You are on page 1of 124

# Control Systems

10ES43

## 10ES43 CONTROL SYSTEMS (Common to EC/TC/EE/IT/BM/ML)

Subject Code:10ES43
No. of Lecture Hrs./Week : 04
Total No. of Lecture Hrs.:52

IA Marks: 25
Exam Hours : 03
Exam Marks : 100

PART A
UNIT 1:
Modeling of Systems: Introduction to Control Systems, Types of control systems, Effect of
feedback systems, Differential equations of physical systems Mechanical systemsFriction, Translational systems (Mechanical accelerometer, Levered systems excluded),
Rotational systems, Gear trains. Electrical systems, Analogous systems.
6 Hours
UNIT 2:
Block diagrams and signal flow graphs: Transfer functions, Block diagrams, Signal Flow
graphs (Statevariable formulation excluded).
7 Hours
UNIT 3:
Time Response of feed back 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.
7Hours

UNIT 4:
Stability analysis: Concepts of stability, Necessary conditions for Stability, Routh-Hurwitz
stability criterion, Relative stability analysis; Special cases of RH criterion.
6 Hours
PART B
UNIT 5:
RootLocus Techniques: Introduction, basic properties of root loci, Construction of root
loci.
6 Hours
UNIT 6:
Stability analysis in frequency domain: Introduction, Mathematical preliminaries, Nyquist
Stability criterion, (Inverse polar plots excluded), Assessment of relative stability using
Nyquist criterion, (Systems with transportation lag excluded).
7Hours
Department of EEE, SJBIT

Page 1

Control Systems

10ES43

UNIT 7:
Frequency domain analysis: Correlation between time and frequency response, Bode plots,
All pass and minimum phase systems, Experimental determination of transfer functions,
Assessment of relative stability using Bode Plots.
7 Hours
UNIT 8:
Introduction to State variable analysis: Concepts of state, state variable and state models for
electrical systems, Solution of state equations.
6 Hours
TEXT BOOK :
1. Control Systems Engineering, I. J. Nagarath and M.Gopal, New Age International (P)
Limited, 4 Edition 2005
2 Modern Control Engineering, K. Ogata, PHI, 5th Edition, 2010.

## Department of EEE, SJBIT

Page 2

Control Systems

10ES43

Sl .no

Contents
UNIT 1:Modeling of Systems:
Introduction to Control Systems

Page number

## Differential equations of physical systems Mechanical

systems- Friction

5 to 12

## Translational systems (Mechanical accelerometer,

Levered systems excluded)

## Rotational systems, Gear trains. Electrical systems,

Analogous systems.

## UNIT 2: Block diagrams and signal flow graphs

Transfer functions
2

Block diagram

13 to 23

## Signal Flow graphs (Statevariable formulation excluded).

UNIT 3: Time Response of feed back control systems:
Standard test signals

24 to 46

## Time response specifications of second order systems,

state errors and error constants.
UNIT 4: Stability analysis:
Concepts of stability

47 to 65

Page 3

Control Systems

10ES43

Introduction
5

66 to 84

## Construction of root loci.

UNIT 6: Stability analysis in frequency domain:
Introduction

Mathematical preliminaries

85 to 102

## UNIT 7: Frequency domain analysis:

Correlation between time and frequency response

Bode plots

103 to 118

## UNIT 8: Introduction to State variable analysis:

Concepts of state
8

119 to 127

## Department of EEE, SJBIT

Page 4

Control Systems

10ES43

UNIT-1
Modeling of Systems
Introduction to control systems
A system is an arrangement of or a combination of different physical components connected
or related in such a manner so as to form an entire unit to attain a certain objective
Control system is an arrangement of different physical elements connected in such a
manner so as to regulate, director command itself to achieve a certain objective
Requirements of good control system are accuracy, sensitivity, noise, stability, bandwidth,
speed, oscillations

## Types of control systems

A system in which the control action is totally independent of the output of the system is
called as open loop system

Example: Automatic hand driver, automatic washing machine, bread toaster, electric lift,
traffic signals, coffee server, theatre lamp etc.

A system in which the control action is somehow dependent on the output is called as
closed loop system
The elements of closed loop system are command, reference input, error detector, control
element controlled system and feedback element

Page 5

Control Systems

10ES43

## Elements of closed loop system are:

1. Command : The command is the externally produced input and independent of the
feedback control system.
2. Reference Input Element: It is used to produce the standard signals proportional to the
command.
3. Error Detector : The error detector receives the measured signal and compare it with
reference input. The difference of two signals produces error signal.
4. Control Element : This regulates the output according to the signal obtained from error
detector.
5. Controlled System : This represents what we are controlling by feedback loop.
6. Feedback Element : This element feedback the output to the error detector for
comparison with the reference input.
Example: Automatic electric iron, servo voltage stabilizer, sun-seeker solar system, water
level controller, human perspiration system.

## Feedback system is that in which part of output is feeded back to input

In feedback system corrective action starts only after the output has been affected
1. Accuracy is very high as any error arising is corrected.
2. It senses changes -in output due to environmental or parametric change, internal
disturbance etc. and corrects the same.
3. Reduce effect of non-linearities.
4. High bandwidth.
5. Facilitates automation.
1. Complicated in design and maintenance costlier.
2. System may become unstable.
Department of EEE, SJBIT

Page 6

Control Systems

10ES43

## Advantages of open loop system:

1. They are simple in construction and design.
2. They are economic.
3. Easy for maintenance.
4. Not much problem of stability.
5. Convenient to use when output is difficult to measure.
1. Inaccurate and unreliable because accuracy is dependent on accuracy of calibration.
2. Inaccurate results are obtained with parameter variations, internal disturbances.
3. To maintain quality and accuracy, recalibration of controller is necessary from time to
time.

Feed forward system is that in which the corrective action is initiated without waiting for
the effect of disturbance to show up in the output
System having multiple inputs and multiple outputs is known as multiple output (MIMO)
control system
A servomechanism is a power amplifying feedback control system in which the controlled
variable is mechanical position or its time derivative such as velocity, acceleration
A regulator or regulating control system is a feedback control system in which the
reference input remains constant for long periods/entire intervals of operation
An adaptive control system is one that continuously and automatically measures the
dynamic characteristics of the plant.
The system which follows the principle of superposition and proportionality is called a
linear system.

The motion take place along a straight line is known as translational motion.
Rotational motion of a body is the motion about a fixed axis.
The elements of rotational system are inertia (J), damping coefficient (B) and torsional
stiffness (K).
Mechanical system
Department of EEE, SJBIT

Page 7

Control Systems

10ES43

## For every mechanical system, there is analogous electrical system.

DAlemberts principle states that, the algebric sum of externally applied forces and the
forces resisting motion in any given direction is zero.
For mechanical network, analogous electrical network can be obtained by using f-v and f-i
analogy.
Force-voltage analogy: In this method force is analogous to voltage.
Similarly,

displacement n charge q.

## Mechanical rotational system:

(a) Force-voltage analogy:

## (b) Force-current analogy :

Mechanical coupling:
Laplace transform of signal x (t) is denoted by X (s)

Example problems:
Q 1. Draw the f-1 analogous mechanical system for the electrical circuit of fig. below:

Page 8

Control Systems

10ES43

## Ans. f (t) is analogous to e (t)

f (t) is analogous to R.
f (t) frictional force f is analogous to r.
Spring constant K is analogous to reciprocal
Mass M is analogous to inductance L

## Q. 2. Draw the mathematical model of the following

system and obtain the transfer function.

Page 9

Control Systems

10ES43

The equation is

## Q. 3. Write force-current analogous quantities.

Ans. Force analogous to current

Page 10

Control Systems

10ES43

## Q. 4. Draw the mathematical model of the following

system and obtain the transfer function.

Page 11

Control Systems

10ES43

At node x

## Department of EEE, SJBIT

Page 12

Control Systems

10ES43

UNIT-2
Block diagrams and signal flow graphs
Transfer function
The ratio of laplace transform of the output to the laplace transform of input under the
assumption of zero initial conditions is defined as the transfer function Of system. It is
denoted by G(s).

## Importance : Transfer function is highly important because of following reasons

1. It is used to give the gain of given blocksystem.
2. The system poles/zeros can be found from transfer function.
3. Stability can be determined from characteristic equation.
4. The system differential equation can be obtained from transfer function by replacing.
s-variable with linear differential operator
Significance of Transfer Function

Where

## C(s) is laplace transform of output

R(s) is laplace transform of input.
Transfer function gives the gain of the given block system.
Properties of Transfer Function
1. The transfer function is independent of the inputs to the system.
2. The transfer function of a system is the laplace transform of its impulse response for zero
initial conditions.
3. The system poles/zeros can be found out from transfer function.
4. The transfer function is defined only for linear invariant systems. It is not defined for
non-linear systems.
Limitations of transfer function are listed below
1 Transfer function is valid only for linear time invariant system.
2 It does not take into account the initial conditions initial conditions loose its significance.
3 It does not give any idea about how the present output is progressing.

## Department of EEE, SJBIT

Page 13

Control Systems

10ES43

Poles are the value of s which when substituted in the denominator of a transfer function,
make the transfer function value as infinity
Zeros are the value of s which when substituted in the numerator of transfer function,
make the transfer function value as zero
The characteristic equation can be obtained by equating the denominator polynomial of the
transfer function to zero
. The highest power of s in the characteristic equation is called the order o system
The number of poles at the origin defines the type of system
Block diagram algebra
Block diagram gives a pictorial representation of a control system by way of short handing
the transfer function Signal flow graph further shortens the representation of a control
system by eliminating summing symbol take-off point and block This elimination is
achieved by way of representing the variables by points called nodes
A pictorial representation of the relationship between input and output of a system is termed
as block diagram.
The direction of flow of signal from one block to other is indicated by an arrow.
The point in a block diagram at which signal can be added or subtracted is termed as
summing point.
Gain is the ratio of laplace transform of output to laplace transform of input .
Blocks in series are algebraically combined by multiplication.
The lines drawn between the blocks to indicate the connections between the blocks are
termed as branches.
The point from which a signal is taken for the feedback purpose is called as take-off point.
The order of summing point can be changed if two or more summing points are in series.
Signal Flow Graphs
Department of EEE, SJBIT

Page 14

Control Systems

10ES43

A signal flow grow is a pictorial representation of a system and it displays graphically, the
transmission of signal in system
Node: It is a point from where branches originate or terminate or pass through.
Branch : It is connecting link between two nodes.
Path : The time traced by connecting two or more node is called path.
Loop : It is a path that originates and terminates on same node and along which node other
node is traversed more than once.

Mason s gain formula is used to find the gain of signal flow graph
According to Masons gain formula

where

## i = Number of forward path

= Gain of ith forward path
= System determinant

= 1 (sum of all individual loops) + (sum of all gain products of two nontouching loops) - (sum of all gain product of three non-touching loops) +.
The gain associated with each branch is called branch transmittance
The independent and dependent variable of a control system are represented by small
circles as nodes.
The relationship between nodes is represented by drawing a line between two nodes Such
lns are called branches.
Example problems:

2010

Dec/Jan

Page 15

Control Systems

10ES43

## Step I Obtain total number of forward paths

There is only one forward path

## Step II. Total number of single loop

There are two loops. Thus

## As there is one forward path which touch all the loops

Step IV. Obtain transfer function

## Q. 2. The transfer function of a system is

Calculate the phase shift at

Ans.

## Hence here is phase shift of zero corresponding to

Department of EEE, SJBIT

Page 16

Control Systems

10ES43

Ans.

Ans.

The term

Page 17

Control Systems

## Q. 5. Represent the following set of equations by a signal

2011
flow graph and determine the overall gain relating

10ES43

June/July
.

Page 18

Control Systems

10ES43

## Step IV. Number of three non-touching loops

--no
Step V. Find the value of

## Applying Masons gain formula

Overall transform function

## Q. 6. Simplify the block diagram in fig and obtain the transfer

2010
function relating C(s) and R(s).

June/July

Ans.

Page 19

Control Systems

10ES43

## There are no loops.

Q. 21. From the block diagram shown in the figure below draw the corresponding
signal flow graph and evaluate close loop transfer function relating
Dec/Jan 2006
the output and input.

## By Masons gain formula:

Individual loop,
Non-touching loops = 0

Page 20

Control Systems

10ES43

## Q. 7. The transfer function of a system is given by

Determine the state model in canonical form using parallel decomposition method.
Ans.

Page 21

Control Systems

10ES43

Ans.

Page 22

Control Systems

## Department of EEE, SJBIT

10ES43

Page 23

Control Systems

10ES43

UNIT 3
Time Response of feed back control systems

## Time-domain Analysis of Control Systems

In time-domain analysis the response of a dynamic system to an input is expressed as a
function of time. It is possible to compute the time response of a system if the nature of
input and the mathematical model of the system are known.
Usually, the input signals to control systems are not known fully ahead of time. In a radar
tracking system, the position and the speed of the target to be tracked may vary in a random
fashion. It is therefore difficult to express the actual input signals mathematically by simple
equations. The characteristics of actual input signals are a sudden shock, a sudden change, a
constant velocity, and constant acceleration. The dynamic behavior of a system is therefore
judged and compared under application of standard test signals an impulse, a step, a
constant velocity, and constant acceleration. Another standard signal of great importance is
a sinusoidal signal.
The time response of any system has two components: transient response and the steadystate response. Transient response is dependent upon the system poles only and not on the
type of input. It is therefore sufficient to analyze the transient response using a step input.
The steady-state response depends on system dynamics and the input quantity. It is then
examined using different test signals by final value theorem.
Standard test signals
a) Step signal:

r (t ) Au(t ).

b) Ramp signal:

r (t ) At; t 0.

c) Parabolic signal: r (t ) At 2 / 2; t 0.
d) Impulse signal: r (t ) (t ).

## Time-response of first-order systems

Let us consider the armature-controlled dc motor driving a load, such as a video tape. The
objective is to drive the tape at constant speed. Note that it is an open-loop system.

## Department of EEE, SJBIT

Page 24

Control Systems

G( s)

10ES43

kk
kk
ak1km
W ( s)
a ak k
1 m ; If r (t ) au(t ) , W ( s) 1 m 1 m
R( s ) m s 1
ms 1 s
s
s 1/ m

t

wr
the motor
k1km

## will eventually reach the desired speed.

We are interested not only in final speed, but also in the speed of response. Here, m is the
time constant of motor which is responsible for the speed of response.
The time response is plotted in the Figure in next page. A plot of et / m is shown, from
where it is seen that, for t 5 m the value of et / m is less than 1% of its original value.
Therefore, the speed of the motor will reach and stay within 1% of its final speed at 5 time
constants.

## Department of EEE, SJBIT

Page 25

Control Systems

Here, T ( s)

10ES43

k1km / ( m s 1)
k1km
kk
W ( s)

1 o
R( s) 1 k1k2 km / ( m s 1) m s (1 k1k2 km ) o s 1
where, ko

km
m
and o
.
1 k1k2 km
1 k1k2 km

## If r (t ) a , the response would be, w(t ) ak1ko ak1ko e

t / o

If a is properly chosen, the tape can reach a desired speed. It will reach the desired speed in
5 o seconds. Here, o m . Thus, we can control the speed of response in feedback system.
Although the time-constant is reduced by a factor (1 k1k2 km ) , in the feedback system, the
motor gain constant is also reduced by the same factor. In order to compensate for this loss
of gain, the applied reference voltage must be increased by the same factor.
Ramp response of first-order system
Let, k1k0 1 for simplicity. Then, T ( s)

## Department of EEE, SJBIT

1
W ( s)
. Also, let, r (t ) tu(t ) .

o s 1 R( s )

Page 26

Control Systems

10ES43

o2
1
1 o

Then, W ( s) 2
;
s ( o s 1) s 2 s o s 1

## w(t ) tu(t ) o (1 et / o )u(t )

The error signal is, e(t ) r (t ) w(t )
Or, e(t ) o (1 et / o )u(t )

ess (t ) o
Thus, the first-order system will track the unit ramp input with a steady-state error o ,
which is equal to the time-constant of the system.

## Time-response of second-order systems

Consider the antenna position control system. Its transfer function from r to y is,

## Department of EEE, SJBIT

Page 27

Control Systems
T ( s)

10ES43

k1k2 km
k1k2 km / m
n2
Y ( s)

R( s) m s 2 s k1k2 km s 2 1 s k k k /
s 2 2n s n2
1 2 m
m

where, we define,

n2 k1k2 km / m and

2n

## . The constant is called the damping

ratio and n is called the natural frequency. The system above is in fact a standard second
order system.
The transfer function T ( s) has two poles and no zero. Its poles are,
s1 , s2 n jn 1 2 jd .

## Here, is called the damping factor, d is called damped or actual frequency.

The location of poles for different are plotted in Figure below. For 0 , the two poles

jn are purely imaginary. If 0 1 , the two poles are complex conjugate. All possible
cases are described in a table shown below.

## Unit step response of second-order systems

n2
s 2n
1
1
1
2
Suppose, r (t ) u(t ), R( s) ; Y ( s) 2
2
s s 2n s n s s 2n s n2
s

## Department of EEE, SJBIT

Page 28

Control Systems

10ES43

Or,
Natural frequency, n
The natural frequency of a second order system is the
frequency of oscillation of the system without damping.

s 2n
s n n
1
1
Y ( s)

2
2
2
s ( s n ) n (1 ) s ( s n )2 (n 1

Performing
inverse Laplace
transform,

Damping ratio,
The damping ratio is defined as the ratio of the damping
factor, to the natural frequency n .

## y (t ) 1 ent cos(n 1 2 )t ent sin(n 1 2 )t

b
.
T (s) 2
s as b

Suppose,

nt

e
2
) 1 standard equation,
1 2 cos(an 21
n and
)t sin(n 1 2 )t
or, y(twith
Comparing
1 2

b.
2
n

y (t ) 1

or,

d n 1 2 and tan 1

ent
1

sin(d t ) ,where,

1 2 / cos 1

n t
e sin(d t )
d

.(01)
y (t ) 1

or,

## The plot of e t sin(d t ) is shown in Figure.

yss (t ) lim y(t ) 1
t

## Thus, the system has zero steady-state error.

The pole of T ( s) dictates the response,
e t sin(d t ) .
Department of EEE, SJBIT

Page 29

Control Systems

10ES43

## Time response specifications

Control systems are generally designed with damping less than one, i.e., oscillatory step
response. Higher order control systems usually have a pair of complex conjugate poles with
damping less than unity that dominate over the other poles. Therefore the time response of
second- and higher-order control systems to a step input is generally of damped oscillatory
nature as shown in Figure next (next page).
In specifying the transient-response characteristics of a control system to a unit step input,
we usually specify the following:
1. Delay time, t d
2. Rise time, tr
3. Peak time, t p
4. Peak overshoot, M p
5. Settling time, t s
Department of EEE, SJBIT

Page 30

Control Systems

10ES43

1. Delay time, t d : It is the time required for the response to reach 50% of the final value in
first attempt.
2. Rise time, tr : It is the time required for the response to rise from 0 to 100% of the final
value for the underdamped system.
3. Peak time, t p : It is the time required for the response to reach the peak of time response
or the peak overshoot.
4. Settling time, t s : It is the time required for the response to reach and stay within a
specified tolerance band ( 2% or 5%) of its final value.
5. Peak overshoot, M p : It is the normalized difference between the time response peak and
the steady output and is defined as,

%M p

c(t p ) c()
c ( )

100%

6. Steady-state error, ess : It indicates the error between the actual output and desired
output as t tends to infinity.
ess lim[r (t ) c(t )] .
t

## Department of EEE, SJBIT

Page 31

Control Systems

10ES43

Let us now obtain the expressions for the rise time, peak time, peak overshoot, and settling
time for the second order system.
1.

Rise

time,

tr :

Put

t tr , sin(d tr ) 0 sin , tr

; cos1 .
d

2.

Put

Peak

time,

tp :

dy
0 and
dt

y(t ) 1

at

for

t tp ;

solve

n t
e sin(d t ) n e t cos(d t ) .
d
tan(d t p )

1 2
d n 1 2

tan , d t p k

k 0,1, 2,

## Peak overshoot occurs at k = 1. t p / d / n 1 2 .

3. Settling time, t s : For 2% tolerance band,

n t
4
e
0.02 , ts 4T .

d
s

4. Steady-state error, ess : It is found previously that steady-state error for step input is
zero.
Let us now consider ramp input, r (t ) tu(t ) .
Then, ess lim s{R( s) Y ( s)} lim s{
s 0

s 0

n2
1 1

}
s 2 s 2 s 2 2n s n2

2
2
2
n2
1
1
s 2n s n n
ess lim {1 2
} lim
s 0 s
s 0 s
s 2n s n2
s 2 2n s n2

2n 2
.
2
n
n

## Steady-state error and error constants

The steady-state performance of a stable control system is generally judged by its steadystate error to step, ramp and parabolic inputs. For a unity feedback system,

E ( s)

R( s )
sR( s)
, ess lim sE ( s) lim
.
s 0
s 0 1 G ( s )
1 G( s)

## Department of EEE, SJBIT

Page 32

Control Systems

10ES43

It is seen that steady-state error depends upon the input R( s) and the forward transfer
function G( s) . The steady-state errors for different inputs are derived as follows:
1. For unit-step input: r (t ) u (t ), R( s)

s 0

s 0

1
s

1
1
1

## ; k p is called position error constant.

1 G( s) 1 G(0) 1 k p

## ess lim sE ( s) lim

s 0

s 0

1
s2

1
1
1
lim
;
s 1 G( s) s 0 sG( s) kv

kv is

called

velocity

error

constant.
3. For unit-parabolic input: r (t ) t 2 / 2, R( s)

## ess lim sE ( s) lim

s 0

s 0

1
s3

1
1
1
lim 2
; k a is called acceleration error
s

0
s 1 G( s)
s G ( s ) ka
2

const.

## Types of Feedback Control System

The open-loop transfer function of a system can be written as,

G( s)

K ( s z1 )( s z2 )( s z3 )
s n ( s p1 )( s p2 )( s p3 )

## K (Tz1s 1)(Tz 2 s 1)(Tz 3 s 1)

s n (Tp1s 1)(Tp 2 s 1)(Tp 3s 1)

If n = 0, the system is called type-0 system, if n = 1, the system is called type-1 system, if n
= 2, the system is called type-2 system, etc. Steady-state errors for various inputs and
system types are tabulated below.

## Department of EEE, SJBIT

Page 33

Control Systems

10ES43

The error constants for non-unity feedback systems may be obtained by replacing G(s) by
G(s)H(s). Systems of type higher than 2 are not employed due to two reasons:
1. The system is difficult to stabilize.
2. The dynamic errors for such systems tend to be larger than those
types-0, -1 and -2.

## Effect of Adding a Zero to a System

Let a zero at s = -z be added to a second order system. Then we have,

( s z )n2 / z
n2
n2
C (s)
s
2

.
2
2
2
2
2
R( s) s 2n s n s 2n s n z s 2n s n
The multiplication term is adjusted to make the steady-state gain of the system unity. This
gives css = 1 when the input is unit step. Let cz(t) be the response of the system given by
the above equation and c(t) is the response without adding the pole. Manipulation of the
above equation gives,
cz (t ) c(t )

1d
c(t ).
z dt

The effect of added derivative term is to produce a pronounced early peak to the system
response which will be clear from the figure in the next page. Closer the zero to origin, the
more pronounce the peaking phenomenon. Due to this fact, the zeros on the real axis near
the origin are generally avoided in design. However, in a sluggish system the artful
introduction of a zero at the proper position can improve the transient response. We can see
from equation (03) that as z increases, i.e., the zero moves further into the left half of the splane, its effect becomes less pronounced.

Page 34

Control Systems

10ES43

## Design Specifications of Second-order Systems

A control system is generally required to meet three time
response specifications: steady-state accuracy, damping factor
(or peak overshoot, Mp) and settling time ts. Steady-state
accuracy requirement is met by suitable choice of Kp, Kv, or Ka
depending on the type of the system. For most control systems
in the range of 0.7 0.28 (or peak overshoot of 5 40%) is
considered acceptable. For this range of , the closed-loop pole
locations are restricted to the shaded region of the s-plane as
shown in Figure.
For

the

system, n k1k2 km / m ;

antenna
1
2n m

; ess

ramp

position
2

; ts

control

## . Here, k 2 is only the adjustable

parameter. If we increase k 2 , n will increase and thus settling time will decrease. At the
same time, will decrease, this indicates the increase in peak overshoot. Thus by merely
increasing gain, we cannot improve both transient and steady-state error specifications. We
need to add additional components to the system. These are called compensators. It will
allow improvement of both transient and steady-state specifications.

Page 35

Control Systems

10ES43

## The classical three-term PID controller

proportional feedback control can reduce error responses but that it still allows a non-zero
steady-state error for a proportional system. In addition, proportional feedback increases the
speed of response but has a much larger transient overshoot. When the controller includes a
term proportional to the integral of the error, then the steady-state error can be eliminated.
But this comes at the expense of further deterioration in the dynamic response. Addition of
a term proportional to the derivative of the error can damp the dynamic response.
Combined, these three kinds of actions form the classical PID controller, which is widely
used in industry.
This principle mode of action of the PID controller can be explained by the parallel
connection of the P, I and D elements shown in Figure 3.1 From this diagram the transfer
function of the PID controller is

(3.1)

## Figure 3.1: Block diagram of the PID controller

The controller variables are
Gain
integral action time
derivative action time

Page 36

Control Systems

10ES43
(3.2)

## These three variables

, and
are usually tuned within given ranges. Therefore, they
are often called the tuning parameters of the controller. By proper choice of these tuning
parameters a controller can be adapted for a specific plant to obtain a good behaviour of the
controlled system.
If follows from Eq. (3.2) that the time response of the controller output is

(3.3)

## Using this relationship for a step input of

, i.e.
, the step response
of the
PID controller can be easily determined. The result is shown in Figure 3.2a. One has to
observe that the length of the arrow
the impulse.

## of the D action is only a measure of the weight of

Figure 3.2: Step responses (a) of the ideal and (b) of the real PID controller
In the previous considerations it has been assumed that a D behaviour can be realised by the
PID controller. This is an ideal assumption and in reality the ideal D element cannot be
realised . In real PID controllers a lag is included in the D behaviour. Instead of a D element
in the block diagram of Figure 3.1 a

## element with the transfer function

Page 37

Control Systems

10ES43
(3.4)

is introduced. From this the transfer function of the real PID controller or more precisely of
the

controller follows as
(3.5)

and

it follows
(3.6)

## The step response

of the
controller is shown in Figure 3.2b. This response from
gives a large rise, which declines fast to a value close to the P action, and then
migrates into the slower I action. The P, I and D behaviour can be tuned independently. In
commercial controllers the 'D step' at
can often be tuned 5 to 25 times larger than the
'P step'. A strongly weighted D action may cause the actuator to reach its maximum value,
i.e. it reaches its 'limits'.
As special cases of PID controllers one obtains for:
a)
the PI controller with transfer function
(3.7)

b)
the ideal PD controller with the transfer function
Department of EEE, SJBIT

Page 38

Control Systems

10ES43
(3.8)

and the

(3.9)

c)
and

## the P controller with the transfer function

(3.10)

The step responses of these types of controllers are compiled in Figure3.3. A pure I
controller may also be applied and this has the transfer function
(8.11)

## Department of EEE, SJBIT

Page 39

Control Systems

10ES43

In addition, the different cases should be compared with respect to the normalised
maximum overshoot

## The different cases are discussed below:

a)
The P controller shows a relatively high maximum overshoot
settling time

## as well as a steady-state error

, a long

b)
The I controller has a higher maximum overshoot than the P controller due to the
slowly starting I behaviour, but no steady-state error.
c)
The PI controller fuses the properties of the P and I controllers. It shows a
maximum overshoot and settling time similar to the P controller but no steady-state
error.
d)
The real PD controller according to Eq. (3.9) with
has a smaller
maximum overshoot due to the 'faster' D action compared with the controller types
mentioned under a) to c). Also in this case a steady-state error is visible, which is
smaller than in the case of the P controller. This is because the PD controller
generally is tuned to have a larger gain
due to the positive phase shift of the D
action. For the results shown in Figure 3.5 the gain for the P controller is
and for the PD controller

## . The plant has a gain of

e)
The PID controller according to Eq. (3.6) with
fuses the properties of a
PI and PD controller. It shows a smaller maximum overshoot than the PD controller
and has no steady state error due to the I action.
The qualitative concepts of this example are also relevant to other type of plants with
delayed proportional behaviour. This discussion has given some first insights into the static
and dynamic behaviour of control loops.

Page 40

Control Systems

10ES43

## Figure 3.5: Behaviour of the normalised controlled variable

at the input to the plant

## for step disturbance

for different types of

controllers
Example problems:
Q. 1. Consider the system having transfer function

Calculate the settling time for 2% tolerance band, for the unit step response.
Ans.

Comparing with

we get

Page 41

Control Systems

10ES43

setting time

## Q. 2. Calculate the natural frequency of a second order system described by the

differential equation

## Ans. The given equation is

(1)
Now standard equation of a second order system is given as
(2)
so comparing eq (1) and (2), we get

Q.3. For a second order system the roots of its characteristic equation are
underdamped natural frequency of the system will be.

the

Ans.

## Department of EEE, SJBIT

Page 42

Control Systems

10ES43

Q. 4. What will be the response of a first order system with unit step input?

Ans. We have

Page 43

Control Systems

10ES43

## Q.6. Measurements conducts on a servomechanism show that response to

be
when subjected to a unit step input. Determine the undamped
natural frequency and damping ratio of the system.

Ans.

## Unit step input means

Comparing, we get

Natural frequency

## Department of EEE, SJBIT

Page 44

Control Systems

10ES43

Determine the value of K and B such that the closed loop until step response has w = 3
Ans. The characteristics equation of the system is 1 + G (s) = 0

## Department of EEE, SJBIT

Page 45

Control Systems

10ES43

UNIT 4
Stability Analysis
Concepts of stability
Because of its feedback structure a control system can become unstable, e.g. oscillations
with increasing amplitudes in the signals can occur .
A linear time-invariant system is called ( asymptotically) stable, if its weighting function
decays to zero, i.e. if
(4.1)

is valid. If the modulus of the weighting function increases with increasing to infinity, the
system is called unstable.
A special case is a system where the modulus of the weighting function does not exceed a
finite value as
or for which it approaches a finite value. Such systems are called
critically stable. Examples are undamped

S and I elements

This definition shows that stability is a system property for linear systems. If Eq. (4.1) is
valid, then there exists no initial condition and no bounded input signal which drives the
output to infinity. This definition can be directly applied to the stability analysis of linear
systems by determining the value of the weighting function for
. If this value exists,
and if it is zero, the system is stable. However, in most cases the weighting function is not
given in an explicit analytic form and therefore it is costly to determine the final value. The
transfer function

## of a system is often known and as it is the Laplace transform of the

weighting function
, there is an equivalent stability condition for
according to
Eq. (4.1). The analysis of this condition shows that for the stability analysis it is sufficient to
check the poles of the transfer function
characteristic equation

of its

(4.2)

Page 46

Control Systems

10ES43

## Necessary conditions for Stability

The following necessary and sufficient stability conditions can be formulated:
a)
Asymptotic stability
A linear system is only asymptotically stable, if for the roots
equation

of its characteristic

for all
is valid, or in other words, if all poles of its transfer function lie in the left-half
plane.
b)
Instability
A linear system is only unstable, if at least one pole of its transfer function lies in the
right-half plane, or, if at least one multiple pole (multiplicity
imaginary axis of the plane.

) is on the

c)
Critical stability
A linear system is critically stable, if at least one single pole exists on the imaginary
axis, no pole of the transfer function lies in the right-half plane, and in addition no
multiple poles lie on the imaginary axis.
It has been shown above that the stability of linear systems can be assessed by the
distribution of the roots of the characteristic equation in the plane (Figure 5.2). For control
problems there is often no need know these root with high precision. For a stability analysis
it is interesting to know whether all roots of the characteristic equation lie in the left-half
plane or not. Therefore simple criteria are available for easily checking stability, called
stability criteria. These are partly in algebraic, partly in graphical form.

## Department of EEE, SJBIT

Page 47

Control Systems

10ES43

Figure 4.2: Stability of a linear system discussed by the distribution of the roots of the
characteristic equation in the plane

Routh criterion
For given coefficients of the characteristic equation the method of Routh, which is an
alternative to the method of Hurwitz, can be applied,. Here the coefficients
will be arranged in the first two rows of the Routh schema, which contains
rows:

The coefficients
first two rows according to

in the third row are the results from cross multiplication the

## Department of EEE, SJBIT

Page 48

Control Systems

10ES43

Building the cross products one starts with the elements of the first row. The calculation of
these values will be continued until all remaining elements become zero. The calculation
of the values are performed accordingly from the two rows above as follows:

From these new rows further rows will be built in the same way, where for the last two rows
finally

and

A polynomial

a)

Page 49

Control Systems
all coefficients

10ES43
are positive,

b)
all coefficients

Example 1

## The Routh schema is:

For proving instability it is sufficient to build the Routh schema only until negative or zero
value occurs in the first column. In the example given above the schema could have been
stopped at the 5th row.
Another interesting property of the Routh scheme says, that the number of roots with
positive real parts is equal to the number of changes of sign of the values in the first
column.
Example 2
Determine the stability of the system whose characteristics equation given by
a(s) s 6 4s 5 3s 4 2s 3 s 2 4s 4.
The above polynomial satisfies the necessary condition for stability since all the
coefficients are positive and nonzero. Writing the Routh array, we have

## Department of EEE, SJBIT

Page 50

Control Systems

s6
s5

1
4

s4

10ES43

3
2

4.3 1.2

1
4

4
.2 4.0
2 2
5

s2

s1

12
3( ) 4.2
76
5

s0

2
5

5 12
2.0 ( )

4.4 1.0
4

.4 4.4
12 2

5
5

15

4
5

s3

4.1 4.1

4
0

2.4 .0
2
4
2

76

.4 0
4 15
76

15

We conclude that the system has roots in the right half plane, since the elements of the
first column are not all positive. In fact there are two roots in the right half plane, since
there are two sign changes. In other words two closed loop poles of the system lie in the
right half plane and hence the system is unstable.

Page 51

Control Systems

10ES43

## Example 3. Determine the stability of the following polynomial.

a(s) s 5 5s 4 11s 3 23s 2 28s 12.
Writing the Routh array, we have
s5
1
11
s4

28

23

12

6.4

25.6

12

s3
s

s1

Since the entire row is zero, we construct an auxiliary equation by taking the coefficients
of the previous row, i.e.,
a1 (s) 3s 2 12.
Differentiating the above equation with respect to s, we get
da1 (s)

6s.

ds

(1)

Page 52

Control System

10ES43

## So the Routh array is continued by taking the coefficients of equation (1).

s1
6
0
0
s
12
Since there are no sign changes in the first column of the Routh array, there are no roots
in the right half plane. However, since one entire row in the Routh array was zero, there
are roots in the imaginary axis. The roots in the imaginary axis can be obtained by
solving the auxiliary equation. Therefore,
3s 2 12 0,
s 2 4 0,
s j2
Example 4. Consider the system shown below. The stability properties of the system are a
function of the proportional feedback gain k. Determine the range of k over which the
system is asymptotically stable.

s 1

s(s 1)(s 6)

## The characteristics equation for the system is given by

1 k

s 1
s(s 1)(s 6)

0,

s 3 5s 2 (k 6)s k 0.
Department of EEE, SJBIT

Page 53

Control System

10ES43

## Therefore the corresponding Routh array is

s3
s2
s1

1
5

k-6
k

(4k 30)
5

s0

For the system to be stable, it is necessary that all the elements in the first column of the
Routh array must be positive. Therefore,
4k 30

5
k 7.5

and
and

k 0,
k 0,

k 7.5

Page 54

Control System

10ES43

.

7 12 1
x
x u,
0 0
1
y 1 2x.

## Department of EEE, SJBIT

Page 55

Control System

10ES43

For the system to be stable, the poles of the system should lie in the left half plane. In other
words all the real poles should be negative or the real parts of complex poles must be
negative. The poles of the system are nothing but the eigenvalues of the A matrix of the
system. The MATLAB code is shown below
a = [-7 -12; 1 0];
[v,d] = eig(a)

The result is
v =
-0.9701
0.2425

0.9487
-0.3162

d =
-4
0

0
-3

The diagonal elements of the matrix d are eigenvalues of the system and columns of the
matrix v represent the corresponding eigenvectors.
Example problem:
Q. 1. Calculate the value of k for which the unity feedback
system

Ans.

Routh array

## To cross imaginary axis

Department of EEE, SJBIT

Page 56

Control System

10ES43

## so at K = 48, it will cross imaginary axis.

Q. 2. A closed loop system is shown in the following fig. Find out the largest possible
value of

Ans. As we know

Now

## or characteristic equation becomes

Now the Routh array can be written as

So,

## The largest possible value for the system to be, stable.

Q. 3. By means of Routh criterion, determine the stability of the system represented by
the following equation
Ans. This equations states

Page 57

Control System

10ES43

## There is one row that becomes zero.

So it is being replaced by

## equation can be written as:

Imaginary roots make system unstable and response is continuous oscillatory. Therefore
system is unstable.
Q. 4. Consider the closed loop feedback system shown in the figure below. Determine
the range of K for which the system is stable.

Ans.

Page 58

Control System

10ES43

## This is the range of K for the system to be stable.

Q. 5. Discuss the Routh Hurwitz criteria for determining the stability of a control
system and calculate the range of K for stable operation of following characteristic
equation

Ans. Routh-Hurwitz criterion helps in determining relative stability of a control system. From
the characteristic equation of control system Rouths array is constructed. In case there is no
Given characteristic equation is

Page 59

Control System

10ES43

## Q. 6. Discuss the Routh-Hurwitz criteria for determining the stability of a control

system and calculate the range of K for stable operation of following characteristic
equation.
Ans. Routh-Hurwitz criterion helps in determining relative stability of a control system.
From the characteristic equation of control system Rouths array is constructed. In case there
is no change of sign in first column and the system is stable.
Given characteristic equation is

## Q. 7. Determine the stability of system having characteristic equation:

using Routh Hurwitz criterion.

Page 60

Control System

10ES43

## Since row becomes zero, we formula auxiliary equation

Differentiate auxiliary equation

Since the last element is negative, there is a sign change in first column, the system is
unstable.

## Ans. Given characteristic equation is

Since terms corresponding to S3 are zero, we will form -auxiliary equation corresponding
to

Page 61

Control System

10ES43

Differentiating
2s = 0

## Thus system is unstable.

Q. 9. Find the range of K for stability of
Ans. Given characteristic equation is
Rouths array for given equation is

- For system to be stable all elements in first column should be greater than zero.

Page 62

Control System

10ES43

## so, value of K for stability will be

0 < K < 0.44.

Q. 10. Determine the stability of the system whose characteristic equation is given by

Ans.
Rouths array is

## As their is sign change in first column thus system is unstable.

Q. 11. A unity feedback control system is characterized by open loop transfer function
Using Rouths criterion, calculate the range of values of K for
the system to be stable.
Ans. The characteristic equation is
1 + G(s). H(s),
H(s)=1
1 + G (s)

Page 63

Control System

10ES43

For s K + 1 > 0
K > 1

## Department of EEE, SJBIT

Page 64

Control System

10ES43

UNIT 5
Root-locus techniques
Introduction
A designer can determine whether his design for a control system meets the specifications if
he knows the desired time response of the controlled variable. By deriving the differential
equations for the control system and solving them, an accurate solution of the system's
performance can be obtained, but this approach is not feasible for other than simple systems.
It is not easy to determine from this solution just what parameters in the system should be
changed to improve the response. A designer wishes to be able to predict the performance by
an analysis that does not require the actual solution of the differential equations.
The first thing that a designer wants to know about a given system is whether or not it is
stable. This can be determined by examining the roots obtained from the characteristic
equation
(5.1)

of the closed loop. The work involved in determining the roots of this equation can be
avoided by applying the Hurwitz or Routh criterion. Determining in this way whether the
system is stable or unstable does not satisfy the designer, because it does not indicate the
degree of stability of the system, i.e., the amount of overshoot and the settling time of the
controlled variable for a step input. Not only must the system be stable, but the overshoot
must be maintained within prescribed bounds and transients must die out in a sufficiently
short time.
The root-locus method described in this section not only indicates whether a system is stable
or unstable but, for a stable system, also shows the degree of stability. The root locus is a plot
of the roots of the characteristic equation of the closed loop as a function of the gain. This
graphical approach yields a clear indication of the effect of gain adjustment with relatively
small effort.
With this method one determines the closed-loop poles in the plane - these are the roots of
Eq.(5.1) - by using the known distribution of the poles and zeros of the open-loop transfer
function
. If for instance a parameter is varied, the roots of the characteristic equation
will move on certain curves in the plane as shown by the example in Figure5.1. On these
curves lie all

Page 65

Control System

10ES43

for

## Values of are red and underlined.

possible roots of the characteristic equation for all values of the varied parameter from zero to
infinity. These curves are defined as the root-locus plot of the closed loop. Once this plot is
obtained, the roots that best fit the system performance specifications can be selected.
Corresponding to the selected roots there is a required value of the parameter which can be
determined from the plot. When the roots have been selected, the time response can be
obtained. Since the process of finding the root locus by calculating the roots for various
values of a parameter becomes tedious, a simpler method of obtaining the root locus is
desired. The graphical method for determining the root-locus plot is shown in the following.
An open-loop transfer function with poles at the origin of the plane is often described by
(5.2)

where
is the gain of the open loop. In order to represent this transfer function in terms of
the open-loop poles and zeros it is rewritten as

(5.3)

or

Page 66

Control System

10ES43

(5.4)

with

and

is

(5.5)

(5.6)

or
(5.7)

root locus.

, represent the

## From the above it can be concluded that the magnitude of

must always be unity and its
phase angle must be an odd multiple of . Consequently, the following two conditions are
formalised for the root locus for all positive values of

## from zero to infinity:

a)
Magnitude condition:
(5.8)

Page 67

Control System

10ES43

b)
Angle condition
(5.9)
for

## In a similar manner, the conditions for negative values of

(
determined. The magnitude conditions is the same, but the angle must satisfy the
c)
Angle condition

) can be

(5.10)
for

## . All points of the plane that fulfil the

angle condition are the loci of the poles of the closed loop by varying

. The calibration of

the curves by the values of is obtained by the magnitude condition according to Eq. (5.8).
Based upon this interpretation of the conditions the root locus can constructed in a
graphical/numerical way.
Once the open-loop transfer function
has been determined and put into the proper form,
the poles and zeros of this function are plotted in the plane.

The plot of the locus of the closed loop poles as a function of the open loop gain K,
when K is varied from 0 to +00.
When system gain K is varied from 0 to +oo, the locus is called direct root locus.
When system gain K is varied from -oo to 0, the locus is called as inverse root locus.
The root locus is always symmetrical about the real axis i.e. x-axis.
The number of separate branches of the root locus equals either the number of open
loop poles are number of open-loop zeros whichever is greater.
A section of root locus lies on the real axis if the total number of open-loop poles and
zeros to the right of the section is odd.
If the root locus intersects the imaginary axis then the point of intersection are
conjugate. From the open loop complex pole the root locus departs making an angle
with the horizontal line.
The root locus starts from open-loop poles.
The root locus terminates either on open loop zero or infinity.
The number of branches of roots locus are:

N if P>Z
Department of EEE, SJBIT

Page 68

Control System

10ES43

and M if P<Z
where N > No. of poles P
M > No. of zeros Z

## Angle of departure is. tangent to root locus at complex pole

Based on the pole and zero distributions of an open-loop system the stability of the closedloop system can be discussed as a function of one scalar parameter. The root-locus method
shown in this module is a technique that can be used as a tool to design control systems. The
basic ideas and its relevancy to control system design are introduced and illustrated. Ten
general rules for constructing root loci for positive and negative gain are shortly presented
such that they can be easily applied. This is demonstrated by some discussed examples, by a
table with sixteen examples and by a comprehensive design of a closed-loop system of higher
order.

## Department of EEE, SJBIT

Page 69

Control System

10ES43

Example Problems:
Q.1. Consider the example

with

and

and

As

and

## the poles of the closed loop

transfer function are identical with those of the open-loop transfer function
values
a)

. For other

:

Both roots

and

and

b)
:

The roots

and

## which does not depend on

, and the imaginary part Im
The curve has two branches as shown in Figure 6.2.

,
.

Page 70

Control System

10ES43

## Figure 5.2: Root locus of a simple second-order system

At
is the breakaway point of the two branches. Checking the angle
condition the condition

and

## have the angles

and

and the

magnitudes
and
. The triangle (
) in Figure 6.2 yields the angle
condition. Evaluating the magnitude condition according to Eq. (6.8)

The value of

## the gain of the open loop is

is

Page 71

Control System

10ES43

Table 5.1 shows further examples of some 1st- and 2nd-order systems.
Table 5.1: Root loci of 1st- and 2nd-order systems
root locus

root locus

## General rules for constructing root loci

To facilitate the application of the root-locus method for systems of higher order than 2nd,
rules can be established. These rules are based upon the interpretation of the angle condition
and the analysis of the characteristic equation. The rules presented aid in obtaining the root
locus by expediting the manual plotting of the locus. But for automatic plotting using a
computer these rules provide checkpoints to ensure that the solution is correct.
Though the angle and magnitude conditions can also be applied to systems having dead time,
in the following we restrict to the case of the open-loop rational transfer functions according
to Eq. (5.3)

(5.11)

or
(5.12)

Page 72

Control System

10ES43

)

and

or

(5.13)

(5.14)

## and from Eq. (5.9) the angle condition

(5.15)
for

follows. Here
and
denote the angles of the complex values
and
,
respectively. All angles are considered positive, measured in the counterclockwise sense. If
for each point the sum of these angles in the plane is calculated, just those particular points
that fulfil the condition in Eq. (5.15) are points on the root locus. This principle of
constructing a root-locus curve - as shown in Figure 5.3 - is mostly used for automatic rootlocus plotting.

Page 73

Control System

10ES43

## Figure 5.3: Pole-zero diagram for construction of the root locus

In the following the most important rules for the construction of root loci for

are listed:

Rule 1 Symmetry
As all roots are either real or complex conjugate pairs so that the root locus is
symmetrical to the real axis.
Rule 2 Number of branches
The number of branches of the root locus is equal to the number of poles of the
open-loop transfer function.
Rule 3 Locus start and end points
The locus starting points (

## ) are at the open-loop poles and the locus ending

points (
) are at the open-loop zeros.
branches end at infinity. The
number of starting branches from a pole and ending branches at a zero is equal to the
multiplicity of the poles and zeros, respectively. A point at infinity is considered as an
equivalent zero of multiplicity equal to
.
Rule 4 Real axis locus
If the total number of poles and zeros to the right of a point on the real axis is odd,
this point lies on the locus.
Rule 5 Asymptotes
There are

(5.16)

For
Figure 5.4.

Page 74

Control System

10ES43

## Rule 6 Real axis intercept of the asymptotes

The real axis crossing

of the asymptotes is at
(5.17)

## Rule 7 Breakaway and break-in points on the real axis

At least one breakaway or break-in point
exists if a branch of the root locus
is on the real axis between two poles or zeros, respectively. Conditions to find such
real points are based on the fact that they represent multiple real roots. In addition to
the characteristic equation (6.1) for multiple roots the condition
(5.18)

## must be fulfilled, which is equivalent to

(5.19)

for
. If there are no poles or zeros, the corresponding sum is zero.
Rule 8 Complex pole/zero angle of departure/entry
The angle of departure of pairs of poles with multiplicity

is
(6.20)

## and the angle of entry of the pairs of zeros with multiplicity

Department of EEE, SJBIT

Page 75

Control System

10ES43

(5.21)

## Rule 9 Root-locus calibration

The labels of the values of

## can be determined by using

(5.22)

For
the denominator is equal to one.
Rule 10 Asymptotic stability
The closed loop system is asymptotically stable for all values of for which the
locus lies in the left-half plane. From the imaginary-axis crossing points the critical
values

can be determined.

## . According to the angle condition of

Eq. (5.10) for negative values of some rules have to be modified. In the following these
rules are numbered as above but labelled by a *.
Rule 3* Locus start and end points
The locus starting points (

## ) are at the open-loop poles and the locus ending

points (
) are at the open-loop zeros.
branches end at infinity. The
number of starting branches from a pole and ending branches at a zero is equal to the
multiplicity of the poles and zeros, respectively. A point at infinity is considered as an
equivalent zero of multiplicity equal to
.
Rule 4* Real axis locus
If the total number of poles and zeros to the right of a point on the real axis is even
including zero, this point lies on the locus.
Rule 5* Asymptotes
There are

(5.23)

Page 76

Control System

10ES43

## Rule 8* Complex pole/zero angle of departure/entry

The angle of departure of pairs of poles with multiplicity

is
(5.24)

## and the angle of entry of the pairs of zeros with multiplicity

(5.25)

The root-locus method can also be applied for other cases than varying

. This is possible as

long as
can be rewritten such that the angle condition according to Eq. (5.15) and the
rules given above can be applied. This will be demonstrated in the following two examples.
Q.2. Given the closed-loop characteristic equation

rewritten as

## to which the rules can be applied.

Q.3.Given the closed-loop characteristic equation

## Department of EEE, SJBIT

Page 77

Control System
it is required to find the effect of the parameter
equation is rewritten into the desired form

10ES43
on the position of the closed-loop poles. The

Using the rules 1 to 10 one can easily predict the geometrical form of the root locus based on
the distribution of the open-loop poles and zeros. Table 6.2 shows some typical distributions
of open-loop poles and zeros and their root loci.
Table 6.2: Typical distributions of open-loop poles and zeros and the root loci

Page 78

Control System

## Department of EEE, SJBIT

10ES43

Page 79

Control System

10ES43

For the qualitative assessment of the root locus one can use a physical analogy. If all openloop poles are substituted by a negative electrical charge and all zeros by a commensurate
positive one and if a massless negative charged particle is put onto a point of the root locus, a
movement is observed. The path that the particle takes because of the interplay between the
repulsion of the poles and the attraction of the zeros lies just on the root locus. Comparing the
root locus examples 3 and 9 of Table 5.2 the 'repulsive' effect of the additional pole can be
clearly seen.
The systematic application of the rules from section 5.2 for the construction of a root locus is
shown in the following non-trivial example for the open-loop transfer function

(5.26)

one zero (

## and we have the

four poles (
,
,
,
2). First the poles (x) and the
zeros (o) of the open loop are drawn on the plane as shown in Figure 5.5. According to rule
3 these poles are just

Page 80

Control System

10ES43

## Figure 5.5: Root locus of

underlined.
those points of the root locus where

. Values of

## are in red and

. We have

branches that go to infinity and the asymptotes of these three branches are lines
which intercept the real axis according to rule 6. From Eq. (5.17) the crossing is at
(5.27)

(5.28)

## Department of EEE, SJBIT

Page 81

Control System

10ES43

i.e.

The asymptotes are shown in Figure 5.5 as blue lines. Using Rule 4 it can be checked which
points on the real axis are points on the root locus. The points

with

and

belong to the root locus, because to the right of them the number of poles and zeros is
odd. According to rule 7 breakaway and break-in points can only occur pairwise on the real
axis to the left of -2. These points are real solutions of the Eq. (5.19). Here we have
(5.29)

or

roots

and

and

. The real

## The angle of departure

of the root locus from the complex pole at
determined from Figure 5.6 according to Eq. (5.20):

can be
(5.30)

## Department of EEE, SJBIT

Page 82

Control System
Figure 5.6: Calculating the angle of departure

10ES43
of the complex pole

With this specifications the root locus can be sketched. Using rule 9 the value of can be
determined for some selected points. The value at the intersection with the imaginary axis is

## Department of EEE, SJBIT

Page 83

Control System

10ES43

UNIT 6
Frequency Domain Analysis
Bode plot
If

the

absolute

value

and

the

phase

of

the

frequency

response

## , one obtains the

Figure 6.1: Plot of a frequency response: (a) linear, (b) logarithmic presentation (
logarithmic scale) (Bode plot)

on a

amplitude response and the phase response. Both together are the frequency response
characteristics.

and

## This representation is called a Bode diagram or Bode plot. Usually

decibels [dB] By definition this is

## The logarithmic representation of the amplitude response

scale in this diagram and is called the magnitude.

## with a linear scale.

will be specified in

## has consequently a linear

The logarithmic representation has some advantages for series connections of transfer
functions. For complicated frequency responses, e.g. with

## Department of EEE, SJBIT

Page 84

Control System

10ES43

for
, it can be represented as series connections of the frequency responses of simple
elements of the form

and

for

with
for

and

## and the phase characteristic

Department of EEE, SJBIT

Page 85

Control System

10ES43

with
for
and
for
. Thus, the frequency response of a
series connection is obtained by addition of the individual frequency response characteristics.
A further advantage of this logarithmic representation is for the determination of the inverse
of a frequency response, that is for

. Here

and

line) and

and

(0-

and

## respectively, the curve of

and that of
can be approximated by line segments. This
approximation by lines allows the analysis and synthesis of control systems using simple
geometric constructions. They are important concepts for the control engineer.
Improtant points:

The magnitude and phase relationship between sinusoidal input and steady state
output of a system is known as frequency response.
The polar plot of a sinusoidal transfer function G (jw) is plot of the magnitude of G
(jw) versus the phase angle of G (jw) on polar coordinates as co varied from zero to
infinity.
The phase margin is that amount, of additional phase lag at the gain crossover
frequency required to bring the system to the verge of instability.
The gain margin is the reciprocal of the magnitude l G(jw) l at the frequency at which
the phase angle as _1800.
The inverse polar plot at G (jw) is a graph of 1/G (jw) as a function of w.
Bode plot is a graphical representation of the transfer function for determining the
stability of control system.
Bode plot is a combination of two plot - magnitude plot and phase plot.

## Department of EEE, SJBIT

Page 86

Control System

10ES43

The transfer function having no poles and zeros in the right -half s-plane are called
minimum phase transfer function.
System with minimum phase transfer function are called minimum phase systems.
The transfer function having poles and zeros in the right half s-plane are called nonminimum phase transfer functions systems with non-minimum phase transfer
function. are called non-minimum phase system.
In bode plot the relative stability of the system is determined from the gain margin
and phase margin. .
If gain cross frequency is less than phase cross over frequency then gain margin and
phase margin both are positive and system is stable.
If gain cross over frequency is greater than the phase crossover frequency than both
gain margin andphase margin are negative.
It gain cross over frequency is equal to me phase cross over trequency me gain marg
and phase margin are zero and system is marginally stable.
The maximum value of magnitude is known as resonant peak.
The magnitude of resonant peak gives the information about the relative stability of
the system.
The frequency at which magnitude has maximum value is known as resonant
frequency.
Bandwidth is defined a the range of frequencies in which the magnitude of closed
loop does not drop 3 db.

Example Problems:

with

## is rearranged into the form

This system can now be decomposed into an integrator, two PD and two
is

## Department of EEE, SJBIT

elements, that

Page 87

Control System

10ES43

From this simple analysis the Bode plot can be determined by adding the Bode plots of the
elements

to

## . In this figure the variable quantities

Figure: Representation of a dynamic system by two frequency response diagrams: (a) Bode plot, (b)
Nyquist plot

in the terms
for
and are the breakpoint frequencies in the
Bode plot. Also shown in Figure 4.20 is the Nyquist plot of the frequency response. Both
representations of Figure 4.20 basically contain the same information about the system.

Based on the example given above the procedure for constructing a Bode plot of a
given system can be recapitulated:

a)
The given transfer function must be put into the form

## Department of EEE, SJBIT

Page 88

Control System

with
where possible poles of
according to their multiplicity .
b) Then for

10ES43

at

## the asymptotes of the elements will be used to approximate

and

.
c) If necessary corrections of the approximations can be performed.

Compensation technique is used to make the unstable system stable by introducing the
poles and or zeros at suitable place.
In control system design, if the designed specifications do not satisfy the
requirements of the system or leads to expensive and conflicting demands, then it is
required to insert an additiohal component within the structure of feedback system.
Compensators are of three types:
(a) Lag phase compensator(b) Lead phase compensator
Depending on the location the compensation is divided in following types:
(b) Parallel or feedback compensation
Load compensations is provided to damp out oscillations in the system having
mechanical output.
Primary function of the lead compensator is to reshape the frequency response curve
to provide sufficient phase lead angle to offset the excessive phase lag associated with
the components of fixed systems.

Page 89

Control System

10ES43

## The primary function of lag compensator is to provide attenuation in the high

frequency range to give a system sufficient phase margin.

## Department of EEE, SJBIT

Page 90

Control System

10ES43

Phase lag network is shown in figure below. The primary function of lag compensator is to
provide attenuation in high frequency range to give a system sufficient phase margin. Various
design steps of phase lag network are

Example Problems:
Q1. The asymptotic magnitude Bode plot of a system is given in the figure below Find
the transfer function of the system analytically It is known that the system is minimum
phase system.

## Department of EEE, SJBIT

Page 91

Control System

10ES43

Q 2. Sketh the Bode Plot for the transfer function given by,
and from Plot find (a) Phase and Gain cross rer frequencies (b) Gain Margin and Phase
Margin. Is this System Stable?

Page 92

Control System

## Department of EEE, SJBIT

10ES43

Page 93

Control System

10ES43

The gain crosses 0db axis at co = 1.24 rad/sec, the gain crossover frequency is co = 1.24
The phase crosses 180 line at co = 0.4 rad/sec, therefore phase crossover frequency
At phase cross over the gain is 20 dB, therefore gain margin is 20 dB.
At gain crossover the phase angle is 2150, the phase margin is 180 + (215) = 35.
As both gain and phase margins are negative, the system is unstable.
Q3. Sketch the bode plot for the transfer function given by

## and from plot find gain margin and phase margin.

Ans.
On 0)-axis mark the point at 23.7 rad/sec. since in denominator (jw) term is having power
one, from 23.7 draw a line of slope 20 db/decade to meet y-axis. This will be the starting
point.
Step 1.
From the starting point to I corner frequency (0.33) the slope of the line is 20 db/decade.
From I corner frequency (0.33) to second corner frequency (1) the slope of the line will be
From II corner frequency to IV corner frequency (2) the slope of the line be 40 + (20) =
20
From III corner frequency to IV corner frequency, the slope of line will be 20 + (20) =
From IV corner frequency (5) to V corner frequency the slope will be 40 (+20) = 20
After V corner frequency, the slope will be (20) (20) = 40 db/decade.
Step 2.
Draw the phase plot.
Step 3.
From graph
Phase margin = +34
Gain margin =infinity

## Department of EEE, SJBIT

Page 94

Control System

10ES43

Q 4. State the advantages of bode plots. Determine the value of K in the transfer
function given below such that
(a) Time gain margin is 20 dB
(b) The phase margin is 30

Page 95

Control System

## Department of EEE, SJBIT

10ES43

Page 96

Control System

10ES43

On the phase plot, we made this poirtt and extend the line tO magnitude plot where it
intersects. For required phase margin shift magnitude plot upwards
Total shift = 19 dB (from graph)
20 log K = 19
K =8.91.1
Q.5.Design a suitable compensator such that the system will have K= 10 and phase
margin = 50.

Page 97

Control System

10ES43

## The bode prot is plotted in fig. The frequency range selected in

From bode plot it is noted that uncompensated system is stable having a phase margin
of 30 and the gain crossover occurs at o = 6.75 rad/sec.
To increase the phase margin to 50 the gain crossover point is to be shifted to a higher
value of frequency and this is possible by introducing a phase lead compensation network.

## compensation network. Allowing an amplification of

compensation network transfer function at

## the magnitude of phase lead

is determined below:

Page 98

Control System

10ES43

Page 99

Control System

## Department of EEE, SJBIT

10ES43

Page 100

Control System

10ES43

UNIT 7
Stability in frequency domain
Nyquist criterion
This graphical method, which was originally developed for the stability analysis of feedback
amplifiers, is especially suitable for different control applications. With this method the
closed-loop stability analysis is based on the locus of the open-loop frequency response
. Since only knowledge of the frequency response
practical approach for the following cases:
a) For many cases
are known.

is necessary, it is a versatile

## b) Frequency responses of the loop elements determined by experiments or

considered directly.

can be

## c) Systems with dead time can be investigated.

d) Using the frequency response characteristic of
not only the stability analysis, but also the
design of stable control systems can be easily performed.

## Nyquist criterion using Nyquist plots

To derive this criterion one starts with the rational transfer function of the open loop

Page 101

Control System

10ES43

## Figure: Poles of the open and closed loop in the

their multiplicity)

## plane (multiple poles are counted according to

To determine
, the locus
can be drawn on the Nyquist diagram and
the phase angle checked. Expediently one moves this curve by 1 to the left in the
plane. Thus for stability analysis of the closed loop the locus
according to Figure 5.5 has to be drawn.

Here

## of the open loop

and

is the continuous change in the angle of the vector from the so called critical point (-1,j0)

for

## . Points where the locus passes through

the point (-1,j0) or where it has points at infinity correspond to the zeros and poles of
on the
imaginary axis, respectively. These discontinuities are not taken into account for the derivation of .
Figure shows an example of a

Page 102

Control System

10ES43

## Figure: Determination of continuous changes in the angle

where two discontinuous changes of the angle occur. Thereby the continuous change of the angle
consists of three parts

## As the closed loop is only asymptotically stable for

the Nyquist criterion follows:

## , then from the general case of

The closed loop is asymptotically stable, if and only if the continuous change in the angle of the
vector from the critical point (-1,j0) to the moving point of the locus

## For the case with a negative gain

the case with a positive
time in the open loop.

Page 103

Control System

10ES43

## Simplified forms of the Nyquist criterion

It follows from that for an open-loop stable system, that is
Therefore the Nyquist criterion can be reformulated as follows:

and

, then

If the open loop is asymptotically stable, then the closed loop is only asymptotically stable, if
the frequency response locus of the open loop does neither revolve around or pass through
the critical point (-1,j0).
Another form of the simplified Nyquist criterion for
'left-hand rule':

with poles at

is the so called

The open loop has only poles in the left-half plane with the exception of a single or double
pole at

(P, I or

behaviour). In this case the closed loop is only stable, if the critical

.

## in the direction of increasing values of

This form of the Nyquist criterion is sufficient for most cases. The part of the locus that is
significant is that closest to the critical point. For very complicated curves one should go back
to the general case. The left-hand rule can be graphically derived from the generalised locus
The orthogonal (
a curve with
hand side of

## )-net is observed and asymptotic stability of the closed loop is given, if

passes through the critical point (-1,j0). Such a curve is always on the left.

## The Nyquist criterion using Bode plots

Because of the simplicity of the graphical construction of the frequency response
characteristics of a given transfer function the application of the Nyquist criterion is often
more simple using Bode plots. The continuous change of the angle
critical point (-1,j0) to the locus of
response of

. From figure

Page 104

Control System

## Figure : Positive (+) and negative (-) intersections of the locus

the left-hand side of the critical point

10ES43

## with the real axis on

it can be seen that this change of the angle is directly related to the count of intersections of
the locus with the real axis on the left-hand side of the critical point between
. The
Nyquist criterion can therefore also represented by the count of these intersections if the gain
of the open loop is positive.
Regarding the intersections of the locus of
with the real axis in the range
,
the transfer from the upper to the lower half plane in the direction of increasing values are
treated as positive intersections while the reverse transfer are negative intersections
(Figure 5.7). The change of the angle is zero if the count of positive intersections
is equal
to the count of negative intersections
. The change of the angle
depends also on the
number of positive and negative intersections and if the open loop does not have poles on the
imaginary axis, the change of the angle is

In the case of an open loop containing an integrator, i.e. a single pole in the origin of the complex
plane (
), the locus starts for
at
change of the angle. For proportional and integral behaviour of the open loop

Page 105

Control System

10ES43

at

## (Figure 5.8), and this intersection would be counted

Figure : Count of the intersections on the left-hand side of the critical point for
open loop
as a negative one if

behaviour of the

## de facto there is for

(and accordingly
) no intersection. This follows from the detailed
investigation of the discontinuous change of the angle, which occurs at
. As only a continuous
change of the angle is taken into account and because of reason of symmetry the start of the locus
at
is counted as a half intersection, positive for
and negative for
analogous to the definition given above For continuous changes of the angle

has

, which is

## plane and possibly a

single (
) or double pole (
) at
. If the locus of
has
positive and
negative intersections with the real axis to the left of the critical point, then the closed loop is
only asymptotically stable, if

is valid. For the special case, that the open loop is stable (
negative intersections must be equal.

## ), the number of positive and

From this it follows that the difference of the number of positive and negative intersections in
the case of
that for

the number

is even, for

the number

## is uneven and therefore in

Page 106

Control System
all cases
valid if

10ES43

is an even number, such that the closed loop is asymptotically stable. This is only
.

The Nyquist criterion can now be transferred directly into the representation using frequency
response characteristics. The magnitude response

## , which corresponds to the locus

, is always positive at the intersections of the locus with the real axis in the range of
. These points of intersection correspond to the crossings of the phase response
with lines

## intersection of the locus, the phase response at the

lines crosses from below to
top and reverse from top to below on a negative intersection as shown in Figure 5.9. In the
following these crossings

## Figure : Frequency response characteristics of

and negative (-) crossings of the phase response

## and definition of positive (+)

with the -180 line

will be defined as positive (+) and negative (-) crossings of the phase response

over the

particular
lines, where
may be valid. If the phase response starts at 180 this point is counted as a half crossing with the corresponding sign. Based on the discussions
above the Nyquist criterion can be formulated in a form suitable for frequency response
characteristics:
The open loop with the transfer function
has poles in the right-half plane, and possibly a
single or double pole at
.
are the number of positive and
of negative crossings of the
phase response
over the
lines in the frequency range where
valid. The closed loop is only asymptotically stable, if
Department of EEE, SJBIT

is

Page 107

Control System

## is valid. For the special case of an open-loop stable system (

10ES43

must be valid.
Table 7.1: Examples of stability analysis using the Nyquist criterion with frequency response
characteristics

## Department of EEE, SJBIT

Page 108

Control System

10ES43

Finally the 'left-hand rule' will be given using Bode diagrams, because this version is for the
most cases sufficient and simple to apply.
The open loop has only poles in the left-half plane with the exception of possibly one single or one
multiple pole at
if

(P, I or

has a phase of

## behaviour). In this case the closed loop is only asymptotically stable,

for the crossover frequency

at

This stability criterion offers the possibility of a practical assessment of the 'quality of
stability' of a control loop. The larger the distance of the locus from the critical point the
Department of EEE, SJBIT

Page 109

Control System

10ES43

farther is the closed loop from the stability margin. As a measure of this distance the terms
gain margin and phase margin are introduced according to Figure below

Bode diagram

and

or

## , respectively, in the (a) Nyquist diagram and (b)

Example Problems:
Q1 The polar plot of the open-loop transter of feedback control system intersects the
real axis at2 Calculate gain margin (in dB) of the system.

Q2. What is the gain margin of a system in decibels if its Nyquist plot cuts the negative
real axis at 0.7?
Ans.
a = 0.7

Q4. Consider a feed lock system with the open-loop transfer function. Given by

## Department of EEE, SJBIT

Page 110

Control System

10ES43

Examine the stability of the closed-loop system. Using Nyquist stability theory
.

Q 5. Draw the Nyquest plot for the open loop transfer function given below:

## Department of EEE, SJBIT

Page 111

Control System

10ES43

Q6. Consider a feed lock system with the open-loop transfer function. Given by

## Examine the stability of the closed-loop system. Using Nyquist

stability theory.

Q7. Sketch the Nyquist plot for the system with the open loop transfer function

Page 112

Control System

10ES43

Page 113

Control System

10ES43

Page 114

Control System

10ES43

## Q 9. How is it possible to make assessment of relative stability using Nyquist criterion?

Construct Nyquist plot for the system whose open loop transfer function is

Ans.

## Nyquist critierion can be used to make assessment of relative stability.

Using the characteristic equation the Nyquist plot is drawn. A feedback system is
sable if and only if, the i.e. contour in the G (s) plane does not encircle the (1, 0)
point when the number of poles of G(s) in the right hand s plane is zero.
If G (a) has P poles in the right hand plane, then the number of anticlockwise
encirciements of the (1, 0) point must be equal to P for a stable system,
N=P0
where N = No of clockwise encirclements about (1, 0) point in C (s) plane
P0 = No of poles G (s) in RHP 0

Page 115

Control System

## Department of EEE, SJBIT

10ES43

Page 116

Control System

10ES43

UNIT 8
Introduction to State variable analysis
State-variale representation of single-input-single-output systems
In the following a short introduction into the representation of systems using state-variable
techniques is given. For this purpose, the example of the
network from Figure is used.
The dynamical behaviour of this network is completely defined for

initial conditions

, if the

and the

input variable

for

and

## can be determined for all

. The variables
and
characterise the 'state' of the network and are therefore
called state variables of the network.
The differential equations describe the dynamical behaviour of this network. Inserting into
one obtains the 2nd-order differential equation according to which completely describes the
system with respect to the input-output behaviour. But one can also use the two original
differential equations and can write them in vector notation so that the 1st-order vector
differential equation

(8.1)
with the initial condition

is obtained. This linear 1st-order vector differential equation describes the connection
between the input variable and the state variables. To complete a state-space system, one
needs an additional equation that describes the dependence of the output variable on the state
variables. In this example, it is the direct relationship
Introducing the state vector

Page 117

Control System

10ES43

and
with the matrix

## and with the scalar variables

and
one obtains the general state-space representation of a linear time-invariant single-inputsingle-output system:
(8.2)

initial condition

(8.3)
The Eq. (8.2) is the state equation, and in the general case it is a linear system of 1st-order
differential equations of

state variables

## , which are combined in the state

vector
. Eq. (8.3) is the output equation, which maps the states and inputs
linearly to the output. This is an algebraic equation, whereas the state equation is a
differential equation.

## State-space representation of multi-input-multi-output systems

The Eqs. (8.2) and (8.3) describe an th-order linear time-invariant single-input-singleoutput system. For linear multi-input-multi-output systems of order with inputs and
outputs these equations become

Page 118

Control System

10ES43
(8.4)

## with the initial condition

(8.5)
where the following notation is used:

state vector

) vector

input vector

) vector

output vector

) vector

system matrix

) matrix

input matrix

) matrix

output matrix

) matrix

feedthrough matrix

) matrix

It goes without saying that the general representation of Eqs. (8.4) and (8.5) also includes the
single-input-single-output case. The matrices
and
have constants elements. If
these elements are time-varying, the matrices of the corresponding time-varying system are
substituted by matrix functions of time, e.g.

Page 119

Control System

10ES43

## The relationship between transfer functions and the state-space

representation
In the following, the Eqs. (8.4) and (8.5) will be transformed into the domain using the
Laplace transform, which will be done analogously to the scalar case . For this, the operator
notation

from section 2.1 is adopted and when applying it to Eq. (8.4), one obtains

or rearranged

## The solution of the state equation in the domain is then given by

(8.6)
with
(8.7)
Similarly, for Eq. (12.5) yields

Substituting

## To obtain the relationship with transfer functions, the initial condition

has to be set to
zero. For a single-input-single-output system according to Eqs. (8.2) and (8.3) the system
output is

Comparing this equation with Eq. (8.3) the transfer function is given by
(8.8)

Page 120

Control System
The matrix
represented by

10ES43

(8.9)

where

## is a matrix with polynomial elements in . From Eq. (8.2) it is obvious that

(8.10)

and
(8.11)
which is the characteristic polynomial of the system. The zeros of this polynomial are the
poles of the transfer function and at the same time eigenvalues of the system matrix . If the
system in the state-space representation is fully controllable and observable (see section 8.6),
then the number of poles are equal to the number of eigenvalues.

## State-space vs transfer function approach

The key advantage of transfer functions is in their compactness, which makes them suitable
for frequency-domain analysis and stability studies. However, the transfer function approach
suffers from neglecting the initial conditions. Not only does state-space representation serve
as an alternative to transfer functions, but also it is not limited to linear and time-invariant
systems and it has the following advantages:
1. Single-input-single-output and multi-input-multi-output systems can be formally
treated equal.
2. The state-space representation is best suited both for the theoretical treatment of
control systems (analytical solutions, optimisation) and for numerical calculations.
3. The determination of the system response in the homogeneous case with the initial
condition
is very simple.
4. This representation gives a better insight into the inner system behaviour. General
system properties, for example, the system controllability or observability can be
defined and determined.

## Department of EEE, SJBIT

Page 121

Control System

10ES43

Initially it sounds paradoxical that the choice of the state variables is not unique. This means
that for one and the same system with the input , the output and state variables, there
exist an infinite number of state-space representations. For each value of time one gets the
state in the -dimensional state space. The values are the cartesian coordinates of the state

where the unit vectors are -dimensional linear independent vectors. Their elements are besides the th element, which has the value of 1 - all zero. For describing the state also other
basis vectors can be used. Candidates are all linear independent and -dimensional vectors
. Therefore, it is always possible, to write the state as
(8.12)

## one can rewrite Eq. (8.12) as

(8.13)

The constant and regular matrix is a so-called transformation matrix. Now, instead of the
state the new state can be used and its behaviour analysed. Describing the system in the
new coordinates the state from Eq. (8.13) is inserted into Eqs. (8.4) and (8.5) and one
obtains
(8.14)

Page 122

Control System

10ES43

(8.15)

## with the initial condition

The benefits of the transformation of systems into different state-space representations are:

Most system properties do not depend on the choice of the state variables. They
remain unchanged after a regular transformation and may be analysed in an
appropriate representation form.
The computational determination and analysis of system properties can be
tremendously simplified if the representation form is specifically selected. In
particular certain canonical forms are of interest.

Example 12.5.1 In order to demonstrate a transformation, the example from Eq. (8.1) is
used with the system parameters
,
and
. The initial condition is assumed
to be zero and therefore omitted. With these values one obtains the state equation as

Page 123

Control System

10ES43

## The new representation

consists of two decoupled differential equations with respect to the state variables

and

The analysis and treatment of a system in such a structured representation form, as shown in
the example above, is doubtless more simple. As the representation form must be specifically
selected depending on the type of analysis or synthesis problem, the different representation
forms, for example, the canonical forms, are not discussed separately and are introduced in
the following sections when they are needed.

Page 124