Module 3

Module -3
Time Response of Feedback Control

Systems
1 Dept.of ECE,MITE,Moodabidri
Contents
 Time Response of feedback control systems
 Standard test signals
 Unit step response of First and Second order Systems.
 Time response specifications
 Time response specifications of second order systems
 steady state errors and error constants.
 Introduction to PI, PD and PID Controllers (excluding
design).
Lesson Plan
S.N DATE TOPIC DATE REMARKS
PLANNED ENGAGED
1.1 Time Response of feedback control systems
1.2 Standard test signals
1.3 Unit step response of First order Systems
1.4 Unit step response of Second order Systems
1.5 Time response specifications

1.6 Time response specifications of second order
systems
1.7 steady state errors and error constants
1.8 Introduction to PI, PD, PID Controllers
(excluding design).
Motivation
 How to know the performance of a control system for any

input signal?
 How to design a control system which meets the desired
response and control requirements ?
Time Domain Analysis and Design Specifications
Time Domain Analysis
 Time domain analysis refers to the analysis of system performance in

time i.e., the study of evolution of system variables (specifically
output) with time
 There are two common ways of analyzing the response of systems:

1.Natural response and forced response
2.Transient response and steady state response
 In both cases, the complete response of the system is given by the

combination of both responses i.e., natural and forced responses or
transient and steady state responses
Natural and Forced Responses

 Natural response (Zero input response) :
–System's response to initial conditions with all external forces set to
zero
E.g. In RLC circuits, this would be the response of the
circuit to initial conditions (inductor currents or capacitor voltages) with
all the independent voltage and current sources set to zero
 Forced response (Zero state response) :

–System's response to external forces with zero initial conditions
E.g. In RLC circuits, this would be the response of the
circuit to only external voltage and current source, and zero initial
conditions
Transient and Steady State Responses
Transient response 𝒚𝒕𝒓(𝒕):

 –Part of the time response that goes to zero as time tends to be large
 –Transient response can be tied to any event that affects the
equilibrium of a system viz. switching, disturbance, change in input,
etc.
lim𝑡→∞𝑦𝑡𝑟(𝑡)=0
Steady state response 𝒚𝒔𝒔(𝒕):

 –Steady state response is the time response of a system after transient
practically vanishes and as time goes to infinity
 𝑦(𝑡)=𝑦𝑡(𝑡)+𝑦𝑠𝑠(𝑡)
Standard Test Inputs

 In most cases, the input signals to a control system are not known prior to design
of control system
 Hence to analyse the performance of a control system, it is excited with standard
test signals
 In general, control system design specifications are also based on the response of
the system to such test signals
 Standard test signals include:
–Unit impulse, unit step (sudden change), ramp (constant velocity),

parabolic (constant acceleration) and sinusoidal
–These inputs are chosen because they capture many of the possible
variations that can occur in an arbitrary input signal
Contd..
Contd..
Contd..
Contd..
Contd..
Contd..
Contd..
Contd..
Time Response Specifications
Time-domain Specifications: Motivation
 These specifications refer to the performance indices of the step

response of a system
 In general, these indices are specified as a part of the design
requirements of control systems
 These indices answer the following questions pertaining to step
response of a system:
–How fast the system moves to follow the input ?
–How oscillatory is the response (indicative of damping) ?
–How long does it take to practically reach the final value ?
Time-domain Specifications
Delay time 𝒕d: –Time required for the response to reach 50% of the final value at first instance
𝑡d=(1+0.7𝜁)/𝜔𝑛
Rise time 𝒕𝒓:
–Time required for the response to rise from 10% to 90% of the final value for
overdamped systems and 0 to 100% of the final value for underdamped systems, at first instance
Peak time 𝒕𝒑: –Time required for the response to reach the peak value of time response
Peak overshoot 𝑴𝒑: –It is the normalised difference between the peak value of time response and
the steady state value
M p = ymax − yss ;
Settling time 𝒕𝒔: –Time required for the response to reach and stay within a specified tolerance
band of its final value or steady state value
–Usually the tolerance band is 2% or 5%
Note: 𝑡𝑝 and 𝑀𝑝 are not defined for overdamped and critically damped systems
Time-domain Specifications
1. Steady-state value: yss
Mp
2. Maximum overshoot: Percent overshoot = 100%
yss
M p = ymax − yss ;
ymax
3. Peak time: tp Mp error band
4. Rise time: tr 0.9

yss
5. Settling time: ts
Is there overshoot in
the time response of a 0.1
first-order system?
Dept.of ECE,MITE,Moodabidri tr ts
35 tp
Expression for Rise time, peak time and

Overshoot
−nt
Unit-step e
y (t ) = 1 − sin(d t +  )，0<  1
response: 1−  2
1. Rise Time
−n tr
ee− wntr  sin( t +  ) = 0
y (tr ) = 1, that is, 1 −  sin( wdd trr +  ) = 1
1 −  22
−  t
1− 
e n r
 0,  d tr +  = n (n = 0, 1, 2, )
1−  2
 −  −
tr = =
tr is the time needed for d n 1 −  2
the response to reach the
steady-state value for For a given wn， ζ ↓ ，tr ↓ ；
the first
36
time,
Dept.of so n=1.
ECE,MITE,Moodabidri For a given ζ，wn↑，tr ↓ .
−nt
e dy (t )
y (t ) = 1 − sin(d t +  )，0<  1 =0
1−  2 dt
dy (t )  2n
2 . Peak time =( + n 1 −  2 )e− n t sin d t
dt 1−  2
n
= e− n t sin d=0
t
1−  2
 sin d t p = 0  d t p = n (n = 0, 1, 2, )
 
tp is the time needed for
d t p =   t p = =
the response to reach the d n 1 −  2
maximum value for the
first time, so n=1. For a given wn， ζ ↓ ，tp ↓ ；
For a given ζ，wn↑，tp ↓ .
−nt  
sin(d t +  )，0<  1 t p = w =
e
y (t ) = 1 −
1−  2 d wn 1 −  2
y (t p ) − y ()
3. Overshoot % = 100%
−  wn t p y ( )
e
y (t p ) = 1 − sin( +  ) Im
1−  2
d
sin( +  ) = − sin  = − 1 −  2 n
− Re
− / 1− 2
 y (t p ) = 1 + e −d
Suppose that y () = 1 1− 2
 = arctan
− / 1− 2

Thus  % = e 100%
Overshoot is a function of damping ratio ζ , independent of wn.

Dept.of ECE,MITE,Moodabidri
38
Expression for Settling time
−nt
e
4 Settling Time y (t ) = 1 − sin(d t +  )，0<  1
1−  2
With the definition of error band
e− n t
 sin( 1 −  2  nts +  )  0.05 or 0.02
1−  2
ts can not be obtained directly, but we can get the

relationship between wnts and ζ.
Relationship between ( ,n ) and (tr ,t p , %, ts )
 −  −
tr = =
d n 1 −  2 For a given n
   tr  t p 
 
tp = =
 d n 1 −  2
   %
 can be calculated by the requirement
 % = e − / 1− 2
 100% on the overshoot  %.   [0.4,0.8].
3 n   ts 
ts  ( = 5%)
n Once  is determined,n can be
4 determined based on the requirement
ts  ( = 2%)
40 
Dept.ofnECE,MITE,Moodabidri on error band %.
Error and steady-state error
Open-loop control system Closed-loop control system
R(s) E (s) Y (s)
R(s) Y (s) G( s )
G( s )
r (t ) y (t ) r (t ) e (t ) y (t )
Error: e(t ) = r (t ) − y (t )
Steady-state error: ess = lim e(t )
t →
Utilizing the final value theorem: lim f (t ) = lim sF ( s )

t → s →0
ess = lim e(t ) = lim sE ( s)

t → s →0
Assuming r(t)=1(t) is a unit-step input, according to the above

43 definition, could you calculate the steady-state error of the open-
loop and closed-loop control systems?
Open-loop control system Closed-loop control system
R(s) Y (s) R(s) E (s) Y (s)

G( s ) G( s )
r (t ) y (t ) r (t ) e (t ) y (t )
Unit-step input r(t)=1(t), 1

E ( s) = R( s) − Y ( s)
R( s) = E ( s) = R( s) − Y ( s)
s
= R( s) − G ( s) R( s) G (s)
= R( s) − R( s)
= [1 − G ( s )]R ( s ) 1 + G ( s)
ess = lim sE ( s ) =
1
R( s)
s →0 1 + G ( s)
1 1 1
= lim s[1 − G ( s )] ess = lim sE ( s ) = lim s
s →0 s s →0 s →0 1 + G ( s ) s
= lim[1 − G ( s )] 1 1
s →0
= lim =
s →0 1 + G ( s ) 1 + G (0)
= 1 − G (0)
Feedback has the effect to reduce steady-state error.
The forward-path transfer function G(s) can be
formulated as
ν is the order of
k ( s + 1) ( 22 s 2 + 21 2 s + 1)
G ( s) = v 1 the pole of G(s) at
s (T1s + 1) (T22 s 2 + 2 2T2 s + 1) s=0
k
= v
G0 ( s ) when s → 0, G0 ( s) → 1
s
System Type: the order of the pole of G(s) at s=0.
When ν=0,1,2, the system is called type 0, type 1, type 2; k is called open-
loop gain.
1
E (s) = R( s)
1 + G ( s)
1
ess = lim sE ( s ) = lim s R( s )
s →0 s →0 k
1 +  G0 ( s)
45 s 45
Example 1: Consider the following unit-feedback system
R(s) 5K A C(s)
- s ( s + 34.5)
System input is the unit-step function, When the

amplifier gains are KA=200, KA=1500, KA=13.5
respectively, can you calculate the time-domain
specifications of the unit-step response ?
Investigate the effect of the amplifier gain KA on the

system response
48
Solution: The closed-loop transfer function is
G( s) 5K A
 ( s) = = 2
1 + G ( s ) s + 34.5s + 5K A
1000
K A = 200,  ( s ) =
s 2 + 34.5s + 1000
n2 = 1000, 2n = 34.5

34.5
n = 31.6(rad / s),  = = 0.545
2n
According to the formula to calculate the performance

indices, it follows that

tp = = 0.12(sec)
n 1 −  2
3
ts  = 0.174(sec)
n
− / 1− 2
% = e 100% = 13%
K A = 1500
If K A = 200, then n = 34.5(rad / s);  = 0.545
 t p = 0.12( s), ts = 0.174( s),  % = 13%
If K A = 1500, then n = 86.2(rad / s);  = 0.2

 t p = 0.037( s), ts = 0.174( s),  % = 52.7%
Thus, the greater the KA, the less theξ,the greater the
wn, the less the tp, the greater theб%, while the settling
time ts has no change.  1
K A = 13.5 Overdamped
When K A = 13.5, n = 8.22(rad / s),  = 2.1

51
When K A = 13.5, n = 8.22(rad / s),  = 2.1

When system is over-damped, there is no peak time,
overshoot and oscillation.
The settling time can be calculated approximately:
t s  3T = 1.46(sec)
1
= n ( −  2 − 1)
T
The settling time is greater than previous cases,
although the response has no overshoot, the transition
process is very slow, the curves are as follows:
52
 c (t )
 = 0.2( K A = 1500 )
1.4
1.2
1.0
0.8
0.6
 = 2.1( K A = 13.5)
0.4  = 0.545( K A = 200)
0.2
0 2 4 6 8 10 12 14 16 18 wn t
Note: When KA increases， tp decreases， tr decreases，
the speed of response increases, meanwhile, the overshoot
increases. Therefore, to improve the dynamic performance
indexes of system, we adopt PD-control or velocity
53feedback control，namely, PD compensation 。
53
1
ess = lim sE ( s ) = lim s R( s)
s →0 s →0 k
1 + 0 G0 ( s)
ν=0, type 0 system s
Step input: Steady-
1
1 ess = state error
r (t ) = 1(t ) R ( s ) = 1+ k exists and
s
is finite.
Ramp input:
1 ess =  Unstable
r (t ) = t R( s) = 2
s
Parabolic input:
1 1
r (t ) = t 2 R( s ) = 3
ess =  Unstable
2 s
56
1
ess = lim sE ( s) = lim s R( s)
s →0 s →0 k
1 + 1 G0 ( s)
Step input: 1
ess = =0 No steady-
1
r (t ) = 1(t ) R ( s ) =
s
1+  state error
Ramp input: 1
1 ess = Steady-state
r (t ) = t R( s ) = 2 k error exists
s
Parabolic input:
ess =  Unstable
1 1
r (t ) = t 2 R( s ) = 3
2 s
Type-1 system can track step signal accurately.

57 57
1
ess = lim sE ( s ) = lim s R( s )
s →0 s →0 k
1 + 2 G0 ( s)
Step input: 1
1 ess = =0 No steady-
r (t ) = 1(t ) R ( s ) = 1+  state error
s
Ramp input:
1 ess = 0 No steady-
r (t ) = t R( s ) = 2 state error
s
1
Parabolic input: ess = Steady-state
1 1 k error exists
r (t ) = t 2 R( s ) = 3
2 s
Type-2 system can track step and ramp signals accurately.

58 58
Steady-state error constants
with step input
- step-error constant
k p = lim G ( s )
s →0
with ramp input
- ramp-error constant
k = lim sG ( s )
s →0
with parabolic input
- parabolic-error constant
ka = lim s 2G ( s)
s →0
59
Type of
System
Error
constants
Steady-state error ess
j k p k v k a r (t ) = R0 1(t ) r (t ) = V0t r (t ) = A0t 2 2
R0
0 k 0 0  
1+ k
V0
Ⅰ  k 0 0 
k
A0
Ⅱ   k 0 0
k
Summary of steady-state error and error constants
for unit-feedback
60
systems (H(s)=1)
60
Example 1
Consider a first-order system with the following parameters.
Time constant  = 0.1s ; DC gain K = 20
The problem is to determine the unit impulse response of a
system that has these parameters.
K 20 − t /0.1
Solution. y (t ) = e−t  = e = 200e −10t
 0.1
Using MATLAB to get its

impulse response
K 20
G (s) = =
 s + 1 0.1s + 1
>> num=20;
>> den=[0.1 1];
>> impulse(num,den)
Example 2 Below is the impulse response of a system -
i.e. the response to a unit impulse.
The system starts with

an initial condition of
zero just before the
impulse comes along at t
= 0, so y(0-) = 0.
CouldDept.of
65 you compute the parameters of the system?
ECE,MITE,Moodabidri
Solution. Transfer function:

The general time response of a K 44
G( s) = =
first-order system is
 s + 1 2.2s + 1
K
y (t ) = e−t  (1) A

From the right figure, we have
at t = 0, y (0) = 20
So we can get
K
= 20 (2) With (2) and (3), we have
 y (2) = 20e−2/ = 8
How to get K ?
Pick up a point on the response curve: e −2/ = 0.4
Solving the equation yields:
A: at t = 2sec; y (2) = 8 −2 /  = ln(0.4) = −0.9163
66 Dept.of K
y (2) e −2  = 8
= ECE,MITE,Moodabidri (3)   2.2sec K = 20 = 44

Example 3 Consider a first-order system with the following
parameters.
Time constant  = 0.1s ; DC gain K = 20
The problem is to determine the unit step response of a
system that has these parameters.
Solution.
Using the general form of the unit-step response of first –order systems
y (t ) = K (1 − e −t / )
With the parameters above, the impulse response is:
y (t ) = 20(1 − e −t /0.1 ) = 20(1 − e −10t )
Example 3 Time constant  = 0.1s ; DC gain K = 20

Solution. Unit-step time response: y (t ) = 20(1 − e −10t )
Transfer function of the K 20
first-order system: G (s) = =
 s + 1 0.1s + 1
Using MATLAB to get its
step response
>> num=20;
>> den=[0.1 1];
>> step(num,den)
If we keep  = 0.1s
but change K,
what will
happen?
Time constant  = 0.1s ;
DC gain K = 20 → 16 → 10 → 4
DC gain K = 20
Time constant  = 0.1s → 0.2s → 0.3s → 0.5s
Note: the time constant reflects the system inertia. The smaller is the
is  and the quicker is the response, vice versa.
system inertia, the shorter
70
Two important properties of the unit-step response of a
first-order system:
a) The time constant  can be used to calculate the
system output
experiment methods
t =  , y ( ) = 0.632 y () = 63.3% final value; -- estimate the
time constant
t = 2 , y (2 ) = 0.865 y () = 86.5% final value;
--judge whether a
t = 3 , y (3 ) = 0.950 y () = 95.0% final value; system is first-
order or not
t = 4 , y (4 ) = 0.982 y () = 98.2% final value;
b) The initial slope of the response curve is 1/T and

the slope decreases with time
dy (t ) 1 dy (t ) 1 dy (t )
= , = 0.368 , = 0.
dt t =0  dt t =  dt t =
It is also a common method to obtain the time constant though the initial
71
slope in control engineering.
Unit-step response of a first-order system
y (t )
y ( )
y (t ) = K (1 − e −t / )
0.632 y ()
1
slope =

0 t

Controllers
73
cs1567
Introduction
 Suppose you have a system that needs to be controlled
 Your software gives commands, the system responds to it
 Turn x degrees to the right
 Move forward 15 wheel rotations
 Can you always trust your commands will be executed
accurately?
74 cs1567
Problem example
overshoot Ripple
setpoint
 Increase the quantity until you get to the setpoint

 Temperature, angle, speed, etc
 If too much, reduce the quantity, until the setpoint
75 cs1567
Closed loop controller
setpoint error signal output

controller system
measured quantity
• closed loop because it has feedback

• output is measured at a certain frequency
• signal is generated at a certain frequency
• which frequency is not smaller?
76 cs1567
PID stands for: P (Proportional)
I (Integral)
D (Derivative)
On-off control
 For some systems, on-off signaling is sufficient
 For example, a thermostat, when the heater is either on
or off, and early cruise-control systems
 Could do airflow or speed control also
 More modern systems do it
 Depending on the frequency of control, overhead of on-
off, etc, this could cause overshoots and undershoots
(ripples)
 Oscillation is a common behavior in control systems
 Need to avoid it at all costs… well, almost all costs
78 cs1567
Proportional control
 Good alternative to on-off control: more “control” ☺

 Signal becomes proportional to the error
 P ( setpoint – output )
 Example, car speed for cruise control
 Need to find out value of constant P
 Tuning the controller is a hard job
 If P is too high, what happens?
 If P is too low, what happens?
 Typically a prop cntrl decreases response time (quickly gets
to the setpoint) but increases overshoot
79 cs1567
Adding derivative control

 To avoid (or reduce) overshoot/ripple, take into account
how fast are you approaching the setpoint
 If very fast, overshoot may be forthcoming: reduce the signal
recommended by the proportional controller
 If very slow, may never get to setpoint: increase the signal
 In general: D ( current measure – previous measure)
 PD controllers are slower than P, but less oscillation,
smaller overshoot/ripple
80 cs1567
Integral control
 There may still be error in the PD controller
 For example, the output is close to setpoint
 P is very small and so is the error, discretization of signal will provide no
change in the P controller
 D controller will not change signal, unless there is change in output
 Take the sum of the errors over time, even if they’re
small, they’ll eventually add up
 I * sum_over_time (setpoint — output)
 Again the main problem is the value of I
 Can we let sum grow to infinity?
81 cs1567
Summary
 Different types of controllers
 PID hardest task is tuning
Controller Response Overshoot Error

time
On-off Smallest Highest Large
Proportional Small Large Small
Integral Decreases Increases Zero
Derivative Increases Decreases Small change
82 cs1567
At the end of this lesson, the student should be
able to:
 Write the input-output relationship of a P-I-D controller
 Explain the improvement of transient response in closed loop
with P-controller
 Explain the presence of offset in presence of simple P-controller
 Define Proportional Band
 Explain the elimination of steady state error with Integral
Control.
 Define the error transfer function and compute steady state error
 Explain the advantages of P-I controller over simple P and I
actions
 Explain the effect of P-D controller
 Recommend a suitable controller configuration for a particular
process.

Module 3

Uploaded by

Document Information

Original Title

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

Module 3

Uploaded by

Copyright:

Available Formats

Module -3

Time Response of Feedback Control

1.4 Unit step response of Second order Systems

1.5 Time response specifications

 How to know the performance of a control system for any

Time Domain Analysis and Design Specifications

 Time domain analysis refers to the analysis of system performance in

 There are two common ways of analyzing the response of systems:

 In both cases, the complete response of the system is given by the

Natural and Forced Responses

 Forced response (Zero state response) :

Transient response 𝒚𝒕𝒓(𝒕):

Steady state response 𝒚𝒔𝒔(𝒕):

Standard Test Inputs

–Unit impulse, unit step (sudden change), ramp (constant velocity),

Time Response Specifications

 These specifications refer to the performance indices of the step

4. Rise time: tr 0.9

Expression for Rise time, peak time and

Overshoot is a function of damping ratio ζ , independent of wn.

With the definition of error band

ts can not be obtained directly, but we can get the

Relationship between ( ,n ) and (tr ,t p , %, ts )

Utilizing the final value theorem: lim f (t ) = lim sF ( s )

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

Assuming r(t)=1(t) is a unit-step input, according to the above

Open-loop control system Closed-loop control system

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

Unit-step input r(t)=1(t), 1

Example 1: Consider the following unit-feedback system

System input is the unit-step function, When the

Investigate the effect of the amplifier gain KA on the

n2 = 1000, 2n = 34.5

According to the formula to calculate the performance

If K A = 1500, then n = 86.2(rad / s);  = 0.2

When K A = 13.5, n = 8.22(rad / s),  = 2.1

When K A = 13.5, n = 8.22(rad / s),  = 2.1

Type-1 system can track step signal accurately.

Type-2 system can track step and ramp signals accurately.

Using MATLAB to get its

The system starts with

Solution. Transfer function:

A: at t = 2sec; y (2) = 8 −2 /  = ln(0.4) = −0.9163

With the parameters above, the impulse response is:

y (t ) = 20(1 − e −t /0.1 ) = 20(1 − e −10t )

Example 3 Time constant  = 0.1s ; DC gain K = 20

b) The initial slope of the response curve is 1/T and

Unit-step response of a first-order system

 Increase the quantity until you get to the setpoint

Closed loop controller

setpoint error signal output

• closed loop because it has feedback

 Good alternative to on-off control: more “control” ☺

Adding derivative control

Controller Response Overshoot Error

You might also like