EE3114

Systems and Control

3. Dynamic Response

Semester B, 2018-19

2018-9-19
 Framework of Lectures
System

Time Domain Frequency Domain

à Laplace Transform à

State-Space Formulation Frequency Analysis

(dynamic responses, transients,
steady-states, root-locus,
PID controllers)
ß mixed à
(State-feedback,
Controllability, controller Today:
Observability, observer
Dynamic responses
Optimal control)
Given Transfer Function G(s) and Input U(s):
U(s) Y(s)
G(s)

Y(s) = G(s)U(s)

Output obtained by Inverse Laplace Transform:

1
y (t ) g (t ) u (t ) L {G ( s )U ( s )}
Convolution
 Impulse Response
Impulse input Impulse response
𝑢(𝑡)= d(t) h(t)=?
G(s)

U(s) = 1 for 𝑢 𝑡 = 𝛿(𝑡)

H(s) = G(s)U(s) = G(s)
h(t ) L 1[G ( s )]

Impulse response of the system is obtained by

Inverse Laplace Transform of Transfer Function
 d(t) h(t)=?
Example: 1
s +1

1
H (s) 1
Impulse Response

s 1 0.9

(t ) e t
hInverse Laplace 0.8
Transform of H(s)
0.7
à
Amplitude 0.6

0.5

0.4

0.3

%MatLab code 0.2

Hs = tf(1,[1 1]); 0.1
impulse(Hs);
0
0 1 2 3 4 5 6
Time (sec)
 Step Response
Step input Step response
𝑢 𝑡 =1 y(t)=?
G(s)

!
U(s) = for u(t) = 1 (unit step)
"
!
Y(s) = G(s)U(s) = G(s)
"
1 G (s)
y (t ) L
s
Step response is the integral of the impulse response
(Integration property of Laplace Transform)
 1 y(t)=?
Example 1
s +1

Step Response
1
1
Y (s) 0.9

s ( s 1) 0.8

t 0.7
y (t ) 1 e 0.6
Amplitude
0.5

0.4

0.3
%MatLab code
0.2
Hs = tf(1,[1 1]);
step(Hs); 0.1

[y,t] = step(Hs,6); %Time [0,6] 0

0 1 2 3 4 5 6
plot (t, y); Time (sec)
 Ramp Response
Ramp input Ramp response
𝑢 𝑡 =t y(t)=?
G(s)

!
U(s) = ! for u(t) = t
"
!
Y(s) = G(s)U(s) = G(s) !
"
1 G (s)
y (t ) L
s
Ramp response is the integral of the step response
(Integration property of Laplace Transform)
 t y(t)
Example: 1
𝑇𝑠 + 1 (Constant 𝑇 > 0)

Exercise: (5 minutes)
𝑦 𝑡 =?
 t y(t)
Example: 1
𝑇𝑠 + 1 (Constant 𝑇 > 0)

Exercise: (5 minutes)
𝑦 𝑡 =?
 t y(t)
Example: 1
𝑇𝑠 + 1 (Constant 𝑇 > 0)

Exercise:
𝑦 𝑡 =?

Answer:
 t 1 y(t)
𝑇𝑠 + 1

u(t)
y(t)
lim 𝑒 𝑡 = 𝑡 − 𝑦(𝑡) = 𝑇
"→$

𝑇
%MatLab code
%lsim(num,den,u,t)
%for any input u u(t)
num = [1];
!
den = [1 1 0 0]; %T=1 y(t) !"
y 𝑡 = 𝑡 − 𝑇 + 𝑇𝑒
t = 0:0.01:3;
u = t;
y = lsim(num,den,u,t);
plot(t,u,'-',t,y,'.')
 Poles and Zeros
Transfer Function
N (s) b0 s m bm 1s bm
G (s)
D( s) s n a1s n 1
an 1 s an
k ( s z1 ) ( s z m )
( s p1 )( s p2 ) ( s pn )
pi are poles:
D(s) = 0 à G(s) à ∞
zi are zeros:
N(s) = 0 à G(s) = 0
 Pole-Zero Plot
• System dynamics can be represented graphically
by plotting the locations of poles and zeros on
the complex s-plane
• A zero location is marked by a circle (o)
• A pole location is marked by a cross (x)

Im

Example:

Re
Example:
A linear system is described by the differential equation

y + 4 y + 8y = u + u
Plot the system transfer function poles and zeros.

Solution:
System transfer function:
Y (s) s 1 s ( 1)
U (s) s 2 4s 8 ( s ( 2 2 j ))( s ( 2 2 j ))
It has
zero: s = −1
poles:
s =−2 − 2𝑗
s = −2 + 2𝑗
 Pole-Zero Map
2

"#!
H = 1.5
" ! #$"#%
1

0.5

Imaginary Axis
zero: s = −1
poles: 0

s =−2 − 2𝑗 -0.5
s = −2 + 2𝑗
-1

-1.5
%MatLab code
H = tf([1 1],[1 4 8]); -2
-2 -1.5 -1 -0.5 0
pzmap(H);
Real Axis
Inverse Problem Pole-Zero Map
2

Example: 1.5

Given the pole-zero 1

plot of a third-order 0.5

system, with gain k

Imaginary Axis
0 (0,0)

-0.5
Find its differential
-1
equation
-1.5

Solution: First, find -2

-2 -1.5 -1 -0.5 0

Zero: −2 Real Axis

Poles: −1, −1 − 2𝑗, −1 + 2𝑗

Then, find transfer function
Zero: −2
Poles: −1, −1 − 2𝑗, −1 + 2𝑗

%MatLab code
s + 2
H = zpk(-2,[-1 -1-2j -1+2j],1); --------------------------
tf(H) s^3 + 3 s^2 + 7 s + 5

OR Y (s) k ( s 2)
à U (s) ( s 1)( s ( 1 2 j ))( s ( 1 2 j ))
k ( s 2)
s 3 3s 2 7 s 5

Finally, Inverse Laplace Transform à

d3y d2y dy du
3
3 2 7 5y k 2ku
dt dt dt dt
Recall:

Final Value Theorem

If all poles of sY(s) are on LHP, then
lim y (t ) lim sY ( s )
t s 0

Note: “on LHP” does not include the boundary

Initial Value Theorem

For any Laplace Transform pair (y, Y):
y (0 ) lim sY ( s )
s
Example:
3( s 2) 3( s 2)
Y (s) sY ( s )
s ( s 2 2 s 10) s 2 2 s 10
Poles of sY(s) are , which are on LHP, so
2 3j
6/10 = 0.6
y(∞) = lim sY (s) = 0.6
s→0

+ 3(s + 2)
y(0 ) = lim 2 =0
s→∞ (s + 2s +10)

%MatLab code % OR
num=[3 6]; num=[3 6];
den=[1 2 10 0]; den=[1 2 10];
Ys = tf(num,den); Ys = tf(num,den);
impulse(Ys)
step(Ys)
Example: s s2
Y (s) 2
sY ( s )
s 1 s2 1
Poles of sY(s) are ± j , which are not on LHP.
So, the final value theorem is not applicable.
But, there is no condition for initial value theorem.
2
s Impulse Response
y(0 + ) = lim 2 =1 1
s→∞ s +1 0.8

0.6
%MatLab code
0.4
num = [1 0];
0.2
den = [1 0 1];
Amplitude

0
Ys = tf(num,den) -0.2
impulse(Ys, 0:0.01:20); -0.4

-0.6

-0.8

-1
0 5 10 15 20
Time (sec)
Exercise: (10 minutes)

lim y (t ) lim sY ( s )
t s 0

y (0 ) lim sY ( s )
s

((*)
(1) Find: Transfer Function +(*) =?
,(*)
(2) Find: Transfer Function +(*) = ?
- -
(3) Given: 𝐺 𝑠 = *.- and 𝑅 𝑠 = *

Find: lim! 𝑒 𝑡 = ?
/→1
lim 𝑒 𝑡 = ?
/→2
Answer:

lim y (t ) lim sY ( s )
t s 0

y (0 ) lim sY ( s )
s

((*) 3(*)
(1) Find Transfer Function +(*) = -.3(*)
,(*) -
(2) Find Transfer Function +(*) = -.3(*)
- -
(3) Given 𝐺 𝑠 = and 𝑅 𝑠 =
*.- *
- -
lim 𝑒 𝑡 = lim 𝑠𝐸(𝑠) = lim 𝑠 " =1
/→1! /→2 /→2 -."!# *

- - -
lim 𝑒 𝑡 = lim 𝑠𝐸 𝑠 = lim 𝑠 " =4
/→2 /→1 /→1 -."!# *
 Complex Conjugate Poles

Closed-loop Transfer Function:

Recall:
𝜔!"
=
(𝑠 − 𝑝)(𝑠 − 𝑝∗ )

2
n, d n 1
Physical Meaning of Complex Conjugate Poles

2
p n j n 1

ζ – Damping Ratio

If ζ < 1 then wn > s (under damped)

If ζ = 1 then wn = s (un-damped) [Also called Critically damped]

If ζ > 1 then wn < s (over damped)

[see next page for details]
More precisely:

(or, critically damped)

Performances of Different Damping Ratios

ß Under
damped

ß Un-
damped

ß Over
damped
What will happen as 𝜁 → 0 ?

(0 < 𝜁 < 1)

𝑦 𝑡 =

( 𝜁 = 0)
= 𝜔8 sin(𝜔8 𝑡)
 Effects of Poles on System Response
Example: Oscillatory Time Response
Show the relation between the poles of
2s 1
H (s)
s 2 2s 5
and the Impulse Response of the system, and find the exact Impulse Response

Solution:
The poles of the transfer function: 𝑝, 𝑝∗ = −1 ± 2𝑗
Recall:

à
𝜎 = 1, 𝜔& = 2
à '
𝜁= = 0.447 < 1 (under damped) and 𝜔) = 5 = 2.24 𝑟𝑎𝑑/𝑠
(
 2s 1 s 1 1 2
H (s) 2 Input: 𝑢 𝑡 = 𝛿(𝑡)
( s 1) 2 2 2 ( s 1) 2 2 2 2 ( s 1) 2 22

Impulse Response: ℎ 𝑡 = 2𝑒 !" cos 2𝑡 − '* 𝑒 !" sin 2𝑡

Plot shows:
(1) How the sinusoidal terms behave
(2) How the envelope attenuates the sinusoidal terms
Impulse Response
2

1.5

1
Amplitude

0.5

-0.5

-1
0 1 2 3 4 5 6
Time (sec)
 Effects of Poles on System Response
Example: Oscillatory Time Response
Show the relation between the poles of
2s 1
H (s)
s 2 2s 5
and the Step Response of the system, and find the exact Step Response

Solution:
The poles of the transfer function: 𝑝, 𝑝∗ = −1 ± 2𝑗
Recall:

à
𝜎 = 1, 𝜔& = 2
à '
𝜁= = 0.447 < 1 (under damped) and 𝜔) = 5 = 2.24 𝑟𝑎𝑑/𝑠
(
Input: 𝑢 𝑡 = 1(𝑡)
Step Response: 𝑦 𝑡 =
'
𝜁= = 0.447 and 𝜔) = 5 = 2.24
(

5 !"
𝑦 𝑡 = 𝑒 sin(2𝑡)
2

Damped Sinewave
Pole Location and System Response

cos 𝜔& 𝑡

nt
e
 Effects of Zeros on System Responses
Example:
" " "
H$ 𝑠 = = − à ℎ 𝑡 = 2𝑒 )* − 2𝑒 )"*
(&'$)(&'") &'$ &'"
"(&'$.$) ,.$- $../
H" 𝑠 = $.$(&'$)(&'") = + &'" à ℎ 𝑡 = 0.18𝑒 !" + 1.64𝑒 !*"
&'$

Both have the same poles and same DC gain (Direct Current)
(DC gain := ratio of the magnitude of the steady-state step
response to the magnitude of the step input
= transfer function value at s = 0)

A zero 𝑠 = −1.1 near a pole 𝑝 = −1.0 significantly reduces

the coefficient (magnitude), from 2 to 0.18
à
A zero near a pole reduces the effect of the pole on response
Example:
A second-order system with zero s = −αζωn = −ασ
𝑠( s+/ 𝛼𝜁𝜔n8) 1
H (𝑠s ) =
𝐻
(𝑠s4/ +n 2𝜁𝜔 s /𝜔84n ) 1
) 2 28 𝑠 (+

If α > 0 à a zero on LHP à

If α ≅ 1
à value of zero is close to
real part of the poles
à substantial influence on (α > 0)

step response
If α is large
à zero is far from poles
à zero has only a little Plots of the step response with
effect on the response zero −ασ (ζ = 0.5)
If α < 0 à a zero on RHP: non-minimum-phase system

Step Response

Overshoot
is reduced

An LTI system is said to be minimum-phase

if the system and its inverse are causal and stable
 Effect of Zeros on Transient Response
Example:
A second-order system with a finite zero and a unit DC gain:
24 (s + z)
H (s) =
z (s + 4)(s + 6)
Determine the effect of the zero location (𝑠 = −𝑧) on the
unit step response when z = {1, 2, 3, 6}
Solution:
24 s 24
Y (s)
z s ( s 4)( s 6) s ( s 4)( s 6)
Step Response:
12 4t 12 6t
y (t ) 1 3 e 2 e
z z
Effect of Zero on Transient Responses

Recall: zero is given by 𝑠 = −𝑧

poles are: 𝑝' = −4, 𝑝* = −6
à Real zero on LHP moving towards
imaginary axis will increase overshoot
Effect of Complex Zeros on Transient Responses

Example:
Consider a third-order feedback system with the
transfer function having a pair of lightly-damped
poles and a pair of complex zeros:
(s )2 2
H (s)
( s 1)[( s 0.1) 2 1]
Determine the effect of the complex zero location
on the unit step response of the system, when
(α,β) = (0.5, 1) 𝛽 = 1 is fixed
(α,β) = (0.25, 1) 𝛼 = 0.5, 0.25, 0.1
(α,β) = (0.1, 1) à Complex conjugate
zeros moves to the right
Effect of Complex Zeros on Transient Responses

Reducing

Moving
 Effect of Real Poles on Transient Responses

H M 1
Example: 𝐻( s𝑠) =
[( s(NOPQR
/ ) 1][(
" s / ) 2
2 !"(]s / ) 1]
n ! )[N OTQR n ! NOR n

Real pole:
−𝛼𝜁𝜔8
Step response of
the system (ζ = 0.5)

Smaller 𝛼 > 0
(Pole still on LHP)
increases rise time
(Takes longer time
to reach value 1.0)
 Summary
Effects of Poles/Zeros on System Responses

• Locations of poles in the s-plane determine

the characteristics of the system responses
• A real zero on LHP, moving toward imaginary
axis, will increase the overshoot
• A pair of complex conjugate zeros on LHP,
moving toward imaginary axis, will decrease
the overshoot
• A real pole on LHP, moving toward imaginary
axis, will increase the rise time
END