4.

Time Response Analysis of control systems

4.1. Introduction
Total response of a control system is made up of
i. the steady- state response
ii. the transient response: is said to exist as long as any
of the variables of the system is changing with time
Transient Response
• Transient response is that part of a response
which occurs when there is a change in input
and which dies away after a short interval of
time
Chapter 4 1
• depend up on system poles only and not on the
type of input, so sufficient to analyze the
transient response to one of the standard test
signals.
• Is also called dynamic response.
Steady-State Response
• is the response that remains after all transient
responses have died down
• steady-state is the equilibrium state attained such
that there is no change with respect to time of
any of the system variables.

Chapter 4 2
4.2. Standard Test signals

• Input signals to control systems are not usually

known fully ahead of time & may be random in
nature.

• The characteristics of actual signal may be a

sudden shock, a sudden change, constant velocity
and a constant acceleration .

• The response of a control system with the

standard test signals has a good correlation with
the response to actual inputs
Chapter 4 3
a) Step signal:

• whose value changes from one level (usually

zero) to another level A in zero time.
• when input change shortly in value, example,
when a voltage is switched on in a circuit.

r(t) r (t) = A u(t) where u(t) = 1 ; t >0

1
Unit Step, A=1 u(t) = 0 ; t < 0
R (s) = A/S
0 t

Chapter 4 4
b) Ramp signal:
• Starts at a value of zero and increases linearly
with time
r(t)
• Steadily increasing input
r (t) = At ; t > 0
= 0 ; t< 0 0 t

R (s) = A/s2
C) Impulse -signal
• Pulse function area is given by h.to f(t)
h
h
(
f ( s) = 1 − e
s
− to s
) t
to
Chapter 4 5
d) Parabolic signal r(t)
• r(t) = At2/2 ; t > 0
=0; t<0
 R(s) = A/s3 0 t

e) Sinusoidal signal
r (t ) = sin t
r(t)


 R( s) = 2 0 t

s + 2

Chapter 4 6
 Laplace transform

Chapter 4 7
4.3. Time Response of 1st and 2nd order systems
First order systems
• When the maximum power of the differential is
one
Ex. 1.
dVc
V = iR + Vc ; but i = c
R

V i c vc dt
dVc
V = Rc + Vc
dt
Ex. 2.
dv
m F m + cv = F
c
v dt
Chapter 4 8
 • Generally a first order system is given by
d0
a1 + ao0 = boi
dt
Where
• a1, ao and bo are constants
• θi is input function to the system and can take
any of the test signals mentioned above
• θo is the out put
The Laplace transform of the equation is:
a1s 0 ( s) + a0 0 ( s) = b0i ( s)
0 ( s) b0 b0 / a1
 = =
i ( s) a1s + a0 s + ao / a1
Chapter 4 9
• Response to the step input
A
i ( s) =
s
 bo   A 
 0 (s) =   
 a1s + a0   s 
Using Partial Fraction Method θo(t)

b0 A  1 1  boA/ao
 0 (s) =  − 
a0  s s + a0 a1  63% of b A/a
o o

the output in time domain will be:

 o (t ) = (
b0 A
a0
1− e )
− ( ao / a1 )t
 t

• The output rises exponentially from zero value to the

final value of bo A/ao.
Chapter 4 10
• The time constant,  is indicative of how fast the
system tends to reach the final value
When t = 0; 0 = 0
b0 A
t →  ; 0 → and that is the steady-state value
a0
steady-state relation is given by:
 b0 
o =   i , where i = A
 a0 
b0
Gss = → Stead-state transfer function
a0

(
  0 (t ) = Gss A 1 − e
Chapter 4
− ( ao / a1 )t
) 11
• When t=a1/ao; θo=0.63 GssA
• This means that the out put has risen to 0.63 of
its steady-state value.
• Time needed to reach this value is called time
constant,  .
• If t = 2 (a1/a0 ) = 2 and hence θo= 0.86 Gssθi
θo/θi
Time, t θo/θi Gss
1 0.63Gss 0.86Gss

2 0.86Gss 0.63Gss

3 0.95Gss
4 0.98Gss
∞ Gss
Chapter 4
1 2 3 4 12
 (
 o (t ) = Gssi 1 − e −t /  )
and the differential equation becomes
d o
a1 + ao o = boi
dt
a1 d o bo
+  o = i
ao dt ao
d o
 +  o = Gss i
dt
from which a Transfer function will be develped as:
o Gss
= where  is the time constant.
i 1 +  s
Chapter 4 13
Steady State Error:
• difference between output and input when all
conditions are in steady state.
Assuming Gss = 1; e( t ) = i ( t ) −  o ( t ) = i e −t / 
 ess = lim e( t ) = 0
t →
Response to the Unit-Ramp input
o Gss Gss
=  o = i
i 1 +  s 1 + s
k  k  G 
For i = 2 and Gss = G  o =  2   
s  s  1 + s 

Chapter 4 14
 Using Partial Fraction Method:
B AC
o = 2 + +
s s 1 + s
where A = kG, B = − kG , C = kG 2
 G G G 2 
 o = k  2 − + 
s s 1 + s 
 G G G 
 o = k  2 − + 
s s s + 1 /  
The response in time domain will then be:

(
 0 (t ) = k Gt − G + G e −t 
)  (
=kG t −  1 − e
 )
−t 


Chapter 4 15
 For G = 1 and k = 1 (Unit Ramp input)
θo(t)
Steady-
 0 (t ) = t −  (1 − e − t / )
state error e(t ) = r (t ) −  0 (t )
− t /
=  (1 − e )
t ess = lim e(t ) = 
t →

• First Order System will track the unit-ramp input

with a steady-state error  , which is the time
constant.
• Hence, reducing the time constant not only
improves its speed of response but also reduces
its steady-state error to a ramp input
Chapter 4 16
Example:
• A thermo couple has the transfer function linking
its output in volts to its inputs in oc of:
30 x 10 −6
G( s ) =
10s + 1
• When the thermocouple is subjected to a steadily
rising temperature input of 5oc/s, what will be
the thermocouple output after 12s and how
much will it lag behind the output it would have
indicated if it had responded instantly to the
input
• What if it is subjected to a sudden input of 5oc
Chapter 4 17
Answer
• G = 30 x 10-6 ,  = 10, k=5
• 0 = 30 x 10-6 x 5 [12-10 (1-e-12/10)] = 7.5 x 10-4 v
• The lag is the difference between the value of
Gkt & Gk (t-  (1-e-t/), since
• Gkt = 30 x 10-6 x 5 x12 = 18.0 x 10-4 v
• The lag is then 10.5 x 10-4 V

Chapter 4 18
Response to Unit-Impulse Input
• Is a response that indicates when a system is
subjected to a sudden input
o ( s ) G
= G( s) = ; If i ( s ) = A
i ( s )  s +1
G G. A AG  1 
o ( s ) = i ( s ) = =
 s +1  s + 1   s + 1/  
AG − t /  θo(t)
 0 (t ) = (e )
 GA/ 

 1  −t /
= GA   e
 
0.37 GA/ 

Chapter 4  t 19
Example:
• Referring to the example above, what will be the
output of the thermocouple 5 sec after it was
subject to a temperature impulse of 100oc by
suddenly and very briefly coming in contact with
a hot object?
Answer:

 o ( t ) = GA( 1 /  )e −t / 
G = 30 x 10 −6 ,  = 10 sec, A = 100
 o ( t ) = 30 x 10 −6 x 100 x ( 1 / 10 )e −5 / 10
= 1.8 x 10 −4 v
Chapter 4 20
Second Order System
k
Ex. 1. d 2x dx
m F
m 2 + c + kx = F
dt dt
c
di dvc
Ex. 2.
R L vs = iR + L + vc , but i = c
dt dt
Vs i c vc
d 2 vc dvc
Lc + Rc + vc = vs
dt 2 dt
• a second order system has the general form
d o ' d o
2
+ ao0 = boi
' '
a2 2
+ a1
dt dt
Chapter 4 21
 d o a1 do ao
2 ' ' '
bo
 2
+ + o =  2
dt a2 dt a2 a2
d o do
2
 2
+ a1 + aoo = boi
dt dt
• And usually is written as:
d o d o
2

2
+ 2 n +  2

n o = bo n i
 2
dt dt
• The transfer function will then be:
o ( s ) bon
2
= 2
i ( s ) s + 2n s + n2
where n is the undamped natural frequency
 is the damping ratio
Chapter 4 22
 −2n  4 n − 4n
2 2 2 2
−a1  a1 − 4ao
where m1,2 = =
2 2
 m1,2 = n  −   − 1 
2
 
• Three cases could be considered here:
Case 1: If  =1 Y(t)
m1,2 = −n , double root
bo
 Y( s ) =
( s + n ) 2

−n t t
 y( t ) = bote
• And this is called a Critically Damped System.
Chapter 4 23
Case 2: If 0< <1

(
m1,2 = n −  j 1 −  2 )
bo
 Y( s ) =
( )
2
( s + n ) 2
+ n 1 −  2

bo
 y( t ) = e −n t sin( n 1 −  2 t )
n 1 −  2
y(t)

Chapter 4 24
• The above is a damped sinusoidal function with a
decreasing amplitude.
• It has got a new frequency called damped natural
frequency, and is given as:
d = n 1 −  2
Case 3: If  >1

(
m1,2 = n −   2 − 1 )
bo A B
 Y( s ) = = +
( s − m1 )( s − m2 ) ( s − m1 ) ( s − m2 )
bo
Where A = and B = − A
Chapter 4
2n  2 − 1 25
  y( t ) =
bo −n t  ( )
 2 −1 t
−e
− ( )
 2 −1 t 
e  e 
2n  2 − 1  
– an exponential function
Summary
Considering all the cases we have seen so far
1. ξ = 0  a1 = 0 → oscillatory system with ωn
2. 0 < ξ < 1  damped oscillation with ωd
3. ξ = 1  critically damped → transition zone
4. ξ > 1  over damped
Exercises

Chapter 4 26
For 0    1
 bon 2  A 
0 ( s) =  2 2  
 s + 2n s + n   s 
 n2 
= Abo  2 2 
 s ( s + 2n s + n ) 
 1 −n t

  o (t ) = Abo 1 − e sin(n 1 −  t +  ) 
2

 1−  2 

1−  2
where  = tan −1 = cos −1 

  
for  = o   0 (t ) = Abo 1 − sin (nt + )  = Abo 1 − cos nt 
 2 
Chapter 4 27
 Summary

Figure: Time functions associated with points in the S-plane

(LHP, left half-plane; RHP, right half-plane)
Chapter 4 28
4.4. Time Response Specification
• Control Systems are generally designed with
damping less than one.
• Specification for control system design often
involve certain requirements associated with the
time response of the system.
• The requirement for a step response are
expressed in terms of standard quantities
illustrated in Fig. shown in the next slide.
• The quantities express:
– how fast the system moves to follow the input,
– how oscillatory it is (indicative of damping) and
– how long does it take to practically reach the final
Chapter 4
value. 29
Chapter 4 30
 Time Response Specification of 2nd Order System
1. Rise time tr :
– Time required for the response to rise
• from 0 to 100% of the final value of under damped system
• From 10 to 90% of the final value for over damped system
– Considering system with ξ <1, for unit step input and
bo=1
– o(tr) = Abo=1
1
1− e −ntr sin n 1 −  2 tr +    1
1− 2  

 −  1 −  2 
 tr  where  = tan −1   = cos −1 
n 1 −  2   
 
Chapter 4 31
Rough approximation
– All the curves rise in roughly the same time.
– Hence, if we consider the curve for ξ = 0.5 to
be an average, the rise time from y=0.1 to 0.9 is
approximately ωntr=1.8. Thus we can say that
1.8
tr 
n
2. Peak time tp: do (t p ) = 0
dt
d o (t ) n −nt
= e sin n 1 −  2 t + cos −   −
dt 1−  2  

e −nt
n 1 −  2 cos n 1 −  2 t + cos −1   = 0
1−  2  
Chapter 4 32
  sin n 1 −  2 t p +   cos  − sin  cos[n 1 −  2 t p +  ] = 0
 
 sin n 1 −  2 t p +  −   = 0
 
n 1 −  2 .t p = n ; n = 0,1, 2...
Since peak time corresponds to the first overshoot

 tp =
n 1 −  2
Qn: When will the first undershot and the second overshot occur.

3. Peak Overshoot Mp :
M p =  o (t p ) − Abo
 1 
sin n 1 −  t p +   
−n t p
 − Abo  e 2
 1−  2  
 
Chapter 4 33
  
− n  
− 
 M p = Abo   + 
1 n 1− 2
e sin n 1 −  .
2
   
 1−
2
  1 −  2
 
 n

but sin( +  ) = − sin  = − 1 −  2


−
1− 2
 M p = Abo e

• The peak percent overshot will then be:


−
1− 2
 o (t p ) −  o () Abo e
M p (%) = 100% = ( )100%
 o ( ) Abo

−
1− 2
 M p (%) = (e )100%

Chapter 4 34
4. Settling time ts :
• This is the time required for the transient to decay
to a small value so that y(t) is almost in the steady
state.
• Various measures of smallness are possible.
• 2 and 5% are possible but here 1% of error is
taken as a reasonable measure.
• The duration of this error is essentially decided by
the transient exponential, hence we can define ts
when decaying exponential reaches 1%:
e −n ts = 0.01  nts = 4.6
4.6
 ts =
Chapter 4 n 35
• Similarly, for 2% and 5 % errors, the settling
time will be given consecutively as:
4 3
ts  and ts 
n n

Examples

Chapter 4 36
 End of Chapter Four

Chapter 4 37