IntroCtrlSys_Chapter4

Lecture Notes
Introduction to Control Systems
Instructor: Assoc. Prof. Dr. Huynh Thai Hoang

Department of Automatic Control
Faculty of Electrical & Electronics Engineering
Ho Chi Minh City University of Technology
Email: hthoang@hcmut.edu.vn
© H. T. Hoang - HCMUT 1
Chapter 4
ANALYSIS OF CONTROL SYSTEM

PERFORMANCE
Content
 Performance criteria
 Steady state error
 Transient response
 The optimal performance index
 Relationship between frequency domain performances and
time domain performances.
Performance criteria
Performance criteria: Steady state error
yfb(t)
Y(s) ess
R(s) E(s) r(t)
+_ G(s)
Yfb(s) e(t) ess

H(s) t
0
 Error: is the difference between the set-point (input) and the

feedback signal.
e(t )  r (t )  y fb (t )  E ( s )  R ( s )  Y fb ( s )
 Steady-state error: is the error when time approaching infinity.

ess  lim e(t )  ess  lim sE ( s )
t  s 0
Performance criteria – Percent of Overshoot (POT)
 Overshoot: refers to an output exceeding its steady-state value.

y(t) y(t)
overshoot
ymax
ymax yss
yss yss
yss
No overshoot t t
0 0
 Percentage of Overshoot (POT) is an index to quantify the

overshoot of a system, POT is calculated as:
y max  y ss
POT   100%
y ss
Performance criteria – Settling time and rise time
 Settling time (ts): is the time required for the response of a

system to reach and stay within a range about the steady-
state value of size specified by absolute percentage of the
steady-state value (usually 2% or 5%)
 Rise time (tr): is the time required for the response of a
system to rise from 10% to 90% of its steady-state value.
y(t) y(t)
(1+)yss
yss yss
(1)yss 0.9yss
0.1yss t
t
0
0
ts tr
Steady-state error
Steady-state error
R(s) E(s) Y(s)

+_ G(s)
Yfb(s)
H(s)
R( s)
 Error expression: E (s) 
1  G(s) H (s)
sR( s )
 Steady-state error: ess  lim sE ( s )  lim
s 0 s 0 1  G ( s ) H ( s )
 Remark: Steady-state error not only depends on the structure

and parameters of the system but also depends on the input
signal.
Steady-state error to step input
 Step input: R( s )  1 / s
1
 Steady-state error: ess 
1 Kp
with K p  lim G ( s ) H ( s ) (position constant)
s 0
yfb(t) yfb(t)
1 1
t t
0 0
G(s)H(s) does not have G(s)H(s) has at least 1
any deal integral factor ideal integral factor
Steady-state error to ramp input
 Ramp input: R( s )  1 / s 2
1
 ess  with K v  lim sG ( s ) H ( s ) (velocity constant)
Kv s 0
yfb(t) yfb(t) yfb(t)

r(t) r(t) r(t)
ess  0 ess = 0
e(t) 
t t t
0 0 0
G(s)H(s) does not have G(s)H(s) has 1 ideal G(s)H(s) has at least
deal integral factor integral factor 2 ideal integral factors
Steady-state error to parabolic input
 Parabolic input: R( s )  1 / s 3
1 2 (acceleration
 ess  with K a  lim s G ( s ) H ( s )
Ka s 0 constant )
yfb(t) yfb(t) yfb(t)

r(t) r(t) r(t)
ess0 ess= 0
e(t) 
t t t
0 0 0
G(s)H(s) has less than G(s)H(s) has 2 ideal G(s)H(s) has more than
2 ideal integral factors integral factors 2 ideal integral factors
Example
R(s) E(s) Y(s)
+_ G(s)
25
G (s) 
( s  2)( s  5)
 Calculate the steady state error to step input and ramp input
 Solution:
 Step input:
40 1 1
K P  lim G ( s )  lim 4  ess    0.2
s 0 s 0 ( s  2)( s  5) 1 KP 1 4
 Ramp input:
40 1 1
K v  lim sG ( s )  lim s 0  e   
s 0 ( s  2)( s  5)
ss
s 0 KV 0
Simulation result
Exercise
R(s) E(s) Y(s)
+_ G(s)
40
G (s) 
( s  2)( s  5)
 Solution:
 Step input:
40 1 1
K P  lim G ( s )  lim 4  ess    0.2
s 0 s 0 ( s  2)( s  5) 1 KP 1 4
 Ramp input:
40 1 1
K v  lim sG ( s )  lim s 0  e   
s 0 ( s  2)( s  5)
ss
s 0 KV 0
Exercise 1
R(s) E(s) Y(s)

+_ GC(s) G(s)
1 40
GC ( s )  0.5  G (s) 
s ( s  2)( s  5)
 Simulate the system by using Matlab to verify the result
Solution
 Step input:
 1 40
K P  lim GC ( s ).G ( s )  lim  0.5   
s 0 s 0
 s  ( s  2)( s  5)
1 1
 ess   0
1 KP 1 
 Ramp input:
 1 40
K v  lim sGC ( s )G ( s )  lim s  0.5   4
s 0 s 0
 s  ( s  2)( s  5)
1 1
 ess    0.25
KV 4
Simulation results
Exercise 2
R(s) E(s) Y(s)

+_ GC(s) G(s)
10
GC ( s )  2 s  1 G (s) 
( s  2) 2 ( s  5)
 Simulate the system by using Matlab to verify the result
Transient response
First-order system
R(s) K Y(s)
Ts  1
K
 Transfer function: G (s) 
Ts  1
1
 First order system has 1 real pole: p1  
T
1 K
 Transient response: Y ( s )  R ( s )G ( s )  .
s Ts  1
 y (t )  K (1  e  t /T )
First-order system (cont’)
Im s y(t)
(1+).K
K
(1).K
Re s
0 0.63K
1/T
t
0
T ts
Pole – zero plot Transient response

of a first order system of the first order
y (t )  K (1  e t /T )
First-order system – Remarks
 First order system has only one real pole at (1/T), its
transient response doesn’t have overshoot.
 Time constant T: is the time required for the step response of
the system to reach 63% its steady-state value.
 The further the pole (1/T) of the system is from the
imaginary axis, the smaller the time constant and the faster
the time response of the system.
 Settling time of the first order system is:
1
ts  T ln 
 
where  = 0.02 (2% criterion) or  = 0.05 (5% criterion)
First-order system
The relationship between the pole and the time response
 The further the pole of the system is from the imaginary axis,
the smaller the time constant and the faster the time
response of the system.
Im s y(t)
K
Re s
0
t
0
Pole – zero plot Transient response

of a first order system of the first order
Second-order oscillating system
R(s) K Y(s)
T 2 s 2  2Ts  1
 The transfer function of the second-order oscillating system:

K Kn2 1
G( s)  2 2  2 (n  , 0    1)
T s  2Ts  1 s  2n s  n
2
T
 The system has two complex conjugate poles:
p1, 2  n  jn 1   2
1 Kn2
 Transient response: Y ( s )  R ( s )G ( s )  . 2
s s  2n s  n2


y (t )  K 1 

e nt
1 2

sin (n 1   2 )t   



(cos    )
Second-order oscillating system (cont’)
y(t)
Im s
cos = 
j n 1   2 (1+).K
n K
(1).K
 Re s
n 0
 j n 1   2
t
0
ts
Pole – zero plot of a second Transient response of a second

order oscillating system order oscillating system
Second-order oscillating system – Remark
 A second order oscillation system has two conjugated
complex poles, its transient response is a oscillation signal.
 If  = 0, transient response
is a stable oscillation signal =0
at the frequency n  n  = 0.2
is called natural oscillation  = 0.4
frequency.
 If 0<<1, transient
response is a decaying
oscillation signal   is
called damping constant,  = 0.6
the larger the value  , (the
closer the poles are to the
real axis) the faster the
response decays.
Second-order oscillating system – Overshoot
 Transient response of the second order oscillating system

has overshoot.
  

The percentage of overshoot: POT  exp  .100%
 1   2 
 
 The larger the value ,
(the closer the poles are
POT (%)
to the real axis) the

smaller the POT.
 The smaller the value ,
(the closer the poles are
to the imaginary axis)

the larger the POT
The relationship
between POT and 
Second-order oscillating system – Settling time
 Settling time:
3
5% criterion: ts 
n
4
2% criterion: ts 
n
Relationship between pole location and transient response
 The 2nd order systems that have the poles located in the same
rays starting from the origin have the same damping constant,
then the percentage of overshoots are the same. The further
the poles from the origin, the shorter the settling time.
Im s y(t)
K
cos =  Re s

0
t
0
Relationship between pole location and transient response (cont’)
distance from the origin have the same natural oscillation
frequency. The closer the poles to the imaginary axis, the smaller
the damping constant, then the higher the POT.
Im s y(t)
n K
Re s
0
t
0
Relationship between pole location and transient response (cont’)
distance from the imaginary axis have the same n, then the
settling time are the same. The further the poles from the real
axis, the smaller the damping constant, then the higher the POT
Im s y(t)
K
Re s
n 0
t
0
Example
R(s) E(s) Y(s)
+_ G(s) 25
G (s) 
s ( s  5)
 Calculate the POT and settling time (2% criterion)
 Solution:
 Characteristic equation:
25
1  G (s)  0  1   0  s 2  5s  25  0
s ( s  5)
 s1,2  2.5  j 4.33
 Complex poles have general form: 1,2
s   n  j n 1   2
n  2.5   0.5

 

 n 1   2
 4.33 n  5
Example (cont.)
 POT:
    0.5   
POT  exp    100%  exp    100%  16.3%
 1 
2
 1  0.5 2

 
 Settling time:
4 4
ts    1.6(sec)
n 0.5  5
Simulation result
1.163
1.6 (sec)
Exercise 1
R(s) E(s) Y(s)
+_ G(s) 50
G (s) 
( s  1)( s  5)
 Solution:
50
1  G (s)  0  1   0  s 2  6 s  55  0
( s  1)( s  5)
 s1,2  3.5  j 6.144
 Complex poles have general form: 1,2
s   n  j n 1   2
n  3.0   0.4

 

 n 1   2
 6.78 n  7.42
Solution
 POT:
    0.5   
POT  exp    100%  exp    100%  25.38%
 1  
2
 1  0.5 2

 
 Settling time:
3 3
ts    1.0(sec)
n 0.4  7.42
Simulation result
ymax=1.135
yss=0.91
ymax  yss
POT   100%  24.97%
yss
Exercise 2
R(s) E(s) Y(s)
+_ G(s) 20
G (s) 
s ( s  3)
 Simulate the system using Matlab to verify the result
Transient response of high order system
 High-order systems are the system that have more than 2 poles
 If a high order system have a pair of poles located closer to the
imaginary axis than the others then the high order system can be
approximated to a second order system. The pair of poles
nearest to the imaginary axis are called the dominant poles.
Im s y(t)
Response of high
order system
Re s
0 Response of second
order system with
the dominant poles
t
0
High order systems A high order system can be
have more than 2 poles approximated by a 2nd order system
Example
R(s) E(s) Y(s)
+_ G(s) 100
G (s) 
s ( s  2)( s  5) 2
 Solution:
 Characteristic equation: dominant Im s
100 poles
1  G (s)  0  1  0
s ( s  2)( s  5) 2
Re s
 s  12 s  45s  50 s  100  0
4 3 2
0
 s1,2  0.253  j1.637

 s3,4  5.747  j1.846
Example (cont.)
s1,2  0.253  j1.637
 Dominant poles have general form: 1,2

s   n  j n 1   2
n  0.253   0.153

 
n 1    1.637
2
n  1.656
 POT:
    0.153   
POT  exp    100%  exp    100%  61.48%
 1  2   1  0.1532

 
 Settling time:
3 3
ts    11.86(sec)
n 0.253
Simulation result
Exercise 1
R(s) E(s) Y(s)

+_ GC(s) G(s)
10 1
GC ( s )  1  G (s) 
s ( s  2)( s  5)
Solution
 Solution:
1  GC ( s )G ( s )  0 Im s
dominant
 10  1 poles
 1  1   0
 s  ( s  2)( s  5)
Re s
 s ( s  2)( s  5)  ( s  10)  0
0
 s  7 s  11s  10  0
3 2
 s1,2  0.863  j1.073


 s3  5.274
Solution
s1,2  0.863  j1.073
 Dominant poles have general form: 1,2

s   n  j n 1   2
n  0.863   0.627

 
n 1    1.073
2
n  1.377
 POT:
    0.627   
POT  exp    100%  exp    100%  7.98%
 1  2   1  0.627 2

 
 Settling time:
4 4
ts    4.64(sec)
n 0.863
Simulation result
Exercise 2
R(s) E(s) Y(s)

+_ GC(s) G(s)
40
G (s) 
s ( s  3)( s  8)
 Simulation the system using Matlab to verify the result
Performance indices
Integral performance indices
 IAE criterion
(Integral of the Absolute Magnitude of the Error )

J IAE   e(t ) dt
0
 ISE criterion
(Integral of the Square of the Error)

J ISE   e 2 (t )dt
0
 ITAE criterion
(Integral of Time multiplied by the Absolute Value of the Error)

J ITAE   t e(t ) dt
0
Optimal systems
 A control system is optimal when the selected performance
index is minimized
 Second order system: J IAE  min when   0.707
J ISE  min when   0.5
J ITAE  min when   0.707
=0.3
y(t)
=0.5
=0.707
=0.9
t
0
Transient response of second order systems

ITAE optimal control
 ITAE is usually used in design of control system

 An n-order system is optimal according to ITAE criterion if the
denominator of its transfer function has the form:
ITAE optimal control (cont’)
 Optimal response according to ITAE criterion
y(t)
1st order system
2nd order system
3rd order system
4th order system
t
0
Relationship between frequency domain
performances & time domain performances
Relationship between frequency response & steady state error
R(s) Y(s)
+ G(s)
K p  lim G ( s ) H ( s )  lim G ( j ) H ( j )
s 0  0
K v  lim s G ( s ) H ( s )  lim jG ( j ) H ( j )

s 0  0
K a  lim s G ( s ) H ( s )  lim( j ) 2 G ( j ) H ( j )
2
s 0  0
Relationship between frequency response & steady state error
R(s) Y(s)
+ G(s)

 Steady state error of the closed-loop system depends on the

magnitude response of the open-loop system at low
frequencies but not at high frequencies.
 The higher the magnitude response of the open-loop system
at low frequencies, the smaller the steady-state error of the
closed-loop system.
 In particular, if the magnitude response of the open-loop system
is infinity as frequency approaching zero, then the steady-state
error of the closed-loop system to step input is zero.
Relationship between frequency response & transient response
R(s) Y(s)
+ G(s)
 In the frequency range  <c , because G ( j )  1 then:

G ( j ) G ( j )
Gcl ( j )   1
1  G ( j ) G ( j )
 In the frequency range  >c , because G ( j )  1 then:
G ( j ) G ( j )
Gcl ( j )    G ( j )
1  G ( j ) 1
 Bandwidth of the closed-loop system is approximate the gain
crossover frequency of the open-loop system.
Bode plot of a open-loop system Bode plot of the corresponding

closed-loop system
R(s) Y(s)
+ G(s)
 The higher the gain crossover frequency of open-loop system,

the wider the bandwidth of closed-loop system  the faster the
response of close-loop system, the shorter the settling time.
 4
 ts 
C C
 The higher the phase margin of the open-loop system, the
smaller the POT of closed-loop system. In most of the cases,
if the phase margin of the open-loop system is larger than 600
then the POT of the closed-loop system is smaller than 10%.
Example 1
R(s) Y(s)
+ G(s)
10
G (s) 
s (0.1s  1)(0.08s  1)
 Estimate POT and settling time using frequency response

 Solution
 Gain crossover frequency:
10
G ( jC )  1  1
jC (0.1 jC  1)(0.08 jC  1)
10
 1  C  7.1(rad / sec)
C (0.1C ) 2  1 (0.08C ) 2  1
Example 1
 Phase margin:
M=180+G ( jC )
 M=180  (90o  tan 1 (0.1C )  tan 1 (0.08C ))
 M  180o  155o  25o
 Overshoot:
POT  10% because M  60 o
 Setting time:
 4  4
 ts    ts 
C C 7.1 7.1
 0.44  ts  1.77(sec)
Example 1: Matlab verification
Example 2
R(s) Y(s)
+ G(s) 6
G (s) 
s (0.1s  1)

 Solution
6
G ( jC )  1  1
jC (0.1 jC  1)
6
 1  C  5.3(rad / sec)
C (0.1C ) 2  1
Example 2
 Phase margin:
M=180+G ( jC )
 M=180  (90o  tan 1 (0.1C ))
 M  180o  118o  62o
 Overshoot:
POT  10% because M  60 o
 Setting time:
 4  4
 ts    ts 
C C 5.3 5.3
 0.59  ts  2.37(sec)
Exercise 1
R(s) Y(s)
+ G(s) 10e 0.1s
G (s) 
s ( s  2)

 Solution
10
G ( jC )  1  1
jC ( jC  2)
10
 1  C  2.87(rad / sec)
C (C ) 2  4
Exercise
 Phase margin:
M=180+G ( jC )
 1  C  0.1C  180 
 M=180   90  tan 
o
 
  2   
 M  180o  162o  18o
 Overshoot: POT  10% because M  60 o
 Setting time:
 4  4
 ts    ts 
C C 2.87 2.87
 1.09  ts  4.36(sec)
Exercise 2
R(s) Y(s)
+ G(s) 20e 0.2 s
G (s) 
( s  1)( s  10)

 Simulate the system using Matlab to verify the result
End of Chapter 4

IntroCtrlSys_Chapter4

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

IntroCtrlSys_Chapter4

Uploaded by

Copyright:

Available Formats

Lecture Notes

Introduction to Control Systems

Instructor: Assoc. Prof. Dr. Huynh Thai Hoang

ANALYSIS OF CONTROL SYSTEM

Yfb(s) e(t) ess

 Error: is the difference between the set-point (input) and the

 Steady-state error: is the error when time approaching infinity.

 Overshoot: refers to an output exceeding its steady-state value.

 Percentage of Overshoot (POT) is an index to quantify the

 Settling time (ts): is the time required for the response of a

R(s) E(s) Y(s)

 Remark: Steady-state error not only depends on the structure

yfb(t) yfb(t) yfb(t)

yfb(t) yfb(t) yfb(t)

R(s) E(s) Y(s)

R(s) E(s) Y(s)

Pole – zero plot Transient response

Pole – zero plot Transient response

 The transfer function of the second-order oscillating system:

Pole – zero plot of a second Transient response of a second

 Transient response of the second order oscillating system

to the real axis) the

n  2.5   0.5

n  3.0   0.4

 Dominant poles have general form: 1,2

n  0.253   0.153

R(s) E(s) Y(s)

 s1,2  0.863  j1.073

 Dominant poles have general form: 1,2

n  0.863   0.627

R(s) E(s) Y(s)

Transient response of second order systems

 ITAE is usually used in design of control system

 Optimal response according to ITAE criterion

2nd order system

3rd order system

4th order system

K v  lim s G ( s ) H ( s )  lim jG ( j ) H ( j )

 Steady state error of the closed-loop system depends on the

 In the frequency range  <c , because G ( j )  1 then:

Bode plot of a open-loop system Bode plot of the corresponding

 The higher the gain crossover frequency of open-loop system,

 Estimate POT and settling time using frequency response

 Estimate POT and settling time using frequency response

 Estimate POT and settling time using frequency response

 Estimate POT and settling time using frequency response

You might also like