UNIT II

CLOSED LOOP CONTROL SYSTEMS

Syllabus:
 Development of block diagram for feed-back control systems
 Servo and regulator problems in closed-loop control system
 Transfer function for controllers and final control element
 Principles of pneumatic and electronic controllers
 Transient response of closed-loop control systems
 Stability of closed-loop control system
 Transportation lag

Open loop system :

Manual mode of Operation

Closed loop system

 Auto mode of Operation
 CONCEPT OF FEEDBACK CONTROL

Open loop system:

d

m y
Process

Closed loop system:

d
Controller Mechanism
ysp e Final m y
+ c
Controller Control Process
- Element

Sensor/
ym transmitter
 Representation of a linear system

 System without controller(Open loop)

Pictorial representation
Mathematical representation
Nth order differential equation
N first order differential equations
matrix differential equation
Transfer function
 Block diagram representation
System with controller(Closed loop)
Feed-back control systems
Positive feed back
Negative feedback
 Components of a control system

Process ( Stirred – tank heater )

Measuring element (thermometer)
 Controller(P,PI,PID Controllers)
 Final control element(Power driver/Control valve)

Heating system with process control components:

CONTinued…

Closed loop block diagram of a simple feed back control system:

TR
Servo problem:
No change in load [Ti] and Changes only in set point[TR]

Regulatory problem:
No change in set point[TR ] and Changes only in set point load [Ti]
 DEVELOPMENT OF BLOCK DIAGRAM

 Each block represents the functional relationship existing between the input
and output of a particular component.

 Variables selected In block diagram representations of control systems are

deviation variables

 The transfer function relating the input – output pair of variables are placed
inside of each block

 Individual components blocks are combined to give the overall block

diagram.

 Process have two input variables and one output variable.

 Procedure of developing transfer function for the process is similar to other

components
 Energy balance on stirred tank heater

q  wC Ti  To   wC T  To   CV
dT
     (1)
dt

1 '
wC Q( s )  i ( s )  T ' ( s )      (2)
T
s  1 s  1
If there is a change in Q (t) only, then Ti’ (t) = 0 and the transfer function relating T’ to Q
1
T ' ( s)
 wC      (10)
Q( s ) s  1
If there is a change in Ti’ (t) only, then Q (t) = 0 and the transfer function relating T’ to Ti’

T ' ( s) 1
      (11)
Ti ( s) s  1
'

1
T ' ( s) 
s  1
 i 
wC Q( s)  wCT ' ( s)      (12)
Continued…
Block diagram of heating process:

Measuring element:

'
Tm ( s ) 1

T ' ( s)  ms 1

Summer:

Error, ϵ = TR – Tm
CONTinued…

Controller and Final Control Element(FCE):

 Controller and the final control element are combined into one block
 For the combined block, Input is error and output is heat
 Controller used the block is proportional controller

The relationship for a proportional controller   TR  Tm

TR  setpo int Temperature
q  K C   A    (1) K C  Controllergain
At steady state, A  Heatinput ; when,   0

TRS  TS  TmS      (2)

Let ϵ’ be the deviation variable

 '     s      (3)

 S  TR  Tm  0      (4)
S S

 '        (5)
CONTinued…

At steady state, equation (1) becomes

qS  K C  S  A 0 A  qS  A

q  K C   qS  q  qS  K C   Q  KC       (6)

Taking Laplace on both sides

Q(s)  KC  (s)      (7)

   ',   TR '  Tm '      (8)

 (s)  TR ' (s)  Tm ' (s)      (9)

Block diagram of Proportional controller and FCE:
CONTinued…

Summary of block diagram closed loop control system:

Block diagram gives:

 Relationship among the process control components
 Feedback relationship between measured variable & desired variable
Block diagram of servo and regulatory problem
 Control Valve (FINAL CONTROL ELEMENT)
Pneumatic control valve:

Selection of the type of valve:

 Selection of the type of valve is made based on safety considerations

 Example: air – to – close Control valve is used in controlling the inlet
flow of cooling water to a cooling jacket on an exothermic chemical reactor
CONTinued…
 Control valve consists of an actuator and a valve.
 Control valve itself is divided into the body and trim.
 The body consists of housing for mounting the actuator
 Valve Connections for attachment to a supply and delivery line.

Electronic control valve:

CONTinued…

Block diagram of transducer, Controller and Converter in Electronic control valve:

Ideal Transfer functions of Pneumatic Control valve:

Q( s ) Kv Kv -- Steady state gain (Constant of

 proportionality between steady state flow
P( s )  v s  1 rate and valve top pressure)

Q( s ) v is the time constant of the valve (Time

 Kv constant of the valve is very small, then it is
P( s)
negligible.
 Pneumatic and Electronic controllers

The choice between electronic and pneumatic control will probably

be decided by the relative costs of installation and maintenance and the
reliability of the instruments.

Difference between pneumatic and electronic controllers:

 CONVENTIONAL CONTROLLERS

Controllers:
 Input to the controller is error signal (ϵ)
 Output of the controller is pneumatic signal (P)
 Transducer and converter are combined with the controller for simplicity

Types of Controllers:
 Proportional control
 Proportional integral control
 Proportional derivative control
 Proportional integral derivative control
 PROPORTIONAL – INTEGRAL (PI) CONTROL

 Integral control mode is introduced to eliminate the residual error

 It is termed as PI or proportional-integral control
The integral mode ultimately drives the error to zero
 The values of Kc and τI are both adjustable

t
KC
Pr oportional  int egralcontroller , p  KC  
I   dt  p
0
s    (4)

where, K C  proportiona lg ain

 I  int egraltime , min
p s  cons tan t (thebiasvalue)

KC
p(t )  K C   ps    (5)
I

P  p  ps
 CONTinued…

 To visualize the step response , Unit step change in error (ϵ =1) is given
 ‘p’ changes suddenly by an amount Kc and linearly with K C /  I
 Reciprocal of I is called as reset rate

Step response of PI Controller:

Transfer function for this controller:

P( s )  1 
Pr oportional  int egralcontr ollertrans ferfunctio n, 
 K C 1      (6)
 ( s)  Is 
 PROPORTIONAL – DERIVATIVE (PD) CONTROL

 Derivative control mode is introduced to reduce the process oscillations

 It is termed as PD or proportional- derivative control
The values of Kc and τD are both adjustable

d
Pr oportional  derivativecontroller , p  K C   K C D  ps    (7)
dt

where, K C  proportional gain

 D  derivative time, min
ps  cons tan t (the bias value)

p(t )  AK C  AK C D  ps    (8)

P  p  ps
 CONTinued…

 To visualize its response , a linear change in error is given [  (t )  At]

 ‘p’ changes suddenly by an amount AKC τD and linearly with AKC.

Step response of PD Controller:

Transfer function for this controller:

 K C 1   D s     (9)
P( s )
Pr oportional  derivativecontroller transferfunction ,
 ( s)
 PROPORTIONAL – INTEGRAL - DERIVATIVE (PID) CONTROL

 It is a combination of the integral and derivative control mode

 It is termed as PID or proportional- integral - derivative control
The values of Kc , τI and τD are adjustable

d K C
t

dt  I 0
Pr oportional  int egral  derivative controller , p  K C   K C D  dt  ps    (10)

Transfer function for this controller:

P( s )  1 
Pr oportional  int egral  derivativecontroller transferfu nction ,  K C 1   D s      (11)
 ( s)   I 
s

 Derivative control action is based on how rapidly the error is changing

 Rapidly changing error signal will induce a large derivative response.
 Derivative control should be avoided for “Noisy” error signals.
 Integral control action will give error free process output
 ON/OFF CONTROL
 A special case of proportional control is ON/OFF control
 For maximum gain Kc , the valve will move from one extreme position to
other even for a slight deviation in set point. i.e. the valve acts as a switch.
 In practice, a dead band is inserted into the controller. With a dead band,
the error reaches some finite positive value before the controller “turns on”.
 Conversely, the error must fall to some finite negative value before the
controller “turns off”.

Step response of ON-OFF Controller:

CONTinued…

Step response of all type of Controllers:

 CLOSED LOOP TRANSFER FUNCTIONS
Closed loop block diagram:

B = measured Variable ϵ = Error

M = Manipulated variable GC = Transfer function of controller
U = Load variable or disturbance G1 = Transfer function of FCE
R = Set point or desired value G2 = Transfer function of Process
C = Controlled variable H = TF of measuring element
Continued…

Block diagram of servo problem :

Block diagram reduction :

 Several Transfer Functions in series are simplified as product of individual TF

 G = GCG1G2

C  G    (1)

B  HC    (2)

  R  B    (3)

Solving these three equations simultaneously , C in terms of R are obtained

Continued…

Overall closed loop transfer function :

C  G( R  B)
 C  G( R  HC)
 C  GR  GHC
 C(1  GH )  GR
C G
    (4)
R 1  GH
Overall closed loop block diagram :
Continued…
Block diagram of regulatory problem :

C  G2 (U  M )      (5) B  HC      (7)

M  GC G1      (6)   B      (8)

Solving these four equations simultaneously , C in terms of U are obtained

C  G2 (U  GC G1 )
 C  G2 (U  GC G1  HC)
 C  G2U  GC G1G2 HC
C G2
C (1  GH )  G2U      (9)
U (1  GH )
 Block diagram reduction rule

 Transfer function relating any pair of variables X, Y

Y f
 ( Negative feedback )      (10)
X 1 l

Y f
 ( Positive feedback )      (11)
X 1  l

Where,
πf = Product of transfer functions in the path between the locations of the signals X
and Y.
πl = Product of all transfer functions in the loop.(ie. πl = GCG1G2H)

 Overall Transfer Function based on the block diagram reduction rule

C GC G1G2 G
 
R 1  GC G1G2 H 1  GH
TRANSIENT RESPONSE OF SIMPLE CONTROL SYSTEMS

1
wC
 Servo Response with P Controller

For proportional control, GC  K C

KC A
T'

s  1      (1)
TR' 1  K C A
s  1
T' KC A
 ' 
TR s  1  K C A


T'

KC A offset  TR'   T '     (3)
 
1  K C A  s  1
'
TR
1  K C A 

KC A 1  KC A  KC A

KC A
 
  1 
A
Let, 1
1  KC A
& 1
1  KC A 1  KC A 1  KC A

1
T' A1 offset     (4)
 '       (2) 1  KC A
TR  1 s  1
 REGULATORY RESPONSE WITH P CONTROLLER
In regulatory problem, TR  0
'

AA 1
T'

s  1
Ti ' 1  KC A
s  1
T' 1
 
Ti ' s  1  K C A

T' 1 1  KC A
 ' 

offset  TR'   T ' 
Ti
s 1
1  KC A
1  1
Let, A2  &1   0
1  KC A 1  KC A 1  KC A
1
offset  
T' A2 1  KC A
 
Ti '  1 s  1
In Proportional control, K C  offset 
 REGULATORY Response with PI Controller
 
PI Controller transfer function is given as GC  KC 1  1 s 
 I 

Overall transfer function for load change:

AA 1
T'

s  1    (1)
Ti '
 1  A
1  K C 1   
  I s  s  1

1

s  1
s  1 I s   KC A I s  1 s  1 I s 
Is Is
 
s  1 I s   KC A I s  1  I s 2   I s  K C A I s  K C A
 I s
Is  K C A 
 T'
 '        (2)
 I s 2   I s1  KC A  KC A Ti   I  2  1 
  s   I s  1  1
 KC A   KC A 
Continued…

I  I  I
A1     (3); 1   1       (4)
2
Let, KC A KC A KC A

 1   I  1 
2 1   I 
  1
  2   I   1
 KC A  KC A  KC A 

 I  KC A  1 
 2  I I  
KC A  K A K A
 C C 

1 I  1  KC A 
       (5)
2   K A 
 C 

T' A1s
Substitute equation (3), (4) & (5)
      (6)
in equation (2) Ti  1 s  2 1s  1
' 2 2
Continued…
1
For a unit step change in load{ Ti  }in equation (6) gives
'

s
1 As A1
T '  2 2 1  T'    (7)
s  1 s  2 1s  1  1 s  2 1s  1
2 2

 Eqn (7), response of the tank temperature to an impulse function of magnitude A1.
The impulse response of this system for  1


1 
1 t /  2 t 
T  A1 
'
e sin 1         (8)

 1 1
 2  1 
 Continued…

The effect of K C &  I on the unit step response :

offset  TR'   T '   0  0  0

One of the most important advantages of PI control is the elimination of offset.

 STABILITY OF CLOSED LOOP SYSTEM

Concept of stability:

 If the system less (or) equal to second order, the response of system is
inherently stable.

 If the system is higher than second order, the problem of stability in a

control system is slightly more complicated.

Stability criterion:

 A linear control system is unstable if any roots of its characteristic equation

are on or to the right of the imaginary axis. Otherwise the system is stable.
 The characteristic equation is 1 + G = 0. Where, G  G1G2 H
 The characteristic equation of a control system determines its stability and
is the same for set point or load changes.
 Continued…
Closed loop block diagram and Transfer function:

C G

R 1  GH

Location of the roots of the Characteristic equation of the system, 1 + GH = 0

.
 ROUTH TEST FOR STABILITY

 Routh test is a purely algebraic method for determining how many roots of the
characteristic equation have positive real parts
 In the characteristic equation, if there are no roots with positive real parts, the
system is stable.
 The test is limited to systems that have polynomial characteristic equations and
not suitable for the system with transportation lag.
 To check the roots of the given system, Write the characteristic equation in the
n 1
form of a0 s  a1s
n
 a2 s n2        an  0 , Where, a0 is positive.
 If any coefficient is negative, the system is definitely unstable and the Routh
test is not needed to answer the question of stability.

 If all the coefficients are positive, the system may be stable (or) unstable.
Routh array table:
Continued…
Routh stability table:

a a  a 0 a3 a a  a 0 a5 a1a6  a0 a7
b1  1 2 b2  1 4 b3 
a1 a1 a1

b1a3  a1b2 b1a5  a1b3

c1  c2 
b1 b1
Problems
Example: 1
Determine the transfer functions C/R, C/U1 and B/U2 for the system show in
Fig. Also determine an expression for C in terms of R and U1 for the situation when
both set-point change and load change occur simultaneously.
Using the rule given by equation (10) we obtain by inspection the results.

C GC G1G2G3
(i)       (12) (ii )
C GG
 2 3      (13)
R 1  GC G1G2G3 H1H 2 U1 1  GH
G

1  GH
Where, B G3 H1H 2
(iii )       (14)
G  GC G1G2G3 & H  H1 H 2 U 2 1  GH
 For separate changes in R and U1, we may obtain the response C from
equation (12) & (13)
GC G1G2G3
C R      (15)
1  GH
and
G2G3
C U1      (16)
1  GH

If both R and U1 occur simultaneously the principle of super position

requires that the overall response be the sum of the individual responses
GC G1G2 G3 GG
C R  2 3 U 1      (17)
1  GH 1  GH
Overall transfer function for Multi loop control systems:
To illustrate how one obtains the overall transfer function for a multi loop
system, in which the method used is to reduce the block diagram to a single –loop
diagram by application of the rules summarized by positive feed back and negative
feed back equation.
 Example: 2 Determine the transfer functions C/R for the system show in Fig. This
block diagram represents a cascade control system, which will be discussed later.

Step: 1 Reducing the inner loop and we may also combine G2 & G3 into a single block
C GC1 Ga Gb

R 1  GC1 Ga Gb H1
C G2G3

U1 1  GC1 Ga Gb H1

C G3
Step: 2 Single loop can be converted into single block. 
U 2 1  GC1 Ga Gb H1

B G3 H1

U 2 1  GC1 Ga Gb H1
Problem: 3
A unit step change in error is introduced into a PID controller. If KC = 10, τ1 =1
and τD = 0.5. Plot the response of the controller P (t).
Solution:
WKT, the transfer function for PID controller is
P( s )  1 
 K C 1   D s        (1)
 ( s)  Is
Given Data:
KC = 10, τ1 =1, τD = 0.5 t Vs P(t)
120
Substitute the above value in equation (1) 100
80
P( s )  1 60
 101  0.5s      (2) P(t)

 ( s)
40
 s 20
0
1
unit  stepchange   ( s)  0 2 4 6 8 10 12

s
Substitute the above value in equation (2)

10  1 10 10  P( s )  5 
10 10
 2    (3)
P( s)  1  0.5s    P( s)   2 5 s s
s  s s s
Taking inverse Laplace
P(t )  101  t     (4)
Problem 1:
The set point of the control system shown in Fig A is given a step change of
0.1 unit. Determine: (i) the maximum value of C and the time at which it occurs (ii) the
offset (iii) the period of oscillation. Draw a sketch of C (t) as a function of time.
(NOVEMBER 2012 & R2007)

The transfer function for the given system is

5
1.6 
C

s  12 s  1
R 5
1  1.6 
s  12 s  1
8

s  12 s  1 
8

8
s  12s  1  8 2s 2  s  2s  1  8 2s 2  3s  9
s  12s  1
C 0.889

R 0.222s 2  0.333s  1
  2  0.222    0.471
0.333
2  0.333     0.354
2  0.471
  0.354    1
So the response equation is at under damped condition.
Step change is
0.1
R(t )  0.1  R( s) 
s
0.1 0.889
C 
s 0.222s 2  0.333s  1

  1 

2
1 t /   2 t
C  0.08891  sin 1    tan 1

e
 1    
2
  

  1  0.354 2 
C  0.08891 
1 0.354t / 0.471 
sin 1  0.354 2 t
 tan 1 

e
 1  0.354  0.471 0.354 
2
  

 
C  0.0889 1  1.069e0.752t sin1.986t  1.209      (1)
(i) Maximum value of ‘C’ occurs at peak time tp:
     0.471
tp    tp   1.581 min
 1  2 1  0.354 2

At t p  1.581min ,the maximum value of “C” can be substituted in equation (1)


Cmax  0.0889 1  1.069e0.7521.581 sin1.986 1.581  1.209 
Cmax  0.117
(ii)offset  R  C 
0.0889
C ()  lt C (t )  lt sC(s)  lt s 
t  s 0 s 0 
s 0.222s 2  0.333s  1 
0.0889
  C ()  0.0889 offset  0.1  0.0889  0.0111
1
(iii) Period of oscillation: t Vs C(t)

C(t)
0.12
1 1  2 1
 f  0.316  T  
1
 3.163 0.08
0.1
f 
2  f 0.316 0.06

Period of oscillation, T  3.163 time units

0.04
0.02
0
Graph is plotted between “t” Vs “C(t)” 0 2 4 t6 8 10
Problem 2:
The location of a load change in a control loop may affect the system
response. In the block diagram shown in Fig C. a unit-step change in load enters at
either location 1 or location 2. (a) What is the frequency of the transient response
when the load enters at location 1 and when the load enters at location 2? (b) What is
the offset when the load enters at location 1 and when it enters at location 2? (c)
Sketch the transient response to a step change in U1 and to a step change in U2.
(NOVEMBER 2013 & R2007, NOVEMBER 2012 & R2008, MAY 2012 & R 2008)

 2  1 
  
C  2 s  1  2 s  1 

U1  2  1 
1  5  
 2 s  1  2 s  1 
 2
C 2 s  1
2
2
U1


2s  12  10  
4s 2  4s  11
2s  12
Apply step change U 1 ( s)  1
s
2
1 2 1 11
C  2  
4 2 4
s 4s  4s  11 s s  s 1
11 11
0.182


s 0.364s 2  0.364s  1 
 2  0.364    0.603; 2  0.364    0.302    1
Taking inverse Laplace

  1  0.302 2 
C (t )  0.1821 
1 0.302t / 0.603 
sin 1  0.302 2 t
 tan 1 

e
 1  0.302  0.603 0.302 
2
  


C (t )  0.182 1  1.049e 0.5t sin1.581t  1.4798 
  1 
 
C  2 s  1 

U2  2  1 
1  5  
 2 s  1  2 s  1 

1
C 2s  1 C 2s  1
  
U2 
2s  12  10  U2 4s 2  4s  11
2s  12
Unit step change in location 2.
1
U 2 (s) 
s

C ( s) 
1

2s  1 11
s 0.364s 2  0.364s  1
2 s 1
C ( s)  11  11
  
s 0.364s 2  0.364s  1 s 0.364s 2  0.364s  1 
 0.182 0.091
 
0.364s  0.364s  1 s 0.364s 2  0.364s  1
2
 
Taking inverse Laplace

 1 
C (t )  0.182
1
 
e 0.5t sin 1.581t   0.0911  1.049e 0.5t sin 1.581t  1.478
 0.603 1  0.302 
2

  
C (t )  0.182 1.781e0.5t sin1.581t  0.091 1  1.049e0.5t sin1.581t  1.478 
Yes, the response equation will vary depend on the load variations in location 1
or location 2.
(a)Frequency of response is same for both location, because the denominator
has second order differential equation in both, So

1 1 2 1 1  0.302 2
f    f  0.251
2  2 0.603

T  3.97time / cycle
 (b) Offset:
For location 1
C ()  lt C (t )  lt sC( s)
t  s 0

0.182
 lt s 
s 0 
s 0.364s 2  0.364s  1   C ()  0.182

offset  R  C   0  0.182

0.3
offset  0.182
0.25
For location 2
C ()  lt C (t )  lt sC(s) 0.2
t  s 0

 lt s 
2s  1 11 0.15 Location 1

s 0 
s 0.364s  0.364s  1
2
 0.1
Location 2

1 11
 C ( )   0.091
1 0.05

offset  R  C   0  0.091

0
offset  0.091 0 2 4 6 8 10

So offset is different for the location 1 & location 2.

(c)Graph for “t” Vs “C(t)”
Problem 3:
A PD controller is used in a control system having a first order process and a
measurement lag as shown in Fig.
(a) Find the expressions for and for the closed loop response.
(b) If 1  1min,  m  10 sec , find KC so that   0.7 for the two cases; (i)  D  0
 D  3 sec
(ii)

© Compare the offset and period realized for both cases and comment on the
advantage of adding the derivative mode.

 1 
K C 1   D s  
C   1s  1 

R  1  1 
1  K C 1   D s   
  1 s  1   m s  1 
Solve the above equation algebraically, we have
KC
1   D s  m s  1
C 1  K
 C

R  m 1 2     m  K C D 
s  s 1   1
1  KC  1  KC 
 m 1  m 1
2   
1  KC 1  KC

 1   m  K C D 1   1   m  K C D 
2     
1  KC 
2  1 m  1  KC 


(a) The expressions for and for the closed – loop response is

 m 1

1  KC

1   1   m  K C D 
   
2  1 m  1  KC 

(b) Given  1  60 sec, m  10 sec,   0.7 : findK C

For  D  0

1  60  10  0 
0.7   
2 60 *10  1  K C 


Solve the above equation algebraically, we have

1  K C  4.166  K C  3.166
For  D  3 sec
1  60  10  3K C 
0.7   
2 60 *10  1  KC 

Solve the above equation algebraically, we have
9K C  756K C  3724  0  K C  78.7(or )5.25
2

(c)Offset:

C ()  lt C (t )  lt sC(s)
t  s 0
 Step change is applied
KC
1   D s  m s  1
1 1  KC KC
 lt s   C ( ) 
s0 s  m 1 2     m  K C D  1  KC
s  s 1   1
1  KC  1  KC 

1  KC  KC
offset  R  C   1 
KC 1
  offset 
1  KC 1  KC 1  KC

1
For  D  0, K C  3.166 : theoffset   0.24
1  3.166
For , D  3 sec, K C  78.7(or )5.25 :
1
K C  78.7 : offset   0.013
1  78.7
1
K C  5.25 : offset   0.16
1  5.25

Thus adding the derivative mode to the proportional action, the offset is
decreased.
 Problem 4:
•For the control system in Fig. Obtain the closed loop transfer C/U.
•Find the value of proportional gain for which closed loop has a damping coefficient is
2.3.
•Find the offset for a unit step change in U if KC = 4.

The transfer function for closed loop

1
 
C
 s
U  s  1  1 
1  KC   
 0.25s  1  s 
Solve the above equation algebraically, we have
1
0.25s  1
C KC

U 0.25 2  1  KC 
s  s   1
KC  KC 
(a) The transfer function for the closed loop system is
1
0.25s  1
C KC
    (1)
U 0.25 2  1 KC 
s  s   1
KC  KC 

(b) Find the value of KC at   2.3 from equation (1)

0.25

KC
1  KC 1  1  KC  1  1  KC 
2       2.3   
KC 2  K C  2  KC 
Solve the above equation we have
K C  3.29K C  1  0  K C  2.95(or )0.339
2

1
(c)U ( s)  ; K C  4
s
1
1
0.25s  1
C (s)   4
s 0.25 2 5
s  s   1
4 4
 1
1
0.25s  1 14
C ()  lt C (t )  lt sC(s)  lt s   4   C ()  0.25
t  s 0 s 0 s 0.25 2 5 1
s  s   1
4 4
offset  R  C   0  0.25
offset  0.25
Problem 5:
For the control system shown in Fig with R = 2/s, Determine (i) C (s) / R (s) (ii)
C(∞) (iii) offset (iv) C(0.5) (v) Whether closed loop response is oscillatory? (MAY
2013 & R2007, NOVEMBER 2012 & R2008, NOVEMBER 2011 & R2007)

(i) The transfer function for closed loop

2
2
C ( s) s s  1

R( s) 2
1 2 
s s  1
Solve the above equation algebraically, we have
C ( s) 1

R( s) 0.25s 2  0.25s  1

(ii )C ()  lt C (t )  lt sC(s)

t  s 0

We know R(s) = 2/s

2 1  C ()  2
 lt s   2
s 0 s 0.25s  0.25s  1
2

(iii )offset  R  C   2  2 offset  0

C ( s) 1
(iv ) 
R( s) 0.25s 2  0.25s  1

 2  0.25    0.5
2  0.25    0.25

  0.25  1,
The response equation is

 1  1  0.25 2 1  0.25 2 
C (t )  21  e 0.25t / 0.5
sin t  tan 1 
 1  0.25 2  0.5 0.25 
 


C (t )  2 1  1.033e 0.5t sin1.936t  1.318 
C (0.5)  0.7853

(v)  0.25  1
Therefore, The closed loop response is oscillatory behavior.

Stability:
 A stable system will be defined as one for which the output response is bounded
for all bounded inputs.
 A system exhibiting an unbounded response to a bounded input is unstable. This
definition although somewhat loose is adequate for most of the linear systems and
simple inputs.
 Bounded is an input that always remains between an upper and a lower limit.
Example 1:
In terms of fig, a control system has the transfer functions
0.5s  1
G1  10 ( PI controller)
s
1
G2  ( stirred tank)
2s  1
H  1(measuring element without lag )

We have suggested a physical element by the components placed in

parentheses. Find the characteristic equation and its rots and determine the system is
stable.
Step 1: 10(0.5s  1)
Open loop transfer function G  G1G2 H 
s(2s  1)
Step 2:
The characteristic equation, 1 + G = 0
10(0.5s  1)
1 G  0  1 0
s(2s  1)
s 2  3s  5  0
3  18
Solving by the quadratic formula s   
2 2
Since the real part of S1 & S2 is negative (-3/2), the system is stable.
Example:1
Given the characteristic equation s  3s  5s  4s  2  0
4 3 2

Determine the stability by the Routh criterion.

Solution:
Since all the coefficients are positive, the system may be stable. To test this,
form the Routh array
Row

15  4 11 60
b1  b2  2
3  3 3
44 26
6 *2  0
3 26 d1  11
c1   26 2
11 11
3 11
Since there is no change in sign in the first column, there are no roots having positive real
parts, and the system is stable.
Example: 2
1 1
Using  1  1, 2  and  3 
2 3
determine the values for KC for which the control system is stable. (b) For the value of
KC for which the system is on the threshold of instability, determine the roots of the
characteristic equation with the help of theorem 3.
Solution:
The characteristic equation, 1 + G = 0 becomes
KC
1 0        (1)
( s  1)(s 2  1)(s 3  1)

Substituting the time constants and solving algebraically,

s 3  6s 2  11s  6(1  K C )  0 - - - - - - - -(2)
Routh array table:
Row
 Since the proportional sensitivity of the controller K C
is a positive quantity, we see that the fourth entry in the first column, 6(1+ KC) is
positive. According to Theorem 1, all the elements of the first column must be positive
for stability. Hence
10  KC  0 and KC  10
It is concluded that the system will be stable only if K C  10
At , KC  10
the system is on the verge of instability and the element in the nth row of the array is
zero. According to Theorem 3, the location of the imaginary roots is obtained by
solving.
Cs  D  0
2

Where C and D are the elements in the (n-1)st row. For this problem with KC  10
we obtain,
6s 2  66  0  6s 2  66  s  i 11

Therefore, two of the roots on the imaginary axis are located at 11 and  11
The third root can be found by expressing eqn (2) in factored form

(s  s1 )(s  s2 )(s  s3 )  0 - - - - - - - - - - - (3)

Where s1, s2 & s3 are the roots. Introducing the two imaginary roots s1  i 11 and s2  i 11
into eqn (3) and multiplying out the terms give
 s  i   
11 s  i 11 s  s3   0  s 2  11 ( s  s3 )  0 
s 3  s 2 s3  11s  11s3  0 - - - - - - - -(4)
Comparing eqn (1) & (4), we see that S3 = - 6. The roots of the characteristic
equation are therefore

s  i
1 11 , s2  i 11 and s3  6 