Frequency Response Analysis

It is how the output response (amplitude, phase shift) change with the frequency of
the input sinusoidal.
The input frequency is varied, and the output characteristics are computed or
represented as a function of the frequency. Frequency response analysis provides
useful insights into stability and performance characteristics of the control system.
Figure below shows the hypothetical experiment that is conducted.

Figure: How frequency response is defined.

Response of a first-Order System to a Sinusoidal Input

Consider a simple first-order system with a transfer function
y(s) Kp
G (s)   (1)
F(s)  ps  1
Let F(t) be a sinusoidal input with amplitude A and frequency ω;
F(t) = A sin (ωt)
Then
A
F(s)  2 (2)
s  2
Sub. F(s) from eq. (2) into eq. (1)
Kp A Kp A
y(s)   2  
 ps  1 s    ps  1 (s  j )(s  j )
2

Expand into partial fraction and find

C1 C2 C3
y(s)   
s  1 /  p s  j s  j
Compute the constants C1, C2 and C3 and find the inverse of laplace transform.
K A  K A p K A
C1  2p 2 p , C2  2 p 2 , C3  2 p2
p  1 p  1 p  1
K p A p  t / p K p A p Kp A
y( t )  e  cos(t )  sin(t )
 p2  2  1  p2  2  1  p2  2  1
t / p
As t, then e  0 , and the first term disappers.

Process Control /Lec. 11 120 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
Thus, after a long time, the response of a first order system to a sinusoidal input is
given by:
K A K A
yss ( t )   2p 2 p cos(t )  2 p2 sin(t )
p  1 p  1
Kp A
yss ( t )  2 2 [  p cos(t )  sin(t )] (3)
p  1
Use the follwing trigonometric identity.
p cos  q sin   r sin(   )
p
r  p2  q 2   tan 1
q

q 1 p    p
r  ( p ) 2  1   2p  2  1
2

p   p
  tan 1  tan 1 ( )  tan 1 (  p )
q 1
Then eq.(3) yield
K A
yss ( t )  2 p2 [(  2p  2  1) sin(t   )]
p  1
Kp A
yss ( t )  sin( t   ) (4)
 p2  2  1
  tan 1 (p )  Phase lag (5)

From eq.(4) and eq. (5), we observe that:

1) The ultimate response (also referred to as s.s.) of a first order system to a sin
input is also a sinusoidal wave with the same frequency ω.
2) The ratio of the output amplitude to the input amplitude is called the “
amplitude ratio” and is a function of the frequency:
K pA
 p2 2  1 Kp
AR  amplitude ratio  
A  p2 2  1
Kp
AR  (6)
 p2 2  1
3) The output wave lags behind (phase lag) the input wave by an angle  >,
which is a function of the frequency  (see eq.(5)).

Process Control /Lec. 11 121 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
 Input = A sin (ωt) 

A B

T Output = B sin (ωt-  )

It is the most important methods for stability analysis and used for design purposes
control system.
Suppose the input to the process is sinusoidal signal
Where: A is amplitude
1
ω is frequency (rad/sec) =
T
T is period of one complete cycle (time)

Frequency Response of a Second Order System

For a second order system the transfer function is:
Kp
G (s)  2 2
 s  2 s  1
Put s=jω then
1) Amplitude Ratio
Kp
AR 
(1   2 2 )2  (2   )2
2) Phase shift
2  
  tan 1 ( )
1   2 2
Which is aphase lag since   0

Frequency Response of a Pure Dead-Time Process

G (s)  e  t d s
Put s=jω
 G ( j )  e  j t d 
AR  1
   t d
That is aphase lag since   0

Process Control /Lec. 11 122 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
Frequency Response of a Feedback Controllers
1- Propertional controller:
The transfer function is G (s)  K c

 AR  K c
 0

2- PI controller:
1
The transfer function is G (s)  K c (1  )
I s
1
 AR  K c 1 
( I ) 2

1
  tan 1 ( ) 0
 I

3- PD controller:
The transfer function is G (s)  K c (1   D s)

AR  K c 1   2D  2

  tan 1 ( D  ) 0
The positive phase shift is called phase lead and implies that the controller output
lead the input.

4-PID controller:
1
The transfer function is G (s)  K c (1    D s)
I s
1
AR  K c ( D  )2  1
 I

1
  tan 1 ( D  )
 I
ϕ is + or – ve depending on the values of τD, τI and ω

Process Control /Lec. 11 123 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
 Bode Diagrams

The bode diagrams consist of a pair of plots showing:

1. How the logarithm of the amplitude ratio varies with frequency.
2. How the phase shift varies with frequency.

First Order system:

Kp
Amplitude ratio AR  (*)
1 2
p
2

Phase lag=   tan 1   p 

AR 1
log   log(1   p2  2 ) (**)
Kp 2
The plot can be carried by considering its asymptotic behaviour as ω0 and as
ω . Then
1. As ω0, then  p   0 and from eq.(*)

AR AR
log  0 or  1.This is the low-frequency asymptote. It is a horizontal line
Kp Kp
AR
passing through the point  1.
Kp
2. As ω, then  p    and from eq.(**)
AR
log   log p  . This is the high frequency asymptote.
Kp
AR
It is a line with slope -1 passing through the point  1for p   1 .
Kp
3. At the corner  p  1    c
1
corner  c 
p
AR 1
The frequency  c is known as the corner frequnecy (and   0.707 )
Kp 2
The phase lag plot
as   0 ,  0
1
as   ,   tan 1 (1)  45o
p
Process Control /Lec. 11 124 Written by Assis. Prof.
Fourth Class Dr. Zaidoon M. Shakor
as    ,   tan 1 ()  90o

𝐴𝑅
𝐾𝑝

𝜙
or
Phase lag

𝜏𝑝 𝜔
Figure:Bode diagram for first-order system.

Second –order system

Kp  2 
AR    tan 1 ( )
(1    )  (2)
2 2 2 2 1   
2 2

ψ=0.1

𝐴𝑅
𝐾𝑝

ψ=0.1

𝜙
or
Phase lag

𝜏𝑝 𝜔

Process Control /Lec. 11 125 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
 1
Figure: Block diagram for second-order system
 s  2s  1
2 2

AR 1
log   log[(1   2 2 ) 2  (2) 2 ]
Kp 2
AR 1
1) as   0 , then log   log(1)  0
Kp 2
AR
  1 stright line of a slope=0 (L.F.A)
Kp
0
  tan 1 0
1
AR 1
2) as    , then log   log ( ) 4  2 log ( ) (H.F.A)
Kp 2
It is a straight line with a slope -2 passing through the pointAR=1 and τω=1
1
3)   c 

Pure dead-time system
For the system
AR  1
   d 

as   0 ,   0
as    ,   
1.0
AR

ϕ -180

-360 | | |
1 10 100
ω
Example: Two systems in series
1 6 0
G1 (s)  and G 2 (s) 
2s  1 5s  1 -180
The overall T.F. is
1 6 -360
G (s)  
2s  1 5s  1
1 6
 AR  
1  4 2 1  25 2
Process Control /Lec. 11 126 Written by Assis. Prof.
Fourth Class Dr. Zaidoon M. Shakor
log AR  log 6  log( AR )1  log( AR ) 2
1
1- Region 1: From ω=0 to   , slope of the overall asymptote =0+0=0
5
(i.e. horizontal * going through the point AR=1)
1 1
2- Region 2: From   to   , slope of the overall asymptote =0+(-1)=-1
5 2
1
going through the point AR=1 ,  
5
1
3- Region 3: From   , slope of the overall asyptote =(-1)+(-1)=-2
2
For ϕ
When as   0 , 1  0 , 2  0,   0
When as    , 1  90 , 2  90,   180
1⁄ 1⁄
1 𝜏2 𝜏1
1 1
5𝑠 + 1 2𝑠 + 1
𝐴𝑅 𝑠𝑙𝑜𝑝𝑒 = −1
6
𝑠𝑙𝑜𝑝𝑒 = −1
𝑠𝑙𝑜𝑝𝑒 = −2
0

-45

𝜙 -90
-135

-180 | |
0.2 ω 0.5

0
Feedback Controller
-180
1-Propertional controller
AR  K c   0-360

𝐴𝑅
1.0
𝐾𝑐

ϕ 0

Process Control /Lec. 11 ω 127 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
0
2-Propertional Integral controller (PI)
1
AR  K c 1 
( I ) 2
1
  tan 1 ( )0
 I
AR 1 1
log( )  log(1  )
Kc 2 (  I )2

1- Low frequency asymptote

1 AR
as   0 ,  1 thenn log( )   log(  I )
(  I ) 2
Kc
Consequently, the LFA is a straight line with slope=-1
1
  tan 1   90o
0
2- High frequency asymptote
1 AR AR
as    ,  0 then  log( )  0 i.e 1
(  I ) 2
Kc Kc
AR
HFA is a horizontal line at value 1
Kc
For the ϕ
as   0 ,   90
as   c ,   45
as    ,   0

Process Control /Lec. 11 128 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
 -1

Figure Bode diagram for PI controller.

2-Propertional Derivative controller (PD)

AR  K c 1   D2  2
  tan 1 (  D )  0
1) Low frequency asymptote

AR 1 AR
as   0 , log( )  log( D2  2 )  0  1 (L.F.A) slope=0
Kc 2 Kc
  tan 1 0  0o
2) High frequency asymptote

AR 1
as    , log( )  log( D2  2 )  log( D ) (H.F.A) slope=+1
Kc 2
  tan 1 0  90o

as   0 ,   0
as   c ,   45o
as    ,   90o

Process Control /Lec. 11 129 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
 +1

ω=1/τD

Figure: Bode diagram for PD controller.

3-Propertional Integral Derivative controller (PID)
AR 1 2
 1  ( D  )
Kc  I
1
  tan 1 ( D  )
 I

AR 1 2
1) as   0 then  1 ( ) PI Controller
Kc  I

AR
2) as    then  1  ( D ) 2 PD Controller
Kc

1 AR
3) as   then  1  ( D  1) 2
I Kc

1 AR 1 2
4) as   then  1  (1  )
D Kc  I
Frequency Response of non-interacting capactine in sereies
G (s)  G1 (s)  G 2 (s)  G 3 (s)  ............  G n (s)
AR t  AR 1  AR 2  AR 3  ............  AR n

log AR t  log AR 1  log AR 2  log AR 3  ............  log AR n

 t  1  2  3  ............  n
Process Control /Lec. 11 130 Written by Assis. Prof.
Fourth Class Dr. Zaidoon M. Shakor
Example: Bode Digram of PID Controller
1
G1 (s)  10(1   5s)
10s
1
c1 (s)   0.1 signal (1)
10
1
c1 (s)   0.2 signal (1)
5

𝐴𝑅
𝐾𝑐 1.0
1⁄ 1⁄
𝜏𝐼 𝜏𝐷

90

+45
ϕ
0

-45

-90
1⁄
𝜏𝐼 ω 1⁄
𝜏𝐷

-180
Example:
Bode plots of
-360open loop system

Gc Gf Gp
YSP(s) 1 10 5e 0.2s
K c (1  ) Y(s)
 Is 0.1s  1 (2s  1)(s  1)

Ym(s)
2
0.5s  1
The Open loop T.F. of the feedback control
G OL  G c G f G p G m
1 1 1 1
G OL  100 K c (1  )   e  0.2s
 Is 0.1s  1 (2s  1)(s  1) (0.5s  1)
With Kc=4 and τI=0.25
K=400

Process Control /Lec. 11 131 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
 1
(1 + )
0.25𝑠

𝑠𝑙𝑜𝑝𝑒 = −1 1
1
0.1𝑠 + 1
𝑠𝑙𝑜𝑝𝑒 = −2
𝐴𝑅 1
𝐾𝑐 0.5𝑠 + 1
1
𝑠𝑙𝑜𝑝𝑒 = −3 𝑠+1
1
2𝑠 + 1
𝑠𝑙𝑜𝑝𝑒 = −4
0.1 0.5 1 2 4 10
0 𝑠𝑙𝑜𝑝𝑒 = −3

-45
𝑠𝑙𝑜𝑝𝑒 = −4
-90
𝜙 .0
-135
0
-180
𝑂𝑣𝑒𝑟𝑎𝑙𝑙
-45

-90

Example: Plot the B.D. for the open loop T.F. for the fig. below
-135

Gc U Gp
-180
R K c (1   Ds) 1
(s  1) (0.1s  1)
2 C

Ym(s) s

e 10

For Kc=10 and τD=0.5 the overall transfer function is

s

10(0.5s  1)e 10
G OL (s) 
(s  1)2 (0.1s  1)
Overall Bode diagram
1  1
G1 (s)  c1   1 stright line slope=-1
1s  1 1
1  1
G 2 (s)  c 2   1 stright line slope=-1
1s  1 1
G 3 (s)  0.5s  1  c3  1  2 stright line slope=+1
0.5
1  1
G 4 (s)  c 4   10 stright line slope=-1
0.1s  1 0.1
s

G 5 (s)  e 10 stright line slope=0

Process Control /Lec. 11 132 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor
Amplitude Ratio Curve Prediction
ω SL1 SL2 SLC SL3 sLd SLTotal
0-1 0 0 0 0 0 0
1-2 -1 -1 0 0 0 -2
2-10 -1 -1 +1 0 0 -1
10- -1 -1 +1 -1 0 -2

Phase Lag Curve Prediction

ω ϕ L1 ϕL2 ϕC ϕ L3 ϕd SLTotal
0 0 0 0 0 - -
1 -45 -45 0 - - -
2 - - 45 - - -
10 - - - -45 - -
 -90 -90 90 -90 - -

Figure: Block diagram for: (a) Amplitude ratio; (b) phase angle.

Process Control /Lec. 11 133 Written by Assis. Prof.

Fourth Class Dr. Zaidoon M. Shakor