mechanical hysteresis of the shafts connecting the
various turbine sections therefore there exists some damping in the system. However, the
damping produced due to this various phenomena is very small in magnitude, we will
develop the dynamic model of the system assuming the certain finite value of the
dampings associated with the with the generator and steam turbine sections. Now let us
just understand the meaning of this stiffness coefficient.
The stiffness coefficient is represented by the symbol K and when we represent the
stiffness coefficient of the particular shaft section, we will add additional subscripts. The
torsional stiffness defines the relationship between the torque transmitted torque
transmitted and the angular twist between the 2 ends of the shaft, this is very important
that that when the torque is transmitted by the shaft right then there is a relationship
between the the torque transmitted and the angular twist between 2 ends of the shaft and
the relationship is represented by this equation T equal to K theta where theta is the
angular twist and K is the stiffness coefficient and torque T stands for the torque
transmitted by the shaft.
Now in order to develop the dynamic mode of the system, let us consider consider the
generator and LPA turbine section in detail the on the rotor of the generator a torque Te
which is the air gap torque is acting. Now this torque acts in a direction opposite to the
direction in which the rotor is rotating. Now this generator rotor is connected to the
turbine section LPA the shaft connecting the turbine section LPA and generator has a
stiffness coefficient K12, the driving torque produced by the shaft is represented as T12
acting in the direction of rotation.
Therefore, when we look actually at the dynamics of the generator rotor we have 2
torques acting on the rotor, 1 is the torque transmitted by the shaft that is T12 another is
the air gap torque Te. In addition to this there is some damping torque acting on the
generator rotor. Now here the position of the rotor will be represented with respect to the
synchronously rotating reference frame the this phasor represents represents the
synchronously rotating reference frame, the speed of this is omega o. The generator rotor
rotates at a speed omega 1 and its axis is displaced with respect to the synchronously
rotating reference frame by an angle delta 1.
Similarly, when I talk about the turbine section LPA, LPA then the position of the rotor
with respect to the synchronously rotating reference frame is noted by an angle delta 2
and the speed is omega 2, under steady state conditions, under steady state conditions the
omega naught, omega 1 and omega 2 they are all equal. However, there is a finite value
of the delta 1, delta 2 and so on and in order to develop or in order to write the equation
of motion for the generator, let us write down the accelerating torque acting on the
generator rotor the accelerating torque acting on the generator rotor is equal to is equal to
T12 minus Te.
Now this T12 can be represented can be expressed as K12 into delta 2 minus delta 1minus
Te therefore, the accelerating torque which is acting on the generator rotor is this. Now
now when we consider the damping of the generator rotor then the damping torque which
acts on the system will be acting in the direction opposite to the accelerating torque
therefore the net accelerating torque will be this K12 will be delta 2 minus delta 1 minus T
minus D1 delta omega 1.
The equation of the equations of motion of the generator rotor, now can be written as 2
times H1 d by dt of delta omega 1 equal to Ta, this is the accelerating torque is equal to
T12 into delta 2 minus delta 1 minus T minus D1 times delta omega 1 right that this is the
net accelerating torque which is acting on the generator rotor. Now in this equation the
speed deviation is represented in per unit, the equation d by dt of delta 1 is equal to
omega naught into delta omega 1 therefore these 2 equations represents the equation of
motion of the generator rotor these are the 2 first order differential equations. Now let us
consider the steam turbine section LPA.
Now when I look at the steam turbine section LPA then we can clearly see that the
accelerating torque acting on the steam turbine section LPA can be written as that the
driving torque is T23 plus the torque developed in the steam turbine section that is T23
plus TLPA is the driving torque while the breaking torque is T12 and therefore the net
accelerating torque will be T23 plus TLPA minus T12 minus D2 delta omega 2 and therefore
the equation of motion for the LPA rotor that is the steam turbine rotor represented as LPA
can be written as 2 times H2 d by dt of delta omega 2 equal to TLPA torque developed in
the steam turbine section plus K23 delta 3 minus delta 2 that is this is your T23 minus K12
into delta 2 minus delta 1 therefore, this quantity is T12 and multiplied by D2 into delta
omega 2 this is the damping torque therefore the net torque acting or net accelerating
torque is TLPA plus T23 minus T12 minus D2 delta omega 2. The second equation the
differential equation for this LPA rotor is d by dt of delta 2 equal to omega naught into
delta omega.
Similarly, we can write down the equations of motion for LPB IP and HP sections. The
equations of motion for LPB rotor are 2 times H3 d by dt of delta omega 3 equal to TLPB
plus plus the T34 which is equal to K34 into delta 4 minus delta 3 minus T23 which is
written as K23 into delta 3 minus delta 2 minus D3 delta omega 3 and the next equation is
d by dt of delta 3 equal to omega naught into delta omega 3.

The equations of motion for IP turbine section are writing in the same way as we have d1
for LPB and LPA sections that is 2 times H4 d by dt of delta omega 4 equal to TIP plus K45
delta 5 minus delta 4 minus K34 into delta 4 minus delta 3 minus D4 delta omega 4 and d
by dt of delta 4 equal to omega naught delta omega 4. Our next step is to write the
equation for the last turbine section that is the high pressure turbine section.
Now when we look at the high pressure turbine section then the there is no shaft on the
left of the high pressure turbine section and therefore the net accelerating torque acting on
the high pressure turbine rotor is THP the torque which is developed in the turbine section
minus K45 minus D4, D5 into delta omega 5 and therefore the equation of motion is
written here as 2 times H5 d by dt of delta omega 5that is equal to THP that is the torque
developed in the high pressure turbine section minus K45 into delta 5 minus delta 4 minus
D5 delta omega 5 and d by dt of delta 5 equal to omega naught delta omega 5.
Now for this 5 mass turbine generator shaft we have written 10 first order differential
equations now these differential equations are non-linear differential equations. The the
equations are non-linear because the the air gap torque Te is non-linearly related to the
angular positions of the generator. Now we will attempt a simple problem the problem
is like this we consider the same 5 mass system, we will consider this same 5 mass
system and we will assume that that the power torque developed under steady state
condition in the high pressure intermediate pressure LPB, LPA have certain fixed ratios.
Now for this particular problem our interest will be to find out to find out that if the
generator is delivering certain amount of power right then how much is the power which
is which is transferred through each of these shaft sections that is the power or the torque
transmitted by each of the shaft sections which are connecting the various turbine
sections and the generator. We will also find out that how much is the total angular twist
which is created from generator to high pressure turbine section.
Now to illustrate this we will consider the that this system has the various stiffness
coefficients K12 is given as 37.25, K23 is 31.31, K34 is 27.66, K45 is 17.78. Now these
coefficients are expressed in per unit torque per electrical radian or per radian right
however we will assume that the power output from the generator is .9 per unit and in per
unit system the air gap torque Te is also equal to .9 per unit.
Now to solve this problem we need 1 more information and that information is that what
is the fraction of torque which is produced in high pressure, intermediate pressure, low
pressure and A and B sections. We will assume here in this say that that the fractions of
the torque which are produced in the HP is 0.28 where the fraction of torque which is
developed in T this intermediate pressure this TIP is equal to TLPB also equal to TLPA equal
to 0.24.
So that when you add these fractions the sum is 1. Now when the torque which is
produced in the by the generator is represented as Te and its value is 0.9 per unit then the
actual actual torque which is produced by HP, IP and LP sections can be obtained by
multiplying these all these coefficients are these fraction by .9 therefore, we can say that
the the THP produced is 0 point 2 8 into 0 point 9 similarly TIP equal to TLPB equal to TLPA
equal to 0.24 into 0.9, what is this value 0.216 and what is this value 0.25.
The meaning here is that the when the generator is delivering .9 per unit power right the
the the torques produced in the IP, LPB, LPA section will be 0.216 and torque produced in
HP section is 0.252. Now our interest to find out how is the torque transmitted through
each of the shaft sections that I want to know what is T12 what is T23 what is T34 and T
what isT45. Now to answer this what we have to do is that we make use of our definition
for definition for stiffness coefficient.
Our stiffness coefficient is that the torque transmitted is equal to the angular twist into
stiffness coefficient. Therefore, the angular twist of the shaft section connecting the
turbine section LPA and generator is delta 2 minus delta 1 and therefore this delta 2 minus
delta 1 is equal to T12 upon K12. Now the T12 under steady state condition is equal to Te
okay because under steady state condition the net accelerating torque is 0 therefore this
angular twist delta 2 minus delta 1 can be obtained as Te divided by K12 and when you
substitute the value of Te equal to 0.9 and value of K12 as 37.25, we will get delta 2 minus
delta 1 as 0.0246 radian.
Now the torque transmitted through the shaft section connecting the connecting the
turbine section LPB to LPA is T12 minus LPA,TLPA, now T12 is .9 and the torque developed
in the turbine section LPA is 0.216 this is the torque which is transmitted and therefore the
angular twist is delta 3 minus delta 2 of the second shaft section that is T23 divided by K23
when we substitute the values of T23 and K23 we will get the angular twist as 0.02185
radians. Similarly, 1 can find out the angular twist delta 4 minus delta 3 delta 5 minus
delta 4 and this come out to be like this and the net angular twist or angular twist of the
high pressure section and the generator or we can say the net angular difference between
the axis of the generator rotor and the axis of this high pressure stream turbine comes out
to be 4.42 electrical degrees.
Therefore when you can you can easily understand here that when when the the steam
turbine is delivering certain power that the there exist a exist a certain net or twist in the
shafts and the net twist is given by this formula delta 5 minus delta 1 and we can see here
actually that the the angular twist is small okay.

When it is delivering 90 percent of the rated power its its twist is hardly of the order of
few electrical degrees. Now the next step will be to develop the the linear dynamic model
of the system now in order to develop the linear dynamic model of the system what we
do is that we linearize these differential equations around the nominal operating
conditions. Further, while doing this linearization where we will assume that the torque
developed in the various turbine sections are constant right.
So that we can write down we can write down now that under small perturbation, when
the system is perturbed the Te the electrical torque or the air gap torque is equal to Teo
plus delta Te. Similarly, since we have assumed this the torque developed in various
turbine sections constant right therefore the THP is equal to THPo there is no change
similarly TIP will be equal to the TIPo. Now this o stands for or 0 stands for steady state
value of the torques and similarly we can write down that T LPB equal to TLPBo and TLPA,
LPA equal to TLPA.
Now these angular positions will represent the angular positions like say angular position
under the dynamic condition as delta 1 equal to delta 1 o plus delta delta 1. Similarly,
delta 2 will be written as delta 2 o plus delta delta 2 and so on we can write down the
delta 5 as delta 5 o plus delta delta 5 similarly when we talk about the speed right. Now
under steady state conditions omega 1, omega 2, omega 5 are equal to omega naught
okay and when the system is perturbed the deviation in the speeds is delta omega 1 delta
omega 2 and delta omega 5. Now using these small perturbations around the operating
condition we can write down the dynamic model of the system around the operating
condition.
Now suppose I consider this equation this is the equation for motion of the generator
rotor. Now here when you substitute substitute delta 2 equal to delta 2 o plus delta delta
delta 1 equal to delta 1 o plus delta delta Te as Teo plus delta delta Te and this this term
remains same.
Now with this the when you substitute the quantities we can write down the equation in
this form 2 times H1 d delta omega 1 by dt of d by dt of delta omega 1 equal to equal to
the the torque which is acting here on this is the accelerating torque which is acting on
this is T12 okay now this T12 can be written as is K, K12 into delta delta 2 o plus delta
delta minus delta 1 o minus delta this is delta delta 2,1 okay minus minus Teo minus delta
Te minus D1 delta omega 1, I remember that when we had written actually derivative of
speed d by dt of delta omega 1 is same as d by dt of omega 1 okay.
Now in this equation under steady state condition under steady state condition K12 delta
2o minus delta 1o minus Teo equal to this is under steady state condition therefore
therefore for considering the small perturbation around the steady operating condition
around nominal operating condition we can write down the linear differential equation as
2 times H1 d delta omega 1 by dt that is d by dt of delta omega 1 equal to K12 delta delta 2
minus delta delta 1 minus delta Te minus D1 omega 1. Our second equation when we
write down delta 1 equal to delta 1 o plus delta delta 1 the linear equation will come out
to be d by dt of delta delta 1 equal to delta omega 1 into 0 therefore these 2 equations are
the linear differential equations.
Now here this torque delta Te is expressed in terms of the deviations of these angles delta
delta 2, delta delta 1 and so on it is a function of the the deviation of the angular positions
of the various rotors okay therefore, these 2 equations represents the linear differential
equations. Similarly, we can write down the or linearize linearize the differential
equations associated with the turbine section LPA, turbine section LPB turbine section
intermediate from the IP and turbine section HP that is all the 9 differential equations can
be linearized and once you linearize these equations we will get the dynamic model of the
system in the form X dot equal to AX that is we can write down the dynamic model of
the complete system in the form X dot equal to AX where X is the state vector written as
delta omega 1, delta delta 1, delta omega 2, delta delta 2, delta omega 3, delta delta 3,
delta omega 4, delta omega delta delta 4, delta omega 5, delta delta 5 transpose that is X
is a vector comprising of these 10 state variables.

Now once we have obtained the linear dynamic model of the system we can further study
the behaviour of the behaviour of the torsional oscillations or we can study the
oscillations of the turbine generator shaft. Now to study the behaviour of the system first
step is to find out Eigen values of the system and these Eigen values are Eigen values will
give you natural frequencies of the shaft system that is when we obtain Eigen values what
we will get is we will get the natural frequencies of the shaft system.
Now the system matrix A, A is a 9 by 10 matrix and the elements of this matrix depend
upon the damping constants, inertia constants, the stiffness coefficients and the
synchronizing torque coefficient, okay. Now to understand the behaviour of the or
dynamic behaviour of the turbine generator shaft system. Let us take an example, we
consider here the same 5 mass system and we represent these masses as 1, 2, 3, 4 and 5
the inertia constant of HP section is .124, inertia constant of IP is .232, LP is 1.155 and LPA
is 1 .192 and generator is .855 that is these are the different values of the inertia constants
of the 5 masses. The stiffness coefficient of the shafts connecting the various masses are
represented as 21.8, 48.4, 75.6 and 63. 62.3 that is in this diagram what I have shown is
that these are the inertia constant of the 5 masses and these are the stiffness coefficients.
Further, we assume that is synchronizing torque coefficient K is is 1.98 right and this K is
depends upon the operating condition okay.
Now the operating condition which we have considered is that the system is delivering 90
percent power okay. Now for this system for this system we develop the dynamic model
and write down the a matrix or determine the a matrix of the system. Now in order to in
order to simplify out understanding, we neglect here the damping coefficients associated
with the system right. When we neglect the damping coefficients and we obtain the Eigen
values of the system matrix, the Eigen values this system matrix A which I have just now
shown is a real matrix right and since we have assumed the damping of associated with
these masses as negligible and with this assumption the Eigen values come out to be all
complex conjugate and the real part is 0.
Now for these Eigen values we can find out the associated natural frequency of
oscillation. Now for this first Eigen value whose imaginary part 9.12 the associated
natural frequency is 1.61 hertz and similarly, the frequency associated of this second
Eigen value is 16.37 for a third Eigen value is 24.123, the next is 30.25 and next is
43.996 hertz, I mean these are the 5 natural frequencies associated with the turbine
generator shaft that is whenever this turbine generator is perturbed right then then the
torsional torsional oscillations are produced and the torsional frequencies are given like
this that is you can see here that we have for this 5 mass system these 5 different
frequencies the the frequency 1.671 can be considered to be the system frequency where
the the entire turbine generator mass oscillates with respect to the system while other 4
frequencies are associated with the different torsional modes of oscillations.
Now to further understand the behaviour or dynamic behaviour of the system, we analyze
the the dynamic model of the system further that is for each of the Eigen values we obtain
the Eigen vector and we can easily understand the meaning of the Eigen vectors, we all
know what we stand for Eigen vector that is for each Eigen value there exist an Eigen
vector and we can understand what the Eigen vector stands for that is for any Eigen value
lambda i the column vector phi i which satisfy this equation A phi i equal to lambda i phi
i is called the right Eigen vector of the matrix A.
Now this phi i is associated with Eigen value lambda i if I put i equal to 1, I will get
Eigen vector 1 if I put i equal to 2, I will get Eigen vector corresponding to Eigen value
lambda 2 and so on now this Eigen vector will have the same dimension as they are the
dimension of state vector okay. The Eigen vector phi i corresponding to lambda i gives
mode shape, you have to understand this very important this is very important term gives
mode shape that is the relative activity of a state variable in a given mode okay.
Now to understand this what we will do is that we will find out we will find out the Eigen
vectors corresponding to each of the 5 Eigen values or 5 complex conjugate Eigen values.
Now Eigen values of the system matrix A can be easily obtained using mat lab software
and similarly, we can find out the Eigen vectors corresponding to each of the Eigen
values. Now these are obtained for the system which we have investigated.
Now here in this table I have shown the Eigen vector, this Eigen vector is associated with
Eigen value 0 plus minus 20.21 i that is for this first Eigen value I have obtained using
the matlab software the Eigen vector and this Eigen vector has has 10, 10 actually entries
or now we can easily see here actually the first these first 2 elements of the Eigen vector
are associated with the rotor of generator that is delta omega 1 and delta delta 1.
Similarly, the next 2 Eigen values are next 2 entries are associated with with the state
variables associated with LPB rotor section and so on. Now here what we have done is the
we are considering, we can consider we can consider the either the elements associated
with the speed or angle for analyzing the mode shape therefore, we have considered the
the elements associated with speed not speed but the angular position these are the
angular positions, okay.
Now next we have plotted the magnitude of these elements the magnitude is same as the
.4346 because its real part is 0 and the angles are also shown the angular for example, the
angle associated with this element is 90 degrees, angle associated with this is also 90,
angle associated with this is also 90 and so on that is all these elements have phase angle
equal to 90 degrees. Now when we plot when we plot the magnitudes magnitudes of
these Eigen vector elements right what we will do is that those magnitudes which are
positive right we will plot actually in the as positive and those which are negative we will
plot in the as negative.
I mean in this case we find that the all these elements have the same angular position and
therefore they are all positive. Now this diagram shows the plot of mode 0, this mode 0 is
the system mode whose frequency is 1.67 hertz and this mode 0 has the magnitudes 1.99
6.99, 3.992 and and .958 that is more or less the the all the all the state variables
associated with with this mode have the same displacement okay and this mode this 0 th
mode or system mode represents that all the masses rotate rotate together with respect to
the system okay or rotate rotors of other machines.
Now the next we consider the torsional mode, the frequency the Eigen value of the
second or the first torsional mode is 102 plus minus 102.89 i and its frequency 16. 37.
Now following the same approach for this Eigen value the Eigen vector is obtained, now
this Eigen vector you can see here it has again the 9 elements and you have real
imaginary real imaginary and so on.
Now this these elements these elements are associated with with angular displacements
delta delta 1, delta delta 2, and delta delta 5. The magnitude and phase angle of these
elements are written here as .4178 minus 90 degrees because minus is here like that then
this vector is normalized when we normalize it the the element of the vector which is
maximum magnitude or highest value right, we use this as 1 right and rest are represented
with respect to this as the base therefore we find here that the the normalized values of
the elements associated with delta delta 1, delta delta 2 and so on are given like this.
Now here the angles are minus 90 minus 90 and then they change over to plus 90 plus 90
plus 90 therefore now if I represent these as positive they will become negative right
therefore the plot which I have shown here for this it is starting from minus 1 minus
.7676 minus .32 that is this is the mass 1 this is mass 2, mass 3, mass 4, mass 5 okay and
you show the magnitudes positive or negative this is our 0 line okay therefore this shows
where we join all these points by a state line and this gives us a mode shape
corresponding to the mode 1 or this is the mode shape of mode 1 whose natural frequency
of oscillation is 16.3 hertz.
Now let us understand what we mean by this mode shape the meaning is can be easily
understood here that if you look at this diagram right then the displacement of
displacement of these these masses 4 and 5 with respect to displacement of mass 1, 2 and
3 are in opposite direction. It means here the mass that is generator the rotor of LPA and
LPB sections oscillate with respect to the rotors of IP and HP sections that is whenever the
generator rotor is perturbed right and the the torsional with this torsional frequency or
with this frequency 16.3 hertz in that case in that case the the generator the LPB, LPA and
LPB they will oscillate with respect to the other 2 masses and in this mode 1 there is 1
crossing 1 changeover of sign that is you can just see here actually this is a 0 line and
there is 1 crossing here.
Now similarly we have obtained the Eigen vector for the mode 2, this is mode 0, we will
call it mode 0, this is mode 1, this is mode 2 that is the frequency of oscillations or
torsional oscillations torsional frequency is 24.123 and the corresponding Eigen vector is
shown here.
Now if you follow the same procedure as we have done for other 2 cases then we obtain
actually the magnitude of the elements associated with delta delta 1, delta delta 2, delta
delta 5. These are the magnitudes and these are the associated phase angles you can easily
say that phase angle of this is 90 degrees phase angle of these 2 elements minus 90 minus
90phase angle of other these again plus 90 plus 90 it means when we plot plot the the
magnitudes of these this displacements right you will find that there will be 2 crossings
right and the mode shape is shown here that is this is the frequency 24.1 hertz and the
mode shape you can see is like this therefore there are 2 crossings here that is the mode
shape will always have 2 crossings okay.
Now in this when you try to analyze here then you can easily understand that the that
these 2 masses 1 and 2, these masses will oscillate with respect to the masses 3 and 4 and
similarly, the 3 and 4 will oscillate with respect to mass 5 that is whenever it is disturbed
or perturbed with the frequency corresponding to this 24.1 right this the shaft will
oscillate in such a fashion. So that the mass 5 will oscillate with respect to 3 and 4 and 3
and 4 will oscillate with respect to 1 and 2 right and frequency of oscillation is so much.
Similar studies performed for the mode 3 whose frequency is 30.25 hertz and this is the
Eigen vector again we are considering the elements corresponding to angular
displacements. Now these are the normalized magnitudes of the elements associated with
with the the speed deviations and you can easily see here actually that we have 3 change
over is that is start from positive go to negative then again positive and again negative
and this shows the mode shape associated with the mode 3 whose frequency is 30.4, 30.3
hertz and again. Now we can easily see here that when the system is perturbed we will
find actually that the mass 1 will oscillate with respect to mass 2 and 3 and 2 and 3
oscillate with respect to mass 4 and 4 and 4 will oscillate with respect to mass 5 and so on
right. Similarly, the mode shape for mass 4 is shown and you can easily see that there are
4 crossings.
Now with this let me summarize that we have studied the modeling of turbine generator
shaft system. We have obtained the non-linear and linear models and we have also
studied the application of Eigen vector concept to study the mode shapes of the system.
Thank you!
Prof. M. L. Kothari
Lecture - 34
Sub Synchronous Oscillations (Contd)
Friends, today we shall continue our study of sub synchronous oscillations. We had developed
the non linear and linear dynamic model of turbine generator shaft system, we have also
analyzed the different torsional modes of oscillations and the mode shapes using model analysis.
The major problems related to the torsional oscillations which we shall discuss today are the
torsional interaction with power system controls, sub synchronous resonance and torsional
fatigue due to network switching as I pointed out in the beginning that the torsional oscillations
may interact in the adverse manner with the power system network and therefore it is necessary
to clearly understand the the torsional interaction with power system controls then sub
synchronous resonance and torsional fatigue due to network switching.
First we will study about the torsional interaction with power system controls. When we talk
about the power system controls I we will address the two important controls one is the
excitation control that is interaction of the torsional oscillations with the excitation control and
second the interaction of the governor of the steam turbine with the torsional oscillations. Now
let us start with the interaction with excitation controls.
Now we have studied that power system stabilizers are used for damping the system modes of
oscillations that is the power system oscillations in the frequency range of .2 to 2 hertz. Now
here when we design the power system stabilizer we ensure that a component of electrical torque
or component of electrical torque produced is in phase with the speed deviation signal and the
parameters of the power systems stabilizers are so chosen so that we get a component of air gap
torque in phase with speed deviation and this is pure damping torque.
Now if you look here in the in this diagram I have shown here the on x axis delta delta and on y
axis delta omega right. In case we produce a torque which is in phase with speed deviations delta
omega that is delta Te is the torque produced in phase with speed deviation delta omega then this
torque is a purely damping torque.
Now in practice what happens is that the that the the the speed signal is sensed by using a speed
transducer and when we mount a speed transducer say particular location, let us say at the
generator end of the turbine generator shaft then at that particular location, location the speed
signal has a high component of of the torsional oscillations. If you examine this diagram that is
for typical for typical turbine generator system this is the mode shape corresponding to 16.3 hertz
and we can easily see here the that the component of speed deviation signal which is sensed at
the generator end of the shaft is quite a large one and this component when it is processed
through the power system stabilizer will produce produce air gap torque a component of air gap
torque which will have a large phase shift between between the speed deviation signal and the air
gap torque produced.
For example, if you see in this diagram in a typical case the air gap torque or the component of
air gap torque produced due to due to torsional component which is actually fed back through
the power system stabilizer that is the speed signal has a torsional component or high magnitude
of torsional component and this signal when it is fed through the power system stabilizer and air
gap torque which is produced this air gap torque comes out in the fourth quadrant, a typical case
now this air gap torque delta Te when it is resolve into two components one in phase with angle
deviation the delta delta another in quadrature with angle deviation that is in phase with speed
deviation signal delta omega. Now the component of the air gap torque in phase with speed
deviation comes out to be negative that is the the power system stabilizer will contribute a
negative damping torque and this negative damping torque if it is more than the normally
existing system damping then it will cause excitation of the torsional modes because they as we
have seen that the damping of the torsional modes is very low right and the moment you have a
large negative damping torque produced due to the due to the power system stabilizer then this
the net damping will become negative and the torsional oscillations may grow in magnitude and
therefore there is a necessity to account for this adverse effect when we design our system.

Now various measures have been suggested to counteract this problem of excitation of torsional
modes due to power system stabilizers, one very interesting example is given in the power
system stability and control by Kundur in the text book power system stability and control by P
Kundur and example 15.2 addresses the various aspects of the interaction of torsional oscillations
with excitation control.
Now to examine this he has considered steam turbine which has high pressure section then they
are three sections which are called low pressure sections he names them as LP1, LP2, LP3 and
each low pressure section has two sub sections called left and right right and they are all
mounted on the same shaft at the end of the generator. Now to overcome the problem of
excitation of torsional modes due to adverse interaction between the excitation controlled and
torsional oscillations the studies were conducted by putting by putting speed sensing device at
the end of the generator that is suppose you put speed sensing device here the speed senses at
delta omega g, another location which was identified for sensing the speed is between LP2 and
LP3 that is delta omega c that is at the location c.
Now at c, c forms a coupling between the LP2 and LP3 sections. Similarly, studies are also
carried out by sensing speed at the coupling between LP1 and LP2. The reason for trying these
difference locations were that at these locations the amplitude of the torsional oscillation or was
very small right therefore, if you pick up the speed here then then the chances of accept some
torsional modes is practically low. Another step which was taken is that that filter was put in the
in the power system stabilizer circuit that is the speed deviation signal is first filtered out by
putting a special filter which will which will filter out or eliminate the torsional modes right.
Now in case actually the you have torsional mode one is having the predominant signal then the
filter is design designed so that so that the the torsional mode one is completely attenuated while
other torsional modes will be highlight attenuated right and the first step which was taken was
that study was conducted considering delta omega PSS by locating this PSS as different
locations, studies were conducted by putting the speed sensing or speed sensor at location B at
location C and then at B and C both right. Then the next stage, the stage were conducted by
booting delta omega PSS with the torsional filter right.
Then the last study was carried out by putting a stabilizer where we do not sense the speed
deviation as the input signal but we sense the deviation in power as the input signal the delta P
stabilization and the study show that that if you have delta omega PSS only without torsional
filter then as you increase the power system stabilizer gain then some of the torsional modes will
get excited or they will have negative damping, then further studies were carried out when the
torsional filter was put then it is seen actually that with the moment where torsional filter
incorporated then the chances of {torse} ((00:12:22 min)) excitation torsional modes is
eliminating.
Further, when the delta P stabilization was put then the then the results obtained were similar to
what we got actually in second case where delta omega PSS with a torsional filter was output.
Now all the details are given in the example 15.2, I will suggest the interested readers to go
through this example in detail.
The next aspect we have to address is the interaction with speed governors, how the speed the
governors will interact with the torsional oscillations now historically the interaction with speed
governors was observed or experienced in 1983 when a 655 MVA, 1800 RPM nuclear unit at
Ontario hydro was commission this this unit, this is a nuclear unit basically the the turbine is
similar to a steam turbine it is steam turbine only okay and the the difference in this unit was that
was equipped with a fast acting electrohydraulic governor system because we may I have the
mechanical mechanical controlled system right and we may have fast acting electrohydraulic
governor system in this electro hydraulic governor systems which are now provided in most of
the modern turbines the the low power functions are performed by using electronic circuits that
is the speed sensing and certain processing is done using electronic circuit, one of the important
processing which is done in electro hydraulic governor system is they use a linearizing circuit
to compensate for the non-linear flow verses valve position characteristic of the steam valves that
is if you plot a characteristic which relates the steam valve position with respect to the steam
flow.
Then this characteristic is a very linear characters non-linear characteristic and to compensate for
the nonlinear characteristic right we provide here electronic linearizing circuit. Then it was
observed that when this machine which was under commission when this machine was loaded to
100 megawatt load its capacity is 635 MVA then then the vibrations of the steam valves was
observed experienced and then immediately immediately the they looked into the electronic
circuits and some retuning of the electronic circuit was done.
With that retuning they could load up to 475 megawatt and at that particular loading condition
again it was seen that the steamed valves were vibrating and then by it was examine in detail it
was found that this non-linear that is the electronic linearizing circuit right. It was behaving in
the different fashion at that particular frequency that is a particular frequency was sensed because
here if the governor sense is the speed of the rotor and for sensing the speed of the rotor the
speed sensor was mounted on the high pressure end of the stem turbine okay and at that
particular end it picks up the various torsional components right and when these torsional
components are process through the electronic circuit it was found that the speed droop at
475megawatt loading came out to be something like 12 types more than the normal 4 percent
speed droop and then the certain action was taken to overcome this problem or to mitigates this
problem.
Therefore this is also very series problem when you when you have to address the torsional
interaction with governor circuits because modern steam turbines are with electro hydraulic
governors and these days, these circuits are so designed, so that all through the loading condition
a four percent speed droop is maintained speed droop .You know all the speed droop parameter
are which is the governor characteristic where where actually the when you change the load from
no load to full load the droop in this speed is 4 percent from the normal value or when the
machine is operating under full load condition and if the load is thrown off the increasing in
speed will 4 percent of the rated speed.
Now this particular problem has been nicely dealt with in the research paper written by D. C. Lee
R. E. Beaulieu and Rogers effect of governor characteristics on turbo generator shaft torsionals
there is specific is specific problem which was examined this has appeared 1985.
In fact this torsional oscillations do interact adversely with the near nearby HVDC controls also
these they needs to be examined in detail but it is observed that the HVDC controls also interact
in adverse matter manner if you do not account for the torsional modes. The next important
aspect will we basically study is the sub synchronous resonance. First let us define what you
mean by the sub synchronous resonance then we will study this phenomena in detail.

The sub synchronous resonance is defined as as an electric power system condition that is sub
synchronous resonance is an electric power system condition where, the electric network
exchanges energy at one or more of natural frequencies of the combined system below the
synchronous frequency of the system that is here the importance is that the frequencies of
interest are the frequencies which are less than the synchronous frequency these are called are
the sub synchronous frequency and there is a condition where the electric network that is the
power system network exchanges energy energy at one or more of natural frequencies right and
therefore to understand this problem very we have to examine the system very carefully and let
us understand the SSR phenomena or sub synchronous resonance phenomena, this this SSR
phenomena was first experienced in 1970 resulting in the failure of turbine generator shaft at
Mohave plant in southern California.
Now till that time till that time people had no idea about the subsynchronous resonance and in
the Mohave plant which is located in southern California and USA right in 1970 the shaft of one
machine completely damaged then the second shaft failure occurred in 1971 then it was a series
problem the studies were conducted the system was modeled and after detailed studies it was
found that the shafts were damaged because of SSR phenomena which occurred in the system,
this SSR phenomena phenomena primarily occurs in series compensator transmission lines okay.
Now there is a necessity for us first to understand the characteristic of the series compensate
transmission lines before we understand the SSR and its impact on the system. Now first let me
just tell you that suppose you have a uncompensated transmission system and in uncompensated
transmission system suppose fault occurs in the transmission line okay, then depending upon the
strength in the voltage wave at which the fault occurs there exist a decaying DC component of
fault current and this decaying DC component of fault current flows through the distributed
stator windings of the system that is when the stator windings which are distributor, 3 phase
stator winding right it will carry it will carry the decaying DC component of the fault current and
when this decaying DC component of fault current is flowing in the stator winding then then in
the air gap torque having frequency equal to 60 hertz will be produced a right.
Now normally under steady state condition the torque which acts on the rotor of the turbine
generator system or is always having 0 frequency, it is unidirectional torque when the it is
carrying three phase balance currents but the moment actually the stator carries a decaying DC
component of current right the torque or air gap torque is produced to its frequencies 60 hertz if
it is a 60 hertz system, if it is a 50 hertz system the frequency will be of 50 hertz. Similarly, when
you know that whenever there is unbalanced fault which takes place on the uncompensated
system then the negative sequence currents will flow in the stator winding of the generator and
these negative sequence currents will produce produce the air gap torque whose frequency will
be 2 times the system frequency, it is the 50 hertz system the frequency will be 100 hertz, if it is
60 hertz system these 120 hertz and therefore therefore there is necessity that when you design
your turbine general shaft none of the torsional modes should be close to 60 hertz or 120 hertz.
Further, if I consider now a series compensator transmission line and the primary purpose of
providing series compensation is to reduce the line impedance, so that we can transfer more
power on the transmission line. Now let us consider machine infinite bus system where we have
a generator, transformer a step of transformer and the transmission line whose parameter RE and
XE, the resistance is RE, reactance is XE.
Now we have provided a series capacitor of reactance XC right. Now this system this system
when a fault occurs in this particular series compensated system then the fault currents will also
have have components of frequency which will be corresponding to the natural frequency of the
inductance and capacitance of the system that is all you know actually that suppose when you rlc
circuit, when this rlc circuit is excited right then the then the current which flows will will have
frequency of oscillation which depends upon the inductance and capacitance of the circuit. Now
since the the capacitance which you put here in the circuit is to partially compensate for the line
reactance at power frequency and therefore at frequency less than the power frequency at
subsynchronous frequency these two reactances will become equal right and at that time we say
that this is series resonance occurs.
Now we can draw the equivalent circuit of temperature system that is generator can be represent
by a voltage in series with the reactance transformer of the reactance XT line with reactance RE,
XE and resistance RE, capacitance XC the this can be simplified then we can write down a voltage
source in series with the resistance the total reactance XL that is XC plus XT plus X double prime,
sub transient reactance and a capacitance connected to infinite bus okay. Now for this circuit the
natural frequency of a oscillation, it is denoted as by the symbol omega n is equal to 1 divided by
squire root of LC, this is the natural frequency of oscillations, okay. This can be written as omega
naught divided by squire root of omega naught into L omega naught into C okay this is exactly
same as this one there is you have multiplied and divided by omega naught this can be expressed
as omega naught multiplied by squire root of XC by XL okay. Therefore, now we have a
relationship between the natural frequency of oscillation fn with the system frequency fo and and
the values of the line reactance and series capacitor, the capacitor, the reactance of the series
capacitor.
Now this relationship is a very important one. Now the problem come something like this that
that whenever some disturbance takes place right and if the currents of this frequency fn flows in
the stator winding of the synchronous generator then the there will be a torque developed in the
air gap because air gap torque the frequency of this air gap torque will be equal to fo minus fn that
is, if fn is the frequency of the currents which are flowing in the stator winding then in the
frequency of the torque developed or frequency of the air gap torque developed due to this
current will be fo minus fn because we have seen that if fn is 0 that is if it is carrying DC current
decaying DC current the frequency of the torque developed is the simply fo right therefore now
if the current flowing has a frequency fn then the air gap torque will have a frequency which is fo
minus fn.
Now in case the frequency frequency of the torque developed that is fo minus fn, if it coincide
with any of the torsional modes then these torsional modes can can grow in magnitude or can be
sustained that is what is the main philosophy of excitation of torsional modes. This table is an
important table which shows which show the percent compensation that is XC by XL this ratio
multiplied by hundred that shows the degree of compensation for example here when I say ten it
is a 10 percent of the line reactance is compensative and so on as shown this 10, 25, 30, 45 and
50 as these frequency the with this with this compensation the natural frequency of oscillation
comes out to be 18 hertz, 30 hertz, 32.6 hertz, 38 and 42.4 hertz, while the slip frequency we so
add 60minus fn right because the when the currents of the natural frequency flows in the stator
winding then the torque developed will have a frequency of 60 minus fn that is the slip frequency
slip frequency torque are produced and this slip frequency torque should not should not have
should not coincide with the frequency of any of the torsional modes, this is one of the important
requirements.
Now you can easily see here actually that at load when the compensation is low the slip
frequency is high right and as the compensation increases the slip frequency decrease, suppose
you have 50 percent compensation this frequency may come very close to frequency of mode
one right and as you decrease this you will find that you will go and match with the higher and
higher as together higher and higher slip frequency and it may match with higher modes. A
series capacitor compensator transmission network can cause sustained or negatively damped
sub synchronous oscillations by 2 distinctive mechanisms.
Now this is the important aspect which will you address in our study that is a series capacitor
compensator transmission network can cause can cause sustained, I am using the word sustained
or negatively damped sub synchronous oscillations that two mechanisms, one is the self-
excitation due to inductance generator effect and second is interaction with torsional oscillations.
Now this these two effects are completely different, now to understand this the self-excitation
due to inductance generator effect we can visualized that is a synchronous generator whose rotor
is running at synchronous speed.
When the stator of the synchronous generator is carrying sub synchronous frequency currents
right then due to the sub synchronous frequency currents a rotating magnetic wheel is produced
whose speed will be less than the synchronous speed while rotor is rotating at synchronous speed
this is situation similar to inductance generator effect that is if you consider only the situation
where the stator is carrying subsynchronous currents right and then the behavior of this
synchronous generator can be examined considering the inductance generator effect because in
induction generator, the rotor of the generator rotates at a speed higher than the synchronous
speed.
Now when you examine this the self-excitation due to induction generator effect. Let us look at
the equivalent circuit of the synchronous generator when it is behaving like a induction generator
right we are not considering the synchronous generator aspect here. We are considering only that
portion of the system which behaves like induction generator this is the armature resistance this
is the leakage reactance this is the resistance of the rotor, this is the leakage reactance of the rotor
and this is the resistance which takes care of the loading that is the load, now here this resistance
Rr is 1minus s by s.
Now this slip is equal to fn the frequency of the current which is flowing through the stator minus
fo the the frequency corresponding to the synchronous speed then fn minus fo divided fn is the slip
because when you define the slip right what we do we, suppose it is inductance motor right the
suppose actual speed is m then the synchronous speed is ns, then the slip is ns minus m and then
this is expression percentage of ns, okay. Now here corresponding to this ns I am using this
frequency fn because the current which is flowing is natural frequency current fn that is fn minus
fo by now since fn is less than fo the slip is negative, when the slip is negative then the this
quantity this part of the resistance will become negative because because s is negative the
denominator right with 1 plus this was positive but this is a negative term here this system
become negative.
Now when if you neglect this magnetizing reactance then equivalent circuit can be shown as R
effective plus X effective this is the equivalent circuit looking into the stator terminals of the
synchronous generator that is when you look into the stator terminals then this circuit or stator
circuit will behaved as if it has effective resistance and effective reactance therefore the net
effective resistance is equal to Ra Ra plus Rr by s and now s is negative right and therefore in case
Rr by s will exceeds Ra when R effective becomes negative. Now the point here is that they we
have when when we look at the stator with side only the stator is connected to the transmission
network through transformer and it is a series compensated line. Now in case the currents of a
natural frequency fn flow in the stator winding right and if the circuit has circuit has net
resistance which is less than 0 or negative resistance then these oscillations can grow that is the
in rlc circuit if r is positive then the oscillation will damp out okay.
They will decay but if r becomes 0 the oscillations will sustain but when r is negative oscillations
will grow right and this is the phenomena where we say actually that due to due to the series
compensated transmission line right and due to induction generator effect, the effective
resistance of the circuit as a whole will become negative or very or 0 then the oscillation can
either grow or again sustained right and when the oscillations are sustained right their
magnitudes are increasing then the then the torque developed on the generator of corresponding
to the slip frequency and as I told you that this torques should not have a frequency which should
match with the any of the torsional modes you want to avoid the damage to the system. The
another aspect which I have to emphasize here is the torsional interaction resulting in SSR.
Now as I have told you that so far the induction generator effect is concerned concern there is no
interaction with the torsional modes right and this phenomena is basically a circuit phenomenon,
when you talk about the torsional interaction resulting resulting in SSR. Now if the complement
of the natural frequency when complement that is the synchronous frequency minus the natural
frequency of the network is close to one of the torsional frequencies of the turbine generator
shaft system the torsional oscillations can be excited this is what we have already seen. This
condition is referred to as sub synchronous resonance that is in case in case the complementive
frequency, complementive frequency you always call the word complementive frequency
because because the slip frequency plus the natural frequency the sum is equal to the system
frequency you called the complementive frequency. In case the complement of natural frequency
is is closer to any of the torsional modes right then the oscillation can be excited and this
phenomena is called SSR nominal.
The consequences of the subsynchronous resonance can be dangerous, if the torsional

oscillations build up the turbine generator shaft will break with disastrous consequences as you
have seen actually that at Mohave the shaft damage took place and it was ultimately identified
due to the occurrence of SSR phenomena even if the oscillations are not unstable I am using the
word. Now oscillations may be stable or unstable system disturbances can the system
disturbances can result in shaft torques of high magnitudes causing loss of fatigue life of the
shaft the idea here is that when the oscillations are produced in the shaft okay even if suppose the
oscillations do not grow right the shaft is subjected to increasing and decreasing stresses right
and due to this there is certain fatigue cause to the shaft and there is the terminology which is
used is called it the fatigue life that is each shaft has some fatigue life right and and when this
shaft is subjected to this oscillatory torques the fatigue life is lost or it decreased.
Now another important aspect is very important the torsional interactions effects or torsional
interactions effects involve energy interchange between the generator shaft system and
inductance of the capacitances of the network that is basically there are basically the there is a
exchange of energy between the generator shaft system and the network inductance and
capacitance because they are the energy storing devices. Therefore, the analysis of SSR problem
requires representation of the both the electromechanical dynamics of the generating units and
the electromagnetic dynamics of the transmission network this is point I want to highlight
because when we studied the power system dynamics okay and when we develop the
synchronous generator model we neglected the stator transients right the stator circuit is
represented by algebraic equations okay and similarly the transmission line equations are also
obtained by steady state analysis or transmission line performance is also obtained through
algebraic equations it means we have not represented the dynamics of the transmission line
network or transmission network or the stator network but here when you want to analyze the the
torsional interaction then we have to develop detailed dynamic model of the electromechanical
system as well as the electromagnetic dynamics of the transmission network and that is why
whenever you develop the model the dimensional of the model will become higher right and a lot
of research work has been done on modeling the the system for analyzing the SSR phenomena.
The generator stator circuits and the network cannot be represented by steady state algebraic
equations as in the system stability studies which is just now I have highlighted the consequently
the system model used for SSR studies is of higher order and greater stiffness than the model
used for stability studies is it clear that is when I say that the when you develop the model of the
system system for SSR study the dimension of the model is high, it is a high order dynamic
model of the system and since the we know about the stiffness what do you mean by the stiffness
that is the smallest Eigen value to the largest Eigen value that this ratio gives us the stiffness
therefore, since actually there will be different type of time constant involved the stiffness also is
different greater stiffness is there and therefore when you solve these equations you may have to
take small integration step.
Now several techniques have been used to analyze the SSR phenomena I will just briefly discuss
these techniques them out of all the technique the simplest technique is Eigen value analysis
where you develop the linear dynamic model of the system and obtain the Eigen values of the
system and you performed the studies whether whether the system is stable or not stable and
which are the modes which may get excited depending upon the dampings associated with or
damping ratios of various modes of oscillations that is when you consider the realistic system
right then the Eigen values will be compressed conjugate or real right and we can find out the
damping the associated with each of the modes and we can find out what is the damping
associated with this in case damping is poor or if it gets excited, if damping is negative then
minus you take some measure.
The the second method is called frequency scanning method this is this method is used actually
to do some preliminary studies what is done is what is done is that looking from the internal bus
of the synchronous generator right, you try to see what are the frequency of the various
components that is you can just to find out the impedance of the network to different frequencies
right and that helps us actually in finding out which the help us in finding out preliminary, the
occurrence of SSR phenomena because we can easily find out actually the what is the impedance
to this particular frequency, what is the impendence to the next frequency called frequencies
scanning right wherever the impedance is very low to particular frequency that that is the
correspond concern.
Another is the frequency response analysis for full system that is we have we can find out the
stability of the subsynchronous modes by applying the multidimensional Nyquist criteria for this
particular system. This is a completely a frequency domain analysis for analyzing the stability of
the sub synchronous modes. Then we also have one simple technique or frequency domain
analysis but this technique is approximate as compared to the frequency response analysis
however the most most efficient technique is time domain analysis that is we can perform the
and obtained a time domain solution or response of the system using EMTP program. In this you
can model the three phase system you can model the non-linearities of the system and the
complete transient behavior can be studies applying the EMTP program or today we are using
the EMTDC or PSK program right. These programs are very useful in analyzing the SSR
phenomena and I triple E has given 2 standard models for studying the SSR phenomena, one is
called the first benchmark model and second is the second benchmark model.
Then we have come to a very important aspect that okay, if suppose SSR problem takes places
then we have to we have to take some counter measures to eliminate the SSR problems. The
various techniques have been reported to to mitigate this SSR problem is some of the methods
for or counter measures are static filter, dynamic filter, dynamic stabilizer, excitation system
damper, protective relays and NGH scheme in state the basic, in the static filter the basic
approaches that we put in series with the stator winding a circuit which is having parallel
resonance to a particular particular torsional mode.
So that the stator circuit will not carry frequency or in current of sub synchronous frequency that
is if this parallel network is tuned for some subsynchronous frequency right then no currents can
circulate in the stator circuit of that particular subsynchronous frequency and therefore we can
mitigate this problem. Another is the dynamic filter in this dynamic filter again we put we inject
a voltage in series with the stator circuit which will be equal and opposite to the sub synchronous
voltage which is generated in the stator winding.
So that we nullify the voltage which is developed in the stator winding of sub synchronous
frequency. So that no currents of sub synchronous frequency will flow. This dynamic stabilizer is
basically shunt connector device right and it is a thyristor control by appropriately controlling
this reactors by by switching on and off the thyristors the this SSR problem can also be
overcome. Again excitation system damper, now we have the power system stabilizer similarly
similarly, you can design a stabilizer right to to damp the SSR problem because power system
stabilizer is designed to to take care of low frequency response to 2 hertz and therefore if you
want to use this PSS or excitation system damper then it is designed for those torsional modes of
oscillation right.
The protective relays are designed so that agent when this SSR problem is sensed the relays will
disconnect the system from the supply that is the machine before it gets damage it is
disconnected, this is what is the meaning of the protective relays they will sense the
subsynchronous phenomena and then disconnect the machine. The last scheme is NGH scheme
this NGH scheme , N stands for Narayan G. Hingorani , this is the first person this is the person
who is the father of fec technology and he has suggested scheme for mitigating the SSR problem.
Here, again the approach is very simple across the series capacitor capacitor a thyristor control
resistor is connected that is across that is there is a shunt resistance which is controlled by by
switching on and off the thyristors right and as we know actually that whenever the sub
synchronous frequency currents flow the voltage across the capacitor will become non sinusoidal
and the period of this voltage wave will be less than 180 degrees or more than 180 degrees.
Therefore this NGH scheme senses the period of the voltage way which is developed across this
capacitor and it will short circuit the capacitor in case period is more than this. So that energy
corresponding to the sub synchronous frequencies will dissipated and that that way it will be
ensure that the voltage across the capacitor remains more or less sinusoidal, if the voltage across
the capacitor is sinusoidal then no current will be half sub synchronous frequency this is what is
the philosophy and students have done projects using NGH scheme. The last aspect which we
should address is the impact of network switching disturbances.
Now this is a very important aspect I will just tell you that in any power system this we have
number of switching phenomena, switching phenomena are planned or unplanned, planned
phenomena means suppose I want to switch off transmission line this is a planned okay but
unplanned means the faults occurrence of the faults clearance of the fault re-closure of the lines
right. Now all these all these switching phenomena will cause a sudden change in the electrical
power output of the generator. Suppose there is a transmission line which is supplying some sort
power and if this line is switched off the electrical powerful will become suddenly 0 right and
because of the sudden change in the electrical power output the shaft of the synchronous
generator is subjected to stresses and that causes the loss of fatigue life right and therefore this
subject is a detail subject of important and that is why number of frequency studies have been
conducted conducted. So that the fatigue loss of fatigue life due to switching phenomena is
reduced.
Now next point is the hydro generator torsional characteristic, till now whatsoever I have talked
is the torsional characteristics of steam turbine generator. The rotor of the hydraulic generating
unit consists of turbine runner and a generator rotor, if the unit has a sub-driven exciter there is in
additional rotor mass consequently there are at most 2 torsional modes of frequencies that is
generator, runner that is the hydro turbine and exciter therefore, there cannot be more than two
modes of torsional oscillations. The inertia of the generator rotor is about 10 to 40 times higher
than then than that of the turbine runner or water wheel the torsional natural frequencies lie in the
range of 6 to 26 hertz that is because of vast difference in the in the inertia constants the torsional
frequencies are in this range 6 to 26 hertz for typical, this is a typical range there are no reported
causes of adverse dynamic interaction between the hydro generator rotor and the electrical
network.
Now there are two principal reasons of of no adverse interaction one is the high value of
generator rotor inertia in relation to the inertias of the turbine runner and exciter rotor. This
effectively shields the rotor mechanical system the as a result it is difficult to excite torsional
oscillations through disturbances on the generator that is whenever some disturbance takes place
in the generator these oscillations are not excited. The last one is that the runner is moving in a
water and therefore there exist a good damping is viscous water wheel damping this makes the
inherent damping of the torsional oscillations for 2 hydro units significantly higher than the
steam turbine generators right.

Do you understand this point that in the steam turbine the the turbine blades are moving in the
steam. Here it is moving in water therefore there is lot of viscous friction right and therefore the
damping is very high. Now let me conclude here that today we have studied the the interaction or
a torsional interaction between the system control as excitation system control, governor control.
We also study the SSR phenomena we also study the techniques for analysis of SSR phenomena
and countermeasures for overcoming the SSR problem, we have also studied that this SSR
problem does not exist in hydro turbines. Thank you!
Prof. M. L. Kothari
Lecture - 35
Voltage Stability
Friends, so far we have studied the angle stability of power systems. In the previous
lectures we have studied the basic concepts of ah transient stability, dynamic stability
modeling of the components and various techniques for analyzing the transient stability
and dynamic stability of a system.
Today, we shall study another very important form of stability that is called voltage
stability of a power system. This subject has been the subject of concern over the last two
decades and over the last two decades a lot of research work has been conducted in the
area of voltage stability and the standard resource material for this subject is available in
chapter 14 of this book power system stability and control written by Prabakan, Prabha
Kundur. This chapter 14 deals with the voltage stability of the power system then there is
one more book which is exclusively written to deal with this subject that is power system
voltage stability by Carson W. Taylor, this book is exclusively written and addresses this
voltage stability problem.
The resource material for my presentation is drawn mainly from these two resources and
we will study this voltage stability problem in somewhat detail in the coming lectures.
The voltage stability sub problem, we will actually examine by studying the following
aspects of the subject that is we will examine the basic concepts of voltage stability and
while while trying to understand the concepts we will study the characteristics of the
transmission lines, characteristic of the generators, load characteristics which forms a
very important important aspect that is the load characteristics affect the voltage stability
to a large extent then we will study the characteristics of reactive compensating devices.
The reactive compensating devices are shunt capacitors; regulated shunt compensation
when we talk about regulated shunt compensation under this category comes the static
var compensators and STATCOM. We shall study the characteristic of series capacitors
and we will also study the effect of control series compensation that is thyristors control
series capacitors and the static synchronous series condenser.
Then after studying the characteristic of these devices which affect the voltage stability
we will examine the voltage collapse phenomenon that is how the voltage collapses takes
place in a power system. The next point which we will examine is the classification of
voltage stability that is the voltage stability problem is generally divided into 2 sub
categories, one is called the transient voltage stability problem and one is called small
signal voltage stability problem.
It is also divided into the two another 2 categories, one is called the static voltage stability
problem and another is the dynamic voltage stability problem. Therefore, this subject will
be dealt in detail then for any given system one should know how to analyze the stability
therefore we will examine the techniques which are available for analyzing the voltage
stability of a system.
Now while analyzing we will deal with the modeling requirements of the system then we
will talk about the dynamic analysis that is the static analysis that is as I have told you
that we divide this stability problem, voltage stability problem into 2 main categories, the
static stability and another is the dynamic stability therefore we will examine the static
and dynamic stability analysis techniques then there is a concept of the determination of
shortest distance to instability that is when you are operating operating at a certain point
how far we are from the instability point that is what called the shortest distance then we
will just address the continuation power flow which is used for analyzing the stability
problem.
Then at the end we will examine the techniques which can be used for prevention of
voltage collapse the various measures which can be used for for avoiding the voltage
collapse are put under major headings like system design measures that is when you
design the system you must take care of that the system is operated and so that so that we
are we avoid the instability of the system that is right at the design stage we incorporate
certain features so that the voltage instability problem does not take place.
The various design measures are application of reactive power compensating devices
control or fixed then control of network voltage and generator reactive power output
because generator becomes a very important source of reactive power. So that we should
know how to control what techniques can be used to control the network voltage at
various points and also the generator reactive power output. Then protection plays very
significant role in the power system stability therefore one has to incorporate certain
features which have a coordination between the protection and control.
When I talk about control means control of generators control of shunt and series
compensating devices like this, then we also examine the control of voltage transformer
tap changers then under voltage load shedding for for improving the transient stability of
the system that is angle stability we have some phenomena called the under frequency
load shedding. Similarly, for for avoiding the voltage instability phenomena one can
resort to under voltage load shedding that is when the voltage is less than the certain level
then you can shed some of the load or disconnect some of the load this is called under
voltage load shedding.
Now these these measures are basically the measures which are to be to be incorporated
right at the design stage of the system that is when you design your system you
incorporate these features. So that you can appropriately control the network and the
voltage instability problem can be avoided.
There is there are certain ah measures which can be adopted that you operate the system
you maintain certain stability margin that is the shortest distance to stability that is you
have to maintain the stability margin, you maintain certain spinning reserve and there
should be trained operators to take appropriate action in time. So that instability problem
does not occur.
Now let us look into what is the basic definition of voltage stability, the voltage stability
is concerned voltage stability is concerned with the ability of a power system to maintain
acceptable voltages at all buses, in the system under normal conditions and after being
subjected to disturbance this is very standard definition that is voltage stability is
concerned with the ability of the power system to maintain acceptable voltages at all
buses in the system under normal operating conditions and after being subjected to
disturbance, suppose some sort of disturbance occurs like some generator dropping or
some line dropping right or some fault occurring and fault being cleared and these are
certain disturbances therefore under normal operating condition as well as after the
disturbance has occurred in under both these conditions the power system should have
the ability to maintain acceptable voltage level.
Then we say that this is the voltage stability of the system let me again read it voltage
stability is concerned with the ability of the power system to maintain acceptable voltages
at all buses in the system under normal conditions and after being subjected to
disturbance.
Now a system enters a stable a system enters a state of voltage instability a system enters
a state of voltage instability when the disturbance the disturbance may be in the form of
increase in load demand or change in system condition causes a progressive and
uncontrollable decline of voltage that is that is suppose you have a system right and in
that system suppose there is a there is a increase in load or there is some sort of system
change there is a dropping line or some change in the system takes place, network
topology right then in case suppose this disturbance will cause a progressive and
uncontrollable decline of voltage then this type of situation is the voltage instability
problem that is the voltage will decrease continuously and progressively and we are not
in a position to control and and stop the decrease or hold the decrease.
The main factors causing instability the main factor causing instability is the inability of
the power system to meet the demand of reactive power this is again very important point
the crux of the whole problem is that the power system is not in a position to meet the
reactive power demand that the any power system as we talk about the angle instability
occurs because of because of unbalance between the generation and load.
Similarly, here if there is a there is a situation where the reactive power demanded from
the network is not equal to the reactive power supplied by the network that is the power
system is not in a position to meet the reactive power demand that is the main problem.
Now here let me again mention that under what circumstance this problem occurs
because generally when a power system is operating right then in case a power system
heavily stressed, heavily stressed in the sense that is heavily loaded right then there is a
possibility of voltage instability occurring that is voltage stability problems normally
occur in heavily stressed systems while the disturbance leading to voltage collapse
voltage collapse may be initiated by a variety of causes the underlying problem is an
inherent weakness of the power system as I told you actually the one is that the power
system is not in a position to meet the reactive power demand.
Now reactive power demand meeting is one of the most requirement but at the same time
suppose the system is heavily loaded right and now you increase the load beyond a
certain point then system is not in a position to maintain the required reactive power and
the voltage will decline and may continuously decline or continuously decrease may not
be controllable. In fact actually in some systems the main cause of voltage instability
used to be the long transmission lines right or if you have a well you know interconnected
system and the system is loaded heavily and if you put additional load then the possibility
of voltage collapse may occur. In fact these two terms voltage instability and voltage
collapse are used interchangeably many voltage collapse phenomena have occurred
during the last 3 decades and where the some sort of disturbance occurs and you will find
actually that the voltage at most of the buses in the system will go below certain
acceptable level, it is called the voltage collapse phenomena and it it occurs actually in a
very in a very very different type of manner.
In the sense you start some you you trigger some sort of disturbance voltage drops and
then it continue to drop we are not in a position to bring it back to the if we are in a
position to control the system after disturbance and bring back the voltage to the required
level then it becomes a voltage stability not the voltage instability problem right therefore
to maintain the system voltage stable we have to have fast reactive power controls in the
system. Now let me just sum up what are the basic factors which contributes to voltage
collapse.
Basic factors which contribute to voltage collapse are the strength of transmission
network and power transfer level. When I talk about the strength of the power power
transmission network in the sense that when the transmission network is short and it has a
it has a high power transfer capability right then we say that the transfer power
transmission network is strong its inherent reactance is low and in if this power
transmission network may be strong and it may be supplying huge amount of power. So
that it may be operating near the limiting condition.
Therefore, one of the reason is the strength of the transmission network and the power
transfer level at which you are operating then the next important component is the
generator reactive power or voltage control limits a synchronous generator synchronous
generator is to be always operated operated such that such that its a certain parameters are
within the limits.
For example, the armature current should be within the limits the field current should be
within the limits and therefore the condition or under what how the power generator is to
be operated is always determined by the the operating chart, there is a operating chart of
the generator right. So that in within that chart one has to always operate right. In case
you are operating such that if you supplying the maximum reactive power output then
there is no scope for further increase in the reactive power output.
I shall discuss this generator reactive power voltage control limits these aspects we will
discuss in detail later on, then comes the load characteristics, most of the loads which we
come across are voltage dependent that is the loads are voltage and frequency dependent
and as the system voltage increases the real power and the reactive power absorbed by
the load varies right and therefore how how the real and reactive power absorbed by the
load varies it affects the voltage collapse phenomena right.
The load characteristic the the load characteristics sometime we have constant power
loads so that the even the voltage is fluctuating the p and q does not change there may be
constant impedance loads where the power consumed is proportional to V square right
there may be constant current loads right. Therefore, we have normally the composite
load where part of the load is constant impedance load, part of the load is constant power
load, part of the load is constant current load and therefore the load characteristics affect
the voltage collapse phenomena to a very large extent.
As I have already mentioned the characteristics of the reactive compensation devices I

have told you the reactive compensating devices are basically synchronous condenser, the
fixed capacitors, switch capacitors and continuously control devices like SVC
STATCOM then other series devices like TCSC, triple SC and so on that the
characteristic of these devices also affect a lot. Then, one of the very important devices
which affect the voltage stability is the action of under load tap changer of the
transformers.

In any power system we have the under load tap changers these are called ULTC or
sometimes we call the on load tap changers and the function these on load tap changer is
to maintain the voltage at the load terminals and many a times it is seen depending upon
the load characteristics, characteristic the the ULTC may help in improving the stability
or sometimes it acts in a adverse way we will discuss this effect of ULTC on voltage
collapse phenomena separately.
Now to understand some of the basic features of a certain concepts of voltage instability
or voltage stability. We consider a very simple network let us say there is a voltage
source Es connected to a load whose impedance is ZLD angle phi and this load is supplied
through a transmission line whose impedance is denoted as ZLN angle theta, this is the
impedance of the transmission line network or transmission line, this is the impedance of
the source or you can even look actually that this is Thevenin equivalent of the network
looking into the two terminals at the load is connected because in that case Es becomes
the Thevenin voltage and ZLN becomes the Thevenin impedance.
We will denote the power consumed or the complex power consumed by this load as P R
plus j times QR. Now, we will first derive the expression for the received voltage VR in
terms of the various parameters of the network. The current I which flows in this network
we will, I will denote the current I I call this I bar right this current I bar is written as the
voltage divided by some of the two impedances the line impedance plus the load
impedance right where we denote the line impedance as ZLN angle theta or in the
complex form ZLN, angle ZLN bar similarly, the load impedance we will denote as ZLD bar
as ZLD angle phi and this voltage Es will be taken as reference that is Es angle 0 degrees.
The expression for the magnitude of current, now here this is the expression for
magnitude of current can be written in terms of the source voltage Es and the parameters
of the line and load impedance that is ZLN cos phi plus ZLD cos this is cos theta cos phi
whole square plus ZLN sin theta plus ZLD sin phi whole square this can be derived by
substituting the expression for ZLN, ZLD and then you can write down the expression for
the magnitude of the current okay. Now this expression for this current can be written as
one by square root of F into Es by ZLN where F is devind defined as F is 1 plus ZLD by
ZLN whole square plus 2 times ZLD by ZLN cos of theta minus phi.
Now the voltage at the load end that is the receiving end voltage VR the magnitude only
can be written as the magnitude of the load impedance multiplied by magnitude of
current, ZLD ZLD into I, you are writing only the magnitude. This can be written as one by
square root of F ZLD by ZLN into Es that is you substitute here the expression for I right
you will get the expression for voltage at the load end in this form and the power received
by the load the real power received by the load is VR into I cos phi where, VR is the
voltage across the load I is the current and since I have consider actually the the phase
angle of the load impedance as phi.
So that the received power is written as VR into I into cos phi okay or the PR can be
written in terms of the other parameters like load impedance and this expression F this
the this this F, Es upon ZLN whole square into cos phi, in fact to understand understand
the behavior of the simple network what we are interested is we want to know that how
the this two important parameters one is the voltage across the load and the power which
is derived by the load drawn by the load right are are are affected by the impedance and
other parameters of the network.
See after deriving the expression for receiving end voltage and the power received by the
load, we plot the variation of the receiving end voltage and receiving end power in terms
of a normalized impedance that is we we use actually this quantity ZLD upon ZLN on the x
axis that is the normalized impedance and then plot the variation of the receiving end
voltage VR and we plot the variation of the receiving end power PR and the current I.
Now these three quantities that is current I and the power and this voltage they are again
again again plotted as normalized quantities that is the receiving end voltage VR, this
graph shown, this graph shows, this blue graph shows the variation of receiving end
voltage with the variation of ZLD by ZLN therefore this receiving end voltage is shown as
VR divided by Es that is the what is plotted is the value of receiving end voltage in per
unit considering the system voltage Es as one per unit or the base quantity therefore you
can easily see here that as the as the impedance ZLD that is the impedance of the load
increases increases the the receiving end voltage decreases continuously.
Now the receiving end power is concerned receiving end power is plotted as a ratio of the
actual power PR divided by PRmax that is the maximum power, what is plotted here is PR
divided by PRmax that is the maximum power and you can see here actually that this graph
shows the variation of variation of this quantity PR upon PRmax that is when this power
received becomes maximum then this ratio is equal to 1 therefore this quantity is one here
this is the maximum power reaches that is what we see here is that as the magnitude of
this impedance ZLD increases the receiving end power first increases at a fast rate and
then attains a maximum value then it start decreasing right.
Now we can see here actually that this is the maximum power which is received by the
load and for power to be received maximum load, maximum power the the ratio of ZLD to
ZLN is 1 that is when this the magnitude of load impedance is equal to the magnitude of
the source impedance right then the power transfer to the load becomes maximum, this is
the maximum power transfer theorem also that is if you have actually a source whose
whose internal impedance is equal to ZLN and supplying power to a load whose
impedance is ZLD then the condition for maximum power transfer is that these two
impedances should be equal in magnitude right.
Therefore this is what is happening when this ratio is equal to one the power transfer is
maximum. Now generally the operating condition is on this region of this curve on this
portion of the curve that the power increases when the load impedance is increased right.
Now when we see that from this point onwards when we increase the load impedance the
power decreases right. Therefore, now suppose when you come actually at any of these
points let us see this point in case the load is offering a constant impedance load, if the
impedance of the load is constant then this is also a stable operating point that is you can
operate anywhere on this curve right in case the load is a stable load is a constant
impedance load.
The another characteristic of interest generally is that we plot the PR versus VR that is we
plot actually the receiving end voltage versus the real power rather than plotting the
expression in terms of the load impedance, we plot the the receiving end power as if
receiving end voltage as a function real power received by the load and how how actually
this PR and VR are affected by the reactive power injection that is at the load point
suppose we inject reactive power then how this reactive power is going to affect actually
the variation of VR with, the variation of variation of PR with variation of VR that is as
you increase the value of PR the voltage will vary.
In fact this graphs which you have plotted earlier they are plotted actually for considering
considering the cos phi as .95 and the value of theta is such that tan theta is equal to 10
right the theta is the phase angle of the source impedance or the line impedance right and
we know that that line impedance has a has a large phase angle therefore tan theta is 10.
Now we show another very important actually form of characteristics where we plot the
variation of receiving end voltage VR as a function of PR by PRmax that is earlier the the
voltage power and current were plotted as a function of load impedance. Now we we plot
here the normalized receiving end power that is PR by PRmax this quantity is PR divided by
PRmax that is the maximum power. This is normalized power on this axis we put the
normalized voltage VR by VR, Es,VR divided by Es is a normalized receiving end voltage.
Now this graph is plotted for different power factors. Now here, if you can see here then
this graph or this curve represents the variation of PR and VR there is a relationship
between PR and VR for power factor equal to .9 lagging and I think let me make it more
dark here. These graphs can be shown in the more 1 typical graph is plotted here for
power factor equal to .9 lagging the graph looks like this, this is the graph I will show
only one graph like this that is what we see here is that as this receiving end power
increases, the receiving end voltage decreases and it comes to a point where the receiving
end power increases and the receiving voltage also drops but if you if you try to you can
say increase the power beyond this point you will find actually you will not be in a
position to increase the power and the power will decrease and the voltage will decrease
that is actually from this point to this point it becomes the unstable operating point, we
cannot operate on this point.
Similarly, for different power factor this graph represents for .95 lagging this graphs
represents the relationship between PR and VR for unity power factor. Then the graphs are
plotted for leading power factor condition .95 and 9 lead therefore if you join these
critical points which has the which is the maximum power which can be transmitted
without without losing voltage, without occurring or without coming to the voltage
instability problem right.
Now these points are when we join together these are given the name of locus of critical
points right. Therefore for a simple power system looking at one particular bus that is you
take a system and represent equivalent Thevenin network then we can always plot this
type of ah graphs which are normally called actually the nose curves, they look like a
nose therefore nose curves right and especially here this is the critical point right
therefore, the operating point will always be away from this critical point.
Now suppose you have now a large interconnected system then for this large
interconnected system how to draw similar type of graphs. Now to illustrate this
phenomena what is done is that one has to perform load flow study for each and every
operating condition therefore to illustrate actually for how we plot this type of graphs for
a multi-machine system or for an interconnected network, a typical network considered is
a 39 bus, 9 generator with 1synchronous condenser network. This is a very of course, this
network is not very clearly visible but this network is given in both the reference books
ah this is discussed actually in a by Prabha Kundur in chapter 14 taking this network that
is 39 bus, 9 generator and 1 synchronous condenser. This network has also been
discussed in this book on power system stability by Taylor right.
Now here to illustrate these voltage stability phenomena what is considered is that a
portion of network is examined. Now a portion of network which is examined is I will
just show you here it is a, it is a big network of having actually the 39 buses and some
generators they are they are all generators and this is the synchronous condenser which is
shown here right.
Now in this network ah in this network, this bus is the furthest bus and denoted as bus
number 530 here in this network therefore to plot this the power output versus the versus
the voltage characteristic what is done is that this is considered as one particular area in
which the load is increased at all the load points slowly that is you increase small steps
the loads at all the load points which are given here where about that all the buses which
are enclosed in this area right and all the loads are increased in small steps. Then for each
each of the loading condition a load flow program is run right and for running the load
flow program what is done is that whatsoever is the increase in the load here the output
from the generators are also increased in proportion to their capacity because we know
that real balance is to be maintained if we increase the load the generator output has also
to be increased right.
Therefore to analyze this this system the voltage at the bus number 530 is plotted for
different loading conditions which are considered in this area. This is the area which is
enclosed right and for this each of the loading condition load flow is run and when you
run the load flow you will get the voltage at this bus 530 right.
Now this is the graph which is plotted for this particular network on this axis of this
graph we are marking the total active power in megawatt in area one that is I have
already specified that this this portion of the power system is considered as area one right
therefore in this area the total total active power is varied right and for each active power
loading right the load flow is carried out but when you have varied the active power the
reactive power are also increased in the same ratio. So that the power factor is maintained
at the same level that is at each bus the same power factor is maintained.
Now the graph which is plotted is here you can see that this is one point on the graph
when the loading is something like 80 plus megawatt as you increase the load from 80
megawatt to 200 megawatt right what we see here is that voltage at this bus, the bus
which is considered is the bus number 530 which is expressed in per unit this voltage
drops it comes to point B, it comes to point C. At point C that is if the the total load in
area one comes to the level of level of something like 190 megawatt.
We will find actually that the load flow will not converse it means actually, we have
reached the critical operating condition and beyond that if you try to you can say load the
system system will become unstable, voltage collapse will take place. This is the another
very important characteristic which can be used to understand the voltage stability
phenomena is the relationship between the voltage and the reactive power injected at a
particular bus that gives actually the suppose I inject a reactive power right then how the
voltage at that bus increases bus varies it may increase or decrease right therefore for the
same system which we have studied that is 39 bus, 10 generator system here. You
consider 10 machine system because 9 generators and 1 synchronous condenser that
becomes the 10 machine system.
For this system one can plot the graphs relating relating the reactive power Qi injected at
a particular bus versus the voltage at that bus. Now this these graphs are generated for the
for the different buses the for example bus number 530 which is the most critical bus in
that area one of the of the 39 bus system on this axis we are marking the bus voltage in
per unit and on this axis we are marking the reactive power Qi in MVR right.
Now the voltage at this bus 530 is concerned, you can see here actually that you the
variation of this voltage with the variation of reactive power which is injected is
practically quite constant here and is the for different voltages you can see that this
voltage is practically constant ah it is something like this. Now from this point onwards
what happens is that the voltage start increasing right therefore you can see that from for
something like this .6 and beyond if you if you increase the injected reactive power the
voltage increases it means the if you take the ratio of ratio of voltage to delta V to delta q
then this is a positive quantity if the voltage increases the reactive power increases.
Now if you see this one typical bus number 200 in the same area then the variation is
very very sharp you can easily see that this is the type of variation like this. Now in this
case at the point this is there is lowest point here in this on this graph that is if you are on
this side of the graph, just wait if you are on the this side of the graph what we see here is
that as the reactive power injected increases you can see the here actually the reactive
power, the reactive power. We consider as a reactive power injected is considered as a
positive and reactive power absorbed is considered as a negative right.
Therefore, if when the you can easily see that here when you increase the reactive power
particularly you can just see here then the voltage decreases that because the graph is like
this that is with the increase in the reactive power injected the voltage is dropped
therefore this part of the graph is represents actually the unstable condition. Now from
this lowest point onwards what we see that when you increase the reactive power which
is injected at that particular bus the voltage also increases as a simple rule actually the
rule is that at a particular bus when you injected the reactive power right the voltage
when it increases that shows the stable condition, if the voltage decreases that shows that
we are moving towards the voltage collapse situation.
The similar graphs are plotted actually for different loading condition. This is the graph
which is plotted for the loading condition which is shown here A, because there are three
different loading conditions which are identified in area one and for this loading
condition the graphs are plotted for different buses in area one and these are the shows
the variation it means you can easily see that except that most of the buses there exist
actually a region where the voltage increases with the increase in the reactive power
injection.
Now these graphs are plotted for the second operating point which is shown here. This is
the operating point the heavy loading condition. For this loading condition, for this
loading condition you consider that loading condition and see the sensitivity of the
voltage at a particular bus by by injecting the reactive power right. These graphs show the
variation of the uh voltage bus voltages of the buses 200, 160, 500,10 ,530 as different
buses in the same area right. You can easily see here actually that yes we have some
minimum point and from the lowest point there exist a region where the voltage is going
to increase with the increase in the injected power.
Now these are the graphs which are plotted you can see here they are the graphs which
are plotted for the third operating condition that is the critical point. At this critical point
you can easily see here actually that can you see here that there is a large region or that
the minimum point is occurring at at these points you can see this this is the minimum
here this is the minimum point here while earlier the minimum points were occurring on
this side see this. Suppose this is the this this minimum points are below .6.
Similarly, you can see the minimum point here again on this side but if you see the
minimum point here, they are occurring somewhere actually around .6, this minimum
point is somewhere here. Therefore at this critical point you will find actually that the
voltage sensitivity or actually the the relationship between the Qi and the bus voltage
there is a large portion where the system will become unstable.
There is a certain portion actually we are operating beyond this particular point then yes,
you will be in a position to operate stability but if you are operating below that particular
voltage condition then you will find actually the voltage will voltage will decrease when
you are injecting the reactive power. Now here I will actually like to mention one very
important point what is the effect of on load tap changer on the h voltage stability
phenomena. Now suppose the, we consider a system a very simple system, a system
right, I am drawing only the one line diagram. Now if this system is supplying a load load
and let us say it is a constant impedance load right and this is supplied through a
transformer, here is a transformer connected, here is a here is a transformer which is
connected okay that we show this transformer here and the load is connected here. At this
point we are connecting the load and we are assuming here that the load is a constant
impedance load. This transformer is provided with with under load tap changer right.
Now when you increase the load here, let us say they are increasing the load impedance, I
am sorry not load impedance the you are actually reducing the load impedance to
increase the power or other way we can say that you are increasing the load admittance.
Now if the if the voltage drops here the action of the ULTC is, what is the action of this
ULTC? It will try to maintain the voltage at this point and how does it maintain the
voltage by changing the transformation ratio suppose the the transformation ratio is
determined by this quantity N1 by N2, N1 is the primary turns N2 is the secondary turns.
Now to boost the voltage at on the secondary of this transformer the ULTC will work so
as to reduce N1. So when you reduce N1then what will happen is that voltage at the other
end will increase, the applied voltage is same right such that it is stepping down by a
different ratio. Suppose you take a ratio let us say the N1 is 1000 and N2is say 100 right
then input voltage is step down by a factor of 10. Now suppose instead of 1000 I make it
say 950 therefore it will be step down by a factor of 9.5 it means this through this effect
that the this transformer is trying to maintain the voltage but now this load impedance
when it is reflected to primary side that is the load impedance when it is reflected to
primary side what happens, the ZLD prime will be equal to N1 upon N2 square into ZLD
that is when you refer actually the secondary impedance to the primary side, this is the
impedance which is referred.
Therefore here what has happened is that ZLD decreased so that power consumed was to
increase the power ZLD was decreased but now happens that system will see that the ZLD
is further decreased because N1 has been decreased. Now when Z there therefore the
impedance which is which is seen by the system system is now decreasing when the
impedance decreases the voltage will further decrease when voltage further decreases
ULTC will again try to increase the voltage by reducing this ratio. When it reduces this
ratio again this ZLD is going to be decreased.
Therefore, this becomes a cumulative phenomenon and therefore in this particular

situation where you have a constant impedance load the operation of under load tap
changer is is adversely affecting the voltage stability phenomena. Normally one should
always think actually that yes the uh the ULTC should be in a position to stabilize the
voltage right but here in this particular case incidentally when the you are operating
operating on the other part of the curve that is we have the curve by the volt when you
increase this decrease this impedance the power is increasing but after certain maximum
value the power decreases and voltage decreases, power increases but voltage decreases
right. If it is a constant impedance load and you do not have this ULTC the you will
operate at a stable point but the movement you a ULTC which is trying to trying to boost
the voltage at the load point. You will find that in a sequence that a ULTC tries to
increase the voltage the impedance in further decreases and so on is it understood like
that therefore ULTCs affect or the under load tap changer operation it plays a very
significant role actually in the stability of the power system.
Now let me sum up what we have discussed today I have introduced the the basic
concepts of voltage stability and I have also introduced that what is all what about we are
going to discuss in the voltage stability problem. Then I have given the given the basic
characteristic of a transmission line. We have plotted the variation of voltage current and
power as a function of variation in load impedance then we generated the nose curves for
a multi-machine system and also a technique how to generate the nose curves for the
multi-machine system and I have introduced the basic concept of the voltage reactive
power sensitivity.
Prof. M. L. Kothari
Lecture - 36
Voltage Stability (Contd)
Friends, we continue with the study of voltage stability.

In the previous lesson we have discussed the basic concept of voltage stability and I have
discussed the transmission line characteristic, I have also talked about the VP curves for
understanding the voltage stability problem then I discuss a problem which I still
reiterated again about the large system. Suppose for a large system if you have to plot the
VP curve how do we plot the VP curve for a large system.
Today, we will specifically discuss the generator characteristics, then load characteristic
and discuss the characteristics of reactive compensating devices such as shunt capacitors,
regulated shunt compensation under this regulated shunt compensation we talk about the
SVC and STATCOM. We will also talk about series capacitors we will also discuss
briefly about the regulated series compensation like thyristor control series compensation
and synchronous, static synchronous series compensators and so on. Then we will discuss
the voltage collapse phenomena, how the voltage collapse takes place in the system what
are the sequence of events that take place which may cause the voltage to collapse. Then
the next point which we have to discuss is the classification of voltage stability, voltage
stability analysis then determination of shortest distance to instability the continuation
power flow analysis and ultimately the prevention of voltage collapse.

I intend to cover these topics in two lectures the the material which I am covering today
and in the subsequent lecture is mainly drawn from this book on power system stability
and control by Prabha Kundur. In this book he has discussed one one large system
voltage stability analysis problem the system considered is shown here this system has 39
buses, 39 buses and 9 generators and 1 synchronous condenser, synchronous condenser.
Now for this system first step is that suppose for a particular area, the area which has
been identified is marked here this is the area called as area one this is the area one.
Now here in this area some some buses which are likely to we can say suffer the voltage
instability problem are identified the buses which are very far off from the loads are likes
a 530, this bus number is 530 this is one typical bus which has been identified.
Now for this system the first step which has been done is to plot the VP curves for this
multi-machine system, how do we perform the obtain the VP curve. The approach which
has been considered here is that in this area one we increase the load in small steps that is
at all the buses you increase the load in this area one and and to meet this load, we
correspondingly increase the generation at all the generating stations in proportional to
that to their sizes right and then a load flow is conducted to find out what is the voltage at
this point that is you perform repeated low flow for different loading condition in this
area area to plot a VP curve corresponding to the bus number 530 okay.
Now this figure shows the VP curve obtained for area obtain for bus 530 considering the
load change in area one that is area one has number of buses right and at each bus the
load is increased in small steps and to to meet that load the generations also increase
correspondingly and then load flow is conducted then for each of the loading condition
you find out what is the voltage at this bus 530 which is considered to be the critical bus
and is the found to voltage in stability.
Now this curve is similar to what we have seen for a a single a single bus system or we
can say radial transmission system where load is at only one bus right. Now we can see
here as the load in area one increases the voltage at the bus number 530 decreases and at
this point C we will find actually that beyond this point if you want to increase the load
we are not in a position to increase the load, the load flow will not converse right and
therefore we have generated actually in this in this example the VP curve for a 39 bus
system considering a particular area and a particular bus.
Now while plotting this plotting this VP curve the power factor at the at the buses the
load power factors are assumed to be constant that is when we increasing p you may
increase correspondingly value of q also. Now here there are three points are identified
one operating point A which is called normal operating load which is less than 100 mega
watt and more than something like 80 megawatt around 85 megawatt in the normal load.
Another point b on this curve is identified where the load is more than 160 megawatt, this
point B is very close to the critical point and C is the critical point in the system.
Now to understand the voltage collapse phenomena one very important approach is to
examine the qv curves because we know actually that the voltage is related mainly to the
reactive power injection at that buses right therefore for this system which we have
examined the QV curves are generated.
Now for generating the QV curve the approach is that we go to the system let us say that
we want to generate QV curve for say bus 530 QV curve what we really do here is that
we inject a curtain amount of reactive power at this bus, curtain amount of reactive power
at this bus or other way around we say that we want to maintain curtain voltage at this
bus then compute what is the reactive power injection required that is you assume that
bus number 530 is a PV bus right and perform the load flow analysis and then compute
what is the reactive power required at this bus to maintain the voltage at that level.
Now this exercise has been conducted for four different buses in the system. The buses
identified are bus number 160, bus number 200, bus number 510 and bus number 530
that is for the complete system under the normal operating condition right the the QV
curves are to be generated. Now for generating this QV curves again 3 different loading
conditions are considered, one is corresponding to the point A which is shown here,
another is corresponding to point B and third is corresponding to the critical point C in
the system for for this the point A which is the normal operating condition right the QV
curves plotted are shown here.
Now these QV curves are shown for different buses like this is the QV curve I will just
redraw it more really here this is the QV curve plotted for bus number 200. Now if you
examine this QV curve that is on this side we have put the injected reactive power and on
this side we are putting the bus voltage in per unit. Now you can see here actually that
this QV curve shows as the bus voltage increases the increases initially the injected
reactive power decreases it comes to a point where we reach a minimum point if it is a
minimum reactive power injection required to maintain this voltage at this level then to
maintain the voltage at any other level beyond this point again we have to increase the
injected reactive power.
Therefore we can see this curve then this curve is basically a V curve type thing where
when the voltage is very low we have to inject curtain amount of reactive power if the if
the desired voltage is more reactive power required injection is less right. Now in this
similarly this QV curves have been generated for other buses likes say bus 530, bus 510,
160 all these curves have similar pattern. Now let us examine this curve that is a QV
curve for bus number 200, this is the QV curve.
Now you can see here actually that this is the point where if you find out the slope that is
dq by dv then this dq by dv is 0 that is at this point the slope of this curve is 0. Now dq by
dv 0 means actually here here if you increase the increase the reactive power injection
voltage does not change right. Now if you go from this point onwards then this dq by dv
is positive that is from this point let me call this mark from this point onwards then dq by
dv is positive and therefore the any operation on this part of the curve will be stable
operation in the sense that dq by dv is positive means that when I increase the value of
reactive power injected the there is a increase in the voltage at that bus but now if I see
the left part of this curve then the slope of this curve slope or slope of this slope that is dq
by dv is negative it means if you when you increase the reactive power injection then the
voltage drops right and therefore this portion of the curve of course we write I take this
extended portion like this then this portion is unstable operating point. On this portion,
we cannot operate because if you just increase the the reactive power voltage drop rather
than increasing.
Now this is the different type of q curves which have been shown and as we have see in
our further discussion that this the slope of the slope of this curve or you can say the qv
sensitivity will be very important term or indicator for for analyzing the voltage stability
of a system okay.
(Refer Slide Time: 14:43 min)
Now this shows the QV curves generated for loading condition corresponding to B on the
nose curve or on the VP curve. Now the you can examine you can see that now the
curves or you can see shape of the curves have drastically changed.
Similarly, for the last case that is the operating condition corresponding to the point C
again QV curves are plotted and you can easily see that the curves have completely
changed therefore the what we conclude here is that the QV curves for different buses
depend upon the loading condition on the system and the QV sensitivity is different at
different loading condition.
Now after this we, let us examine the characteristics of various components of the power
system which would affect the voltage stability of the system. In fact the voltage stability
depends upon the complex interaction between the various components of the power
system. Now we start with the characteristic of generators it is very well known actually
that a generator is a source of reactive power, it can generate reactive power, it can
observe reactive power and it can generate real power also. Therefore, now let us to to
study the study the generator characteristics particularly in response to voltage stability
phenomena phenomena we we see that all generators have automatic voltage regulators
and these generators are the generator AVRs are the most important means of voltage
control in a power system.
When we talk about the voltage control in a power system then the devices which are
available for voltage control are automatic voltage regulators of the generators that is by
regulating the excitation of the generators. We control the voltage by tape changing of the
transformers we control the voltage by by reactive power injections. The reactive power
may be injected by synchronous condenser or may be by SVC or may be consider by
shunt compensators or static capacitors or so on but these are the means for regulating the
voltage therefore in generators the AVRs will regulate the voltage under normal
conditions the terminal voltage of generators are maintained constant that is as the
loading condition changes the terminal voltage is maintained constant by the
Now, during conditions of low system voltage because because voltage stability problem
comes when the system voltage becomes low. The reactive power demand on generators
may exceed their field or armature current limits this is very important point I will
explain here that that whenever you increase the reactive power demand on the generator
then we have to increase the excitation current right and therefore in case suppose
reactive power demand becomes such that the you cross the field current limit right.
Now the field current limit is an important limit actually one has one should not violate
the field current limit similarly, the armature current limit should not be violated
therefore any generator is concern. In case in case suppose you come to a situation where
we exceed the field or armature current limits then then we will not be in a position to
regulate the voltage beyond that. Suppose the field current reaches the upper limit it
means now we will not be in a position to regulate the terminal voltage of the generator
current limit reaches its maximum value.

When reactive power output is limited the terminal voltage is no longer maintained
constant, see the in any system the generators are the are very important voltage control
devices but movement actually you you reach the field current limit then that particular
generator will no more be voltage controlling device it will have actually fixed internal
voltage right. In which basically you are you are in a position to maintain a voltage which
is constant behind a synchronous reactance right.
Now to explain this point I will just briefly talk about the operating chart of a large turbo
alternator, all generators generators when they are operated right an operating chart is
used. So that the operator always maintains the the generator operation within the
operating chart.
The typical operating chart of a generator is shown here where, on this on this axis we
have shown the reactive power reactive power supplied by the generator that is this is the
base point I will just highlight some lines. This is our x axis and this is the y axis, the on
x axis we have marked the real power input in megawatts on y axis we mark the reactive
power output in MVR that is this positive part of this axis so the lagging reactive power,
lagging reactive power and on this side it is the leading reactive power the the the region
in which the generator is to operate is confined by some considerations, one consideration
is the the maximum real power output that is the that is that is determined by the turbine
capability.
For example, here this line which is drawn here shows that the the this is the maximum
real power output with a generator cantilever right let me just draw it here. Okay that is
this is the maximum real power output which one can generate then then in the same
mark we have in same actually operating chart this portion of the curve shows, this
portion of the curve shows the stator current limit that is in case you go beyond this side
then the stator current will be more than the rated current then this portion you can see
here this portion of the curve I will show by a different color here shows basically the the
field current limit right, field current limit. Then this part of the curve which is shown
here is is to represent the stability consideration.
In fact this graph is plotted in fact this line, this horizontal line shows actually the
situation where the phase angle between the terminal voltage and the internal voltage will
become 90 degrees and that is we cannot operate when the the power angle is more than
ninety degrees right and this graph is plotted considering a stability margin of 10 percent
and therefore the operating reason is inside this graph or inside this curve which is shown
here which has we consider a four different segments, one segment of the curve
represents the stability stability limit consideration and this is basically we are operating
in the leading power factor condition.
This vertical lines shows actually the real power output condition that is the turbine limit,
this segment q to n is corresponding to the stator current limit and this from n to this point
n to m it represent the field current limit right. The moment you operate within this
within this the generator will be regulating the regulating the terminal voltage but the
moment you reach actually the limiting condition, terminal voltage can no more be
regulated right.
The next important components of the power system which affect the voltage stability is
the loads and their characteristic I have discussed actually in the previous lecture the load
characteristics there are certain loads which which can be consider as a constant p and
constant q loads that is equal to p loads then there are certain loads which which are
which can be characterize as a constant impedance loads, there are certain loads which
can be characterized as constant current loads and in general a load in a power system
will have some component which may behave as a constant impedance load, some
component may behave like constant current load, some may behave as a constant pq
load right.
Therefore, normally the composition of the load is very important and it affects the
affects the voltage stability to a great extent load characteristic and distribution system
voltage control devices are among the key factors influencing the voltage stability that is
we have talked about the load characteristic another important aspect which we have to
understand is the the voltage regulating devices in the distribution system. In any
distribution system we have some voltage regulating devices that is the voltage regulating
devices along with the loads have a significant affect on voltage stability.
Now when I talk about the voltage control devices the one of the most important voltage
control device is the is the under load tape changers, in the distribution system we have
under load tape changers and and as a as and when the terminal voltage or the voltage at
the load point decreases then this control device will tried to bring the voltage to the
required value or try to regulate the voltage. Therefore, the behavior of this voltage
control devices and the load characteristics have a very significant affect on the on the
voltage stability. Loads whose active and reactive components vary with voltage interact
with the transmission characteristics by changing the power flow through the system.
This is very important point actually those you have done some studies of the power
system and if you want to account for the load loads depending upon the voltage that is
load dependent voltage, I am sorry voltage dependent loads.
If you perform load flow analysis, load flow analysis consisting the voltage dependent
loads right then the then the solution of the load flow will be different then a case where
you consider consider the constant pq loads right because in the moment the load at a
particular bus depends upon the applied voltage then then the the net voltage which you
get resultant voltage will depend upon the composite characteristic of the transmission
line and the load characteristic.
The system voltages settle at value determined by composite characteristic of the

transmission system and the load. Now another aspect is when the distribution voltages
remain low for a few minutes which is another interesting aspect by suppose due to some
reason the voltage are at low level right. Then what happens is that we have many loads
which are thermostat control the examples are particularly even the you can say air
conditioning load or the fridge they are all thermostat control.
Now when the system voltage is low right then these devices will remain on the system
for longer time to achieve the same level of temperature right. Therefore, this the when a
distribution voltage remains low for a few minutes thermostats and other load regulation
devices as well as manual controls tend to restore load restore loads means actually the
more loads will be there on the system.
Consequently, more such devices will be operating at any given time and this will cause
further drop in voltage, more the devices are operating in the system obviously the
voltage will drop and if the voltage was low because the low voltage devices more
devices are on the system, this in turn will cause further drop in voltage.
Therefore one has to examine the characteristic of the loads which are thermostatically
control another aspect is that we have large number of induction motor loads and the
induction motors will stall at a voltage which is below 85 to 90 percent of the normal
voltage and suppose induction motor stalls and remains in the system it will cause further
dropping voltage. However, what happens is that most of this induction motors have
magnetically operated contactors so that what happens at that moment the voltage is low
then the contactors will open and trip the load when the contactors open and take the load
the voltage will be restored again right. Therefore, this type of phenomena takes place in
the system that is voltage drops when the loads trip voltage will again go up. For voltage
stability analysis loads need to be modeled accurately this is very important point when I
talk about the voltage stability analysis one has to model the loads that is industrial loads
may be induction motor loads have to be modeled then the thermostatically controlled
loads have to be model.
We have to model actually the operation of the under load tape changers right then in
addition to this one has to model the the behavior of the shunt capacitors which are there
in the distribution system for for reactive power compensation. When the voltage is low
then the reactive power injected by this capacitors which are there in the system will also
becomes low it is proportional to be square of the voltage right therefore this whenever
you examine the voltage stability problem one has to one has to model these loads in
detail this is very important and the the dynamics of this load models is slow.
For example actually the thermostat opens and closes it is not very fast it takes few
seconds and even sometimes minutes right therefore when you model this the voltage
stability phenomena is to be examined over few minutes, did not like angle stability
where we examine in first few seconds, 2, 3, 4 seconds system is stable or not that can be
examined.
The next important device in the power system is the reactive power compensating
devices and one has to model the behavior of the shunt capacitors and the shunt
capacitors are the most inexpensive means of providing reactive power and voltage
support that is although we have today more sophisticated reactive power compensation
devices but the fixed shunt capacitor or series shunt capacitor they are the most
inexpensive devices. When you use this shunt capacitors the one very important
advantage which we get is that the generator spinning reactive power is made free that is
suppose a reactive power demand is made by shunt capacitors naturally the generator the
demand on the generator will be reduced.
So that the reactive power will have some spinning reactive power reserve which can be
which can be used in a very short time okay, that is why the what you this the application
of shunt capacitor will free up a spinning reactive power reserve because generator is
already operating or spinning reactive power reserve.
I have already mentioned that the reactive power generated by a shunt capacitor is
proportional to the square of the voltage and during system conditions of low voltage, a
very interesting part when the system voltage is low we want more of reactive power but
on the other hand what happens if the reactive power injected by the capacitor decrease
that is why during system conditions of low voltage the var support drops and thus
compounding the voltage stability problem, say the voltage becomes low right the
capacitors further you can say reduce actually their share of reactive power and therefore
the problem gets compounded.
The next category of reactive power sources come under the heading regulated shunt
compensation under the regulated shunt compensation we know about the static var
compensators SVCs the another device is the STATCOM today which is available. Then
these devices the difference between these devices is that within certain voltage suppose
you talk about SVC I hope everybody knows about the characteristic of SVC that in the
normal voltage regulating range range the SVC will be in a position to maintain the
voltage constant but the moment.
You go outside this voltage regulating range then the SVC will become a constant a
capacitor a constant capacitor it will behave like a capacitor only just like a shunt
capacitor because it is now no more in a position to regulate actually the reactive power
right and the reactive power consumed by the shunt capacitor is proportional to the
square of the applied voltage. Therefore that limit will become but within the voltage
control limits it maintain the voltage constant this is very important point. Then another
device is particularly the static synchronous condenser the STATCOM and synchronous
condenser.

Now the characteristic of the synchronous condenser and characteristic of STATCOM
are exactly identical, only difference is one is rotating device, one is static device. But the
advantage of this devices is that even when the voltage is low they are capable of
injecting reactive power that is have an internal voltage source and this continue to
supply reactive power down to relatively very low voltage that is if you plot actually the
plot the reactive power versus the voltage characteristic then the state common
synchronous condensers have better reactive power injecting capability and they are fast
also.
The another device which is very important here is the series capacitors, when I talk
about the series capacitors we do not use series capacitors in distribution system but we
have series compensation in transmission network the advantage of the series capacitors
is that the reactive power generated by the series capacitors is proportional to the square
of the current and therefore the series capacitors are self-regulating the reactive power
supplied by series capacitor is proportional to the square of line current and is
independent of bus voltage.
This has a favorable effect on voltage stability whenever you have actually certain
problem of reactive power demand being not made because I have already talk to you
that voltage stability is the problem of insufficient reactive powers available in the
system, in insufficient reactive power sources right and particularly here series capacitors
will be in a position to give you reactive power even the voltage is low but voltage it does
not depend upon depends upon the current which is flowing and they are very helpful
actually in.
Now we will take one example of generating QV curves because as I have mentioned
actually the QV curves play very significant role and I have also mentioned actually that
the the the voltage at a bus will depend upon the interaction between the between the
transmission line characteristic, it will depend upon the load characteristic and it will also
depend upon the reactive power sources which are therefore to understand understand the
basic concepts of voltage stability where I have mentioned that the QV sensitivity is very
important that is dq by dv right therefore here. Let us study how do we generate the
generate the a QV curves for this the example considered is transmission line supplying a
load through a transformer, step down transformer.
Now the voltage V1 this is the this is the infinite bus the voltage of this source is assume
to remain constant equal to one and its phase angle is 0 degrees that is the voltage is one
per and its phase angle is 0 the transmission line considered is a long transmission line
and this example is discussed in the book on power system stability and control by
Kundur in chapter 14 the we will discus the salient points of this example at the load bus
where load is connected to a transformer we are injecting a reactive power Qi, Qi is the
reactive power which injected at this bus our exercise is to plot the QV curves QV curves
for fixed values of real and real power supply real and reactive power supply.
For this particular example the the loadings considered are at different levels like say
1300, 1500, 1700 and 1900 megawatt. These are the four different loading conditions are
considered. For each loading condition the power factor of the load is assumed to be
unity and the cubic curves are generated. Now for generating the cubic curve the
approach is again what I have discussed earlier that is that is to maintain a certain voltage
voltage with the given load you perform the load flow analysis and find out what is the
reactive power injection required right. Now this diagram shows the QV curves plotted
for this example considered.
Now this graph I will again highlight it here and see this graph I will take this graph this
is the QV curve plotted for load equal to 1300 megawatt on this axis. We are marking the
reactive power injected and on this axis we are marking the voltage magnitude of V2 that
is magnitude of the voltage V2 at the load point as we have seen actually the graph is
having the same pattern as we have seen for a multi-machine system.
Similarly, we can plot curves for different loading conditions 1500, 1700, 1900. Now on
the same on the same graph we are plotting the characteristic of the capacitors that is the
reactive power injected by a fixed capacitor that is here what has been considered is that
you take a capacitor whose value is say 300 MVR, it means this what we have assumed is
that there is a capacitor who we can supply a reactive power of three hundred MVR when
the voltage is one per unit right.
Now for this capacitor if we plot the voltage versus the reactive power injected then this
is going to be it means its shape is going to be parabola right because as the voltage
increases the reactive increase proportional the square of the voltage right. Therefore, let
me just take one one particular characteristic. We can see this is the characteristic this is
the QV characteristic of the shunt capacitor. Now when you I have already mentioned
very clearly that for this when I specify that shunt capacitor has a 300 MVR rating means
it it supply three hundred MVR, if the voltage across it is one per unit right therefore if
the voltage deviates from one for example this is the one the voltage deviates from one it
will absorb it will supply more reactive power and it when is less it will supply the
negative power therefore this is the graph okay.
Now the operating point is determined by the intersection of intersection of QV curve of

the load with with QV curve of the shunt capacitor. Now this point at this point if you see
then it is slope of the QV curve of the capacitor then the slope of the QV curve of the
capacitor at this intersection point is less than the slope of the QV curve of the of the load
right that is at this point this point is a stable point reason is that suppose you deviates
slightly from this point. Let us say that you deviate and come to this point it means what
happens what happens the to stay at this point it will require more reactive power while
the shunt capacitor is supplying less reactive power therefore it will move and come back
to A right. However, if you take this point then this is not a stable point right this can be
examined.
Now point here is that when you go from from loading equal to 300, 1300 megawatt to
fifteen hundred megawatt you get the similar operating point when you go to 1700
megawatt then the characteristic of this MVR device which is put as 7, 625 MVR device
this become just transient but if you have a MVR device or actually the shunt capacitor
whose MVR rating is 950 MVR and the load is 1900 hundred megawatt the this is the
intersection but this point is not stable. We can examine it what do we find here is that
that at this point is slope of the slope of the QV curve of the system is less than the slope
of the shunt capacitor that is if you deviate from slightly from this we will find that
voltage will keeping on increasing right and therefore the the point which I want to
highlight here is highlight here is that S by injecting a certain amount of reactive power
we can operate the system at a stable operating point.
However, this reactive power sources of large capacity large capacity right then we will
not be in a position to operate at a at a stable operating condition. Next point we want to
illustrate through the same example is the effectiveness of SVC reserve the effectiveness
of a fixed capacitor.
In this graph we have shown the QV curves of the same system at four different loading
conditions 1300 megawatt, 1500 megawatt, 1700 megawatt and 1900 megawatt for
different loading conditions and one typical SVC is shown here that SVC will be in a
position to maintain the voltage equal to one per unit, you can just see in SVC this is the
SVC characteristic shown.
So for the around this voltage equal to one per unit depending upon slight slope right. We
will find actually that till the reactive power of the SVC is a reactive power limit of the
SVC is reached it will be in a position to maintain the voltage at the same level whether
the load is 1300, if it is 1300 this will be the operating point, if it is 1500 this is the
operating point, 1700 this is the operating point. Now suppose the you reach at this point
B where the SVC reaches its limiting condition, the moment it is said limiting condition it
becomes a fixed capacitor and the capacitor SVC as a capacitor device as this
characteristic as a capacitor.
So that this point B, B in case it is a it has reached the limit when this B again becomes a
unstable operating point because the the slope of the capacitor characteristic or QV
characteristic of the capacitor is more than the slope of the a system QV characteristic
that is why actually the the we have importance of regulated reactive power sources right.
Now I will just take one typical scenario of voltage collapse how the voltage collapse
phenomenon takes place in the system the, first we have to understand what we mean by
voltage collapse?
The meaning of this voltage collapse is voltage collapse is the process by which the
sequence of events accompanying voltage instability leads to low unacceptable voltage
profile in a significant part of the system. This is a very general definition that there may
be certain sequence of events may takes place which will which will lead to unacceptable
voltages in the significant part of the system, then say that the voltage collapse taken
place. In fact voltage instability and voltage collapse are considered to be the synonyms
now let me just explain how the voltage collapse takes place in any particular system.
Suppose a system is operating, now in the case the system is initially operating where its
loading is heavy heavy loading in the sense actually the transmission lines are heavily
loaded. We have limited reactive power source is in the system the generators are fully
loaded or generators are quite away from the source. Now suppose a heavily loaded
system is now subjected to some sort of contingency.
Now example of contingency may be that there is a one heavily loaded transmission line
is tripped the movement this happens happens the other lines will be further loaded and
the reactive power consumed by these lines will increase and this will cause reduction in
voltage. Now the when the reduction in voltage comes the immediate thing which will
happen is happen is that the under load tap changers will come into operation to retain the
voltage and in case in case actually the ULTC in a portion to to stop the voltage drop
voltage will be at a curtain level voltage drop.
Suppose voltage is dropping and ULTC start functioning and the voltage drop is stopped
another thing which happens is that suppose the ULTCs reach their limits then they will
not be in a position to regulate the voltage further. There is a possibility that generators
may also reach the field current limits they are not in a position to further supply the
reactive power, this will cause naturally the reaction in voltage was earlier generators
were in a position to maintain the terminal voltage but the movement actually the reactive
power demand has increased they reach the reactive power demand and therefore the the
moment of reactive power demand limit is reached or reactive power limit of the
generator which is reached then they will no more be regulating the terminal voltage.
Then the effectiveness of the shunt capacitors have reduced because when the voltage is
low shunt capacitors supply less of reactive power and all these phenomena put together
may cause the system to lose the voltage stability. Although some times what happens is
that when the voltage drops then the some of the loads starts consuming less power
because the the power consumed by the loads depends upon the voltage.
When they consume less power right there is a possibility that that the system will
stabilize to a low voltage but at the acceptable level, voltage may be low but with the
acceptable level. But the the sequence of events that take place in a very complex manner
right and in that whole process process the if reactive power sources are weak we have
less of reactive power then voltage may come down to very low level and we say it is a
voltage collapse phenomena.
Now for studying the voltage stability we always divide this stability phenomena as a
large disturbance voltage stability phenomena and small disturbance voltage stability
phenomena the this classification is similar to that for angle stability we just define the
large disturbance voltage stability. The large disturbance voltage stability is concerned
with a systems ability to control voltages following large disturbances such as faults loss
of load or loss of generation or loss of heavy loaded transmission lines right. But they are
called large disturbances similar to our angle stability problem. We will discuss actually
the large disturbance voltage stability analysis in the next lecture. I will talk about the
modeling requirement of the requirements when we study the large stability problem.
The determination of this form of stability requires the examination of dynamic

performance of the system over a period of time sufficient to capture interaction of such
devices as ULTCs and generator field current limiters, large disturbance stability can be
studied by using nonlinear time domain simulation which include proper modeling. In
fact actually the angle stability and voltage stability the mathematical model required is
identical accept that curtain details which are necessary for voltage stability may not be
relevant for angle stability.
Small disturbance voltage stability, the small disturbance voltage stability is concerned
with the systems ability to control voltages following small disturbances such as gradual
changes in the load whenever, we talk about small disturbances means gradual changes in
the load. Idea here is that the system is never never in static condition it is always in
Quasi-static condition loads keeps on changing and when the loads keeps on changing
right then if we are in a position to control the voltage, this small changes in load we say
it is a it is a the system is small signal voltage stable. With this let me sum up what we
have discussed today.
We have primarily discussed 2 important aspects, one is how to generate the QV curves
for a system I have given you the example of generation of QV curves for a large system
we have also plotted the QV curve of a of a machine infinite bus system or a load
connected to a long transmission line. We have also seen the interaction between the QV
characteristics of the system with the QV characteristics of the of the shunt regulating
devices like fixed capacitors and and and actually regulated regulated static var devices. I
have also define the voltage stability as small signal voltage stability and large signal
voltage stability, I have also briefly mentioned about the voltage collapse phenomena
how the voltage collapse takes place in the system we will continue with this in the next
lecture. Thank you!
Prof. M. L. Kothari
Lecture 37
We shall examine today the classification of voltage stability. We will also examine the
techniques for voltage stability analysis, under this voltage stability we may have to study
determination of shortest distance to instability the continuation power flow analysis and
prevention of voltage collapse. Today, we will be addressing the first two points pertaining to
the voltage stability.
In the previous lecture, I have discussed about the classification of voltage stability as a large
disturbance voltage stability and small disturbance voltage stability. This classification is
similar to what is done for for angle stability analysis. The large disturbance voltage stability is
concerned with the systems ability to control voltages following large disturbances such as
faults, loss of load or loss of generation that is whenever actually the the the disturbance is
large in magnitude right then we call this as a large disturbance voltage stability problem.
Now when examining the large disturbance voltage stability of a system, one has to one has to
model and exclusively include the operation of some of the devices such as under load tap
changer generator field current limiters. Generally, what happens is that these devices ULTC
they operate slowly, now suppose the voltage is low voltage is at low level then from that stand
till the first step changing takes place the time delay is quite large right and therefore when you
model the system dynamics or to capture the complete dynamics of the system one has to
include the the dynamics of some of these devices like ULTCs and generator excitation current
limiters and so on. In fact the large disturbance stability can be studied by using non-linear time
domain simulation which include proper modeling.
Now when we talk about proper modeling it means it should include the modeling of all the
devices when we have discussed the angle stability we we, our thrust of modeling is the
modeling of the generator, modeling of excitation system, modeling of controls on the
generator and excitation system but we hardly can say concentrate on modeling of loads in the
detail right but here when we talk about the the voltage stability one has to model the devices in
more detail because this voltage stability is is relatively a slow phenomena and one has to
capture the the dynamics of the slow devices like say ULTCs similarly, the operation of
thermostats right.
The small disturbance voltage stability, the definition of the small disturbance voltage stability
is that small system voltage stability is concerned with a systems ability to control voltages,
systems ability to control voltages following small disturbances such as gradual changes in
load. This form of stability can be effectively studied with steady state approaches that utilizes
linearization of the system dynamic equations at a given point that is we may we make use of
the the linear model of the system for analyzing the small signal stability.
Now when we talk about voltage stability analysis, the analysis of voltage stability for a system
state involves the examination of two aspects. Now here when I say we want to analyze the
voltage stability is means we have to analyze the voltage stability at one particular operating
condition. Suppose the system is operating at a one nominal condition or some condition then
you want to you want to find out find out if a certain type of contingency occurs how the
system is going to behave, whether the system is going to remain stable or not like that it means
basically analysis of voltage stability for a system state for a system state means particular
operating state involves the examination of two aspects, one is proximity to voltage stability or
proximity to voltage not in stability, voltage instability, proximity to voltage instability that is
how how much how close we are or how close is the system to voltage instability this is what is
the meaning, how close that is the.
Suppose I am operating at a particular loading condition and the system becomes voltage
instable at a particular loading condition critical load condition then what is the distance
between these 2 that is called actually the margin also, same approach is used actually in angle
stability also where when we are operating at a certain loading condition right then for a given
condition for a given contingency contingency we can find out what is the what is the critical
clearing time of the system and we know the actual actual fault clearings time.
So that the difference between the critical clearing time and the actual fault clearing time that
gives the margin of stability because that is the margin suppose the critical clearing time for a
given operating condition with a given contingency shall comes out to be, let us say .2 second,
if the actual fault clearing time you say .05 second then we have a margin of .15 second.
Similarly, we are also interested in knowing through the analysis that how close we are to
voltage instability that is what is the distance distance to voltage stability you can say.
Then we are also interested in finding out what is the mechanism of voltage instability, the
second aspect is to understand the mechanism of mechanism of voltage instability that means
how and why does instability occurs what are the what are the parameters or what are the main
main reasons for the voltage instability that is what we are interested in knowing through the
analysis then what are the key factors contributing to voltage instability what are the voltage
weak areas, in a complete system there may be certain portion of the system which will be
more prone to voltage instability then we can take corrective measures to make that system
strong by injecting or by installing more of reactive power sources, what are the measures or
what measures are most effective in improving voltage stability that is what I was mentioning
that the analysis should, when we do the voltage analysis our objective is also to find out what
are the effective measure that can be used for improving voltage stability.
Now when we perform the voltage stability analysis. First, I will discuss the the large
disturbance voltage stability analysis in when you want analysis this voltage stability of the
system then the most effective method is the simulation study that is you have to do time
domain simulation of the system and when you do the time domain simulation requirement of
modeling is important and therefore the models of power system elements that have significant
impact on power system stability are as I have discussed the load characteristic, generator and
their excitation controls, static var systems and their controls, automatic generation control
AGC.
Generally, we do not model the AGC for for transient stability analysis however, since the
voltage stability or voltage collapse phenomena is a slow phenomena and therefore, the the
slow controls are also required to be model the AGC is a slow control automatic generation
control which actually is there to regulate the frequency flows by regulating the the generation
of the system then we may also have to see the protection and controls what protections are
there in the system and what controls are to be modelled along with actually the complete
system.
So that we have complete modeling of the controls protection aspects AGC, then the reactive
power sources loads generators everything has to be properly models. When I talk about loads
means again we have to model the regulating devices in the distribution system, we have to
model the thermostats right and so on when we talk about generators we have to model the over
excitation limits limiters. Similarly, we have to model the over load limits of the generator and
so on.
Now we go to the dynamic analysis the general structure of the system model for voltage
stability analysis is similar to transient stability analysis. This is very general statement that the
mathematical model which we develop for dynamic analysis is similar to what we have put
transient stability analysis, only difference is that we have some more additional detailed
models of the system. The model comprising a set of first order differential equations may be
expressed in the following general form that is when you look into the mathematical model of
the system we have we have a set of differential equations, a set of non-linear differential
equations along with a set of non-linear algebraic equations and this these two sets of equations
have to be solved to get the complete time domain simulation.
The basic equations of the system differential equations the set of first order differential
equations are written as x dot equal to f of x,V where, f is a general non-linear function, x is the
state vector and V is the bus voltage vector. This is the set of algebraic differential equations
the first order non-linear differential equations then we have associated with this a set of
nonlinear algebraic equations which is written in the form i equal to Y into V where, i is
function of state vector and the bus voltage vector. YN is actually the network admittance
matrix and V is the bus matrix. This set of equations are same as the load flow equations which
we come across where we write down i bus is equal to Y into V bus right. Here, also it is same
thing only emphasis is here that yes the current vector is function of the state variables as well
as the voltages.
Now about this YN which is the network admittance matrix. Now here when we model the
system for stability analysis of voltage stability analysis we we do consider the tap changers or
or say under load tap changers therefore when you have the under load tap changers operating
as a function of time then then this YN will also change at each time you have actually the
change in the tap setting that is YN has to be modified. Therefore whatsoever change that takes
place actually in the system have to be accommodated in network admittance matrix. Initial
operating condition is determined as xo Vo where, initial state value state vector is xo and initial
bus voltage vector is Vo. You already specified that x is the state vector of the system, V is the
bus voltage vector, I is the current injection vector and YN is network node admittance matrix
already studied.
Now to illustrate the voltage stability problem, we will consider this system this system has
been discussed by several researchers and it is also discussed by Prabha Kundur in a in this
book power system stability and control in chapter 14 on voltage stability this system has been
the system comprises system comprises of generator 1, generator 2 and generation they are
three generators generator one is connected to this bus 7 through a transformer T1 and a
transmission line there are total 5 transmission lines in parallel ((00:17:36 min)) G2 is
connected through transformer T2 then there are 2 loads, one load is connected at bus 8 through
transformer T4 and another load is connected at bus 11 through a through a ULTC transformer
at bus 11.
Now this system is considered for for understanding the basic phenomena of voltage stability.
Here we are talking about the large disturbance voltage stability and the voltage stability is to
be analyzed by time domain simulation and therefore all the generators have been modelled in
detail dynamics of the generators, dynamic to the excitation system, automatic voltage
regulators and their limiters they are all modelled in detail.
Similarly, the loads at bus eight and bus eleven are also modelled in detail the the tap changer
at bus the tap changer that is under load tap changer on transformer T6 is such that a first tap
changer occurs in a time with the time delay of around 30 seconds and the second tap changer
tap change will take place in 5 seconds and this it has total of 16 plus plus minus 16 steps in tap
changing and the moment actually the all these steps that it reaches the upper limit or lower
limit then the tap changer limits are reached for this ULTC is also model.
Now with this complete modeling the voltage stability analysis is carried out considering 3
different levels of loads that is the loads at bus 8 and bus 11 total put together are are classified
as load level 1, load level 2, load level 3, load level 1is light load then to a further increase load
3 is a further increased load. Now these three load conditions are developed, now here the when
the system is operated under normal conditions a contingency occurs and the contingency
considered here is that out of this 5 transmission line, 1 is line tripped right then the complete
simulation has been carried out and the the response of voltage at bus 8, bus 11, bus 7 are
recorded in addition to the voltages at these places the 4 generator 3, the reactive power
generated by the generator 3 is recorded, the field current of the generator 3 is recorded and the
voltage at the terminal generator three is also recorded. Now let us examine the responses
which are obtained for this particular contingency.
Now here this diagram shows the that this is the response for voltage at bus 11, bus 11 for load
level 1. Now we can see here that initially the system is operating under normal conditions the
moment one line is tripped the loading on other lines increases and it will cause a drop in
voltage. Now with this drop in voltage the first thing which happens is that the generator AVRs
will come into operation and they bring back the voltage to the original level, if you see
actually the responses of voltages at all the three buses that is bus 11 this is the response for
voltage at bus 10 in the level 1.

Similarly, this is the response for voltage at bus 7, I told you that the responses for the voltages
voltages for these 3 buses, 7, 10 and 7, 10 and 11 are recorded okay this is 11 this is 7 and this
is bus number 10, these are recorded right. You can if you could see actually that if loading is
at this level then the voltage is regulated to the initial value only for a short time there is a dip
in voltage.

However, if we examine now the response of the generator 3 for the generator 3 the 3
parameters were recorded, one is the field current in per unit. Now for the load level 1 the field
current the moment there was a dip in voltage the field current this side you will find that when
the voltage is regulated back to the regular value the field current has risen increase but
however the field current remains below the below the limiting value right and therefore there
is no action of any other controller and the voltages is restored back to the initial value.
Now you can look at the reactive power output in MVR, the reactive power output of the
generator three has increased from the initial value to this value because the moment, the
moment one line is taken out additional lines are loaded and therefore additional reactive power
demand is posed on the system and the generator reactive power output has increased but still
this remains below the limiting condition of the generator, things are fine.
You can see actually the a generator terminal voltage in per unit the generator terminal voltage
is regulated to the desired value when the loading condition 1 is there and you have actually the
particular contingency. With the same contingency occurring, if we are now at the load level
two this is a higher level in that case what happens is that the moment one line is out first the
voltage dip take place the generator will regulate the voltage increase its excitation and bring it
to the the voltage is brought up to the initial value.
Now this happens and this remains at that level for a time equal to something like 180 seconds.
Now during this period if you see the voltages what happens at other buses this is the voltage at
bus 10, this is at the lower level then what it was earlier same thing is happing for the voltage at
bus number 7. Now when this load level voltages are there the immediate thing which happens
is that the the generator generator will try to bring the voltages and will try to generate more
and more reactive power.
So far the generator three is concerned with this load level 2, load level 2 you can see here the
field current has increased to this level initially of course it was at this level with load level, one
it has gone up it remains there but at this increase in increased current level since this field
current is more than the rated value what happens is that the overload or over over current field
limiter is there and this the field current will be ramped down. It will made to decrease from
this level therefore this field current is made to come down that is you can easily see see this is
the level at 2 and it is brought down it is brought down to the level. So that is below the
permissible limit.
Now during this period when the voltage is coming down when when the field current is
brought down voltage comes down now in during this period the OLTC or you can say under
load tap changers will start taking into coming into action because they ULTC will try to
restore the voltage at the load points. Now when the ULTC tries to restore the voltage right you
will find actually that the more of reactive power flow on the lines because the voltage is
restored load will draw more reactive power and therefore more reactive power flows and with
this more reactive flowing on the line, you will find actually the that the voltage will further
drop.
When voltage drops ULTC will again try to boost it up if you see this, this response is here at
this point somewhere at one hundred eighty seconds or so the ULTC comes into operation and
at every step when it is trying to raise the voltage, the voltage is coming down and at this point
it has reached the upper limit that is the limiting condition of the tap changer has reached. So
that the voltage now settles down to a low level of the order of .95 same things happens for the
voltages other buses you can see the voltage at bus 10 this is 11 at this particular point has
started happening this is the action of ULTC okay.

Now we can see the what is happened happened to the reactive power output of the generator
the reactive power output has increased from this level, this level because this this is the
reactive power output at load level 2, the moment actually one line is out this it has increased.
Now this is now has gone beyond the limit it remains there and then it is stepped down or it is
ramped down comes down like this okay now let us look at the what happens at load the load
level 3, now at load level 3 if you just see here this is the voltage at bus 11.
At load level three the moment actually the fault occurs not fault but the contingency takes
place, line is taken out the voltage dip takes place this voltage dip is more here because the
initial load is more right. The first, the first thing which happens is that the AVR will try to
regulate the voltage and bring to the desired value but in this case what happens is that in a very
short time the AVR has thrust its field excitation limit and therefore the AVR will no more
regulate the voltage then immediately the ULTC comes into action and tries to regulate the
voltage but in every step when ULTC is trying to regulate the voltage instead of voltage being
regulated it it it drops down because for the every time the voltage is brought up more reactive
power demand is there and that causes further voltage drop.
So you can see here actually the voltage drops to a very low value of the order of something
like this .83 plus or .838 or so 8.845 or so something like that you can see the level brought
down to this level.
Similarly, if you see the voltage at bus 10 its happening like this comes down this is much less
than .81 the voltage at bus 7 is going down to this level. You can just see that is .92 or so the
voltage at bus 7 is going down to this level you can just see that is .92 or so. If we examine the
field current field current of the generator three then the moment this generator has come into
action the field current is increased to high level then it is going to be ramped down and it
comes to this level.

Now during this period the ULTC comes into action and the voltage, the reactive power
demand, reactive power demand you can easily see it goes up very high then it is ramped down
therefore, what we see here is actually that the through this example what I wanted to illustrate
is that whenever some contingency occurs in the system system gets overloaded. The first
action is that the AVRs come into action tries to bring the voltage to the desired value however,
in that process the field current may go beyond the required limit.

Similarly the reactive power generated by the generator may go up so that the over excitation
limiters will come into action and will ramp down the excitation from that higher value to the
safe value. Similarly, reactive power will come down when the voltages come down the
ULTCs will start functioning to restore the voltage but in that process instead of succeeding in
restoring the voltage, the voltage keeps on going down till the ULTC settled down to the
limiting condition. This is the type of phenomena which has been recorded in this particular
example. The next thing which we will discuss is the static analysis you know so far the
dynamic analysis is concerned, the dynamic analysis is very useful but its very time consuming
it requires actually the large number of time domain simulations because for for each operating
condition for the each topology and and the for each type of contingency the behaviour of the
system is going to be different.
Therefore, one has to do large number of such studies however the the dynamic study which is
performed gives you clear picture about the voltage stability causes what were the key factors
which effected the voltage instability and so on. However, it does not give the information
about how much how much my system is stable, how far I am from the from the voltage
instability that is the distance from voltage instability is not is not can say determined through
dynamic analysis it gives the complete information system is stable or not this is how the
system collapses complete information you get it but not this quantitative information that yes
what is the additional load which I can put on the system without losing voltage stability right.
The static analysis s simpler and give some more insight into the system the static analysis is
possible only because because the dynamics of the system is slow and therefore for when you
perform the static analysis we you perform or we capture the snap shots over a certain time
frame and during this time frame we assume actually that the system is practically in the steady
state condition and therefore, we can apply apply static analysis tools for determining the
system stability. The two important techniques which have been developed over the years one
is called V-Q sensitivity analysis, another is called Q-V model analysis. I have already
discussed actually the significance of the Q-V curves, we have also seen actually the
importance of slope of these curves for stability and therefore this V-Q sensitivity is on the
similar lines and we will study the V-Q sensitivity first.
To analyze this V-Q sensitivity, we develop the steady state model of the system similar to
what we do for load flow analysis using Newton Raphson technique where you can down write
down this as delta P delta, Q vector is equal to a Jacobian into error vector see delta theta and
delta V. Now here, we will not call it as error vector as you thing we will say this is the
deviation in voltage, this is the deviation in angle, this is the deviation in real power, this is
deviation in reactive power therefore, this equation which is which is similar to our load flow
equation is used for performing the V-Q sensitivity analysis.
Here this delta P is the incremental change in bus real power, delta Q is the incremental change
in bus reactive power, delta theta is the incremental change in bus voltage angle and delta V is
the incremental change in bus voltage magnitude. Now here to obtain or to obtain this V-Q
sensitivity what we do is we assume the the deviation in real power at the bus equal to zero that
is we make delta P0. Now once we make delta P0 then by matrix simplification we can
eliminate this term delta theta that is you can we can write down the expression for delta Q in
terms of delta V only.
Now this approach is similar to the bus elimination technique which we make use of in the fault
analysis, when you develop the Z bus matrix and so on the similar approach is required that is
we can write down here delta Q in terms of delta V and this matrix J r this matrix Jr is obtained
in terms of the 4 elements of the Jacobian matrix Jqv Jq theta Jp theta Jpv I will advise these
people to derive this expression and satisfy that we can obtain the reduced Jacobian matrix Jr, Jr
as Jqv minus Jq theta Jp theta inverse Jpv.
Now using this equation we can write down delta V as J r inverse delta Q it means this equation
relates the the incremental change in bus voltages to incremental change in bus reactive power
through this matrix Jr inverse this Jr inverse is a full matrix right and therefore I can find out the
sensitivity of ith voltage or sensitivity of voltage at ith bus with with the reactive power change
or incremental change in reactive power at ith bus in terms of diagonal elements of Jr inverse
because Jr inverse is a full matrix but I am interested in finding out the the, let us say delta Vi
versus a delta VQ not delta VQ, a delta Qi that is incremental change in voltage at bus i with
incremental change in reactive power at ith bus.
This ratio is nothing is but the the diagonal ith diagonal element of the Jr inverse right and thus
I can find out the the V-Q sensitivity right. Now here our requirement is for this V-Q sensitivity
basically you know the QV curves which we have generated from the QV curves one can find
out the V-Q sensitivity but however to generate those V-Q curves is a cumbersome process but
now we get the information about this the voltage V-Q sensitivity by through this matrix Jr
inverse straight away.
Now for the system is stable system is stable when this sensitivity is positive and in case it is
low it is more stable, in the sense that in case suppose this this quantity is very large that is if
the sensitivity is very large the meaning of this is a small change in mall incremental change in
reactive power at bus I will make a large change in voltage at bus i that will be instability
therefore our requirement is it should be positive and small for the instability in case this
quantity is negative that indicates that system is unstable.
Now generally generally instead of taking the inverse of this matrix that is J r inverse one can
find out this information right from this matrix itself because this is delta Q, Jr delta V therefore
we get the information through the elements of Jr through the elements of Jr how how these
how this change in delta Q or change in Q is related to change in V or incremental change in Q
is related to incremental change in V but whichever way you determine one can find out the V-
Q sensitivity and using this V-Q sensitivity one can make out the the voltage instability or find
out voltage stability. Another technique which has been developed is called Q-V model
analysis.
Now in this Q-V model analysis, we we make use of the information about the about the Eigen
values and Eigen vectors of the Jr matrix that is this matrix Jr this is the reduced Jacobian
matrix this Jr is used to further obtain the information about the voltage instability using Q-V
model analysis. The approach is very simple this matrix Jr can be written as as matrix zeta zeta
is a matrix into lambda lambda is also a matrix into eta this zeta is the right. Eigen vector
matrix of Jr that is what is do we that take this matrix Jr then find out the Eigen values and for
each Eigen value you find out the Eigen vector right Eigen vector and since this matrix will
have n Eigen values if it is n by n matrix then there will be n Eigen vectors therefore, you you
obtain this right Eigen vector matrix making use of n Eigen vectors that is these Eigen vectors
will become the columns of this eta matrix, this zeta matrix.
Okay this is the similarly, is eta is the left Eigen vector matrix of Jr we one knows about what
we left Eigen vector and right Eigen vector. This lambda is the diagonal Eigen value matrix of
Jr that is Jr suppose has a n Eigen values then this lambda is a diagonal matrix the elements on
the diagonal are Eigen values of the matrix Jr. Now at this point I want to mention here that in
case you neglect the resistance of the system network and the the if the and system is
symmetrical then then these Eigen values come out to be real.
Generally you will find actually that we have a real complex conjugate Eigen values in the
system but for this particular system, power system where if you the resistances are negligible
as compared to the reactance of the network system, if you neglect the resistance right and you
will if you do not consider again the the tap changers position because the tap changers will
make this matrix or bus admittance matrix unsymmetrical like that but however that effect is
small.
Now this is a real the the this Eigen values are real and therefore actually the diagonal Eigen
value matrix of JR is written as lambda. Now our interest is again to find out the the how
incremental change in Q is related to the incremental change in V so that we make use of this
expression. Now here Jr inverse can be written as eta lambda inverse I am sorry, zeta lambda
inverse into eta.
Now this is possible because because this relationship exists between this right Eigen vector
matrix and left Eigen vector matrix that is the zeta inverse is equal to eta and because of this
relationship that is inverse of Jr can be written as zeta lambda inverse into eta one can verify
this relationship. Now you substitute the value of Jr inverse in our original equation that is delta
V is equal to Jr inverse delta Q therefore this Jr inverse I am substituting this expression.
So when I substitute for Jr inverse zeta lambda inverse eta we get this relationship delta V equal
to zeta lambda inverse eta into delta Q this is very important relationship how it is important let
us just look into it we can further simplify this. We multiply both the sides of this equation by
eta you multiply both sides by eta then you will get eta delta V equal to lambda inverse eta
delta Q why because because eta into zeta inverse is one identity matrix because that is you can
say eta into zeta this product is an identity matrix right okay.
So that we have this relationship which I have shown here that eta into delta V is equal to
lambda inverse eta into delta Q that is when we talk about the Q model Q-V analysis this is the
most important expression. Now in this expression we see here actually what is eta eta is
actually the left Eigen vector matrix is a matrix suppose, that when this matrix is multiplied
with this vector we will get a vector this vector we denote as small v and this vector will be
called actually the vector of modal voltage variations this delta v is incremental change in
voltage but this product is called modal voltage variations. In fact actually here this v is a
vector right and this vector contains the information information or the the of all modes in the
system all the modes are represented right Q is represented as eta into delta Q which is called as
modal reactive power variations that is instead of having the vector of vector of reactive powers
incremental change in reactive powers.
Now we have got here modal voltage variations is a vector of modal voltage variations this is a
modal reactive power variations and therefore we get this information here very interesting
expression. Now since this lambda is a diagonal matrix where diagonal elements are the Eigen
values of Jr these are the Eigen values of Jr and therefore if I write down here the ith component
of the modal voltage variation then this ith component can be written as one by lambda i into Qi
where Qi is the Qi is the ith ith reactive power ith modal reactive power okay.
Now what we see here is that this Vi is inversely proportional to lambda i right in the sense
actually here we get a relationship which shows the the relation of the modal voltage vector,
modal voltage of ith bus in terms of modal reactive power change at ith bus in terms of Eigen
value of the Jr matrix right. Now here suppose lambda i is a positive what does it show positive
lambda i shows that that the modal voltage variation vector is along with the modal reactive
power variation.
In the sense that where the reactive power varies in a positive direction then the modal vector
will also vary in positive direction and that shows that system is a stable but suppose if lambda
i is 0, 0 that is the case of instability because that means a very small because there is a this
becomes Vi becomes infinite when Qi is small and therefore if lambda i is 0 it is unstable while
if lambda i is positive positive instable and in case lambda i is small positive it is more stable if
large positive it means lambda i zero is unstable lambda is positive with stable right and if
lambda i is negative system is unstable.
Now here I will just give you the relationship between V-Q sensitivity and lambda Eigen value
the V ,V the the relationship is like this delta Vk one can establish this delta Vk divided by delta
Qk is equal to summation over I, zeta ki, eta ik divided by lambda i. In fact this is this gives you
a relationship between the 2 approaches, one approach is based upon V-Q sensitivity another
approach is based upon the Eigen value right the modal analysis is based upon the Eigen value
and as I have told is Eigen value is positive you can see that this this V-Q sensitivity is
inversely proportional to lambda i right.
Now one simple example we can consider here the example is that you have a strong system
infinite bus system which is supplying a load through a transmission line to a load through to a
transformer. At this bus you have a reactive power source where I am injecting the reactive
power Qi and this becomes a simple system therefore this voltage is fixed V1 this is equal to 1
per unit, its phase angle is 0 and only quantity which is varying is V2 right. Now this for this
simple problem this is solved in example 3 of the example 14.3 in the book power system
stability and control by Prabha Kundur.

I will give the results what has been done is that for a particular load load is about 1500
megawatt the V-Q sensitivity analysis and Q-V modal analysis have been applied. Now in this
particular case this Jr becomes simply a one by one matrix, Jr becomes one by one matrix and
therefore the problem becomes a simpler one however, actually the the results are very
interesting the results are shown here. The studies are carried out for different values of reactive
power injected at bus 2. Suppose it is 500 megawatt MVR is 400, 306and 305.9 just to show
the difference as we go from 306 to 305.9 for this the voltage is computed V2 in fact for each
value of the injected reactive power there are 2 solutions for the voltages, one is 1.024, another
is .671 the lower one there are two solutions always because the it is a nose curve the phase
angle is computed.
Now you can see here in this diagram this is a lambda i not very clearly shown here. This is
lambda and this is dV by dQ right. The lambda for this 500 MVR injected is 17.03 and the
corresponding dV by dQ is small of the order of .059 although when you talk about the other
part other solution lower solution system is unstable because lambda is negative and dV by dQ is
also negative right.
Now through this example what we see here is as the injected reactive power decreases from
500 to 306 there is a drastic change in the value of lambda that is 17 to .52 the larger the value
of lambda larger the value of lambda right system is more stable it is similar to a damping ratio
in a small signal stability analysis right. Now this 306 and 305.6, if you see here the lambda is
changed from .52 to .02 by small change in Q by point one MVR right like this.
Now with this let me sum up what we have discussed today. We have discussed the 2 important
aspects, one is the classification of a voltage stability phenomena, then we have discussed the
techniques for analyzing the voltage stability of the system. I have illustrated the application of
application of a small signal stability analysis through an example and we have seen that how
how actually the system functions and how the system voltage collapses then we have studied
the two two static voltage stability analysis tools, one is the V-Q sensitivity another is modal
Qu, QV approach a simple example we have illustrated the application of these techniques to a
simple system. Thank you!
Prof. M. L. Kothari
Lecture - 38

Today, we shall studythe determination of shortest distance to instability, the continuation power
flow analysis and at the end we will study the techniques which can be used to prevent the
voltage collapse or voltage instability. Now let us try to understand briefly what do we mean by
the shortest distance to instability.
The purpose of this study is to determine the smallest stability margin that is whenever you are
operating at a certain loading condition then we would like to know that how much increase in
load can be made made so that the system becomes unstable that means how far we are from the
from the instability but our interest is to find out the shortest distance.
In fact actually suppose you are operating at a certain loading condition then then the load can
vary in a variety of manners right. Then for each particular combination of loading if you in
stretch the load in a particular direction then you will reach even particular point where the
system becomes unstable. We want to know what is the direction in which when we stretch the
loading so that the instability occurs with the minimum distance.
Now to understand this basic point, we consider here a simple radial system radial system. In this
radial system you have one source, one transmission line and say let us say the load here, we will
call it P, I am considering there simplest system but any multi machine system and large system
the there are several loads connected at various buses and and therefore whenever you want you
can say increase the loading, the loading can be increased in variety of manners right. But we
know actually that when you increase the load at a one particular loading condition condition the
system becomes voltage instability occurs. So system becomes voltage instable and at that
loading condition if you find out the Jacobian of the matrix then that Jacobian becomes singular
right.
Now here for this simple radial system we plot here actually in the PQ plane. This is the load
reactive this is the load in per unit that is the active power P on this side is the reactive power
that is you put P on this side P here and the reactive power Q on this axis. Now we can find out
one one surface if you any point on that surface surface will lead to if you operate at that point
then it will lead to voltage instability. In case it is a large system with number of buses right that
surface comes out to be a hyper surface right and in this particular case the the simple radial
system right the we get actually a curve right.
Now this is the curve, now suppose you are operating initially at this point called Po Qo that is
initial loading condition is denoted as Po Qo and suppose you increase the load P and Q in at
random. Let us say you increase in this direction, so that you come and reach the this this surface
as or actually the the curve as this you call as surface S or curve S right. Now at this point the
system becomes unstable.
Now suppose now the question is actually that what is the shortest distance to instability to find
out the shortest distance, obviously in this particular case the shortest distance one can find out
first you find out the normal at this point one to the surface, let us call this is the normal at this
point call it say eta 1 then the again go back to the initial loading condition and stretch the load in
parallel or in the direction of this normal that is second time when you are stretching you are
stretching in this direction this is parallel to this you reach this .2.
Then once you reach the point two again you find out the normal at this point eta 2 and then
again go back to the initial condition and stretch the loading in the direction parallel to the
normal. Then you will lend to a point where the distance between the operating initial operating
condition and the new operating condition will be minimum that is one can find out, suppose you
you represent actually the initial operating condition by as a vector rho, a parameter vector rho.
Now this parameter vector rho is the initial operating condition and let us say rho psi is the final
operating condition then our interest is to find out this distance that is the rho psi minus rho norm
norm of this quantity. This will be minimum right and the for a simple two machine system not
two machine systems for a single radial system I have illustrated that you start from initial
condition stretch the loading in any direction. Here, P is increasing while Q is decreasing like
that and you reach to a new point then at this point one the see this point one lies on this surface
this is the which represent the unstable operating condition then the then the normal which you
find out at this point right. Now you stretch the in the direction of the normal and therefore you
keep keep on doing this till you get the minimum distance therefore in the multi machine system.
We denote actually the the loading condition loading condition in terms of initial P and Q
through a parameter and we find out the final as rho star and in case the rho star is such a loading
condition.

So that the moment you are operating this rho star lies on the hyper surface which is which
represent the unstable operating condition and and that our objective is to find out the the norm
of this quantity that rho naught minus, rho star minus rho naught norm of this line and this rho
minimum and this process is a documented in various papers, how to find out for a multi
machine or a large system this node, the purpose of this obtaining the shortest distance is to to
note that at what stability margin we are operating because whenever you are operating system
you would like to maintain certain stability margin all the time.The next point we will study is
the continuation power flow analysis. The continuation power flow analysis is very important
and is a very important tool which has been developed for for obtaining the critical operating
conditionin. In fact as we know that if you operate if you operate near the critical point or at the
critical point then the Jacobian becomes singular that is the Jacobian matrix, the Jacobian matrix
becomes singular at the voltage stability limit.
Consequently the conventional power flow algorithms are prone to convergence problems that is
the the moment the Jacobian becomes singular right the convergence problems arises and
therefore if you try to plot actually the um we call PV curve right then we will be in a position to
get certain points on the PV curve by conventional load flow analysis but the moment you reach
closer to the instability poin, you will find the Jacobian becomes singular.
Therefore to solve this problem the continuous power flow algorithm has been developed that is
the continuous power flow analysis overcome this problem by reformulating the power flow
equations. So that they remain well conditioned at all possible loading conditions this allows the
solution of power flow problem for a stable as well as unstable equilibrium points that is when
you look at the at the PV curve right then the upper portion of this PV curve represent this stable
operating points or equilibrium points but the lower portion right, they are the unstable. Now this
these points when you cannot compute or we cannot get by conventional power flow equations
or conventional power flow algorithm.
Now this continuation power flow approach has been well documented and the most important
research paper where this continuation power flow has been discussed is written by V. Ajjarapu
and C. Christy. The continuous power flow, the continuation power flow a tool for steady state
voltage stability analysis. IEEE PICA conference proceedings page numbers 304 to 311, it
appeared in may 1991 these very standard reference paper for understanding the continuation
power flow algorithm. We will briefly briefly study the salient features of this continuation
power flow algorithm basic principle.

The continuation power flow analysis uses uses iterative process involving predictor and
corrector steps as we will discuss just now that is the if the basic principle in this continuation
power flow is that the two steps. First is suppose we are operating at a initial point, first we find
out through the predictor step right the and then from that predictor step we go to the into the
character step that two steps involved and these steps are followed attentively to get the solution.
To understand basic concept let us look at this graph, this graph shows actually the the load on
this axis and the bus voltage on this axis now this is a PV curve which is plotted in plotted part of
it can be plotted by repeated application of load flow analysis right. Now suppose we are
operating at this initial operating condition at A, the first step will be we we find out at this
operating condition A a tangent vector that is there is a tangent vector which we find out, for
example this is the tangent vector at this operating condition. Then the next step is persue you
find out the tangent vector we want we want to find out what is the step length therefore the step
length for example in this case shown is A to B therefore the step length is also a important
decision to be taken.
Now once you come to this point B, next step is at this from this point we come to the actual
solution C that is called character step from B to C that is at this if you see here actually that this
graph which is the firm graph that shows the actual solution that of this is the exact solution, this
graph shows the exact solution I will just draw it with the different color to make it more clear
that is this blue graph is I have shown here is the exact solution that is we want to come from A
to C but how do we come from A to C the step is at A you find out a tangent vector then you
select a judiciously the step length then from this point B is not on the solution exact solution. At
point B you apply character step and come to point C right.
Now this this approach is to be applied to find out the solution for different points on this graph
here it is possible to get solution so that we can get the critical point. For example this is the
critical point and then you can get points which are below the critical point also that is the
complete part of the complete you can say PV curve can be obtained for a particular bus in the
multi machine system.
Now let us study actually the algorithm the mathematical formulation the basic equations are
similar to standard power flow analysis except that the increase is in load is added as a
parameter. Generally, in a conventional power flow what we really do is that loads are known to
us and we solve the problem. Now here what we do is that increase in load is added as a
parameter that is increase in load is added as a parameter this is one difference here the
reformulated power flow equations with provision of increasing generations as the load is
increased may be expressed as.
Now the point here is that is any system when I increase in the load right then I may have to
increase the generations also right. So that the load generation match is achieved right therefore
whenever you increase the load in a particular fashion then you have to increase the generations
also that is the what it writes here is the you reformulate the power flow equations with provision
of increasing generations as the load is increased. The the load flow equations can be formulated
in this fashion F theta V equal to lambda times K where, F is a non-linear function of theta and V
it means F theta V is represents actually F theta V equal to lambda K, it represents the set of non-
linear algebraic equations okay. Here here F is a function in fact general F is a function and this
equation is a is actually a vector equation a large number of equations are there in vector
equation.
Here lambda is called the load parameter, theta is the vector of bus voltage angles, V is the vector
of bus voltage magnitudes and K is the vector representing percent load change at each bus that
is K is the vector representing percent load change at each bus. In fact here here when you look
at the equations what you find here is that is lambda into K. We can fix actually some percent
load changes and multiply by a parameter lambda. The magnitude of lambda varies from 0 to
some critical value right. It means this the amount of load change which we put here can is
adjustable.
The above set of non-linear equations is solved by specifying a value for lambda such that
lambda is between 0 and lambda critical where, lambda 0 stands for base load condition, when
lambda equal to 0 it means we are not giving any increment in the loading condition and lambda
equal to lambda critical deposit at the critical load that is they may be value of lambda is equal to
lambda critical then we have reached the critical loading condition.
The equations which we have reforming for put here that is the equations which we are put as F
of theta V equal to lambda K right. This equation can be reformulated or re arranged in the form
F F theta V lambda equal to 0, K is a constant lambda is a variable parameter right these are the
set of non linear algebraic equations. Now here to apply the continuation load flow algorithm as
I have discussed they are two steps, one is called predictor step, another is called corrector step.
First, let us examine the predictor step once a base solution has been found lambda equal to 0.
A prediction, a prediction of the next solution can be made by taking an appropriately sized step
in the direction tangent to the solution path that is our interest is to find out find out actually the
predictor step it is, let us find out next solution, next solution find out next solution a prediction
of the next solution can be made it is a prediction by taking an appropriate sized step in the
direction tangent to the solution path that is we start one initial condition, find out the tangent
that is the direction that is the first task in predictor process is to calculate the tangent vector. It is
the what we have discussed here the tangent vector is very important in and therefore deep in the
predictor step the step is how to calculate the tangent vector the approach is very straight
forward. We we have obtained the differential of this equation that is d of f theta v lambda, this
function can be written as f naught d theta plus fv dv plus f lambda d lambda, this is equal to 0.
Now here this f theta, fv and f lambda these are the the matrices of partial derivatives and this
equation can be arranged in this form f theta, fv, f lambda that is this matrices of partial
derivatives are arranged and into this vector d theta, dv, d lambda equal to 0. Since actually we
have theta the dimensional of theta is equal to number of buses in the system right similarly or
you can say actually d theta and dv right these are n vectors and lambda is of course one scalar
quantity.

Therefore, this vector d theta, dv and d lambda this vector is given the name tangent vector that
is tangent vector is defined as d theta, dv, d lambda. Our interest is to find out this tangent vector
to find out the tangent vector we have just now seen what we really do here is this is the case but
now the question is how to solve this problem. Now if you look here here so far the partial
derivatives are concerned right this there are the number of unknowns I have now increased by
one lambda is the additional parameter which we have added and therefore there is a necessity to
add one more equation here.
So that you can solve this for this tangent vector that is one of the components of tangent vector
is set equal to plus 1 or minus 1 that is we have first our step is how to solve this problem how to
get the tangent vector right therefore the approach is that first we we add a rho vector ek ek to the
set of differential equations or do this Jacobian matrix ek. This vector ek accept the k th element
all other elements are 0 this is a rho vector and and our interest is that we want to make one of
the component of the tangent vector equal to plus 1 or minus 1.
So that when this rho vector accept one quantity others are all 0 certain when you multiply this
you will find actually that the delta lambda will come out to be equal to plus 1 that is here we are
making actually this ek ek in such a fashion so that the you get d lambda. One of the elements of
one of the elements of the tangent vector equal to one, for example if you want to make one of
the elements of voltage component right that is say let us say the voltage of bus 1 dv1 or dv2 any
one or similarly phase phasing allot any of the bus voltages that is in this you can make one of
the elements equal to 1, it may plus 1 or minus 1, plus 1 is corresponding when you are operating
on the upper part of the PV curve and minus 1, when you are operating on the lower part right.
So in this reformulated equation we have now, we have now a situation where where we can
compute this tangent vector okay.
Now the this imposes a non-zero norm on the tangent vector is that by by this approach by
adding this rho vector ek and making one of the elements as plus minus 1 right what was have
achieved is that two things which you have achieved first thing is that this rho vector right. It it
becomes a non-zero norm tangent vector that if a find a find in norm of this tangent vector then it
will come out to be non-zero, it will not become zero right. Second is guarantees the augmented
Jacobian will be non-singular that is the all Jacobian which you have augmented. Now this
Jacobian becomes non-singular and you can find out the tangent vector that is one tangent vector
is found once not one once once tangent vector is found.
The prediction of the next solution is as follows that is what we really do is that we first compute
the tangent vector and the moment you compute the tangent vector the new tangent vector is
obtained as theta not not tangent vector this is the tangent vector d theta, dv and d lambda, this is
the tangent vector. The new operating condition that is theta v and lambda obtained as the initial
values of this vector plus sigma times the tangent vector this sigma is the step length sigma is the
step length step size okay.
By step size sigma is chosen so that the power flow solution exist with the specified continuation
parameterthe idea here is that the step size sigma should be such that such that now you can
perform the normal power flow solution and obtain the obtain the complete solution that is you
can find out the all the voltages and their phase angles. I will just see here, generally you will
find out actually that in case suppose the step size is large then from that predicted solution you
will not be in a position to get the corrected that is actually the low power flow solution when
you try to perform right the solution will not give actually this the the exact solution or the
required solution.
In that case we we reduce the value of sigma again repeat till we find out actually that in a
character step. We get the solution if for a given step size a solution cannot be found in the
character step the step size is reduced and the and the character step is repeated until a successful
solution is obtained. This is actually a very important at what should become we do not know
what should be characterized the what should be the step size. Suppose you take a very small
step size then you will require large number of repetitions due to a large step size, it will not
converse and therefore one has to take very judiciously these a step size next step step is the
character step the character step is idea of the character step is to really perform actually the load
flow and get the solution.
Now again the, if you see here actually this function f theta, v lambda right. The number of
unknowns are theta v and lambda right but the number of equations will be one equation less
because the the when you have formulated the number of equations you have added a term
lambda into k but number of equations were still the same. Now to get the solution what is to be
done is we add one more equation this equation is written as Xk minus eta equal to 0 that this Xk
this is very important point here the Xk is the state variable selected as continuation parameter
what are the state variables all the all the phase angles, all the bus voltages right and lambda they
are the state variables here right and therefore what we do is that Xk is one of the state variables
and this is chosen as a continuation parameter and and eta stands for value of Xk value of Xk that
is the predicted value of Xk because Xk is one of the one of the state variables right and therefore
if you look here which is our process then we start from the initial operating condition there we
find out the tangent vector and suppose you start from point A and go to B, B is the predicted
solution therefore at this B we know all the values of we know the value of the complete state
vector right and therefore Xk, Xk is is the continuation parameter which is chosen and it is one of
the state variables and value of Xk is chosen equal to equal to lambda which is the which is the
predicted value of Xk predictor value of continuation parameter.
Therefore, the number of equations which now we have will be one more additional equations
you added and therefore this set of equations can be solved by a normal normal load flow and
you will get the solution therefore this is what is called character step. Now the question
basically is that how to solve this problem, how to choose actually the continuation continuation
parameter or continuation variable the Xk that is because Xk is one of the state variable to be
chosen as the continuation parameter the tangent component of lambda d lambda is positive for
the upper portion of the VP curve 0 for the critical point and negative for the lower point, so to
be already mentioned.
The question is how to select the continuation parameters right obviously one can choose some
some any one of the any one of the state variables as continuation parameters but the but the for
optimum solution the approach is a good practice is to choose the continuation parameter as the
state variable that has the greatest rate of change. In fact for to find out what we do is that we
find out the the tangent vector and in this tangent vector the element which has highest value is
chosen corresponding to that the chosen as the continuation parameter or that state variable
which is highest or greatest rate of change is chosen as the continuation parameter.
Let us look at the a flow chart of this continuation power algorithm. The flow chart is as usual
you start your problem. First step is run the conventional power flow on base case, run the
conventional power flow and the on the base case and find out actually the initial solution then
specify continuation parameter. Now here at this stage how to specify the continuation parameter
we have not calculated the tangent vector and therefore the suggestion is you can choose lambda
itself which is the which is the load parameter as a continuation parameter right.

The next step is calculate tangent vector for which the equation has already been specified that is
you calculate the tangent vector, next is if critical point has been reached that is you have to
check whether you have reached the critical point or not if we at if you have reached the critical
point you stop there. Otherwise, otherwise choose continuation parameter for next step.
The first step the continuation parameter was chosen arbitrarily and I told you that we can choose
the log parameter as the continuation parameter. But here, now we once you have this tangent
vector we can choose the continuation parameterthen after choosing this continuation parameter
do the complete the predict the solution out of this prediction you go for the correction that is
perform the correction part and then return back to this point again and this process continues till
you reach the critical point and you can go below the critical point also and got get the other
points operation. This is a very important algorithm which we have discusses.
Now the problem here is that should we use the continuation load flow algorithm right from
beginning or should we use it actually when we reach the near the critical point thus the answer
is obviously, you first solve this continuation solve this problem by normal conventional load
flow that is you are operating at initial condition, you give increment in load again perform the
load flow analysis find out the new solution and till you come very close to the critical point, you
can keep on using the conventional load flow but the moment you reach the critical point or near
the critical point the Jacobian becomes so at critical point become singular and therefore both or
or closure to the critical point you can now resort to the continuation load flow algorithm and
obtain the points operating points which are the critical operating points.
The the idea here is that the continuation load flow algorithm it takes more time to get the
solution because it has two important steps the predictor and corrector steps and so far the the
conventional load flow is concerned you can you can increase the load in a particular direction
and find out the solution.
Next point, we will discuss is what can be done and how the system can be designed and
operated. So that we avoid the voltage collapse situation that is we want to operate the system so
that the we can prevent the voltage collapse. Now here we will divide this our discussion into
two different headings one is the system design measures that is when you design the system you
you take care so that it has a inherent stability margin right and then next is operating measure
that is when your system is already there how to operate the system, so that the voltage collapse
is avoided.
System design measures what are the design measures which can be adopted for for having a
system which will have a large so voltage stability margin and voltage collapse would not take
place there. Obviously, when we have discussed this voltage stability problem, the voltage
stability problem is basically the problem of inadequate reactive power supply. It is a situation
where the system has has less reactive power supply capability, it cannot supply the required
amount of reactive power and the voltage collapse takes place therefore obviously the first
solution or first design the approach should be design measure should be have apply apply
reactive power compensating devices or adequate reactive power compensating devices, apply
we should have various sources of reactive power and that these reactive power sources should
be applied very judiciously.
We will discuss this application part in the next step second will be control the network voltage
and generator reactive power output that is that is when a system is under the design stage design
measure itself right you provide certain certain controls which can be used for for controlling the
network voltage and the generator reactive power output that is you should have the controls so
that you can increase or decrease the reactive power output from the generating devices then
there may be certain certain protections and controls we should have proper coordination,
sometimes the protection there was use when they are controlling a load let us say right and
when the arrangement is there whenever the voltage is decreasing you can do load shedding and
voltage will recover.
Therefore, load shedding is one of the controls and therefore whatsoever protection and control
devices you incorporate in the system, load control devices or voltage control devices all should
be properly have proper coordination with the protection systems then also have proper control
of the transformer tap changers where they are the voltage control devices you have to provide
certain arrangements so that you can control actually the control the tap changer positions. Then
next is under voltage load shedding if there is a last resort in case you find their voltage is
sinking voltage is going low you you do that load shedding.
For example, under frequency load shedding all of you know right when the system frequency is
going down we do load shedding so that the frequency is recovered right. Similarly, when the
voltage is low one can resort to what side call under voltage load shedding but this this should be
done very judiciously and this should be neither than last resort for for for saving the system
from voltage collapse right therefore the point is that these measures which I have put here as
design measures and therefore you should have a provision in the system so that you have a
provision of under voltage load shedding.
As a as a built of strategy to to avoid voltage collapse phenomenon are voltage instability

phenomenon. In case suppose it is provision is not exist it will not work right and generally you
will find that the that the human intervention may sometime success be successful there will not
be successful but here when I talk about this this is the automatic under voltage load shedding
scheme.
Now let us discuss this things in slightly more detail, application of reactive power compensating
devices. We know that we have the various reactive power sources our interest here is that you
should design your system so that you have adequate stability margin that is that is when you
plan for the reactive power sources, you plan in such a fashion so that adequate voltage stability
margin by proper selection of compensating devices.

In fact here the selection of sizes what should we size of the compensating devices what should
the ratings, what should be the locations all these things should be done by by a detailed system
study covering the most onerous system condition for which the system is required to operate
satisfactorily, idea here is idea is that when you are doing this design you have to you can say
determine what should be the location of the reactive power sources what should be their sizes
and what should be the type in the sense that is if control or is it fixed or whether is way which
should be continuously control or else we discrete control that the idea here is that when you
right at the design stage you have to perform a detailed system study or detailed system
optimized study you have to be done. So that you find out actually the sizes locations type of
devices and that should all include the most onerous or most most actually the most heavily
loaded condition right.
We wanted at that condition system should be stable, third is design criteria on maxim allowable
voltage drop following a contingency are not often satisfactory from voltage stability point of
view. Sometimes whenever, we design the application of reactive power sources the one
approach will be that okay suppose the certain contingency occurs right what is the maxim
permissible voltage drop and based upon that you design the reactive power sources but the
suggestion is that instead of going for this type of situation one should design from the point of
view of stability margin that is the stability margins should be based on mega watt and MVAr
distances to instability that is how much real power, inductive power increment can be put on
that system right till instability occurs that is what is that point actually I have already discussed
the minimum distance to instability therefore that criteria should be we must we must design our
system. So that there is a minimum inbuilt voltage instability.
Next the control of network voltage and generator reactive power output again the design step.
Now here, so far actually the reactive power sources are concerned the generator is one of the
most important reactive power source and the generators have automatic voltage regulators, one
way is that you you a regulate the voltage at the terminal at the generator, another way is that
using load compensators you you regulate the voltage at a point away from the generator
terminal that is the point which you choose for regulation may be not the terminal voltage of the
generator suppose this is generator, this is the step of transformer that is you can regulate
actually the field current of the generator through AVR. So that you maintain the voltage on the
high voltage side of the transformer constant this is normally done with the help of load
compensation devices the load compensators are provided to take care of this, this the AVR
regulate the voltage on the high tension side of or part away through the step of transformer.
You can always regulates through that some voltage may not be at this point may not be at this
point but in between that what you manage to decide that is the load compensation of the
generator AVR regulate the voltage on the high tension side or part away through the step of
transformer. This is beneficial effect on voltage stability by moving by moving the point of
constant voltage electrically closure to the loads, in the sense that what is happening that this
constant voltage point is now coming closure to the loads right. This is one approach that this
instead of regulating the terminal voltage you regulate the voltage at some other point using the
AVR load compensation devices.
Then many utilities are developing secondary control or secondary voltage control schemes for
centrally controlling voltages and generator reactive power output. In the sense that instead of
actually having the this they dis-analyzed voltage controlled you an take this system as a whole
right and monitor the system and control the voltages of the network from a central point of
view, control all the generators all actual the reactive power sources centrally, so that you
maintain actually a good voltage profile and maintain a voltage stability margin. This is this is a
new modern approach where we go for system wide control.
The next is the coordination of protection and controls because we have provided certain
protections so that when is suppose system is overloaded the protections will will trip the
equipment right. Therefore, tripping of equipment to prevent an overload condition should be the
last resort not in the necessary that suppose system is operating and there is a transmission line is
the overloaded right the protection system look at the overloading and may trip the line but this it
should be considered as a last resort that is your protection particularly the tripping or protection
system wherever possible adequate control measures, automatic or manual controls may be
automatic or manual. So the provided for relieving the overload condition before isolating the
equipment from the system this point is very important here idea is that here is a system which is
operating.
Let us say overloading takes place the protection system will try to look at the overloading and
trip the line. Our interest is that we should try to relieve the overload by say by say tripping some
loads before before tripping the generator or transmission line or so on that is the equipment
before isolating the equipment here the equipment means basically is the generator transformer
and transmission lines. This is what is the meaning of coordination of protection and controls.
Next about the control of tap changers but this is also playing very important role tap changers
can be controlled either locally or centrally so is to reduce the risk of voltage collapse in fact
what happens is that under certain situation the operation of tap changer is helpful under others
operating condition of situation right the tap changer will not be in a position to increase the
voltage but but that is going to reduce the voltage on the contrary that when you tried to you can
see increase the voltage the reactive power flow increases voltage drops right therefore idea here
is that where the tap changing is detrimental.
So detrimental a simple method is to block the tap changing when the source side voltage sags
and unlock when the voltage recovers right that is again actually very centralized control we
have to, there is also potential for the application of microprocessor based ULTC control
strategies such a strategies need be develop based on the knowledge of the load and distribution
system characteristic that is here microprocessor base means basically it is going to be intelligent
type of control systems right. The next step is discussed was the under voltage load shedding.
Now to cater to unplanned or extreme situations it may be necessary to use under voltage load
voltage load shedding schemes. The load shedding provides a low cost means of preventing
widespread system collapse in the sense the before before the system collapse is right it is better
to can say do some load shedding and rely the system from the overloading. The main cause of
you can say the collapse is again the overloading more reactive power demand than what can be
supplied and therefore if you reduce some of the loads or if you disconnect some of the non-
critical loads right then one can save the system this approach is also used actually even for our
a for improving the angle stability of the system under load, under load shedding.
Then we talk about the system operating measures, what measures can be adopted so that when
we operate the system, operate the system right the system will not go into into voltage collapse
phenomena. The approach is very simple you operate the system so that it is a sufficient stability
margin it means right now at the operating stage you design that yes loading should not go below
a above a particular limit so that you have sufficients voltage stability margin. This point is also
very important is called the spinning reserve, the spinning reserve the concept is the generators
generators are the reactive power sources right and this reactive power can be obtained quickly
by regulating the excitation right and therefore what is said is that you should you should operate
initially the system.
So that other reactive power sources are used and the reactive power capability of the generator
is kept reserve because you have various reactive power sources like shunt capacitors,
synchronous condensers, may be SVC, may be state comment so on right, use then first and the
load the generator. So that generator is operating at nearly the very use the power factor
condition is supplies the real power so it has a reactive power capability and use this reactive
power capability to save the system from voltage collapse that is actually the right at the design
steps. Next is operators action, idea here is that these are the two which you can plan at the at
the operating stage when you operate your system, next is the operators reaction, operators
intervention, how the operator can can actually monitor the system, look at the complete system
as a whole and take necessary action.
I will talk about the operator reactions slight in more detail, the operators must be able to
recognize voltage stability related symptoms and take appropriate remedial action such as
voltage and power transfer controls and possibly a last resort load curtailment. This is the most
onerous task which is put on the operator that is the operator should be in a position to to
recognize the voltage stability related symptoms the, right. So that this operator will take
remedial action before the voltage instability occurs, it means the operator has to be well trained
and he can find out that yes there is a possibility of voltage instability which may take place and
therefore he takes certain measures and it should have certain certain authority to take those
measures. Of course the last resort is the load curtailment if you feels that yes without load
curtailment the system cannot be saved he should have the authority to do the load curtailment
and he should have that type of experience and training.
So that he can make out actually that yes yes at this parse at this operating condition system is
vulnerable is prompt to voltage instability and you should know the load remedial measures to be
taken to save the system. It is very difficult task and that is why actually we require a highly
trained personal to operate the system, operating strategy that prevent voltage collapse need be
established in the sense here here the meaning of this is the operating strategies that prevent
voltage collapse need be established.
Right at the beginning when you are operating your system you have to establish that what
should be the what should be the strategy so that we if this happens this measures should be
taken that should be available to the operator an online monitoring and analysis to identify
potential voltage stability problems and possible remedial measures should be invaluable in this
regard is always online monitoring. With this I have I have completed the discussion on voltage
stability what we have discussed in the voltage stability the how to analyze the dynamic stability
problem this is done by complete time domain simulation then one can apply the certain tools
static tools to to analyze the voltage instability problem.
We have seen the PQ sensitivity analysis, we have also seen the QV model model QV analysis i
have also discussed the continuation load flow for so and so forth. At the end we have discussed
the various measures that can be taken to prevent the voltage collapse in the system. Thank you
very much!
Prof. M. L. Kothari
Lecture - 39
Methods of Improving Stability
Today, we shall study the methods of improving power system stability. We have so far studied
the modeling of the system and various techniques for analyzing the stability of the system. We
also examine the techniques which are suitable for analyzing transient stability and small signal
stability and we also studied the voltage phenomena in a power system. Today, we will discuss
certain special methods which are used for improving stability or enhancing stability of the
system.
The several methods have been evolved over the years for enhancing the stability of the system
and whenever you apply certain techniques or methods for enhancing stability only one method
is not adequate judiciously chosen a combination of several methods may have to used. Today,
we shall try to examine the techniques for transient stability enhancement.
Under this category we will study importance of high speed fault clearing reduction of
transmission system reactance under this category we will study how we can use series
compensation to enhance the enhance the stability of the system by reducing the transmission
line reactance then we shall also study about the affect of regulated shunt compensation on the
stability of the system.
By applying regulated shunt compensation, we will be in a position to maintain voltage at a
desired level at certain points in the system and hence we can improve the power transfer
capability of the system. Then we will also study very special technique which is suitable for
improving the transient stability it is called dynamic braking.
Another techniques which can be useful for enhancing the stability are the reactor switching then
in terms of operation of the circuit breaker, there are two ways which can be used for enhancing
the stability one is the independent pole operation of circuit breaker and the single pole switching
these are the two different approaches for clearing the fault by operating the circuit breakers. In
the independent pole operation as we will see the 3 poles of the circuit breaker are independently
opened and closed that is there is a mechanism is so that we can independently open and close.
One arrangement maybe that all the three poles have one common mechanism to open and close
but here the independent pole operation means we have 3 separate mechanisms for closing and
opening the circuit breaker. However, however the whenever fault occurs irrespective of type of
fault all the three poles will be open. In single pole switching arrangement the as we will see the
3 poles have separate independent mechanism for opening while whenever the fault occurs that
is line to ground fault then only that faulted phase pole will be open.
In case any other type of fault occurs then all the 3 poles will be open this is what is the basic
different between the independent pole operation of circuit breaker and single pole switching.
We will see actually in detail under what circumstances the these techniques are suitable for
enhancing the stability. Then the next important method for improving stability is steam turbine,
fast valving this technique we try to control the mechanical power delivered by the turbine. We
will see actually how we control the mechanical power.
We will discuss under this technique the fast valving logic and typical valve closing and opening
sequence, how the valve is closed, how it is opened this sequence we will see. Then we will see
what is the affect of fast valving on stability of fossil fuel, fired, fossil fuel fired station that
means basically coal fired thermal power plant although the fast valving can be used for nuclear
power plants also.
Then there are some other techniques which are suitable for enhancing stability are generator
tripping, controlled system separation and load shedding, high speed excitation systems or high
speed control of excitation. Then we will examine the performance of two types of excitation
systems from the point of view of transient stability enhancement and at the end I will talk about
the control of HVDC lines for improving the stability of the system these are several these are
the large number of alternative approaches which are available for enhancing the system
stability.
However, as I mentioned that for any particular system for a particular contingency one has to
analyze the system in advance and and incorporate a set of techniques or set of devices which
can be operated to enhance the stability of the system, these all these techniques are primarily for
enhancing the transient stability or we can say the large perturbation stability of the system,
small signal stability enhancement is done by by in installing the power system stabilizers and I
have discussed about the designed and application of power system stabilizers for enhancing
small signal stability in detail earlier we will not go into details about this small signal stability
enhancement now.
Now let us discuss these techniques one by one and see their effectiveness in enhancing the
transient stability. Now when we try to see enhancement of transient stability or transient
stability enhancement all the methods all the methods which are which have been evolved over
the years all the techniques which have been evolved over the years are primarily aimed at
achieving the following following affects, one is reduction in the disturbing influence by
minimizing the fault severity and duration.
Normally you know power system instability occurs occurs when this faults occurs in the system
and as we see that if the duration of fault can be reduced then then the the affect of disturbance is
minimized because after all what happens is the when the fault occurs in the system during the
faulted condition, the power output from the machines decreases, electrical power output.
However, the mechanical input remains same and therefore all these machines where electrical
output has decreased experience accelerations and they gain certain amount of kinetic energy.
The amount of kinetic energy gained depends upon the duration of the disturbance duration for
which the system is under fault condition and therefore one of the approach is to reduction of
disturbing influence by minimizing the fault severity and duration, the one is the fault severity
another is the duration over the years lot of efforts have been made to to reduce the fault severity
and its duration, second is I will talk about what technique can be used but second is increase of
restoring synchronizing forces. See the system is subjected to disturbance and then when has to
be restored back to the initial operating condition or to the new operating condition right for that
we need synchronizing forces.
Therefore the first is the reduction in disturbing influence and second is increase of restoring
synchronizing forces, all the techniques we have to do is that we have to restore. We have more
synchronizing forces that system is brought back to the new operating condition or the previous
operating condition depending upon the the system.
Another very important approach is reduction of accelerating torque that is we want to reduce the
accelerating torque this can be reduced by by prime mover mechanical power control or through
control of prime mover mechanical power you control this prime movers so that the mechanical
power is is reduced when system is accelerating another is by applying artificial load.
Now here when I talked about this mechanical power the technique is called fast valving where
we operate the valves to reduce the mechanical power output from the turbines in a very special
manner I will discuss the fast valving in detail. Similarly, another approach is that we apply
artifical load that is called braking we apply actually at the generator terminals, the braking
resistors and the resistors are applied in a very judicious manner so that the the transient stability
of the system is enhanced. We will talk about this braking resistors in detail.
Therefore, let me sum up the basic approach are is like this one is that you reduce the reduce the
severity of the fault, duration of the fault, increase restoring forces, control mechanical power
you reduce actually the acceleration accelerating torque or apply artificial load to reduce actually
the electrical to increase the electrical power output and to enhance or enhance stability that is to
to reduce the gap between the mechanical power and electrical power or say accelerating power
that is accelerating power can be can be reduced by either reducing the mechanical power or
increasing the electrical power.
Now here let us talk about high speed fault clearing which is primarily for for reducing the
duration of fault on the system enhance the hence the affect disturbing affect the amount of
kinetic energy gained by the generators during fault is directly proportional to the fault duration
the quicker the fault is cleared the less disturbance it causes this is very very fundamental you
can say concept here right and all the efforts have been directed towards rapid fault clearing of
the system. So for the severity of the fault is concerned, the severity of fault will depend upon the
type of fault location of fault right.

But the by by first fault clearing or high speed fault clearing we can reduce the duration of fault.
So that the total duration for which actually the disturbing affect is there on the system is reduced
as it has been discussed actually when we discussed the course on power system protection, our
intention is to reduce the fault clearing time by reducing the operating time of the protective
system and reducing the operating time of the circuit breakers because the fault clearing time is
sum of the relay time, relay operating time plus the circuit breaker operating time.
The relays which we use in the power system today, particularly on EHV system or high voltage
system these are high speed relays and as we know actually that whether you use a distance
protection or whether you use actually the unit protection system for the feeders particularly the
important feeders we use carrier protection system right under this case actually the operating
time of the relay is very low may be of the order of half cycle right some ultrahigh speed
relaying systems have also been reported which are best upon traveling waves where the
operating time of the relays is even less than the 10 milli second which I have just now
mentioned therefore this the high speed relaying is one of the very important aspect or important
technique for enhancing the stability.
So far the circuit breakers are concerned over the years there have been developments in the
circuit breaker technology to reduce the fault clearing time. Today we have circuit breakers
where the total operating time or fault clearing time maybe of the order of two cycles means in
50 hertz system, 40 milli seconds and there have been certain reports where one cycle circuit
breakers or the circuit breakers which operate or complete their fault clearing in one cycle time
are reported and therefore, we can see actually that the so far, so far actually the high speed fault
clearing is concerned, the high speed relay operation and high speed circuit breaker operation is
used for for enhancing the speed stability of the system.

The next important aspect is that is from the design point of view how you design your system
right from beginning so that the system has has high power transmission, power transmission
capability. If you see the power angle characteristic, power angle characteristic of a machine a
simple machine connected to infinite bus, a simple power angle characteristic we plot, then this
power angle characteristic is given by this equation, electrical power output Pe is equal to E1 E2
divided by X sin delta or we say this is equal to Pmax sin delta where, we know actually that this
Pmax this is the this is the maximum power which can be transferred over a transmission line if
you increase the load in small steps right. Therefore, this P maximum becomes the actually the
steady state stability limit now by increasing this Pmax this can be increased by increasing E1 E2
or reducing X therefore if we reduce X or actually the transmission line reactance then we have a
higher power transfer capability in the system that is the steady state stability limit is high and
then correspondingly the transient stability limit will also be high.
The reduction of transmission line or transmission system reactance we are talking of

transmission system reactance. Now here the point is that series inductive reactance of
transmission networks are primarily or primary determinants of stability limits the reduction of
reactance is of various elements of transmission networks improves the transient stability by
increasing post fault synchronizing power transfers right.
The reduction of transmission network reactance is achieved by reducing the transmission lines
reactance by itself. Generally, generally when you design the transmission line right this will not
be the only criteria for for or reducing the transmission line and reactance may not be the
primary criteria of transmission line design but we know that in EHV transmission systems, if we
use the bundle conductors the the line reactance reduces right and therefore these are the certain
steps which can be used actually to reduce the line reactance or the bundle reactance is bundle
conductor is required for for EHV transmission lines from the consideration of corona losses and
but the byproduct of this use of bundle conductor is that it reduces the transmission line
reactance and hence it helps in improving the the power transfer capability of the lines then in all
transmission system, we will have transformers and therefore you design the transformer use
transformers of low per unit reactance.
The transformers of per unit reactance will be of .1 to .15 are normally used in the EHV systems
and then wherever, you find that length of transmission line is high or a long transmission lines
are involved then series compensation is used for for reducing the line reactance. Now today
there has been a lot of lot of you can say thrust on this aspect one is you have a fixed series
compensation another is by using FACTS technology, you can go for for control series
compensation by using control series compensation like thyristor control series, compensator and
capacitors or static static synchronous series compensator the SSSC right.
We are in a position to control actually the line impedance in a dynamic fashion and improve the
transient stability this is what is the normal technology here therefore most promising technique
for improving stability is to make use of either fixed series capacitors or sometimes we may use
switch series capacitors or the most suitable is actually the continuous control series capacitors.
Now here, when we go for series compensation one has to one has to study or learn some
important aspects one is one is that generally whenever the fault occurs in the system system the
series series capacitors need to be bypassed during the fault the series capacitors need be
bypassed reason is actually that when a fault occurs in the system the voltage across the series
capacitor becomes very large and to protect this capacitors.
The capacitors need to be bypassed and then the when the fault is cleared, we have to reinsert
this series capacitors if you can reinsert quickly then only then they will be effective that is the
series capacitors need to be present in the post fault system for enhancing the system stability.
This point is that for for enhancement of transient stability depends on the facilities provided for
bypassing the series capacitors during fault and then reinsertion after fault clearing earlier earlier
simple approaches were used where they were using actually the spear gaps across the series
capacitors to bypass the series capacitors but there was the problem with this spear gaps is that
yes they when voltage across the gap increases it will spark over and bypass the series capacitor
but till the till the ionized space between the air gap is you know cleared or it recovers its
strength the series capacitor is not reinserted
Therefore, the new technology which is used today is the use of non-linear resistors which are
basically the zinc oxide non-linear resistors and the terminology used is called metal oxide
varistors MOVs, MOVs are used across the capacitors and whenever fault occurs depending
upon the voltage which is developed the MOVs will bypass this circuit the capacitors but so far
actually the MOVs are concerned the the voltage drop across the MOV will will not become 0,
it is not going to completely bypass but it will shunt this capacitor and the voltage across the
capacitor is restricted or limited to safe value right but the interesting thing is that the reinsertion
time is instantaneous.
The movement the fault is cleared the MOV recovers its original high resistance condition and
the capacitor is reinserted instantaneously therefore this this technology is very you know
prevalent in all systems wherever we have series compensation I would mention that even if you
use fixed series capacitors we require protection system for the series capacitors that is MOVs
even when we use the FACTS devices like thyristor control series capacitors or SSSC, SSSC, S
cube C or we use UPFC then we have to can say have an arrangement.
So at the voltage drop across these series component is reduced right and is safe that is series
compensation MOVs are always provided but the advantage of MOV is that you you you can
quickly reinsert them, re insertion re insertion time is practically negligible. As I have mentioned
here the series capacitors directly offset the lines is reactance, the reactance becomes small the
maximum power transfer capability of a transmission line may be significantly increased by use
of series capacitor banks. Then the transient stability enhancement depends upon upon how
quickly you reinsert the capacitor when it is bypassed by by bypassing devices under fault
conditions.
For transient stability enhancement one can used switch series capacitor I just now mentioned
either it can be switch series capacitor so that some of the capacitors is a fixed one some is
switched one but this switch series capacitors were considered attractive when the FACTS
technology was not available but with the FACTS technology available available the same thing
is achieved using the FACTS technology. Now when you go for series compensation there are
some problems and one of the important problem is the the protective relaying for the
transmission system becomes more complicated because the distance relay which we normally
use where we we measure the impedance of the transmission line between the relay location and
the fault point to find out the distance of the fault from the relay location.
Now here this this simple approach cannot be used because because part of the line reactance is
compensated by the series capacitor therefore this is a subject of research and many people are
pursuing research on this subject that is the protection of series compensated transmission lines.
Next is we go for regulated shunt compensation, the shunt compensation capable of maintaining
voltages at selected points of the transmission system can improve system stability by increasing
the flow of synchronizing power among interconnected generators following devices have been
popularly used.
I will explain this point slightly in more detail that is for for regulating the voltage at selected
points in the transmission network. The devices available are are synchronous condensers which
has a rotating device but it is very fast in response and it is capable of maintaining the voltage at
the certain locations. Another new technology is we use the static VAR compensators, the SVC
we call it SVC static VAR compensators this is part of the FACTS technology and the next is the
STATCOM which is actually the voltage source converter based technology which can be used
to inject the controlled reactive power in the system and maintain the voltage constant.
I have not mentioned here UPFC but unified power flow controllers also have a capability to
maintain the voltage constant at a certain bus in addition to controlling the power flow on the
line. In fact each of these techniques are very very powerful and lot of research work has been
done to understand how much how much transient stability enhancement can be done but here
here the capability of these devices can be harnessed by appropriately designed controllers.
So that during the dynamic condition whenever the system is accelerating we we reduce
mechanic we reduced we increase electrical power output when the system is accelerating and
when the system is decelerating we we we reduce electrical power output that is the approach
basically for all these devices is this electrical power output Pe is increased, increased when d
delta by dt this is the or you can say d2 delta by dt2 that is the acceleration or you can call it d
omega by dt is positive whenever, you find that this system is accelerating means acceleration is
positive what we have to increase Pe, so that we have to reduce the acceleration and therefore we
require actually the controllers which will sense the acceleration and accordingly act, so that we
can reduce the or increase the electrical power in response to whether the system is accelerating
or decelerating.
Now here I will discuss the two important techniques one suitable for artificially increasing the
electrical power output of the machines during during dynamic conditions, another is
dynamically controlling the mechanical power input, dynamic braking is a technique which is
used for for artificially controlling the electrical power output from the machine during the
dynamic conditions, during the during the condition when this machine is oscillating dynamic
braking uses the concept of applying artificial load.
Now here because we have to understand load is anyway it is a load but it is not the normal load
put on the system by the consumers we have a load actually which is specifically put to to
observe certain electrical power from the machines it is why it is called artificial it is it is not
actually there useful power, it is not going to do any useful work for us it is going to help in
improving the stability that is why the word is used is artificial load during a transient
disturbance to increase electrical power output of generators and thereby reduce the rotor
acceleration.

Now here actually I will just discuss slightly in detail about these braking resistors, braking
resistors used to-date are shunt devices although one can use series devices or series connected
braking resistors for enhancing the stability but most convenient method or what technology
which has been used is is a shunt connected devices this point I will explain slightly in detail
here. Let us look let us look at a synchronous machine. This machine is connected to a large
network, external portion of this system to be generator or to which it is connected I will not
show in detail but let us say that this is the power system network with all loads and generating
only I have taken out one particular machine at the terminal of this machine what we do is that
we keep a bank of resistors which will be connected connected by switching on this switch that
is what we have is that in the generating plant we have a bank of resistors and these resistors are
connected through through a some switching arrangement.
Now here first I will tell how this helps in improving stability and then I will talk about the the
new technology which is used for switching. Here basically these resistors are switched on only
when the machine is accelerating that is you sense the acceleration and when the machine rotor
is accelerating it means electrical power output is less you close the close the switch. So that this
resister absorbs power we usually that when the system is under fault condition the voltage at the
bus of the generator becomes low therefore, normally during fault condition the these resistors
are less effective but when the fault is cleared the terminal voltage is restored and the system is
swinging right during that period you switch on so that this resistance will absorbs observe
power right and the power which is absorbed by the resister will apply electrical artificial
electrical load, you can consider it and reduce the acceleration right.
Now the movement actually the system is decelerating you disconnect this resister. Now today,
today the improvement here is that instead of using a mechanical switch for closing the braking
resistors or inserting the braking resistors, we use what is called actually the thyristor controlled
braking resistors, the terminology is called TCBR, thyristor controlled braking resistors are used.
You can put it here in thyristor controlled controlled braking resistors. On this braking dynamic
braking this is called dynamic braking where we we is make use of actually thyristor controlled
braking resistors people have done lot of research work, the research is basically in terms of
designing the controller so that what signal we what signal is to be sensed so that these resistors
are inserted in proper time and taken out.
Then another point which is very important here is that what should be the dose of braking dose
of braking in the sense actually that how much power this resister should absorb, on that also a
lot of research has been carried out generally these resistors are short duty resistors because these
resistors are used only for a fraction of second and therefore these resistors are capable of
absorbing huge power with their small size. Generally, today they use actually the stainless steel
strip of large size right and that is capable of absorbing huge amount of power because they they
are switched on only for a very short time of the order of fraction of a second right. Now when
you use this thyristor controlled control then by adjusting the firing angle you can you can
control the dose of braking and therefore, today is technology is that we can use actually the
thyristor controlled braking resistors and they are very effective.
Now let us talk some few aspects about this. The series resistors may be used to provide the
braking effect this point i will explain separately, another form of braking resister application
that enhances system stability for enhanced for unbalanced ground faults only consist of resistor
connected permanently between the ground and the neutral of the Y connected high voltage
winding of the generator step up transformer.

In fact here this point can be understood say you have a generator then at the generator, we put
output of the generator straight away comes to a high voltage or you can say step up transformer
and then it goes to the transmission network. This transformer is delta Y transformer, you you
can ground this through a resistance, now this resistance remains permanently right. So that
whenever there is a ground fault right the fault current flows through the resistor and when this
the because of the fault current is flowing through the resistor certain amount of power is
dissipated in this resister and this will act as a load on the system and generator will be
decelerated when it is trying to accelerate it because when fault is occurred right the the power
output from the terminal of the machine is reduced but now we are increasing the power loss in
the neutral connected resistor, this is permanently connected right.
So that you certain amount of power is absorbed here this is called actually the neutral braking
resistor, you can call that neutral resistor. This is what the point here another form of braking
resistor application that enhances system stability for unbalanced ground faults only, there
should be ground faults right consist of a series connected permanently a resistor connected in
series permanently between the ground and the neutral of the Y connected high voltage winding
of the generator step up transformer. This is a very import technique which has been used
actually, series resistors the approach is like this.
To understand this concept of series resistors I will draw a simple diagram, I will draw a single
line diagram only but let me just this is the stator of the machine, in series with this you connect
on the neutral side these resistors and these resistors will have a permanently by passed by
putting a by passing switch that is this switch is closed here this switch is permanently closed
and this neutral is of a grounded one right you can understand. This is the these are the buses
terminals of the generator.
So that under normal conditions this series connected resistors are bypassed but whenever you
find actually the machine is accelerating you you open these switches which are bypassed
switches. So that this resistance are inserted in series with the stator winding and so that these are
very effective during the during the fault condition, when the fault is on right when the system
output becomes very low right at that time the fault the fault current flows through these resistors
and a huge amount of energy can be absorbed but this alternative arrangement has been talked
about but it is not very popular from the application point of view.
The another point here is these braking resistors have been applied only to hydraulic generating
stations which are remote from load centers the hydraulic units in comparison with thermal units
are quite rugged therefore they can withstand the sudden shock from the switching in of in of
resistors without any adverse effect on the units. In fact what is actually is actually happening is
that you you you just close them switch right then suddenly the electrical power output will
increase and this causes actually a what we call actually the transient break on the system and in
for so far this thermal units are concerned there is a danger actually so that the the shaft may get
damaged because of frequent stroke with the which which may occur on the shaft by by closing
and opening the circuit resistors.
The braking resistors are applied to thermal units, then this shaft fatigue life must be carefully
examined that is before you apply apply these braking resistors for the thermal units you must
examine this thing that what will be the effect of these braking resistors on, on shaft fatigue life
there is certain life of the shaft and due to this phenomena its life gets reduced. With the
switched form of braking resistors the switching time should be based on detailed simulation that
is how you switch on how you switch off duration of the time for the switch is on then off this all
has to be done by doing lot of simulation studies and then designing the appropriate controllers.
Now this this is a danger if the resistors remain connected too long there is a possibility of
instability on the back swing. This can be this is a self explanatory that is if the resistors remain
when when the system is no more accelerating right then this is putting additional load on the
system and therefore system will decelerate backswing that is why actually the whenever you use
this type of sophisticated techniques sometimes they use the word heroic techniques and this
heroic measures when you want to apply you have to take special precautions.

This is a very simple approach for enhancing stability that is reactor switching, shunt reactors
near generators provide a simple and convenient means of improving transient stability. The
reactor normally remains connected to the network the resulting reactive load increases the
generator internal voltage, this is beneficial to the stability in a sense what happens is that you
have at the terminal of the generator a reactor connected right it means actually the the generator
is seeing a low power factor at its terminals and therefore to maintain the terminal voltage at the
desired level desired level the internal voltage is large excitation will be more internal voltage is
high and when the internal voltage is high right and the system is subjected to disturbance what
we do is that we disconnect actually the reactor and because of high internal voltage the
synchronizing power becomes more and that is following a fault switching out of the reactor
further improves the stability. Then let us talk about the independent pole operation of circuit
breakers.
Now this is again a measure which is used for enhancing the stability as I have mentioned
actually that high speed fault clearing is the target or aim. Sometimes what happens is that due to
some failure in the in the circuit breaker mechanism right the circuit breaker may remain struck
right. Now suppose you have a arrangement arrangement where all the three poles are controlled
by one common mechanism if the circuit breaker one pole gets struck the complete circuit
breaker remains struck, closely struck right therefore and and actually the stability design of the
system considering a three phase fault with a struck circuit breaker is considered to be the most
severe contingency.
Suppose, you design your system that how I should load the system, so that my system remains
stable with three phase fault with a struck circuit breaker right that is very very onerous condition
for which system may be designed. Now, generally independent pole operation refers to the use
of separate mechanisms for each phase of the circuit breaker so that the three phases are opened
and closed independently of each other. As a result failure of one pole will not restrict the
operation of other poles this is what I was telling you that this is the that there maybe some
mechanical failures right and if all the poles remain struck right then the the fault is not cleared 3
phase fault still remains therefore if suppose actually if you have a independent pole operation, if
one pole remains struck other two poles clear then the fault fault actually severity is reduced.
Although the break is another important point although the breaker poles operate independently,
the relaying system is normally arranged to trip all 3 poles for any type of fault this was again
this was mentioned earlier also when I talked about the outline of the lesson that in this
independent pole operation although the mechanism for opening these poles are independent, 3
separate mechanisms but the relaying system is such that such that for any type of fault whether
it is line to ground fault or three phase fault or any type of fault all the poles are opened
simultaneously right this is what is the meaning of independent, operation is independent but the
all these poles are opened that is line is completely opened.
The system design criteria include three phase fault compounded by circuit breaker or 3 phase
fault with a stuck circuit breaker suppose you have designed a system with this as the criteria
design criteria is is a 3 phase fault compounded by stuck circuit breaker right. Maintaining
stability for this contingency is extremely difficult in generally you will find actually that we are
not in a position to design the system with this type of contingency, if you design the system
with this type of contingency means the system system loading will be very low for a given for
this type of contingency, you have to very low loading .Otherwise, this type of contingency when
it occurs system will lose stability but since this type of situation is 3 phase faults are very rare in
the occurrence and the possibility of the circuit breaker poles getting stuck right. Looking into
the probability and using independent pole operation we can we can mitigate this type of
situation that is that is we can design the system for lesser or less severe contingency.
With circuit breakers designed for independent pole operation a failure of all 3 poles is highly
improbable, it all the three mechanisms will fail. Simultaneously, that is very that probability is
very low that is called it is why it is improbable and used the another is actually that as we have
discussed when we talked about the protection system to increase the reliability of the protection
system, we make use of duplicate relays or duplicate protection systems right and lot of you can
say features are added.

So that the failure of relays and the failure of circuit breaker is reduced, use of duplicate relay
systems duplicate circuit breaker coils, tripping coils guarantee is actually that at least two poles
will open that is you use certain certain features. So that actually whenever some some defects
develop in the system possibility of all the three poles getting struck is very low and at least two
poles will open. In fact actually one need to examine how the how this independent pole
operation with actually one pole remaining closed and two poles open will help in improving the
stability can be easily examined because what happens is that three phase fault has reduced to a
single phase fault and the next is the single pole switching this is different from independent pole
switching in single pole switching uses separate operating mechanisms for each phase for single
line to ground faults relaying is designed again.
Now the relaying relaying is designed to trip only the faulted phase followed by fast reclosure in
.5 to 1 point second for multiphase faults all 3 phases are tripped this is what is the mechanism
that in single pole switching when there is a line to ground fault may be on a phase or b phase or
c phase whichever phase is faulted you open that pole and then reclose it because many of
manlier faults many faults are transitive in nature and the reclosure will be successful right
therefore the this by using this single pole switching with the reclosure facility, we can enhance
the transient stability to a great extent this is very important practice used actually in all EHV
transmission network today.
The reason is I have again let me reiterate most of the faults on transmission systems are line to
ground faults opening and re-closing only the faulted phase results in improvement of transient
stability over 3 phase tripping and re-closing. I think I will just mention about the steam turbine
fast valving this is very promising technique for improving the stability of the system, we will
discuss this remaining methods for enhancing transient stability in the next lecture.
Let me sum up the points which we have discussed today. We have discussed the basic
philosophy of improving the transient stability then we have talked about high speed fault
clearing system, then design aspects to reduce the reactance of the transmission system,
transmission network, one can reduce this by using series compensation or one can go for a
controlled series compensation, a very powerful techniques.
I talked about the shunt compensation and particularly the controlled shunt compensation which
maintains the voltage at certain points and increase the power transfer capability of the system by
increasing the the synchronizing torque of the system and then we have discussed about braking
resistors, application of braking resistors and particularly the the thyristor control braking
resistors are very suitable today for enhancing the stability and then I have discussed about the
two different mechanisms of opening this operating the circuit breakers, one is the independent
pole operation and another is the single pole switching.
Prof. M L Kothari
Lecture - 40
Methods of Improving Stability (Contd)
Friends, we continue with the study of methods of improving power system stability. In the
previous lecture, we have studied some of the techniques which help in enhancing the stability of
system and the techniques which we have already discussed last time are high speed fault
clearing, reduction of transmission line system reactance we also talked specifically about the
series compensation regulated shunt compensation, we talked about dynamic braking reactor
switching independent pole operation of circuit breakers and we came to single pole switching.
Today, we will discuss the reaming techniques which add in enhancing the stability of the power
system. The techniques which we shall discuss will be single pole switching, steam turbine fast
valving, under this we will discuss the fast valving logic and typical valve closing and opening
sequence. Then, we will see the typical result through simulation studies the effect of fast
valving on the stability of fossil fuel fired station, then we will talk technique like generator
tripping, control system separation and load shedding, high speed excitation systems. Under this
category we will investigate or we will see the performance of two types of excitation systems
the ac exciter and bus fed thyristor, exciter then at the end we will talk about the the effect of
high speed control of HVDC lines for enhancing stability and at the end small signal stability.
As we have seen earlier the single pole switching uses separate operating mechanism on each
phase for single line to ground faults relaying is designed to trip only the faulted phase. Now this
is what is the important point that is when there is a line to ground fault right then the protection
system will trip only the faulted phase then this will follow a re closer within .5 to 1.5 and in case
there is a multiphase fault like say to phase to ground or line to line fault or 3 phase fault then the
all the 3 phases will be opened. This is very powerful technique for enhancing stability and it is
used in EHV transmission systems.
We take the advantage of this fact that most of the faults on transmission lines are line to ground
faults and opening and reclosing only on the faulted phase results in improvement of transient
stability. As compared to three phase tripping and reclosing some of the some of the features
which one has to study about this single pole switching are
that single pole switching is attractive where where you have a single line connects or wherever
the single line connects the 2 systems because if there is only one line connecting the two system
then then the power which can be transferred with the stability comes out to be 0 because the
moment its one line is faulted and you disconnect that line then the power transfer become 0
right therefore wherever there is a single line which connects the two systems then if you can
resort to the single single pole switching then some power is transferred on the healthy phases
and the and the the amount of power which can be transferred with the stability is improved that
is the transient stability limit is increased another is that if you have a single line connecting a
generating station to rest of the system.
Suppose, you have actually a generating station located at far of point and you have connected
that station to the system through a single line normally we connect through more than one line
but is the only one line connecting then again there is a advantage in going for single fault
switching, the problems associated with single pole switching are secondary arc extinction. I will
discuss this secondary arc extinction what it is but if it is a problem secondary arc extinction, the
fatigue due to turbine generator shafts and turbine blades if you are using because because
whenever you go for single fault switching the operation of the system becomes unbalanced
right.
Then the because of unbalanced operation the the thermal duty on nearby generators due to
negative sequence currents because when this one pole is taken out and only two lines are
connected then a stator windings of the synchronous generators carry negative sequence current
and that causes the additional heating in the machine then we talked about the power system
protection, we did discuss about the capability of generators that is the negative sequence current
withstand capacity of the synchronous generators for few seconds synchronous generators can
withstand this situation but not on long time basis.
Now when I talked about the secondary arc extinction the problem is something like this that if
you have a 3 phase system and you disconnect one phase at both the ends right and there was a
fault on the phase which has been disconnected. Now this this phase which has been separated or
disconnected from the system remains electrostatically and electromagnetically coupled with
other phases and therefore there is certain potential which is developed across the conductor
which is isolated from the line and therefore that it is because of this potential the the for at the
fault point right the arc will be sustained is called secondary arc this they you know what
happened is that when the fault occurred system was healthy before the occurrence of fault and
this arc was arc was fed by the system energy but the moment you have taken out the faulted
phase right.
Since this is is isolated we we feel actually that there should be no more arc arc should be
cleansed but because it remains coupled to the remaining two phases through electrostatic
coupling and through electromagnetic coupling there will be certain amount of voltage induced
and this is a problem which is very serious problem. So far actually the the single pole switching
is concerned because you cannot re-close till that secondary arc persist because if you re-close it
again. You will find that it will be unsuccessful re closer because whenever you adapt a
technique there with some problems and one has to sort out those problems and accept adapt that
technique only when you are in a position to overcome those problems.
Now we take now another very powerful technique for improving the stability of the system and
that is steam turbine fast valving. Historically, the fast valving technique was recognized long
back somewhere in 1930 it self as a very powerful technique for enhancing stability but because
of certain unresolved problems up to 1960 no attempt was made to apply the fast valving
technique for stability enhancement. Let us see the firs fast valving is a technique applicable to
thermal units to assist maintaining power system transient stability it involves rapid closing and
opening of steam valves in a prescribed manner to reduce the generator accelerating power
following following the recognition of a severe transmission system fault that is if you find that
there is a severe transmission system fault. So the electrical power out power power output from
the generating unit has reduced and mechanical input is same therefore to decrease the
accelerating power right one has to control the valves of the steam turbine. So as to reduce the
mechanical power generated by the steam turbines.
Let us discuss, this several utilities have tested and implemented fast valving on some of their
units for illustration of fast valving application, let us consider a fossil fuel generating unit with a
tandem compound single reheat steam turbine. In fact steam turbines have different
configurations you may have a single reheat double, reheat tandem, compound, cross compound
there are so many different configurations are there further the this fast valving technique can be
applied to nuclear power stations also right but just for illustrating this technique I have
considered this particular system that is fossil fuel generating unit that is a coal fired unit here we
will have a single reheat turbine and it is a tandem compound. In tandem compound actually all
the all the units are on the same shaft the high pressure intermediate pressure low pressure all
these stages of that turbine they are mounted on the same shaft.
Now let us consider this configuration if you see this configuration then the steam enters from
the boiler at this point we have a main safety valve here, main inlet safety valve there is a control
valve then the steam enters the high pressure turbine after expanding in the high pressure turbine,
steam is sent to re heater, this is re heater. Then we have intercept valve this is this valve is called
intercept valve and then we have again here this RSV right, this is also a safety valve right a re-
heater safety valve we call it re heater safety valve. The steam goes the IP section there are 2
sections shown here after working in the IP the steam will enter the LP stage of the steam turbine
and you can easily see that high pressure low, intermediate pressure, this is low pressure, low
pressure. All these stages are mounted on the same shaft and at the end is the generator
connected.
The steam which leaves the intermediate low pressure section goes to condenser and then it is
recirculated to the turbine this is what is the configuration of the tandem compound single reheat
steam turbine, the fossil fuel it is a coal fired unit.

Now here we we have two possible controls one is that you close this control valve if you close
this control valve then the steam flow will stop and the mechanical power output will be reduced
second is that yes you control not the control valve but you you close this intercept valve this
intercept valve when you close again the mechanical power output from the turbine will reduce.
We will see the different configurations which can be used that is the main inlet control valves
and re-heater intercept valves this I have told you that these 2 valves, one is the main inlet
control valve, another is the reheater intercept valve they provide a convenient means of
controlling the steam turbine power because we want to control the steam turbine power for a
short time momentarily. So that the the system is saved from losing synchronism.
In fact there are two alternative schemes which have been discussed, one is the intercept valves
are rapidly closed and then fully opened after a short time, it is not actually open the the main
control valve only intercept valve is opened and then closed again that is the other way around
intercept valve is closed and then opened fully okay the intercept valves, now since the intercept
valves control nearly seventy percent of the total unit power this method results in fairly
significant deduction of steam turbine power.
The another arrangement is that or you can say there are more pronounced temporary reduction
in turbine power can be achieved through action of both control and intercept valve that is you
control both the intercept valve as well as the control valve, you close both the valves. Now this
procedure of rapid closing and subsequent full opening in both the cases that is what you do is
that intercept valve is closed and opened. Similarly, the control valve is closed and opened okay
this particular type of arrangement is call momentary fast valving then there is another scheme
which is called sustained fast valving in sustained fast valving what is done is that intercept valve
is closed and reopened fully while the main control valve is closed and partially reopened, well
what happens is that suppose the post fault system is weak right. So once you partially reopen it
means you are reducing the mechanical power input to the system as a whole right. Therefore,
that that type of arrangement is called called sustained fast valving valving is fast it has to be fast
because you know that if you want to save the system then this mechanical power or the controls
should be fast So that they they reduce the acceleration and deceleration okay.
(Refer Slide Time: 17:14 min)
The generally fast valving has been found to be effective and economical method of meeting the
performance requirements of power systems where design and operating criteria require stability
to be maintained for a 3 phase fault with delayed clearing because of stuck circuit breakers.
Suppose, you design your system so that system stability is maintained for this type of
contingency where you consider a 3 phase valve with stuck circuit breaker it means the fault is
going to be clear with certain delay under such contingency the fast valving is very effective in
enhancing the system stability because here the fault remains on the system for a longer time.

The logic which is used for fast valving are one is for generating the valve control sequence and
other is for generating unloading signal right and let us look at the typical valve closing and
opening sequence.
Now this diagram shows a typical valve closing and opening sequence. This is this diagram is a
simple diagram in practice you will find that there will always some deviations that is this
diagram shows the valve position as a function of time. Now suppose a time T equal to 0, you
will sense that yes valve is to be closed then you take a initiation or take take initiate the closing
of the valve then for a time equal to T1, T1 there is a certain time delay which is required to start
the movement of the valve that is you will find at T1 is a delay between the time of initiation and
the time where the valve begins to close. These valves are generally very heavy and there is lot
of inertia is involved and the therefore you will find that the moment you give start initiate the
action there is certain time delay that is called T1 then from T1 to T2 the valve position is closed,
So that actually slowly the valve is closed at T2, you come that is T2 is the valve closing time you
can say at this time that is from this point to this point the time is wall closing time T3 that is
from this instant to this instant this is the time T3 the time during which the wall remains closed
and T4 is the wall opening time that is you a time T equal to 0 you imitate the action for for
closing and opening the valve. Then this is the time delay no opening closing takes place while
in the same position this is the time which requires for closing the valve from this point to this
point that is for time T3 valve remains closed then from this point onwards it reopens and this is
the fully opened condition this is partially open condition this is fully closed position that is
when during this period when its partially closed right again you will find actually that during
this period the steam is flowing right because its a partially closer and certain mechanical power
is developed. The mechanical power in the intermediate pressure and in the low pressure section
is not developed during this period.
Now here the ability of steam valves to rapidly close and reopen depends upon the governor
system which we have used because after all this is going to be actuated through the governor
action and electrohydraulic turbine governing or electrohydraulic turbine governor system is
capable of this rapid control because you have to close very fast.
I will tell you typical timings for for closing and opening of this valve but this times are very
short. Now you I will just show you the effect of fast valving on stability of fossil fuel fired
station. Let us take a typical illustration to see the effectiveness of fast valving on improvement
of stability.
Now, here this really the the example which I have considered is that there is a plant which has 2
units of 500 megawatts and they are feeding power to network and there is a close by a severe
fault on the transmission network following that following that the fast valving is resorted and in
this diagram shows shows the rotor angle in degrees that is delta delta power angle delta versus
time t in seconds I think it is the swing curve of the machines without fast valving this is the
graph you can see this is the graph this graph is the graph or a swing curve without fast valving
and with fast valving this is the swing curve, this is the swing curve with fast valving.
You can easily see that without fast valving the maximum swing is something of the order of 120
degrees while the maximum swing is now now around 70 degrees and therefore whenever we try
to investigate the effectiveness of certain devices for improving stability then the swing curve
conveys lot of information and what is the maximum swing that is important right.
Now this graph shows this graph shows the power in per unit on 100 MVA base versus time on
this side that is what has been done is to investigate or to show the effect of fast valving, effect of
fast valving on stability of a fossil fuel station exhibiting a slow inter area swing that is the graph
which I have shown earlier, you can see actually that the the the swing curve shows that there
exists a inter area swing this is very slow, thing type of thing but the point which I wanted to
make here is we show here with this dotted graph this dotted graph you can see this this one I
will just highlight it here again.

This graph shows the mechanical power with fast valving that is at this point you have initiated
the fast valving phenomena, for some time mechanical power does not change then it decreases it
comes to from something like 9.8 per unit it comes down to something like 5.5 per unit, 6.5 per
unit then slowly it increases and you can see that this whole thing closer and opening has been
completed practically in less than 1.5 second. You can see this thing that this graph, this this is
the 2 second time in 2 seconds you can easily see that practically it has reached the same level of
mechanical output therefore there is a dip in mechanical power which is generated and this dip
has helped actually in improving the dynamics of the system that we have already seen.
Now let us look into certain special features of fast valving, the fast valving imposes a relatively
severe transient on the turbine and steam generator system what is happening is that you
suddenly close suddenly close the interceptor valve then what happens, the steam which is which
is coming from the high pressure side and it is actually trapped in the re heater right that steam is
stopped suddenly because of sudden movement there will there will be a shock to the reheater
and to the turbine side right that is what is the point that fast valving impose a relative severe
transient on turbine and steam generator system.
This is steam generator means boiler here boiler re heater and it also imposes on the turbine
because in the turbine we are carrying normal steam you will find that actually suddenly
mechanical input reduces and therefore this sudden change in the steam flow effects actually the
life of the turbine blades and therefore I would mention here that whenever fast valving is
applied as the technique for enhancing stability, one has to evaluate evaluate that what will be
the consequences of this technique in terms of in terms of the steam generating unit that is boiler,
re heater, on turbine that is when we talk about turbine turbine blades fatigue of the shaft all
these things have to taken care of and only when certain situation demands that this is effective
and can used is applied. Then there are some other techniques which can be used for enhancing
stability.
The generator tripping is another powerful technique which can be used but we have to be very
careful in using this technique that is let us understand how this generator tripping helps.
Suppose there is a large system right and there is severe fault on the system. Now this machines
in the system will be accelerating or decelerating depending upon what is the excess power right
and all the techniques which we have discussing have mechanical mechanical reduction in power
that is you control the steam valves or you apply actually the artificial load in the system right
are basically to reduce the acceleration under some situation you will find that you can trip
certain amount of generation if you trip certain amount of generation again actually you will find
this say the excess power which was there to accelerate the system is reduced right the this is
what is the approach here.
A selective tripping of generating units for severe transmission system contingencies have been
used as a method of improving stability for many years but again you know the the consequences
of tripping a unit have to be examined suppose you trip a unit because tripping a unit is easy, you
open the circuit breaker general circuit breaker you will find actually that the whole unit transfer
of generator is disconnected from the system tripping is not than take much time time taken is
the same circuit breaker and relay operating time right.
But the moment actually the system generating unit which are supplying certain load right
immediately you will find that there are certain ill effects to transient on that system speed will
increase governor will try to control the speed right and all these effects are there one
arrangement is that suppose you you to the machine and follow the normal procedure of shutting
down you close the steam turbine you stop the machine another is that you use.

Let the machine continue to run for some time and supply certain auxiliary power or power
which is required for meeting the auxiliary requirement right and then resynchronize this
machine when the system is becoming healthy that is historically there are practice of generator
tripping as a stability aid was confined to hydro units and this practice has been extended to
fossil fuel, fired units and nuclear units in 1917 because this is considered to be a very powerful
actually in technique for saving the system when system is subjected to very severe
contingencies that is not in a position to save the system, sometimes blackout takes place if you
cannot take certain measures which have happened in our Indian power system also.
Evidently tripping of a unit subject in in its subjects it to sudden changes in mechanical and
electrical loadings you can easily understand this aspect the unit which are supplying certain load
suddenly, there is electrical power output is 0 mechanical output is reduced mechanical loading
right. With the associated impact on generator prime mover and energy supply system again
again as I have discussed actually about the fast valving phenomena. Similarly, when you talk
generator tripping right the it affects the steam generating system it affects the turbine it affects
the generator all these units and one has to examine very carefully the impact of tripping before
you resort to this type of practice in fact the thermal units are not designed for frequent full load
rejection. Generally, whenever whenever the machine is unloaded the load is slowly reduced full
load rejection means machine is having an 100 percent loading condition and you trip the
machine that type of situation is very dangerous to the thermal units.
Now this is this is warning that generator tripping technique for improving system stability
should not be used indiscriminately okay, you understood actually the following are the major
turbine generator concerns the over speed resulting from tripping of the generator, thermal
stresses caused by the rapid load changes and the shaft fatigue life consideration, all these are the
major considerations they have to be evaluated examined before you make use of this this
approach. But this is considered to be one of the very powerful approach or for enhancing system
stability under certain certain, I would say severe contingencies.

Another approach is controlled system separation and load shedding. Controlled system
separation may be used to prevent a major disturbance in one part of interconnected system from
propagating into the rest of the system and causing severe system break up. Now this it is
something like this this normally we call this islanding islanding therefore you have a large
interconnected system and if a disturbance has occurred in part of the system you do you resort
to islanding. So that the the disturbance does not propagate to the remaining part of the system
and today certain steps are been initiated in our Indian power system also so that we can save
part of the system whenever some disturbance occurs in a certain part of the system because in
2002 January the the northern regional grid completely collapsed and no part of the system could
be saved from that point onwards efforts are been made to to resort to certain techniques so that
at least you can save a certain part of the system and complete blackout is avoided. Globally also
certain approaches are been adopted to save the system from complete blackout.
Today the terminology which is used is called wide area protection system which requires
actually the synchronized phase measurement units to be used so that you have a common time
for the complete system then all system operates in a completely synchronized manner, the
timings are all synchronized and they have a synchronized arrangement you can have the
synchronized phasor measurements and monitor from central point. So that you can find out
whether the disturbance is coming and what to do to save the system from blackout.
The impending system instability detected by monitoring one or more of the following quantities
in the sense that how do we monitor that yes a disturbance is coming in a particular system and it
is going to propagate into the remaining part of the system, one is sudden changes in power flow
through specific transmission circuits there will be one indication on a particular line some
power is flowing and suddenly you find that there is wide change in the power flow that is one
indicator, change in bus voltage angles rate of power changes these are some of the indicators
that can be used to initiate the system separation, a very complex problem in fact it is not that
easy to talk about the separation.
Now upon detection of impending instability impending instability control system separation is
initiated by opening appropriate tie lines before cascading outages can occur. Few few selected
tie lines have to be opened in some instances it may be necessary to shed selected load in the
separated system that is when you separate the system problem may come there may be
generation load unbalance wherever you have a more generation and load less you have to shed
some generation wherever you find that load is more than generation you have to shed some load
right.
Therefore this is another this is to be approached there is no way out you will find that whenever
you do this islanding you will not be in a position to make the island to have a load generation
balance impossible, wide area protection and synchronized phasor measurement units are applied
for this type of protection purpose or for this type of islanding okay.
Now next is the high speed excitation systems how high speed excitation systems helps in
improving the transient stability of the system, over the years there has been development in the
in the excitation system technology I have discussed the different excitation systems in my
previous lectures I have discussed the excitations systems models also. Now here significant
improvement transient stability can be achieved through rapid temporary increase of generator
excitation, idea is something like this actually we make use high initial response high ceiling
voltage excitation system.
Now normal normal fault conditions what happens is the terminal voltage drops you will find
actually that whenever the system fault is there the voltage at the terminals of the generator will
be low the automatic voltage regulators will quickly take action and apply to the field winding
maximum voltage this is equivalent to the celling voltage which will cause in turn increase in
field current and internal voltage. Now this internal voltage which is which is increased by this
action will help actually in increasing the synchronizing power when the fault is cleared.
There is a mechanism here that is increase of generator field voltage during a transient
disturbance has the effect of increasing the internal voltage of the machine and this in turn
increases the synchronizing power. Because we have discussed in the beginning itself that we
have to take measures so as to increase the synchronizing power which can restore back the
system to the new operating condition, this is one of the approaches. The AVR response to this
condition by increasing the generator field voltage and this has beneficial effect on transient
stability.
The effectiveness of this type of control depends on the ability of the excitation system to
quickly increase the field voltage to its highest possible value that is, if suppose the excitation
system is slow you will find will not be effecting it has to be fast that is why the effectiveness of
this technique depends on the ability of the excitation system to quickly increase the field voltage
to its highest possible value that is to the field we apply the highest possible value that is the
celling voltage. Here the high initial response excitation system with high ceiling voltage are
most effective in this regard.
The ceiling voltage again then what should be the ceiling voltages what is the maximum voltage
which can be applied to the field winding of the synchronous generator that is called the ceiling
voltage? All the excitation systems are designed to have certain ceiling voltage. The ceiling
voltage are however limited by generator rotor insulation consideration for thermal units the
ceiling voltage are limited to about 2.5 to 3 times the rated load field voltage. Suppose a
generating unit has a rated field voltage of say 250 volts right then the ceiling voltage may be
750 volts is three times right this is what is the meaning and this voltage can be raised from 250
is the nominal value to 750 in a very short time.
The fast excitation response to terminal voltage variations required for improvement of transient
stability often leads to degrading the damping of the power system oscillation, this is the another
ill effect of of high response or initial response high ceiling voltage excitation system that causes
the degrading of the damping that sometimes it adds negative damping and therefore we need
actually the application of PSS to overcome this problem, power system stabilizers a lot of
research work has been done on this and we, I have discussed actually the effect of PSS in my
previous lectures and the effect of high high or we can say high initial response excitation
systems on the stability in previous lectures also.
The high the use of high initial response excitation system supplemented with PSS is by far the
most effective and economically method of enhancing the system stability. In fact actually out of
all the techniques which we have discussed so far this is a very accepted and tool for for
enhancing the system stability you will find today large capacity machines will have high initial
response excitation system, high ceiling voltage and along with power system stabilizers the ill
effect of high initial response excitation system which enhances transient stability but degrades
the damping is overcome by providing PSS.
Let us just look the comparison of transient stability with a AC exciter and a bus fed thyristor
exciter system in fact here I will make the comparison between the two excitation systems one is
the AC exciter system, where you have a AC generator rotate rotating or stationary rectifier
system and another is that you you have a static excitation system which is taking which take
input from the bus other than taking from the terminal of the synchronous generator the bus in
fed that is the bus fed thyristor excitation system.
Thyristor word of course is used thyristors are required in both the cases but here the meaning is
one is static excitation system another is a rotating excitation system.

For the purpose of illustrating this thing we have considered a system with 2 generating plants or
2500 mega watt plant there are 2 units of 500 megawatt capacity, a 3 phase fault of 60 milli
second duration on major transmission line is considered. For example, I considered certain
improvement by using the fast valving same system is here also you have a power plant with
2500 megawatt units and there is a major fault actually in the transmission system and 3 phase
fault is considered. The fault duration considered is a 60 milli second if we use for this particular
system, the critical clearing time with AC exciter that is 47.5 milli second only.

For this particular system with the particular type of fault considered with AC excitation system
AC exciter, the critical clearing time is 47.5 milli second while the critical clearing time with
static excitation system is 62.5 milli seconds. Obviously, the fault duration is 60 milli second
with this AC exciter the system will become unstable with thyristor excitation system system
will remain stable. This is illustrated through swing curve again here, rotor angle here, you can
see that this firm line shows the the swing curve with thyristor excitation system with PSS and
this is AC exciter you can easily. See that the system with AC excitation system has become
unstable while the while the static excitation systems remains stable it is a beauty right because
normally the stability margin is always judged by knowing actually the critical clearing time and
the actual fault clearing time. For example, in this case the with AC exciter there is no margin
system is unstable however with the static excitation system the we have a margin of 2.5 milli
seconds.
Now we discuss some more techniques one is the control of HVDC lines. Now in some systems
we have HVDC lines and wherever HVDC lines are available the we can take the advantage of
HVDC lines because HVDC line is one where you can quickly quickly change the power flow
on the line the controls are such that you can increase the power flow at a very fast rate. You can
say the ramp of the power output or ramp of the power flow on the line or ramp down the power
flow on the line this control is very fast this is only clearly by adjusting the firing angle of the
thyristor walls you can appropriately ramp up or ramp down the power flow on the line and this
advantage can be taken wherever you have HVDC link. I mean I am taking about here actually
that this is a AC system with additional HVDC link already operating, it is not actually that we
have 2 systems connected by HVDC link it is not that way they are not two asynchronous
systems same synchronous system but there is a HVDC link.
Now when a HVDC transmission link is highly controllable it is possible to take advantage of
this unique characteristic of HVDC link to augment the transient stability of the AC system.
During a transient disturbance the dc power can be ramped down rapidly to reduce generation
load unbalance of the system on both sides, on some situations it may be necessary to ramp up
the dc power to assist system stability by taking advantage of short term overload capability of
HVDC system normally the HVDC systems have short term overload capacity and all these
control measures are only for a short duration.

Now just before I close my discussion on the methods of transient stability enhancement. Let us
discuss about small signal stability enhancement the problem of small signal stability is one of
the insufficient damping of power system oscillations the use of PSS to control generator
excitation systems is the most cost effective method of enhancing small signal stability of the
power systems. The application of power system stabilizers and their capability in enhancing
stability I have discussed in my earlier lectures I am not going to repeat the discussion on PSS
but today today all excitation systems which are manufactured by by excitation system
manufacturers they provide power system stabilizers as a integral part of the excitation system.
Now the equality companies have to tune the parameters of their excitation systems or power
system stabilizer. So that the local as well as internal modes of oscillations are damped out. Now
my before I close the discussion on methods of improving the stability of the system. Let me
mention that the new technology that is the flexible AC transmission system technology is
evolved primarily to have flexible controls on the system. For example, if you whatsoever we
have talked till now is either on the generator or on the turbine or on the HVDC line but very
little discussion I have made on transmission line.
Now here when I talked about the application of series capacitors or for compensating the
transmission line reactances or TCSE for for for control {comp} ((00:50:07 min)) control
compensation of the transmission lines in fact actually the application of these FACTS devices
whether it is a TCSE or SVC or STATCOM or UPFC or other FACTS devices. All these FACTS
devices have the capability of controlling power flow on the transmission line and we can
regulate the power flow on the transmission line rapidly and on these systems also we can add
certain devices to damp out the oscillations that is I can use TCSC for improving transient
stability as well as dynamic stability I can control actually the line impedance for enhancing the
stabilities.
Similarly, you can control the STATCOM or shunt connected devices or UPFC where you can
enhance both both transient as well as the dynamic stability of the system or improve the
damping of the system. Today the lot of research work is been carried out to coordinate the
tuning of the FACTS devices as well as the the power system stabilizers. So that we want
coordinate the tuning so that the the devices are tuned in coordinated fashion. Now with this I
have just complete discussion of techniques of improving the power system stability a variety of
techniques have been used for improving the transient stability of the system and dynamic
stability of the system and these techniques need to be judiciously used taking care of the system
considerations and one only one technique is not sufficient one has choose a a a group of
techniques or a combination of certain techniques they have to be used to improve the system
stability.
With this I have completed the series of 40 lectures on on power system dynamics or say power
system stability during this we have discussed the modeling of the system synchronous
generating modeling excitation system modeling, turbine modeling, governor modeling. There
are various techniques of analyzing the stability then we have discussed the voltage stability, we
have discussed about the power system stabilizers and at the end the techniques for improving
the transient stability of the system.

Power System Stability

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Power System Stability

Uploaded by

Copyright:

Available Formats

Power System Dynamics

(Refer Slide Time: 01:07)

(Refer Slide Time: 07:49)

(Refer Slide Time: 09:46)

(Refer Slide Time: 10:23)

(Refer Slide Time: 11:06)

(Refer Slide Time: 11:26)

(Refer Slide Time: 14:01)

(Refer Slide Time: 16:22)

(Refer Slide Time: 18:35)

(Refer Slide Time: 19:27)

(Refer Slide Time: 20:08)

(Refer Slide Time: 21:18)

(Refer Slide Time: 22:45)

(Refer Slide Time: 24:08)

(Refer Slide Time: 24:33)

(Refer Slide Time: 26:30)

(Refer Slide Time: 28:23)

(Refer Slide Time: 28:48)

(Refer Slide Time: 30:47)

(Refer Slide Time: 34:19)

(Refer Slide Time: 37:18)

(Refer Slide Time: 37:43)

(Refer Slide Time: 41:37)

(Refer Slide Time: 43:58)

(Refer Slide Time: 45:39)

(Refer Slide Time: 47:29)

(Refer Slide Time: 01:09)

(Refer Slide Time: 03:24)

(Refer Slide Time: 05:06)

(Refer Slide Time: 07:02)

(Refer Slide Time: 09:47)

(Refer Slide Time: 14:05)

(Refer Slide Time: 14:21)

(Refer Slide Time: 18:28)

(Refer Slide Time: 20:43)

(Refer Slide Time: 22:36)

(Refer Slide Time: 23:22)

(Refer Slide Time: 27:15)

(Refer Slide Time: 29:36)

(Refer Slide Time: 32:46)

(Refer Slide Time: 34:24)

(Refer Slide Time: 37:45)

(Refer Slide Time: 40:06)

(Refer Slide Time: 42:41)

(Refer Slide Time: 45:02)

(Refer Slide Time: 45:58)

(Refer Slide Time: 49:00)

(Refer Slide Time: 01:07)

(Refer Slide Time: 03:25)

(Refer Slide Time: 05:35)

(Refer Slide Time: 07:23)

(Refer Slide Time: 08:32)

(Refer Slide Time: 10:38)

(Refer Slide Time: 14:25)

(Refer Slide Time: 15:43)

(Refer Slide Time: 16:50)

(Refer Slide Time: 19:09)

(Refer Slide Time: 20:46)

(Refer Slide Time: 21:51)

(Refer Slide Time: 23:06)

(Refer Slide Time: 24:31)

(Refer Slide Time: 29:42)

(Refer Slide Time: 31:36)