DESIGN VIA ROOT LOCUS

PART 1
 Course Outcome
• Be able to interprete design specifications and objectives of
control systems in time and frequency domain. (Aplication)
• Be able to determine the control system strategy based on its
objectives in time and frequency domain. (Analysis)
• Be able to design suitable compensator system/control in
time and fequency domain (Synthesis)
• Be able to evaluate and give comments or opinion about the
developed scheme/control systems (Evaluation)
 Topic-Objectives
• Students should be able to:
– Use the root locus to design cascade
compensators to improve the steady-state error
– Use the root locus to design cascade
compensators to improve the transient response
– Use the root locus to design cascade
compensators to improve both the steady-state
error and the transient response.
 Performance Specifications
• Performance specifications are divided by two
categories:
– Transient Response (maximum overshoot, settling
time)
– Steady State Requirement (steady state error)

**If the accuracy at steady state operation is of prime

importance in a given control system, then we should not
require unnecessarily rigid performance specifications on
the transient response and vice versa.
Flowchart of design
 Design by Root Locus
• The design by root locus is based on reshaping the
root locus of the system by adding poles and zeros
to the system’s OLTF and forcing the root loci to
pass through desired closed loop poles in the s-
plane.
• (Design is being based on the assumption that the
closed loop system has a pair of dominant closed
loop poles)
 Design by Root Locus - Gain Adjustment

Responses from poles at A

and B

• Setting the gain is the first step in adjusting the system for satisfactory
performance. In some systems, simple gain adjustment may move the
closed-loop poles to desired locations.
• Possible design point via gain adjustment (A) from point O. So design can
be made by choosing appropriate gain value.
• Design point that cannot be met via simple gain adjustment (B) (because
point B not on the RL). So - design by addition of a compensator to the
system.
PART 1: IMPROVING STEADY
STATE USING LAG AND PI
COMPENSATOR

CASCADE COMPENSATION
 Improving steady state error via
cascade compensation
• One objective of this design is to improve the steady-
state error without appreciably affecting the
transient response (the neighborhood of the
dominant closed-loop poles should not be changed
appreciably)

• Two techniques to improve the steady state error:

– Lag Compensation
– Ideal Integral Compensation (PI)
 Lag Compensator
• Transfer Function:
s  zc
Gc ( s )  K c   1 Kc  1 zc  pc
s  pc

• Pole and zero are placed very closed to origin.

Example of RL Plots of Compensated and
Uncompensated systems
Design Procedure for Lag Compensator

KV  lim sG ( s)
s 0
Design Procedure for Lag Compensator
 Lag Compensator

KC K̂

Kc  1

(a) Original system before compensation

(b) System after compensation
(a) Uncompensated system
(b) Compensated system using lag compensator

The effect on the transient response

 The uncompensated system’s root locus is shown in Figure (a), where the point P is
assumed to be the dominant pole.
 Figure (b), where the compensator has been added, point P is still at approximately
the same location on the compensated root locus.
(a) Uncompensated system
(b) Compensated system using lag compensator

The effect on the required gain

 After inserting the compensator, we find that K is virtually the same for the
uncompensated and compensated systems (not much difference), since
lengths of the vectors drawn from the lag compensator are approximately
equal, and all other vectors have not changed appreciably.
(a) Uncompensated system TYPE 1
(b) Compensated system using lag compensator KV  lim sG ( s )
s 0

Kz1 z 2 ...
K v ,unc 
p1 p2 ...

( Kz1 z 2 ...)( zc )
K v ,comp 
( p1 p2 ...)( pc )

K v ,comp zc
 
K v ,unc pc
The effect in the steady-state error
• Since we established that the gain, K, is about the same for the uncompensated
and compensated systems, we obtain:
zc 1
K v ,comp  K v ,unc ess ,comp  Steady state error
pc K v ,comp decrease
K v ,comp  K v ,unc
 Example 2
• Compensate the system below to improve the steady-state
error by a factor of 10 if the system is operating with a
damping ratio of 0.174. Find the lag compensator transfer
function.

Type = 0

**RL from Example 1 in Revision 1

 Steps:
(1) Draw the uncompensated system root locus and analyze the system at =0.174

K
G(s) 
( s  1)( s  2)( s  10)

At  = 0.174, s = -0.694 +j3.926

1
K unc  K
G ( s) H ( s)
( s  1)( s  2)( s  10)

1 s  0.649  j 3.926

 164.6
 K
Steps: G (s) 
( s  1)( s  2)( s  10)
(2) Find the static position constant & its steady state

K P ,unc  lim G ( s) 1
s 0 ess ,unc 
1 KP
K
 1
( s  1)( s  2)( s  10) 
1  8.23
164.6
  8.23  0.108
1x 2 x10

OLTF :
164.6
G ( s ) unc 
( s  1)(s  2)( s  10)
We may derive the
OLTF & CLTF for CLTF :
uncompensated 164.6
T ( s ) unc 
system s 3  13s 2  32 s  184.6
164.6

( s  0.694  j 3.96)( s  0.694  j 3.96)( s  11 .61)
 Steps:
(3) Specs: improve the steady-state error by a factor of 10

0.108
ess _ desired   0.0108
10

Type 0
1  e() 1  0.0108
1 K p _ desired    91.59
e ( )  e( ) 0.0108
1 KP

zc
K p _ desired  K p _ unc
pc
zc K p _ desired 91.59
   11 .13
pc K p _ unc 8.23
 Steps:
(4) The location of the pole and zero should be close to origin and z c  pc

s  zc
Gc ( s )  K c Kc  1
s  pc

Just choose
pc  0.01
 so zc  11 .13 pc  0.111
s  0.111
Glag ( s )  K C . Kc  1
s  0.01
(5) Draw the RL for compensated system and analyze

OLTF for compensated system:

At   0.174, s  0.678  j 3.836
K  158.1
158.1( s  0.111)
GC ( s )G ( s ) 
( s  0.01)( s  1)( s  2)( s  10)

CLTF for compensated system:

158.1( s  0.111)
Tcomp ( s) 
( s  0.678  j 3.836)( s  0.678  j 3.836)( s  11 .55)( s  0.101)
 Conditions and approximations
Conditions for reducing orders:

• If the negative real part of the pole of the system is further away from the
imaginary axis, the corresponding items of the pole will attenuate faster in
the transient response. On the other hand, the closed-loop pole nearest
to the imaginary axis, which correspond to the item that attenuates most
slowly in the transient response, is called a dominant pole. Typically, in
engineering, pole A can be ignored in system analysis if it is five times
further from the imaginary axis than pole B
• In the transfer function of the system, if the numerator and denominator
have a zero a pole that have negative real parts and are similar in values,
then the zero and pole can be cancelled together. Which is called dipole
cancellation.

Valid 2nd order approximations

 Evaluate your develop compensator &
Comment
• Comparing:

CLTF  Uncompensated
164.6
T (s) 
( s  0.694  j 3.96)( s  0.694  j 3.96)( s  11 .61)

CLTF  Compensated
158.1( s  0.111)
TC ( s ) 
( s  0.678  j 3.836)( s  0.678  j 3.836)( s  11 .55)( s  0.101)
• Valid 2nd order? Comment!
• The remaining closed-loop poles of the compensated system very close in value to
the closed-loop poles of the uncompensated system.
• So, the transient response of both system is approximately the same as is the gain.
• Steady state error improve??Verify! Transient response changes??
• RL for compensated system
• Characteristics of uncompensated and lag-compensated systems for
Example 2

91.59
• Step responses of uncompensated and lag-compensated
systems for Example 2
 K̂

K  Kˆ

K̂

K  K c Kˆ

K  K c Kˆ
158.1
Kc 
164.6
 0.961
1
 Exercise 2
• The feed forward transfer function is

1.06
G(s) 
s ( s  1)( s  2)
• The system is operating with a damping ratio of 0.491 and its dominant
closed-loop poles is s=-0.3307+j0.5864.
• It is desired to increase the static velocity error constant Kv to about 5sec-1
Place the zero of the lag compensator at s=-0.05.
• Find the open-loop transfer function of the compensated system (Given
the dominant closed-loop pole from the new root locus (compensated
system) is s = -0.31+j0.55
• Verify that the steady state error has decreased.
• Uncompensated S d  0.3307  j 0.5864

1.06
KV ,unc  sG ( s )   0.53
s 0 (1)(2)
ess  1.88

• Compensated
KV ,desired  5.0 sec1

 zc 
KV ,desired  K v ,unc  
 pc 
KV ,unc 0.53  0.05
pc   zc   0.005
KV ,desired 5.0
  s  0.05 
Lag Compensator TF: Gc ( s )  K  
 s  0.005 

( s  0.05)
OLTF for compensated system: Gc ( s )G ( s )  K
s ( s  0.005)( s  1)( s  2)

s ( s  0.05)( s  1)( s  2)
K
OLTF Gain: ( s  0.05) sd  0.31 j 0.55

K  1.0235

( s  0.05)
Gcomp ( s )  Gc ( s )G ( s )  1.0235
s ( s  0.005)(s  1)(s  2)

K v ,comp  lim sGcomp ( s )  5.12

Verify: s 0

ess  1  0.195
K v ,comp
 PI Controller
 KI 

KPs  
1  K P  K (s  zc )
Transfer function : Gc ( s)  K p  K I  
s s s

R(s)＋ Kp ＋ C(s)
G(s)
－ ＋
1
KI GC (s )
s zc pc

Block Diagram of PI Controller

• In practical, the PI controller zero is placed very close to its pole

located at the origin so that the angular contribution to the root
locus is almost zero.
 PI Controller
Pole at A is:
a. on the root locus without compensator
b. not on the root locus with compensator pole at origin added
c. on the root locus with PI controller (pole at origin and zero)
 Example 3
• Given the system below operating with a damping ratio of
0.174. Show that the addition of the ideal integral
compensator with zero at -0.1 will yield the following
specifications:
– steady-state error will be zero for a step input
– Percent overshoot less than 57.4%
 Steps:
1) Draw the uncompensated system root locus and analyze the system at
=0.174 (same as Example 2)

ess , unc  0.108

CLTF uncompensated
164.6
Tunc ( s ) 
s 3  13s 2  32s  184.6
164.6

( s  0.694  j 3.96)( s  0.694  j3.96)( s  11.61)
 Step response for uncompensated system
Step Response
1.4

System: T
Peak amplitude: 1.37
1.2 Overshoot (%): 54
At time (sec): 0.88

0.8 System: T System: T

Amplitude

Settling Time (sec): 5.14 Final Value: 0.892

0.6 num=164.6;
den=poly([-1 -2 -10]);
0.4
G1=tf(num,den);
G=zpk(G1)
T1=feedback(G,1);
0.2
T=zpk(T1)
step(T)
0
0 1 2 3 4 5 6 7 8
Time (sec)
2) Block diagram of compensated system

Type = 1
Pole at the origin will increase the system type
Choose zero at -0.1 (close the compensator pole, so that the angular contribution
is approximately zero)
3) Draw the compensated system root locus and analyze

1
K
G (s) H (s)
s ( s  1)( s  2)( s  10)

( s  0.1) s  0.678 j 3.837

 158.2

1
ess  lim sR ( s )
s 0 1  G (s)
s 1
ess  lim 0
CLTF compensated
s 0 s 1  ( s  0.1)G0 ( s)
s
158.2( s  0.1)
T (s) 
s ( s  1)( s  2)( s  10)  158.2
158.2( s  0.1) Reduces steady state

( s  0.678  j 3.837)( s  11 .55)( s  0.0902) error to zero
 Steps:
4) Compare & comment

CLTF uncompensated
164.6
T ( s)  3
s  13s 2  32 s  184.6
164.6

( s  0.694  j 3.96)( s  11 .61)

CLTF compensated Pole zero

158.2( s  0.1) cancelation
T ( s) 
s ( s  1)( s  2)( s  10)  158.2
158.2( s  0.1)

( s  0.678  j 3.837)( s  11 .55)( s  0.0902)

Assumptions: CL poles and gain of uncompensated and compensated

are approximately the same …so, indicates the transient response of
the compensated is about the same as the uncompensated.
Valid 2nd order approx.
 Verify by step response-compensated system
Step Response
1.4
System: T1
Peak amplitude: 1.4
1.2 Overshoot (%): 39.6
At time (sec): 0.901 System: T1
Final Value: 1
1
System: T1
Settling Time (sec): 17.8 %pole @ -0.1
0.8
num1=158.2*[1 0.1];
Amplitude

den1=poly([0 -1 -2 -10]);
0.6 G11=tf(num1,den1);
G1=zpk(G11)
T11=feedback(G11,1);
0.4
T1=zpk(T11)
step(T1)
0.2 rltool
ltiview
0
0 5 10 15 20 25 30 35
Time (sec)
Specs:
1) steady-state error will be zero for a step input 
2) percent overshoot less than 57.4% 
 Comparison: Uncompensated (T2) and
PI compensated system (T)
System: T
Peak amplitude: 1.4
Overshoot (%): 39.6
At time (sec): 0.901 Step Response
1.4

System: T2
Peak amplitude: 1.37 PI compensated
Overshoot (%): 54
1.2 At time (sec): 0.88
%pole @ -0.1
num=158.2*[1 0.1]
System: T System: T den=poly([0 -1 -2 -10]);
Final Value: 1
1
Settling Time (sec): 17.8
G1=tf(num,den);
G=zpk(G1)
T1=feedback(G,1);
System: T2
0.8
System: T2
Final Value: 0.892
T=zpk(T1)
Settling Time (sec): 5.14
Amplitude

Uncompensated
0.6
num2=164.6
den2=poly([-1 -2 -10]);
G22=tf(num2,den2);
0.4
G2=zpk(G22)
T22=feedback(G2,1);
T2=zpk(T22)
0.2
step(T,T2)

0
0 5 10 15 20 25 30 35
Time (sec)
 Effect on different zero location
Step Response
1.6

1.4

1.2
System: T2
Settling Time (sec): 8.61
1
System: T3 System: T1
Amplitude

Settling Time (sec): 6.95 Settling Time (sec): 17.8

0.8
T1 (B):zero @ -0.1
0.6
T2(H): zero @ -0.2
0.4 T3(M):zero @ -0.4

0.2

0
0 5 10 15 20 25 30 35
Time (sec)
%zero @ -0.1
num1=158.2*[1 0.1];
den1=poly([0 -1 -2 -10]);
G11=tf(num1,den1); %zero @ -0.4
G1=zpk(G11) num3=159.5*[1 0.4];
T11=feedback(G11,1); den3=poly([0 -1 -2 -10]);
T1=zpk(T11) G33=tf(num3,den3);
G3=zpk(G33)
T33=feedback(G33,1);
%zero @ -0.2 T3=zpk(T33)
num2=158.7*[1 0.2];
den2=poly([0 -1 -2 -10]); step(T1,T2,T3)
G22=tf(num2,den2);
G2=zpk(G22)
T22=feedback(G22,1);
T2=zpk(T22)
 Exercise 3
• The system below operates with a damping ratio
of 0.5. Find its steady state error .
K
G(s) 
( s  1)( s  3)( s  10)
• Then, design a PI controller to drive the step
response error to zero for the unity feedback
below:
• Compare the specifications of the uncompensated
and compensated systems.
 Exercise 3_RL_Part 1
10
9
8
7
6
5
4
3
2
Imag Axis

1
0
-1
-2
-3
-4
-5
-6
-7 Scale
-8
-9
X:Y
-10 1:1
-15-14-13-12-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5
Real Axis
 Uncompensated
Root Locus Editor for Open-Loop 1 (OL1)
10
0.76 0.64 0.5 0.34 0.16
0.86
8

6
0.94

0.985 num1=1
2
den1=poly([-1 -3 -10]);
G11=tf(num1,den1);
Imag Axis

14 12 10 8 6 4 2
0
G1=zpk(G11)
T11=feedback(G11,1);
-2
0.985 T1=zpk(T11)
rltool
-4

0.94
-6

-8
0.86
0.76 0.64 0.5 0.34 0.16
-10
-15 -10 -5 0 5
Real Axis
 Uncompensated
System: Closed-Loop r to y
I/O: r to y
Peak amplitude: 0.819 Step Response
0.9 Overshoot (%): 15.5
At time (sec): 1.31
System: Closed-Loop r to y
0.8 I/O: r to y
Final Value: 0.709
0.7
System: Closed-Loop r to y
0.6 I/O: r to y
Settling Time (sec): 2.71

0.5
Amplitude

0.4

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Time (sec)
 Compensated
Root Locus Editor for Open-Loop 1 (OL1)
8
0.85 0.74 0.62 0.48 0.32 0.16

6
0.93

0.98
2
zero@-0.1
num2=[1 0.1];
Imag Axis

12 10 8 6 4 2
0
den2=poly([0 -1 -3 -10]);
G22=tf(num2,den2);
-2
0.98
G2=zpk(G22)
T22=feedback(G22,1);
-4
T2=zpk(T22)
0.93 rltool
-6

0.85 0.74 0.62 0.48 0.32 0.16

-8
-12 -10 -8 -6 -4 -2 0
Real Axis
 Compensated
System: Closed-Loop r to y
I/O: r to y
Step Response Final Value: 1
1

System: Closed-Loop r to y
0.9 I/O: r to y
Settling Time (sec): 36.2
0.8

0.7

0.6
Amplitude

0.5

0.4

0.3

0.2

0.1

0
0 10 20 30 40 50 60 70
Time (sec)