Automatic Control Systems

-Lecture Note 7-
Time Domain Analysis of
Control Systems

-1-
Automatic Control Systems
 Introduction

Time domain analysis of control system for

1st order system
2nd order system

Reference inputs used

Unit-step input is mainly used
Impulse, ramp input

-2-
Automatic Control Systems
 Introduction

Time response
Transient response : temporary time response before entering
the steady state
Steady state response : time response following after the
transient response

-3-
Automatic Control Systems
 Typical Test Signals

□ Typical Test Signals for Time Response

Step-Input
ìR t ³ 0
r (t ) = í or r (t ) = Rus (t )
î0 t < 0

Ramp-Input
r (t ) = Rtus (t )

Parabolic-Input
Rt 2
r (t ) = us (t )
2
-4-
Automatic Control Systems
 Typical Test Signals

-5-
Automatic Control Systems
 Unit Step Response
Transfer function : G(s)
Y ( s) = G ( s) R( s)
y (t ) = L-1{Y ( s )}= L-1{G ( s ) R ( s )}

G (s) ü
Unit-step response : ys (t ) = L-1 ìí ý
î s þ

Impulse and ramp input responses from unit-step response

ì G (s) ü d
yimpulse (t ) = L-1 {G ( s )} = L-1 í s ý = ys (t )
î s þ dt
-1 ì G ( s ) ü -1 ì 1 G ( s ) ü
t
yramp (t ) = L í 2 ý = L í ý = ò0 ys (t )dt
î s þ îs s þ

-6-
Automatic Control Systems
 Unit Step Response

< Note >

① 3rd and higher order systems are not easy to analyze manually

② Simulation tools (e.g., Matlab or CEMTool) can be used

Ex.) >>step(num,den)
>>impulse(num,den)

-7-
Automatic Control Systems
 Unit Step Response

【Example】 Usage of Matlab or CEMTool

Sol) 4
G (s) = 3
s + 2 s 2 + 3s + 4
1) Impulse response
>> num=4;
>> den=[1 2 3 4];
>> impulse(num, den);

2) Unit step response

>> step(num, den);

-8-
Automatic Control Systems
 Unit Step Response
3) Ramp response
>> den=[1 2 3 4 0]; // G(s) : 4th order system
>> step(num, den);

-9-
<Fig> impulse response
Automatic Control Systems
 Unit Step Response

<Fig> unit step response

- 10 -
Automatic Control Systems
 Unit Step Response

<Fig> ramp input response

- 11 -
Automatic Control Systems
 Unit Step Response
Typical form of the Unit step Response

<Fig> time response of stable system to unit step input

- 12 -
Automatic Control Systems
 Unit Step Response

- 13 -
Automatic Control Systems
 Unit Step Response

Transient performance criteria

Rise time ( tr) : time required for the output to rise from
10% to 90% of steady state value

Peak time ( tp ) : time when the output arrives at its

maximum

% Maximum Overshoot (Mp )

y (t p ) - R
Mp= ´100 [%]
R

- 14 -
Automatic Control Systems
 Unit Step Response

Steady state performance criteria

Settling time (ts ) : the first time when the output variation
gets in and stay there within ±2(or ±5)% of reference
value. In general, it equals 4(or 3) times of rise time.

Steady state output (Ys ) : the average output value after

the settling time

Steady-State Error( Es or ess) :

R - Ys
Es = ´100 [%]
R
- 15 -
Automatic Control Systems
 Unit Step Response

【Example】 Find the rise time, peak time,

maximum overshoot, steady state output, and
steady state error.
4
G (s) = 3
s + 2 s 2 + 3s + 4

Sol.) rise time : 0.96s, peak time : 3.56s,

maximum overshoot : 48.6%,
steady state output : 1, steady state error : 0%
- 16 -
Automatic Control Systems
 Steady-State Errors

Definition of the Steady-State Error

r + u y
G(s)
-

H(s)

e(t) = reference signal – y(t)

- 17 -
Automatic Control Systems
 Steady-State Errors
The steady-state error is defined as

ess = lim e(t ) = lim sE ( s )

t ®¥ s ®0

When the system has unity feedback, H(s) = 1

r + e y
G(s)
-

- 18 -
Automatic Control Systems
 Steady-State Errors

In this case,
e(t ) = r (t ) - y (t )

E ( s) = R( s) - Y ( s)
G (s)
= R( s) - R( s)
1 + G (s)

=
[1 + G ( s ) ]- G ( s )
R( s) =
1
R( s)
1+ G ( s) 1+ G ( s)

sR ( s )
lim e(t ) = ess = lim sE ( s ) = lim
∴ t ®¥ s ®0 s ®0 1 + G ( s )

- 19 -
Automatic Control Systems
 Steady-State Errors

Consider the unity feedback system

r + e y
G(s)
-

Type of Control Systems

lim e(t ) = ess = lim sE ( s )
t ®¥ s ®0

sR ( s )
= lim
s ®0 1 + G ( s )

⇒ ess depends on the characteristic of G(s)

- 20 -
Automatic Control Systems
 Steady-State Errors

K (1 + T1s ) (1 + T2 s )L (1 + Tm1s + Tm 2 s 2 ) -Td s

G (s) = N e
s (1 + Ta s ) (1 + Tb s )L (1 + Tn1s + Tn 2 s )
2

⇒ N : Type Number (= the order of pole of G(s) at s=0)

【Example】

K (1 + 0.5s )
G (s) = 2
: Type 1
s (1 + s )(1 + 2 s )(1 + s + s )
K (1 + 2 s )
G (s) = 3
: Type 3
s
- 21 -
Automatic Control Systems
 Steady-State Errors
- ess of Systems with a Step Input
□ ess of Systems with a Step Input
With r(t) = Rus(t) → R(s) = R/s
R
s( )
sR ( s ) s = lim R R
ess = lim = lim =
s ®0 1 + G ( s ) s ®0 1 + G ( s ) s ®0 1 + G ( s ) 1 + lim G ( s )
s ®0

Define K p º lim G ( s ) = G (0) : Step Error Constant

s ®0

R
If N = 0 (type 0 system), ess = ← Constant
1+ K p
If N ≥ 1, since Kp = ∞, ess = 0

- 22 -
Automatic Control Systems
 Steady-State Errors
- ess of Systems with a Step Input

- 23 -
Automatic Control Systems
 Steady-State Errors
- ess of Systems with a Ramp Input
□ ess of Systems with a Ramp Input

With r(t) = Rtus(t) → R(s) = R/s2

R
s( )
sR ( s ) s 2
R R
ess = lim = lim = lim =
s ®0 1 + G ( s ) s ®0 1 + G ( s ) s ®0 s + sG ( s ) lim sG ( s )
s ®0

Define K v º lim sG ( s ) : Ramp Error Constant

s ®0

Type 0 (N=0) system, ess = ∞

R
Type 1 (N=1) system, ess = ← Constant
Kv

Type 2 or higher (N≥2) system, since Kv = ∞, ess = 0

- 24 -
Automatic Control Systems
 Steady-State Errors
- ess of Systems with a Ramp Input

- 25 -
Automatic Control Systems
 Steady-State Errors
- ess of Systems with a Parabolic Input
□ ess of Systems with a Parabolic Input
Rt 2
With r (t ) = us (t ) ® R ( s ) = R / s 3
2
R
s( 3 )
sR ( s ) s R
ess = lim = lim =
s ®0 1 + G ( s ) s ®0 1 + G ( s ) lim s 2G ( s )
s ®0

Define K a º lim s 2G ( s ) : Parabolic Error Constant

s ®0

Type 0, 1 (N=0, 1) system, ess = ∞

R
Type 2 (N=2) system, ess = ← Constant
Ka
Type 3 or higher (N≥3) system, since Ka = ∞, ess = 0

- 26 -
Automatic Control Systems
 Steady-State Errors
- ess of Systems with a Parabolic Input

- 27 -
Automatic Control Systems
 Steady-State Errors
Summary of the Steady-State Errors Due to Step, Ramp and
Parabolic Function Inputs for Unity Feedback Systems

⇒ To be valid, the closed-loop system must be stable

- 28 -
Automatic Control Systems
 Steady-State Errors

【Example】
K ( s + 3.15)
G (s) = , H ( s ) = 1 : Type 1 system
s ( s + 1.5)( s + 0.5)

Step input : R
K p = ¥, ess = =0
1+ K p
R R
Ramp input : K v = 4.2 K , ess = =
K v 4.2 K
R
Parabolic input : K a = 0, ess = =¥
Ka

⇒ All of these results are valid only if K stays within stable range
⇒ 0 < K < 1.3043 (by Routh-Hurwitz Condition)
- 29 -
Automatic Control Systems
 Steady-State Errors

【Example】

K
G (s) = 2
, H ( s ) = 1 : Type 2 system
s ( s + 12)

⇒ The closed-loop system is unstable for all values of K

⇒ Error analysis is meaningless

- 30 -
Automatic Control Systems
 Steady-State Errors

【Example】

5( s + 1)
G (s) = 2 , H ( s ) = 1 : Type 2 system
s ( s + 12)( s + 5)

⇒ The closed-loop system is stable

Step input : R
K p = ¥, ess = =0
1+ K p
Ramp input : R
K v = ¥, ess = =0
Kv
Parabolic input : K a = 1 , ess = R = 12 R
12 Ka
- 31 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
□ Relationship between ess and the Closed-Loop TF

Let us assume the following condition :

lim H ( s ) = H (0) = K H = constant

s ®0

⇒ H(s) cannot have poles at s = 0

⇒ bss = KH × yss
- 32 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
⇒ steady-state error would be zero
if rss – bss = rss – KH × yss = 0

⇒ Define the error signal e(t) as (in order to focus on yss)

1
e(t ) = r (t ) - y (t )
KH
or in the transfer domain,

1 1 Y (s) 1
E (s) = R( s) - Y ( s) = [1 - K H ] R( s) = [1 - K H M ( s )] R ( s )
KH KH R( s) KH

Where M ( s) (= Y ( s) ) is the closed-loop transfer function.

R( s)
- 33 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
Let us assume that M(s) does not have any poles at s = 0,
and is of the form:
Y ( s ) bm s m + bm -1s m -1 + L + b1s + b0
M (s) = = n , n>m
R( s) s + an -1s + L + a1s + a0
n -1

⇒ Assume that all the poles of M(s) are in the left-half s-plane
(that is, stable)
⇒ The steady-state error is
1
ess = lim sE ( s ) = lim [1 - K H M ( s )]sR ( s )
s ®0 s ®0 KH
1 s n + L + (a1 - b1 K H ) s + (a0 - b0 K H )
= lim sR ( s )
K H s ®0 s + an -1s + L + a1s + a0
n n -1

- 34 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
We consider the three basic types of inputs for r(t)

1. Step Input. R(s) = R/s

1 æ a0 - b0 K H ö
ess = ç ÷R
KH è a0 ø
b0 1
⇒ ess can be zero only if a0 - b0 K H = 0 or M (0) = =
a0 K H
⇒ This means that for a unity-feedback system (KH=1), the
constant terms of the numerator and the denominator of M(s)
must be equal, that is, b0=a0 , in order for the steady state
error to be zero.
- 35 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
2. Ramp Input. R(s) = R/s2

1 s n
+ L + (a1 - b1 K H ) s + (a0 - b0 K H )
ess = lim R
K H s ®0 s ( s + an -1s + L + a1s + a0 )
n n -1

The following values of ess are possible :

ess = 0 if a0 - b0 K H = 0 and a1 - b1 K H = 0
a1 - b1 K H
ess = R = constant if a0 - b0 K H = 0 and a1 - b1 K H ¹ 0
a0 K H
ess = ¥ if a0 - b0 K H ¹ 0

- 36 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
3. Parabolic Input. R(s) = R/s3

1 s n + L + (a2 - b2 K H ) s 2 + (a1 - b1 K H ) s + (a0 - b0 K H )

ess = lim R
KH s ® 0 s ( s + an -1s + L + a1s + a0 )
2 n n -1

The following values of ess are possible :

ess = 0 if ai - bi K H = 0 for i = 0, 1 and 2

a2 - b2 K H
ess = R = constant if ai - bi K H = 0 for i = 0 and 1
a0 K H
ess = ¥ if ai - bi K H ¹ 0 for i = 0 or / and 1 or both

- 37 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
【Example】
5( s + 1)
Assume G (s) = 2
and H ( s ) = 1, then K H = H (0) = 1
s ( s + 12)( s + 5)
G (s) 5( s + 1)
Þ M (s) = = 4
1 + G ( s ) s + 17 s 3 + 60 s 2 + 5s + 5
Þ M ( s ) is stable

Step input: ess = 0 since a0 = b0 (= 5)

Ramp input: ess = 0 since a0 = b0 (= 5) and a1 = b1 (= 5)

a2 - b2 K H 60
Parabolic input: ess = R= R = 12 R
a0 K H 5
⇒ Since this is a Type 2 system with unity feedback, the same
results are obtained with the error constant method
- 38 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
【Example】
1 5( s + 1)
G (s) = 2
and H ( s ) = , then K H = H (0) = 1
s ( s + 12) s+5
Y (s) G (s) s+5
Þ M (s) = = = 4
R ( s ) 1 + G ( s ) H ( s ) s + 17 s 3 + 60 s 2 + 5s + 5
Þ M ( s ) is stable
Þ a0 = 5, a1 = 5, a2 = 60, b0 = 5, b1 = 1, b2 = 0

a0 - b0 K H
Step input: ess = =0
a0 K H
a1 - b1 K H 5 - 1
Ramp input: ess = = = 0.8
a0 K H 5

Parabolic input: ess = ¥ since a1 - b1 K H ¹ 0

- 39 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
(Remark)
For a unit-step input (r(t) = us(t)) :

y (t ) = 1 - 0.00056e -12.05t - 0.0001381e -4.886t

- 0.9993e -0.0302t cos 0.2898t - 0.1301e -0.0302t sin 0.2898t , t ³ 0
Þ y ss = 1 Þ ess = 0

For a unit-ramp input (r(t) = tus(t)) :

y (t ) = t - 0.8 + 4.682 ´ 10 -5 e -12.05t + 2.826 ´ 10 -5 e -4.886t

+ 0.8e -0.0302t cos 0.2898t - 3.365e -0.0302t sin 0.2898t , t ³ 0
⇒ steady-state portion of y(t) is t - 0.8 (yss = t - 0.8)
⇒ ess = 0.8
- 40 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
For a unit-parabolic input (r(t) = t2us(t)/2):

y (t ) = 0.5t 2 - 0.8t - 11.2 - 3.8842 ´10 -6 e -12.05t - 5.784 ´10 -6 e -4.886t

+ 11.2e -0.0302t cos 0.2898t + 3.9289e -0.0302t sin 0.2898t , t ³ 0

⇒ steady-state portion of y(t) is

0.5t2 – 0.8t – 11.2 (yss = 0.5t2 – 0.8t – 11.2)

⇒ ess = 0.8t + 11.2⇒ ess = ∞ as t → ∞

- 41 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
【Example】
1 10( s + 1)
G (s) = 2 and H ( s ) = , then K H = H (0) = 2
s ( s + 12) s+5
Y (s) G (s) s+5
Þ M (s) = = = 4
R ( s ) 1 + G ( s ) H ( s ) s + 17 s 3 + 60 s 2 + 10 s + 10
Þ M ( s ) is stable
Þ a0 = 10, a1 = 10, a2 = 60, b0 = 5, b1 = 1, b2 = 0

a0 - b0 K H 10 - 5 ´ 2
Step input : ess = = =0
a0 K H 10 ´ 2
a1 - b1 K H 10 - 1´ 2
Ramp input : ess = = = 0.4
a0 K H 10 ´ 2
Parabolic input : ess = ∞ since a1 – b1KH ≠ 0
- 42 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
(Remark)
For a unit-step input (r(t) = us(t)) :
y (t ) = 0.5u s (t ) + transient terms
1
Þ e(t ) = r (t ) - y (t ) = - transient terms
KH
Þ ess = 0

For a unit-ramp input (r(t) = tus(t)) :

y (t ) = [0.5t - 0.4]u s (t ) + transient terms
1
Þ e(t ) = r (t ) - y (t ) = 0.4u s (t ) - transient terms
KH
Þ ess = 0.4
- 43 -
Automatic Control Systems
 Steady-State Errors
- Relationship between ess and the Closed-Loop TF
For a unit-parabolic input (r(t) = t2us(t)/2) :

y (t ) = [0.25t 2 - 0.4t - 2.6]u s (t ) + transient terms

1
Þ e(t ) = r (t ) - y (t ) = (0.4t + 2.6)u s (t ) - transient terms
KH
Þ ess = ¥ as t ® ¥

- 44 -
Automatic Control Systems
 Steady-State Errors
- ess of Nonunity Feedback Systems: H(s) has Nth-Order Zero at s=0

□ ess of Nonunity Feedback Systems:

H(s) has Nth-Order Zero at s=0
This corresponds to desired output being proportional to
the Nth-order integration of the input in the steady-state
⇒ Reference signal can be defined as R(s) / KHsN

⇒ Error signal may be defined as

1
E (s) = N
R( s) - Y ( s)
KH s

where H (s)
K H = lim
s ®0 sN
- 45 -
Automatic Control Systems
 Steady-State Errors
- ess of Nonunity Feedback Systems: H(s) has Nth-Order Zero at s=0
Consider only the results for N = 1
⇒ In this case, the transfer function of M(s) will have a pole
at s = 0, or a0 = 0
1 é s n -1 + L + (a2 - b1 K H ) s + (a1 - b0 K H ) ù
ess = lim ê ú sR ( s )
KH s ® 0
ë s + an -1s + L + a1s
n n -1
û
⇒ For a step input of magnitude R
1 é s n -1 + L + (a2 - b1 K H ) s + (a1 - b0 K H ) ù
ess = lim ê ú R
KH s ® 0
ë s + an -1s + L + a1s
n n -1
û
⇒ The steady-state error is
ess = 0 if a2 - b1 K H = 0 and a1 - b0 K H = 0
a2 - b1 K H
ess = R = constant if a1 - b0 K H = 0 and a2 - b1 K H ¹ 0
a1 K H
ess = ¥ if a1 - b0 K H ¹ 0
- 46 -
Automatic Control Systems
 Steady-State Errors
- ess of Nonunity Feedback Systems: H(s) has Nth-Order Zero at s=0
Remark 1)
Y ( s ) bm s m + bm -1s m -1 + L + b1s + b0
M (s) = = n
R( s) s + an -1s n -1 + L + a1s + a0
If a0 ¹ 0, the problem does not make sense.
(That is, the output shows the usual step response.)
Y ( s ) bm s m + bm -1s m -1 + L + b1s + b0
Remark 2) For M ( s ) = =
R( s) s n + an -1s n -1 + L + a1s
æ 1 ö
ess = lim sE ( s ) = lim s ç R( s) - Y ( s) ÷
s ®0 s ®0 K
è H s ø
1 1 æ1 ö
= lim ( R ( s ) - sK H Y ( s ) ) = lim ç - K H M ( s ) ÷ sR ( s )
K H s ®0 K H s ®0 è s ø
1 é s n -1 + L + (a2 - b1 K H ) s + (a1 - b0 K H ) ù
= lim ê ú sR ( s )
KH s ® 0
ë s + an -1s + L + a1s
n n -1
û
- 47 -
Automatic Control Systems
 Steady-State Errors
- ess of Nonunity Feedback Systems: H(s) has Nth-Order Zero at s=0

【Example】
1 10 s H (s)
G (s) =2
and H ( s ) = , then K H = lim =2
s ( s + 12) s+5 s ® 0 s
Y (s) G (s) s+5
Þ M (s) = = = 4
R ( s ) 1 + G ( s ) H ( s ) s + 17 s 3 + 60 s 2 + 10 s

⇒ The velocity control system is considered to be stable,

although M(s) has a pole at s = 0, since the objective is to
control velocity with a step input
⇒ a0 = 0, a1 = 10, a2 = 60, b0 = 5, b1 = 1

For a unit-step input,

a2 - b1 K H 60 - 1´ 2
ess = = = 2.9
a1 K H 10 ´ 2
- 48 -
Automatic Control Systems
 Steady-State Errors
- ess of Nonunity Feedback Systems: H(s) has Nth-Order Zero at s=0

(Remark)
For a unit-step input (r(t) = us(t)) :

y (t ) = (0.5t - 2.9)us (t ) + transient terms

1
Þ E (s) = R( s) - Y ( s)
KH s
1 t 1
Þ e(t ) =
KH ò0 r (t )dt - y(t ) = K H tus (t ) - y(t )
= 2.9us (t ) - transient terms
Þ ess = 2.9

- 49 -
Automatic Control Systems
 Simple Definition of ess
□ Simple Definition of ess for Nonunity Feedback
Systems

E ( s) = R( s) - Y ( s)
G (s)
= R( s) - R( s)
1 + G (s) H (s)
G (s)
Let M (s) =
1 + G (s) H (s)
then E ( s ) = R ( s ) - M ( s ) R ( s ) = [1 - M ( s )]R ( s )

\ lim e(t ) = ess = lim sE ( s ) = lim s[1 - M ( s )]R ( s )

t ®¥ s ®0 s ®0

- 50 -
Automatic Control Systems
 Simple Definition of ess
【Example】 1 2.1
G (s) = , H (s) =
s ( s + 1) s+2
r + u y
G(s)
-

H(s)

Sol. : é G (s) ù s 3 + 3s 2 + s + 0.1

E ( s ) = ê1 - ú R( s) = 3 2
R( s)
ë 1 + G (s) H (s) û s + 3s + 2 s + 2.1
1
Steady state error for the unit step input R( s) =
:
s
s 3 + 3s 2 + s + 0.1
ess = lim 3 2
= 0.0476
s ®0 s + 3s + 2 s + 2.1
- 51 -
Automatic Control Systems
 Simple Definition of ess
【Example】
K 2
G (s) = , H (s) =
s+2 s+4
1
unit step input : R ( s ) =
s
K
s + 2 K ( s + 4)
Þ M (s) = =
K 2 ( s + 2)( s + 4) + 2 K
1+
s+2 s+4
4K 8 - 2K
\ ess = 1 - M (0) = 1 - = (1)
8 + 2K 8 + 2K
To achieve zero steady-state error, K = 4
(1) is valid only if “K > -4” (by R-H Criterion)
- 52 -
Automatic Control Systems
 Poles and Transient Response

RHP poles are unstable and time response diverges

LHP poles are stable and time response of system

is bounded and converges

The larger the magnitude of real part of LHP pole

is, the faster the time response reaches the steady
state

- 53 -
Automatic Control Systems
 Time Response of 1st Order System
Standard 1st order system
Gp A
G (s) = or G ( s ) =
Tc s + 1 s + 1/ t
Gp : DC gain (= At )

Tc or τ : Time Constant ; for stable system

<Note> Time Constant : the time when the output reaches 63.2% of
the steady-state value
b
i ) x& (t ) = - ax(t ) + u (t ) ii ) u ® ® y
s+a
y (t ) = bx(t )

- 54 -
Automatic Control Systems
 Time Response of 1st Order System
if u (t ) = us (t ) = 1
b æ b ö
Þ y (t ) = (1 - e - at ), t ³ 0 ç Þ y ss = ÷
a è a ø
1
Þ 0.632 = 1 - e - at Þ t =
a
Unit step response
-1
y (t ) = L {Y ( s )} = G p 1 - e ( - t / Tc
), t ³ 0
Time domain performance criteria
- rising time : t r » 2.2Tc - delay time : td » Tc
- overshoot : M p = 0 - settling time : t s » 4Tc

- 55 -
- steady state error : ess (= Es ) = 1 - G p
Automatic Control Systems
 Time Response of 1st Order System

<Note> ① time constant determines transient response

② DC gain determines steady state response

- 56 -
Automatic Control Systems
 Time Response of 1st Order System

【Example】1st order system

1
G (s) =
s +1

Sol.) Gp=1 and Tc=1[sec],

t r » 2.2[sec]
td » 1[sec]
Mp = 0
t s » 4[sec]
Es = 0
- 57 -
Automatic Control Systems
 Time Response of 1st Order System

- Unit step response for 0~10[sec] using Matlab or

CEMtool

>>num = [1];
>>den = [1 1];
>>t=[0:0.1:5];
>>step(num,den,t);

- 58 -
Automatic Control Systems
 Transient Response of 2nd Order System
□ Transient Response of Standard 2nd Order System

r + e y
G(s)
-

2
Y (s) wn
G (s) = =
E ( s ) s ( s + 2zw n )
2
Y (s) wn
Þ = 2
R ( s ) s + 2zw n s + wn 2

- 59 -
Automatic Control Systems
 Transient Response of 2nd Order System

- wn : undamped natural frequency

- z : damping ratio
z ³ 1 : over damped
0 < z < 1 : under damped
z = 1 : critically damped

- 60 -
Automatic Control Systems
 Transient Response of 2nd Order System

unit step response, R(s) = 1/s

wn 2 1 s + 2zwn
Y (s) = 2 2
= - 2 2
s ( s + 2zwn s + wn ) s s + 2zwn s + wn
1 s + 2zwn 1 ( s + zwn ) + zwn
= - = -
s ( s + zwn ) 2 + wn 2 - z 2wn 2 s ( s + zwn ) 2 + (wn 1 - z 2 ) 2

1 ( s + zwn ) zwn (wn 1 - z 2 )

= - -
s ( s + zwn ) + (wn 1 - z ) wn 1 - z 2 ( s + zwn ) 2 + (wn 1 - z 2 ) 2
2 2 2

- 61 -
Automatic Control Systems
 Transient Response of 2nd Order System

-zwn t z
⇒ y (t ) = 1 - e 2
cos(wn 1 - z t ) - e -zwnt sin(wn 1 - z 2 t )
1- z 2
æ
-zwn t 2 z 2
ö
= 1- e ç cos(wn 1 - z t ) + sin(wn 1 - z t ) ÷
ç 1- z 2 ÷
è ø
e -zwnt æ 1 - z 2 ö
= 1- sin ç wn 1 - z 2 t + tan -1 ÷
1- z 2 ç z ÷
è ø
e -zwnt
= 1- sin(wn 1 - z 2 t + cos -1 z ), t ³ 0
1- z 2
æ 1 ö
Þ ç Time constant : t = ÷
è Vwn ø

- 62 -
Automatic Control Systems
 Transient Response of 2nd Order System
a
;use the summing rule a cos w t + b sin w t = a 2 + b 2 sin æç w t + tan -1 ö÷
è bø
b
= a 2 + b 2 cosæç w t - tan -1 ö÷
è aø

- 63 -
Automatic Control Systems
 Response of 2nd Order System

□ Damping Ratio and Damping Factor

Characteristic Equation
2
Y (s) wn s1 , s2 = - zw n ± jwn 1 - z 2
= 2
R ( s ) s + 2zw n s + wn 2
= -a ± jw
where a = zw n : damping factor
w = wn 1 - z 2 : damped frequency
wn : natural undamped frequency
- z : damping ratio

a actual damping factor

z = =
wn damping factor at critical damping
- 68 -
Automatic Control Systems
 Response of 2nd Order System

□ Natural Undamped Frequency

z = cos q

- 69 -
Automatic Control Systems
 Response of 2nd Order System

- 70 -
Automatic Control Systems
 Response of 2nd Order System

Classification of system dynamics w.r.t. the value of z

0 < z <1 : s1 , s2 = -zwn ± jwn 1 - z 2 , (-zwn < 0) underdamped

z =1 : s1 , s2 = -wn critically damped
z > 1 : s1 , s2 = -zwn ± wn z 2 - 1 overdamped
z = 0 : s1 , s2 = ± jwn undamped
z <0 : s1 , s2 = -zwn ± jwn 1 - z 2 , (-zwn > 0) negatively damped

- 71 -
Automatic Control Systems
 Response of 2nd Order System

- 72 -
Automatic Control Systems
 Response of 2nd Order System

- 73 -
Automatic Control Systems
 Response of 2nd Order System

- 74 -
Automatic Control Systems
 Response of 2nd Order System
Poles and impulse response (without finite zero)
imaginary

real

<Fig> poles and impulse response

<Note> with zero, the response differs from them

- 75 -
Automatic Control Systems
 Response of 2nd Order System

□ Maximum Overshoot

- 76 -
Automatic Control Systems
 Response of 2nd Order System
e -zwnt
y (t ) = 1 - sin(wn 1 - z 2 t + q ), t ³ 0 (1)
1- z 2
-zw t
dy (t ) wn e n
Þ = [z sin(w t + q ) - 1 - z 2 cos(w t + q )], t ³ 0 (2)
dt 1- z 2

Where w = wn 1 - z 2 , q = cos -1 z

Inside the bracket

z sin(w t + q ) - 1 - z 2 cos(w t + q )
2
2
= z + - 1- z( 2
) sin[(w t + q ) - cos -1 z ] (3)

= sin w t
- 77 -
Automatic Control Systems
 Response of 2nd Order System
(2) → (1) ⇒ dy (t ) = wn e -zw nt sin wn 1 - z 2 t , t ³ 0
dt 1- z 2
dy (t )
Setting =0
dt
Þ we have the solutions : t = ¥, and wn 1 - z 2 t = np , n = 0, 1, 2, L
np
Þt = , n = 0, 1, 2, L (4)
2
wn 1 - z
p
Þ tmax (or t p ) = (5)
2
wn 1 - z

Note that although the unit-step response for z ¹ 0 is not periodic,

the overshoots and the undershoots of the response do occur at
- 78 -
periodic intervals Automatic Control Systems
 Response of 2nd Order System
The magnitudes of the overshoots and the undershoots : (4)→(1)
- npz / 1-z 2
e
y (t ) max or min = 1 - sin( np + q ), n = 1, 2, K
1- z 2
or
n -1 - npz / 1-z 2
y (t ) max or min = 1 + (-1) e , n = 1, 2, K

The maximum overshoot (Mp) is

pz
-
1-z 2
M p = y (t p ) - 1 = ymax - 1 = e

pz
-
1-z 2
% M p = 100e
⇒ Mp is function of the damping ratio z only
- 79 -
Automatic Control Systems
 Response of 2nd Order System

- 80 -
Automatic Control Systems
 Response of 2nd Order System

□ Delay Time td and Rise Time tr

Normalized delay time versus damping ratio

- 81 -
Automatic Control Systems
 Response of 2nd Order System

1 + 0.7z
td @ , 0 < z <1
wn
or a better approximation is
1.1 + 0.125z + 0.469z 2
td @ , 0 < z <1
wn

- 82 -
Automatic Control Systems
 Response of 2nd Order System

Normalized rise time versus damping ratio

- 83 -
Automatic Control Systems
 Response of 2nd Order System

0.8 + 2.5z
tr @ , 0 < z <1
wn
or a better approximation is
1 - 0.4167z + 2.917z 2
tr @ , 0 < z <1
wn

- 84 -
Automatic Control Systems
 Response of 2nd Order System

Settling Time ts
approximate settling time is 3~4 times of time constant
4
ts @
zw n

When 0 < z < 0.69 , the unit-step response has a

maximum overshoot greater than 5%
The larger z , the smaller the overshoot
The larger z , the larger the rise time and slower
response
- 85 -
Automatic Control Systems
 Response of 2nd Order System

- 86 -
Automatic Control Systems
 Transient Response of 2nd Order System

【Example】Overdamped 2nd-Order Systems

G (s) 9 Ü ( V = 1.5, wn = 3)
M (s) = = 2
1 + G ( s) s + 9s + 9

Sol.) unit step response using Matlab or CEMtool

>>num = 9;
>>den = [1 9 9];
>>step(num,den);

- %Mp : 0%, steady state error : 0, tr = 1.9 [sec]

- ts = 3.5 [sec], td = 0.8 [sec]
- 87 -
Automatic Control Systems
 Transient Response of 2nd Order System

- 88 -
Automatic Control Systems
 Transient Response of 2nd Order System
【Example】Underdamped 2nd-Order System :
G (s) 9 Ü ( V = 0.5, wn = 3)
M (s) = = 2
Sol.) 1 + G ( s ) s + 3s + 9
1.8
tr » = 0.6[sec]
3
- ( 0.5p / 1- 0.5 2 )
Mp =e ´ 100 @ 16.3033[%]

p
tp = = 1.2092 [sec]
2
3 1 - 0.5
4
ts » = 2.6667 [sec]
1.5

ess = 1 - M (0) = 0
- 89 -
Automatic Control Systems
 Transient Response of 2nd Order System
- unit step response using Matlab or CEMtool
>>num = 9;
>>den = [1 3 9];
>>step(num,den);

- 90 -
Automatic Control Systems
 Transient Response and Zeros

Standard 2nd order system with a simple zero

Y (s) wn2 (- s / a + 1)
M (s) = = 2 2
, z >0
R ( s ) s + 2zwn s + wn

Unit step response

é w ù
y (t ) = L-1{Y ( s )} = 1 - e -zw nt êcos wd t + n (z + 1 / a ) sin wd t ú
ë wd û
æ
-zw n t ç z 1/ a ö
= 1- e cos wd t + sin wd t + sin wd t ÷
ç 1 - z 2
1 - z 2 ÷
è ø
e -zw nt é 1 ù, t ³ 0
= 1- sin( w t + q ) + sin w t
1 - z 2 êë
d d ú
a û

- 91 -
Automatic Control Systems
 Transient Response and Zero

- 92 -
<Fig> LHP zero and overshoot
Automatic Control Systems
 Transient Response and Zero

- 93 -
<Fig> RHP zero and undershoot
Automatic Control Systems
 Transient Response and Zero
imaginary

overshoot undershoot
no effect increases increases no effect

0 real

<Fig> zero and transient response

<Note> RHP zero and undershoot à Intractable problem

- 94 -
Automatic Control Systems
 Dominant Poles of Transfer Function
Dominant Pole
poles at near the imaginary axis
Poles more than 4 times farther from the imaginary axis than
dominant poles are negligible

- 95 -
Automatic Control Systems
 Dominant Poles of Transfer Function

Regions of dominant and insignificant poles for design

purposes

- 96 -
Automatic Control Systems
 Dominant Poles of Transfer Function

【Example】
Dominant Pole Approximation
s+5
G (s) =
( s + 20)( s 2 + 2 s + 2)

Sol)
- 0.04 0.04 s + 0.254
G (s) = + 2
s + 20 s + 2s + 2
- 0.04 0.02 + j 0.107 0.02 - j 0.107
= + +
s + 20 s +1+ j s +1- j
0.04 s + 0.254
» 2
s + 2s + 2
- 97 -
Automatic Control Systems
 Dominant Poles of Transfer Function

num3 = [1 5];
den3 = conv([1 20], [1 2 2]); //3rd order system
num2 = [0.04 0.254];
den2 = [1 2 2]; //2nd order approximation
t = [0:0.1:20];
y3 = step(num3, den3, t);
y2 = step(num2, den2, t);
plot(t, y3, t, y2);
- 98 -
Automatic Control Systems
 Dominant Poles of Transfer Function

- 99 -
Automatic Control Systems
 Dominant Poles of Transfer Function

□ Relative Damping Ratio

Damping ratio of the dominant second-order system

【Example】
Y (s) 20
M (s) = =
R ( s ) ( s + 10)( s 2 + 2 s + 2)

⇒ s2 + 2s + 2 is dominant part of the system

⇒ Relative damping ratio : z = 0.707

- 100 -
Automatic Control Systems
 Dominant Poles of Transfer Function

□ Proper Way of Neglecting the Insignificant Poles

Neglecting the insignificant poles with consideration of
the steady-state response
【Example】 M (s) = 20
( s + 10)( s 2 + 2 s + 2)
= 20 @ 20
2
s 2
10( + 1)( s + 2 s + 2) 10 ( s + 2 s + 2)
10

<Note> In steady state, sà0, two DC gains are equal to M(0)=1.

- 101 -
Automatic Control Systems