Using UPFC in order to Power flow control

Maryam Hashemi Namin Power Electrical Engineer Iran University of science and Technology Tehran University T.R.E.C

Abstract- This test case is made to verify the current injection model of the UPFC. The UPFC is installed in a sample network. Then a fault study apply to this network by monitoring active power flow in the faulted line for the system with and without the UPFC. The parameters of the UPFC are chosen based on static behavior of the UPFC. With the control of the variables r and y, improvements in damping of the oscillations are shown obviously The general form of the UPFC control system has been proposed. The UPFC should operate in the automatic power flow control mode keeping the active and reactive line power flow at the specified values. This can be achieved by the linearizing the line power flow. Figures show the first preliminary results of the proposed control method if the specified value of the active and reactive power be chosen .An alternative control strategy for the UPFC is based on the series voltage injected by the UPFC. If this injected voltage is instantaneous by the UPFC, the components can be related to the control variables r and y. The further studies will investigate these control methods with respect to performance and robustness.
1. Unified Power Flow Controller (UPFC) The flexible AC transmission systems (FACTS) initiative was originally launched to solve the emerging problems in the late 1 980s due to restrictions on the transmission line constructions, and to facilitate the growing power export/import and wheeling transactions among the utilities. FACTS devices can enhance transmission system control and increase line loading in some cases all the way up to thermal limits thereby without compromising reliability. These devices can be an alternative to reduce the flows in heavily loaded lines, resulting in an increased loadability, low system loss, improved stability of the network, reduced cost of production and fulfilled contractual requirement by controlling the power flows in the network. These capabilities allow transmission system owners and operators to maximize asset utilization and execute additional bulk transfer with immediate bottom-line benefits. FACTS devices provide new control facilities, both in steady state power control and dynamic stability control. The unified power flow controller (UPFC) is arguably the most comprehensive device to have emanated so far from the FACTS initiative. In principle at least, the UPFC offers new horizons in terms of power system control, with the potential to independently control three power system parameters such as bus voltage, line active and reactive power. Provided no operating limits
1 -4244-0726-5/06/$20.OO '2006 IEEE 1486

are violated, the UPFC regulates all three variables simultaneously or any combination of them. Using controllable components of UPFC, the line flows can be changed in such a way that thermal limits are not violated,losses minimized, stability margin increased, contractual requirement fulfilled etc, without violating specified power dispatch. Power flow (or load flow) analysis involves the calculation of power flows in lines/transformers and voltages of a power system for a given set of bus bar loads, active power generation schedule and specified bus voltage magnitude at generating buses. Such calculations are widely used in the analysis and design of steady state operation as well as dynamic performance of the system. The power flow problem is formulated as a set of nonlinear equations. Many calculation methods have been proposed to solve this problem. Among them, NewtonRaphson (NR) method and fast-decoupled load flow method are two very successful methods. In general, the decoupled power flow methods are only valid for weakly loaded network with large X/R ratio network. For system conditions with large angles across lines (heavily loaded network) and with special control devices (FACTS devices such as UPFC) that strongly influence active and reactive power flows, NR method may be required. Besides transformers, the general structure of UPFC contains also a "back to back" AC to DC voltage source converters operated from a common DC link capacitor, Figure 1. First converter (CONVI) is connected in shunt and the second one (CONV2) in series with the line. The shunt converter is primarily used to provide active power demand of the series converter through a common DC link. Converter 1 can also generate or absorb reactive power, if it is desired, and thereby provide independent shunt reactive compensation for the line. Converter 2 provides the main function of the UPFC by injecting a voltage with controllable magnitude and phase angle in series with the line via an voltage source, Figure 2. The reactance xs describes a reactance seen from terminals of the series transformer and is equal to (in p.u. base on system voltage and base power) I XS = k S. )/$S

where xk denotes the series transformer reactanse, rmax the maximum per unit value of injected voltage magnitude, SB the system base power, and SS the nominal rating power of the series converter.The UPFC

injection model is derived enabling three parameters to be simultaneously controlled. They are namely the shunt reactive power, Qconvl,and the magnitude, r,y and the angle,y ,of injected series voltage Vse.The series connected voltage source is modeled by an ideal series voltage Vse which is controllable in magnitude and phase, that is, Vse=rV kej < <; l) O= where 5where U 9 r < -~~andIZ r < r and
CoeTEr "meh, 2

= -. P. V=rb : + ui-rj- ia. A

(8)

-separate

_I

T
Figure 1: Implementation of the UPFC by back-to-back voltage source converters =
T}

=-rb9o-g + rbV w+rb-2 I v s the series Afterwards, voltage source is coupled with the shunt part of the UPFC, which can be modeled as a controllable shunt reactive source. Here it is assumed that QCONV 1 = O,but to allow for QCONV1I 0 in the model is straight forward. Consequently, the UPFC injection model is constructed from series connected voltage source model with the addition of power equivalent to PCONV 1 + jO to node i. The UPFC injection model is shown in Figure 6.

_<<J 2

1.1 Injection model of UPFC To obtain an injection model for UPFC, it is first necessary to consider the series voltage source Figure 3.
zise
J

Figure 2: The UPFC electric circuit arrangement

Q =

P1<

Figure 6: Injection model of the UPFC (l0I) =rbj- s-in(OZ- !+ )

Q ~=rb.

-rbg1K Q

2:co;~ (r1 P;

cos(&

13)

The injection model is obtained by replacing the voltage source Vse by a current source Iinj -jbs Vse in parallel with xs, Figure 3.
-1

Figure 3: Representation of the series connected voltage source

Xgp

xm

JA. -

lo..

F_

(. r n....- _

Besides the bus power injections, it is useful to have expressions for power flows from both sides of the UPFC model defined. At the UPFC shunt side, the ~~~~~~~~~~injection iction mel define. wAthesUPFC pwrlows ar gvna and reactive I 1, 'cos X ~ _ b,K2 + Qi1 =- -r k'J1 1.5) J? C08-Y + Qco3uiwhere as at the series side they are + = 2 IrbIj VJsin (116) s Oi, ( V_ The UPFC injection model is thereby defined by the constant series branch susceptance,bs,which is included in the system bus admittance matrix,and the bus power
.
y

where r and y are the control variables of the UPFC.

b.

Figure 4: Transforme The current source Iinj corresponds to injection powers Si and Sj which are defined by
_ ii24 -m \i-I:, ==

2Ct

b =I i/ h Where X = 6 q Figure 5 shows the injection model of the series part of UPFC, where
1g = -r&alt

Sj

=_

v- ~(I m O

rbJ 1

sinl(-n

cus(&
c

-v)

(3)j

c n

through changes

to be

injections Psi,Qsi,Psj andQsj.If there is a control objective achived, the bus power injections are modified
of the UPFC parameters r,y, and

Qi =-nS

4L)

7j) P = -realc)d , =-i Having the UPFC losses neglected,

* the series voltage source The apparent power supplied by covre*sclc-ae rm converter iS calcu-lated from: - i X,
.......................St'F

ONVI= F1XiVV2

6)

3 Power flow with the UPFC The performance of the UPFC injection model is tested on the two area four generator power system shown in Figure 8. The 230 km interconnecting tie line carries 400MW from area 1 (generators I and 2) to area 2 (generators 3 and 4) during normal operating conditions. The injection model of the UPFC is placed at the buses 8 and 12 in of the llower line ~~~~~~~~~beginning i between buse the wee begrnig sf 12ai through order to see the influence on the power flow 8hand that line as well as on the bus voltages. According to the algorithm for rating of the UPFC, rmax, installed
X WD 6
7

11

Active and reactive power supplied by Converter 2 are distinguished as:

Figure 5: Injection model of the series part of the UPFC

2C

Figure 8: Two area system with the UPFC

1487

SS and maxjPconvlj are defined, although maxjPconvlj basically is not needed in this test because the shunt part is inactive For the value of rmax 0:15pu,the corresponding powers SS and maxjPconvlj are equal to 0.40 pu, and 0.2737 pu, respectively. That value of rmax is usually estimated to be acceptable for voltage/power flow control purposes, [2].Having the UPFC shunt part inactive(Iconvl 0),the UPFC has two control parameters , r and, the magnitude and the phase of the injected voltage respectively. Thereby, the shunt side voltage Vi cannot be controlled. Figures 9, 10, and 11 show active power flow in line 8, where the UPFC is located. Figure 9 shows the power flow in line 8 where is kept constant at various values while r varies from 0 to 0.15. It can be seen that the controllability of the power flow with r is maximal when S = 7J2for increasing power flow and when =a3'/2for decreasing load flow. The relationship between r and active power flow is monotonic for fixed. Figure 10 shows the same active power flow in line 8 but with respect to rotational change in r and . That means, r is kept constant at some values for a full circle of the angle (WO 360)jIs it obvious that the active power flow is maximal when r is maximal. The active power flow in the system without UPFC in line 8, is Pbase = 1:9526, whereas the maximum change in positive directions equal to +0:6012 pu, and in negative direction -0:6711 pu. It means that by inserting the maximum value of the magnitude r(0: 15pu), the active power in line 8, could be changed by maximum 54.47 MW in positive direction or by 48.98 MW in negative one, if the angle is appropriately adjusted. The maximum active power flow conditions occur around 700 and 2500. Figure 11 shows the relations of the both parameters in single three dimension picture. Figures 12 and 13 show the bus voltages at the series and shunt side of UPFC, with respect to the rotational change in r and . Because the third parameter of the UPFC, Qconvl, is inactive, Vi is not controlled in this case.As can be seen the voltage magnitude have opposite directions. One of them has magnitude increased when the other one is decreased.
2r _

Figurell:ActivepowerflowinlinewithUPFC
i.S LDS

~gnma

[1 'OA

"'fl

Figure 12: Series side bus voltage magnitude Vi f(r;y)

models of a synchronous machine ~~~~~~~~~Mathematical from very from elementary classical models to

Figurel3: Shunt side bus voltage magnitudeVj=f(r;) 4 System models 4.1 Synchronous machine model
more

vary

detailed ones. In the detailed models, transient and subtransient phenomena are considered. Here, the transient models are used to represent the machines in the system, according to following equations.Stator winding

equations:
-nd

+ZdE
Ed

(8
t EWli` i.J NritunienSt I) h 4d k
,

t VL= -rd

li 1

l.i

4 uw p

Figure 9: Active power flow in line with UPFC;

const:

Rotor winding equations: dJd)

i20)

T9dt

EdE+dr=Ef- (

t qj~

)o{2D

;T~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I
T_Eq 2,f
+ A -

+z

E2pX

Fiur 10 Acive pwr flo in ine withUPFC r cost

1. .

Rotr.eqaion:...i.iP.ii
gii .i i iCDi g '.littgi4iXg9.

1488il9ii

Id

TJ Figure 15: Block diagram for computation of torque

d.-I
T X3 tj 6 ilw nsom&A3 tromtw .U

K[)N2Ib
((l

f ti 1k

DawD
g

I'd igi

,mlpxgtk hiilig lwidw

Figure 17. The first block is the stabilizer Gain block, which determines the amount of damping. The second is the Washout block, which serves as a high-pass filter, xi 8with a time constant that allows the signal associated with tmoscillations in rotor speed to pass unchanged, but does not allow the steady state changes to modify the terminal voltages. The last one is the phase-compensation block, which provides the desired phase-lead characteristic to compensate for the phase lag between the AVR input and the generator torque.
the 15,4ock BptLU torq

D is t1w- daultiniE, weffikieu.mVSMAX 7k6 d wuid q~&As b1ok diA,nA of'1te amT fimm CZ thpo trawknSt

nlioo

L gead I Lg

Id'-t

Lg

ES

and speed in the transient generator model For time domain simulation studies, it is necessary to include the effects of the excitation controller. Automatic voltage regulators (AVRs) define the primary voltage regulation of synchronous machines [3]. AVR and exciter model for synchronous generator is modeled as the standard IEEE model, Figure 16.
I

Figure 14: Block diagram for the transient generator model

where It is the in phase with Vi and Iq is the V31 %Ei.AX Ncurrent in quadrature with Vi. In Figure 19 the voltage source V se is replaced by the current source inj . ..................... ....

4.4 UPFC Injection model of the UPFC is described in the static part --of the analysis, where the power injection model is used. However, for a dynamic analysis,due to model requirements, current injection model is more appropriate. Figure 18 which illustrates the UPFC electric circuit arrangement is repeated here due to clarity. In Figure 18, Ih + += 2 i 2 =57 )

Figure 17: PSS block diagram

VSMIN

current

4.2 Load model T.- V J J 'A The loads can be modeled using constant impedance, constant current and constant power static load models [3]. Thus 1. Constant impedance load model (constant Z): A static load model where the real and reactive power is proportional to the square of the<2m voltage magnitude Fgr 9 rnfre eisvlaesuc 2. Constant current load model (constant I): A static load model The active Figure 19:uranforedsyterishvoltag sourentsuc a where the real and reactive power is directly proportional to the powacua ed rosplemyth hn uretsuc a b acltdFo Costn =meac load model (constan -) A rbati loa sine voltage magnitude. 1.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~N 7 3. Constant power load model (constant PQ): A static load model Prv R {-7 ) 26) where the real and reactive powers have no relation to the voltage From the static part we have equations: magnitudeAll these load models can be described by theimpdaVI =onstant 27A m ] t following equations: p
P
P

Figure 16:AVR and exciter model for synchronous generator

Figure 18: The UPFC electric circuit arrangement

2! SS )

whereFrom last three equations we have

~~~~~~~

,(TV

mb I-

'J

-r~JTz

The shunt current source is calculated from where P0 and QO stand for the real and reactive powers + consumed at a reference voltage VO. The exponents Lk and j,+o IA t hdepend ba S is being asrbed Thenithree.l fadthetpca llutrthedi on esic theblok loadmodel,anb that for type of represented; constant power load models = = L, for constant current From the Figure 19 can be defined, !P;VI = XE2 {20)148 load models and * * and for constant impedance load can be=vieed asanaddtionacontrlblokuse towhere, sta tic fErom the part, Aodels

h=-r~~~~~~~~~~~~~~~~~~~~~~~~~~1 smnOj

b~I )+ (29)

Inserting Equations 3Oand 33 into Equations 3 land 32 yields

rb.I
=r -,

ssn F e? k ty

'Jq,#

(34)fC.-

iGu

9s

where Iq is independently controlled variable, like a shunt reactive source from the power injection model of UPFC. Based on previous Equations current injection model can be presented as in Figure 20.
.......... ~~~ ~~ ~ ~ ~ ~ ~ ~ ~ ~~

A tl I1'2 14It18 s 2. ~6 ~

Figure 22: The active power ow in the line 8 with UPFC installed after fault applied

4.5 Results The two area system is shown again here, due to clarity in Figure 21. The system data can be found in [3]. The system model is used as it is described above,but without PSS. The active and reactive components of loads have constant current characteristics =3=1) The UPFC is installed in line 8, according to figure 21. Suppose that the fault occurs in the system at point F. The fault is cleared after 100 ms by opening the faulted line. Figure 22 illustrates the active power flow in line 8 in that case, for the system with and without the UPFC. The UPFC is not controlled. The parameters of the UPFC are chosen based on static behavior of the UPFC. This test case is made to verify the current injection model of the PFC. With the control of the variables r and, improvements in damping of the oscillations should be obvious .Figure 23 proposes the general form of the UPFC control system. The UPFC should operate in the automatic power flow control mode keeping the active and reactive line power flow at the specified values. This can be achieved by the linearizing the line power flow equations 16 and 17 around the starting point resulting in the gain matrix in Figure 23. A2rand Ar are the changes in the control variables, assuming that the third control variable Iq is inactive. Figures 24 and 25 show the first preliminary results of the proposed control method if the specified value of the active power is Psp = 2:5[pu] and the reactive power, Qsp 0:02[pu], see Figures 24 and 25. The starting point is defined at Pbase = 2:1526[pu] and Qbase = - 0:1798[pu]. An alternative control strategy for the UPFC to be investigated is based on the series voltage udq injected by the UPFC. If udq is the instantaneous voltage injected by the UPFC, the components ud and uq can be related to the control variables

Figure 20: The UPFC current injection model

.....

PQ

4-r X
;

Figure 23: General form of the UPFC control system

r
J

Iri-Q0r
V

Aipes.c.

li- 8

24
3

Z.2

SK

Figure 24: Controlled active power flow

Ra- tiNr

Figure 24: Controlled active power flow

.Is

E .S

6 UPFC Controller Performance

the potential of the UPFC controller in damping power system oscillations,a simple proportional UPFC controller based on the generator angular speed as a stabilizing signal is considered. For simplicity, the UPFC controller is considered to employ a unity feedback proportional controller as shown in Figure 26.
Ara

a=rXS1n

q=-rS1n:^!)

g)

and hence (37 = = (3'"t-7 The further studies will investigate these two control methods with respect to performance and robustness.
0 1 63l;gIa

damp power system

Generator angular speed is a commonly used signal to

oscillations. In order to

investigate

Figure 21: The two area system with UPFC installed in line

Figure 26: A UPFC proportional feedback controller.

The performance of this controller is examined with respect to the performance of the power system without a

1490

7 UPFC Controller Design The proposed design of the UPFC controller is based on a technique that combines the pole shifting method and the Genetic Algorithm principles. In such a technique, a fitness function is defined such that it will be maximum when the dominant eigenvalues are located in the left half of the s-plane as far from the imaginary axis as the physical constraints permit in order to achieve high damping. The fitness function used by the proposed technique searches for the optimum controller design that depends upon the system parameters. It is selected such that it combines the effect of the real and imaginary parts of the eigenvalues on the time response. Moreover, this fitness function does not depend on the order of the system but on the dominant eigenvalues. In selecting such a fitness function, the closed loop system must satisfy the

UPFC installed. Two system disturbances are considered for each case, a step change in the mechanical input power (ATm = 0.1 p.u.) and a step change in the reference voltage (AVref= 0.1 p.u.). The power system dynamic performances due to these disturbances are shown in Figures 27 and 28 It can be seen from these figures that the Proportional UPFC Controller (P-UPFC) significantly damps power system oscillations.

damping ratio. When wd/sd is less than tanb, the fitness function will have a bigger value, and when wd/sd is greater than tanb, the fitness function will have a smaller value.The second constraint is applied when the most dominant eigenvalues are conjugate complex roots. In the case of a real dominant eigenvalue, the fitness function is Isdl (maximizing settling time as possible).Two UPFC controller structures are considered in these investigations. The first controller is a Lead-Lag Controller, shown in Figure 30 Aw is used as an input signal for the Lead-Lag controller. The second controller is a Proportional Integral (PI) controller with Aw and Ad as input signals as shown in Figure 31 The controller parameters are optimized using the GA technique described above.

following requirements: 1- The system is stable. 2- The damping ratio is maximum and preferably greater than a prespecified design value. 3- The settling time is minimum (maximize the negative real v X part). The fitness function is defined as follows:
0

Figure 29: Design area for eigenvalues. Aw ii, AtE SerieS VotuiWge

1 2l..... Figure 30 Lead-Lag controller.

KIT:[

Shunt Voltage
AiiiB

ta.n1 . F (F- I f = (F Ir +| {0 1 :.6 Where:sd is the real part of the dominant eigenvalues wd is the imaginary part of the dominant eigenvalues cosb is the lower limit of the desired damping coefficient. ATa pOx.tp

for

> d0

Figure 31: PI controller. It can be seen that the dominant system eigenvalue is shifted from a near critically stable location to a more stable location in the case of the power system equipped with the UPFC controller.It can also be seen that the Lead-Lag and the PI UPFC controllers have higher damping coefficients. 8 Appendix
The generator parameters in per unit are as follows:
= =8 H = 6.5 (ifr Gl and G2) s .1s T= H = 6.1 li for G3 and GC) D., = 0 The exciter parameters in per unit are as follows:

v..X.=
We
=

i0.2

7 X=i.A

XA

R.5. =R JI125

Figure 27: Power angle deviation during a step change in the mechanical input power.

LOS
_____

= = ia 1 .=1
_ & __ F=

____

'P-.PhC

__4__ B r = 1.0 75 TR = 0 Y05 0056 The UPFC parameters in per unit are as follows:
=

[1] R. Sadikovic, Single-machine in_nite bus system, internal report, July 2003

9 References

T~ 02

|I).09

Kr

~02z

9S0 S~ =, 4

T,~ 02

IF

K, 2

Voltage Stability Problem, Doctoral thesis, Zagreb, 2001 [3] P. Kundur, Power System Stability and Control, McGraw-Hill,

[2] N. Dizdarevic, Uni_ed Power Flow Controller in Alleviation of

Power System Dynamic Performance The area shown in Figure 29 repres Power angle deviation Power System Toolbox, Version 2.0, Cherry Tree Scienti_c the reerenc voltgeent the esign[5] h Software, On-tario, Canada_ e1n a step change during [6]A.A.Eldamaty,S.O.Faried,S.Aboreshaild,"Dampig Power system area defined by the fitness function.For a stable system, the oscillations using a fuzzy logic based UPFC second constraint of the fitness function is to maximize the

Figure 28:

Inc.,1993

[4] Z.J. Meng, P.L.So, A Current Injection UPFC Model for Enhancing

interfrnevlaeet

1491