BENJAMIN
C.
KUO
utomatic
control
THIRD EDITION
Syste
HI OX 2D
'.
PRLNIlUt HALL
Automatic
Control
Systems
Third Edition
BENJAMIN
University
C.
KUO
of Illinois at UrbanaChampaign
PRENTICEHALL,
INC., Englewood
Cliffs,
New Jersey
(,^s
Library of Congress Cataloging
in Publication
Data
Kuo, Benjamin
Automatic control.
2.
Control theory.
Title.
629.8'3
7426544
METROPOLITAN
BOROUGH OF W1GAN
DEPT.
OF
LEISURE
LIBRARIES
*?
:
Ace. No,
10067*2 A
5m^?1^;,
1975 by PrenticeHall, Inc.
Englewood
Cliffs,
o :
New Jersey
No
in
part of this
book
may be reproduced
without permission in
10 9 8 7 6 5 4
OF OF OF OF
CANADA,
JAPAN,
LTD., Toronto
New
Delhi
Tokyo
Contents
Preface
Introduction
1.1
Control Systems
1.2
What
1.3
Feedback and What Are Its Effects ? Types of Feedback Control Systems 11
Is
2.
Mathematical Foundation
2. 7
15
Introduction
15 15 18
2.2
ComplexVariable Concept
Laplace Transform
2.3
2.4 2.5
21
25
2.6
2.7 2.8
26
32
zTransform
39
~\
vi /
Contents
3.
51
Introduction
51
3.2
51
55
Block Diagrams
Signal Flow Graphs
58
64
Summary
Flow Graphs 67
66
10
Examples of the Construction of Signal Flow Graphs 75 General Gain Formula for Signal Flow Graphs
Transfer Functions of DiscreteData
71
3.11
80
3.12
Systems
81
4.
95
95
97 99
4.2
4.3 4.4
4.5
103
4.6
Form 109 Between State Equations and Transfer Functions 117 Characteristic Equation, Eigenvalues, and Eigenvectors
Transformation to PhaseVariable Canonical
115 118
4.10
4.11
4. 4.
Diagonalization of the
123
12
13
Diagram 126 136 Decomposition of Transfer Functions 141 Transformation into Modal Form
Controllability of Linear
Systems
144 152
Relationship
Among
Controllability, Observability,
and
4.
Transfer Functions
156
158
161
4.19 4.20
4.21
165
4.22
4.23
167 Diagram for DiscreteData Systems 171 State Diagrams for Samp/edData Systems State Equations of Linear TimeVarying Systems
State
173
5.
187
Introduction
187
5.2
188
190
5.3 5.4
5.5
5.6
Modeling of Mechanical System Elements 203 Equations of Mechanical Systems ErrorSensing Devices in Control Systems
Tachometers
208
219
Contents
vii
5.7 5.8
5.9
Motors in Control Systems 220 TwoPhase Induction Motor 225 Step Motors 228
TensionControl System 235 EdgeGuide Control System 237 Systems with Transportation Lags 242 SunSeeker System 243
DC
5.10
5.11
5. 5.
12
13
6.
259
Introduction
259
6.2
Response of Control Systems 260 TimeDomain Performance of Control SystemsSteadyState Response 262 TimeDomain Performance of Control Systems Transient Response 271 Transient Response of a SecondOrder System 273 Time Response of a Positional Control System 284
Typical Test Signals for Time
Effects of Derivative Control
295
6.9
6.
on the Time Response of Feedback Control Systems 300 Rate Feedback or Tachometer Feedback Control
Control by State Variable Feedback
302
10
305
7.
Stability of Control
7.1
Systems
316
Introduction
316
Equation,
7.2
7.3
Stability. Characteristic
Transition Matrix
317
319
321
7.4
7.5 7.6 7.7 7.8
Methods of Determining
RouthHurwitz
Criterion
Systems
322
Nyquist Criterion
330
344
the
7.9
and Zeros G(s)H(s) on of the Nyquist Locus 352 Stability of Multiloop Systems 356
Stability of Linear Control
Shape
7.10
7.11
Stability of
8.
375
Introduction
375
8.2
376
380
8.3
8.4
412
8.5
Some
417
8.6 8.7
8.8 8.9
Root Contour MultipleParameter Variation 424 Root Loci of Systems with Pure Time Delay 434 Relationship Between Root Loci and the Polar Plot Root Loci of DiscreteData Control Systems 447
444
viii /
Contents
9.
459
Introduction
9.2
9.3
9.4
9.5 9.6
OpenLoop OpenLoop
Transfer Function
Transfer Function
and
Mp
473
9.7
9.8 9.9
As Related to the Slope of Bode Plot 483 485 Loci in the G(jOi) Plane Constant 489 Constant Phase Loci in the G{jCO)Plane
9.10
9.11
9.
Constant
VersusPhase Plane
ClosedLoop Frequency Response Analysis of Nonunity Feedback Systems 497 Frequency Domain
496
10.
504
Introduction
Classical Design of Control
10.2 10.3
Systems
504 510
557
11.
572
Introduction
572
Analytical Design
1.3
Parameter Optimization
1 1
1.4 1.5
11.6
11.7
Feedback
599 615
APPENDIX A FrequencyDomain
A.
1
Plots
627
633
643
626
A.2 A.3
Bode
MagnitudeVersusPhase Plot
APPENDIX B
645
APPENDIX C
Index
Lagrange's Multiplier
Method
650
653
Preface
The
first edition of this book, published in 1962, was characterized by having chapters on sampleddata and nonlinear control systems. The treatment of the analysis and design of control systems was all classical.
The two major changes in the second edition, published in 1967, were the inclusion of the state variable technique and the integration of the discretedata
systems with the continuous data system. The chapter on nonlinear systems was eliminated in the second edition to the disappointment of some users of that text. At the time of the revision the author felt that a comprehensive treatment on the subject of nonlinear systems could not be made effectively
with
The third edition is still written as an introductory text for a senior course on control systems. Although a great deal has happened in the area of modern
control theory in the past ten years, preparing suitable material for a course on introductory control systems remains a difficult task.
it is difficult
to
developments in modern control theory at situation in control systems has been that many of the practical problems are still being solved in the industry by the classical methods. While some of the techniques in modern control theory are much more powerful and can solve more complex problems, there are often more restrictions when it comes to
modern The problem is teach the topics concerned with new the undergraduate level. The unique
However, it should be recognized that control engineer should have an understanding of the classical as well as the modern control methods. The latter will enhance and broaden one's perspective in solving a practical problem. It is the author's opinion that one should strike a balance in the teaching of control systems theory at the beginning
a
modern
Preface
emphasis and intermediate levels. Therefore in this current edition, equal classical methods and the modern control theory. placed on the A number of introductory books with titles involving modern control
is
have attempted to theory have been published in recent years. Some authors according integrate the classical control with the modern control, but unify and and reviews, most have failed. Although such a goal is highly the critics
to
desirable, if only
to
be a good
solution. It
may
new
theories
remains that
learning
and new techniques are developed for this purpose. The fact of control systems, in some way, may be regarded as a science
to solve one
problemcontrol, in many different ways. These against each other, different ways of solution may be compared and weighed approach used in but it may not be possible to unify all the approaches. The
how
method and the modern approach indepenconsidered as alternadently, and whenever possible, the two approaches are are weighed. Many tives, and the advantages and disadvantages of each
this text is to present the classical
illustrative
Many
many
text
examples are carried out by both methods. for not existing text books on control systems have been criticized
book
One reason for this is, perhaps, that who lack the practical background and
that the experience necessary to provide reallife examples. Another reason is fact that most realthe difficulty in the control systems area is compounded by
life
problems are highly complex, and are rarely suitable as illustrative examples is lost by simplifying at the introductory level. Usually, much of the realism techniques developed in the the problem to fit the nice theorems and design system material. Nevertheless, the majority of the students taking a control
text
must put course at the senior level do not pursue a graduate career, and they new employment. It is extremely their knowledge to immediate use in their an important for these students, as well as those who will continue, to gain
what a real control system is like. Therefore, the author has text. The introduced a number of practical examples in various fields in this the attempt of this text to provide more realhomework problems also reflect
actual feel of
life
problems.
The following
with the
first
features of this
new
two editions
1.
2.
3.
Equal emphasis on classical and modern control theory. Inclusion of sampleddata and nonlinear systems. Practical system examples and homework problems.
in this
book
is
an outgrowth of a
seniorlevel
Illinois at control system course taught by the author at the University of written in a style UrbanaChampaign for many years. Moreover, this book is
adaptable for selfstudy and reference. Chapter 1 presents the basic concept of control systems. The definition of feedback and its effects are covered. Chapter 2 presents mathematical founda
Preface
/ xi
and preliminaries. The subjects included are Laplace transform, ztransform, matrix algebra, and the applications of the transform methods. Transfer function and signal flow graphs are discussed in
tion
Chapter
3.
Chapter 4 intro
duces the state variable approach to dynamical systems. The concepts and definitions of controllability and observability are introduced
at the early stage
transducers and control systems used in practice are illustrated. The treatment cannot be exhaustive as there are numerous types of devices and control systems. Chapter 6 gives the time response considerations of control systems. Both the classical and the modern approach are used. Some simple design considerations in the time domain are pointed out. Chapters 7, 8, and 9 deal with topics on stability, root locus, and frequency response of control systems. In Chapter 10, the design of control systems is discussed, and the
These subjects are later being used for the analysis and design of linear control systems. Chapter 5 discusses the mathematical modeling of physical systems. Here, the emphasis is on electromechanical systems. Typical
approach
subjects
is
basically classical.
Chapter
11 contains
which, in the author's opinion, can be taught at the undergraduate level if time permits. The text does contain more material than can be covered in one
semester.
One of
subjects to cover.
To keep
which were in the original to be left out of the final manuscript! These included the treatment of signal flow graphs and timedomain analysis,
of discretedata systems, the second method of Liapunov's stability method! describing function analysis, state plane analysis, and a few selected topics on implementing optimal control. The author feels that the inclusion of these subjects would add materially to the spirit of the text, but at the cost of a higher price.
to a reasonable length,
The author wishes to express his sincere appreciation to Dean W. L. Everitt (emeritus), Professors E. C. Jordan, O. L. Gaddy, and E. W. Ernst, of the University of Illinois, for their encouragement and interest in the project! The author is grateful to Dr. Andrew Sage of the University of Virginia and
Dr. G. Singh of the University of Illinois for their valuable suggestions. Special thanks also goes to Mrs. Jane Carlton who typed a good portion of the manuscript and gave her invaluable assistance in proofreading.
Benjamin C.
Urbana,
Illinois
Kuo
1
Introduction
1 .1
Control Systems
In recent years, automatic control systems have assumed an increasingly important role in the development and advancement of modern civilization and technology. Domestically, automatic controls in heating and air conditioning systems
regulate the temperature
living. Industrially,
and the humidity of modern homes for comfortable automatic control systems are found in numerous applications, such as quality control of manufactured products, automation, machine tool control, modern space technology and weapon systems, computer systems,
Even such problems as inventory control, and economic systems control, and environmental and hydrological systems control may be approached from the theory of automatic control.
social
The
may be
The
control system.
In more common terms, the controlled variable is the output of the system, and the actuating signal is the input. As a simple example, in the steering control of an automobile, the direction of the two front wheels may be regarded as the
controlled variable
c,
the output.
e.
The position of the steering wheel is the input, The controlled process or system in this case is composed
of the steering mechanisms, including the dynamics of the entire automobile. if the objective is to control the speed of the automobile, then the amount of pressure exerted on the accelerator is the actuating signal, with the speed regarded as the controlled variable.
However,
Chap.
Introduction
Actuating
signal e
Controlled
(Input)
Control system
variable c
(Output)
There are many situations where several variables are to be controlled simulmultivariabk taneously by a number of inputs. Such systems are referred to as
systems.
The word automatic implies that there is a certain amount of sophistication is usually in the control system. By automatic, it generally means that the system
capable of adapting to a variety of operating conditions and is able to respond system has to a class of inputs satisfactorily. However, not any type of control feature. Usually, the automatic feature is achieved by feeding the the automatic
output variable back and comparing it with the command signal. When a system does not have the feedback structure, it is called an openloop system, which is the simplest and most economical type of control system. Unfortunately, openloop control systems lack accuracy and versatility and can be used in none but
the simplest types of applications. Consider, for example, control of the furnace for
home
heating. Let us
assume that the furnace is equipped only with a timing device, which controls the on and off periods of the furnace. To regulate the temperature to the proper level, the human operator must estimate the amount of time required for the furnace to stay on and then set the timer accordingly. When the preset time is up, the furnace is turned off. However, it is quite likely that the house temperature is either above or below the desired value, owing to inaccuracy in
the estimate. Without further deliberation,
control
is
it is
One
in the
fact that
one
may
not
know
The other
no control over the outdoor temperature, which has a important definite bearing on the indoor temperature. This also points to an openloop control system, in that the disadvantage of the performance of an
factor
is
not capable of adapting to variations in environmental conditions or experito external disturbances. In the case of the furnace control, perhaps an control for a certain desired temperature in the house; enced person can provide but if the doors or windows are opened or closed intermittently during the system
is
operating period, the final temperature inside the house will not be accurately
regulated by the openloop control.
An
electric
washing machine
A and estimation of the human machine should have the means of checking the cleanliness of the clothes continuously and turn itself off when the desired degree of cleanliness is reached. Although openloop control systems are of limited use, they form the basic
another typical example of an openloop is entirely determined by the judgment true automatic electric washing operator.
is
Sec. 1.1
Control Systems / 3
elements of the closedloop control systems. In general, the elements of an open12. An input applied to the controller, whose output acts as the actuating signal e; the actuating signal then actuates the controlled process and hopefully will drive the controlled variable c to the desired value.
command
r is
Reference
input r Controller
Actuating
signal e
(Output)
Fig. 12.
What is missing in the openloop control system for more accurate and more adaptable control is a link or feedback from the output to the input of the system. In order to obtain more accurate control, the controlled signal c(t) must be fed back and compared with the reference input, and an actuating signal
proportional to the difference of the output and the input must be sent through the system to correct the error. system with one or more feedback paths like that just described is called a closedloop system. Human beings are probably the most complex and sophisticated feedback control system
in existence.
human
be considered to be a control system with outputs, capable of carrying out highly complex operations.
being
may
many
inputs and
being as a feedback control system, let us consider an object on a desk. As one is reaching for the object, the brain sends out a signal to the arm to perform the task. The eyes serve as a sensing device which feeds back continuously the position of the hand. The distance between the hand and the object is the error, which is eventually brought to zero as the hand reaches the object. This is a typical example of closedloop control. However, if one is told to reach for the object and then is blindfolded, one can only reach toward the object by estimating its exact posithat the objective
is
To
illustrate the
human
to reach for
may be
is
being
is
The example of the reaching of an object by a described by the block diagram shown in Fig. 13.
illustrative
human
As another
Fig. 14
Error
Input
detector
command
Reach
for object
f x
Error
Controller
(brain)
Controlled process
1 Controlled
variable
Position
of hand
Fig. 13.
/ Introduction
Chap.
Rudder
Fig. 14.
Rudder
control system.
shows the block diagram of the rudder control system of a ship. In this case the objective of control is the position of the rudder, and the reference input is applied through the steering wheel. The error between the relative positions of the steering wheel and the rudder is the signal, which actuates the controller
When
the rudder
is finally is
zero. Let us
given a sudden rotation of R units, as shown by the time signal in Fig. l5(a). The position of the rudder as a function of time, depending upon the characteristics of the system, may typically be one of the responses shown
in Fig. l5(b). Because all physical systems
have
electrical
the position of the rudder cannot respond instantaneously to a step input, but will, rather, move gradually toward the final desired position. Often, the response
will oscillate
about the
it is
final position
before settling.
It is
rudder control
0,(0
6e
*t
(b)
*~t
(a)
Fig. 15. (a) Step displacement input of rudder control system, (b) Typical
output responses.
Sec. 1.1
Control Systems / 5
Error sensor
Input
~^
Error
Controller
Controlled process
Output
Feedback
elements
are
The basic elements and the block diagram of a closedloop control system shown in Fig. 16. In general, the configuration of a feedback control system
may not be constrained to that of Fig. 16. In complex systems there may be a multitude of feedback loops and element blocks. Figure l7(a) illustrates the elements of a tension control system of a windup
process.
The unwind
which
is
reel may contain a roll of material such as paper or cable to be sent into a processing unit, such as a cutter or a printer, and then
collects it by winding it onto another roll. The control system in this case is to maintain the tension of the material or web at a certain prescribed tension to avoid such problems as tearing, stretching, or creasing.
web is formed into a halfloop by passing it The roller is attached to a pivot arm, which allows free upanddown motion of the roller. The combination of the roller and the pivot arm is called the dancer. When the system is in operation, the web normally travels at a constant speed. The ideal position of the dancer is horizontal, producing a web tension equal to onehalf of the total weight of the dancer roll. The electric brake on
the
roller.
the
unwind
reel is to generate
all
horizontal position at
times.
During actual operation, because of external disturbances, uncertainties and irregularities of the web material, and the decrease of the effective diameter of the unwind reel, the dancer arm will not remain horizontal unless some scheme is employed to properly sense the dancerarm position and control the
restraining braking torque.
obtain the correction of the dancingarmposition error, an angular used to measure the angular deviation, and a signal in proportion to the error is used to control the braking torque through a controller. Figure l7(b) shows a block diagram that illustrates the interconnections between the
sensor
is
To
Chap.
Introduction
Unwind
reel
(decreasing dia.)
Web
processing
Windup
reel
(increasing dia.)
Drive system
(constant
web
speed)
(Current)
Reference input
~"\ Error
Controller
Electric
Unwind
process
Tension
brake
Dancer
arm
(b)
diagram depicting the and interconnections of a tension control system. basic elements
Fig. 17. (a) Tension control system, (b) Block
1 .2
What
Is
Its
Effects ?
control systems.
We demon
closedloop strated in Section 1.1 that feedback is a major requirement of a control system would not be able to achieve control system. Without feedback, a applications. the accuracy and reliability that are required in most practical standpoint, the definition and the significance However, from a more rigorous feedback are much deeper and more difficult to demonstrate than the few
of
carry examples given in Section 1.1. In reality, the reasons for using feedback than the simple one of comparing the input with the output far more meaning error is merely one of the in order to reduce the error. The reduction of system now show that effects that feedback may bring upon a system. We shall many
Sec.
.2
What
Is
Its
Effects ? /
feedback also has effects on such system performance characteristics as bandwidth, overall gain, impedance, and sensitivity.
stability,
To understand the effects of feedback on a control system, it is essential that we examine this phenomenon with a broad mind. When feedback is deliberately
its existence is easily identified. However, numerous situations wherein a physical system that we normally recognize as an inherently nonfeedback system may turn out to have feedback
there are
when
it is observed in a certain manner. In general we can state that whenever a closed sequence of causeandeffect relation exists among the variables of a system, feedback is said to exist. This viewpoint will inevitably admit feedback
number of systems that ordinarily would be identified as nonfeedback systems. However, with the availability of the feedback and control system theory, this general definition of feedback enables numerous systems,
in a large
with or without physical feedback, to be studied in a systematic way once the existence of feedback in the abovementioned sense is established.
We shall now investigate the effects of feedback on the various aspects of system performance. Without the necessary background and mathematical foundation of linear system theory, at this point we can only rely on simple
system notation for our discussion. Let us consider the simple feedback system configuration shown in Fig. 18, where r is the input signal, c the output signal, e the error, and b the feedback signal. The parameters G and ZTmay be considered as constant gains. By simple algebraic manipulations it is simple to
static
show
is
G M = t = FTW
Using
this basic relationship
(li)
some of the
significant effects
+
r
o
b +
H
o
Fig. 18.
Feedback system.
As seen from Eq. (11), feedback affects the gain G of a nonfeedback system by a factor of 1 + GH. The reference of the feedback in the system of Fig. 18 is negative, since a minus sign is assigned to the feedback signal. The quantity GH may itself include a minus sign, so the general effect of feedback is that it
may
and
H are
Introduction
Chap.
may be greater than 1 in functions of frequency, so the magnitude of 1 one frequency range but less than 1 in another. Therefore, feedback could
increase the gain of the system in one frequency range but decrease
it
+ GH
in another.
whether the system will be able to follow the input command. In a nonrigorous manner, a system is said to be unstable if its output is out of control or increases without bound. To investigate the effect of feedback on stability, we can again refer to the 1, the output of the system is infinite for any expression in Eq. (11). If GH
Stability is a notion that describes
=
finite input.
Therefore,
originally stable to
we may state that feedback can cause a system that is become unstable. Certainly, feedback is a twoedged sword;
when it is improperly used, it can be harmful. It should be pointed out, however, that we are only dealing with the static case here, and, in general GH = 1 is
not the only condition for instability. It can be demonstrated that one of the advantages of incorporating feedback is that it can stabilize an unstable system. Let us assume that the feedback 1. If we introduce another feedsystem in Fig. 18 is unstable because through a negative feedback of F, as shown in Fig. 19, the inputback loop
GH =
G
is
+GH+GF
G
and
( "
H are
1,
GH =
+
r
+
e
+
+
o+
c
+
o
o
H
o
o
o
o
F
Feedback system with two feedback loops.
Fig. 19.
an important
all
physical elements have properties that change with cannot always consider the parameters of a control
Sec
12
What
Is
Its
Effects? / 9
system to be completely stationary over the entire operating life of the system. For instance, the winding resistance of an electric motor changes as the temperature of the motor rises during operation. In general, a good control system
still
able to follow
command responsively. We
shall investigate
what
effect
sensitivity to
parameter variations.
we
consider
G as
The
sensitivity
M to the variation in G
a parameter that
may
is
defined as
io
_ dM/M ~
~dGjG
^ 3 >
where dM denotes the incremental change in due to the incremental change G; dM/M and dG/G denote the percentage change in and G, respectively. The expression of the sensitivity function Sg can be derived by using Eq. (11). We have
in
SM _ dM G _ io
~lGM~l+GH
>
This relation shows that the sensitivity function can be made arbitrarily small by increasing GH, provided that the system remains stable. It is apparent that in an openloop system the gain of the system will respond in a onetoone fashion to the variation in G.
In general, the sensitivity of the system gain of a feedback system to paramdepends on where the parameter is located. The reader may derive the sensitivity of the system in Fig. 18 due to the variation of H.
eter variations
The
effect
and brush or commutator noise in electric motors. of feedback on noise depends greatly on where the noise is intro
situations, feedback
duced into the system; no general conclusions can be made. However, in many can reduce the effect of noise on system performance.
mand
Let us refer to the system shown in Fig. 110, in which r denotes the comsignal and n is the noise signal. In the absence of feedback, 0, the output c is
H=
=GGe+G
x
2n
(15)
is
where e
r.
The
defined as
e
l
_GGe_c ~~
x
G2 n
~n
'
'
To increase the signaltonoise ratio, evidently we should either increase the magnitude of G, or e relative to n. Varying the magnitude of G would have
2
no
effect
whatsoever on the
ratio.
to r
and n acting
10
Introduction
Chap.
\h
+ +
r
+
e2
Gi
G2
__
_.
Fig. 110.
signal.
simultaneously
is
_
is
Gl G 2
T + G,G 2 H
+ +
_
b3
G,G 2 H
(17) K '
ponent
is
with Eq. (15) shows that the noise component in reduced by the factor 1 + Gfi,H, but the signal comalso reduced by the same amount. The signaltonoise ratio is
(17)
_GG
i
2 rj(\
2 n/(l
+ G^G^H + G G H)
1
___
g
1
r_
(1%}
and is the same as that without feedback. In this case feedback is shown to have no direct effect on the output signaltonoise ratio of the system in Fig. 110. However, the application of feedback suggests a possibility of improving the signaltonoise ratio under certain conditions. Let us assume that in the system of Fig. 110, if the magnitude of G is increased to G\ and that of the input r to r', with all other parameters unchanged, the output due to the input signal
t
acting alone
is
at the
same
level as that
when feedback
is
we
let
'1 ^
=
to noise acting alone
(1 ' 9)
becomes
which
is
when G
is
tonoise ratio
now
G 2 nl{\
which
is
G\G 2 H)
^ +
^^>
(111)
(1
G\G 2 H).
In general, feedback also has effects on such performance characteristics
Seo
Types
of
11
become known
as
and frequency response. These one progresses into the ensuing material of this text.
1.3
Feedback control systems may be classified in a number of ways, depending upon the purpose of the classification. For instance, according to the method of analysis and design, feedback control systems are classified as linear and nonlinear,
time varying or time invariant. According to the types of signal found in is often made to continuousdata and discretedata systems,
or modulated and unmodulated systems. Also, with reference to the type of system components, we often come across descriptions such as electromechanical
control systems, hydraulic control systems, pneumatic systems, and biological control systems. Control systems are often classified according to the main purpose of the system. positional control system and a velocity control system
control the output variables according to the way the names imply. In general, there are many other ways of identifying control systems according to some
special features of the system. It
is
common
is
ways of
known
This classification
is
made according to
do not
and
design.
tems are nonlinear to some extent. Linear feedback control systems are idealized models that are fabricated by the analyst purely for the simplicity of analysis and design. When the magnitudes of the signals in a control system are limited to a range in which system components exhibit linear characteristics (i.e., the principle of superposition applies), the system is essentially linear. But when the magnitudes of the signals are extended outside the range of the linear operation, depending upon the severity of the nonlinearity, the system should no longer be
considered linear. For instance, amplifiers used in control systems often exhibit saturation effect when their input signals become large; the magnetic field of a motor usually has saturation properties. Other common nonlinear effects
found in control systems are the backlash or dead play between coupled gear members, nonlinear characteristics in springs, nonlinear frictional force or torque between moving members, and so on. Quite often, nonlinear characteristics
are intentionally introduced in a control system to improve its performance or provide more effective control. For instance, to achieve minimumtime control, an onoff (bangbang or relay) type of controller is used. This type of control is
found in many missile or spacecraft control systems. For instance, in the attitude control of missiles and spacecraft, jets are mounted on the sides of the vehicle to provide reaction torque for attitude control. These jets are often controlled in a fullon or fulloff fashion, so a fixed amount of air is applied from a given
jet for a certain
12
Introduction
Chap.
For
and graphical
tech
niques for design and analysis purposes. However, nonlinear systems are very that may be difficult to treat mathematically, and there are no general methods
When
time during the operation of the system, we have a timeinvariant system. Most physical systems contain elements that drift or vary with time to some extent. the If the variation of parameter is significant during the period of operation, of the unwind system is termed a timevarying system. For instance, the radius
reel
is
of the tension control system in Fig. 17 decreases with time as the material being transferred to the windup reel. Although a timevarying system without
is still
nonlinearity
a linear system,
its
analysis
is
usually
electrical engineering,
ac and dc control systems carry special significances. When one refers to an ac control system it usually means that the signals in the system are modulated by some kind of modulation scheme. On the other hand, when a dc control
system is referred to, it does not mean that all the signals in the system are of dc control the directcurrent type; then there would be no control movement. system simply implies that the signals are unmodulated, but they are still ac by common definition. The schematic diagram of a closedloop dc control system
is
shown
in response to a step
^^^
6*^) "r
Reference input
6,
Error detector
Controlled
variable
diagram of a
typical
Sec. 1.3
13
function input are shown in the figure. Typical components of a dc control tem are potentiometers, dc amplifiers, dc motors, and dc tachometers. The schematic diagram of a typical ac control system is shown
sys
in Fig. 112.
In this case the signals in the system are modulated; that is, the information is transmitted by an ac carrier signal. Notice that the output controlled variable still behaves similar to that of the dc system if the two systems have the same control objective. In this case the modulated signals are demodulated by the lowpass characteristics of the control motor. Typical components of an ac control system are synchros, ac amplifiers, ac motors, gyroscopes, and accelerometers.
control systems are strictly of the ac or the dc type. of ac and dc components, using modulators and demodulators to match the signals at various points of the system.
all
In practice, not
system
Synchro
transmitter
ac
servomotor
Reference input
0.
Fig. 112.
For instance, the error signal in a control supplied only intermittently in the form of pulses, in which case the control system receives no information about the error signal during the periods between two consecutive pulses. Figure 113 illustrates how a typical sampleddata system operates. continuous input signal r(t) is applied to the
system
may be
Chap.
14
Introduction
Input
r(t)
eg)
*c >
Data hold
(filter)
hit)
Controlled process
c(f)
Sampler
Fig. 113.
sampler, and is sampled by a sampling device, the sampler is a sequence of pulses. The sampling rate of the samthe output of the advantages of incorporating pler may or may not be uniform. There are many control system, one of the most easily understood of these being sampling in a among several that sampling provides time sharing of an expensive equipment
system.
The
control channels.
flexibility, Because digital computers provide many advantages in size and Many aircomputer control has become increasingly popular in recent years. discrete borne systems contain digital controllers that can pack several thousand shows the 114 elements in a space no larger than the size of this book. Figure
Digital
coded
input
Digital
missile
computer
, ,
analog converter
Airframe
Analogto
con\ erter
Fig. 114. Digital autopilot system for a guided missile.
2
Mathematical Foundation
2.1
Introduction
The study of control systems relies to a great extent on the use of applied mathematics.
subjects as
For the study of classical control theory, the prerequisites include such complex variable theory, differential equations, Laplace transform,
and ztransform. Modern control theory, on the other hand, requires considerably more intensive mathematical background. In addition to the abovementioned subjects, modern control theory is based on the foundation of matrix
theory, set theory, linear algebra, variational calculus, various types of mathematical programming, and so on.
2.2
ComplexVariable Concept
Complexvariable theory plays an important role in the analysis and design of When studying linear continuousdata systems, it is essential that one understands the concept of complex variable and functions of a complex
control systems.
variable
when
method
is
used.
Complex Variable
complex variable j is considered to have two components: a real component a, and an imaginary component co. Graphically, the real component is represented by an axis in the horizontal direction, and the imaginary
component measured along a vertical axis, in the complex jplane. In other words, a complex variable is always defined by a point in a complex plane that has' a a axis and ayco axis. Figure 21 illustrates the complex jplane, in which any
is
15
Chap. 2
16
Mathematical Foundation
/co
splane
OJ]
i i
Fig. 21.
Complex
jplane.
arbitrary point, s
= su
is
== a,
and
co
a> or
simply
Si
ffj
+y'coi.
The function
G(s)
is
G(j)
= ReG+yImC
(21)
where Re
part of
represents the imaginary denotes the real part of G(s) and Im G by the complex GThus, the function G(s) can also be represented
and whose vertical axis plane whose horizontal axis represents Re G every value of s (every point in the sthe imaginary component of G{s). If for value for G(s) [one corresponding point plane) there is only one corresponding a singlevalued function, and the mapping in the G^plane], G(s) is said to be points in the G(s)plane is (correspondence) from points in the jplane onto there are many functions for described as single valued (Fig. 22). However, complexvariable plane is which the mapping from the function plane to the
xplane
/co
S,
.
measures
ImG G 0)plane
ReG
=0, +/C0,
a,
Gfri)
the G(plane.
Sec 2 2

ComplexVariable Concept
17
G(J)=
,(7TT)
<2
"
2)
it is apparent that for each value of s there is only one unique corresponding value for G(s). However, the reverse is not true; for instance, the point G(s) oo is mapped onto two points, s and j 1, in the jplane.
Analytic Function
A function G(s) of the complex variable s is called an analytic function in a region of the splane if the function and all its derivatives exist in the region.
For instance, the function given in Eq. (22) is analytic at every point in the splane except at the points s and s 1. At these two points the value of the function is infinite. The function G(s) s 2 is analytic at every point
= +
Singularities
theory.
The
order r at s
can be stated as: If a function G(s) is analytic and of s except at s it is said to have a pole of
t
,
lim [0
 s,)
G(s)]
has a finite, nonzero value. In other words, the denominator of G(s) must include r s,) so when s s the function becomes infinite. If r 1, the pole at j s, is called a simple pole. As an example, the function
the factor (s
G(s)
has a pole of order 2 at s
W=
is
l0(s
s(s
IX*
+ 2) + 3)* =
=
n (2
" xi
3>
= 3
Zeros of a Function
The
is
definition of a zero of
analytic at s
=s
t,
it is
a function can be stated as: If the function G(s) said to have a zero of order r at s s l if the limit
!S [(* ~
has a finite, nonzero value. has an rthorder pole at s
zero at s
J '>" ,<7
W]
( 2 4)
Or
s,.
simply, G(s) has a zero of order r at s s, ifl/G(s) For example, the function in Eq. (23) has a simple
2.
under consideration is a rational function of s, that is, a quotient of two polynomials of s, the total number of poles equals the total number of zeros, counting the multipleorder poles and zeros, if the poles and
If the function
18
Mathematical Foundation
r.
Chap. 2
function in Eq. (23) has zeros at infinity and at zero are taken into account. The 2, but is one finite zero at s 0, 1, 3, 3; there four finite poles at s
limGO)
s.
= lim^=0 S
s<*>
(2
"5
four zeros in the entire Therefore, the function has a total of four poles and
plane.
s
2.3
is
one of the mathematical tools used for the solution comparison with the classical method
transform method has the of solving linear differential equations, the Laplace
one operation.
s.
2.
s equation by simple algebraic rules to obtain the solution in the solution is obtained by taking the inverse Laplace domain. The final
transform.
Definition of the Laplace Transform
satisfies
the condition
(26)
r\f(t)e'\dt<oo
J
for
some
finite real a,
is
defined as
(27)
F(s)=
or
\~ f(t)e"dt
m=
The
variable s
is,
[/(')]
(2
g
is
which
is
a complex
variable; that
a +jco. The
is
also
known
to oo
is as the onesided Laplace transform, as the integration This simply means that all information contained in /(f) prior to
evaluated from
t
is
ignored
any serious limitation or considered to be zero. This assumption does not place linear system problems, since on the applications of the Laplace transform to often chosen at the instant in the usual timedomain studies, time reference is 0, when an input is applied at t 0. Furthermore, for a physical system t 0; that is, response start sooner than t the response of the system does not
does not precede excitation. The following examples serve as illustrations on how Eq. (27) for the evaluation of the Laplace transform of a function /(f).
may
be used
Sec. 2.3
Laplace Transform / 19
Example
21
Let/0) be a
of unity for
is
>
for
< 0.
Or,
(29)
= u.(t)
Then
is
F(s)
= [u (t)] =
s
j~ us )e" dt
e~"
is
(210)
valid if
Of course,
ut)e" dt

e
dt
<
co
real part
we simply refer to the Laplace transform of the unit step function as lis, and rarely do we have to be concerned about the region in which the transform integral converges
absolutely.
Example
22
= e",
written
2;
where a
is
a constant.
is
=/: e'e" dt
+a
+a
(211)
The operation of obtaining /(?) from the Laplace transform F(s) is termed the inverse Laplace transformation. The inverse Laplace transform of F(s) is
denoted by
f(t)
ZVis)]
(212)
and
is
/(0
where
2^7
'
me"ds
(213)
c is a real constant that is greater than the real parts of all the singularities of F(s). Equation (213) represents a line integral that is to be evaluated in the jplane. However, for most engineering purposes the inverse Laplace transform operation can be accomplished simply by referring to the Laplace trans
form
table,
The
by the
many
presented in the following in the form of theorems, and no proofs are given.
20
Mathematical Foundation
Chap. 2
Multiplication
by a Constant The Laplace transform of the product of a constant k and a time function f{t) is the constant k multiplied by the Laplace transform of f{t);
is,
that
[kf(t)]
= kF(s)
(214)
2.
where F{s) is the Laplace transform of f{t). Sum and Difference The Laplace transform of the sum {or difference) of two time functions
is
the
sum
tions; that
[fi(0
(21 5)
where
3.
F {s)
t
and
2 {s)
and/2 (r),
respectively.
Differentiation
is
The Laplace transform of the first derivative of a time function f(t) s times the Laplace transform of f(f) minus the limit of f(t) as t
0\; that
is,
approaches
df(ty
dt
= sF(s) =
sF(s)
lim /(/)
(216)
/(0+)
[d'fW] _
L
df
s"F(s)  lim
(0 +
n s F(s)
1
sif(t)
s~
ld l^\
 ^r /(o+)
5"" 2
(217)
4. Integration
first integral
S\
f(j)dr
F{s)
s
(218)
5.
J
Shift in
... rr...r.fir) dx A, J Jo
o
o
dt n
^
is
(219)
Time
[fit
T)us {t
T)]
= e T *F{s)
where u s {t
Sec 2 4

6.
InitialValue
Theorem
is
F(s), then
lim sF(s)
(221)
FinalValue Theorem
If the Laplace transform of fit) is F(s)and ifsF(s) is analytic on the imaginary axis and in the right half of the splane, then
lim f(t)
lim sFis)
(222)
The finalvalue theorem is a very useful relation in the analysis and design of feedback control systems, since it gives the final value of a time function by determining the behavior of its Laplace transform as s tends to zero. However,
if sFis) contains any poles whose real part equivalent to the analytic requirement of sFis) stated in the theorem. The following examples illustrate the care that one must take in applying the finalvalue theorem.
is
not valid
is
is
Example
23
+s+2)
and
in the right half of the .splane, the
Since sFis)
finalvalue
is
analytic
on the imaginary
axis
theorem
may
lim fit)
5
,
,
,
= 1
f2231
Example
24
F& = J2"^p
(224)
which is known to be the Laplace transform of /(f) = sin cot. Since the function sFis) has two poles on the imaginary axis, the finalvalue theorem cannot be applied in this case. In other words, although the finalvalue theorem would yield a value of zero as
the final value of fit), the result
is
erroneous.
2.4
'
In a great majority of the problems in control systems, the evaluation of the inverse Laplace transform does not necessitate the use of the inversion integral of Eq. (213). The inverse Laplace transform operation involving rational functions can be carried out using a Laplace transform table
and
partialfraction
expansion.
When the
function in
s, it
is
a rational
X^ = W)
22
Mathematical Foundation
Chap 2

where P(s) and Q(s) are polynomials of s. It is assumed that the order of Q(s) in s is greater than that of P(s). The polynomial Q(s) may be written
Q(s)
a^"'
+a
n
(226)
where
an are real coefficients. The zeros of Q(s) are either real or in a,, complexconjugate pairs, in simple or multiple order. The methods of partialmultiplefraction expansion will now be given for the cases of simple poles,
. .
X( S)
M
,
.
P
.
(227)
s are considered to be real numbers in the where the poles s s 2 present case. Applying the partialfraction expansion technique, Eq. (227) is
t ,
.
written
X(S) v '
= ^ + 4* + s + s s + i,
2
+ T^V s s
1
<2
28 )
The
coefficient,
(i
1, 2,
n), is
(s
sl
;
for instance,
we
that
is,
K
As an
s\
(S
+ s )?&
(s 2
Pis,
s t )(s 3
Si)...
(s
(229)
Si)
illustrative
(s
l)(,s
2)(s
3)
which
is
The
coefficients
K_ u K_ 2 and K_ 3
,
*.
*2
A" 3
+3 + W*)],i = ( 2~lX3  l) _i ~ 5( = + 2)X(s)} ^ = (1 2) + 3 2) = 7 2X3 5(= [(* + 3)*(j)],, (13) +3 3) 6 3X2
:
[('
[(.s
(233)
._
,=
(234)
7 W) = tti + s + 2  t^ j + 3
2  35 >
Sec. 2.4
23
PartialFraction Expansion
If r of the
s,
is
of
multiplicity
r,
X(s)
written
/>(*)
*(*)
g(5)
(5
A*)
s t Xs
+
:
**)...(*
*,)'(*
(236)
Then
JT(j)
can be expanded as
X(s)
= ^^ J + s
<
I
*.
t
+s +
2
...+
s
AT,,
r) terms +
/
of simple poles
A,
A2
(*
(237)
A,
J,)'
>

which correspond to simple poles, K K n K, 2 may be evaluated by the method described by Eq. (229). The determination of
r coefficients,
,
. .
The n
is
described below.
(238)
Ar
= [(S + *,)'*(*)],__
Ar^Ks + s^Xis)]
(239)
(240)
(241)
Example
25
l)3(j
+ 2)
z
(242)
(237),
X(s)
is
written
*
J
'
+ ^3 l)2T(7+Tp
,
(243)
Then
K =
K 2
As =[(j
"^
[sX(s)] s=0
[(j
=$
i
(244)
+ 2)Y(j)],__ 2 =
1
<wLt(j
{
(245)
1) JT(j)]
,._,=
(246)
ls=i
2).
~(2j
s 2 (y
+ 2) + 2)2
1)
3
=
^)]
2(J
1
(247)
and
Al
=2T^^ +
is +
1
d1
d[ 2(s +
l)
i
1)
,
2 dsls 2 (. (s
2(s
"+"
+ 2) 2
(248)
2y
^ sHs + 2)3
,
1)
F(JT2p.
24
Mathematical Foundation
Chap. 2
The completed
partialfraction expansion
is
X = 27 + 2(7+2) ~ 7TT ~
(s
l) 3
(2 " 49)
handle and are of special interest in controlsystems studies, they deserve sepa
5= tx+jcu
Then
and
s= a jco
,
= = K...J.
K.+j.
(s (s
(250) (251)
Example
26
(2  52)
Let us assume that the values of C and are such that the nonzero poles of X(s) are complex numbers. Then X(s) is expanded as follows
* = T + , + 1% + , f +%
where
(2
" 53
>
a
and
oo
Ccu
(254)
= coVl  C 2
s
(255)
The
coefficients in
K = sX(s)\
Kx+Ja>
, = = (j + ay()A'(j),._, co? _ _ G)
\
(256)
+./
(257)
j(e+i/2)
2co
= tani[f]
Also,
*/,
(258)
(*
+a
+yo))A'(j),._
<
,. yto
(259)
2jco( a /a))
The complete expansion
is
1
2co
to
ejw+*n)
Sec 2 5

25
= =
e ne+nm e {*Mt\
(261)
or
x(t)
=
*
viC
~ 2 ' ~ 6)
<2
" 62
>
where 6
is
2.5
With the aid of the theorems concerning Laplace transform given in Section 2.3 and a table of transforms, linear ordinary differential equations can be solved by the Laplace transform method. The advantages with the Laplace transform method are that, with the aid of a transform table the steps involved are all algebraic, and the homogeneous solution and the particular integral
solution are obtained simultaneously.
Example
27
Consider the
differential
.
equation
,
d$T + 3 ~^ +
where u s (t)
is
x(t)
?dx(t)
2*(f)
... = 5u,(t)
(263)
is
defined as
The
initial
dx(t)/dt
,. 0+
= 2. To solve
of Eq. (263);
we first
on both
sides
we have
s*X(s)
 sx(0+)  *<(0+) +
3sX(s)
 3*(0+) + 2*0) =
and solving
(265)
for X(s),
we
* Y(s)
Equation (266)
is
w=
(
j(j*
~ sl ~ s + 5  ~ ja ~ s + 5 + 3s + 2) ~ s(s + 1)0 + 2)
partialfraction expansion to give
Kt\ (2_66)
expanded by
* *)= 2!j4t+20T2)
Now
taking the inverse Laplace transform of the last equation,
(2 " 67)
we
solution as
x(t)
5e<
is
+ le 1
'
>
and the
(268)
last
The
first
term in the
last
equation
two
terms are the transient solution. Unlike the classical method, which requires separate
26
Mathematical Foundation
Chap. 2
and the
method
one operation.
is
of
theorem
may be
Thus
lim x{t)
(.00
= lim
s0
sX(s)
4"
I
(269)
where we have
left
first
checked and found that the function, sX(s), has poles only in the
Example
differential
)
equation
s (t)
(270)
where u s (t) is the unit step function. The initial values of x(t) and dx{t)jdt are assumed to be zero. Taking the Laplace transform on both sides of Eq. (270) and applying zero initial conditions, we have
s 2 X(s)
34.5sX(s)
1000X(s)
(271)
we
2
obtain
(2 " 72)
The
fore,
X s) = s(s +
(
1000
17.25 j26.5)(s
is
17.25
+ J26S)
g/<e+s/2)
(2 " 73)
One way
(273), giving
X(s)
= + 5^63) U +
31.6
e J(+*/2) 17.25
i
/26.5
+s+
17.25
+726.5J
(2 " 74)
where
tani(:=
J!)

56.9
(275)
1.193e I7
2 ='
sin(26.5f
 6)
(276)
Another approach
is
to
compare Eq.
CO n
= ^1000 = = 0.546
31.6
(277)
and
C
(278)
x(t)
is
2.6
2' 6
it is
The simplifying matrix notation reduce the amount of work required to solve the mathematical equations, but it usually makes the equations much easier to handle and manipulate.
may not
Sec. 2.6
As a motivation to the reason of using matrix notation, the following set of n simultaneous algebraic equations:
n*i
21*1
let
us consider
+ + +
a, 2 *2
22*2
a 2 x 2
+ +
+
+ a x =y + a 2n X = J2
ln
n
l
(279)
a.i*i
...+
ax
=y
=y
(280)
The symbols A, x, and y are defined as matrices, which contain the coeffiand variables of the original equations as their elements. In terms of matrix
which will be discussed later, Eq. (280) can be stated as: The product of the matrices A andx is equal to the matrix y. The three matrices involved here are defined to be
algebra,
21
(281)
a2
x2
X
=
x
>l'
(282)
=
JV
(283)
coefficients
and
variables.
Thus, we can
Definition of a Matrix
A matrix is a collection of elements arranged in a rectangular or square array. Several ways of representing a matrix are as follows
A=
1
10"
.
2
3
A =(
\1
2
:)
10
2
A=
[aj 2i3
In this text
we
28
Mathematical Foundation
Chap. 2
It is
Determinant
An
with n rows
An
value, although
square),
m) has a
deter
Has a
Some important
Matrix elements.
When
a matrix
is
written
a, 2
fl
an
#21
a 13 a 23 a 31
(284)
22
a 31 au
is
a 32
As a
first
row and thejth column of the matrix. and the column last.
of Order of a matrix. The order of a matrix refers to the total number example, the matrix in Eq. (284) has three rows and columns of the matrix. For matrix. rows and three columns and, therefore, is called a 3 X 3 (three by three) and m columns is termed " x m" or "n by m." In general, a matrix with n rows
Square matrix.
as columns.
square matrix
is
is
an
m X
1 matrix,
m>
simply Quite often, a column matrix is referred to as a column vector or m rows. The matrix in Eq. (282) is a typical column an mvector if there are matrix that is n X 1, or an nvector.
Row
vector.
matrix.
is,
row matrix
is
column, that
a 1 X n matrix.
one that has one row and more than one row matrix can also be referred to as a row
Diagonal matrix.
all
i
diagonal matrix
is
for
"5
0"
a 22
(285)
a 33 .
matrix with Unity matrix (Identity matrix). A unity matrix is a diagonal unity matrix is often main diagonal (i =j) equal to 1. all the elements on the designated by I or U. An example of a unity matrix is
Tl 1 =
1
0" (286)
1
Sec. 2.6
Null matrix.
for example,
all
equal to zero;
o=
Symmetric matrix.
the condition
"0
(287)
symmetric matrix
is
satisfies
au
for
all
i
<*jt
(288)
andj.
A
its
changed with
symmetric matrix has the property that if its rows are intercolumns, the same matrix is preserved. Two examples of symT6
5
1
r
1
4".
(289)
10
10
1
the
a square matrix
Determinant of a matrix. With each square matrix a determinant having same elements and order as the matrix may be defined. The determinant of
is
designated by
detA
As an
illustrative
=
1
A^
A
(290)
r
3
2
oj
(291)
1
The determinant of A
is
1
1
3
A
=
1
= 5
(292)
Singular matrix.
determinant
is zero.
A square matrix is said to be singular if the value of its On the other hand, if a square matrix has a nonzero deter
called a nonsingular matrix. a matrix is singular, it usually means that not all the rows or not all the columns of the matrix are independent of each other. When the matrix is used to represent a set of algebraic equations, singularity of the matrix means that these equations are not independent of each other. As an illustrative
it is
minant,
When
exam
ple, let us
2x
x,
x,
3x2 + x = + x + x =  2x + 2x =
3
(293)
that in Eq. (293), the third equation is equal to the sum of the first two equations. Therefore, these three equations are not completely independent. In matrix form, these equations may be represented by
Note
AX =
30
Mathematical Foundation
Chap. 2
where
2
3
1
A=
1
1
(294)
2
X=
A3
(295)
and
is
a 3
null matrix.
The determinant of A
1
is
3
1
A= 1
1
43+2
(296)
2
2
(294) is singular. In this case the
A of Eq.
rows of A are
dependent.
Transpose of a matrix. The transpose of a matrix A is defined as the matrix that is obtained by interchanging the corresponding rows and columns in A. matrix which is represented by be an n X Let
A=
Then the transpose of A, denoted by
A'
[%], m
is
(297)
A',
given by
[a] m>
= transpose of A =
is
(298)
n.
Example
29
As an example of the
r
5.
1
The transpose of A
is
given by
3
A'
2
Ll
1
5.
Skewsymmetric matrix.
that equals
its
skewsymmetric matrix
;
is
a square matrix
negative transpose
that
is,
A=
Some
Operations of a Matrix Transpose
1.
A'
(299)
(A')'
2.
3.
4.
(2100)
(2101)
(2102)
(2103)
Sec. 2.6
31
A
is
n.
The adjoint
A=
[if
(2104)
where the
ting the
ij
z'th
row and
the jth
by (l) i+J
Example 210
first
let
us consider
A = n
a2
i
fll2
#22
The determinant of A
is
A
= an
a%i
(1, 1)
12
o 22
element of A,

The
1,1 cofactor,
is
.
flu!
a xl
A
a 22 the
;
1,2 cofactor
is
and the
2,2 cofactor
is
A is
1,1 cofactor
1,2 cofactor
adj
2,1 cofactor
1,1
2,2 cofactor.
2,1 cofactor"
cofactor
J ,2 cofactor
#22
#21
(2105)
2,2 cofactor.
#12
auj
adjoint matrix, consider
Example
21
aj 2
an
#23 #3 3
(2106)
#22
O a
a 32
Then
1,1
cofactor
2,1 cofactor
3,1 cofactor"
adj A =
1,2 cofactor
1,3
2,2 cofactor
2,3 cofactor
3,2 cofactor
3,3 cofactor_
cofactor
(#22#3 3
(#21#33

(#21#32
~ #23#32) (#12#33 ~ #1 3#32) (#12#23 #1 3#22)~ #23#3l) (#U#33 #13#3l) ~ (#1 1#23 #13#2l) 22#3l) (#U32 ~ #123l) (#1 1#22 #12#2l)(2107)
Conjugate matrix. Given a matrix A whose elements are represented by a tJ the conjugate of A, denoted by A, is obtained by replacing the elements of A by their complex conjugates; that is,
,
Chap. 2
32
Mathematical Foundation
A=
conjugate matrix of A
(2108)
=
where a tl
[<y
.
complex conjugate of a u
2.7
Matrix Algebra
When
define matrix algebra in carrying out matrix operations it is necessary to division, and other necessary the form of addition, subtraction, multiplication, operations are to point out at this stage that matrix operations. It is important operations for scalar quantities. defined independently of the algebraic
Equality of Matrices
Two
matrices
and
following conditions:
1.
2.
is,
=
a 12
b iJ
/and j
For example,
A=
implies that a n
a ti
_2
1
bn
Pl\
bn
022_
(2109)
fl 22.
b u a 12
,
b 12 a 21
,
b 2U and a 22
Addition of Matrices
Two
order.
matrices
Then
+
=
[b,jl,,
=C=
[c,7 L,,
(2110)
where
ClJ
au
+
is
b 'J
preserved after addition.
(2111)
for all
and/ The
Example 212
As an
A=
_
1
B = 1
_
1
1_
0^
is
given by
5"
3+0
C=A+B
_
11
~ 3 + 3" = 2 4 + 2 _ 1+0.
(2112)
1
Sec. 2.7
Matrix Algebra /
.,
33
Matrix Subtraction
The rules governing the subtraction of matrices are similar to those of matrix addition. In other words, Eqs. (21 10) and (21 1 1) are true if all the plus signs are replaced by minus signs. Or,
C=AB=
_ l = a ul,m + [b
[,,]
[blJ
m m
(2113)
iJ ] n .
=
where
Cti
C l tjh,m
= a b
for
all /
and /
(2114)
Associate
The
subtraction. Therefore,
(A
B)
+C=A+
(B
C)
(2115)
The commutative law for matrix addition and subtraction states that the following matrix relationship is true:
A+B+C=B+C+A
=A+C+B
Matrix Multiplication
(2 " 116 >
The matrices A and B may be multiplied together to form the product AB they are conformable. This means that the number of columns of A must equal the number of rows of B. In other words, let
if
B=
Then
[*>],,,
and
C = AB = [a] P [b,j\q m = [c J, The matrix C will have the same number of rows
,
(2H7)
as
A and
as B.
important to note that A and B may be conformable for AB but they not be conformable for the product BA, unless in Eq. (2117) n also equals m. This points out an important fact that the commutative law is not generally valid for matrix multiplication. It is also noteworthy that even though A and B are conformable for both AB and BA, usually AB BA. In general,
It is
may
made with
AB = A postmultiplied
by
AB = B premultiplied
by
34
Mathematical Foundation
Chap. 2
Having established the condition for matrix multiplication, let us now turn to the rule of matrix multiplication. When the matrices A and B are conformable to form the matrix C = AB as in Eq. (2117), the yth element of C, c, is given by
Cii
=
. .
Li a.kbkj
,
(2118)
for
1, 2,
n,
andj
1, 2,
m.
Example
213
A=
we notice
Thus,
that these
[fl,
7] 2
B=[6
; ,]
3 ,i
AB
Hn U
On
0n 12
Oil
"An"
fln 13
bn
631(2119)
a 13i
12^21 #22621
tfllAll
a2lAu
+ +
+ 013631 + O2363I
Example 214
r
1
A
J2
B=
o_
are conformable for multiplication.
we
AB
1
and
"i
BA
r
3.
1
AB =
_2
r
o_
o_
"(3X1)
(0)(1)
_(2)(1) _
1
(3)(0) (0)(0)
(2X0)
(2120)
1
2
_2
"1
2_
"3
1"
1
BA
_2
1
11
Oj
_2
0_
"(1X3)
_(2)(3)
1
+ +
1
(0)(0) (1)(0)
+ +
(1X2)
(0)(2)
(1)(1)
(2)(l)
+ +
(0)(1)
(1)(1)
h
1
(1X0)"
(0X0)

(2121)
n 1"
j>
1.
Sec. 2.7
Matrix Algebra
36
Although the commutative law does not hold in general for matrix multiand the distributive laws are valid. For the distributive
law,
we
state that
A(B
if
C)
= AB + AC = A(BC)
(2122)
For the
associative law,
(AB)C
if
(2123)
the product
is
conformable.
Multiplication
by a Scalar k
Multiplying a matrix
by any
scalar
is
element of A by
k. Therefore, if
A=
[a u ] m
kA
(2124)
[ka u ]. m
when we
write
(2125)
ax
it
=y
1
leads to
x=y
or
(2126)
a *y
(2127)
Ax
then
it
(2128)
may be possible
to write
x
where
= A _1 y
A
1
(2129)
exists are:
1.
2.
If
exists, it is
A'
Example 215
Given the matrix
adjA
(2130)
A=
the inverse of
"11
Ozi
12
a 2 2
(2131)
A is
given by
22
a 12
(2132)
A'
adj
A
dllOll
1221
Chap. 2
36
Mathematical Foundation
where for
A to
be nonsingular, A

^ 0, or
1221
=
is
011^22
If
(2133)
obtained by interchanging the two elements on see that for a 2 x of A. signs of the elements on the off diagonal the main diagonal and changing the
we pay
we
2 matrix, adj
is
Example
216
A=
L
the determinant of
10
l l
(2134)
U
=
(2135)
A is
1 1
IA
1
1
"0
1
2
1
Therefore,
A has
is
given by
"2
A"
1
1
1
2
lj
(2136)
Some
1.
AA" =A"A
1 1
1
(2137) (2138)
2.
3.
=A
AB = AC
does not necessarily imply
(2139)
easily construct
C.
AAB = AAC
f
(2140)
IB
which leads to
= IC =C
and are nonsingular, then
(2141)
B
4. If
and
(AB)" 1
=B'A'
(2142)
Rank of a Matrix
maximum number of linearly independent in the largest nonsingular matrix contained is the order of of a matrix are as follows: on the rank
is
the
Sec. 2.7
Matrix Algebra /
37
"0
.0
"3
rank
o.
"0514
1
rank
_3
2,
9
3
"3
rank
1
2,
2
1
rank
The following properties on rank are useful in the determination of the rank of a matrix. Given matrix A,
annxm
= = =
1.
2.
3.
Properties 2 and 3 are useful in the determination of rank; since A'A and AA' are always square, the rank condition can be checked by evaluating the deter
/(x) which
is
=2
We
a,jx,xj
(2143)
can write
this
equation as
(2144)
/(*)
Let
= 2 *,!>,,*,. = % =
a,jXj
yt
(2145)
Then Eq.
(2144)
becomes
/00
2 X&
~y\
(2146)
Now
if
we
define
Xi
x2
x
yi y
=
xn _
=
_j_
/(*)
x'y
(2147)
(2145),
= Ax
[atf ],.
(2148)
A=
(2149)
38
Mathematical Foundation
Chap. 2
Finally, /(x)
becomes
/(x)
x'Ax
(2150)
quadratic form as in a n for i j, given any with a symmetric matrix. In other words, Eq. (2150), we can always replace A given any A, we can always define a symmetric matrix B such that
is
bn
=?ii+,
i*j
(2151)
are often used as performance indices in control conveniences in the systems design, since they usually lead to mathematical
design algorithms.
Definiteness
Positive definite.
An
X n matrix
AI
A

is
all
A =
(2152)
of A, are positive. Equation (2152) is called the characteristic equation eigenvalues of A. roots are referred to as the
Positive semidefinite.
its
and the
The matrix
(n
n)
is
zero.
if all its
Negative
definite.
eigenvalues are
Indefinite.
The matrix A (n X n) is negative semidefinite nonpositive and at least one of the eigenvalues is zero.
n)
is
indefinite if
some of
of a square matrix
is
principal the signs of all the leading principal minors of the matrix. The leading are defined as follows. Given the square matrix n matrix minors of an n X
"an
a 12
...
0i
a
a lt
021
0n
12
^21
<*22
012 22 032
013
a23
"33
a 31
Then
the definiteness of
A is
determined as follows:
A is A are
are
positive (negative).
A is positive semidefinite if A =
 
and
all
nonnegative.
Sec. 2.8
_ __ zJransform / 39
,
,
A
of
is
negative semidefinite
if

A
A are nonnegative.
We may
The quadratic form, x'Ax (A
and
all
is symmetric), is positive definite (positive semidefinite, negative definite, negative semidefinite) if the matrix is positive definite (positive semidefinite, negative definite, negative semidefinite).
2.8
zTransform 1213
The Laplace transform is a powerful tool for the analysis and design of linear timeinvariant control systems with continuous data. However, for linear systems with sampled or discrete data, we may find that the ztransform is more appropriate.
Let us first consider the analysis of a discretedata system which is represented by the block diagram of Fig. 23. One way of describing the discrete nature of the signals is to consider that the input and the output of the system are sequences of numbers. These numbers are spaced T seconds apart. Thus, the input sequence and the output sequence may be represented by r(kT) and c(kT), respectively, k 0, 1, 2, .... To represent these input and output sequences by timedomain expressions, the numbers are represented by impulse functions in such a way that the strengths of the impulses correspond to the values of these numbers at the corresponding time instants. This way, the input sequence is expressed as a train of impulses,
'*(')
%/^T)5it
kT)
(2153)
c(kT)
'('>
system
r^ _X. p>
(
,pK ',?(')
Fig. 23.
Fig.
24.
system
Block diagram of a
iinitepulsewidth sampler.
system.
A sampler is a device
data.
Another type of system that has discontinuous signals is the sampleddata A sampleddata system is characterized by having samplers in the system.
that converts continuous data into
For example,
Fig. 24
some form of sampled shows the block diagram of a typical sampler that
closes for a very short duration of/> seconds once every seconds. This is referred to as a sampler with a uniform sampling period T and a finite sampling duration p. Figure 25 illustrates a set of typical input and output signals of the sampler. With the notation of Figs. 24 and 25, the output of the finitepulseduration sampler is written
'J(0
'(0
[u.{t
*r)
u,(t
kT p)]
(21 54)
where ut)
is
40
Mathematical Foundation
Chap. 2
r(t)
ire)
T 2T
and output
For small p, that is, p <^T, the narrowwidth pulses of Fig. 25 may be approximated by flattopped pulses. In other words, Eq. (2 54) can be written
1
r*(t)
(t
s {t
(2155)
Multiplying both sides of Eq. (21 55) by l/p and taking the limit as p approaches
zero,
we have
Hm r*(r)
p0
Urn f]
P^O
k=o
r{kT)[u p

s (t
kT)
u,{f
 kT  p)]
=
or
lim
t, r{kT)8(t
kT)
r*(0  r\i)
we have made
6(t)
(2156)
lim L [k,(0
J>0
u,(t
~ p)]
(2157)
The significance of Eq. (2156) is that the output of the finitepulsewidth sampler can be approximated by a train of impulses if the pulsewidth approaches zero in the limit. A sampler whose output is a train of impulses with the strength of each impulse equal to the magnitude of the input at the corresponding sampling instant
is called an ideal sampler. Figure 26 shows the block diagram of an ideal sampler connected in cascade with a constant factor p so that the combination is an approximation to the finitepulsewidth sampler of Fig. 24 if p is very small. Figure 27 illustrates the typical input and output signals of an ideal sampler; the arrows are used to represent impulses with the heights representing
r*(t)
*.
Ideal sampler
V
*
(t)
Fig. 26.
pler
ideal
sam
Sec. 2.8
zTransform
41
r*(t)
T IT 3T 4T
HT
and output
signals of
an ideal sampler.
view of these considerations we may now use the ideal sampler to represent the discrete data, r(kT). This points to the fact that the signals of the system in Fig. 23 can essentially be treated as outputs of ideal samplers. we are ready to investigate the application of transform methods to discrete and sampleddata systems. Taking the Laplace transform on both sides of Eq. (2153), we have
In
Now
R*(s)
r(kT)e
(2158)
(21 58) contains the exponential term e~ kTs reveals the difficulty of using Laplace transform for the general treatment of discretedata systems, since the transfer function relations will no longer be algebraic as in the continuousdata case. Although it is conceptually simple to perform inverse Laplace transform on algebraic transfer relations, it is not a simple matter to perform inverse Laplace transform on transcendental functions. One simple fact
is
the
that
tables
do not have
entries
with trans
cendental functions in s. This necessitates the use of the ztransform. Our motivation here for the generation of the ztransform is simply to convert transcendental functions in s into algebraic ones in z. The definition of ztransform is given with this objective in mind.
Definition of the zTransform
The ztransform
is
defined as z
(2159)
Equation
= In z
is
(2160)
written
r (kT)z~
R*(s
or
R(z)
= L in z) =
R(Z)
(2161)
ztransform of
/*(/)
(2162)
= [Laplace transform
of r*(t)] s=1/Tlnz
42
Mathematical Foundation
ap
'
Therefore, z =
e
Ts
.
we have
= e kT = S
= 0,1,2,...
(2163)
is
From
e~ kT5(t
kT)
(2164)
Then
r*(s)
f; e *=o
' kT
e kTl
(2165)
< + )T and subtract the resulting equation from Multiply both sides of Eq. (2165) by e that R*(s) can be written in a closed form, ; now we can show easily Eq. (2165)
*>(*)
for
!__}<...
<2
" 166)
g(a+)r
<
._.
(2167)
/*(/) is
where
<T is
Then
the ztransform of
*
itfy>
(2168)
forle^z'l
<
1.
Example
218
In Example 217,
if
r(*D
l,
= 0, we have * = 0,1,2,...
all
(2169)
equal to unity.
~ e kTs
Then
(2 " 170)
= S A=
1
Biz)
= z* = k=a
R(z)
(2171)
This expression
is
written in closed
form
y4^
i*m<i
Z_1
<
2  172)
or
^) = F=1
I<1
(2
 173
>
functions are obtained by In general, the ztransforms of more complex preceding two examples. If a time use of the same procedure as described in the of finding its ztransfunction r(t) is given as the starting point, the procedure Eq. (2161) to get R(z). and then use is to first form the sequence r(kT)
form
An
is
an
then take the Laplace transform ideal sampler whose output is r*(t). and R(z) is obtained by substituting of r*(t) to give R*(s) as in Eq. (2158),
z for
e
We
Ts
.
Sec. 2.8
zTransform
43
Table 21
Table of zTransforms
Laplace Transform
Time Function
Unit impulse
Unit step
S(t)
zTransform
(/)
znT)
z
e~
\_
StU)
=2,6(tn=0
zTz
(Z
s2
1)2
2
1
T*z{z
1(Z lim
1)
1)3
s n+l
1
11
n\
(1)"
n\
d (
daAz
z
er)
+
1
a
te~ a <
e T
U+
a
s(s
a) 2
1
Tze~' T {z e" T ) 2
(1
+
co
a)
sin
e~<"
e*r)z
(z
l)(z
e'T)
S2
+ CO 2
CO a) 2
of
z2
(*
+
S2
+
2
co 2
z 2 e 2<,T z2
(s
+ CO +a + a) 2 + co 2
s
cos
cot
z(z
cot
z2 z2
Table 21 gives the ztransforms of some of the time functions commonly used in systems analysis. A more extensive table may be found in the
litera
ture. 12 13

Inverse zTransformation
Just as in the Laplace transformation, one of the major objectives of the is that algebraic manipulations can be made first in the zdomain, and then the final time response is determined by the inverse
ztransformation
ztransfor
mation. In general, the inverse ztransformation of R(z) can yield information only on r(kT), not on r(t). In other words, the ztransform carries information only in a discrete fashion. When the time signal r(t) is sampled by the ideal sampler, only information on the signal at the sampling instants, t kT, is
retained.
With
this in
2.
3.
The partialfraction expansion method. The powerseries method. The inversion formula. The ztransform function R{z)
is
ex
panded by
sum of simple
recognizable terms,
Chap. 2
44
Mathematical Foundation
used to determine the corresponding r{kT). In slight difference between carrying out the partialfraction expansion, there is a With reference to the zthe ztransform and the Laplace transform procedures. functions have the transform table, we note that practically all the transform
table
is
into the form term z in the numerator. Therefore, we should expand R{z)
**)
For
this,
= r a + r +"
.
<
2  174)
we should
expand R(z)jz into fractions and then multiply z across illustrate this desired expression. The following example will
first
recommended procedure.
Example 219
Given the ztransform function
R( ^ R{z >
it is
_ ~
(z
l)(z
~ e "T z T  e)
(2175)
)
we have
(2176)
R(z) z
= _J z
I
e" T
Thus,
n( .\
(2177)
From
R(z)
is
found to be
r(kT)
 e~'kT
is
(2178)
Powerseries method.
series in
of
r(t) at
powers t = kT,
into a
21( Z )
the value of z" In view of Eq. (2161), the coefficient of z~* is for the R(z) in Eq. (2175), we or simply r(kT). For example,
1
.
expand
it
power
series in
powers of z'
by long
division; then
we have
(i
_
. .
>"
.
(1
e
2or
)z 2
...
(1
e~
3 T
)zi
(2
" 179
+
or R(z)
+ (l _ e  kT)z k +
e
kT
>
=S *=
(1
)z
( 2_18() )
ThUS
'
r(kT)=le
is
(218D
which
the
same
Inversion formula.
r(kT)=
which
\z\
is
^U 2ff
<f
R{z)z k 'dz
(2182)
J J r
cT
3 a circle of radius a contour integration along the path T, where T is zplane, and c is of such a value that all centered at the origin in the
circle.
Sec. 2.8
zTransform / 45
One way of evaluating the contour integration of Eq. (2182) is by use of the residue theorem of complex variable theory. Equation (2182) may be written
r(
kT )
= 2jp
R(z)z k
~1
dz
l
= 2 Residues
For simple
Residue of R(z)z
k'1
(2183)
at the pole z
=z
is
obtained as
(2184)
at the pole z,
= (z
z J )R(z)z" 1  x ,
Now let us consider the same function used in Example 219. The function R(z) of Eq. (2175) has two poles: z 1 and z e~ r Using Eq. (2183), we
have
r(kT)
[Residue of iJ(z)z* at z
)z (1 (zer
) T
k
=
T
1]
+
k
[Residue of R(z)z k
at z
= e~'T
(1
)z
r
,.,
(z1)
,_,^,
(2185)
=  e" kT
1
If r x {kT)
and
r 2 {kT)
J? 2 (z), respectively,
then
gUr^kT)
2.
r 2 (kT)]
=R
(z)
2 (z)
(2186)
Multiplication
by a Constant
S[ar(kT)]
as[r(kT))
aR(z)
(2187)
where a
3.
is
a constant.
Real Translation
$[r(kT
nT)]
z~"R(z)
(2188)
and
g[r(kT
nT)]
= Ar{z) 
"jj
r(kT)z k
~]
(2189)
is a positive integer. Equation (2188) represents the ztransform of a time sequence that is shifted to the right by nT, and Eq. (2189) denotes that of a time sequence shifted to the left by nT. The reason the righthand side of Eq. (2189) is not z"R(z) is because
where n
is
Chap. 2
46
Mathematical Foundation
Eq. (2189) Thus, the second term on the righthand side of that is lost after it is shifted to the simply represents the sequence
>
0.
left
4.
by nT.
Translation
Complex
Z[e*
5.
Initial Value
kT
r(kT)\
R{ze^ T )
(2190)
Theorem
lim r(kT)
if
lim R(z)
(2191)
6. FinalValue
Theorem
lim r{kT)
lim
(1
z' l )R{z)
(2192)
if
z" ')*(*). has no poles on or outside the unit the origin in the zplane, \z\= 1 circle centered at the function,
(1
The
Example 220
Let
translation
te,
0.
KO =
?>0;then
R(z)
*[/.(01
g(kT)
= ^jjj
= Jl^.ry
(2193)
glte'uM]
R{ze")
(2
"
194 )
Example
221
=
(z
l)(z
0.792z 2 0.416z
+ 0.208)
(2 . 195)
infinity.
(l^W) z2 _
7 9 2Z
41 6 2
does not have any pole on or outside the unit circle Hence, value theorem of the ztransform can be applied.
+ 0.208 in \z\ =
1
(2496)
the zplane, the final
S K*r> +
6
92
lim
z2
_o 416 z + 0.208
R(.z) in
="
'
^
l
This result
is
easily
checked by expanding
i.i21z
powers of z~
R(z)
0.792zi
1.091*'
1.013*"*
0.986z=
+ 0.981z" +
It is
0.998z~ 7
^^
final
series
value of unity.
Chap. 2
Problems
47
REFERENCES
Complex
1.
wood
2.
i
1952.
to
R. Bellman, Introduction
one,
i
"dO.
*<?/
New
3.
C.
Kuo,
Ai*,*, an/
York, 1967.
4.
R. Legros and A. V.
PrenticeHall, Inc.,
5.
%I? lttt
D
Matrices ' Polynomials, and Linear ' TimeInvariant Systems," Trans. Automatic Control, Vol. AC18, pp. 110, Feb. 1973.
ETT
Poles,
8
'
Hazony and J. Riley, "Evaluating Residues and IRE Trans. Automatic Control, Vol. AC4,
Coefficients of
High Order
L tI Partial Fraction u7 Poles by !f. Multiple fu Digital Computer," pp. 161162, Mar. 1964.
IEEE
CT11
9
'
IEEE
10
'
QdCk ChCCk n Fraction Expansion ^ix Automatic Control, Vol. AC11, pp. 318319, Apr. 1966. Trans.
UNIS
Partial
;
"
Coefficients,"
Feb. 1968.
n Expami0n of RationaI F ^tions wi't^r^n h B One HighOrder Pole," IEEE Trans. Automatic Control, Vol. AC13
with
f:.
'
R R
"
Partkl FraCti
p 133
'
"'
IEEE Trans.
Automatic Control,
B. C.
Kuo
Hall, Inc.,
13.
Analysis and Synthesis of Sampled Data Control Systems, Englewood Cliffs, N.J., 1963.
Discrete
Prentice
C.
Kuo,
Champaign,
Illinois, 1970.
PROBLEMS
2.1.
Find the poles and zeros of the following functions (include the ones
W
'
at infinity)
(a)
G(s) ()
= =
(s
5(s
*Hs
(b) G(s)
+ 5) + 1) + 2)(s> + 3* +
2Xs
* 2 (*
1)
2)
: :
48
Mathematical Foundation
Chap. 2
(C)
G(s)
= =
:
f'
+
^u
+ 2)
Ke"
(j
(d) G(s)
DC*
2.2.
(b) git)
(c)
(d)
fcT); <*(<)
*=o
2.3.
gU)
1
1 1
sl
(a)
irOT
l
(b)
Figure P23.
2.4.
fit)
t
'
3<f<4
4
<t
2.5.
wm^j dM + 4/(0 = ^
dt*
5 J dt
,
uAt)
Assume
2.6.
(a) Gis)
is
+ 2)is +
1
3)
(b) Gis)
(.?
+ D 2 (s +
10
4)
(c)
Gis)
 s(j2
(d) Gis)
+ 4)(j +  s(s2 +s 2) +
1}
Chap. 2
Problems
49
2.7.
"3
6"

_0
(b)
"
5_
+
r 20"
0~
10_
3
15"
1
.
 4
_
3_
5.
(c)
s
1
10
s3
2.8.
AB
and
Determine if the following matrices are conformable for the products BA. Find the valid products.
(a)
rii
A=
_3_
(b)
B=
"2
[6
1]
r
0.
B=
set
10
9"
1
1
0.
2.9.
1
1
2.10.
set
of
form (t)
= Ax(f)
+ Bu(:
*i(0
x 2 (t)
x 3 (t)
2.11.
"
.
u 2 (t)
5"
JO
(b)
1^
3
r
_
11
1 1
2
l
1_
41
(c)
3
1
l
_i
2.12.
(a)
1_
7 _0 "2
3_
(b)
4
2
6
8"
Ll
3_
50
Mathematical Foundation
Chap. 2
2.13.
_l
(b)
"
1
2_ 5
r
i_
2
_
2.14.
The following
of
sampled by an ideal sampler with a sampling period Determine the output of the sampler, /*(/), and find the Laplace transform of/*0), F*(s). Express F*(s) in closed form.
signals are
T seconds.
(a) /(/)
(b) fit)
2.15.
G(s)
is
+ a)n
1
+
1
5)
(d)
(e)
2.16.
G(z)
(z
l)(z 2
+ 1) +z
1)
expansion
3
Transfer Function and
Signal Flow
Graphs
3.1
Introduction
ponents.
One of the most important steps in the analysis of a physical system is the mathematical description and modeling of the system. A mathematical model ot a system is essential because it allows one to gain a clear understanding of the system in terms of causeandeffect relationships among the system com
signal flow graphs are valuable tools for analysis as well as for design. In this chapter we give the definition of transfer function of a linear system and demonstrate the power of the signalflowgraph technique in the analysis of linear systems.
by a schematic diagram the system components. From the mathematical standpoint, algebraic and differential or difference equations can be used to describe the dynamic behavior of a system In systems theory, the block diagram is often used to portray systems of all types. For linear systems, transfer functions and
In general, a physical system can be represented that portrays the relationships and interconnections
among
3.2
Transfer function plays an important role in the characterization of linear timeinvariant systems. Together with block diagram and signal flow graph transfer function forms the basis of representing the inputoutput relationships ot a linear timeinvariant system in classical control theory. The starting point of defining the transfer function is the differential
equa51
62
/ Transfer
Chap. 3
dynamic system. Consider that a linear timeinvariant system described by the following nthorder differential equation
tion of a
is
d m r{t)
dm
'/(?)
, .
/,
dr(t)
,.
where a au
,
c(r) is
.
and
r(t) is
The
coefficients,
m. b m are constants, and n The differential equation in Eq. (31) represents a complete description of the system between the input r(t) and the output c(t). Once the input and the initial conditions of the system are specified, the output response may be
.
and b
bt
>
obtained by solving Eq. (31). However, it is apparent that the differential equation method of describing a system is, although essential, a rather cumbersome one, and the higherorder differential equation of Eq. (31) is of little practical use in design. More important is the fact that although efficient subroutines are available
computers for the solution of highorder differential equations, the important development in linear control theory relies on analysis and design techniques without actual solutions of the system differ
on
digital
ential equations.
of describing linear systems is made possible by the use of transfer function and impulse response. To obtain the transfer function of the linear system that is represented by Eq. (31), we take the Laplace transform on
A convenient way
initial
conditions,
we have
a^""
+ a ^s + a)C(s) = (b s m + b.s'"' +
n
...
b m .,s
L
b m )R(s)
The
is
therefore,
s"
+... +...
+ bm . s + b m + a . s a.
1
(3 _ 3)
,
state
strictly,
3.
4.
output to the Laplace transform of the input. All initial conditions of the system are assumed to be zero. A transfer function is independent of input excitation.
is
given to illustrate
how
Example
31
series
RLC
network
e,{t).
designated by
31.
is
this case
can be defined as
the voltage across any one of the three network elements, or the current
Sec. 3.2
53
+ o
R v\M
Tm
'W
.
i(t).
is
written
=Ri(t)+L^ + j /(,) dt
(34)
ei<"
^)
;;C
e c (t)
Taking the Laplace transform on both sides of Eq. and assuming zero initial conditions, we have
EAs)
(34)
= (r+Ls+^)I(s)
i
(35)
Fig. 31.
RLC network.
./(*)
If
we regard
1
the current
e,{t)
(/)
as
an output
simply
function between
and
I
i(t) is
E,(s)
Cs
^'^
is
E (s)=~Ks)
c
(3.7)
E (s) _
c
,(*)
+ RCs + LCs
is
<3
8
easily extended to a system with a system of this type is often referred to as a multivariable system. In a multivariate system, a differential equation of the form of Eq. (31) may be used to describe the relationship between a pair of input and output. When dealing with the relationship between one input and one output, it is assumed that all other inputs are set to zero. Since the principle of superposition is valid for linear systems, the total effect on any output variable due to all the inputs acting simultaneously can be obtained by adding the
The
multiple
outputs.
individual effects.
simple illustrative example of the transfer functions of a multivariable us consider the control of a turbopropeller engine. In this case the input variables are the fuel rate and the propeller blade angle. The output variables are the speed of rotation of the engine and the turbineinlet temperature. In general, either one of the outputs is affected by the changes in both inputs. For instance, when the blade angle of the propeller is increased, the speed of rotation of the engine will decrease and the temperature usually increases. The following transfer relations may be written from steadystate tests performed on the system:
system,
let
As a
C,(j) Ci(s)
2 (s)
2 (s)
(39)
where
(310)
= transformed variable of speed of rotation = transformed variable of turbineinlet temperature Ri(s) = transformed variable of fuel rate R (s) = transformed variable of propeller blade angle
C,(i)
2 (s)
some
reference levels.
54
Chap. 3
Since Eqs. (39) and (310) are written with the assumption that the system
is
that
is,
R 2 (s) =
0.
made
In general,
if
j'th
defined as
(311)
C (s)
t
with
R k (s) =
0,
l,2,...,p,
is
k^j. Note
defined with
z'th
effect,
output
by
Cls)
= G n {s)R
{s)
+G
i2
{s)R 2 {s)
...+
G, p (s)R p (s)
(312)
= t Gu {s)Rj(s)
where
(i=l,2,...,q)
tJ
(s) is
It is
G(s)R(s)
(313)
where
cm
C2 {s)
C(*)
(314)
Cq{s)j
is
a q
Rii.s)
R(*)
=
RJis)}
(315)
is
&p X
'G^is)
...
G lp (s) G 2p {s)
(316)
G(s)
G ql (s)
is
G q2 (s)
...
G(s)_
aq
X p
ec 3 3
' '
55
3.3
The impulse response of a linear system is defined as the output response of the system when the input is a unit impulse function. Therefore, for a system with a single input and a single output, if r(t) = d(t), the Laplace transform of the
system output
is
is,
G(s)
(317)
since the Laplace transform of the unit impulse function is unity. Taking the inverse Laplace transform on both sides of Eq. (317) yields
c(0
= 8(t)
(318)
Laplace transform of G(s) and is the impulse response (sometimes also called the weighing function) of a linear system. Therefore, we can state that the Laplace transform of the impulse response is the transfer
function.
where
is
linear systems,
means that if a linear system has zero initial conditions, theoretically, the system can be described or identified by exciting it with a unit impulse response and measuring the output. In practice, although a true impulse cannot be generated physically, a pulse with a very narrow pulsewidth usually provides a suitable approximation.
For a multivariable system, an impulse response matrix must be defined and
is
given by
g(0
'[G(*)]
(319)
where the inverse Laplace transform of G(s) implies the transform operation on each term of the matrix.
The
differential equation,
is based on the knowledge of the system and the solution of C(s) from Eq. (33) also assumes that
all available in analytical forms. This is not always possible for quite often the input signal #(/) is not Laplace transformable or is available only in the form of experimental data. Under such conditions, to analyze the system we would have to work with the time function r(t) and g(t).
Let us consider that the input signal r(j) shown in Fig. 32(a) is applied to a whose impulse response is g{t). The output response c(t) is to be determined. In this case we have denoted the input signal as a function of r which is the time variable; this is necessary since t is reserved as a fixed time
linear system
all practical
purposes, r(r)
is
from minus
assumed to extend
Now consider that the input r(r) is approximated by a sequence of pulses of pulsewidth At, as shown in Fig. 32(b). In the limit, as At approaches
zero
become impulses, and the impulse at time kLx has a strength or area equal to Atz^At), which is the area of the pulse at kLx. Also, when At decreases, k has to be increased proportionally, so the value of &At remains constant and equals t, which is a particular point on the time axis. We now compute the output response of the linear system, using the
impulseapproxi
these pulses
56
Flow Graphs
Chap. 3
r(T)
r(kAr)
(b)
Fig. 32. (a) Input signal of a linear system, (b) Input signal represented
by
sum of rectangular
pulses.
mated
signal.
is
When
= kAx is considered,
the system
response
given by
At
which
is
r(kAx)g(t
kAx)
(320)
strength Axr(kAx).
By use of
due to r(x) is obtained by adding up the responses due to each of the impulses from oo to +co. Therefore,
c{t)
= =
lim
AT oo
r(kAx)g(t
kAx) Ax
(321)
or
c{t)
^_j(x)g(tx)dx
(322)
For
all
Thus
(323)
=
applied at
/
for
< 0,
is
0.
Or
(324)
g (t
x)
t<z
Sec 3 3
57
the system
c (0
is
now
written
= /'__ r(T)g(t 
T)
<*t
(325)
Further,
if r( T )
for T
<
C
0,
W = J' *<*)*('  t) A
o
(325)
is
and
so
(3. 27 )
is
interpreted as
c(t)
r(t) * g(t)
r(t)
(328)
The positions of
r(t)
and
g{t) in the
convolution operation
changed, since basically there is no difference between the two functions Therefore, the convolution integral can also be written as
c (0
may be
inter
f'
g(i>(t
T ) dx
(329)
/(?)
may be caused by simple deterioration of components due to wear and tear, drift in operating environments, and the like. Some systems simply have parameters that vary with time a predictable or unpredictable fashion. For instance, the transfer characteristic of a guided missile in flight will vary in time because of the change of mass of the nmsile and the change of atmospheric conditions. On the other hand, for a simple mechanical system with mass and friction, the latter may be subject to unpredictable variation either due to "aging" or surface conditions thus the control system designed under the assumption of known and fixed parameters may fail to yield satisfactory response should the system parameters vary. In order that the system may have the ability of selfcorrection or selfadjustment in accordance with varying parameters and environment it is necessary that the system's transfer characteristics be identified continuously or at appropriate intervals during the operation of the system. One of the methods of identification is to measure the impulse response of the system so that design parameters may be adjusted accordingly to attain optimal control at all times In the two preceding sections, definitions of transfer function and impulse response of a linear system have been presented. The two functions are directly related through the Laplace transformation, and they represent essentially the same information about the system. However, it must be reiterated that
The evaluation of the impulse response of linear a system is sometimes an important step in the analysis and design of a class of systems known as the adaptive control systems. In real life the dynamic characteristics of most systems vary to some extent over an extended period of time. This
transfer
58
Chap. 3
function and impulse response are denned only for linear systems and that the
initial
3.4
Block Diagrams
Because of
its
simplicity
and
versatility,
all types.
block diagram
is
can be used,
together with transfer functions, to represent the causeandeffect relationships throughout the system. For instance, the block diagram of Fig. 33 represents a turbinedriven hydraulic power system for an aircraft. The main components of
the system include a pressurecompensated hydraulic an electronic speed controller, and a control valve.
figure depicts
how
these
Current
Controller
Control
valve
Inlet
Turbine torque
Turbine
Output
Mass
flow
rO
Pump
torque
Load
pressure
Hydraulic
pump
Load flow
Fig. 33.
If the mathematical and functional relationships of all the system elements known, the block diagram can be used as a reference for the analytical or the computer solution of the system. Furthermore, if all the system elements are assumed to be linear, the transfer function for the overall system can be obtained by means of blockdiagram algebra. The essential point is that block diagram can be used to portray nonlinear as well as linear systems. For example, Fig. 34(a) shows the block diagram of a simple control system which includes an amplifier and a motor. In the
are
is
depicted by
its
nonlinear
Sec. 3.4
Block Diagrams
59
Kry
s(s+a)
(a)
(b)
Block diagram of a simple control system, (a) Amplifier shown with a nonlinear gain characteristic, (b) Amplifier shown with a linear gain
Fig. 34.
characteristic.
system as
relation. The motor is assumed to be linear and its dynamics are represented by a transfer function between the input voltage and the output displacement. Figure 34(b) illustrates the same system but with the amplifier characteristic approximated by a constant gain. In this case the overall system is linear, and it is now possible to write the transfer function for the overall
inputoutput
E (s)
t
Em (s)
E{s)
s(s
a)
(3
"3
We
shall
trol systems
now define some blockdiagram elements used frequently in conand the blockdiagram algebra. One of the important components
for
s lg nal
and sometimes combinations of these. The blockdiagram elements of these operations are illustrated as shown in Fig. 35. It should be pointed out that the signals shown in the diagram of Fig. 35 can be functions of time t or functions of the Laplace transform variable s. In Fig. 34 we have already used blockdiagram elements to represent inputoutput relationships of linear and nonlinear elements. It simply shows that the blockdiagram notation can be used to represent practically any inputoutput relation as long as the relation is defined. For instance, the block diagram of
tiplication,
synchros, resolvers, differential amplifiers, multipliers, and so on In general, the operations of the sensing devices are addition, subtraction mul
ometer,
is the sensing device that acts as a junction point comparisons. The physical components involved are the potenti
60
/ Transfer
Chap. 3
*
+~
+c
(a)
Subtraction
(b)
Addition
*
e = rl
r2
+~ e
=rc
(c)
(d) Multiplication
systems, (a) Subtraction, (b) Addition, (c) Addition and subtraction, (d)
Multiplication.
u(t)
x(t)
u(f)
x(0
x=
f(x, u)
x =ax + bu
R(s)
G(s)
C(s)
(a)
(b)
(c)
Fig. 36.
systems.
is
tion
x(t)
ax{t)
bu{t)
(331)
Figure 36(b) illustrates the inputoutput relation of a system described by the vectormatrix differential equation
{t)
= f [x(0, u]
(332)
where
x(f) is
Fig. 36(c)
an n X 1 vector and u(t) is an r X 1 vector. As another example, shows a block diagram which represents the transfer function of a
Sec. 3.4
Block Diagrams
61
is,
C(s)
G{s)R(s)
(333)
where G(s)
is
Figure 37 shows the block diagram of a linear feedback control system.
The following terminology often used in control systems is defined with reference
to the block
diagram
R(s)
C(s)
/(/)>
c(t),
= = =
reference input
b(i),
B(s)
e(t), &(s)
e(t),
E(s)
 C(s) = error signal G(s) = g?^ = openloop transfer function or forwardpath transfer function
(s)
The
= j4 = closedloop transfer function H(s) = feedbackpath transfer function G(s)H(s) = loop transfer function closedloop transfer function, M(s) = C(s)/R(s), can be expressed as
From
C(s)
Fig. 37
C(s)
we
write
(334)
= =
G(s)&(s)
and
B(s)
H(s)C(s)
(335)
The actuating
signal
is
written
6(j)
= R(s)  B(s)
G(s)B(s)
(336)
C(s)
G(s)R(s)
(337)
C(s)
G(s)R(s)
G(s)H(s)C(s)
(338)
Solving C(s) from the last equation, the closedloop transfer function of the
R(s)
r(t)
.y
,
/)
ew
'
eit)
G(s)
C(s)
c(t)
bit) B(s)
H(s)
62
Chap. 3
system
is
given by
walgebraic
^
(339)
Gis)His)
may contain many feedback loops, and from the block diagram by means of the
method described above may be tedious. In principle at least, the block diagram of a system with one input and one output can always be reduced to the basic singleloop form of Fig. 37. However, the steps involved in the
reduction process
may
We
shall
show
transfer function of
any
block
multivariable system
Two
block diagram of Fig. 38(b), the multiplicity of the inputs and outputs is denoted by vectors. The case of Fig. 38(b) is preferable in practice because of its
simplicity.
MOMultivariable
 c l
(t)
system
'
(0
*~c q (t)
(a)
Multivariable
i(0"
system
* c(/)
(b)
Fig. 38.
system.
Figure 39 shows the block diagram of a multivariable feedback control The transfer function relationship between the input and the output of
is
the system
Sec. 3.4
Block Diagrams
63
R(s)
S)
i
80)
G(s)
^
Bfr)
Fig. 39.
C&)
Hfc)
Substituting Eq. (342) into Eq. (341) and then from Eq. (341) into
yields
Eq
(340)
C(s)
G(s)R(s)
G(s)H(s)C(s)
(343)
[I
G(j)H(j)] G(s)R(s)
(344)
I G(5 )H(j) is nonsingular. should be mentioned that although the development of the inputoutput relationship here is similar to that of the single inputoutput case, in the
It
provided that
present
improper to speak of the ratio C(s)/R(s) since C(s) and R(s) are matrices. However, it is still possible to define the closedloop transfer
it is
situation
matrix as
(345)
M(j)
[I
+
=
G(*)H(s)]G(j)
Then Eq.
(344)
is
written
C(j)
M(a)R(j)
(346)
Example
32
Consider that the forwardpath transfer function matrix and the feedbackpath transfer function matrix of the system shown in Fig. 39 are
r
l
_j_i
l
1
G(s,=
and
H(s)
respectively.
+
2
(347)
2J
0"
i_
Lo
The
is
is
evalu
ated as follows:
1
G(s)H(s)
s+
2
1_
1
s
1 1
'
+
2
s+j,
s
(348)
2J
+ 2J
64
/ Transfer
Chap. 3
The closedloop
transfer matrix
is
M(a)
[I
G( S )H(s)]'GM
= ^
+ +
_1_
2
5
2
where
+ +
(349)
2
1
2J
A_s+2s+3+ S+1S2
,
2 __ s 2
S
+
4
5s
5(5+1)
(350)
Thus
35 2
s{s
+95 +
1)(5
M(5)
s(s
52
+
55
1)
_ ~5
35
5(5
1_
2)
+ +
(351)
2
1).
3.5
signal flow graph may be regarded as a simplified notation for a block diagram, although it was originally introduced by S. J. Mason as a causeandeffect
representation of linear systems. In general, besides the difference in the physical appearances of the signal flow graph and the block diagram, we may regard
more rigid mathematical relationships, whereas the rules of using the blockdiagram notation are far more flexible and
less stringent.
A signal flow graph may be defined as a graphical means of portraying the inputoutput relationships between the variables of a set of linear algebraic
equations.
is
N algebraic equations
(352)
in the
S
=
o kJ y k
1,2,
,N
form of cause
andeffect relations
N
k=l
jth effect
XI ( am fr
cause)
(353)
or simply
output
This
is
= 2 (gain)(input)
in the construction of the set
(354)
the single
of algebriac
equations from which a signal flow graph is drawn. In the case when a system is represented by a
equations,
set
of integrodifferential
we must
first
form of Eq.
(352), or
W
When
= S G kj (s)Yk (s)
.
7=1,2,
(355)
ys and y k The nodes are connected together by line segments called branches, according to the causeandeffect equations. The
Sec 3 5
Signal Flow Graphs /
65
causeandeach variable in terms of itself and the other variables For instance, consider that a linear system is represented by the simple equation
effect relations relating
general, given a set of equations such as those of Eq. (352) or Eq. (355), the construction of the signal flow graph is basically a matter of following through the
branches have associated branch gains and directions. through a branch only in the direction of the arrow. In
where j, is the input variable, y 2 the output variable, and a l2 the gain or transmittance between the two variables. The signalflowgraph representation of Eq. (356) is shown in Fig. 310. Notice that the branch
n
*
directing
ence of y 2 upon
should be reiterated that Eq. (356) and Fig. 310 represent only the dependence of the out>,. It
12 >''
An
P ut variable upon the input variable, not the reverse. important consideration in the application of
is
signal
that the branch between the two nodes y, and y 2 should be integrated as a unilateral amplifier with gain a i2 , so that when a signal of one unit is applied at the input y u the signal is multiplied by a l2 and a signal of strength a l2 is delivered at node>> 2 Although algebraically Eq (356) can be rewritten
.
flow graphs
is
the signal flow graph of Fig. 310 does not imply this relationship If Eq (357) valid as a causeandeffect equation in the physical sense, a new signal flow
set
of algebraic
ys
= tf12.F1 = a 23 y = Oi*yi = a 2s y
2
+ a 32 y + a43 y + a 34 y + +a
3
a 44 y 4
(3 " 58)
4 <,y 4
graph for these equations is constructed step by step as shown in Fig. 311, although the indicated sequence of steps is not unique The nodes
representing the variables
right.
The
signal flow
The
;
Similarly, with the consideration of the third equation, Fig. 31 1(c) is obtained when the last equation of Eq. (358) is portrayed, the complete signal flow graph is shown in Fig. 31 1(d). The branch that
Finally,
drawn as shown in y 3 depends upon a 23 y 2 and a 43y 4 therefore on the signal flow graph of Fig. 31 1(a), a branch of gain a 23 is drawn trom node y 2 to y 3 and a branch of gain is drawn from y t to y 3 with the 43 directions of the branches indicated by the arrows, as shown in Fig 31 1(b)
Fig. 31 1(a).
<hiy 3
y u y2 y 3 y 4 and y s are located in order from left to equation states that 2 depends upon two y signals, a liyi and the signal flow graph representing this equation is
, , ,
first
states that
y4
66
/ Transfer
Chap. 3
o
>4 (a)y 2
O
y$
=a n y + a 31 y3
l
a 43
o
y\
yi
(b)>>2
^3
y*
=fl23> 2
,
>"5
=a n y\ +032^3. ^3
+ 43>'4
O
^5
=a 24>
+ a 34>
fl
44>"4
(d)
Complete
signal
flow graph
y2
ai 2 yi
a 3 zy3 (b) yi
+ any*, (c) yz = a\zy\ + a 3 zy3, yi + a t,yi + cmy,. (d) Complete signal flow graph.
3
y*
= miyi = auyi
called a loop,
3.6
Summary
Flow Graphs
At
this point
it is
best to summarize
properties of the
1.
2.
3.
The equations based on which a signal flow graph is drawn must be algebraic equations in the form of effects as functions of causes. Nodes are used to represent variables. Normally, the nodes are
Sec. 3.7
Definitions for Signal
Flow Graphs
67
4.
left to right, following a succession of causes and through the system. Signals travel along branches only in the direction described by the arrows of the branches.
arranged from
effects
5.
6.
The branch directing from node y k to j, represents the dependence of the variable y, upon but not the reverse. yk A signal y k traveling along a branch between nodes yk and y. is multiplied by the gain of the branch, a kj so that a signal
,
aj
k is
3.7
Flow Graphs
In addition to the branches and nodes defined earlier for the signal flow graph the following terms are useful for the purposes of
identification
and
reference.
Input node {source). An input node branches. (Example: node yi in Fig. 311.)
is
p Mg.
Output node (sink). An output node is a node which has only incoming branches. (Example: node y 5 in Fig. 311.) However, this condition is not always r eadiy b an out P ut node For instance, the signal flow
312(a) does not have
i
^/
J
graph shown in
satisfies
node. However,
r
it may be necessary to regard nodes y 2 and/or 3 as output y nodes In order to meet the definition requirement, we may simply introduce
eS
fS
in Fig.
graph
it is
flow graph
68
/ Transfer
Chap. 3
are added. In general, we can state that any graph can always be made an output node by noninput node the aforementioned operation. However, we cannot convert a noninput node into an input node by using a similar operation. For instance, node y 2 of the signal flow graph of Fig. 312(a) does not satisfy the definition of an input node. If we attempt to convert it into an input node by adding an incoming branch of
equations y 2
=y
and y 3
=y
of a signal flow
would
result.
y 2 the signal flow graph of Fig. 313 However, the equation that portrays the relationship at node y 2
,
now
reads
yi
= yi +
yi
<*\%y\
a nyi
(3 59 )
which
is
different
from the
from
Fig. 3 12(a),
a i2 yi
a 32 y 3
(360)
Vi
Fig. 313.
an input node.
node.
Since the only proper way that a signal flow graph can be drawn is from a set of causeandeffect equations, that is, with the causes on the right side of the equation and the effects on the left side of the equation, we must transfer y 2
if it were to be an input. Rearranging Eq. (360), equations originally for the signal flow graph of Fig. 312 now become the two
y%
y
"12
(361)
12
y3
a 2i y 2
is
(362)
The
two equations
shown
y2
as an input node.
any collection of continuous succession of branches The definition of a path is entirely general since traversed in the same any node to be traversed more than once. Therefore, as it does not prevent simple as the signal flow graph of Fig. 312(a) is, it may have numerous paths.
Path.
path
is
direction.
Forward path.
forward path
is
ends at an output node and along which no node is traversed more than once. For example, in the signal flow graph of Fig. 31 1(d), y is the input node, and there are four possible output nodes in y 2 y 3 j 4 and y s The forward path
t , , ,
.
Sec. 3.8
SignalFlowGraph Algebra
69
between j t and y 2 is simply the branch connected between y and y 2 There are two forward paths between y L and y 3 one contains the branches from y t to y 2 to y 3 and the other one contains the branches from y to y 2 to y 4 (through the branch with gain a 2i ) and then back to y 3 (through the branch with gain a 43 ).
t
.
The reader may determine the two forward paths between j>, and j 4 there are also two forward paths between j, and y s
.
Similarly,
Loop.
loop
is
there are four loops in the signal flow graph of Fig. 31 1(d). These are
shown
y*
"44
Fig. 315.
Four loops
graph of Fig.
31 1(d).
Path gain. The product of the branch gains encountered in traversing a path is called the path gain. For example, the path gain for the path
J>i
yz  y  y*
3
a 12 a 23 a 34
is
is
y2
y4 y y
3
in Fig. 315
is
a 2i a i3 a 32
3.8
SignalFlowGraph Algebra
Based on the properties of the signal flow graph, we can manipulation and algebra of the signal flow graph.
1
.
variable represented by a
node
is
equal to the
sum
the signals entering the node. Therefore, for the signal flow
70
Chap. 3
Fig. 316.
Node
as a
summing
graph of Fig. 316, the value of y is equal to the sum of the signals transmitted through all the incoming branches; that is,
{
Jl
2.
"li}>2
31^3
4lj4
Osijs
(363)
is
variable represented by a
node
transmitted
Fig. 316,
we have
ye
yn
y%
3.
= = =
tfi6.Fi
a xl y
(364)
tfisJi
same
a single
two
branch with gain equal to the sum nodes can be replaced by the parallel branches. An example of this case is of the gains of
illustrated in Fig. 317.
X
y\
Fig. 317. Signal flow
single branch.
avb +
>
X.
yi
Sec. 3.9
71
a 12
a 23
"34
<*45 45
a "56
O
V\
O
vi
O
y3
o
y4
O
ys
y6
a 12<*23tf34 fl 45fl56
o
Fig. 318. Signal flow
o
re
4.
5.
connection of unidirectional branches, as shown in Fig. can be replaced by a single branch with gain equal to the product of the branch gains. Signal flow graph of a feedback control system. Figure 319 shows
series
318,
whose block
may be regarded as a simplified notation for the block diagram. Writing the equations for the signals at the nodes &(s) and C(s), we have
diagram
is
S(j)
= =
R(s)
H(s)C(s)
(365)
and
C(s)
G(s)&(s)
is
(366)
The closedloop
equations,
transfer
function
C(s) R(s)
_
1
G(s) G(s)H(s)
(367)
R(s)
e(s)\_
ms)
Fig. 319. Signal flow
^y
<Xs)
C(s)
For complex
formula will be introduced which allows the determination of the gain between an input node and an output node by mere inspection.
3.9
ical
was emphasized earlier that the construction of a signal flow graph of a physsystem depends upon first writing the equations of the system in the cause
we
shall give
two simple
illustrative
examples.
72
/ Transfer
Chap. 3
Owing to
electric
5,
the lack of background on systems at this early stage, we are using two networks as examples. More elaborate cases will be discussed in Chapter where the modeling of systems is formally covered.
33
Example
The
passive network
shown
in Fig. 320(a)
is
considered to consist of
R, L, and
C elements
by impedance functions, Z(s), and admittance functions, Y(s). The Laplace transform of the input voltage is denoted by Ein (s) and that of the output voltage is EJjs). In this case it is more convenient to use the branch currents and node voltages designated as shown in Fig. 320(a). Then one set of independent equations
representing causeandeffect relation
Ii(.s)
is
E2 (s)
/,(,)
Ea (s)
YAs)
= [E Js)E (s)]Y S = [h{s)  /,(*)]Z,(j) = [E1 {s)  Ea (s)]Y (s) = Z(j)/ (j)
i
( )
Y3 (s)
n
(s)
(a)
(*)
YAs)
Z2 (s)
Y3 (s)
work.
3 (s), and E(s) arranged from left to right graph of the network is constructed as shown in Fig. 320(b). It is noteworthy that in the case of network analysis, the causeandeffect equations that are most convenient for the construction of a signal flow graph are neither the loop equations nor the node equations. Of course, this does not mean that we cannot
With the
variables
construct a signal flow graph using the loop or the node equations. For instance, in Fig. 320(a), if we let Ii(s) and I3 (s) be the loop currents of the two loops, the loop
equations are
Sec. 3.9
73
Etn(s)
E (s)
= [Zi(s) + Z2 (j)]/,(j)  Z (s)I (s) = Z Gs)/,0) + [Z (s) + Z (s) + Z4 (s)]I = Zt(s)I (s)
2
3
(372)
3 (s)
(373)
(3 . 74)
However, Eqs. (372) and (373) should be rearranged, since only effect variables can appear on the lefthand sides of the equations. Therefore, solving for /,0s) from Ea (372) and I3 (s) from Eq. (373), we get
71
=Z
= Z
i(s)
I Z2 (s) E
^ + zjfzjs) ^
1
'
(375)
h(s)
X {?)
+ z\(s) + Z4 (*) 7
W
is
(3
 76
>
Eqs. (374), (375), and (376) are in the form of causeandeffect equations. The signal flow graph portraying these equations is drawn as shown in Fig 321 This
exercise also illustrates that the signal flow graph of a system
Now,
not unique.
(s)
Z 2 (s)
Z2 fc)
(s)
Z2 (s) + Z 2 (s)
Fig. 321. Signal flow graph of the network in Fig. 320(a) using the loop equations as a starting point.
Example
34
current in
network shown in Fig. 322(a). We shall and the voltage et) as the dependent variables of the network. Writing the voltage across the inductance and the the capacitor, we have the following differential equations:
define the current /(f)
RLC
^W =eM ~ Ri W ~ ecO
.
(377)
r dec (t) = C
~1T
'<'>
(378)
signal flow graph using these two equations since they are differential equations. In order to arrive at algebraic equations, we divide Eqs
and
(378)
by
and C,
respectively.
When we
have
sl(s)
we
i(0+)
(379)
(3 _ 80)
where i(0+) is the initial current and e (0+) is the initial c voltage at t = 0+ In these last two equations, * e (0+), i(0+), and E (s) are the input variables. There are several possible ways of constructing the signal flow graph for these equations. One way is to solve for I(s) from Eq. (379) and (*) from Eq. (380); we get
t
74
Chap. 3
R
1
L
nfw*
WWno
i(')Q )
e c (t)
(a)
EiU)
EJs)
LU + R/L)
(c)
RLC network,
flow graph.
' =
E (s) =
c
7tW
y^c(0+)
(0+)
+ as +W)] lW  WTTRim EM
(3 " 81)
+ gr/O)
(382)
graph using the last equations is drawn as shown in Fig. 322(b). graph in Fig. 322(b) is of analytical value only. In other words, we can solve for I(s) and Ec (s) from the signal flow graph in terms of the inputs, e c (0+), j'(0+), and Ei(s), but the value of the signal flow graph would probably end here. As an
The
signal flow
The
signal flow
alternative,
we can
I(s),
sEc (s)
as the noninput variables. These four variables are related by the equations
Sec
3.10
/(*)
c
=5 [*/(*)]
(383)
(384)
E (s) = s^sEds)]
The
significance of using s~
l
is that it represents pure integration in the time domain. a signal flow graph using Eqs. (379), (380), (383), and (384) is constructed as shown in Fig. 322(c). Notice that in this signal flow graph the Laplace transform variable appears only in the form of s' 1 Therefore, this signal flow graph may be used
Now,
in
computer solution of the problem. Signal flow graphs Chapter 4 as the state diagrams. 5
3.10
Given a
is
solve for
graph or a block diagram, it is usually a tedious task to inputoutput relationships by analytical means. Fortunately, there a general gain formula available which allows the determination of the inputsignal flow
its
output relationship of a signal flow graph by mere inspection. The general gain formula is
V M = na = *=1 Mj^k A
Jin
(385) '
where
yout jln
= output node variable = input node variable N = total number of forward paths Mk = gain of the kth. forward path A =  S Pmi + Pm ~ S Pmi + m m m Pmr = gain product of the mth possible combination
1 i
M = gain between y
in
and yout
(386)
of
nontouching* loops
or
A=
(sum of
all
gain products of
possible combinations of
(sum of two
(387)
nontouching loops) (sum of the gain products of all possible combinations of three nontouching
loops)
A*
the
is
which
forward path
first
may seem
formidable to use at
is
glance.
However,
A; but
in practice, systems
made with
having a large number of nontouching loops are rare. An error that is frequently regard to the gain formula is the condition under which it is valid. It must be emphasized that the gain formula can be applied only between an input node and an output node.
*Two
mon
node.
76
/ Transfer
Chap. 3
Example
35
Consider the signal flow graph of Fig. 319. We wish to find the transfer function C(s)/R(s) by use of the gain formula, Eq. (385).
There is only one forward path between R(s) and C(s), and the forwardpath gain is
M, =
2.
G(s)
is
(388)
There
is
Pu =
3.
G(s)H(s)
(389)
There are no nontouching loops since there is only one loop. Furthermore, the forward path is in touch with the only loop. Thus Ai = 1, and
A=
/,,=
G(j)#(j).
By use of Eq.
system
is
obtained as
C(s)
R(s)
= Mi At = A
G(s)
1
(390)
G(s)H(s)
result
Example
36
Consider, in Fig. 320(b) that the functional relation between Eia and E is to be determined by use of the general gain formula. The signal flow graph is redrawn in Fig. 323(a). The following conclusions
:
1.
There
is
Eia
and
as
shown
in Fig.
323(b).
The forwardpath
M, = FiZ2 y3 Z4
2.
(391)
323(c); the
loop gains
(392) (393)
(394)
/>
3.
There
is
in Fig. 323(d);
the loop
Z Y
2
and
Z4 Y
Thus
Pu =
4.
product of gains of the first (and only) possible combination of two nontouching loops = Z 2 Z4 Yx
(395)
}
0,
A= =
l 1
(J
2
{
+P +P )+P
zl
3l
2
3
3
l2
+ Z Y + Z Y + Z4 Y + Z Z4 Yt Y
2
(396)
Sec. 3.10
(a)
O
(b)
O h
(d)
Ein
graph of the passive network in Fig. 320(a). (b) and Ea (c) Three individual loops, (d) Two
.
5.
All the three feedback loops are in touch with the forward path ; thus
A,
(397)
Substituting the quantities in Eqs. (391) through (397) into Eq. (385),
we
obtain
M
Example
37
Ai
1
+Z
1 Y
+Z
r3 z2 z4 Y + Z<Y + Z1 Zt Y Y 2
y,
3 2
1
(398)
Consider the signal flow graph of Fig. 322(c). relationships between /and the three inputs,
Similar relationship
is
It is
lt
desired for
Ec
linear, the
The gain between one input and one output is detergain formula to the two variables while setting the rest of the
is redrawn as shown in Fig. 324(a). Let us first consider / The forward paths between each inputs and / are shown in
The
signal flow
graph
and
(d), respectively.
78
/ Transfer
Chap. 3
(0+)
ec
(0+)
<(0 +)
o
"
1
osl
O
/
(c)
graph of the
RLC network
Forward path between Ei and /. (c) Forward path between /(0+) and (d) Forward path between e c (0+) and /.
Sec. 3.10
79
The
signal flow
A is given by
(399)
'
A=
+#*' + 1 L LC
two loops; thus A,
for
all
cases
we have
/(0+)
= 0,
0,
* c (0+)
=
=
(3100)
,
'(0+)
e c (0+)
(3101)
<HL)sec (0+)
i(0+)
= 0,
we
E =
x
(3102)
When
all
write
(3103)
c is
considered as the
output variable,
Lc
we have
(3104)
Notice that the loop between the nodes si and / between e c (0+) and Ec
Example
38
Consider the signal flow graph of Fig. 325. The following inputoutput relations are obtained by use of the general gain formula:
yi
>i
y_3
yi
+ dy A _ ag(l + d) + A
a(i
(3105)
abc
(3106)
where
A=
+eg + d +
bcg
+ deg
(3107)
38.
80
/ Transfer
Chap. 3
3.11
Because of the similarity between the block diagram and the signal flow graph, the general gain formula in Eq. (385) can be used to determine the inputoutput relationships of either. In general, given a block diagram of a linear system we can apply the gain formula directly to it. However, in order to be able to identify all the loops and nontouching parts clearly, sometimes it may be helpful if an equivalent signal flow graph is drawn for a block diagram before applying the
gain formula.
To
illustrate
how
the signal flow graph and the block diagram are related,
shown
in Fig. 326.
since a
is
interpreted as a
(a)
(b)
(b)
Equivalent signal
flow graph.
Sec. 3.12
incoming signals to the node, the negative feedback paths in represented by assigning negative gains to the feedback paths.
The closedloop transfer function of the system is obtained by applying Eq. (385) to either the block diagram or the signal flow graph
C(s)
R(s)
Similarly,
+ G G H + G G H + G G G + G H2 +
l
G G2 G
i
j
G
X
Gj
2
3
G,G 4
(3108)
E(s) R(s)
__..
G,(? 2 ff,
+ G G H + G4 H
2
3
A
_ ~
1
(3109)
Y
where
3 (s)
+ G G H + G,H
2
3
R(s)
A
2
3
(3110)
A=
3.12
+ G^H, + G G H +
2

G,G 2 C 3
+ G4 H + G Gi
2
1
(3111)
shown in Chapter 2 that the signals in a discretedata or sampleddata system are in the form of pulse trains. Therefore, the Laplace transform and the
transfer functions defined for continuousdata systems, in the ^domain, cannot
be used adequately to describe these systems. Figure 327(a) illustrates a linear system with transfer function G(s) whose input is the output of a finitepulsewidth sampler. As described in Section 2.8,
the finitepulsewidth sampler closes for a short duration of p seconds once every T seconds, and a typical set of input and output signals of the sampler is
in Fig. 25. Since for a very small pulse duration p, as compared with the sampling period T, the finitepulsewidth sampler can be approximated by an ideal sampler connected in cascade with a constant attenuation p, the system of Fig. 327(a) may be approximated by the system shown in Fig. 327(b).
shown
r{t)
R(s)
?
(P)
*0)
G(s)
c(t)
R?(s)
(a)
C(s)
KO
R(s)
*
r*(t)
'
>*<,t)
c(0
G(s)
* *P
(s)
C(s)
Ideal sampler
(b)
Fig. 327. (a) Discretedata system with a finitepulsewidth sampl i^iaiiGiGuata aysicm wiui iiuiicpuiscwiuin sampler, (b)
an
82
/ Transfer Function
Chap. 3
,y\
d
T
c*(t)
_^_
C*(s)
"
rit)
R(s)
<*
'
>
c(t)
/*(/)
G(s)
R*(s)
C{s)
an
ideal sampler.
it is
is
included in the transfer function of the process, G(s). Therefore, the block diagram of Fig. 328 is considered as that of a typical openloop discretedata
or sampleddata system.
the system of Fig. 328. In the following
There are several ways of deriving the transfer function representation of we shall show two different representa
assume that
r*(t),
the output
of the ideal sampler 5",, is a unit impulse function. This may be obtained by or if r(t) is a unit impulse funcsampling a unit step function u,(t) once at t = tion.* Unless stated otherwise, the samplers considered in the remaining portion
of this text are ideal samplers. The output of G{s) is the impulse
is
g{i). If
fictitious ideal
sampler
S2 which
,
synchronized with Sj and has the same sampling period as that of 5,, is placed at the output of the system as shown in Fig. 328, the output of the sampler S z
may
be written as
c*(t)
g*(t)
A=
t g(kT)S(t ~ kT)
(3112)
where c(kT)
function
is
= g(kT)
is
G*(s)
[g*(f)]
= S
(3113)
g(kT)e*T
which is defined as the pulse transfer function of the linear process. At this point we can summarize our findings about the description of the
discretedata system of Fig. 328 as follows.
When
is
is
sampled by a fictitious ideal sampler S2 and the output of of the process. The Laplace transform of the
weighting sequence impulse train gives the pulse transfer function G*(s).
Although from a mathematical standpoint the meaning of sampling an impulse function questionable and difficult to define, physically, we may argue that sending a pulse through a finitepulsewidth sampler will retain the same identity of the pulse.
is
Sec. 3.12
83
Once
input
r(t),
is
the system,
and the sampled output, c*(t), which is due to any arbitrary can be obtained by means of the principle of superposition.
at
Consider that an arbitrary input r(t) is applied to the system of Fig. 328 0. The output of the ideal sampler is the impulse train,
r*(f)
= A=
T)
r(kT)S(t
kT)
is
(3114)
By means of superposition,
c(t)
due to
. .
.
r*(t), is
r(0)g(t)
+ r(T)g(t +
r{T)g[{k
+
. .
r(kT)g(t
kT)
(3115)
At
c(kT)
r(0)g(kT)
1)7]
r[(k
l)T]g(T)
+
is
r(kT)g(0)
(3116)
where
it is
assumed that
its
<
0, since
the process
a physical
system so that
Multiplying both sides of Eq. (3116) by e~ kTs and taking the summation
from k
to
oo,
we have
c(kT)e kT
=
+
r(0)g(kT)e kT
r(T)g[(k
l)T]e~ kT
...
<t
r[{k
is
l)T]g(T)e^
+ *=0 r(kT)g(0)e^
is
(3117)
simplified to
*=
c(kT)e kT <
[r(O)
r (T)e~ T
r{2T)e~>
(3118)
or
c(kT)e Ts
r{kT)e kT *
g(kT) e
 kT >
(3119)
Therefore, using the definition of the pulse transfer function, the last equation is written
C*(s)
R*(s)G*(s)
(3120)
which
is
shown
in Fig. 328.
The ztransform
relationship
is
Ts
,
Eq. (3119)
also written
c(kT)z k
r{kT)z' k
g{kT)z~ k
(3121)
G(z)
= g{kT)z' k k=0
(3122)
which implies that the ztransfer function of a linear system, C(z), is the ztransform of the weighting sequence, gikT), of the system. Equation (3121)
is
written
C(z)
R{z)G(z)
(3123)
84
/ Transfer
Chap. 3
discretedata system is important to point out that the output of the However, the pulse transform of the output, continuous with respect to time. C(z), specify the values of c{t) only C*(s) and the ztransform of the output, wellbehaved function between sampling at the sampling instants. If c(t) is a description of the true output:c{t). instants c*(0 or C(z) may give an accurate sampling instants, the zHowever if c(t) has wide fluctuations between the only at the sampling instants, will transform method, which gives information
It is
yield misleading or inaccurate results. The pulse transfer relation of Eq. (3120)
can also be obtained by use of C(s), which is given in the literature the following relation between C*(s) and
:
C*(*)
=i S
C(s+jnco
s)
O 124
=
where
co s is
From
2njT. second and ca s the sampling frequency in radians per continuousdata output c(t) Fig. 328, the Laplace transform of the
is
C(s )
G(s)R*(s)
(3125)
C*(s)
G(s
+ jnaJR^s + jnw,)
+
(3126)
We
can write
R*(s
^
(3127) (3128)
E
fc
r(kT)e kTs
A:
and
n,
jnkTo>,
j2xnk
Thus Eq.
(3127)
becomes
R*(s
+ jnco =
s)
J*(5)
(3
" 129
Using
this identity,
Eq. (3126)
C*(s)
is
simplified to
= i{*(4 S
= J?*WG*(*)
=
G(5+7t
s)
(3130)
C*(j)
3431)
where
G*(j)
4 S
J
G^+jnco,)
(3132)
= oo
in z of Eq. (3123) e
Ts
.
by use of z
is
In conclusion,
the input to a linear system is sampled Laplace transform of the continuous output unsampled, the
we
note that
when
given by
C(s)
G(s)R*(s)
is
(3133)
If the continuousdata
output
is
synchronized with
Sec. 3.12
85
and has the same sampling period as the input, the Laplace transform of the discretedata output is given by
C*(s)
G*(s)R*(s)
it is
(3134)
The
is
is
natural, since
The expression
it
in Eq. (3134)
can be interpreted as being obtained directly from Eq. (3133) by taking the pulse transform on both sides of the equation. In other words, in view of Eq. (3129), we can write, from Eq. (3133),
C*(s)
= =
[Gis)R*(s)]*
(3135)
G*(s)[R*(s)]*
R*(s)
(3136)
The
elements
is slightly more involved than that for continuousdata systems, because of the variation of having or not having any samplers in between the elements. Figure 329 illustrates two different situations of a discretedata
system which contains two cascaded elements. In the system of Fig. 329(a), the
is
"2
(0
T
d(t)
D* R*(s)
d*(t)
c(t)
(s)
D(s)
D*(s)
G 2 (s)
C(s)
w)
(0
C*Cs)
Kt)
R(s)
**
r*(t)
Gi(s)
dU)
D(s)
(b)
R*(s)
G 2 (s)
~
C(s)
c(t)
and sampler
separates the
two elements,
86
Chap. 3
same period as the sampler Si. The two elements with transfer functions and G 2 (s) of the system in Fig. 329(b) are connected directly together. In discretedata systems, it is important to distinguish these two cases when deriving
the
(s)
Let us consider
first
is
written
(3137)
D(s)
Gi(s)R*(s)
is
C(s)
=G
(s)D*(s)
(3138)
C(s)
=G =
(s)GKs)R*(s)
last
(3139)
equation gives
(3140)
in Eq. (3136).
is
The corresponding
=G
2 (z)Gi(z)R(z)
(3141)
equal to the product of the ztransforms of the two individual transfer func
tions.
in Fig. 329(b) is
Gi{s)G 2 (s)R*(s)
is
(3142)
last
equation
C*(s)
[Gi(s)G 2 (s)]*R*(s)
(3143)
where
[Gi(s)G 2 (s)]*
= i
f)
G (.s+jmo.yG 1 (.s+jna>.)
1
(3144)
C?i(s) and G 2 (s) are not separated by a sampler, they have to be treated as one element when taking the pulse transform. For simplicity, we
[Gi(s)G 2 (s)]*
= GiG^s) = G GHs)
2
(3145)
Then Eq.
(3143)
becomes
C*(s)
= =
GiG^(s)R*(s)
(3146)
C(z)
1
GiG 2 (z)R(z)
(3147)
where G G 2 (z) is defined as the ztransform of the product of Gi(s) and and it should be treated as a single function.
G 2 (s),
Sec. 3.12
87
It is
G Gt(s)^Gf(s)GKs)
t
(3148)
and
G.G^z)
Therefore,
^ G^Gziz)
(3149)
we conclude
is
sampler in between
In this section the transfer functions of simple closedloop discretedata systems are derived by algebraic means. Consider the closedloop system shown
in Fig. 330.
is
G(s)E*(s)
(3150)
"X,
c*(t)
t)
R(s)
^
_\y
!
e(t)
V
j.
C*(s)
c(t)
e*(t)
G(s)
E(S )
E*(s)
C(s)
H(s)
error function
is
R(s)
H(s)C(s)
(3151)
E(s)
R(s)
G(s)H(s)E*(s)
(3152)
E * (S) _
into Eq. (3150);
R*(s)
(3153)
is
C(s)
1
G(s)
GH*(s)
,R*(s)
(3154)
on both
^TTmwr
G*(s)
{s}
(3155)
88
/ Transfer
Chap. 3
In this case
it is
GH*(s)
(3156)
The
is
C(z) R(z)
_ G{z) ~ + GH(z)
1
(3 " 157)
We
shall
show
although
it
is
possible to define a
transfer function for the closedloop system of Fig. 330, in general, this
may
not be possible for all discretedata systems. Let us consider the system shown in Fig. 331. The output transforms, C(s) and C{z), are derived as follows:
C(s)
E(s)
= =
G(s)E(s)
R(s)
(3158)
(3159)
H(s)C*(s)
~>"
C*(t)
KO
R(s)
(%
~?)
e(t)
c(t)
G(s)
\J
i
E(s)
C(s)
c*(t)
Hds)
C*(s)
C{s)
G(s)R(s)
G{s)H(s)C*{s)
the last
(3160)
sides of
GH *{s) + **L
(3161)
Note that the input and the transfer function G(s) are now combined as one function, GR*(s), and the two cannot be separated. In this case we cannot define a transfer function in the form of C*(s)/R*(s). The ztransform of the output is determined directly from Eq. (3161) to be
C(z)
= r GR(z) GH(z) = =
#[G(*)fl(s)]
(3162)
where
it is
GR(z)
(3163)
(3164)
and
GH(z)
g[G(s)H(s)]
Chap. 3
Problems
89
To determine
we
substitute
We
have
(3 " 165)
G(s)R(s)
GR * (s)  G +gh*\s)
t
and
without
may become
tedious.
and 33 1 by algebraic means more complex system configurations, the algebraic method The signalflowgraph method is extended to the analysis
may
REFERENCES
Block Diagram and Signal Flow Graphs
1.
T. D. Graybeal, "Block
70, pp. 985990, 1951.
2.
S. J.
Properties of Signal
Sept. 1953.
Flow Graphs,"
Flow Graphs,"
S. J.
Mason, "Feedback Theory Further Properties of No. 7, pp. 920926, July 1956.
Signal
L. P. A.
Robichaud, M. Boisvert, and J. Robert, Signal Flow Graphs and Englewood Cliffs, N.J., 1962.
B. C.
New
York, 1967.
6.
N. Ahmed, "On Obtaining Transfer Functions from GainFunction Derivatives," IEEE Trans. Automatic Control, Vol. AC12, p. 229, Apr. 1967.
B. C.
Hall, Inc.,
8.
Kuo, Analysis and Synthesis of Sampled Data Control Systems, Englewood Cliffs, N.J., 1963.
Discrete
Prentice
B. C.
Kuo,
Champaign,
Illinois, 1970.
PROBLEMS
3.1.
The following
where
r(t)
(b)
*%P +
104g>
2cit)
r(
2)
3.2.
The block diagram of a multivariate feedback control system P32. The transfer function matrices of the system are
is
shown
in Fig.
90
Chap. 3
G0)
0"
1.
HO)
.0
Rfe)
y
i
/\
C(s)
,
G(s)
H(s)
Figure P32.
3.3.
is
shown
in Fig. P33.
C2 (s)
Ci(j)
C2 (s)
*i(*)
Ri(s)
G(s)R(s)
R 2 (s)
Figure P33.
3.4.
Draw
set
of algebraic equations:
+ x 2 + 5x = + 2x 2 4x = 2 x 2 x =
3 3
Chap. 3
Problems
91
3.5.
Draw an equivalent
signal flow
in Fig. P35.
Find
Figure P35.
3.6.
Find the
P36.
gains,
y 6 /yi,
y%ly\,
signal flow
Figure P36.
3.7.
Find the gains y<,\y x and y 2 lyi for the signal flow graph shown
0.5
in Fig. P37.
Figure P37.
92
/ Transfer
Chap. 3
3.8.
In the circuit of Fig. P38, e s (t), e^t), and i,(t) are ideal sources. Find the value of a so that the voltage e (t) is not affected by the source ed (t).
o+
Figure P38.
3.9.
in Fig. P39(a)
and
Oyj
(a)
(b)
Figure P39.
3.10.
Given the signal flow graph of Fig. P310(a) and the transfer functions G it G 2 G 3 G 4 and G 5 find the transfer functions GA GB and (7C so that the three systems shown in Fig. P310 are all equivalent.
, ,
,
(a)
Figure P310.
Chap. 3
Problems
93
(b)
(c)
Construct an equivalent signal flow graph for the block diagram of Fig. P311. (a) Evaluate the transfer function C/R when 0. (b) Determine the relation among the transfer functions G u G 2 , G 3 G4 u and 2 so that the output C is not affected by the disturbance signal N.
N=
Figure P311.
3.12.
multivariate system
is
relations
C(s)
S(.y)
where
C(s)
= c^sy
i
R^sT
Ri(s).
G(s)
+
_1_
H( S)

'I
0" 0_
.0
94
/ Transfer
Chap. 3
(a)
= +
M(.$)RO)
G(j)H(5)]'G(i)
by using
M(s)
(b)
[I
Draw a signal flow graph for the system and find M(s> from the signal flow graph using Mason's gain formula.
in Fig. P313.
3.13.
Find the transfer function relations C(s)/R(s) and C(s)jE(s) for the system
shown
Figure P313.
3.14.
P314.
Find the transfer function C(z)/R(z) of the discretedata system shown The sampling period is 1 sec.
r(t)
in Fig.
^
T
r*(t)
s(s
c(t)
1
+ 2)
Figure P314.
3.15.
in
KO
^
^
T
r*(.t)
s
cd)
1
2
s
(a)
r(t)
r*(t)
^
s+
X1
c(t)
s +
(b)
Figure P315.
4
StateVariable Characterization
of
Dynamic Systems
4.1
In Chapter 3 the classical methods of describing a linear system by transfer function, impulse response, block diagram, and signal flow graph have been presented. An important feature of this type of representation is that the
system
relations.
For
fer function describes the inputoutput relation in the Laplace transform domain. However, the transform method suffers from the disadvantage that all
when one is intimedomain solution, which depends to a great deal on the past history of the system, the transfer function does not carry all the necessary
information. Transfer function
js
The greatest advantage of transfer compactness and the ease that we can obtain qualitative information on the system from the poles and zeros of the transfer function.
function
is
An
is
method of describing a
It
linear system
The
statevariable representation
to linear systems
The
statevariable
method
is
often referred to as a
modern approach.
However, in reality, the state equations are simply firstorder differential equations, which have been used for the characterization of dynamic systems for many years by physicists and mathematicians.
To
As the word
96
StateVariable Characterization of
Dynamic Systems
Cna P 4
the past, present, and future conditions of the system. It is interesting to note that an easily understood example is the "State of the Union" speech given by
the President of the United States every year. In this case, the entire system encompasses all elements of the government, society, economy, and so on. In
general, the state can be described
of numbers, a curve, an equation, or something that is more abstract in nature. From a mathematical sense it is convenient to define a set of state variables and state equations to portray systems. There are some basic ground rules regarding the definition of a state
by a
set
variable
state equation.
set
of variables,
characteristics of a ., x(t) is chosen to describe the dynamic Xl (t), x 2 (t), the state variables of the system. Then system. Let us define these variables as these state variables must satisfy the following conditions
1.
At any time
define the
),
x 2 (t a ),
x(t
initial states
initial time.
2.
and the initial states defined t for t above are specified, the state variables should completely define the future behavior of the system.
>
Therefore,
Definition
set
we may
of state variables. The state variables of a system are defined as ., x n (t) such that knowledge of these of variables, x^t), x 2 (t), plus information on the input excitation subsequently variables at any time t t > 1 applied, are sufficient to determine the state of the system at any time
a minimal
.
.
confuse the state variables with the outputs of a system. An output of a system is a variable that can be measured, but a state variable does not always, and often does not, satisfy this requirement. However, an
output variable
is
R
^AM
e(f)
1
Example
41
state vari
RL
i(0
fJ
completely specified by the initial current of the inductance, 0, a constant input voltage of ampli0. At / j'(0+), at t tude Ei is applied to the network. The loop equation of the
network for
Fig. 41.
>
is
RL
network.
^=R
(t)
+ L dm
we
get
E(S ) =*h.
(42)
m~
.,
s{R+Ls)'r
E,
1/(0+)
R+Ls
(43)
The
current
/(/)
for
is
Sec. 4.2
Dynamic Equations
97
We have
'(0
= ^ (1  e1 +
)
i(0+)e R,/L
(44)
Once
is
determined for
> 0,
is
apparent that the current i(t) in this case satisfies the basic requirements as a state variable. This is not surprising since an inductor is an electric element that stores kinetic energy, and it is the energy storage capability that holds the information on the history of the system. Similarly, it is easy
interval. Therefore,
it is
same time
to see that the voltage across a capacitor also qualifies as a state variable.
4.2
The
firstorder differential
equation of Eq.
(41),
between the
e{t),
This firstorder differential equation is referred to as a state equation. For a system with p inputs and q outputs, the system may be linear or nonlinear, time varying or time invariant, the state equations of the system are
written as
^
i
=fix
(f),
x 2 (t), ...,
r p (t)]
(46)
=
. .
.
1,2, ... ,n
,
xn (t)
/,(/),
r 2 (t),
rp {t) are
functional relationship.
. .
.
The outputs of the system c k (t), k = 1, 2, q, and the inputs through the output equation,
,
c k {t)
&[*,(*),
x 2 (t),
...,
x(t),
rM
r 2 {i),
..., r p (t)]
(47)
k=\,2,...,q
where gk denotes the kth functional relationship. The state equations and the output equations together form the set of equations which are often called the dynamic equations of the system.
contain only the
Notice that for the state equations, the left side of the equation should first derivatives of the state variables, while the right side should have only the state variables and the inputs.
Example
42
Consider the
tional
RLC
network shown in Fig. 42. Using the convennetwork approach, the loop equation of the network is
written
e{t)
Ri(t)
+L^ + [
equation
is
i{t)
dt
(48)
Fig. 42.
RLC network.
is not in the form of a state a time integral. One method of writing the state equations of the network, starting with Eq. (48), is to let the state variables be defined as
We notice that
this
98
StateVariable Characterization of
Dynamic Systems
Chap. 4
*i(0
Xl(t)
= iff)
=
f /(/) dt
(49)
(410)
= Rx (t)+L
1
we have
+ ^x 2 (t)
(411)
of Eq. (410),
Rearranging the terms in Eq. (411) and taking the time derivative on both sides we have the two state equations of the network,
^
}
= T*'<'>nr*<'> + r*>
*,(/)
(4 " 12)
^ =
which are
(413)
We have demonstrated how the state equations of the RLC network may be written from the loop equations by defining the state variables in a specific way. The objective, of course, is to replace Eq. (48) by two firstorder differential equations. An alternative approach is to start with the network and define the state variables according to the elements of the network. As stated in Section 4.1, we may assign the current through an inductor and the voltage across a
capacitor as state variables. Therefore, with reference to Fig. 42, the state
variables are defined as
*i(o
*2
= (0 =
m
efy)
(4i4)
(415)
Then, knowing that the state equations would have to contain the first derivatives of x^t) and x 2 (t) on the left side of the equations, we can write the equations directly
by inspection from
Voltage across L:
Fig. 42
^p =
Ri(t)
e e (t)
<?(0
(4
"
16 )
Current in C:
C^2 =
at
i(t)
(417)
sides of the last
Using Eqs. (414) and (415), and dividing both by L and C, respectively, we have
two equations
**M =  Xl
(f)
 i*
(0
e(t)
(418)
4*M =
^ Xi
(f)
(419)
which are the state equations of the RLC network. We notice that using the two independent methods, the only difference in the results is in the definition of the second state variable x 2 (t). Equation (419) differs from Eq. (413) by
the factor of the capacitance C.
two simple examples we see that systems, the state equations can generally be written
From
these
Sec. 4.3
99
dxt)
dt
% a u xj(0 +
d
1
2l b ik r k (i)
i=l,2,....,n
(420)
coefficients.
are written
(421)
kJ Xj(t)
j=
+ m^ t,
e km r m {i)
1
k=l,2,...,q
where dkj and ekm are constant coefficients. For a linear system with timevarying parameters, the (420) and (421) become time dependent.
4.3
coefficients of Eqs.
The dynamic equations are more conveniently expressed us define the following column matrices
"*i(0'
x 2 (t)
x(r)
(n
1)
(422)
where
x(t)
is
>i(0"
r 2 (t)
r(/)
(px
1)
(423)
WO.
where
r(/) is
"c,(0"
cz(t)
c(0
(qX
_c,(0_
1)
(424)
where
c(r) is
Then
can be written
M)=f[x(0,r(0]
where
f
(425)
denotes an n
f as elements,
x 1 column matrix that contains the functions fu f2 and the output equations of Eq. (47) become
c(r)
g[x(r), r(r)]
(426)
where g denotes a q x
column matrix
g u g 2 ,...,
gq
as elements.
100
StateVariable Characterization of
Dynamic Systems
Chap. 4
For a
State equation
^
c(/)
fl
Ax(0
+ Br(r)
(427)
Output equation:
where
= Dx(0 + Er(/)
(428)
A is an
a 11 d 21
12
a, "In
#22
a 2n
(429)
A=
a,
B is
an n X
p matrix given by
t>u
b 12 b2 2
a.;
bn
B
Pi
b 2
(430)
D is
a q
n matrix given by
~d u
"21
d i2
"22
du
d2
(4.31)
D=
dql
and
dq2
i2 e 12
"2p
is
a q
p matrix,
'e
en
E =
Zql
(432)
C2
Example
43
The
and
dxi(ty dt
r
'
R L
j_
V
'
\
r*iw"
+
\x
{t)
L
e(0
(433)
dx 2 (t)
L dt J
be
_R
A=
L
J_
L
(434)
Sec. 4.4
B= L
(435)
4.4
is
homo
^
Let
<(f)
= Ax
(436)
be an n
x n matrix
equation
it
must
satisfy the
^
x(t)
= AKO = 0; then
<f>(t) is
(437)
also defined
Furthermore,
let
initial state at t
<Kr)x(0+)
state equation for
t
(438)
which
is
homogeneous
fy(t) is
>
0.
sX(s)
x(0+)
= AX(s)
get
(439)
we
(440)
last
equation yields
/
x(0
, [(si
A)" ']x(0+)
>
is
(441)
identified
Comparing Eq.
to
be
(>(0
 [(jIA)']
1
(442)
An
alternative
way of
solving the
'x(0+)
series
(443)
for
> 0, where e
At
represents a
AI
power
4
=1+
At
^A*/ 2
is
+1A
(444)*
It is
easy to
show
a solution of the
homogeneous
state
^=
*It
Ae Kt
uniformly convergent.
(445)
this
power
series is
102
StateVariable Characterization of
Dynamic Systems
Cha P 4
expression for
eK<
=I+
At
+ ~A
2 2 /
+ 1A
(446)
directly
from Eq.
(442).
This
is left
as
an
Example 44
Consider the
that
is,
RL
network of
e(t)
written
(447)
*=*)
The solution of the last equation for t Thus
i(t)
>
is
= 0.
(448)
this case is
is
given by
(449)
0(0
which
is
e*" L
>
Since the state transition matrix satisfies the homogeneous state equation,
it
it
governs the
response that
excited
by the
initial
and
matrix
As
implies, the state transition matrix <f>(0 completely defines the transition of the states from the initial time t to any time t.
the
name
The
state transition
1.
matrix
<J>(0
<f)(0)
(450)
Proof:
f
0.
2.
0(O
Proof:
4(O
Ac
,
(451)
we
get
<K0e"
At
e x 'e A '
=
<J>
I
_1
(')>
(452)
by
we 8 et
(453)
e A '
(f,i( f )
<Kr)
<f~'(0
e~ At
is
(454)
An
interesting result
from
this property
of
<f>(t)
arranged to read
x(0+)
<J>(0x(0
(455)
Sec. 4.5
State Transition Equation /
103
time.
which means that the state transition process can be considered as bilateral in That is, the transition in time can take place in either direction.
3.
<J>(7 2
0<K>i
<f>(t 2
for any
tQ
t2
(456)
Proof:
<K'2
'i)<K'l
t )
= A e A( "') =e = M2 <'">
A(r,)
(4.57)
t )
This property of the state transition matrix is important since it implies that a state transition process can be divided into a number of sequential transitions. Figure 43 illustrates that the transition from t t Q to t t 2 is
equal to the transition from t to t and then from U to t 2 In general, of course lt the transition process can be broken up into any number of parts Another way of proving Eq. (456) is to write
.
) ) )
(458)
(4.59)
(4 _ 60)
The proper
4
result is obtained
result
by substituting Eq.
(459) into
comparing the

Eq
(458)
and
MO]" =
Proof:
[<t>(t)Y
<Kkt)
for
= integer
(k terms)
(461)
oAr
e kAt = Mt)
4.5
(462)
The
geneous
RL
input voltage
is
>
0.
104
StateVariable Characterization of
Dynamic Systems
Chap. 4
&j& = Ax(0 +
can be solved by using either the
presented in the following.
classical
Br(r)
(463)
we have
(464)
x(0+)
AX(s)
BR(s)
t
initial state
vector evaluated at
0+. Solving
for
(si
A)" 'x(0+)
(si
A)" 'BR(s)
(465)
The
state transition
is
"'[(5l
A)" ]x(0+)
1
JC'Kil
A) BR(s)]
1
(466)
written
<K0x(0+)
+ Jf
<f>(?
T)Br(r) dx
is
>
when
(467)
The
time
is
state transition
useful only
the initial
defined to be at
0.
In the study of control systems, especially disoften desirable to break up a state transition
it
is
more
ta
flexible initial
time must
initial
initial
time be represented by
r(f) is
t
by
x(/
),
We
start
applied for
,
>
r.
get
x(0+)
<J>(/ )x(/o)
<(>(f)
<K'o)
f"
Wo 
T)Br(t) di
(468)
x(0
(J>(0<K>o)x('o)
,,
<KO<K'o) Jf<Wo
T)Br(T)rfr
(469)
+
J
<(>(? T)Br(r) di
last
Now
two
integrals,
<f>(t
)x(t
+
is
<j)(t
~ r)Br(r) dx
i
(470)
0.
Once
initial state
from Eq.
and the input vector simply by subThus the output vector is written
c(0
Dcf>('
)x(t
f ' D<f>('
T)Br(r) dx
'
Er(f)
(471)
Sec. 4.5
State Transition Equation
/
105
Example
45
dt
dx 2 (t)
dt
+
2 *2(0
1
#(/)
(472)
The problem
t
is
> 0; that
is,
r(t)
x(0 for t when the input r(/) u s (t). The coefficient matrices A and B are identified to be
>
for
r
_
~0~
3.
"
B=
_1_
1"
(473)
Therefore,
0~
si s
~s
1
s
"1
_2
3.
_2
3_
The matrix
inverse of (si
A)
is
(si A)'
The
state transition matrix of
+
2
1"
3s
(474)
5
A is found
=
is
Thus
2e~'
state transition equation for
t
e' 1 + 2e~ 2
'
'
2t
(475)
>
x(0
B and
'
e~ 2 '
x(0+)
2e~ 2 \
(it)
<
2fi('')
e 2(tt)
__
2<t)
2g(rr)
2 e 2('r)
_ e (fT)
g'
<
__2e 2
~2'
'
(''>
rfT
(476)
x(0
2e~'
2c"'
+
c
c 2 '
2e~ 2

e<
x(0+)
2e" 2 '.
(477)
+
As an
alternative, the
<?'
e 2t
>>0
second term of the state transition equation can be obtained by  A)'BR(s). Therefore
A)iBR(s)]
~ Ksl1
=i
's

1"
2
s.
= '

J_l
1
3s
(478)
1_
2>
/>0
106
StateVariable Characterization of
Dynamic Systems
Chap. 4
Example
46
In this example
we
method
RL
is
as
shown
in Fig. 44.
IE
e(t)
waveform
is
for the
network
in Fig. 41.
The
state
rf
= r /(/ + rw
)
(479)
Thus
A
The
state transition matrix
is
R L
#(0
B
e~ R " L
(480)
(481)
One approach
voltage as
to the
;'(?)
for
>
is
Eu s (t)
Eu
(t
?,)
is
(482)
where u s (t)
is
(s)=^ (!+*")
Then
(si
(483)
 A)'BR(j)
s
E
s
R/L L
d+e"')
(484)
*
Rs[l
E + (L/R)s] (1
t
+<?'")
is
>
obtained:
e R "H(0+)u s (0
+ d 
e"*)u,(t)
(485)
e Mw]u.(tt 1 )
state transition
t
=
is
to
t,,
and
to
<
<,
t,,
the input
e(t)
Eu,(t)
0<t <
r,
(486)
Then
(487)
j
Rs[l
+ (L/R)s]
Sec. 4.6
Relationship
107
< <
t
is
i(0
(488)
Substituting
tx
i( tl )
e Ru/L i( Q
+ ) + j (1 
(489)
The value of
period of
tt
i(t) at t
ti is
now
used as the
initial state
< < co. The magnitude of the input for this interval
t
is
2E. Therefore,
i(0
>
t,
(490)
where
In the
i(ti) is
This example illustrates two possible ways of solving a state transition problem.
first
is
in
more
approach requires only one operation, the second and it often presents computational advantages. Notice that in the second method the state at / = t is used as the initial state for the next transition period, which begins at t
first
Although the
method
x
4.6
In preceding sections
we
is
always possible to
from the schematic diagram of a system, in practice the system may have been described by a highorder differential equation or
write the state equations
transfer function. Therefore,
it is
necessary to investigate
how
state equations
can be written directly from the differential equation or the transfer function. The relationship between a highorder differential equation and the state equations
is
de
dn
~l
c(i)
d"~ 2 c(t)
dc(i)
,..
,,..
lA ,,.
where
c(t) is
is
the input.
The problem
the output c(0
by n
state equations
and an output
equation. This simply involves the defining of the n state variables in terms of
and
its
derivatives.
We
have shown
we
convenient
state variables as
met.
108
StateVariable Characterization of
Dynamic Systems
Chap. 4
it is
x,(0 *a(0
c(t)
dt
(492)
*(o
fife""
Then
dx z (t)
dt
=*
(0
(493)
<fai(0
<//
xJLO
dx(t)
dt
= a x,(0 B
a_,x 2 (0
...
a 2 Jfi(0
i^(0
+ KO
where the last state equation is obtained by equating the highestordered derivative term to the rest of Eq. (491). The output equation is simply
c{t)
x,(0
is
(494)
written
Br(t)
= Ax(0 +
and
(495)
where
x(t) is the
state vector
r(i) is
The
coefficient
matrices are
1
n)
(496)
a
0"
ai a
n 2
A,3 fl_4
(xl)
(497)
Sec 4 7

in vectormatrix
form
is
c(0
= Dx(r)
0]
(1
(498)
D=
The
section.
state
[1
...
ji)
(499)
Example
47
(4100)
is
equated to
The
*i(')
= c(0
(4102)
x.(0=*>
**<*)
~ ~^mt) dt*
d 2 c(
Then
with
the state equations are represented by the vectormatrix equation of Eq. (495)
"010
1
(4103)
2
and
1
5
B=
The output equation
is
(4104)
c(t)
= Xl (t)
(4105)
4.7
when
and
are given
is
by Eqs.
(496)
and
sys
canonical form. It
is
tem with
single input
shown in the following that any linear timeinvariant and satisfying a certain condition of controllability
(see
Theorem
given by
41.
4*j&
= Ax(0 + Brit)
(4106)
110
StateVariable Characterization of
Dynamic Systems
Chap. 4
where x(t)
is
an n
1 state vector,
r(t)
an n
a scalar
[B
input. If the
an n X 1
S
is
AB A 2 B
...
Anl B]
(4107)
= Px(r)
(4108)
or
x = P'yCO
which transforms Eq. (4106) to the phasevariable canonical form
(4109)
y(0
where
1
=A
y(0
B,r(0
(4110)
...
1 1
... ...
A,
(4111)
a
and
(4112)
is
given by
Pt
PA
t
(4113)
_P A"1
where
P,
Proof: Let
[0
1][B
AB A B
2
A" B]]
(4114)
Xi(f)
x(0
=
_*.(0
(4115)
Sec. 4.7
>i(0'
y(0
=
(4116)
and
U.(0.
Pll
Pl2
Pin
Pin
P.
Pi
PlZ
P=
Pnl
(4117)
Pnl
where
Pnn.
P =
<
[Pa
Pa
P in ]
1, 2,
(4118)
=P
il
(t)
+ p 12 x
2 (t)
...+ Pu xn(t)
last e 4 uation
tim e
um
, and
J ^"
1
n b th sMeS f the
and
(4111),
in view of
M*)
=y
5
2 (t)
P,x(0
P,Ax(0
pJ=
Eq
TSre^
^^
Mt)
+ p.BKO
f
(4120)
in
'
"
fUnCti n
0nly
Ji(0 = y7.it) = V Jaif) Taking the time derivative of the last equation once again leads to
1
(4 _ 12 i)
with
P AB =
t
=y
3 (t)
P,A 2x(?)
(4122)
0.
= >'.(0 = PiA"'x(0
we have
Pt
(4123)
with
P,A*B
0.
P.A
y(f)
Px(0
=
PjA"1
x(0
(4124)
or
P,
P.A
P=
.Pi A"'.
(4125)
112
Dynamic Systems
Chap. 4
and P, should
satisfy the
condition
PjB
= P,AB =
Px(0
P,A" 2 B
=
we
get
(4126)
PAx(f)
+
'
PBr(0
(4127)
Comparing Eq.
we
obtain
(4128)
A,
= PAP
and
B = PB
Y
(4129)
P,B
'
'0'
P AB
t
PB =
_P,A" B.
1
__
(4130)
_1
Since Pj
is
an
P,[B
AB A 2 B
...
A"B]
=
2
[0
...
1]
(4131)
Thus P!
is
obtained as
P,=[0
...
1][B
1
AB A B
...
A^B]
if
2 A""'B] is nonsingular. This is the the matrix condition of complete state controllability. Once P! is determined from Eq. (4132), the transformation matrix P is given by Eq. (4125).
(4132)
Example
48
1
1
(4133)
It is
desired to transform the state equation into the phasevariable canonical form.
S
is
[B
AB]
ri
'
o~
(4134)
1
may
Therefore, Pi
obtained as a
be expressed in the phasevariable canonical form. contains the elements of the last row
of
S 1
that
P,=[l
Using Eq.
(4125),
1]
(4135)
(4136)
_PjA_
Sec. 4.7
Thus
A,
=PAP'
= PB = ro
i
"0
1
(4137)
B!
(4138)
The method of defining state variables by inspection as described earlier with reference to Eq. (491) is inadequate when the righthand side of the differential equation also includes the derivatives of r(/). To illustrate the point we consider the following example.
Example
49
Given the
differential
equation
^^ +
~dk + ~dT +
L
2c
^^dr +
2r(t)
(4139)
it is desired to represent the equation by three state equations. Since the right side of the state equations cannot include any derivatives of the input r(t), it is necessary to include /(?) when defining the state variables. Let us rewrite Eq. (4139) as
d3c(0
dr(t)
,d*c(t)
dc(t)
,,
,
(4140)
The
now
defined as
*i(0
= c(t)
(4141)
x2
(0=^
(4142)
(4143)
Using these
last three
^
dx 3 (t)
dt
=**(
= *.(0+r(0
= ~ 2x iW ~
,
.
(4144)
~3J~
In general,
it
x 2(0
5x 3 (t)
3r(f)
differential
equation
dt
+
'
dr>
+a"dt"
~dt
+
...
a c
W
1
*.^ +
the state variables should be defined as
+ *._
^ +M0
dt
(4145)
114
StateVariable Characterization of
Dynamic Systems
Chap. 4
x,(t)
c(t)
 V(0
x
x 2 (t)
xAt)
= ^f>h
r{t)
= **M 
h 2 r{t)
(4146)
*(0
^^A
..'(0
where
A,
h2
>h
= a,6 = {b a b = (* a b  * A bj,
flj/71
aJh
(4147)
(b
ab
a_ x h x
a_ 2 h 2
a
2 hn
_2
aj\ n _ x
differential
equa
^
d
x 2 (r)
3 (t)
+
+
MO
h 2 r(t)
^f>
=x
(4148)
^M =
^1 =
xm (t)
ax,(t)
is
...
+
3
fl 2
x._,(0
first
a iX {t)
+ hr(0
(4149)
x,(0
r{t)
49,
we have
/j,
When we
and
(4147),
we have
Sec. 4.8
115
the same results for the state variables and the state equations as obtained in Example 49. The disadvantage with the method of Eqs. (4146), (4147), and (4148) is that these equations are difficult and impractical to memorize. It is not ex
pected that one will always have these equations available for reference.
ever,
How
we
function.
4.8
We
have presented the methods of describing a linear timeinvariant system by transfer functions and by dynamic equations. It is interesting to investigate the relationship between these two representations.
In Eq. (33), the transfer function of a linear singlevariable system
is
Eq. (316) gives the matrix transfer function relation for a multivariate
p inputs and q outputs. Now we shall investigate the transfer function matrix relation using the dynamic equation notation.
system that has
is
^
c(/)
= Ax(0 + =
X
X X
Dx(t)
1
Br(?)
(4150)
(4151)
+ Er(/)
where
x(/)
=n r (0 = P =q c
(t)
state vector
input vector
output vector
and A, B, D, and E are matrices of appropriate dimensions. Taking the Laplace transform on both sides of Eq. (4150) and solving
for X(s),
we have
X(s)
(si
 A)" >x(0+) +
(4151)
is
(si
~ A)'BR(s)
(41 52)
Q = DX(j) + ER(s)
Substituting Eq. (4152) into Eq. (4153),
(4153)
we have
C(s)
D(sl
A)x(0+)
D(jI
 A)'BR(5) + ER(s)
(4154)
Since the definition of transfer function requires that the initial conditions be set to zero, x(0+) 0; thus Eq. (4154) becomes
C(s)
(41 55)
defined as
(4156)
which
is
a q
x p
A)
is
matrix.
Of
matrix (si
nonsingular.
116
StateVariable Characterization of
Dynamic Systems
Chap. 4
Example 410
is
+4
dt
dt
c1
iCl
(4157)
^+^ +
The
state variables of the
+2c 2
=/ 2
(4158)
Ci
(4159)
x2 x3
These
state variables
= dci ^ =
c2
(4160)
(4161)
differential
equations, as
no
particular reasons for the definitions are given other than that these
Now equating the first term of each of the equations of Eqs. (4157)
the rest of the terms
(4161),
and (4158) to
form:
we
in matrix
dxi
dt
xi
dx 2
dt
04
1
ri
_o
x2
x3
ri
1 Li z J
(4162)
dx 3
ldt_
pi~
_C2.
1
01
1_
2
~Xi~
x2
_*3_
Dx
(4163)
To
formulation,
determine the transfer function matrix of the system using the statevariable we substitute the A, B, D, and E matrices into Eq. (4156). First, we form
A),
~s
1
j
(sIA)=
1
+4
1
3
s
(4164)
+
s
2_
The determinant of
Thus
(si
A) is
jIA =
s3
+6s 2 +
11
1]
+
s
(4165)
~s z
(.51
A)
 jI _ A
is,
+6s + 3
(*
case,
4)
+2 s(s + 2) (s + 1)
" f41
fifi^
3i
s(s+4)_
The
in
tllis
GCs)
= T>(sl ~ S* +
A)'B
"
1
s+2
3
!)
(4167)
6s 2
+ Us +
initial
_(
i(5
+ 4)J
The
resulting transformed
Sec. 4.9
and Eigenvectors
117
sis
_ s
+ 4) +
1
3
s
nrc,(j)i
2 is)j
+ 2j[_C
obtain
= VR^s)'
\_R 2 (s)_
(4168)
we
C(.r)
G(j)R(i)
(4169)
~
where
Sis
Gis)
_ s
+ 4) +1
3
s
+ 2_
(4170)
is
obtained
carried out.
4.9
The
an important part in the study of from the basis of the differential equation, the
is
linear trans
df
df
dF^
+
,
+
,
an \
dc + ac
,
(4171)
'df
df =
1,2, ... ,m
By
p
P
as
k
dt
Mi
A;
(4172)
Eq. (4171)
is
written
a 2 pn
..
=
Then the
s"
ib P"
+
1
*,/>""'
+ +
(4173)
defined as
+ a^" +
a 2 s" 2
o,J
an
=
s.
(4174)
which
is
setting the
is
operator p
The
G(s) {S)
...
. .
+ b . iS + b + a _ lS + a
n
n
u (4
'
175)
is
to zero.
From
we can
G( , )
= D adj(5lA) B + E si A
I 
D[adj(sl
A)]B
jl
A
(4176)

UIAI
118
/ StateVariable Characterization of
Dynamic Systems
Chap. 4
we
get
*I
A =

(4177)
which
is
Eigenvalues
The
roots of the characteristic equation are often referred to as the eigenIt is interesting
in
the last
row of
the
That is, if A is given by Eq. (496), the characteristic equation is readily given by Eq. (4174). Another important property of the characteristic equation and the eigenvalues is that they are invariant under a nonsingular transformation. In other words, when the A matrix is transformed by a nonsingular transformation x = Py, so that
elements of the
A matrix.
A = P"'AP
then the characteristic equation and the eigenvalues of of A. This
is
(4178)
proved by writing
si A = ^IP'AP
or
(4179)
si A
The
characteristic equation of
= ^""PP'AP
(4180)
\sl
A is AlHsP'PP'API
=
Since the determinant of a product
(4181)
P'(jI A)P
is
UIEigenvectors
A
The n X
vector
p,
which
satisfies
a,I
 A)p, =
is
where
X, is
A associated with
the eigenvalue X t
Illustrative
examples of
how
4.10
Diagonalization of the
A matrix is
all
that if
to
A is
a diagonal be
X,
assumed
be
distinct, are
Sec. 4.10
Diagonalization of the
Matrix
119
nonzero elements given by e Xl ', e M e*'. There are other reasons for wanting to diagonalize the A matrix, such as the controllability of a system (Section 4. 1 5). We have to assume that all the eigenvalues of A are distinct, since, unless it is real and symmetric, A cannot always be diagonalized if it has multipleorder eigenvalues.
diagonal, with
its
,
.
x(0
= Ax(0 +
Bu(f)
(4184)
where x(t) is an n vector, u(t) an r vector, and A has distinct eigenvalues X u X2 X, it is desired to find a nonsingular matrix P such that the transformation x(0 = Py(?) (4185)
,
.
y(0
with
Ay(r)
ru(f)
(4186)
...
h
23
... ...
(n
n)
(4187)
...
K
state
This transformation
is
also
is
P 'AP
and
P'B
in the following that
(n
r)
(4189)
We show
is,
P=
where
p, (/
[Pi
P2
P 3 ...pj
is
(4190)
=
t
.
eigenvalue k
This
is
written
(4191)
= Ap,. .
.
i=
1,2,.
,n
Now
UiPi
or
[Pi
A 2p 2
Ap]
= [Ap, = A[p, =
A[p,
Ap 2
p2
.
Ap]
(4192)
p]
P2
P]A
p2
P.J
(4193)
Therefore,
if
we
let
P=
[Pi
P2
P.]
(4194)
120
StateVariable Characterization of
Dynamic Systems
Chap. 4
PA = AP
or
(4195)
A = PAP
which
the
is
(4196)
If the matrix
K
X\
X\
XI
(4197)
X~\
where X u X 2
Since
it
of A,
we
shall
show
P contains as its columns the eigenvectors column of the matrix in Eq. (4197) is the
eigenvector of
that
associated with
Xi=
1,2,...,
n.
Let
Pn
P,
(4198)
Pin.
be the
rth eigenvector
of A Then
.
U,i
or
A)p, ==
(4199)
~x
1
x
t
Pn
1
x<
Pn
1
(4200)
1
_0
a!
a n i
a 3
Cl\_
Pin
XiPn
XiPn
Pn = Pn =
(4201)
XtPi,l
fl//i
a n\Pn
+ (X
Pin= + a )p =
t
tm
Sec. 4.10
Diagonalization of the
Matrix
121
Now
we
arbitrarily let
pn
1.
Then Eq.
(4201) gives
X,
Pn
Pn
= =
tf
(4202)
Pt,n1
xr Ar
Pin
rth
in Eq. (4197).
equation
Example
411
on
i
A=
6
(4203)
11
6_
t
which is the phase variable canonical form, the eigenvalues of A are X = 1, A 2 = 2, X 3 = 3. The similarity transformation may be carried out by use of the Vandermonde matrix of Eq. (4197). Therefore,
"1
1
1"
P =
The canonicalform
X\
_Aj
"I
A3
A3
2
4
3
9
(4204)
Ai
is
state
equation
"1
A =piAP
Example 412
Given the matrix
2
A2
A3.
(4205)
r
6
(4206)
A=
it
11
11
5j
can be shown that the eigenvalues of A are Xi = 1, k 2 desired to find a nonsingular matrix P that will transform A, such that A = P~ 1 AP.
is
We shall follow the guideline that P contains the eigenvectors of A. Since of the phasevariable canonical form, we cannot use the Vandermonde matrix. Let the eigenvector associated with a, = 1 be represented by
pn
Pi
A is not
= Pn
(4207)
Then
pi
must
satisfy
a,i
 1^1
==
(4208)
122
StateVariable Characterization of
Dynamic Systems
Chap. 4
or ~X 1
6
A,
1
"
Pu
Pi\
J>31.
+
11
11
_6
The
last
6 A 5_
t
(4209)
Pu Pzi + P3l =0
(4210)
can letpn
= /> 3 =
i
1,
and
Pi
(4211)
=
"
1
A2
6
6 or
+
11
ll
6
A 2 5.
(4212)
2?12 ^22 + />32 = 6p 32 = 6p n + 9p = 6/>i2 + HP22 7/? then/ 2 = 2 andp 32 = three equations we let/>i 2 = In these
22
32
(4213)
4.
Thus
P2
(4214)
4
Finally, for the eigenvector p 3 ,
we have
1
"
~X 3
6
1
A3
Pl3
+
11
11
6
A 3 5.
P2 3
P33.
(4215)
_6
or
(4216)
Now
if
we
arbitrarily let
i3
1,
6 and
p 33 =9.
Therefore,
p3
=
"1
(4217)
The matrix P
is
now
given by
1
r
6
9_
P =
[Pi
P2
P 3]
_1
2 4
(4218)
Sec. 4.11
123
It is
easy to
show
that
"A,
1
on
A=P
4.11.
AP
_0
A2
A3
2
3_
(4219)
when the A matrix has multipleorder eigenvalues, unless the symmetric and has real elements, it cannot be diagonalized. However, there exists a similarity transformation
In general
matrix
is
A = P'AP
ri)
(4220)
such that the matrix A is almost a diagonal matrix. The matrix A is called the Jordan canonical form. Typical Jordan canonical forms are shown in the following examples
[A,
1
A,
A=
_0
"A:
1
A,
(4221)
A2 A3
A,
A=
_0
The Jordan canonical form
1.
A2 A3 A4
(4222)
2.
3.
main diagonal of are zero. the elements immediately above the multipleordered eigenvalues on the main diagonal are Is, such as the cases illustrated by Eqs. (4221) and (4222).
Some of
4.
The Is, together with the eigenvalues, form typical blocks which are called Jordan blocks. In Eqs. (4221) and (4222) the Jordan blocks are enclosed by dotted lines.
5.
When
its is
the nonsymmetrical matrix has multipleorder eigenvalues, eigenvectors are not linearly independent. For an n X n A, there
(r
only r
<
6.
The number of Jordan blocks is equal to the number of independent eigenvectors, r. There is one and only one linearly independent
eigenvector associated with each Jordan block.
7.
the
r.
124
StateVariable Characterization of
Dynamic Systems
Chap. 4
is
as
distinct eigenvalues
among n
eigenvalues. In the
first
place,
manner from
(A I
f
A)p,
=
/
(4223)
where
A,
2,
q.
The mined by
...
1
A,
...
(m X m)
...
A,
(4224)
.0
Xj_
Then
must hold:
0"
...
1
Ay
[Pi
Vz
PJ
o
kj
A[p,
p2
PJ
(4225)
Lo
or
Ay.
AyPi
p
P2
+ AyP + AyP
= Ap, = Ap = Ap
(4226)
Pmi
A,p m
= Ap m
The
also
vectors Pi, p 2 ,
be written
(4227)
(Ayl
A)p m
"0
= P m _l
5"
2
(4228)
Example
413
A=
4_
Sec. 4.11
125
the determinant of Al
A is
A
6
5
IAI
Al
1 3
2
A3
 4A 2 + 2)(A
5A
2
(4229)
2 A 4
=
Therefore,
(A
l) 2
To
such
A has a simple eigenvalue at A! =2 and a double eigenvalue at A 2 = 1. Jordan canonical form of A involves the determination of the matrix _1 that A = P AP. The eigenvector that is associated with X = 2 is deterfind the
t
mined from
(A,I
A)p,
5"
=
/>n ^21
_/>31_
(4230)
Thus
2
6
2
1 3
Setting
fore,
2
2_,
=

(4231)
2
pn
p zl
and pi
2.
There
(4232)
(4227).
For the eigenvector associated with the secondorder eigenvalue, we turn We have (the two remaining eigenvectors are p 2 and p 3 )
(A 2 I
to Eq.
 A)p 2 =
A)p 3
5"
(4233)
and
(A 2 I
= p
P\i
(4234)
6
1
1 3
Setting
2
3_
Pn
_/>3 2_
(4235)
2
p 12
arbitrarily,
wehave/? 22
= ^ andp 32
1"
Pz
(4236)
6
 5" Pl3
2
r
7
3
1 3
from which we have
P23
=
L
(4237)
3_ _P33_
7J
Pl3 P3 Pl3
_P33_
"
1"
n
_46
(4238)
Thus
7 t
n
li
(4239)
126
StateVariable Characterization of
Dynamic Systems
Chap. 4
is
now
obtained as
~2 0"
1
1
A = P'AP
_0
(4240)
1_
is
Note
4.12
State Diagram
The signal flow graph discussed in Section 3.5 applies only to algebraic equations. In this section we introduce the methods of the state diagram, which represents
an extension of the signal flow graph to portray
equations.
state equations
and
that
differential
it
The important
diagram
diagram
is
forms a
close relationship
among
state
is
constructed fol
lowing
all
the rules of the signal flow graph. Therefore, the state diagram
may
it is
The fundamental
can be performed on an analog computer are multiplication by a constant, addition, and integration. These are discussed separately in the following.
Multiplication by a constant. Multiplication of a machine variable by a
constant
tion
is
x 2 (t)
where a
late
is
a Xl (t)
(4241)
a constant. If a
lies
between zero and unity, a potentiometer is used An operational amplifier is used to simu
if a is a negative integer less than 1. The negative value due to the fact that there is always an 180 phase shift between the output and the input of an operational amplifier. The computer block diagram symbols of the potentiometer and the operational amplifier are shown in Figs. 45 and 46, respectively.
Eq. (4241)
of a considered
is
*i(0
x 2 (t)
xAt)
x 2 )
<z>
x 2 (t) = ax
Fig. 45.
l
(t)
0<a<\
x 2 {t)=ax
Fig. 46.
{t)
Analogcomputer blockdiagram
symbol of a potentiometer.
of an operational amplifier.
Sec. 4.12
State Diagram
127
Algebraic sum of two or more variables. The algebraic sum of two or more machine variables may be obtained by means of the operational amplifier. Amplification may be accompanied by algebraic sum. For example, Fig. 47
*i(0
x 2 (t)
x 3 (t)
*
x,(t)
x 4 (0 = a, x l
(t)
+ a 2 x 2 (t) + a } x 3 (t)
summer.
illustrates the
plifier
analog computer block diagram of a summing operational amwhich portrays the following equation
x 4 (t)
Integration.
is
a^At)
a 2 x 2 {t)
+a
x 3 (t)
(4242)
The
on an analog computer
is
the
condition Xi0
given at
and x 2 (t)
x,0)
>
f ax 2 {x)
to
dz
JCi(f )
<
(4243)
The block diagram symbol of the integrator is shown in Fig. 48. The integrator can also serve simultaneously as a summing and amplification device.
*i('o)
x 2 (t)
 x,(t)
*i( f )=
ax 2 (r)dT +
*,(?<))
Fig. 48.
integrator.
We shall now show that these analog computer operations can be portrayed
by
signal flow graphs
state
we
take the
X (s) = aX^s)
2
(4244)
The
is
shown
in Fig. 49.
128
StateVariable Characterization of
Dynamic Systems
Chap. 4
*i(0
*2(0
a
x 4 (t)
X2 (s)
x 2 (t)
XAs)
x,(0
^i(s)
X2 (s)
Signalflowgraph
Fig. 410. Signalflowgraph rep
Fig.
49.
representation
axi(t) or
2
of
x 2 (t)
aX\(s).
X (s) =
X (s) =
4
,*,(*)
a 2 X2 (s)
a s X3 (s)
is
(4245)
The
It is
shown
in Fig. 410.
important to note that the variables in the signal flow graphs of Figs. 49 and 410 may be in the time domain or the Laplace transform domain. Since the branch gains are constants in these cases, the equations are algebraic in
both domains.
sides
of Eq. (4243). In this case the transform operation is necessary, since the signalflowgraph algebra does not handle integration in the time domain. We have
X 2 (T)dT\
+ *i(*o)
x 2 (t) dx
x,(t )
x 2 {x) dx
o
Jo
(4246)
x 2 (r) dx
However, since the past history of the integrator is represented by x 2 (t ), and < x < t a Thus Eq. the state transition starts from t = t x 2 (x) =0 for (4246) becomes
, .
X^s)
It
^^ + ^5)
is
>
to
(4247)
>
X^s)
is
now
algebraic
411.
signal
An
Sec. 4.12
State Diagram /
129
*i('o)
*i('o)
<
as 1
O
X2 is)
Xi(s)
>
Xds)
<>
Xi
(s)
X2 (s)
Fig. 412. Signalflowgraph rep
[aX 2 (s)ls]
[xiOo)/*]
resentation of Xi(s)
[aX 2 (s)/s]
[xi(.to)ls].
Thus we have established a correspondence between the simple analog computer operations and the signalflowgraph representations. Since, as shown
in Fig. 412, these signalflowgraph elements
may
state
directly
This allows the determination of the state variables and the state equations once the differential equation of the system is given.
differential equation.
2.
state
function.
diagram can be constructed from the system's transfer This step is defined as the decomposition of transfer
The
4.
state diagram can be used for the programming of the system on an analog computer. The state diagram can be used for the simulation of the system on
a digital computer.
5.
equation in the Laplace transform domain obtained from the state diagram by means of the signalflowgraph gain formula.
state transition
The
may be
The
6.
from the
state
diagram.
7.
The
from the
diagram.
The
details
From
Diagram
a linear system is described by a highorder differential equation, a diagram can be constructed from these equations, although a direct approach is not always the most convenient. Consider the following differential
state
When
130
StateVariable Characterization of
Dynamic Systems
Chap. 4
equation
d"c
,
d"
dc
._
(4248)
we rearrange
the equa
d"c _
dt"
d" c
dc a. iw acrr
(4249)
o R
O
s"C
s
o
n~l
o
C
s" 2
<
o c
(a)
(b)
("!)/"
(n2>,
(1) (f
c ( ? o)
(c)
differential
equation of Eq.
Sec. 4.12
State Diagram /
131
W
is
.
.
'=1.2,...,b
used to represent dcjdt.
c
(n>
(4250)
Now
as
the variables
r, c,
c (1> , c (2) ,
shown
denoted by R{s), C(s), sC(s), s 2 C(s), s"C(s), respectively. As the next step, the nodes in Fig. 4 13(a) are connected by branches to portray Eq. (4249). Since the variables c and c _I> are related through integration with respect to time, they can be interconnected by a branch with gain 1 s and the elements of Figs. 411 and 412 can be used. Therefore, the complete state diagram is drawn as shown in Fig. 4 13(c).
,
(, '> ('
When
the input
is
is
on the
diagram
directly
not so straightforward.
We
show later that, in general, it is more confunction from the differential equation first and
shall
Example 414
differential equation:
g + 3 +
d
2c
2c
=r
rest
(4251)
of the terms,
we have
(4252)
is
dT^ 2c  3 dl
_
,
dc
Following the procedure outlined above, the state diagram of the system
Fig. 414.
shown
in
c (1) a +)
9*o+)
"
(4251).
to
was mentioned earlier that the state diagram is essentially a block diaprogramming of an analog computer, except for the phase reversal through amplification and integration.
gram
for the
132
StateVariable Characterization of
Dynamic Systems
Chap. 4
Example 415
An
(4251)
analog computer block diagram of the system described by Eq. is shown in Fig. 415. The final practical version of the com
from shown,
since amplitude
and time
scaling
different
Fig. 415.
Eq. (4251).
to Digital
Computer Simulation
on the digital comThe solution of differential equations by puter has been well established. However, from the standpoint of programming, a convenient way of modeling a system on the digital computer is by CSMP (Continuous System Modeling Program). 32 In many respects CSMP serves the
is
FORTRAN
computer program, except that the scaling problem state diagram or the state equations form a natural basis for the solution by CSMP. The following examples illustrate
same purpose
as an analog
practically eliminated.
The
typical
CSMP
Mathematical Equations
c
CSMP
C=
Y=
A\
Statements
*
a x
1
a 2x 2
XI
A2 * X2
x.
JT1/2.
Xl
f'
*,(t) dx
x,(0)
XI
INTGRL(Z2,
JT10)
Example 416
From
written
f t
dt
c=\cdt c = r  3c  1c
Sec. 4.12
State Diagram /
133
c
c
r
=C = CI = C2 =R
c(0)
<?(0)
= CO = CIO
CSMP representation
of the system
is
(4256)
(4257)
(4258)
We
have shown
A).
With the state diagram, the equivalence of the matrix inverse operation is out by use of the signalfiowgraph formula. The state transition equation in the Laplace transform domain is
X(s)
carried
(jI
A) !x(r J)
+
i
(jI
A)" 'BR(i)
>
(4259)
Therefore, the last equation can be written directly from the state diagram by
1, 2,
. .
n, as
and
and Rj(s),j
1, 2,
Example 417
Consider the state diagram of Fig. 414. The outputs of the intediagram is redrawn as shown in Fig. 416.
diagram
2
X (s)
t
and
2 (*o)>
anc*
>
Xl (s)
=J
"'
(1
3;rl)
+ v.(t*\ _L *l('o )
XlOZ)
r 2
*fr)
(4260)
X (s) =
2
2s~
(4261)
34
StateVariable Characterization of
Dynamic Systems
Chap. 4
where
A= +
+3s~ +2s~ 2
i
(4262)
After simplification, Eqs. (4260) and (4261) are presented in matrix form:
1
s
1
xiin)
Xz(s)
(j
IX*
2)
2
+
s
(s
+ +
l)(s
+ +
2)
R(s)
2)J
(4263)
xAn)
L(j
is
iXj
The
>
r{t) is
Then
the
 (1)
u,{t
to)
t>t
t>t
a
(4264)
"
The
~
'
(jTTi)
(4265)
is
2e~
2e
(, ' )
Mt).
( '~''
e 2 + 2e~ 2(t_
*' '
'
gco)
e 2dr
_
xiCt)
' o)
gc'o)
2e~ 2( '~' )
x 2 (t$)_
(4266)
t
t u s {t
t )
 c"
<''>
+ e
2 <''>"
(llo)
p2(tlo)
?>
this result
> 0.
to Transfer Function
transfer function
is
diagram by
418
and
Example
Consider the state diagram of Fig. 416. The transfer function C(s)/R(s) is obtained by applying the gain formula between these
setting x^{t J)
and
2
Xi(t J) == 0. Therefore,
C(s) t_ R(s) s2
+ 3s +
+
2
(4267)
The
is
3s
(4268)
When the state diagram of a system is already given, the state equations and the output equations can be obtained directly from the state diagram by
clarification seems necessary here, since the determined from the state diagram by use of the gain formula. However, when writing the state transition equations from
state transition equations are
Some
2,
n,
and the
states,
xi(t
regarded as the input nodes. Furthermore, the state transition equations, as written directly from the state diagram, are necessarily
)>
'
1 2,
n, are
The
domain
Sec. 4.12
State Diagram /
135
The left side of the state equation contains the firstorder time derivative of the state variable x (t). The right side of the equation contains the state
t
variables. There are no Laplace operator s or initial a state equation. Therefore, to obtain state equations from the state diagram, we should disregard all the initial states and all the integrator
variables
state variables in
To avoid confusion, the initial states and the branches with the gain s' 1 can actually be eliminated from the state diagram. The state diagram of Fig. 416 is simplified as described above, and the result is shown in
branches with gains s~ l
.
Fig. 417.
r as
input nodes,
initial states
and
the integra
^ = 2x
dx 2
dt
(4269)
x
3x 2
Example 419
As another example of
illustrating the determination of the state equations from the state diagram, consider the state diagram shown
in Fig. 41 8(a). This illustration will emphasize the importance of using the gain formula. Figure 41 8(b) shows the state diagram with the initial states
or
(a)
136
StateVariable Characterization of
Dynamic Systems
Chap. 4
O
r
*3
(b)
diagram
and
inte
grators eliminated.
and the integrators being eliminated. Notice that in this case the state diagram in Fig. 418(b) still contains a loop. Applying the gain formula to the diagram of Fig. 418(b) withii,;c 2 and x 3 as the output node variables and r,x\,x 2 and x 3 as the input
,
,
dxi
1
dt
Xi
1 1
dx 2
dt
(a 2 + a +aa
1
3)
O 2
aa a 3
x2
*3
(4270)
dx 3
dt
4.13
Up
this
methods of characterizing a
linear system
have
been presented.
be useful to summarize briefly and gather thoughts at point, before proceeding to the main topics of this section. It has been shown that the starting point of the description of a linear
system
be the system's differential equation, transfer function, or dynamic is demonstrated that all these methods are closely related. Further, the state diagram is shown to be a useful tool which not only can lead to the
equations. It
solutions of the state equations but also serves as a vehicle of translation
may
block diagram
is
drawn
as
from shown in
with the differential equation of a system, one can get to the solution by use of
the transfer function
also
method or the state equation method. The block diagram shows that the majority of the relationships are bilateral, so a great deal of flexibility exists between the methods.
Sec. 4.13
137
Differential
Dynamic
equations
equations
1 ,
'
State
transition
equation
^
Transfer function
State
diagram
Fig.
419. Block diagram showing the relationships among various methods of describing linear systems.
step remains to be discussed. This involves the construction of the diagram from the transfer function. In general, it is necessary to establish a better method than using Eqs. (4146) through (4148) in getting from a
state
One
highorder differential equation to the state equations. The process of going from the transfer function to the state diagram or the state equations is called the decomposition of the transfer function. In general,
there are three basic
tion,
ways of decomposing a transfer function direct decomposicascade decomposition, and parallel decomposition. Each of these three schemes of decomposition has its own advantage and is best suited for a par:
ticular situation.
Direct Decomposition
The direct decomposition scheme is applied to a transfer function that not in factored form. Without loss of generality, the method of direct decomposition can be described by the following transfer function
is
C(s) R(s)
s2
b s*
+as+a +bs+b
x
(4271)
The
objective
is
state equations.
The
fol
it has only negative powers accomplished by multiplying the numerator and the denominator of the transfer function by the inverse of its highest
of
s.
This
in
is
power
2.
For the transfer function of Eq. (4271), we multiply and the denominator of C{s)jR{s) by s~ 2 Multiply the numerator and the denominator of the transfer function by a dummy variable X(s). Implementing steps 1 and 2, Eq. (4271) becomes
j.
the numerator
C(s)
a
b
R(s)
3.
+ a s + a s~ h*~ + b,s' + b s~
' t
2
2
X{s) X{s)
(4272)
138
StateVariable Characterization of
Dynamic Systems
Chap. 4
to each other,
From
Eq.
= R(s) =
C(s)
4.
a,s~
a 2 s 2 )X(s)
b t s*)X(s)
(4273)
(4274)
(b
+ *,*
they must
first
relation. It is ap
we have
b2 ba
o b
'X(s)
X{s)
(4275)
The
diagram.
tors.
state
Eqs. (4273)
diagram is now drawn in Fig. 420 using the expressions in and (4275). For simplicity, the initial states are not drawn on the
usual, the state variables are defined as the outputs of the integra
As
C(s)
R(s)
direct
decomposition.
Following the method described in the written directly from the state diagram:
dxi
dt
1
equations are
Xl
dx 2
.dt
b 2
6i
b
+
1
(4276)
x2
J
The output equation is obtained from Fig. 420 by applying the gain formula with c(t) as the output node and x^t), x 2 {t), and r(t) as the input nodes.
(4  277)
Sec. 4.13
Decomposition
of Transfer Functions /
139
Cascade Decomposition
Cascade decomposition
may
is in
the factored form. Consider that the transfer function of Eq. (4271)
may be
factored in the following form (of course, there are other possible combinations
of factoring)
gi
where z u
z 2 ,p u
R(s)
= ^lIilJi b s + p s + p
t
(4278)
and p 2
Then
it is
The state diagram of each of the firstorder transfer functions is realized by using the direct decomposition method. The complete state diagram is obtained by cascading the
two
firstorder transfer functions. firstorder
two
R(.s)
~Pi
Fig. 421. State
decomposition.
integrators
on the
state
state variables.
The
state
dt
ctx^
Pi
Pi
+
Pi
*2
a b a
(4279)
_*oJ
(z 2
p 2 )x +(z,
l
 Pi)x +
2
%
on the
(4280)
state diagram.
when
the poles
are varied.
Parallel
Decomposition
the denominator of a transfer function
is
When
possible to
in factored form,
it
is
expand the transfer function by partial fractions. Consider that a secondorder system is represented by the following transfer function
Cjs)_
R(s)
(s
Pis)
+ Pl )(s + p
less
(4281)
2)
where P(s)
is
a polynomial of order
than
2.
We
140
StateVariable Characterization of
Dynamic Systems
Chap. 4
and p 2 may be complex conjugate for analytical purposes, but it is difficult to implement complex coefficients on the computer. In this case if p and p 2 are equal it would not be possible to carry out a partialfraction expansion of the transfer function of Eq. (4281). With p and
x t
p2
is
written
C(s) R(s)
Kr
s
K,
s
+ Pi
+p
(4282)
2
where
and
state
are constants.
The
the state
diagram for the system is formed by the parallel combination of diagram representation of each of the firstorder terms on the right
shown
in Fig. 422.
The
state equations
written
dxC
dt
P\
(4283)
dx 2
\_dt _
Pi
X%
X!(f +)
R(s)
parallel
decomposition.
is
[K,
2]
(4284)
that for transfer
One of
Therefore,
is
A matrix is always
a diagonal matrix.
we can
may be
diagonalization of the
A matrix.
When
that the state diagram, as obtained through the parallel decomposition, contain the minimum number of integrators. To further clarify the point just made,
its
partialfraction expansion:
Sec. 4.14
141
an
R(s)
2s 2
(s
6s
iy(s
+5 + 2)
(s
iy
+ s+l
(4285)
Note that the transfer function is of the third order, and although the total order of the terms on the right side of Eq. (4285) is four, only three integrators should be used in the state diagram. The state diagram for the system is drawn as shown in Fig. 423. The minimum number of three integrators are used, with one in
paralled decomposition.
state equations
written
1
dt
*1
dx 2
dt
01
x2
x3
(4286)
dx 3
dt
02
Therefore, the
A matrix is
4.14
When
form
the
it
it
it is desirable to avoid matrices with complex elements. When A has complex eigenvalues, in general, the matrix can be transformed into a nondiagonal matrix which is called the modal form by the transformation
putation,
A = P"'AP
Let us assume that
(4287)
is
x 2 and has
is
eigenvalues A,
+ jco
and X 2
jco.
given by
142
StateVariable Characterization of
Dynamic Systems
Chap. 4
A=
CO
CO
,
. . .
(4288)
The elements of the P matrix may be determined by brute force using Eqs. (4287) and (4288). If A has m real distinct eigenvalues in X u X 2 X m and n sets of complexconjugate eigenvalues in X, = a jcoi, i = 1 2, ,
, , t ,
.
. .
is
given by
..
.
x2
A=
Am
..
o
(4289)
..
..
..
..
where
(4290)
\CO,
If they'th
On
of multiplicity m, then
is
is
written
Tj
o o
r,
o
r,
(m X
blocks)
(4291)
ooo
where
"
Oj
COj~
(4292)
iCOj
ri
CTj]
oi
(4293)
lj
is
Lo
in Eq. (4289)
easily
modified for
real
and multiple
order eigenvalues by use of the Jordan canonical form. Although the modalform matrix is not diagonal and does not represent
decoupling of the states from the standpoint of the state diagram, it still has the components of the eigenvalues as its matrix elements. To determine the transformation matrix P for the matrix A of Eq. (4288),
we
let
P=
where p and p 2 are 2 x
(
[Pl
pj
is
(4294)
vectors.
Equation (4287)
written
a
[Pi
co
P2]
A[p,
co
p2 ]
(4295)
Sec. 4.14
143
or
(4296) (4297)
coll ~pr
rf_ J>2_
"A
_0
0~
Pi
(4298)
col
A_
_P 2
where
denotes a 2
2 identity matrix.
Let qi and q 2 denote the eigenvectors that are associated with the two complexconjugate eigenvalues, X 1 a jco and X 2 a ja>, respectively.
+
1
satisfy
(^+yco)q
(<r
 jco)q
= =
<X,
=Aq = Aq
(4299)
(4300)
(4301)
Let
q,
q2
(4302)
Then
become
(4303)
(4304)
Equating the
real
last
col "i
ctI
"A
01 rii
_coI
LpJ
and
a\
A_ LpJ
(4305)
coT "2~
<rl_
A
_0
0" "2~
_ col
Comparing Eq.
LP 2 J
A_ LpzJ
the identity
(4306)
we have
P=[P.
The
by taking the
with Xi
real
P2]
[i
Pi]
(4307)
+ jco.
Consider the state equation
Example 420
where
i = Ax
1"
+ Br
B=
2 l
(4308)
2
The eigenvalues ofA are A
qi
i
2
q2
_i J.
=
or
qi
a! +/Pi
q2
= a 2 +/P2
+ J 1
144
StateVariable Characterization of
Dynamic Systems
Chap. 4
Therefore,
P =
[*i
Pt]
i
ii~
(4309)
_i
i
A=
P!AP
= ri
(4310)
~i_
(4311)
T = piB =
The
original state equation of Eq. (4308)
is
transformed to
(4312)
= Ay + Tr
~ cos
e~<
t
Then
is
given by
sin i
= eM =
(4313)
sin
cos
4.15
Controllability of Linear
Systems
observability introduced
first
by Kalman 24
The conditions on
controllability
However, one should not associate these conditions with all optimal control problems require
will
that the system be controllable and/or observable in order to achieve the control objectives.
These points
be
this section
we
of controllability.
The concept of
diagram of
Fig. 424.
The process
is
Control u(0
State x(t)
G can
simple
any one of the state variables is independent of the control would be no way of driving this particular state variable to a desired u(r), there state in finite time by means of a control effort. Therefore, this particular state is said to be uncontrollable, and as long as there is at least one uncontrollable
to understand that
state, the
system
is
said to
trollable.
Sec. 4.15
Controllability of Linear
Systems
145
?*i('o+)
?*i (>o+)
u(t)
is
As a
state
simple example of an uncontrollable system, Fig. 425 illustrates the diagram of a linear system with two state variables. Since the control
i/(0 affects
is
it
initial state
x 2 (t
to a desired state
would x 2 (tf)
ta
by any control
u(t).
said to be uncontrollable.
The concept of controllability given above refers to the states and is sometimes referred to as the state controllability. Controllability can also be defined
for the outputs of a system, so there
is
and output
controllability.
is
dynamic equations
x(0
c(0
= =
Ax(0
(4314) (4315)
where
x(0
state vector
input vector
output vector
n coefficient matrix
r coefficient
matrix
n coefficient matrix
r coefficient
matrix
The state x{i) is said to be controllable at t / // there exists a piecewise continuous input u(r) that will drive the state to any final state x(t for a finite f) time {t f t ) 0. If every state x(t ) of the system is controllable in a finite time interval, the system is said to be completely state controllable or simply state
>
controllable.
146
StaleVariable Characterization of
Dynamic Systems
Chap. 4
The following theorem shows that the condition of controllability depends on the coefficient matrices A and B of the system. The theorem also gives one
way of testing
state controllability.
Theorem 42. For the system described by the state equation of Eq. (4314) be completely state controllable, it is necessary and sufficient that the following n X nr matrix has a rank of n:
to
S
[A, B]
[B
AB A
B.
A"'B]
(4316)
Since the matrices A and B are involved, sometimes we say that the pair is controllable, which implies that S is of rank n.
Proof:
The
is
=
t
<j>(?
is
)x(t )
+
0.
>
ftt
t)Bu(t) dx
(4317)
final
to
for
>
tQ
state for
some
f >
x(t f )
Then Eq.
(4317) gives
x('o)
=~
f"
Mo m
t)Bu(t) dx
(4318)
From
A*
= 2
a* m A
for any
(4319)
Then
<K0
= =
'=
S A"t *=o k}
icl
(4320)
Zj TT Kk=
nl
m=0
oo
21
&km
or
ftf)
m0
"t A"
2 a km U k=0
,k
ft!
(4321)
Thus
<(?)
<K0
A"
we have
(4322)
=  2 Am B
f"
a m (f
t)u(t)
</t
(4323)
Let
U m = r m 0o^)u(T)^T
J
to
(4324)
Then Eq.
(4323)
becomes
x('o)
=  2 A m BU m
771
(4325)
which
is
Sec. 4.1 5
Controllability of Linear
Systems
147
x('o)
= [ AB A B = SU
2
A""'B]U
(4326)
where
U=
trollability
[U
U....U,.,]'
(4327)
Equation (4325) represents n equations with nr unknowns, and the conproblem may be interpreted as: Given any initial state x(t ), find
is
x(^)
S, solve
is completely state controllable if and only if there exists a set of n independent column vectors in S. For a system with a scalar input, r = 1, the matrix S is square; then the condition of state controllability is that S must be nonsingular.
the system
linearly
Although the
tems.
quite straightforward,
Even with r = 2, number of possible combinations of n x n matrices. A practical way may be to use one column of B at a time, each time giving an n X n matrix for S. However, failure to find an S with a rank for n this way does not mean that the system is uncontrollable, until all the columns of B are used. An easier way would be to form the matrix SS', which is n X n then if SS' is nonsingular, S has
;
by Theorem 42 is not very easy to implement for multipleinput systhere are In columns in S, and there would be a large
rank
n.
Example
421
Consider the system shown in Fig. 425, which was reasoned earlier to be uncontrollable. Let us investigate the same problem using the
condition of Eq. (4316).
The
rki(Oi
dt
2
iW
dx 2 (t)
I
+
1
J
x%{t)
u(t)
(4328)
dt
S
which
is
.
[B
AB]
"1
2"
(4329)
_o
singular, anc the sys tem
is
o.
e.
Example 422
Determine the
state equation
dxM
dt
1
*i(0
dx 2 Q)
.
+
1
J
"(0
(4330)
dt
xi(0
From
Eq. (4316),
S
which
is
[B
AB]
1
(4331)
is
148
StateVariable Characterization of
Dynamic Systems
Chap. 4
is
i(0
If the eigenvalues
Ax(0
Bu(?)
of
A are
distinct
n,
into a
such that
"A,
P'AP =
A,
(4333)
0.
Let the
new
state variable
be
y
P'x
(4334)
Then
is
y
where
Ay
Tii
(4335)
r = p'B
The motivation
states
(4336)
for the use of the similarity transformation is that the of the system of Eq. (4335) are decoupled from each other, and the only way the states are controllable is through the inputs directly. Thus, for state controllability, each state should be controlled by at least one input. Therefore,
eigenvalues
an alternative definition of state controllability for a system with distinct is: The system is completely state controllable if T has no rows that
are all zeros.
It
should be noted that the prerequisite on distinct eigenvalues precedes all square matrices with
eigenvalues can
be diagonalized.
The
is
natural question
is:
no.
We
must not
lose
any state x(t ) is brought to any state \(t f ) in finite time. Thus the question of independent control must enter the picture. In other words, consider that we have two states
sight of the original definition
on
state equations:
^ ^
a Xi (t)
+ bMO
b 2 u{t)
(4337)
= ax&)
(4338)
Sec. 4.15
Controllability of Linear
Systems
149
This system
is
S
is
~b,
ab
[B
AB]
bz
(4339)
ab 2
is
diagonal, and
controllable.
mean
is
that
has multipleorder eigenvalues. has multipleorder eigenvalues and cannot be diagonalized, there
When A
is
P which transforms A into a Jordan canonical form The condition of state controllability is that all the elements of r = P _1 B that correspond to the last row of each Jordan block are nonzero. The reason behind this is that the last row of each Jordan block corresponds
a nonsingular matrix
A = P'AP.
is
The elements
values,
in the other
all
all
the matrix
A has
four eigen
X it X u X u
Xi, three
is
a nonsingular
which transforms
A into
form
0~
'
P AP
A,
1
!
(4340)
A,
A2 J
selfexplanatory.
and
matrices are,
"2
L
It
r
i_
B=
T
l_oJ
Let us check the controllability of the system by checking the rows of the matrix T.
A is
p=
Therefore,
_o
r
The transformed
=
is
'B
r r =
i_ _o_
"1"
(4341)
_o
state equation
_0_
~2
0~ y(0
no
Since the second
+
_0_
u(t)
(4342)
row of
is
x 2 (0
*s
uncontrollable,
is
uncontrollable.
Example 424
2
1
r
B=
3
"0
A=
1
4
150
StateVariable Characterization of
Dynamic Systems
Chap. 4
Then
1
S
41
(4343)
[B
AB A
B]
1
8_
Since
A2
S is singular, the system is not state controllable. Using the alternative method, the eigenvalues of A are found to be Ai = 2, and A 3 = 1. The Jordan canonical form of A is obtained with
1
2,
0"
1
(4344)
2_
Then
[2
1
01
(4345)
1_
A = P'AP
_0
T = P'B
Since the last
this
(4346)
row of T
is
y3
is
uncontrollable. Since
x2
=y
3,
Example 425
state
i(0
r
o
l
x(0
+
1"
u(t)
(4347)
= [B
AB]
1
(4348)
nonsingular.
Let us
formation,
now
The system is completely controllable. check the controllability of the system from the rows of T. The
x
eigenvalues of
~
1
J
=
i]
A
and
AP
7
_o
"
/_
r=
T
p'b
=
L2/J
Since
all
the rows of
is
controllable.
In general, when the eigenvalues are complex, which occurs quite frequently in
control systems,
it is
more
difficult to
Sec. 4.15
Controllability of Linear
Systems
151
use the modal form so that only real matrices are dealt with. In the present problem
A may
a
_a>
by the transform matrix
CO"
r
(4349)
a_
"1
.1
o_
P =
_1
r
i_
Then
r=p
Since the
'B
2" 1
1
modal form
not
lability is that
all the
rows of
are zeros.
Output Controllability 1 *
is
referred
controllable if every
by an unconstrained
sary nor sufficient for the existence of a solution of the problem of controlling
of output
if
)
controllability.
system
is
said to be completely
output controllable
drive the output y(?
(t f
t )
>0.
Theorem 43. Consider that an nthorder linear timeinvariant system described by the dynamic equations of Eqs. (4314) and (4315). The system completely output controllable if and only if the p X (n T)r matrix
is is
T = [DB DAB DA 2 B.
is
DA nl B
E]
(4350)
of rank p. Or,
is
similar to that of
Theorem
42.
is
described
d 2 c(t)
dt 2
dc(t)
~dT
c{t)
_ du{i) ~ dT
(4351)
The
state controllability
gated.
We shall show that the state controllability of the system depends upon how
the
x2
= =
c
c
152
StateVariable Characterization of
Dynamic Systems
Chap. 4
The
=
_*2_
"
r
2_
C
_i
+
X%
(4352)
l
is
= X\
=
~
1
(4353)
The
is
S
which
matrix
is
= B
[
AB]
r
(4354)
i.
_l
[1
singular.
The system
is
From
is
D=
0]
and
E=
0.
The output
0]
controllability
written
T = [DB DAB
which
is
E]
[1
1
(4355)
is
of rank
let
1,
the
output con
trollable.
Now
By
the
form
(4356)
"
"0"
_*2_
.1
C
2. L*2_
Xi
["1 +
_1_
is
+x =
(4357)
is
now
S
which
is
"0
[B
AB]
_i
r ZJ 2,
1
(4358)
nonsingular.
is still
The system
which
of rank
T = [DB DAB
is
1
E]
[1
0]
(4359)
have demonstrated through this example that given a linear system, state depends on how the state variables are defined. Of course, the output controllability is directly dependent upon the assignment of the output variable. The two types of controllability are not at all related to each other.
controllability
We
4.16
The concept of
tially,
observability
is
is
a system
completely observable
some of the outputs. In other words, it is often desirable to obtain information on the state variables from measurements of the outputs and the inputs. If any one of the states cannot be observed from the measurements of the outputs, the state is said to be unobservable, and the system is not completeaffects
ly observable,
or
is
simply unobservable. Figure 426 shows the state diagram of x 2 is not connected to the output c in any way.
x ,
Once we have measured c, we can observe the state x since x, = c. However, x 2 cannot be observed from the information on c. Thus the system is
Sec. 4.16
153
?*2('0+)
?M'o+)
u(t)
is
not observable.
is
described by the dynamic equations of Eqs. (43 14) and (4315), the state \(t ) is said to be observable if given any input u(f), there exists a finite time t tQ
f
>
such that the knowledge ofu(t)for t t ; the matrices A, B, D, and E; and f the output c(t) for t f are sufficient to determine x(t ). If every state of the system is observable for a finite t f we say that the system is completely observable, or simply observable.
<t <
,
<t<t
The following theorem shows that the condition of observability depends on the coefficient matrices A and D of the system. The theorem also gives one method of testing observability.
Theorem 44. For the system described by the dynamic equation of Eqs. (4314) and (4315) to be completely observable, it is necessary and sufficient that the following n x np matrix has a rank of n:
V=
[D'
A'D'
(A') 2
D\
(A')" 1 D']
(4360)
The condition is also referred to as the pair [A, D] being observable. In particular, if the system has only one output, is an 1 X n matrix; of Eq. (4360) is an n X n square matrix. Then the system is completely observable if is nonsingular.
we have
B(f>(t
)x(t
+D
<f>(?
t)Bu(t) dx
Eu(?)
(4361)
Based on the definition of observability, it is apparent that the observability ) depends essentially on the first term of the right side of Eq. (4361). With u(t) = 0, Eq. (4361) becomes
of x(f c(0
D$(r
)x(t
(4362)
Making
c(0
m=
a m (0DAx(r
(4363)
154
StateVariable Characterization of
Dynamic Systems
Chap. 4
or
c(0
D DA DA
(a I
a I
t
,!)
xOo)
(4364)
DA"
Therefore, knowing the output
is
c(t)
< <
t
f , x(t )
and only
if
the matrix
D DA DA
(np
w)
DA"
has rank
n.
Or
the matrix
V=
has a rank of
n.
[D'
A'D'
(A') 2
(A')"'D']
(4365)
Comparing Eq.
following observations
1.
may
be
made
Controllability
[A', B'].
2.
Example 427
Consider the system shown in Fig. 426, which was earlier defined to be unobservable. The dynamic equations of the system are written
directly
from the
state diagram.
=
t
'2
0"
~3"
1_ L*2_
~x{~
[1
+
_1_
(4366)
0]
(4367)
_*2_ L*2j
Therefore,
D=
A'D'
and, from Eq. (4360),
[1
0]
D'
0"
2
1
[D'
V=
Since
A'D]
"1
2'
(4368)
is
is
unobservable.
Sec. 4.16
155
Example 428
1"
1.
Xl
(4369)
Xl.
L*2.
1 1
ci
1
For the
test
(4370)
1
Lxil
of observability,
we
r
evaluate
i
AD'
The
observability matrix
n ri
i J)
ii
i_
i
2_
(4371)
_i
becomes smes
)'
_l
1
V
Since
AD] =
is
"i
l
l
0"
(4372)
_o
1
2_
is
has a rank of
2,
which
the
number of
completely
observable.
Example 429
Let us consider the system described by the differential equation of Eq. (4351), Example 426. In Example 426 we have shown that
state controllability of
a system depends on
how
are defined.
on the
definition
of the state variables. Let the dynamic equations of the system be defined as in Eqs.
(4352)
and
A=
Then
o
1
D=[l
0]
V=
[D'
AD]
(4373)
and thus the system is completely observable. Let the dynamic equations of the system be given by Eqs. (4356) and Then
1"
(4357).
1
2
D=
"1
1
[l
1]
Then
V=
which
is
[D'
AD] =
']
1 1
we have shown
that given the
singular.
is
unobservable, and
inputoutput relation of a linear system, the observability of the system depends on how the state variables are defined. It should be noted that for the system of Eq.
(4351),
one method of state variable assignment, Eqs. (4352) and (4353) yields a system that is observable but not state controllable. On the other hand, if the dynamic equations of Eqs. (4356) and (4357) are used, the system is completely state controllable but not observable.
There are
definite reasons
behind these
results,
and we
shall
investigate these
phenomena
Alternative definition
values,
it
of observability. If the matrix A has distinct eigencan be diagonalized as in Eq. (4333). The new state variable is
y
P'x
(4374)
156
StateVariable Characterization of
Dynamic Systems
Chap. 4
are
y
where
Then the system
c ==
+ Tu Fy + Eu F = DP
Ay
is
The reason behind the above condition is that if the /th (j = 1,2, ... ,n) column of F contains all zeros, the state variable ys will not appear in Eq. (4376) and is not related to the output z{i). Therefore, y, will be unobservable.
In general, the states that correspond to zero columns of
Example 430
Consider the system of Example 427, which was found to be unobservable. Since the
matrix, as shown in Eq (4366), is already a diagonal matrix, the alternative condition of observability stated
above requires that the matrix D = [1 0] must not contain any zero columns. Since the second column of D is indeed zero, the state x 2 is unobservable, and the system
is
unobservable.
4.17
Relationship
Among
Controllability, Observability,
Although controllability and observability are concepts of modern control theory, they are closely
related to the properties of the transfer function.
Let us focus our attention on the system considered in Examples 426 and
429. It
was demonstrated
state controllable or
in these two examples that the system is either not not observable, depending on the ways the state variables
We
have
. ,
_
s*
1 1
_
at s
+ 2s +
~(s
which has an
identical pole
and zero
( <.<>
Theorem 45. If the inputoutput transfer function of a linear system has polezero cancellation, the system will be either not state controllable or unobservable, depending on how the state variables are defined. If the inputoutput transfer function of a linear system does not have polezero cancellation, the system can always be represented by dynamic equations as a completely controllable and observable system.
Proof: Consider that an thorder system with a single input and single
is
x(?)
= c(t) =
Bu(t)
(4379) (4380)
Dx(0
Sec. 4.17
Relationship
Among
Controllability, Observability
and Functions
157
Let the
A matrix
be diagonalized by an n
1 1
x
1
Vandermonde matrix
...
1
P,
/]
ki XI
A3
X\
...
X\
A?
(4381)
1nl
.Ai
1nl A2
A3
ln1
. .
].
The new
state
is
y(0
where
=
c (0
Ay(?)
+
is
Tu(t)
(4382)
A= =
is
transformed into
(4383)
= =
Fy(0
y(f) are related
where F
DP. The
and
by
(4384)
is
Py(f)
Since
MO =
where
n
a, is
*#&)
ytff)
(4385)
and y, is the /th element of T, where T is an Taking the Laplace transform on both sides
initial
conditions,
we obtain
the transfer
Jl. U(s)
(4386)
A,
(4383)
is
C(s)
FY(5)
DPY(i)
(4387)
Now
then
if it is
assumed that
D=
F
where
/,
[rf 1
d2
[/
I
...
d]
(4388)
= DP =
</,
f2
...
/J
(4389) (4390)
d 2 X,
+
is
djrr
for
written as
Gto
[/i
/JY(j)
[/.
U(s)
(4391)
f,7,
U(s)
X,
158
StateVariable Characterization of
Dynamic Systems
Chap. 4
of the form
U(s)
(s
 A,)(j 
A2)
A)
which
is
expanded by
v$>
t
&^
; x
(4  393)
where a denotes the residue of C(s)jU(s) at s = A,. It was established earlier that for the system described by Eq. (4382) to for be state controllable, all the rows of T must be nonzero that is, y, i = 1, 2, ...,. If C(s)/U(s) has one or more pairs of identical pole and zero, for instance in Eq. (4392), a = A 1; then in Eq. (4393), a = 0. Comparing
we
(r,=/,y, Therefore,
lable.
(4394)
0,
when
a,
0, y, will
be zero if/, =
and the
state y, is uncontrol
For
observability,
it
was established
earlier that
F must
i
for
2,
n.
However,
from Eq.
(4394),
f =
t
f
t
(4395)
a,
When
a,
0.
if y,
^ 0.
4.18
When
(4396)
where
f [x(r),
n
1
p x
denotes an n
is
Being able to represent a nonlinear and/or timevarying system by state is a distinct advantage of the statevariable approach over the transfer function method, since the latter is defined strictly only for linear timeinvariant
equations
systems.
As a simple
linear
:
illustrative
= x,(/) +
x\{t)
(4  397)
Sec. 4.18
159
it
would
be desirable to perform a linearization whenever the situation justifies. A linearization process that depends on expanding the nonlinear state equation into a Taylor series about a nominal operating point or trajectory
is
now
described. All the terms of the Taylor series of order higher than
are
discarded,
and
linear
nominal point results. Let the nominal operating trajectory be denoted by x (f), which corresponds to the nominal input r (?) and some fixed initial states. Expanding the nonlinear state equation of Eq. (4396) into a Taylor series about x(f) = x (f) and neglecting
all
x (t)=f (x
i i
,i
(Xj
x 0J )
(4398)
r 0J )
*
i
Sfjjx, r)
0y
1,
2,
n.
Let
Axi
X,
Oj
(4399)
and
Ar,
r,
r 0/
(4400)
Then
Ax,
Since
x,
X 0!
o>
(4401)
x oi
Equation (4398)
is
= /i(x
r o)
(4402)
written
~~
'
j=x
dMx,r)
dx,
io,r,
j"
or
An
(4403)
The
last
equation
may be
Ax
where
= A* Ax +
9fx
x
B*Ar
df{ dx n
(4404)
'df dx
dx 2 dx 2
dh
A*
dx
dx
(4405)
El
dx
t
dL
dx 2
dli dr z
d_h dr 2
"
df
dx
dfx dr,
df~
drp
df2
df2
'
B*
dr t
dr
(4406)
dh
dr,
dL
dr 2
d_L dr a
160
/ StateVariable Characterization of
Dynamic Systems
Chap. 4
should be reiterated that A* and B* are evaluated at the nominal point. Thus we have linearized the nonlinear system of Eq. (4396) at a nominal operating point. However, in general, although Eq. (4404) is linear, the ele
where
it
ments of A* and B* may be time varying. The following examples serve to illustrate the linearization procedure
described.
just
Example
431
Figure 427 shows the block diagram of a control system with a saturation nonlinearity. The state equations of the system are
Xi *i=fi=xi X 2 = f2 = u
(4407) (4408)
1
It
x = Ax + Bu
*l
Fig. 427.
represented by
(4409)
(1
e*i*'i)
SGN
jc,
jf,
where
+1 SGN xi =
[
>
1
<0
(4410)
Substituting Eq. (4409) into Eq. (4408) and using Eq. (4403),
linearized state equation
we have
the
Aii
= =
MAx 2 = Ax 2
ax 2 ax
At
(4411)
Ax 2
^Ax
t
x
.
= KeV'^Axi
(4412)
where x 01 denotes a nominal value of x Notice that the last two equations are linear and are valid only for small signals. In vectormatrix form, ihese linearized state
equations are written as
~Ax{
"Axi"
(4413)
Ax 2
where
a
Ax 2
Ke K \*"\
constant
(4414)
It is of interest to check the significance of the linearization. If x oi is chosen to be at the origin of the nonlinearity, x 01 = 0, then a = K; Eq. (4412) becomes
Ax 2 =KAxi
Thus the
linearized
(4415)
model
is
Sec. 4.19
gain K.
lie
On
if
is
on the saturated portion of the nonlinearity, and a = 0. This means that any small variation in x, (small Ax,) will give rise to practically no change in Ax 2
.
Example 432
In the
last
invari
of a nonlinear system often leads to a linear timevarying system. Consider the following nonlinear system:
*i
= ^t
(4416)
x2
x,
(4417)
would like to linearize these equations about the nominal trajectory [x ,(0, X02O)], which is the solution of the equations with the initial conditions x,(0) = x 2 (0) = 1 and the input u{t) = 0.
Integrating both sides of Eq. (441 7),
We
we have
Xi
x 2 (0)
=
1
(4418)
Then Eq.
(4416) gives
xi
= / +
(4419)
Therefore, the nominal trajectory about which Eqs. (4416) and (4417) are to be linearized is described by
*oi(0
x Q i(t)
= =
~t
1
+
we
(4420) (4421)
get
Now
=
dxi
V1 = dx
2
<?x.
du
Xi
Ax,
= 3Ax X02
oAx,
(4422)
Ax 2 =
Substituting Eqs. (4420)
,Ai/
(4423)
(4423), the linearized
Ax,
Ax, _Ax 2 _
"
"
_Ax 2
which
is
_o
o_
J '_
(4424)
4.19
by means of discrete
is
state equations.
when
situations.
The
first
one
that the
data elements, but the signals at certain points of the system are discrete or discontinuous with respect to time, because of the sampleandhold operations.
In this case the components of the system are
still
described by differential
may be
162
Dynamic Systems
Chap. 4
generated from the original differential equations. The second situation involves systems that are completely discrete with respect to time in the sense that they
receive
discrete data only, such as in the case of a digital controller computer. Under this condition, the system dynamics should be or digital
Let us consider the openloop discretedata control system with a sampleandhold device, as shown in Fig. 428. Typical signals that appear at various points in the system are also shown in the figure. The output signal, c(t),
rit)
J\J
(t)
Zeroorder hold
h(t)
c(t)
r*(t)
'
i:
T IT 3T AT 5T6T IT
*
<TTTv
*
^liJ
*
: :
Sec. 4.19
163
ordinarily
is
is a continuousdata signal. The output of the sampleandhold, a train of steps. Therefore, we can write
h{t),
h(kT)
r(kT)
0,
2,
(4425)
Now we let the linear process G be described by the state equation and output equation
ML = Ax(0 + Bh(t)
c(t)
(4426)
Dx(0
+ EA(0
are the scalar input
(4427)
where
signals, respectively.
earlier.
x(t)
tft
)\(t
("
<f>(t
x)Bh(x) dx
(4428)
for
>
If
we
we
VT
let
(k
l)rand
=
dc
kT. Then
l)r]
4>(T)x(kT)
+ ^
$[{k
+l)Th{kT)
t]BA(t)
(4429)
where
(k
Since h(t)
is,
= r(kT)
for
kT<t<
sign.
Equation (4429)
is
x[(*
\)T]
= =
${T)x{kT)
+ 2 +
+Ur
*K*
l)r
 t]B dx r(kT)
(4430)
or
x[(fc
l)r]
tt7>(*r)
9(T)
Q(T)r(kT)
(4431)
where
=
J kT
^k +
1}
T~
T]B dZ
Equation (4431)
form. Since
it
is
represents a set of firstorder difference equations, to as the vectormatrix discrete state equation.
referred
The discrete state equation in Eq. (4431) can be solved by means of a simple recursion procedure. Setting k 0, 1, 2, ... in Eq. (4431), we find that the following equations result
k * k
= = =
0:
1
2:
k= k
1:
x(kT)
= (K7>p 
1)T]
*(T)r[(k
\)T]
(4436)
164
StateVariable Characterization of
Dynamic Systems
Chap. 4
Substituting Eq. (4433) into Eq. (4434), and then Eq. (4434) into Eq. (4435), solution for Eq. (4431): , and so on, we obtain the following .
.
x(kT)
Equation (4437)
data system.
is
(4437)
It is interesting to
defined as the discrete state transition equation of the discretenote that Eq. (4437) is analogous to its continu(467). In fact,
(467) describes the state of the system of Fig. 428 with or without sampling. The discrete state transition equation of Eq. (4437) is more restricted in that it
= kT (k =
0, 1, 2,
.),
and only
if
a sampleandhold device such as in Fig. 428. With kT considered as the initial time, a discrete state transition equation similar to that of Eq. (470) can be obtained as
x[(k
+ N)T] =
4>
N (T)x(kT)
+ 2
4>'\T)B(r)r[(k
+ i)T]
as
(4438)
where
is
is left
an exercise
The output of the system of Fig. 428 at the sampling instants is obtained by substituting t = AT and Eq. (4437) into Eq. (4427), yielding
c(kT)
= =
Dx(kT)
+ Eh(kT)
D4>*(7>(0)
+ D 2 p'KTWTWT) +
=
*i
(4439)
Kh{kT)
method
important advantage of the statevariable method over the ztransform is that it can be modified easily to describe the states and the output (k  A)T, where between sampling instants. In Eq. (4428) if we let /
An
<
A<
and
A)r
r]B dx r(kT)
^^
between
and
1,
One
completely described by Eq. (4440). of the interesting properties of the state transition matrix $(t)
is
+*(T) = ftkT) which is proved as follows. Using the homogeneous solution of the we have
x(t)
(4441)
= #t 
foM'o)
(4442)
Let
= kT and
ta
becomes
(4443)
x{kT)
Also, by the recursive procedure
= ftfcTXO) = {k + with
t
\)T and
kT, k
0, 1,
Sec. 4.20
165
2,
x(kT)
<J>*(T)x(0)
(4444)
identity in Eq. (4441).
In view of the relation of Eq. (4441), the discrete state transition equations
x(kT)
x[(A:
(4445) (4446)
These two equations can be modified to represent systems with r into a vector r. When a linear system has only discrete data throughout the system, its dynamics can be described by a set of discrete state equations
respectively.
x[(k
1)7]
= Ax(kT) +
Dx(kT)
Br(A:r)
(4447)
+ Er(kT)
same form
is
(4448)
where A, B, D, and
difference in the
basically of the
is
as Eq. (4431).
The only
two
situations
$(T) and
8(7") are
and
B matrices
an outsignals.
itself represents
which has only discrete The solution of Eq. (4447) follows directly from that of Eq.
is
(4431).
written
(4449)
x(kT)
A*x(0)
A*''BrOT)
i
where
A*
= AAAA...A
k
(4450)
4.20
The
x[(k
1)7"]
= Ax(kT) +
Br(&r)
(4451)
can be solved by means of the ztransform method. Taking the ztransform on both sides of Eq. (4451) yields
zX(z)
zx(0+)
= AX(z) + BR(z) +
(zl
(4452)
(zl
A)" zx(0r)
A)" 'BR(z)
(4453)
166
StateVariable Characterization of
Dynamic Systems
Chap. 4
The
is
x(kT)
g '[(zl
A)" z]x(0)
l
g*[(zl
 A)
BR(z)]
(4454)
In order to carry out the inverse ztransform operation of the last equation,
write the ztransform of
we
A*
k=
as
g(A")
f] A*z"*
=I+
Az"
+A
z" 2
(4455)
Az 1 and
subtracting the
from the
last equation,
we
get
(Ig(A k )
or
Az')S(A*)
(4456)
=
=
(I
Az"
)"
=
1
(zl
A)"'z
(4457)
A*
[{zl
A
(4458)
Equation (4458) also represents a way of finding A* by using the ztransform method. Similarly, we can prove that
5t(2l
 A)'BR(z)] = S
and
A^''BrOT)
(4459)
Now we
x(kT)
A*x(0)
+ 2 =
i
A*''Br(ir)
(4460)
which is identical to the expression in Eq. (4449). Once a discretedata system is represented by the dynamic equations of Eqs. (4447) and (4448), the transfer function relation of the system can be
expressed in terms of the coefficient matrices.
Setting the initial state
x(0+)
X(z)
A)BR(z)
(4461)
When
this
equation
is
we have
C(z)
[D(zl
A)" 'B
+
is
E]R(z)
(4462)
Thus the
G(z)
D(zl
A)'B
(4463)
D[adj(zIA)]B
is
+ lzIAE
{A _
m)
The
defined as
(4465)
zIA
In general, a linear timeinvariant discretedata system with one input and one output can be described by the following linear difference equation with
constant coefficients:
Sec. 4.21
167
c[(k
1071
fl,c[(fc
VP + m)T]
+ n  2)71 + + a _ lC [(k + l)r] + a c(kT) + VP + m 1)7] + n>m + b m . A(k + 1)T] + b m r{kT) + nl)r]
a 2 c[(k
(4466)
sides
is
written
l
C{z)
R(z)
,. v
46? ;
The
fljz" 1
a_ ,z
(4468)
Example 433
is
2)
5c(k
1)
+
+
3c(/c)
r(k
1)
2r(/t)
(4469)
initial
z^C{z)
5zC(z)
3C(z)
= zR(z) +
(4470)
From
easily written
R(z)
z2
5z
+ =
^ *' l)
The
characteristic equation
is
5z
(4472)
The
Xl (k)
XiQc)
Substitution of the last
(4469) gives the
= c(k) = x,{k +
(4473)
1)
 r(k)
(4474)
two
two
system as
(4475)
Xl (k
x 2 (k
from which we have the
3r{k)
(4476)
(4477)
3
5_
is
The same
obtained by using zl

A = 0.
4.21
When
a discretedata system
is
state equations,
may
168
StateVariable Characterization of
Dynamic Systems
Chap. 4
Some of the
operations
of a digital computer are multiplication by a constant, addition of several machine variables, time delay, or shifting. The mathematical descriptions of
these basic digital computations
sions are as follows:
and
1.
Multiplication by a constant:
x 2 (kT)
2
(4478)
(4479)
2.
Summing:
x 2 (kT)
(4480)
X (z)
2
(4481)
3.
x 2 (kT)
2
(4482) (4483)
or
X^z)
The
a
*,(<)+)
(4484)
state
(z)
o
OX
(z)
The
initial
t
time
rj.
= 0+
can be generalized to
Then Eq.
X2 (z) = aX X
(z)
(z)
(4484)
is
written
Z,(z)
z'
X (z) + x, {n)
2
(4485)
X2 (z)
Example 434
c(k
2)
5c(k
1)
3c(k)
r(k
1)
2r(k)
(4486)
(z)
One way of
X2 (z) = XQ (z) + X
(z)
system
is
9*i(0+)
x,(k
x 2 (k
+ +
1) 1)
(4487)
2 (k)
3r(k)
(4488)
X2 (z)o
X
1
O^(r)
= ziX 2 (z) + x 1 (0+)
Using
essentially the
and (4488)
z
_1
is
is
used to relate
x^k +
will
always appear as
The time delay unit The state variables outputs of the delay units on the state
1) to Xi(k).
diagram.
Sec. 4.21
169
R{z)
C{z)
3
diagram of the system described by the difference equation of Eq. (4486) or by the state equations of Eqs. (4487) and
Fig. 430. Discrete state
(4488).
As an alternative, the state diagram can also be drawn directly from the difference equation by means of the decomposition schemes. The decomposition of a discrete transfer function will be discussed in the following section, after we have demonstrated some of the practical applications of the discrete state diagram.
state transition equation of the system can be obtained directly from the {z) and diagram using the gain formula. Referring to z (z) as the output nodes and to xi(0+), * 2 (0+), and R(z) as input nodes in Fig. 430, the state transition equations are written in the following vectormatrix form:
The
state
~ A
where
"1
_
+5z'
3Z1
"*i(o+r
1
rzi(l
+5z')3z 2
R{z)
(4489)
_* 2 (0+)_
A_
_ 3z i_3 z 2
3z
A=
5z'
(4490)
The same transfer function between R(z) and C(z) as in Eq. (4471) can be obtained directly from the state diagram by applying the gain formula between these two nodes.
Decomposition of Discrete Transfer Functions
The
systems can be applied to transfer functions of discretedata systems without the need of modification. As an illustrative example, the following transfer
function
is
diagrams are
decomposed by the three methods, and the corresponding shown in Fig. 431
<2f)
^
state
(4.491)
Equation (4491) is used for direct decomposition after the numerator and the denominator are both multiplied by z~ 2 For cascade decomposition, the trans.
x 2 (0+)
Riz)
C(z)
3
(a) Direct
decomposition
x, (0 +)
R(z)
C(z)
(b)
Cascade decomposition
*i(0+)
RU)
(c) Parallel
decomposition
= (z
2)/(z 2
+
170
5z
3)
tion, (b)
by the three methods of decomposition, (a) Direct decomposiCascade decomposition, (c) Parallel decomposition.
Sec. 4.22
form
z
as
C{z)
R(z)
(z
+ 4.3)(z + 0.7)
(4492)
For the
by
form
0.64
,
C(z) R(z)
4.22
~ z + 4.3 +
0.36
0.7
(4493)
When
is
desired for
all
times.
diagram of the zeroorder hold. Consider that the input of the zeroorder hold is denoted by e*{i) which is a train of impulses, and the output by h{t). Since the zeroorder hold simply holds the magnitude of the input impulse at the sampling instant until the next input comes along, the signal h{t) is a sequence of steps. The inputoutput relation
Let us
first
in the Laplace
domain
is
written
H(s)
In the time domain, the relation
e*
E*(s)
(4494)
is
simply
h(t)
e(kT+)
(4495)
for
kT<t<(k +
this
l)T.
we need
e(kT+). For
e(kT+)
H(s)
= e(fcr+)
The
state
is
(4496)
O
Fig. 432. State
for
kT < t <
illustrative
(k
\)T.
diagram rep
shown
As
an
state
let
diagram of a
sampleddata system
constructed,
us consider the
We
shall
demonstrate the
written
(4497)
is
=
s s
Fig. 433.
Sampleddata system.
172
/ State Variable Characterization of
Dynamic Systems
Chap. 4
(4498)
e(t)
state
technique. Figure 434 illustrates the discrete state diagram of he system through
x, (0 +
e(kT+)
Fig. 433.
from
this state
diagram
*,[(*
= c(kT) =
1)7]
e T x,(kT)
(1
e T)e{kT)
(4499)
x (kT)
x
(4500)
Therefore, the output response of the system can also be obtained by solving
the difference equation of Eq. (4499).
If the
c(t) is
desired for
all
t,
diagram shown in Fig. 435. This state diagram is obtained by cascading the state diagram representations of the zeroorder hold and the process G(s).
state
x (kT+)
{
,'
o
His)
*>
e(kT+)
interval
kT<t<(k +
Sec. 4.23
173
To
determine
c(t),
which
is
also *i(f)>
we must
first
We
have
(4 " 501)
*(*>
for
kT <
< (& +
l)r.
last
equa
tion gives
*i(0
e
' kT>
]e(kT+)
e *"jc,(Jtr)
t is
kT< < (k +
t
one sampling period, whereas the x,(r) only at the sampling instants.
in Eq. (4502), the latter
information on
let t
if
we
{k
Y)T
becomes Eq.
(4499).
4.23
When
it
^
c(?)
= =
A(t)x(t)
+ +
B(/)r
E(r)r(?)
(4503) (4504)
D(0x(r)
where
x(r)
r (0
c (0
A{t), B(/), D(?),
=n =P =9
X x x
state vector
input vector
output vector
and E(t) are coefficient matrices of appropriate dimensions. The elements of these coefficient matrices are functions of t.
Unlike in the timeinvariant case, timevarying differential equations do not have closedform solutions. Let us investigate the properties
generally
^ ^
first
a(t)x(t)
(4505)
a{t)dt
(5506)
sides to get
In x(t)
In x(t Q )
=
J'
a(z)
dx
(4507)
Therefore,
x{t)
exp
a{x) dx\x(t a )
(4508)
where
denotes the
initial
time.
we can define a state transition matrix for the timevarying state equation. For the scalar case under consideraJust as in the timeinvariant situation,
174
StateVariable Characterization of
Dynamic Systems
Chap. 4
matrix
is
<f>(t, t 9 )
exp
a(x) dx
(4509)
Notice that for the timevarying case, the state transition matrix depends upon
t
state equation
(t)
A(0x(r)
(4510)
it is
ftr,
)x(t
(4511)
) is
satisfies
is
in general.
The question
related
exp
let
H"
A(t) dx
(4512)
To answer
into a
power
series,
exp T
A(t) dx)
=I+
A(t) dx
+^
[ A(t) dx f
A(<r) </*
...
(4513)
*[T.
A(t) dx
A(0
m L'
=
A(0
A(t)
do
j
A(t)
rft
A(f)
(4514)
we have
j"
A(0 exp
["
('
A(t) dx
A(0
A(t) dx
(4515)
By comparison of Eqs.
(4514)
and
(4515),
we
see that
^ exp
or
P
J
A(t) dx
= =
A(0 exp
A(t) Jt
(4516)
${t,
dt
if
/ )
A(0<K?,
* o)
(4517)
and only
if
A(r)
f A(t) dx
to
f J la
'
A(t) dx A(t)
(4518)
that
is,
A(0
commute.
and
f
to
A(t)
rfr
The requirement
that A(r)
and
its
integral
commute
is
evidently a very
(4512) will
not be valid.
Sec. 4.23
175
Most of
4>(t
t ),
1.
<KMo)
1
=
)
I
2.
3.
<f.
(?,?
,
<f.(/
,0.
)
${t 2
t )
for any
t2
Solution of the
moment, we
${t, to)f[{t)
(4519)
where
satisfies
an n x 1 vector, and <^(/, t ) is the state transition matrix that Eq. (4517). Equation (4519) must satisfy Eq. (4503). Substitution of Eq. (4519) into Eq. (4503) yields
r\(t) is
4>U,
toMO + W,
t )r\(t)
(4520)
we
get (4521)
t,
*oM) = Bu(0
B(f)u(0
Thus
f(?)
<f"'(f, t )
(4522)
and
0 =
The vector
r\(t ) is
f *
_i
(t,
?o)B(t)u(t) dx
tl(f)
(4523)
Thus
substitut
we have
= W,
W,
r
)x(t
+
r )
$(r,
$(*,
)B(t)u(t)
rft
(4524)
Since
)*'(T,
= W, /)<K'o, t) =
)
<K', t)
(4525)
ftf,
)x(r
f'
$(t, t)B(t)u(t)
dt
(4526)
which
is
timevarying systems can be solved by using not readily available. It is possible to discretize the system with a time increment during which the timevarying parameters do not vary appreciably. Then the problem becomes that of solving a set of
many
is
t )
One method of
by
discretizing the
system
is
176
Dynamic Systems
Chap. 4
i(t)
~ L{x[(k +
l)T]
x(kT)}
kT<t<(k+
l)T
is
(4527)
where T is a small time interval. The state equation of Eq. mated by the timevarying difference equation
x[(k
(4503)
approxi
= A*{kT)x(kT) + B*(kT)r(kT) over the time interval, kT < < (k + l)T, where A.*(kT) = TXQcT) + I
+
l)T]
t
(4528)
B*(kT)
TB(kT)
much
the same
way
as in the time
REFERENCES
State Variables and State Equations
1.
L. A. Zadeh, "An Introduction to State Space Techniques," Workshop on Techniques for Control Systems, Proceedings, Joint Automatic Control Conference, Boulder, Colo., 1962.
2.
B. C.
New
York, 1967.
3.
D. W. Wiberg, Theory and Problems of State Space and Linear Systems (Schaum's Outline Series), McGrawHill Book Company, New York, 1971.
R. B. Kirchner,
"An
Explicit
Formula
W. Everling, "On
p. 413,
the Evaluation of e A by
'
Power
Series," Proc.
IEEE, Vol.
55,
Mar. 1967.
by Truncated Power
6.
T. A. Bickart, "Matrix Exponential: Approximation Series," Proc. IEEE, Vol. 56, pp. 872873, May 1968. T.
7.
Amer.
M.
of Evaluating eA!
in
Closed Form,"
IEEE
A C. G. Cullen, "Remarks on Computing e '," Vol. AC16, pp. 9495, Feb. 1971.
J.
IEEE
10.
"An Algorithm for the Computation of the IEEE Trans. Automatic Control, Vol.
Chap. 4
References
177
11.
Vol. 120,
Transformations
12.
C. D. Johnson and
ical
the Transformation to
Canon
IEEE
H. Mufti, "On the Reduction of a System to Canonical (Phase Variable) Form," IEEE Trans. Automatic Control, Vol. AC10, pp. 206207, Apr. 1965.
14.
IEEE Trans.
15.
to (Phase Variable) Canonical Form," Automatic Control, Vol. AC10, pp. 492495, Oct. 1965.
L.
M. Silverman, "Transformation
IEEE
W. G. Tuel, Jr., "On the Transformation to (Phase Variable) Canonical Form," IEEE Trans. Automatic Control, Vol. AC11, p. 607, July 1966.
D. G. Luenberger, "Canonical Forms for Linear Multivariate Systems," IEEE Trans. Automatic Control, Vol. AC12, pp. 290293, June 1967.
S. J.
17.
18.
tems,"
19.
B. Ramaswami and Ramar, "Transformation to the Phase Variable Canonical Form," IEEE Trans. Automatic Control, Vol. AC13, pp. 746747, Dec. 1968.
20.
W. B. Rubin, "A Simple Method for Finding the Jordan Form of IEEE Trans. Automatic Control, Vol. AC17, pp. 145146, Feb. 1972.
K. Ogata, State Space Analysis of Control Systems, PrenticeHall,
a Matrix,"
21.
Inc., Engle
wood
State
22.
Cliffs, N.J.,
1967.
Diagram
B. C.
Kuo,
"State Transition
Systems,"
Controllability
23.
WESCON Convention
and Observability
Y. C. Ho, "What Constitutes a Controllable System?" Control, Vol. AC7, p. 76, Apr. 1962.
IRE
Trans. Automatic
24.
R. E. Kalman, Y. C. Ho, and K. S. Narendra, "Controllability of Linear Dynamical Systems," Contribution to Differential Equations, Vol. 1, No. 2, pp.
189213, 1962.
25.
E. G. Gilbert, "Controllability and Observability in Multivariate Control Systems,"/. SIAM Control, Vol. 1, pp. 128151, 1963.
L. A. Zadeh and C. A. Desoer, Linear System Theory, McGrawHill Book Company, New York, 1963.
26.
27.
R. E. Kalman, "Mathematical Description of Linear Dynamical Systems," Soc. Ind. Appl. Math., Vol. II, No. 1, Ser. A, pp. 151192, 1963.
/.
178
Dynamic Systems
Chap. 4
28.
E.
Kreindler and
P.
ability
of Linear Systems,"
Sarachik, "On the Concept of Controllability and ObservIEEE Trans. Automatic Control, Vol. AC9, pp.
A. R. Stubberud,
"A
IEEE
30.
Trans. Application
and
Nov. 1964.
R. W. Brockett, "Poles, Zeros, and Feedback: State Space Interpretation," IEEE Trans. Automatic Control, Vol. AC10, pp. 129135, Apr. 1965.
R. D. Bonnell,
31.
"An
IEEE Trans.
System/360 Continuous System Modeling Program (360ACX16X) User's Manual, Technical Publications Dept., International Business Machines Corporation,
White
Plains,
N.Y.
PROBLEMS
4.1.
electric
networks shown
in Fig. P41
e(t)
4.2.
The following
Chap. 4
Problems
179
(a)
*!)
(b)
^
rf
dt 1
+ 3*0
4
rfc(/)
c(t)
r(t)
5c(t)
= =
r(t)
(c)
c(Q
dr(t)
c{t)
(d)
rfM)
(e)
^
<ft
"+
c(t)
c(T)dT
/(/)
f.
1_u +
"
rf/
6^ +
rfr
5c(0
that
= ,H0 +
<fr
2^
,(/)
4.3.
Using Eq.
(442),
show
((>(0
Ar
Klf2 . T\ ji^
r
J,,
4.4.
Ax(0
+ B(0
for the following cases:
<((/)
0"
"1"
1
(b)
2_
B =
_1_
r
2
3_
0"
"0"
B=
_1_
(c)
2
1
2_
(d)
1
B =
0"
1
"10"
_ 1_
ro
1
4.5.
B=
l
1
Find the
for
4.6.
t
Problem 4.4
> 0.
is
is
Given the
where
state equation
(t)
= Ax(0 +
r
1
1
B(0
~0
A=
B
becomes
_2
0_
y(0
where A! and
4.7.
A,x(/)
B!(0
For the
Problem
4.6, if
B=
can the state equation be transformed into the phase variable form? Explain. Given the state equations of a linear timeinvariant system as
x(/)
4.8.
= Ax(?) +
B(0
180
StateVariable Characterization of
Dynamic Systems
Chap. 4
where
1
0~
1
ro
B=
l
2
eigenvalues of A.
4.9.
3.
determine the transfer function relation between X(s) and U(s). Find the
For a linear timeinvariant system whose state equations have coefficient matrices given by Eqs. (4111) and (4112) (phasevariable canonical form),
show
that
1
adj (si
 A)B =
and the
characteristic equation
s"
is
+ ais"
1
+ a 2 s" 2 +
is
+ aa is +
a == 0.
4.10.
A closedloop control
system
(t)
described by
where
r
x(t)
= //vector,
matrix.
n,
is
r,
and
is
the
x n feedback
eigenvalues of
A BG.
0"
1
1
A=
2
B=
10_
g*\
5
gi
G=
[i
g2
Find the characteristic equation of the closedloop system. Determine the elements of G so that the eigenvalues of A BG are at 1, 2, 1 jl, and
1
4.11.
+/1. Can
all
the eigenvalues of
A BG
problem?
is
tion
d 2 c(t)
dt 2
(a)
Ml)
dt
c(t)
= r(f)
Find the
<J>(/).
1,
t(0)
= 0,
and
r(t)
4.12.
system
is
equations
Chap. 4
Problems / 181
rf'cKQ
rf/
2
+
,
dc,(rt
rf/
+ 2c,(02c 2 (0=r,(0
ci(0
dt 2
(a)
c 2 (0
r 2 (f)
Write the state equations of the system in vectormatrix form. Write the output equation in vectormatrix form.
(b)
Find the transfer function between the outputs and the inputs of the system.
state transition
4.13.
Given the
equation (t)
= Ax(/), where
a
A=
a and
(a)
jo
co are real
numbers.
<^(/).
(b)
(c)
Find the state transition matrix Find the eigenvalues of A. Find the eigenvectors of A.
state equations of
4.14.
Given the
a linear system as
(t)
= Ax(/) + B(0
1
where on
1
ro
B=
_1.
6
11
6J
1,X 2
2, X 3
A into
a diagonal matrix
4.15.
Given a
where
1(0
'
= Ax(f) +
1
Bm(0
0'
1
B=
ri
_1_
.25
The eigenvalues are X
x(/)
x
35
I, X 2
11_
5. Find the transformation transformed into the Jordan canonical form. The
5, X 3
HO = Ay(0 + T(0 A
and T.
diagrams for the following systems
Bu(t)
Draw
(a)
state
(0
= Ax(/) +
3 A = 1 5
(b) (r)
0"
1
ro
1
B=
l
2
1
Ax(f)
4.17.
P417.
182
StateVariable Characterization of
Dynamic Systems
Chap. 4
Figure P417.
(b)
(c)
Draw
a state diagram for the system. Find the state transition equations for the system. Express the equations in matrix form. The initial states are represented by x(t'b), and the input r{t) is a unit step function, u s {t / ), which is applied at t == t
.
4.18.
Draw
(a)
state
decomposition
G(s)
5s 2
+ 4s +
1)
10
<
b>
W = s(s + 1F+
state
:
6(s
3)
Write the state equations from the state diagrams and express them in the phasevariable canonical form.
4.19.
Draw
(a)
position G(s)
s(s
+ 1) + 2)(s +
3)
<i3P + ^>
Draw
5c(0
f +
Write the state equations from the state diagrams and show that the states are decoupled from each other.
4.20.
state
decomposition.
4.21.
linear system,
Draw
4.22.
state
position.
The
diagrams for the system using three different methods of decomstate diagrams should contain a minimum number of integrators.
is
The
state
shown
in Fig. P422.
Figure P422.
Chap. 4
Problems / 183
(a)
Assign the state variables and write the dynamic equations of the system.
(b)
4.23.
Draw
The
electric
network shown
is
in Fig. P41.
4.24.
state
shown
in Fig. P424.
Figure P424.
Assign state variables on the state diagram; create additional nodes is not altered. (b) Write the dynamic equations for the system.
(a)
if
4.25.
Given the
where
state equation
(0
= Ax(?)
1
2
0"
1
2
.0
(a)
2_
(b)
<J)(0
4.26.
(/)
=
1
Ax(t)
B(0
on
1
ro
B =
_1.
3
2
4
3J
A are Ai = 1, A 2 = 1 j\, A = 1 +jl. Find the transformation x(?) = Py</) which transforms A into the modal form
The eigenvalues of "1
0"
1 1
4.27.
=P"iAP
1_
Given the
linear system
(0
where
(/) is
= Ax(0 +
B(0
= Gx(0
is
1
The
state transition
<J>(/)
e (ABG)
![(,,!
_ a + BG)
e (ABG)t
gAtgBGt
184
Dynamic Systems
Chap. 4
where
e At e BGr
Determine the
system shown
in Fig. P428.
Figure P428.
(a)
(b)
a, b, c,
is
not
4.29.
Figure P429 shows the block diagram of a feedback control system. Determine
the state controllability
and
by the following
It
\X
3
I
\
.9
x
1
\
+ 2
' '
x2
(a)
matrices.
Coupling of
states.
is
4.30.
The
given by
C(s)
s s3
+a
6
is
R(s)~
(a)
+6s 2 + Us +
either uncontrollable or
one of them
is
uncontrollable.
is
unobservable.
4.31.
= Ax(0 + B(0
Chap. 4
Problems
185
where
"
r
a_
"1"
B =
Ll
Find the region
controllable.
4.32.
L*J
such that the system
is
completely
Draw the state diagram of a secondorder system that is neither controllable nor
observable.
4.33.
b\,
b2
d u and d2
Bk(/)
is
(0
= Ax(0 + ~i r
_o
i_
c(0
=
[/,
Dx(0 d2 ]
D=
4.34.
The block diagram of a simplified control system for the Large Space Telescope (LST) is shown in Fig. P434. For simulation and control purposes, it would be desirable to represent the system by state equations and a state diagram.
Gimbai
controller
K,
~\+ K
K,
Command
Kp s +
s s
+ Kj
1
H
Js 2
position
J
JG s
Vehicle
dynamics
KN
1
s
Figure P434.
(a)
Draw a state diagram for the system and write the state equations in
matrix form.
vector
(b)
(c)
Find the characteristic equation of the system. modern control design scheme, called the state feedback, utilizes the concept of feeding back every state variable through a constant gain. In this case the control law is described by
 gjXi g 2 x 2 g]X  g x Xi
3
,
Find the values of gi, g 2 gi, and g^ such that the eigenvalues of the overall system are at s = 100, 200, 1 +/1, 1 jl. The system parameters = 600, Kj = 9700, JG = 2, Jv = 10 5 Kp = 216, and are given as
KN = 300.
186
/
StateVariable Characterization of
Dynamic Systems
Chap. 4
4.35.
The
difference equation of
c[(k
is
given by
2)T]
+ 0.5c[(A: +
1)T]
0.1c(kT)
(a)
(b)
The k =
initial
2, 3,
10 by
= 1 and c(T) = 0. Find c{kT) for means of recursion. Can you project the final value of
state equations,
1) 1)
x 2 (k
+ +
2 (k)
+ r(k)
Kz
(z
l)(z 2
z
3)
Draw
(b)
a state diagram for the system. Write the dynamic equation for the system
in vectormatrix form..
is
4.38.
shown
in Fig. P438.
>it)
tO ^T
e(t)
 e*(0
z.o.h.
2Q +
s(s
0.5)
+ 0.2)
c(t)
*
Figure P438.
(a)
Draw
(b)
a state diagram for the system. Write the state equations in vectormatrix form,
x[(fc
1)71
= <J>(r)x(/cD + B(T)r(kT)
(c)
T=
0.1 sec.
5
Mathematical Modeling
of Physical
Systems
5.1
Introduction
One
of the most important tasks in the analysis and design of control systems
is
duced a number of wellknown methods of modeling linear most common methods are the transfer function approach and the statevariable approach. However, in reality most physical systems have nonlinear characteristics to
some
extent.
if
linear control systems have been well developed, their counterparts for nonlinear systems are usually quite complex. Therefore, the control systems engineer often has the task of determining not
how to accurately describe a system mathematically, but, more important, to make proper assumptions and approximations, whenever necessary, so that the system may be adequately characterized by a linear mathematical model.
only
how
important to point out that the modern control engineer should place on the mathematical modeling of the system so that the analysis and design problems can be adaptable for computer solutions. Therefore, the main objectives of this chapter are
It is
special emphasis
1.
2.
To demonstrate the mathematical modeling of control systems and components. To demonstrate how the modeling will lead to computer solutions.
187
188
Chap. 5
The modeling of many system components and control systems will be However, the emphasis is placed on the approach to the problem, and no attempt is being made to cover all possible types of systems
illustrated in this chapter.
encountered in practice.
5.2
The
classical
way of
is
the
loop method and the node method, which are formulated from the two laws of Kirchhoff. However, although the loop and node equations are easy to write,
they are not natural for computer solutions.
network equations
in this section.
is
networks
More detailed discussions on the state may be found in texts on network theory. 12
RLC network
R
+
of Fig. 51 to
L
T1KP
V\AA<
HO
+
e(t)
J?
e c (t)
Fig. 51.
RLC network.
It is relatively
^l + R d = Ld
is
f + L q
(t)
(51)
where
q(t)
is
i(t)
by
(52)
q{t)=\'
It is
shown
in
differential
equation in
Eq. (51) can be replaced by two firstorder differential equations called the state equations. In this case it is convenient to define the state variables as
Xl (t)
= ^P =
=
e c (t)
(53)
where ec (t)
is
and
(54)
* 2 (0
/(0
= C^)
= L **M + Rx
2 (t)
Xl (t)
(55)
Sec. 5.2
Thus, from Eqs. (54) and (55) the state equations of the network are
dx
(t)
dt
*,<0
(56)
^r in the inductor L,
variables.
i(t),
x,(t)
Rx
2 (t)
e(t)
(57)
A more direct way of arriving at the state equations is to assign the current
and the voltage across the capacitor C,
e c (t), as the state
Then
C and
way
the voltage across L in terms of the state variables and the input source. This
the state equations are written by inspection from the network. Therefore,
Current in C:
Voltage across
Since x^t)
C^4^ = dt
L
L di(f)_
i(i), it is
i(t)
(58)
e c (t)
Ri(t)
e(t)
(59)
e c (t)
and x 2 (t)
and
(57).
In general,
it is
currents in the inductors as state variables in an electric network, although there are exceptions. 12
One must
electric
recognize that the basic laws used in writing state equations for
still the KirchhofFs laws. Although the state equations in Eqs. (58) and (59) are arrived at by inspection, in general, the inspection method does not always work, especially for complicated networks. However, a general
networks are
method using
Example
51
is
available. 1
As another example of writing the state equations of an electric network, consider the network shown in Fig. 52. According to the
foregoing discussion, the voltage across the capacitor e c and the cur/, and i2 are assigned as state variables, as shown in Fig. 52.
Fig. 52.
Network
in
Example
51.
The state equations of the network are obtained by writing the voltages across the inductors and the currents in the capacitor in terms of the three state variables. The
state equations are
e(.t)
(510)
L2
dh(t)
dt
Rihit)
+ e (t)
c
(511)
190
Chap. 5
c^ = mm
Rearranging the constant
canonical form:
coefficients, the state
(512)
diAtY]
dt
_*i
+
u
1
hit)
dhU)
dt

C
hit)
e(t)
(513)
de c (t)
dt
edt)
5.3
Most feedback
ponents.
From
comand
mechanical elements are analogous. In fact, we can show that given an electrical device, there is usually an analogous mechanical counterpart, and vice versa. The analogy, of course, is a mathematical one that is, two systems are analo;
they are described mathematically by similar equations. The motion of mechanical elements can be described in various dimensions as translational, rotational, or a combination of both. The equations governing the motions of mechanical systems are often directly or indirectly formulated
if
Motion
The motion of translation is defined as a motion that takes place along a The variables that are used to describe translational motion are acceleration, velocity, and displacement. Newton's law of motion states that the algebraic sum offorces acting on a rigid body in a given direction is equal to the product of the mass of the body and its acceleration in the same direction. The law can be expressed as
straight line.
forces
= Ma
(514)
denotes the mass and a is the acceleration in the direction considered. where For translational motion, the following elements are usually involved:
1.
Mass: Mass is considered as an indication of the property of an element which stores the kinetic energy of translational motion. It is denotes the analogous to inductance of electrical networks. If
is
given by
M
where g
(514)
is
W
g
(515)
body due
sets
tion of free
Three consistent
and
Sec. 5.3
191
Units
Mass
Weight
Acceleration
Force
MKS
CGS
British
newtons/m/sec 2
dynes/cm/sec 2
lb/ft/sec 2 (slug)
newton dyne
lb
m/sec 2
cm/sec 2
ft/sec 2
newton dyne
lb
y(t)
is
is
M
Fig. 53.
fit)
written
fit)
Ma(t)
= M<?M
Mdv(t)
r
(516)
Forcemass system.
dt
where
a{t)
is
and
the acceleration,
applied force.
2.
Linear spring
to be
considered
an element that
extent.
is
analogous to a
is
some
Ky{t)
(517)
where
is
The
three unit
K
newtons/m dynes/cm
lb/ft
MKS
CGS
British
is
directly
K

y{t)
is
AAAAA/v
fit)
f{t)T=Ry{t)
Whenever
there
is
(518)
motion or tendency
The
frictional forces
encountered in physical systems are usually of a nonlinear nature. The characteristics of the frictional forces between two contacting surfaces often depend
on such factors as the composition of the surfaces, the pressure between the surfaces, their relative velocity, and others, so that an exact mathematical description of the frictional force is difficult. However, for practical purposes,
frictional forces
friction,
192
Chap. 5
static friction,
and Coulomb
friction.
in
the following.
1.
is
The The
is
often repre
sented
by a dashpot such
H
Fig. 55.
f(f)
B&
(519)
Dashpot
where
is
The dimen
sions of
Units
B
newton/m/sec
dyne/cm/sec
lb/ft/sec
MKS
CGS
British
and
velocity.
to prevent
can be
(F,),_
(520)
where (^)^,
is
when
the
body
is
The
sign
of the friction
initial
is
direction of velocity.
The
static
and other frictions take over. Coulomb friction. Coulomb friction is a retarding force that has a
f
+ F.
Slope =
(b)
(c)
forces,
Viscous
Coulomb
friction.
Sec. 5.3
193
constant amplitude with respect to the change in velocity, but the sign of the frictional force changes with the reversal of the direction of velocity. The mathematical relation for the Coulomb friction is
given by
/w'GF/ISD
where
<5  2,)
is
the
Coulomb
friction coefficient.
is
Rotational Motion
axis.
The rotational motion of a body may be defined as motion about a fixed The variables generally used to describe the motion of rotation are torque; angular acceleration, a; angular velocity, <y; and angular displacement, 6. The
following elements are usually involved with the rotational motion.
Inertia. Inertia, /, is considered as an indication of the property of an element which stores the kinetic energy of rotational motion. The inertia of a given element depends on the geometric composition about the axis of rotation
and
its
density.
For
its
geo
metric axis
given by
J
where
= \Mr
shaft
(522)
r is its radius.
is
the
and
Example
52
is 1 in.
and weighing 5
inertia
is
Wr 2
g
^ =
(5oz)(l
in) 2
386 in/sec 2
(523)
0.00647 ozinsec 2
given in weight per unit volume. Then, for a inertia is proportional to the fourth
length. Therefore,
if
is
W
p
is
the
is
expressed as
W = p(nr
where
length, then Eq. (522)
is
2 h)
(524)
pn^ = omQ6p h/
(525)
(525)
becomes
(526)
J
For aluminum, p
is
= 0.0184 hr 4
Eq. (525) becomes
1.56 oz/in 3
= 0.00636 hr*
(527)
194
Chap. 5
When
T(^
<' >
a torque
is
as
shown
written
(528)
no = Mt) = J ^r = J ~$P
The
tities
Units
Inertia
Torque
Angular Displacement
radian radian
MKS
CGS
English
kgm 2
gcm 2
slugft 2
newtonm dynecm
lbft
or lbftsec 2
ozinsec 2
ozm
radian
found useful
rad
= 11? =
57.3
Angular
velocity.
1
rpm rpm
= ? = 0.1047 rad/sec oU
=
6 deg/sec
Torque.
1
= lbft = ozin =
gcm
0.0139 ozin
192 ozin
0.00521 lbft
Inertia.
1
gcm 2
2
1.417
10" 5 ozinsec 2
lbftsec
=192
ozinsec 2
ozinsec 2
gcmsec 2
1
lbftsec
= = =
Torsional spring.
As with
7X0
of a rod or a shaft
torque.
when
it
is
subject to
an applied
M/WW
0(0
Fig. 58. Torquetorsional spring
Figure 58 illustrates a simple torquespring system that can be represented by the following equa
tion:
system.
7X0
KB{t)
(529)
Sec. 5.3
195
The dimension
for
K
newtonm/rad dynecm/rad
ozin/rad
MKS
CGS
British
If the torsional spring is
is
modified to
TP= Kd(t)
(530)
Eqs. (519),
parts
:
The three types of friction described for motion can be carried over to the motion of rotation. Therefore, (520), and (521) can be replaced, respectively, by their counter
no _
n dftt)
dt
(531)
(532) (533)
no
where
C\)<to
{F )d=0
s
B is
is
and
is
the
Coulomb
friction coefficient.
Relation
Between
Translational
For
instance, a load
may be
controlled to
momove
along a straight line through a rotary motor and screw assembly, such as that shown in Fig. 59. Figure 510 shows a similar situation in which a rack and
no,
Motor
HO
W
x(0
x(0
/VWWWWWWWWWWWWWVAA
Drive
motor
196
Chap. 5
pinion
control
as that
all
is is
used as the mechanical linkage. Another common system in motion the control of a mass through a pulley by a rotary prime mover, such
in Fig. 511.
shown
and
511
can
For instance, the mass in Fig. 51 1 can be regarded as mass which moves about the pulley, which has a radius r. Disregarding a point the inertia of the pulley, the equivalent inertia that the motor sees is
directly to the drive motor.
/=M/ 2 =
If the radius
(534)
is r, the equivalent inertia which the by Eq. (534). Now consider the system of Fig. 59. The lead of the screw, L, is defined as the linear distance which the mass travels per revolution of the screw. In principle, the two systems in Fig. 510 and 511 are equivalent. In Fig. 510, the distance traveled by the mass per revolution of the pinion is 2nr. Therefore,
motor
using Eq. (534) as the equivalent inertia for the system of Fig. 59,
2
7
where in the British system
= fte)
2
)
(5 " 35)
= inertia (ozinsec W weight (oz) L = screw lead (in) g = gravitational force (386.4 in/sec
/
Mechanical Energy and Power
and potential energy controls the dynamics of the system, whereas dissipative energy usually is spent in the form of heat, which must be closely controlled. The mass or inertia of a body indicates its ability to store kinetic energy. The kinetic energy of a moving mass with a velocity v is
kinetic
form of
Wk = \Mv
The following
tion:
(536)
consistent sets of units are given for the kinetic energy rela
Sec. 5.3
197
Units
Energy
joule or
Mass
newton/m/sec 2
dynecmsec 2
lb/ft/sec 2
Velocity
MKS
CGS
British
m/sec
cm/sec
ft/sec
newtonm dynecm
ftlb
(slug)
is
written
(537)
W = ya>
k
where /
is
the
moment
of intertia and
co the
Units
Energy
joule or
Inertia
Angular Velocity
rad/sec
MKS
CGS
British
kgm 2
newtonm dynecm
ozin
gmcm 2
ozinsec 2
rad/sec
rad/sec
work required
by y
Potential energy stored in a mechanical element represents the amount of to change the configuration. For a linear spring that is deformed
in length, the potential energy stored in the spring
is
W = \Ky*
p
(538)
where
stored
K
is
the spring constant. For a torsional spring, the potential energy given by
is
Wp = \KQ*
When
(539)
dealing with a frictional element, the form of energy differs from the previous two cases in that the energy represents a loss or dissipation by the sys
tem in overcoming the frictional force. Power is the time rate of doing work. Therefore, the power dissipated in a frictional element is the product of force
and
velocity; that
is,
P=fi>
Since
(540)
/=
Bv, where
B
is
is
P
The
Bv*
(541)
MKS
unit for
power
CGS
system
it is
represented in ftlb/sec or
horsepower
(hp).
Furthermore,
1
hp
(542)
Since power
is
W = BJv
d
dt
(543)
198
Chap. 5
Gear
Trains, Levers,
a mechanical device
from one part of a system to another in such a way that force, torque, speed, and displacement are altered. These devices may also be regarded as matching devices used to attain maximum power transfer. Two gears are shown coupled together in Fig. 512. The inertia and friction of the
that transmits energy
0,
The relationships between the torques 7^ and T2 angular displacements and 2 and the teeth numbers JV", and N2 of the gear train are derived from
, > ,
The number of teeth on the surface of the gears radii r, and r 2 of the gears; that is,
is
proportional to the
r,N z
2.
=r N
1
(544)
is
The distance
Therefore,
the same.
6',!,
3.
62 r2
(545)
=T
62
JV,
(546)
Ti
Ft
r,,,
N,
i
4
Tiy
d7
JV,
*
T, 0,
M^4
JV,
<C2 "2
Fig. 512.
Gear
train.
Fig. 513.
two
gears, co 1
and
co 2 , are
T2
teeth
#2
CO!
r2
In practice, real gears do have inertia and friction between the coupled gear which often cannot be neglected. An equivalent representation of a gear
elements
is
shown
in Fig. 513.
Coulomb friction, and inertia considered as lumped The following variables and parameters are
Sec. 5.3
199
T=
6 1, 6 2
applied torque
Ti,T2 /, J2
,
= =
angular displacements
torque transmitted to gears
= inertia of gears
F,i,
c2
for gear 2
T2 (t) = J2
The torque equation on the
r(0
^ ^
is
written
2
+B
+ Fc2 ^
(548)
side of gear
1 is
= /,
^ + *, *M> + F M + 7\(?)
ei
(549)
By
is
converted to
w
T
(a
2
W ST + n F
ddti)
>
2
if
<>
e2
Nl
(s 50) (5
^y
Equation (550) indicates that it is possible to reflect inertia, friction, (and compliance) torque, speed, and displacement from one side of a gear train to the
other.
when
reflecting
from gear
2 to gear
Inertia:
(i)V2
:
(^1
B2
Torque:
^T
jrrd 2
~co 2
Jy 2
Coulomb
If torsional spring effect
frictional torque:
rrFc2
c 2
. ,
a>2l
(NJN2 )
in reflecting
were present, the spring constant is also multiplied by from gear 2 to gear 1. Now, substituting Eq. (550) into
Eq. (549),
we
get
HO = /,.
where
^0 + B U ^M + T
2 2
x
(551)
= Ji + J (jff Bu = B + B
Ju
(jfy
(552)
(553)
200
Chap. 5
Example
53
inertia
Coulomb
side).
fric
and
5 gear train
is
reflected inertia
on
on the load
.
The
ozinsec 2
The
reflected
Coulomb
friction is (^)2
= 0.4
ozin.
Timing
belts
serve the
same purposes
except that they allow the transfer of energy over a longer distance without using
an excessive number of gears. Figure 514 shows the diagram of a belt or chain drive between two pulleys. Assuming that there is no slippage between the belt and the pulleys, it is easy to see that Eq. (547) still applies to this case. In fact,
the reflection or transmittance of torque, inertia, friction, etc.,
is
similar to that of
a gear train.
*~/i
T2
02
force in the
shown in Fig. 515 transmits translational motion and same way that gear trains transmit rotational motion. The relation between the forces and distances is
The
lever system
A
Backlash and Dead Zone
(555)
x{t)
\
.HO
Backlash and dead zone usually play an important and similar mechanical linkages. In a great majority of situations, backlash may give rise to
In addition,
it
~*~ 2
Output
ical
model of backlash
an input and an output member is shown in Fig. 516. The model can be used for a rotational system as well as
Sec. 5.3
201
is
reference position.
upon the
motion
In general, the dynamics of the mechanical linkage with backlash depend relative inertiatofriction ratio of the output member. If the inertia of
the output
is
member
is
very small compared with that of the input member, the means that the output
is
member
will
When
stand
the output
member
member
will
it is
up on the other
side, at
which time
assumed that the output member instantaneously takes on the velocity of the input member. The transfer characteristic between the input and the output
displacements of a backlash element with negligible output inertia
Figure 517.
is
shown
in
To
members,
let
is
The displacements and velocities of the input and output members are illustrated as shown in Fig. 518. Note that the reference position of the two members is taken to be that of Fig. 516, that is, with the input member
respect to time.
starting at the center of the total backlash.
For
Fig. 518,
it is
when motion
member
is
b/2. At the other extreme, if the friction it may be neglected, the inertia of the output member remains in contact with the input member as long as the acceleration is in the direction to keep the two members together. When the acceleration of the input member becomes zero, the output member does not stop immediately but leaves the input member and coasts at a constant velocity that is equal to the maximum velocity attained by the input member. When the output member has traversed a distance, relative to the input member, equal to the full width of the backlash, it will be restrained by the opposite side of the input memthe right, so that x(0)
and y(0)
is
so small that
202
Chap. 5
and velocity waveforms of input and output members of a backlash element with a sinusoidal input displacement.
Fig. 518. Displacement
without
friction.
ber. At that time the output member will again assume the velocity of the input member. The transfer characteristic between the input and the output displacement of a backlash element with negligible output friction is shown in Fig. 519. The displacement, velocity, and acceleration waveforms of the input and output members, when the input displacement is driven sinusoidally, is shown in
Fig. 520.
between those of
and
520.
Sec. 5.4
203
Displacement
Velocity
Acceleration
Output
Input
and acceleration waveforms of input and output members of a backlash element when the input displacement
Fig. 520. Displacement, velocity,
is
driven sinusoidally.
5.4
The equations of a linear mechanical system are written by first constructing a model of the system containing interconnected linear elements, and then the system equations are written by applying Newton's law of motion to the freebody diagram.
Example
54
Let us consider the mechanical system shown in Fig. 52 1(a). The freebody diagram of the system is shown in Fig. 521 (b). The force
d 2 y(t)
y)
*/(/)
(a)
(b)
Freebody diagram.
204
Chap.
is
written
(556)
This secondorder differential equation can be decomposed into two firstorder state equations, using the method discussed in Chapter 4. Let us assign x x = y and jc 2 =
dy/dt as the state variables.
Then Eq.
**(/)
(556)
is
written
dxj(t)
dt
(557)
T
It is
i^<)M^> + ]>>
mechanical system
is
(558)
not
analogous to a
series
RLC
electric
ly
it is
from the mechanical system using a different set of state variables. If we consider mass is analogous to inductance, and the spring constant K is analogous to the inverse of capacitance, 1/C, it is logical to assign v(t), the velocity, and f (t), the force k acting on the spring, as state variables, since the former is analogous to the current in an inductor and the latter is analogous to the voltage across a capacitor.
that
Then
Force on mass
M Mp =
l
_ Mt)
A(0
i/(0
(559)
Velocity of spring
dfk (t)
dt
K
is
_ v{t)
(560)
first state
equation
is
on the voltage
through a capacitor. This simple example further illustrates the points made in Chapter 4 regarding the fact that the state equations and state variables of a dynamic system are not unique.
like that of the current
Example
As a second example of writing equations for mechanical systems, consider the system shown in Fig. 522(a). Since the spring is deformed when it is subjected to the force /(f), two displacements, Vi and y 2 must be assigned to the end points of the spring. The freebody diagrams of the system
55
,
From
y 2 U)
VA
yi(0
Md
y 2 (t)
2
dt
y 2 U)
yi(t)
CF
nptfs
M
fU)
nnnnK(y
l
dy 2 (t)
y 2
/(
~dT
(a)
(b)
Fig. 522.
55. (a)
Massspringfriction
Sec. 5.4
205
fit)
= K\y
(t)
 y 2 (t)]
(561)
K[yi(t)
(562)
Now
let
us write the state equations of the system. Since the differential equation of
is
the system
way
is
to
decompose
this
(t)
=y
dt
2 (t)
and x 2 (t)
2 (t)
= dy 2 (t)/dt,
dxi(t)
=x
(563)
dt
*(')
+ 3^/(0
body with mass
state variable, so
(564)
As an
alternative,
we can
M as one
we have
(565)
state variable,
,,,
,.
and
/*(')
=/(')
(566)
two equations
in Eqs. (565)
and
(566) are
seems that only Eq. (565) is a state equation, but we do have two state variables in v(t) and f (t). Why do we need only one state equation k here, whereas Eqs. (563) and (564) clearly are two independcorrect as state equations, since
The
situation
is
network analogous
network of the system, shown in Fig. 523. It is clear that although the network has two reactive elements in L and C and thus there should be two state variables, the capacitance in this case is a "redundant" element, since e c (t) is equal to the applied voltage e(t). However, the equations in Eqs. (565) and (566) can provide only the solution to the velocity of once /(f) is specified. If we need to find the displacement y^it) at the point where f(t) is applied, we have to use the relation
ydt)
= &jp + y 2 (t) = + P
v(t)
dx
+ y 2 (0+)
(567)
where
yi(t)
.^(Ol)
we can
is
displacement of the body with mass M. On the other hand, from the two state equations of Eqs. (563) and (564), and then determined from Eq. (561).
is
the
initial
Example
56
In this example the equations for the mechanical system in Fig. 524(a) are to be written. Then we are to draw state diagrams and derive transfer functions for the system.
for the
two masses are shown in Fig. 524(b), with the y and y 2 as indicated. The Newton's force
t
206
Chap. 5
^^^^^
K,
M,
y 2 (f)
K,
*i(>i
M,
Vlit)
no
(a)
/u)
(b)
Fig. 524.
56.
equations for the system are written directly from the freebody diagram:
(568)
(569)
dt*
dt
differential
We may now
equations
x,
= yi
dy\
(570)
Xl
dx
dt
(571)
dt
x 3 =yi xt
(572)
dy 2
dt
dx 3
dt
(573)
Equations (571) and (573) form two state equations naturally; the other two are obtained by substituting Eqs. (570) through (573) into Eqs. (568) and (569), and
rearranging
we have
dxi dt
= x2
(574)
Sec. 5.4
207
dx 2
dt
Xi)
~ Ml {Xz ~
Xi)
Wm
x
(575)
dx 3
dt
Xi
(576)
dt
If
M
Xl
A",
X2
+
2
AT2
"3 *3
x
A/ 2
(,
+ Bi )x i
(577)
we
and
written
= *i(0
* = *s(0
x
The state diagram of the system, according to the equations written above, is drawn as shown in Fig. 525. The transfer functions Y (s)IF(s) and Y2 (s)/F(s) are
By/Mj
524.
obtained from the state diagram by applying Mason's gain formula. The reader should verify the following results (make sure that all the loops and nontouching loops
are taken into account)
m
x
(s)
M
B
t
2 s*
(Bi
+ B )s + A
2
(K
+ K2
(580)
MO F(s)
where
+K
(581)
A =
MM
x
2 s*
+B
system.
1 (B l
(582)
The state equations can also be written directly from the diagram of the mechanical The state variables are assigned as v = dy /dt, v 2 = dy 2 /dt, and the forces on the two springs, /x! and f K2 Then, if we write the forces acting on the masses and the velocities of the springs, as functions of the four state variables and the external force,
t x
.
Force on
x :
dv
x
~di
B lVl
+B v fK1 +f
x
(583)
208
Chap. 5
Force on
M
.
%*
= fi,^i
#,(*>i
(2?i
B2 )v 2 +/ki
fx:
(584)
Velocity on
df Ki "^
v2 )
(585)
Velocity
on
K2
^p = K v
2
(586)
Example
The rotational system shown in Fig. 526(a) consists of a disk mounted on a shaft that is fixed at one end. The moment of inertia of the disk about its axis is /. The edge of the disk is riding on a surface, and the viscous friction coefficient between the two surfaces is B. The inertia of the shaft
57
negligible,
is
but the
stiffness is
K.
<s^^^^
^^^^
(a)
(b)
Example
57.
Assume that a torque is applied to the disk as shown; then the torque or moment equation about the axis of the shaft is written from the freebody diagram of Fig.
526(b):
T(t) = J 4%P +
Notice that
equations
this
B^ +
K0(t)
(587)
system
is
may be
6(t)
and dx
{f)\dt
= x 2 (t).
exercise.
an
5.5
ErrorSensing Devices
in
Control Systems 4
'
In feedback control systems it is often necessary to compare several signals at a certain point of a system. For instance, it is usually the case to compare the
reference input with the controlled variable; the difference between the two
signals
is
The
error signal
is
The
sum of
several signals
is
defined in
potentiometer or combination of potentiometers, a differential gear, a transformer, a differential amplifier, a synchro, or a similar element. The mathematical
is
Sec. 5.5
ErrorSensing Devices
in
Control Systems
209
is
proportional
when
a voltage
is
applied across
its
fixed terminals,
compare two
may be
and
potentiometer.
When
a constant voltage
is
potentiometer, the voltage across the variable and the reference terminals will be proportional to the difference between the two shaft positions. The arrangeing devices
ment shown in Fig. 527(b) is a onepotentiometer realization of the errorsensshown in Fig. 527(a). A more versatile arrangement may be obtained by using two potentiometers connected in parallel as shown in Fig. 527(c). This
(a)
u(0
(b)
o_
e(0
(c)
Blockdiagram and signalflowgraph symbols for an error one potentiometer, (c) Position error sensor using two potentiometers.
Fig. 527. (a)
t
T
o
Load
e(f)
(a)
Gear
train
dc amplifier
e a (t)
Input
shaft
U >
Permanentmagnet dc motor
Fig. 528. (a) Direct current control system with potentiometers as error
waveforms of
210
Sec. 5.5
ErrorSensing Devices
in
can be ac or dc, depending upon the types of transducers that follow the error sensor. If v(t) is a dc voltage, the polarity of the output voltage e(t) determines the relative position of the two shafts. In the case of an ac applied voltage,
e(t) acts
the phase of
case the transfer relation of the two errorsensor configurations can be written
e(t)
= K [O (t) s
6 c (t)]
(588)
where
e(t)
s
= K =
and the
total displacement
is
V volts
and each of the potentiometers is capable of rotating 10 turns, Ks = F/2O71 V/rad. A simple example that illustrates the use of a pair of potentiometers as an error detector is shown in Fig. 528(a). In this case the voltage supplied to the error detector, v(t), is a dc voltage. An unmodulated or dc electric signal, e(t), proportional to the misalignment between the reference input shaft and the
controlled shaft, appears as the output of the potentiometer error detector. In control system terminology, a dc signal usually refers to an unmodulated signal.
The
is
usually
much
is
referred to as the
v(t)=V sin
Analytically, the output of the error sensor
e(t)
co c t
is
(589)
given by
(590)
= K,0JMt)
where 0,(0 is the difference between the input displacement and the load displacement, 0/0 r (?) c (f). For the 0/0 shown in Fig. 529(b), e(t) becomes
a suppressedcarrier modulated'signal.
occurs whenever
212
Chap. 5
6c (t)
Load
Gear
train
Twophase ac motor
(a)
(b)
the signal crosses the zeromagnitude axis. This reversal in phase causes the ac
motor
error 6t).
fact that
when a
signal
modulated by a
Sec. 5.5
ErrorSensing Devices
in
Control Systems
213
no longer contains the original carrier frequency assume that 8t) is also a sinusoid given by
0,(0
.
co c
To
sin o>X?)
(591)
relations, substitut
where normally, co s <C co c By use of familiar trigonometric ing Eqs. (589) and (591) into Eq. (590) we get
e(t)
$K
V[cos(co c
co s )t
cos (co c
+ co
s )t]
(592)
no longer contains the carrier frequency co c or the signal frequency co s but it does have the two side bands co c + co s and co c co s Interestingly enough, when the modulated signal is transmitted through the system, the motor acts as a demodulator, so that the displacement of the load will be of the same form as the dc signal before modulation. This is clearly seen from the waveforms of Fig. 529(b).
Therefore,
e(t)
,
.
should be pointed out that a control system need not contain alldc or components. It is quite common to couple a dc component to an ac component through a modulator or an ac device to a dc device through a demodulaIt
allac
tor.
For instance, the dc amplifier of the system in Fig. 528(a) may be replaced by an ac amplifier that is preceded by a modulator and followed by a demodula
tor.
Synchros.
Among
most widely used is a pair of synchros. Basically, a synchro is a rotary device that operates on the same principle as a transformer and produces a correlation between an angular position and a voltage or a set of voltages. Depending upon the manufacturers, synchros are known by such trade names as Selsyn, Autosyn, Diehlsyn, and Telesyn. There are many types and different applications of synchros, but in this section only the synchro transmitter and the
shaft errors, the
Synchro Transmitter
shown
in Fig. 530.
singlephase ac voltage
applied to the
Stator
Slip
*1
ac voltage
R,
rings
Rotor
Stator
(a)
(b)
Fig. 530.
214
Chap. 5
which
is
The symbol G is often used to designate a synchro sometimes also known as a synchro generator.
=E
sin co c t
(593)
530, which is defined as the induced across the stator winding between S 2 and the
is
neutral n
is
maximum and
written
e s,Sf)
= KE
sin cot
(594)
where
A"
is
a proportional constant.
the terminals 5[
and
S3 n
are
e Sm (t)
e sm(0
= KE = KE = = =
(595) (596)
e S,Si
 es = e Ss = e Sin ~
,
\.5KE
1.5KE,
(597)
sin cot
(598)
Consider
mitter
is
now
tion, as
shown
will
in Fig. 531.
The
winding
displacement 6; that
ESl = KE
Fig. 531.
cos (6
240)
(599)
Rotor position
of a synchro transmitter.
Es = KE ESin = KE
cos 9
cos (0
(5100)
120)
(5101)
The magnitudes of
become
240)
120)
(5102)
(5103)
sin (9
Es
shown
,s<
^fJKE
sin
(5104)
is
each rotor position corresponds to one unique set of stator voltages. This leads to the use of the synchro transmitter to identify angular positions by measuring and identifying the set of voltages at the three
stator terminals.
Sec. 5.5
ErrorSensing Devices
in
Control Systems /
215
Volts
synchro transmitter as
is
two
apparently
A typical arrangement of a synchro error detector involves the use of two synchros: a transmitter and a control transformer, as shown in Fig. 533.
inadequate.
Synchro
transmitter
Control transformer
R,
Rotor
dc )
Fig. 533.
positions, a proportional generated at the rotor terminals of the control transformer. Basically, the principle of operation of a synchro control transformer is
is
cylindrically
is
essential for a control transformer, since its rotor terminals are usually
The change
216
Chap. 5
The symbol
CT
is
now
magnetic construction, the flux patterns produced in the two synchros will be the same if all losses are neglected. For example, if the rotor of the transmitter
is
and
in the control
When
in the position
shown
in Fig.
The
its
shafts of the
two
is
When the
transformer
is
terminal voltage
known
as the
is
at
two null positions of the error detector. If an angle a from either of the null positions,
Control transformer
~~
/
L/X a = or
!
180
/f
Rotor voltage =
Flux pattern
(b)
Control transformer
Control transformer
Flux
pattern
Rotor voltage
is
at
maximum
(d)
(c)
among
Sec. 5.5
ErrorSensing Devices
in
Control Systems /
217
(d), the magnitude of the rotor voltage is can be shown that when the transmitter shaft
in
shift accordingly,
any position other than that shown in Fig. 534(a), the flux patterns will and the rotor voltage of the control transformer will be pro
of the control transformer versus the difference in positions of the rotors of the transmitter and the control transformer is shown in Fig. 535.
Rotor
voltage
Vr
360
a=(dr dc
Fig. 535.
dif
From
device.
Fig. 535
it is
is
a nonlinear
However, for small angular deviations of up to 15 degrees in the vicinity of the two null positions, the rotor voltage of the control transformer is approximately proportional to the difference between the positions of the rotors of the transmitter and the control transformer. Therefore, for small deviations,
the transfer function of the synchro error detector can be approximated by a
constant
Ks
K
where
^e~rri
" (5 105)
E=
error voltage
shaft position of synchro transmitter, degrees shaft position of synchro control transformer, degrees
= = = Ks =
Br 9C 9e
The schematic diagram of a positional control system employing a synchro error detector is shown in Fig. 536(a). The purpose of the control system is to make the controlled shaft follow the angular displacement of the reference input shaft as closely as possible. The rotor of the control transformer is mechanically
mitter
connected to the controlled shaft, and the rotor of the synchro transis connected to the reference input shaft. When the controlled shaft is
angular misalignment
exists,
is zero and the motor does not an error voltage of relative polarity
When an
218
Chap. 5
Twophase
induction
motor
Reference
input
Controlled output
(a)
or
r
+ ^1
ac
amplifier
em
Motor
Gear
load
ec
employing synchro
(a).
appears at the amplifier input, and the output of the amplifier will drive the motor in such a direction as to reduce the error. For small dev iations between the controlled and the reference shafts, the synchro error detector can be represented by the constant Ks given by Eq. (5105). Then the linear operation of the positional control system can be represented by the block diagram of Fig. 536(b). From the characteristic of the error detector shown in Fig. 535, it is clear that Ks has opposite signs at the two null positions. However, in closedloop systems, only one of the two null positions is a true null; the other one corresponds to an
unstable operating point.
system shown in Fig. 536(a), the synchro positions are and the controlled shaft lags behind the reference shaft; a positive error voltage will cause the motor to turn in the proper direction to correct this lag. But if the synchros are operating close to the false null, for the same lag between 9 r and 6 C the error voltage is negative and the motor is driven in
Suppose
that, in the
Sec. 5.6
Tachometers / 219
still
further
motor to rotate
is reached. In reality, the error signal at the rotor terminals of the synchro control transformer may be represented as a function of time. If the ac signal applied to the rotor terminals of the transmitter is denoted by sin a> c t, where co c is known
in the
same
is
given by
c
= Ks O (i) sin a
e(t) is
(5106)
5.6
Tachometers 4
Tachometers are electromechanical devices that convert mechanical energy into electrical energy. The device works essentially as a generator with the output voltage proportional to the magnitude of the angular velocity.
In control systems tachometers are used for the sensing of shaft velocity and
improvement of system performance. For instance, in a control system with the output displacement designated as the state variable *, and the output
for the
velocity as the state variable
Fig. 537.
variable may be excessed to by monitored by a tachometer. In general, tachometers may be classified into two types: ac and dc. The simplified schematic diagrams of these two versions are shown in Fig. 537. For the ac tachometer, a sinusoidal voltage of rated value is applied to the primary winding. o A secondary winding is placed at a 90 angle mechanically with respect to the primary winding. When the rotor of the tachometer is stationary, the output voltage at the secondary winding is zero. When the rotor shaft is rotated, the output voltage of the tachometer is closely
,
x2
the
first state
is
'
Schematic diagram of a
tachometer.
The inputoutput
dependent upon the direction of rotation, relation of an ac tachometer can be represented by a firstorder differential equation
volta
is
e l (t)
=K
^jp
and K,
is
(5107)
where
e,{t) is
defined as
rpm
or volts
per 1000 rpm. The transfer function of an ac tachometer is obtained by taking the Laplace transform of Eq. (5107); thus
g> =
K,s
(5108)
described above.
dc tachometer serves exactly the same purpose as the ac tachometer One advantage of the dc tachometer is that the magnetic field of the device may be set up by permanent magnet, and therefore no separate
220
Chap. 5
excitation voltage
is
and
by dc
a control system
ac.
Similarly, a
is
phasesensitive demodulator
used
5.7
DC Motors
in
Control Systems
Direct current motors are one of the most widely used prime movers in industry.
The advantages of dc motors are that they are available in a great variety of types and sizes and that control is relatively simple. The primary disadvantage
of a dc motor relates to its brushes and commutator. For general purposes, dc motors are classified as seriesexcited, shunt
and separatelyexcited, all of which refer to the way in which the field is excited. However, the characteristics of the first two types of motor are highly nonlinear, so for control systems applications, the separatelyexcited dc motor
excited,
is
The separatelyexcited dc motor is divided into two groups, depending on whether the control action is applied to the field terminals or to the armature terminals of the motor. These are referred to di% field controlled and armature controlled, where usually the rotor of the motor is referred to as the armature
(although there are exceptions).
In recent years, advanced design and manufacturing techniques, have produced dc motors with permanentmagnet fields of high field intensity and rotors of very low inertia motors with very high torquetoinertia ratios, in other words. It is possible to have a 4hp motor with a mechanical time constant as low as 2 milliseconds. The high torquetoinertia ratio of dc motors has opened new applications for motors in computer peripheral equipment such as tape drives, printers, and disk packs, as well as in the machine tool industry. Of course, when a dc motor has a permanentmagnetic field it is necessarily arma
ture controlled.
FieldControlled
DC Motor
in Fig. 538.
The schematic diagram of a fieldcontrolled dc motor is shown The following motor variables and parameters are defined
ea {t)
armature voltage
Ra
0(f)
armature resistance
airgap flux
e b (t)
Kb
Sec. 5.7
DC
Motors
in
/.
= constant
K,
4(f)
i (t)
torque constant
armature current
field
current
e f (t)
field
voltage
Tm (t)
Jm
motor
Bm
OJt)
To
made:
2.
The armature current is held constant, The airgap flux is proportional to the
0(0
/'
/.
is,
= Kf
i {t)
(5109)
3.
flux
The torque developed by the motor and the armature current. Thus
TJt)
where
is
= Km IJ(t)
(5H0)
Km
is
gives
TJt)
If
= Km Kf IJf {t) Km Kf I =K
t
<t
(5111)
we
let
Kt
(5112)
TJf)
f {t)
(5113)
Referring to the motor circuit diagram of Fig. 538, the state variables are
assigned as if (t), cojt), and OJt). The firstorder differential equations relating these state variables and the other variables are written
L/^ = RMt) + e f
(t)
(5114)
222
Chap. 5
Jn
dcoJt)
dt
= _ BmCOm(t) + = co m (t)
TJf)
(5115)
dBJf)
dt
(5116)
in
By proper
equations:
'
dif {t)
'
f
~
1
dt
i/O
d(Q m (t)
dt
Kj
J
Bm
J
1
~.
coJJ)
efy)
(5117)
ddjt)
.
dt
[9Jf)
in Fig. 539.
The transfer
"sl
ULf
oe
Ki/Jm
R L f' f
Fig. 539. State
~B m Um
diagram of a fieldcontrolled dc motor.
is
obtained from
Ef (s)
or
Om(s)
_
Lf Jm s 3
K,
(R,JM
+ Bm Lf )s + R B n s
2
(5118)
Ef(s)
where
r_
K
R a Bm s(l +
T m s)(l
(5H9)
x f 8)
=P
xf
ArmatureControlled
DC Motor
is
shown
in
relates the
Sec. 5.7
DC
Motors
in
Control Systems
223
+o
WvV
Constant
field
current
armature current,
i a (t)
thus
TJt)
where
K,Ut)
is
(f>
(5120)
is
Of
constant also.
The back
emf voltage
is
=K
dOJf)
dt
co m (t)
(5121)
With reference
Ra
r
=
_
La Bm
1
ao
co m (0
ej,t)
(5122)
dBJf)
dt
OJt)
The state diagram of the system is drawn as shown in Fig. 541 The transfer function between the motor displacement and the input voltage is obtained from
.
e m (s)
K,
E (s)
a
La Jm s>
(R a Jm
+ B m L )s +
2
(Kb K
+ R Bm )s
a
(5123)
Although a dc motor is basically an openloop system, as in the case of the fieldcontrolled motor, the state diagram of Fig. 541 shows that the arma
motor has a "builtin" feedback loop caused by the back emf. back emf voltage represents the feedback of a signal that is proportional to the negative of the speed of the motor. As seen from Eq. (5123), the back emf constant, Kb represents an added term to the resistance and factional coefficient. Therefore, the back emf effect is equivalent to an "electrical friction," which tends to improve the stability of the motor.
turecontrolled
Physically, the
,
It should be noted that in the English unit system, K, is in lbft/amp or ozin/amp and the back emf constant Kb is in V/rad/sec. With these units,
and
We
224
Chap. 5
Kb IL a
= e (t)i
b
a (t)
watts
(5124)
e b {t)ia {t)
746
hp
we
get
p
This power
is
K Tm ddjjt) ~ 146K, dt
h
hp
(5125)
P ~ Tm de m (t) 550 dt
Therefore, equating Eqs. (5125) and (5126),
hp
(5126)
we have
b
K =
t
0.737
(5127)
We
curves.
motor
A typical set of torquespeed curves for various applied voltages of a dc is shown in Fig. 542. The rated voltage is denoted by E At no load,
r.
Tm =
0, the
speed
is
voltage.
Then
the back
emf constant
b is
given by
(5128)
r
Ka
where in
this case the rated values are
225
Speed
co m
rad/sec
dc motor.
When
the
motor
is stalled,
is
designated by
T
,
(t).
Let
blockedrotor torque at rated voltage T ~~ rated voltage ~Er
K, = K ijj) = ^E
t
_
~~
(5129)
Also,
{t)
(5130)
is
determined:
(5131)
K<
kR a
5.8
motors
Control phase
more
rugged and have no brushes to maintain. Most ac motors used in control systems are of the twophase induction
r^WS
which generally are rated from a fraction of a watt up to a few hundred watts. The frequency of the motor is normally at 60, 400, or 1000 Hz.
type,
A
motor
6
e l (t)
stator with
trical
shown two
in Fig. 543.
The motor
consists of a
degrees apart.
Reference phase
Fig. 543.
voltage source
reference phase.
is
induction motor.
226
Chap.
is
voltage is usually supplied from a servo and the voltage has a variable amplitude and polarity. The direction of rotation of the motor reverses when the control phase signal changes its sign. Unlike that of a dc motor, the torquespeed curve of a twophase induction motor is quite nonlinear. However, for linear analysis, it is generally considered an acceptable practice to approximate the torquespeed curves of a twophase induction motor by straight lines, such as those shown in Fig. 544. These curves are assumed to be straight lines parallel to the torquespeed curve at a rated control voltage (E2 = E = rated value), and they are equally spaced for equal
t
The controlphase
Speed
cj m
twophase induc
tion motor.
The
as follows. Let
is,
k be the
=
x
E,
_
~~
7\,
(5132)
Let
m be a negative number which represents the slope of the linearized torquespeed curve shown in Fig. 544. Then
blockedrotor torque noload speed
T
fi
is
(5133)
Tm
represented by the
Tm {t) = mmjf) +
where cojf)
is
ke t (t)
voltage.
(5134)
Now,
if
we
Sec. 5.8
TwoPhase
Induction Motor
227
may be
obtained
from
Jm
^P=B
= <o m
n o, m (t)
+ Tm (t)
and recognizing that
(5135)
m (t)
is
the
we have
d9 m {t)
dt
(t)
&>UK. + $. M
The
state
(5137)
is
shown
in Fig. 545.
The
(mB m )/Jm
Fig. 545. State
placement
is
obtained as
e m (s)
k
(Bm
E 2 (s)
or
m (s)
m)s[\
+ JJ(B m 
m)s]
(5138)
E
where
_
s(l
2 (s)
Km + r m s)
(5139)
(5140)
Bm
(5141)
Since is a negative number, the equations above show that the effect of the slope of the torquespeed curve is to add more friction to the motor, thus
stability
228
Chap. 5
is
of a dc motor. However,
if
is
damping
occurs for
m > Bm
it
can be shown
that the
5.9
Step Motors 8
step
motor
is
digital
pulse inputs to analog output shaft motion. In a rotary step motor, the output
rotates in equal increments in response to a train of input properly controlled, the output steps of a step motor are equal in pulses. When number to the number of input pulses. Because modern control systems often
shaft of the
motor
have incremental motion of one type or another, step motors have become an important actuator in recent years. For instance, incremental motion control is found in all types of computer peripheral equipment, such as printers, tape
drives, capstan drives,
variety of
and memoryaccess mechanisms, as well as in a great machine tool and process control systems. Figure 546 illustrates the
mechanism using a
step motor.
application of a step
case the motor
is
motor
in the paperdrive
mechanism of
a printer. In this
coupled directly to the platen so that the paper is driven a increment at a time. Typical resolution of step motors available comcertain mercially ranges from several steps per revolution to as much as 800 steps per
revolution,
come in a variety of types, depending upon the principle of The two most common types of step motors are the variablereluctance type and the permanentmagnet type. The complete mathematical analysis of these motors is highly complex, since the motor characteristics are very nonlinear. Unlike dc and induction motors, linearized models of a step motor are usually unrealistic. In this section we shall describe the principle of operation and a simplified mathematical model of the variablereluctance motor. The variablereluctance step motor has a wound stator and an unexcited rotor. The motor can be of the single or the multiplestack type. In the multiplestack version, the stator and the rotor consist of three or more separate sets of teeth. The separate sets of teeth on the rotor, usually laminated, are mounted on the same shaft. The teeth on all portions of the rotor are perfectly aligned. Figure 547 shows a typical rotorstator model of a motor that has three sepaStep motors
operation.
Sec. 5.9
Step Motors
229
Stator
"C
Rotor "A'
Fig. 547.
The motor
shown
to have 12 teeth
rate sections
on the rotor, or a threephase motor. A variablereluctance step motor must have at least three phases in order to have directional control. The three sets of rotor teeth are magnetically independent, and are assembled to one shaft, which is supported by bearings. Arranged around each rotor section is a stator core with windings. The windings are not shown in Fig. 547. Figure 548 is a schematic diagram of the windings on the stator. The end view of the stator of one phase, and the rotor, of a practical motor is shown in Fig. 549. In this
Phase
Phase
Phase
<
.
'*
<
<
\ ,
,,
Fig. 549.
End view
230
Sec. 5.9
is
shown
at a position
where
its
stator
tooth pitch on the stator and the rotor are the same.
pitch, in the
To make
the
motor
rotate,
the stator sections of the threephase motor are indexed onethird of a tooth
same
shows
this
"Phase
Fig. 550.
Rotor and stator teeth arrangements of a multiplestack threephase variablereluctance step motor. The rotor has 10 teeth.
on one
stator
phase are displaced 12 with respect to C of the stator are shown to be aligned
with the corresponding rotor teeth. The teeth of phase of the stator are displaced clockwise by 12 with respect to the teeth of phase C. The teeth of phase B of the stator are displaced 12 clockwise with respect to those of phase A, or
12 counterclockwise with respect to those of phase C. It
is
minimum
four
of three phases
is
and fivephase motors are also common, and motors with as many as eight phases are available commercially. For an nphase motor, the stator teeth are
displaced by l/ of a tooth pitch from section to section.
forward. Let any one phase of the windings be energized with a dc signal.
rotor section under the excited phase are aligned opposite the teeth on the excited
stator.
This
is
the position of
minimum
in a stable equilibrium.
232
Chap. 5
If
phase
is
is
positioned as shown. It can also be visualized from the same figure that
signal
if the dc switched to phase A, the rotor will rotate by 12, clockwise, and the
A of the stator.
Continu
ing in the same way, the input sequence wise in steps of 12.
CABCAB
motor clockis,
Reversing the input sequence will reverse the direction of rotation. That
the input sequence
tion in steps of 12.
The
phase.
Fig. 551.
steadystate torque curve of each phase is approximately as shown in The 0 line represents the axis of any tooth of the energized stator
The nearest rotor tooth axis will always lie within 18 on either side of The corresponding starting torque exerted when this phase is energized can be seen in Fig. 551. The arrows mark the direction of motion of the rotor.
this line.
Torque
Fig. 551.
step motor.
Let positive angular displacements represent clockwise motion. Suppose C has been excited for a long time. This means that the initial
shown
in Fig. 550. If
and
As soon
as phase
A is energized,
is
and
may be
is
point. This
zero. It
is,
however,
a position of unstable equilibrium since the slightest shift from this position will
to
lie
exactly at the
8
1
point, theoretically
in construction,
and the
resulting
asymmetry
We now look upon the stepping motor from a singlestep point of view and
try to develop the equations that govern
will
its
be made
initially to simplify
Sec. 5.9
Step Motors
233
may be made if any of these assumptions are found to be invalidated. We by writing the equation for the electrical circuit of the stator winding. Let
start
angular displacement
stator phase
is
written
= Ri{t) + =
RKt)
[iL(d)]
(5 ' 142)
or
e(t)
Ri(t)
L(9)
f+
t
JJgUp)
f
(
(5143)
The term, L(9)(di/dt) represents transformer electromotive force, or selfinduced electromotive force, and the term i[dL(6)/d9](d6/dt) represents the back emf due to the rotation of the rotor. We have assumed above that the inductance is a function of 9(t), the angular displacement, only. No dependence
getting the torque developed
on the current has been assumed. This will reflect by the motor. The energy in the air gap can be written
directly
W = \L{9)i\t)
From
elementary electromechanical energyconversion principles,
is
(5144)
we know
given by
(5145)
T=s [W(i,0)]
where
6{t).
is
i(i)
and
Therefore,
T = \i\t)j
This torque obtained as
is
[L{e)]
(5146)
is
T
where Jm
also
is
= jJ^ + Bjf
Bm
the viscous frictional coefficient.
(5147)
Jm and Bm
may include the effects of any load. To complete the torque expression
of Eq. (5146),
we need
to
know
the
form of the inductance L(d). In practice the motor inductance as a function of displacement may be approximated by a cosine wave; that is,
L(6)
Z,,
L2
cos rd
(5148)
234
Chap. 5
r is the
Now
rium position be the situation when phase and torque for phase A are given by
A is
2
energized.
Then
the inductance
= TA =
LA
L,
+L
cos rO
(5150) (5151)
Ki\
sin rd
^K>
eB
s bB
=K>
/
RB
+ sL B
ebB
~o
Fig. 552.
Sec. 5.10
TensionControl System
235
respectively.
= 10, phase B has its assuming that the sequence behind the reference point, ABC represents forward motion. Similarly, the equilibrium position of phase C is 12 ahead of the reference point. Therefore, the inductances and the torques
For the
10teeth rotor considered earlier, r
equilibrium position 12
for phases
B and C
= Lc = TB = Tc =
LB
The
electrical circuits
L,
 120) L, cos (100 + 120) Ki\{t) sin (100  120) Kil(t) sin (100 + 120)
+L +L
2
cos (100
(5152)
(5153)
(5154)
(5155)
its
differential
The
total torque
is
the algebraic
sum of
torques
T = TA + TB + Tc
The
for the portrayal of a step motor. Therefore, realistic studies of a step
(5156)
motor
blockdiagram representation
digital
analog or
computer simulation,
made only through computer simulaof the motor, which may be used for is shown in Fig. 552.
5.10
TensionControl System
exists in a great variety of winding and unwinding industrial processes. Such industries as paper, plastic, and wire all have processes that involve unwinding and rewinding processes. For example, in the paper industry, the paper is first wound on a roll in a form that is nonsaleable, owing to nonuniform width and breaks. This roll is rewound to trim edges, splice breaks, and slit to required widths. Proper tension during this rewinding is mandatory for several reasons slick paper will telescope if not wound tightly enough and the width will vary inversely as the tension. Conversely, during
:
storage, a roll
wound
it
to explode. Similar examples could be cited, but the need for proper tension
control
is
Most rewind systems contain an unwind roll, a windup roll driven by a motor, and some type of dancer and/or pinchroller assemblies between the two. Some systems employ springwithdamper idlers with feedback to motor drives to control tension. Some use tensionmeasuring devices and feedback to a motorgenerator or brake on the unwind reel to hold tension at a constant
value.
investigated.
reel.
As shown in Fig. 553, the system has a dcmotordriven windup The tension of the web is controlled by control of the armature voltage ea {i)
of the motor.
236
Chap. 5
+ o
Pinch
rolls
is
Armature:
eJLQ
=
b
R.Ut)
+ L **p + K co m {t)
a
h
(5157)
where
co m (t)
K = = =
Torque equation:
TJf)
where
^Vnetojt)]
+ BmeCOm (t) +
nrT(f)
(5158)
= effective radius of windup reel Tm {t) = motor torque = K,i (t) n = geartrain ratio T(t) = web tension n JL = equivalent inertia at motor shaft jme = Jm Bme = Bm + n BL = equivalent viscous friction coefficient
r
a
\
at
motor
shaft
= BL =
JL
Since the
ceeds, the
effective inertia
of windup reel
windup
reel
Jme co m
is
Furthermore,
(5159)
Jr=&>
Thus
(5160)
Ti =
where
dJ,
W
i
dr
(5161)
Tt
Sec. 5.11
237
Assume now
that Hooke's law
that the
is
obeyed.
a coefficient of elasticity of
C and
(5162)
%p =
where v s {t)
is
C[v w (t)
 ,(/)]
Assuming
s.
the
web
are driven at constant speed, v,(f) = constant = V Also, vjt) = roojt) = nrajf)
It is
is
(5163)
apparent
now
that because r
the
web
is
very
thin,
stant.
~ 0, we may consider that over a certain time period r and JL are con^dF =  r>(0 "
T aJf) + T
m
e  (t)
(5 " 164)
%^ =
}
jtUt)
*^
*"
j^cojf)
me
"17X0
me
(5165)
me
^
5.11
Cnrmjt)
 CV
(5166)
Whenever there
trol.
is
is
become very important in modern paper manufacturing, steel strip processing lines, flexographic and textile industries, and similar fields. To maintain optimum sheet quality and maximum process line speed, the moving web must be maintained at a constant lateraledge position. In general, there are many different ways of measuring and tracking the edge of a moving web. However, to achieve stable and accurate
Therefore, the problem of edge guiding has
edge guiding, a feedback control system should be used. The schematic diagram of an edgeguide system using the pivotroll method is shown in Fig. 554. The pivot roll is controlled to rotate about the pivot point in guiding the direction of the web. The source of controlling the motion
of the pivot
roll
may be a dc motor
or a linear actuator.
rollers
Figure 555 shows the side view of the edgeguide system. The axes of the 1 and 2 are assumed to be fixed or uncontrolled. S^ and S 2 represent
sensors that are placed at the indicated points to sense the centering of the at the respective points.
web
Let
v(t)
^r(0 Z;(0
= linear velocity of web = initial error of web position in the z direction at roll = error of web position in the z direction at the leading
side of the pivot roll (see Figs. 554
and
555)
Roll 2
an edgeguide system.
238
Sec. 5.1
239
Assuming
that there
is
is
roll,
(t):
^2 = t<0tana
at
If the angle
(5167)
is
small,
it is
seen that
(5. 1 68)
~ z (0  *i(0
m,
z R (t)
Thus Eq.
dzgt)
ox
If the linear velocity
in 1
z.(f)
v(t)
(5 . 169)
of the web
is
constant, v(t)
v,
and Eq.
(5169) is
written
^
at
t
+ *<'> = *<'> Wj m
x
(5 " 170)
relation
^rr\ Z x(j)
where
Tj = mjv. Assuming that
= Tr Ti5 +
1
(5
"
171 )
there
is
no
from
Fig. 555,
zz(t)
w
2
z R (t)
{t)
(5172)
or
zz(t)
(t)
(5173)
ZR {s),
and solving
for
Z
Substitution of
tion between
^=
+ \lZ)r,s
(5
" 174
>
(s) x
and
ZR (s) from Eq. (174) into Eq. (171) Z (s), z i(J ) = Z 2 (s) + (mJm^TtS
2 \ 1
(5175) v
by an angle L from its reference position, the D sin 9L (t). With no slippage of the the error z 3 (t) due to the error z {f) is written
x
z 3 (t)
= Zl T)(t
DOAt)
is
(5176)
where
T=
L {t), sin
L (t)
on both
Z,(j)
e T 'Z t (s)
D6 L (s)
x ,
240
Chap. 5
z 3 {t)
and z 5 (t)
is
Z Z
where t 3 = z 3 and z 4 is
5 (s) 3 (s)
(5178)
r3s
3 /v.
Z 4 (s) = Z {s)
3
+ (mjm )z +rs
3
3s
(5179)
is
the drive
an armaturecontrolled
is
Us)
where Ia{s)
resistance,
Ejjs)
 sK
OJs)
(5180)
is the armature current, Kh the back emf constant, R a the armature and 6m (s) the motor displacement. The torque is given by
Tm(s)
where
= KJJ^s)
is
(5181)
is
TJs)
where
(Jm s 2
+ Bm s)9 m (s) +
LF(s)
(5182)
Jm
= =
inertia of
motor
Bm = L
F(s)
and
F(s)
= 2^(^ + B^ s + klWl(s)
2
(5183)
where
= inertia of pivot roll about pivot point = viscous frictional coefficient at pivot point KL = spring constant at pivot roll due to tension
JL
BL
of
web
Combining Eqs.
(5182)
and
(5183),
we have
1
BJs)
TJs)
where
'
/me s 2
me
"
+ B me s + Kme
\2nr)
(5184)
(5185)
B,me
B *~m
(4)^
Kl
(5186)
K = (^)
Also,
(5187)
X(s)
r9 L {s)
(5188)
(5189)
9 L (s)
= *(,)
556 using
T3 J*
5
M
241
242
Chap. 5
The blocks
the determination
Hp (s) and H
minimized.
5.12
Thus
(t)
and z 3 (t)
is
that of a pure
(s)
and
3 (s)
may be encountered
in various types
of systems, especially systems with hydraulic, pneumatic, or mechanical transmissions. In these systems the output will not begin to respond to an input until
after a given time interval. Figure 557 illustrates
Metering point
s
(
S~^
[*
d
(a)
^>
Valve
Solution
/I
Roller ( o 3)
^\
Solution
II
\
zs
~*
Steel plate
Roller ( o
5)
d
(b)
an arrangement
in
which two
be mixed in appropriate proportions. To assure that a homogeneous solution is measured, the monitoring point is located some distance from the mixing point. A transportation lag therefore exists between the mixing point and the place where the change in concentration is detected. If the rate of flow of the mixed solution is v inches per second and dis the distance
between the mixing and the metering points, the time lag
is
given by
(5190)
T= d

sec
at the
If it is
is
mixing point
is c(t)
and that
it
Sec. 5.13
SunSeeker System
243
measured quantity
is
6(0
c(t
T)
is
(5191)
last
equation
B{s)
e~ *C(s)
(5192)
Thus the
transfer function
between
b(t)
and
e
c(i) is
m
The arrangement shown
between the thickness at the
(5193).
~ Ts
(5  193)
in Fig. 557(b)
may be
thought of as a thickness
is
As
given by Eq.
Other examples of transportation lags are found in human beings as control systems where action and reaction are always accompanied by pure time delays. The operation of the sampleandhold device of a sampleddata system closely resembles a pure time delay; it is sometimes approximated by a simple timelag
term, e~ Ts
.
In terms of state variables, a system with pure time delay can no longer be described by the matrixstate equation
^p =
Ax(?)
Bu(0
(5194)
A general state description of a system containing time lags is given by the following matrix differentialdifference equation
^
UL
states.
= t=l A,Ht t
77)
+ j=l Bju(t
Tj)
(5195)
where T, and Tj are fixed time delays. In this case Eq. (5195) represents a general situation where time delays may exist on both the inputs as well as the
5.13
SunSeeker System
In this section
we
shall
is
to
sun with high accuracy. In the system described, tracking of the sun in one plane is accomplished. schematic diagram of the system is shown in Fig. 558. The principal elements of the error discriminator are two small rectangular silicon photovoltaic cells
will track the
mounted behind a rectangular slit in an enclosure. The cells are mounted in such a way that when the sensor is pointed at the sun, the beam of light from the slit overlaps both cells. The silicon cells are used as current sources and connected in opposite polarity to the input of an operational amplifier. Any difference in the shortcircuit current of the two cells is sensed and amplified by the
operational amplifier. Since the current of each cell
is
proportional to the
illu
mination on the
when
the light
an error signal will be present at the output of the amplifier from the slit is not precisely centered on the cells. This error
cell,
244
Chap. 5
Error discriminator
dc tachometer
Fig. 558.
voltage,
when
fed to the servoamplifier, will cause the servo to drive the system
is
given as follows.
The
and all rotations are measured with respect to this axis. The solar axis or the line from the output gear to the sun makes an angle d r with respect to the reference axis, and 9 denotes the vehicle axis with respect to the reference axis. The objeca, near tive of the control system is to maintain the error between 6 r and O
,
zero:
6r
(5196)
When
or Ia
the vehicle
is
0.
and Ia
Ib
I,
Ib
0.
From
shown
in Fig. 558,
we have
oa
w = ^ + L tan a ^
Ltana
(5197)
(5198)
Sec. 5.13
SunSeeker System
245
Fig. 559.
where oa denotes the width of the sun's ray that shines on cell A, and ob is the same on cell B, for a given a. Since the current Ia is proportional to oa, and Ib to ob, we have
/
+ ^tan a
2LI
IT
(5199)
and
I
= It
t jp tan
(5200)
for
For W/2L
pletely
J
A, and
= 21,
Ib
= 0.
< tan a < (C  W/2)/L, the sun's ray comFor (C  W/2)/L < tan a < (C + W/2)/L, = I = for tan a > (C + W/2)/L. zero.
is
may be
of Fig. 560, where for small angle a, tan a has been approximated by a
abscissa.
on the
Fig. 560.
is
tan a but
Nonlinear characteristic of the error discriminator. The abscissa is approximated by a for small values of a.
246
Chap. 5
Operational Amplifier
The schematic diagram of the operational amplifier is shown in Fig. 561. at point G and assuming that the amplifier does not draw
/.
h
7f
e + SLjrI* = j
(5201)
VWV
WWR
JvVW
1
1Fig. 561. Operational amplifier.
Since e
= Aee
eg
= e
/ /
'
becomes
+ (i + i +
e
i>
I)
=
A may
" (5 202)
If
approaches
is
infinity, as in
operational amplifiers,
reach 10 6
then
Eq. (5202)
written
= R F(I a
(5203)
Ib
and e
The gain of
the servoamplifier
is
is
a.
With
expressed as
= K e
a
.
(5204)
s
Tachometer
the
The output voltage of the tachometer eT is related to the angular motor through the tachometer constant KT that is,
;
velocity of
eT
= KT co m
is
(5205)
related to the
9,
(5206)
00
247
248
Chap. 5
ArmatureControlled DC Motor
earlier.
The equations
(5207) (5208)
(5209)
Tm =
where / and
shaft.
J^ +
Bco m
(5210)
B are
the inertia
in the
The inductance
in Fig. 562.
and viscous friction coefficient seen at the motor motor armature is neglected in Eq. (5207).
A block diagram that characterizes all the functional relations of the system
is
shown
REFERENCES
StateVariable Analysis of Electric Networks
1.
B. C.
Kuo, Linear
Circuits
New
York, 1967.
2.
R. Rohrer, Circuit Analysis: An Introduction to the State Variable Approach, McGrawHill Book Company, New York, 1970.
Mechanical Systems
3.
New
New
5.
E. Gibson and F. B. Tuteur, Control System Components, McGrawHill Book Company, New York, 1958.
W. A. Stein and G. J. Thaler, "Effect of Nonlinearity in a 2Phase Servomotor," AIEE Trans., Vol. 73, Part II, pp. 518521, 1954.
B. C.
7.
Vol. 7,
Kuo, "Studying the TwoPhase Servomotor," Instrument No. 4, pp. 6465, Apr. 1960.
Soc. Amer.
J.,
Step Motors
8.
B. C.
1972.
Kuo (ed.),
Proceedings,
I,
9.
B. C.
Kuo
(ed.),
1973.
Chap. 5
Problems / 249
10.
B. C.
St.
PROBLEMS
5.1.
Write the force or torque equations for the mechanical systems shown in Fig. from the force or torque equations.
K
Af,
B,
777777m777m777m7777mm7m7777MM77M7M777,
(a)
M,
B,
F(t)
* Fit)
WW///////////////////
(b)
(c)
Figure P51.
5.2.
Write a
set
On
the
first try,
of state equations for the mechanical system shown in Fig. P52. one will probably end up with four state equations with the
2,
0} 2 , 0i,
minimum
order of three.
no,
h
6l
co
h
t
62
cj 2
Figure P52.
250
Chap. 5
(a)
Write the state equations in vectormatrix form with the state variables defined as above.
Redefine the state variables so that there are only three state equations.
(b)
(c)
Draw
state
(d)
(e)
Determine the controllability of the system. Does the fact that the system can be modeled by three state equations mean that the fourstate model is
uncontrollable? Explain.
in Fig. P53,
5.3.
(s)/Tm (s).
The potentiometer
potentiometer terminals
E volts.
I
Tm it)
Potentiometer
i_
is
B2 =
Figure P53.
5.4.
Write the torque equations of the geartrain system shown in Fig. P54. The
moments of inertia of the gears and shafts are Jit J2 and J 3 Tit) torque. N denotes the number of gear teeth. Assume rigid shafts.
,
.
the applied
Chap. 5
Problems
251
5.5.
Figure P55.
(a)
(b)
among
the inputs 0i a
(c)
(o s
and
(O x
and (O a
5.6.
The block diagram of the automatic braking control of a highspeed shown in Fig. P56a, where
train
is
^
>
e
.
Amplifier
eb
Brake
v(t)
Train
Tachometer
(a)
(Sec)
(b)
Figure P56.
252
Chap. 5
= voltage representing desired speed = velocity of train K = amplifier gain = 100 M = mass of train = 5 x 10* lb/ft/sec K, = tachometer constant = 0.15 volt/ft/sec e, = K,v
V,
shown
in Fig.
P56b when e b
volt.
a block diagram of the system and include the transfer function of each block. (b) Determine the closedloop transfer function between V, and velocity v of
the train.
(c)
Draw
of the train
is
to
be maintained at 20
ft/sec,
Figure P57.
ea
= armature current
ia
s,
is
constant.
is,
The
of the
web
material
is
is
assumed to
satisfy
the dis
and the
proportional constant
is
K (force/displacement).
Chap. 5
Problems
253
(a)
i,
co m ,
and T as
(b)
Assuming
that r
is
constant,
draw a
state
V
5.8.
as inputs.
Write state equations and output equation for the edgeguide control system whose block diagram is shown in Fig. 556.
5.9.
steel rolling
process
is
shown
in Fig. P59.
Twophase
induction motor
Om
U),Tm {t)
Steel
plate
bit)
ait)
M0
Figure P59.
Ks dt)
Describe the system by a set of differentialdifference equation of the form of Eq. (5195). (b) Derive the transfer function between c(t) and r{t).
(a)
5.10.
Figure P510a shows an industrial process in which a dc motor drives a capstan and tape assembly. The objective is to drive the tape at a certain constant speed.
Another tape driven by a separate source is made to be in contact with the primary tape by the action of a pinch roll over certain periods of time. When the two tapes are in contact, we may consider that a constant frictional torque of
TF is
are defined
e, = applied motor voltage, volts 4 = armature current, amperes = back emf voltage = K com volts K = back emf constant = 0.052 volt/rad/sec Km = torque constant = 10 ozin./amp Tm = torque, ozin.
et,
254
Chap. 5
'm
'm
"m
^_
00
<
:
^ J
>
e,
um
G
l
(s)
G 2 (s)
Integral
control
volt/rad/s
Feedback transducer
(b)
Figure P510.
(O m
= motor displacement, rad = motor speed, rad/sec Ra = motor resistance = Q Jm = motor inertia = 0.1 ozin./rad/sec 2 (includes capstan inertia) Bm = motor viscous friction = 0.1 ozin./rad/sec KL = spring constant of tape = 100 ozin./rad (converted to rotational) JL = load inertia = 0.1 ozin./rad/sec 2
8m
1
(a)
(b)
(c)
a state diagram for the system. Derive the transfer function for C0 L (s) with E (s) and TF (s) as inputs. (d) If a constant voltage e,(t) = 10 V is applied to the motor, find the steadystate speed of the motor in rpm when the pinch roll is not activated. What
t
Draw
is
(e)
When
speed
activated,
in contact, the
TF
is 1
ozin.
in the steadystate
10 V.
is
(f)
suggested that a
shown by
Chap. 5
Problems
255
motor speed
is
compared with
The
and the
(g)
a state diagram for the closedloop system. Determine the steadystate speed of the load when the input is 1 V. consider that the pinch roll is not activated, and then is activated.
tional torque.
Draw
First
5.11.
When traveling
through the atmosphere, a missile encounters aerodynamic forces that usually tend to cause instability in the attitude of the missile. The basic concern from the
flight
control standpoint
is
air,
missile
about
its
C is traveling, as shown in
Figure P511.
with the angle 6 (9 is also called the angle of attack), a side force is produced by the resistance of the air through which the missile is traveling. The total force
Fa may
be considered to be centered at the center of pressure P. As shown in Fig. P511, this side force has a tendency to cause the missile to tumble, especially if the point is in front of the center of gravity C. Let the angular accelera
Normally, a F
is
Ct F .
= ad
where a
is
a constant described by
Kr
is a constant that depends on such parameters as dynamic pressure, velocity of the missile, air density, and so on, and
= d =
J
t
missile
moment
of inertia about
distance between
C and P
is
The main
control
means
is
to
256
Chap. 5
tail
parameters.
(b)
the system
Assume
S.
that
T is
(c)
Repeat
(a)
and
(b)
C and P interchanged.
and
(5166).
5.12.
(a)
Draw
a state diagram for the tensioncontrol system of Fig. 553, using the
(b)
T(s), with
EJs) and
as inputs
and
x(t)
i,(t)
= =
v{t)
k(v)
 g(x) +
T(t)
where
x(t)
v(t)
k{v)
g{x)
T{t)
= linear displacement of train linear velocity of train = train resistance force [odd function off, with the properties and dk(v)\dv > 0] k(0) = = force due to gravity for a nonlevel track or due to curvature of track = tractive force
motor
that provides the traction force
is
The
electric
relations
e(t)
T(t)
Ria {t)
where
R =
<j>(t)
armature resistance
i
= magnetic flux = Kf f (i) e(t) = applied voltage Km Kb = proportional constants a {t) = armature current i/(f) = field current
,
(a)
is
= Bv(t), and R =
x(t)
0.
a dc series motor so that a {t) = i(t); g(x) = 0, The voltage eU) is the input. Show that the system
i
v(t)
(b)
Consider that
system. g{x)
ia (t)
f (t)
is
the input
0,
k(v)
= Bv(t).
0, k(v) =
Bv{t),
(c)
Consider that
<f>(t) is
state
Figure P514 shows a gearcoupled mechanical system. (a) Find the optimum gear ratio n such that the load acceleration,
(b)
<X L ,
is
(a)
when
TL
is
zero.
Chap. 5
Problems
257
=
1
N,
\
Jl
V
1
/
4
V
N, EE
Figure P514.
5.15.
(a)
in Fig.
j'k8
where 8 is a 3 x 1 vector that contains all the displacement variables, 9 U J is the inertia matrix and K contains all the spring constants. 2 and 8 3 Determine J and K.
,
.
7
(b)
7
%
Figure P515.
Show
of the form
i
where
= Ax
A=roL!.i _j'k o_
;
(c)
set
K2 =
5.16.
3000, Ji
1,
J2
5,
J3
2,
and
K =
3
Figure P516 shows the layout of the control of the unwind process of a cable reel with the object of maintaining constant linear cable velocity. Control is
established by measuring the cable velocity,
signal,
comparing
signal.
it
with a reference
A tachometer is used to
sense the cable velocity. To maintain a constant linear cable velocity, the angular reel velocity 9 R must increase as the cable unwinds; that is, as the effective radius of the reel decreases. Let
D=
jR
cable diameter
= 0.1
2
f
ft
W = width of reel
=
(empty
reel)
2 f
258
Chap. 5
v:
Motor
'OrIM^
Tachometer
+
em
+
Amplifier
e
+
~
/
Figure P516.
Rf = R=
JR
vR
of reel of reel
(full reel)
= moment of inertia of reel = 18 J? 4 = linear speed of cable, ft/sec e, = output voltage of tachometer, volts Tm (t) = motor torque, ftlb emit) = motor input voltage, volts K = amplifier gain
Motor
inertia
200
ftlbsec 2
and
The tachometer
1
transfer function
is
VR(s)
and the motor
transfer function
is
+ 0.5s
50
s
Tm (s)
Em {s)
(a)
.
Write an expression to describe the change of the radius of the reel R as a function of 6 R (b) Between layers of the cable, R and JR are assumed to be constant, and the system is considered linear. Draw a block diagram for the system and indicate all the transfer functions. The input is e r and the output is vR
.
(c)
VR (s)/E (s).
r
6
TimeDomain Analysis
of Control
Systems
6.1
Introduction
is used as an independent variable in most control systems, it is usuof interest to evaluate the time response of the systems. In the analysis problem, a reference input signal is applied to a system, and the performance of the system is evaluated by studying the response in the time domain. For
Since time
ally
is
follow the input signal as closely as possible, and the output as functions of time.
the transient response
it is
The time response of a control system is usually and the steadystate response.
it
two
parts:
may
be written
c(t)
c,(t)
c(t)
(61)
where
The
analysis
been entirely standardized. In circuit sometimes useful to define a steadystate variable as being a constant
with respect to time. In control systems applications, however, when a response has reached its steady state it can still vary with time. In control systems the steadystate response is simply the fixed response when time reaches infinity.
Therefore, a sine wave
ior
is
fixed for
is
a response
is considered as a steadystate response because its behavany time interval, as when time approaches infinity. Similarly, if described by c(t) = t, it may be defined as a steadystate response.
259
260
TimeDomain Analysis
of Control Systems
Chap. 6
Transient response
as time
is
becomes
lim
c,(f)
=
is
(62)
It can also be stated that the steadystate response which remains after the transient has died out.
phenomenon
to
some
extent before a
reached. Since inertia, mass, and inductance cannot be avoided in physical systems, the responses cannot follow sudden changes in the input instantaneously, and transients are usually observed.
steady state
is
The
is
of importance, since
it is
part
of the dynamic behavior of the system; and the deviation between the response
and the input or the desired response, before the steady state is reached, must be closely watched. The steadystate response, when compared with the input, gives an indication of the final accuracy of the system. If the steadystate
response of the output does not agree with the steady state of the input exactly,
the system
is
6.2
many electrical circuits and communication systems, the input excitations to many practical control systems are not known ahead of time. In many cases, the actual inputs of a control system may vary in random fashions with respect
Unlike
to time.
and speed of the an unpredictable manner, so that they cannot target to be tracked may be expressed deterministically by a mathematical expression. This poses a prob
For
vary in
lem for the designer, since it is difficult to design the control system so that it will perform satisfactorily to any input signal. For the purposes of analysis and design, it is necessary to assume some basic types of input functions so that the performance of a system can be evaluated with respect to these test signals. By selecting these basic test signals properly, not only the mathematical treatment
of the problem is systematized, but the responses due to these inputs allow the prediction of the system's performance to other more complex inputs. In a design problem, performance criteria may be specified with respect to these test
signals so that a system
may be designed to meet the criteria. the response of a linear timeinvariant system is analyzed in the frequency domain, a sinusoidal input with variable frequency is used. When the input frequency is swept from zero to beyond the significant range of the system
When
and output are drawn as functions of frequency. It is possible to predict the timedomain behavior of the system from its frequencydomain characteristics. To facilitate the timedomain analysis, the following deterministic test
signals are often used.
Step input function. The step input function represents an instantaneous change in the reference input variable. For example, if the input is the angular
Sec. 6.2
261
position of a mechanical shaft, the step input represents the sudden rotation of
the shaft.
step function
is
K0 =
where
>0 <0
(63)
R is
a constant.
Or
r{t)
Ru
s (t)
(64)
is
where u s (t)
is
The
is
step function
not defined at
0.
The
shown
in Fig. 6l(a).
(a)
(b)
timedomain
test signals
= Ru
s (t).
(b)
Ramp
Rtu s {t).
Rt 2 u s (t).
Ramp
a
ramp
is
con
ramp
function
is
represented by
K0 =
or simply
/(f)
L [0
\Rt
(6
5
(66)
is
shown
form of
ramp input
input function
is
\Rt 2
>
262
TimeDomain Analysis
of Control Systems
Chap. 6
or simply
<0 =
The
These
test signals all
Rt*ut)
is
(68)
shown
in Fig. 6l(c).
have the
common
scribe mathematically,
become progressively
as a test signal since
instantaneous
deal about a system's quickness to respond. Also, since the step function has, in
principle, a
its
is
how
the system
would respond to
than a ramp function. In practice, we seldom find it necessary to use a test signal faster than a parabolic function. This is because, as we shall show later, to track or follow a highorder input, the system is necessarily of high order, which may mean that stability problems will be encountered.
6.3
we
system
may be
response or,
following.
alternative, by a performance index that gives a qualitative measure on the time response as a whole. These criteria will be discussed in the
as an
SteadyState Error
It
was mentioned
measure of system
accuracy
when a
specific type
of input
is
sical system, because of friction and the nature of the particular system, the steady state of the output response seldom agrees exactly with the reference input. Therefore, steadystate errors in control systems are almost unavoidable,
and in a design problem one of the objectives is to keep the error to a minimum or below a certain tolerable value. For instance, in a positional control system, it is desirable to have the final position of the output be in exact correspondence with the reference position. In a velocitycontrol system, the objective is to have
the output velocity be as close as possible to the reference value. If the reference input r{t) and the controlled output c(t) are dimensionally the same, for example, a voltage controlling a voltage, a position controlling a
position,
are at the
same
level
simply
e(t)
= r(t) 
c{t)
(69)
Sec. 6.3
TimeDomain Performance
of Control
Systems
263
However, sometimes
reference input that
is
it
may be
same
it
at the
level or
may
source for the control of the output of a highvoltage power source; for a
it is
more
Under
error signal cannot be defined simply as the difference between the reference
input and the controlled output, and Eq. (69) becomes meaningless. The input
signals
at the
same
level
usually incorporated
62.
The
r(t)
/\
e(0
>
cit)
G(s)
R(s)
as)
bit)
H(s)
Bis)
Fig. 62.
control system
is
defined as
e (0
= =
r{t)
b(t)
(610)
or
&{s)
R(s)
is
B(s)
R(s)
H(s)C(s)
(61 1)
For example,
is
if
a 10v reference
is
a constant and
is
equal to
0.1.
When
exactly 100
v,
the
error signal
f(0= 100.1100 =
As another example
let
(612)
in Fig. 62
c(t)
we need a
K,s.
defined as
e (0
r{t)
b{t)
r{t)
k;
dc(t)
(613)
dt
The error becomes zero when the output velocity dc(i)\dt is equal to r(i)\K The steadystate error of a feedback control system is defined as the error when time reaches infinity; that is,
t
.
steadystate error
lim
e(t)
(614)
264
TimeDomain Analysis
of Control Systems
Chap. 6
With reference
is
^ = r+W)W)
lim
e(t)
(6 ' 15)
By use of the
is
lim s&(s)
axis
(616)
where
s&(s) is to
lie
on the imaginary
and
we have
(617)
lim
*fffr).
,
,
>
which shows that the steadystate error depends on the reference input R(s) and the loop transfer function G(s)H(s). Let us first establish the type of control system by referring to the form of
G(s)H(s). In general, G(s)H(s)
may
be written
i
G(s)H(s) {)
s)
s'(l
+ T s)(l + T s)
a b
(1
+ T s)
n
(6m
{t 1V)
where
K and all the Ts are constants. The type of feedback control system refers = 0. Therefore, the system that is described by the G(s)H(s) o(Eq. (61 8) is of type/, wherey = 0, 1,2, ... The values
.
of m,
and the Ts are not important to the system type and do not affect the value of the steadystate error. For instance, a feedback control system with
n,
G <*M*>
is
= /+^(1+1)
(6 " 19)
of type
1,
since j
1.
Now let us consider the effects of the types of inputs on the error. We shall consider only the step, ramp, and parabolic inputs.
system of Fig. 62
input
is
steadystate
Steadystate error due to a step input. If the reference input to the control
is a step input of magnitude R, the Laplace transform of the Equation (617) becomes
R/s.
i:
"
sR(s)
1
~~
+ G(s)H(s)
R
1
G(s)H(s)
~~
1
R_ lim G(s)H(s)
fi
?m
'
Kp =
where K
is
lim G(s)H(s)
(621) (620)
is
Then Eq.
written
(6 " 22)
"
= TTK,
the input
is
We
when
a step function,
see that for
Kp
must be
infinite. If
G(s)H(s)
is
we
to be infinite,
j must be at least equal to unity; that is, G(s)H(s) must have at least one pure integration. Therefore, we can summarize the steadystate error due to a step
Sec. 6.3
TimeDomain Performance
of Control
Systems
265
system
1
"
e ss
~ + =
1
R
K,
constant
ramp
r(t)
system
of Fig. 62
is
Rtu,(t)
r(t) is
(623)
where
is
(624)
we have
ss
Mm
V"o s
sG(s)H(s)
R
lim sG(s)H(s)
(625)
If
we
define
K =
v
lim sG(s)H(s)
(626)
K.
which is the steadystate error when the input due to a ramp input is shown in Fig. 63.
c(t)
i
is
ramp
function.
typical e s
Reference input
r(l)
e ss
R/Kv
= Rtu s (t)
function,
Equation (627) shows that for e ss to be zero when the input Kv must be infinite. Using Eq. (618) and (626),
is
ramp
K =
v
lim sG(s)H(s)
s0
lim
J s0 S
j=
0,1,2,.
(628)
to be infinite, j
must be
may be
stated
266
TimeDomain Analysis
of Control Systems
Chap. 6
system
system:
e ss e ss e ss
= = =
oo

constant
is
described by
(629)
= \u 2
s {t)
r(t) is
R(s)
= J
is
(630)
The
R
"
lim s 2 G{s)H(s)
(631)
K =
a
lim s 2 G(s)H(s)
(632)
is
e ss
e ss e ss
system:
1
type
system:
= = =
=
oo
oo
e
e ss
= constant
among the
error constants,
As
the relations
the types of the system, and the input types are tabulated in Table 61.
transfer function of Eq. (618)
The
used as a reference.
Table 61
Summary
Due
to Step,
Ramp,
Ramp
Input,
^**
Parabolic
Input,
^ SS
Type of System
J
Input,
Ap
is
Ay
js
Aa
v
&ss
_
\
R
if 1 ~r Jy p
\
_ R
A;,
is
_ R
T^~ Aa
K
1
=
1
"
+K
e ss e ss e 3S
oo oo oo
K
oo oo
= = =
= = =
^
e ss
2
3
K
oo
e ss
e ss
Sec. 6.3
TimeDomain Performance
of Control
Systems
267
should be noted that the position, velocity, and acceleration error constants are significant in the error analysis only when the input signal is a step function, a ramp function, and a parabolic function, respectively.
It It
is
should be noted further that the steadystate error analysis in this section
conducted by applying the finalvalue theorem to the error function, which is defined as the difference between the actual output and the desired output signal. In certain cases the error signal may be defined as the difference between the output and the reference input, whether or not the feedback element is unity. For instance, one may define the error signal for the system of Fig. 62 as
e(t)
lit)
c(t)
(634)
Then
() and
e
G(s)[H(s)
l]
R()
(6 . 35)
lims l
G(s)[H(s)
l]
R(s)
(636)
It
here
relies
should be kept in mind that since the steadystate error analysis discussed on the use of the finalvalue theorem, it is important to first check to
is, of course, that they do not give information on the steadystate error when inputs are other than the
One
is
that
when
the steadystate
We shall
We
start
S ^>
r+Mm
w e (t)r{t
(6 " 37)
or of Eq. (635), as the case may be. Using the principle of the convolution integral as discussed in Section
the error signal e(t)
3.3,
may be
written
(r)
J" _
 t) dx
(638)
where
we (r)
is
W^= +G \s)H(s)
l
<
W9
>
which
is
known
If the first
values of
/,
t)
268
TimeDomain Analysis
of Control Systems
Chap. 6
series; that
is,
T)
r(t)
xr(t)
if(/)
 ^r(t) +
...
(640)
where
first derivative of r(t) with respect to time. considered to be zero for negative time, the limit of the convolution integral of Eq. (638) may be taken from to t. Substituting Eq. (640)
r(t)
represents the
r(t) is
Since
we have
= =
J
lit)
w.(T)[r(0
*H?)
+ ff (0  yj'C) +
xw e (x) dx
J
is
.]
rfr
(641)
w e (r) dx
J
r(t)
r(t)
f ^w.(i) dx
e(t) as
As
approaches
thus
= lim e(t) =
e,(f)
(642)
is
given by
r,(t)\
w e (x)dxf
(t)\
xw e (x)dx
r s (t)\
~^\{x)dx
(643)
and
r,(t)
^(0
^w,(x)dz
r(t).
Let us define
C =
Ci
{x)
dx
=
J
xwx) dx
x 2 w e (x) dx
C2 =
(644)
= (l)TT"W (TVT J
e
Equation (642)
e,(0
is
written
=C
r,(t)
C,r,(0
+ ^(f) +
and the
+ ^r(/) +
,
(645)
which
is
coefficients,
C C C2
];
C are
The
error coefficients
may be
transform,
we have
W. is)
= P w,(t>" rft
J
(646)
Sec. 6.3
TimeDomain Performance
of Control
Systems
269
Taking the
limit
on both
we have
lim
s^o
e (s)
lim
io
>
f w (T)e"" dx
e
(647)
=C
The
derivative of
Ws) of Eq.
f^ =
ds
Tw e (r)e " di Jo
(648)
= de'"
from which we
get
C,
lim
.o
4EMds Therefore,
(649)
The
rest
Cz = C =
3
lim
,^o
^M
ds 2
3
(650)
lim) ds
*^o
(651) v
of the error
series
Example
In this illustrative example the steadystate error of a feedback control system will be evaluated by use of the error series and the error coefficients.
G(s)
= j^j
(653)
is
of type
Thus the
follows
e ss e ss
2
=
1
ramp
input, tu s (t)
t
u s (t):
e ss
= =
oo oo
Notice that
state error
is
when
the input
is
either a
it
ramp or a
infinite in
magnitude, since
apparent
manner
in
if the steadystate response of this system due to a ramp or parabolic input is desired, the differential equation of the system must be solved. We now show that the steadystate response of the system can
actually be determined
series.
^
Chap. 6
270
TimeDomain Analysis
of Control Systems
Using Eq.
(639),
we have for
this
system
(6 " 54)
C = c
lim W.(s)
j^.0
=K+ jrxf
I
(6
55 >
(6 " 56)
(6 " 57)
(1
Although higherorder
coefficients
will
is
become
written
>
"''
less significant
The
error series
= YTK r*W +
the input signal
r,(f)
(1
+K)*
W+
=
(1
+^)
(?)
(6 " 58)
Now
let
When
tives
is
r s (t)
=u
(t),
and
all
deriva
of
are zero.
The error
*,(0
(659)
result given
is
When
and
is
all
a unit ramp function, r s (t) = tu s (t), r s {t) = u s (t), higherorder derivatives of r s (t) are zero. Therefore, the error series
 (0
_1
TTTF t + K'
+ TTZT^I (1 +KY
(')
(6
60)
linearly
with time.
infinite
but
fails
For a parabolic
all
input, rt)
(t
tu s (t), r s (t)
u s {t), and
(6 " 61)
4.
In this case the error increases in magnitude as the second power of /. Consider that the input signal is represented by a polynomial of t and an
exponential term,
r(f)
[o
+a +
t
^+
s (t)
e~jujf)
(662)
where a
r,(f)
[a
+ <M + ^],(0
s
(663) (664)
(665)
fit)
rXt)
= (i + a 2 t)u = a 2 u (t)
K
(1
=
1
+ K n(t) +
+K)
M
K
(i
+^)3 ^(f)
(6
" 66
>
Sec. 6.4
TimeDomain Performance
Example
62
In this example
stant
is
we
shall consider
61
is
a sinusoid,
r(t)
sin (D
(667)
where
(D
rjit)
(668)
t
The
e,(t)
= [c 
2y C0
4T o
sin
V +
[Cio
yfcoS
cos (D
(669)
Because of the sinusoidal input, the error series is now an infinite series. The conis important in arriving at a meaningful answer to the steadyconvergence of the error K to be 100. Then
series
C =
Cl
YTH = 00099
K +Ky = 0000194
5.65
= (1 +K? = 00098
c* = ~ (i
3
" C = (i 6K, = + xy
x 10" 8
first
et) =i ro.0099
L
(6
70 )
=
or
e,(0
0.01029 sin It
0.0196 cos It
(671)
by Eq.
(671).
6.4
The
is
Of course,
when
referred to, since for an unstable system the response does not
out of control.
272
TimeDomain Analysis
of Control Systems
Chap. 6
The
is
usually characterized by
the use of a unit step input. Typical performance criteria thai are used to characterize the transient response to
rise time,
and
settling time.
a unit step input include overshoot, delay time, Figure 64 illustrates a typical unit step response of
Maximum
overshoot
with respect to
Maximum
sient state.
overshoot.
is
defined as the
largest deviation of the output over the step input during the tran
measure of the
shoot
is
The amount of maximum overshoot is also used as a relative stability of the system. The maximum overis,
per cent
maximum
overshoot
= maximum overshoot
final
value
X 100%
(672)
2.
Td is
r
value.
The
rise
time
is
value.
Sec. 6.5
273
reciprocal of the slope of the step response at the instant that the
response
4.
is
its final
value.
Settling time.
The
settling
time
is
of its
final value.
The four
sure
tities
relatively easy to
mea
when
a step response
is
Performance Index
Since the general design objective of a control system is to have a small overshoot, fast rise time, short delay time, short settling time, and low steadyit is advantageous to use a performance index that gives a measure of the overall quality of the response. Let us define the input signal of a system as r(t) and the output as c(t). The difference between the input and the output
state error,
is
r(t) is
referred to as
signal
In trying to minimize the error signal, time integrals of functions of the error may be used as performance indices. For example, the simplest integral
is
dt
J
(673)
is used to designate performance index. It is easy to see that Eq. (673) not a practical performance index, since minimizing it is equivalent to minimizing the area under e{t), and an oscillatory signal would yield a zero area and
where /
is
thus a zero
/.
Some of the
J o
\"\e{f)\dt
[te{t)dt Jo
JO
f e\t)dt
others.
is
The subject of the design of control systems using covered in Chapter 11.
6.5
Although true secondorder control systems are rare in practice, their analysis generally helps to form a basis for the understanding of design and analysis
techniques.
state
Consider that a secondorder feedback control system is represented by the diagram of Fig. 65. The state equations are written
"x (0i
1
"*i(0"
r(t)
(674)
Ji(t)J
\_<ol
2Cco,
where and
co
are constants.
274
TimeDomain Analysis
of Control
Systems
Chap. 6
x 2 (0+)/s
x,(0+)/s
c(0
colx^t)
(675)
Fig. 65, the state transi
state
diagram of
2Cco
*i(0+)"
col
R(s)
(676)
s_\ix 2 (0+)
where
A=
The
the Laplace transform table.
s2
2cos
col
is
(677)
For a
we have
2 1
*l(0
sin (co*s/l
C
1
+
2
y/)
co n
sin
awT
*i(0+)
.* 2 (0+)J
x 2 (t) J
Vi C
co sin g>V
1
sin
(eoV 1
$)
+
where
col
yic
1
e^'sinKVl C 2 'sin
co,,
(678)
<t>\\
r>0
VI C 2
,
w*J\
=e"
fo "'
'
y/
=
=
tan
yic
c
(679)
2
tan
,yic
(680)
Although Eq.
terms of the
initial states
it is
a rather formidablelooking
is
order. However, the analysis of control systems does not rely completely on the
Sec. 6.5
275
and the
characteristic equation.
a great
deal can be learned about the system's behavior by studying the location of the
roots of the characteristic equation.
Fig. 65.
. 81
=
s
col
2
2gv
is
+ col
=
(6 K
) }
The
zero
that
is,
A=
is
2Ccos
col
(682)
For a unit step function input, R(s) I Is, the output response of the system determined by taking the inverse Laplace transform of
C(s)
=
s{s
+
2
*
f' 2cos
+ col)
2
t
(683)
Or,
c(t) is
initial states
c(0
= +
1
7r=c sSin
co*/l
tan ,v/lC
c
c(t).
>
(684)
It is interesting to
The two
su
s2
= =
C
and
co is
(083)
a.jco
, co n , a,
The
now
described
as follows:
As
multiplied to
rate of rise
in the exponential
a>, and a appears as the constant that is term of Eq. (684). Therefore, a controls the
and decay of the time response. In other words, a controls the "damping" of the system and is called the damping constant or the damping
factor.
The
inverse of a, 1/a,
is
When
call the
the two roots of the characteristic equation are real and identical
critically
we
damping occurs when f = 1. Under this condition the damping factor is simply a = co. Therefore, we can regard as the damping ratio, which is the ratio between the actual damping factor and the damping factor when the damping is critical. co is defined as the natural undamped frequency. As seen from Eq. (685),
system
damped.
From
Eq. (685)
we
when
the damping is zero, 0, the roots of the characteristic equation are imaginary, and Eq. (684) shows that the step response is purely sinusoidal. Therefore, con corresponds to the frequency of the undamped sinusoid.
= aw^ 
(686)
0,
the response of Eq. (684) is not a periodic funcnot a frequency. For the purpose of reference co is
276
TimeDomain Analysis
of Control
Systems
Chap. 6
l'u
Root
) ;,
splane
/
e
i
\
\ N
= >/!r 2
"
/
*a =
$u
Root
Figure 66 illustrates the relationship between the location of the characteristic
C> cu,
and
ca.
shown,
the radial distance from the roots to the origin of the splane.
damping factor a is the real part of the roots; the conditional frequency is the imaginary part of the roots, and the damping ratio is equal to the cosine of
the angle between the radial line to the roots and the negative real axis; that cos 9
is,
(687)
loci,
the constant^
loci,
the constanta
and the constantcu loci. Note that the lefthalf of the jplane corresponds to positive damping (i.e., the damping factor or ratio is positive), and the righthalf of the jplane corresponds to negative damping. The imaginary axis corresponds to zero damping (a = 0, = 0). As shown by Eq. (684), when the damping is positive, the step response will settle to its constant final value because of the negative exponent of e~ Ca*. Negative damping will correspond to a response that grows without bound, and zero damping gives rise to a sustained sinusoidal oscillation. These last two cases are defined as unstable for
linear systems. Therefore,
teristic
equation roots plays a great part in the dynamic behavior of the transient
Sec. 6.5
277
splane
Positive
damping
\ ^ \\ XyA J
'
/to
splane
f = o
>N
Negative
damping
A
/
Positive
/] 1
J
r = o
Negative
damping
damping
'
1
(b)
f2>fl
/CO
/OJ
,
splane
splane
co 2
Positive
Negative
damping
damping
00,
2
t
Oij
3
to,
Positive
Negative
damping
damping
co 2
a2
>a >0
l
<*
(c)
(d)
damping
undamped frequency loci, (b) Constant damping factor loci, (d) Constant condi
is further illustrated by Figs. 68 and 69. In Fig. 68, co is held oo to +oo. The following constant while the damping ratio C is varied from classification of the system dynamics with respect to the value of is given
order system
<C<
s u s2
Ccon
j(O *J\~
n
underdamped case
critically
damped
case
case
C>i C = o
overdamped case
Si, s 2
s
t ,
C<o
s2
undamped
C
2
negatively
damped
case
/oo
xplane
r=o
_^
<?<!/'
r>i
5l
0>f>l
?=
*t
f<I
f
yo>r>i
r
=o
Fig. 68.
damping
a> is
/"
c(,t)
splane
X
*
f>l
/OJ
xplane
f=l
xplane
/
c(/)
o<r<i
*f
X
Fig. 69.
in the xplane.
278
Sec. 6.5
279
/co
c(t)
splane
>~ct
f =
/CO
splane
0>f>l
/CO
splane
r<i
locations.
Figure 69 illustrates typical step responses that correspond to the various root
In practical applications only stable systems are of interest. Therefore when f is positive are of particular interest. In Fig. 610 is plotted the variation of the unit step response described by Eq. (684) as a function of the normalized time cot, for various values of the damping ratio f. It is seen that the response becomes more oscillatory as decreases in value When r i there is no overshoot in the step response; that is, the output never exceedsThe value of the reference input.
the cases
>
The exact relation between the damping ratio and the amount of overshoot can be obtained by taking the derivative of Eq. (684) and
zero.
Thus
dc(t)
tcoe
Zcon t
dt
CcOnt
sin (cot
t4)
(688)
CQa
~ /i
_^
^/1
cos (a>t
4>)
t>0
280
TimeDomain Analysis
of Control
Systems
Chap. 6
Fig. 610. Transient response of a secondorder system to a unit step function input.
where
(j>
tan
,
yic
(689)
Equation (688)
dc(t)
is
simplified to
co
dt
Jl 
e "
&,<
si n
q^ ^i
t
_
=
2,
>
(690)
we have
n
oo
and
(691)
cVlC 2 ' =
0,1,2,...
Sec. 6.5
281
or
nn
The first maximum value of the step response c(t) occurs at n the time at which the maximum overshoot occurs is given by
max
/,
..,
1.
Therefore,
(693)
In general, for all odd values of n, that is, n 1 3, 5, Eq. (692) gives the times at which the overshoots occur. For all even values of n, Eq. (692) gives the times at which the undershoots occur, as shown in Fig. 611. It is interesting to note that, although the maxima and the minima of the response occur at periodic intervals, the response is a damped sinusoid and is not a periodic function.
,
.
c(t)
v7~p vT^p"
vT^T1
vT^F
at
The magnitudes of the overshoots and the undershoots can be obtained by Thus
Im.x or
<\0
or
mm
r Vi C 2
,
== sin (
nn
tan"
^ZI^)
[
il
"=1,2,3,...
(694)
C(OIorml.
= +
1
(I)""'* "f=F
is
(695)
obtained by letting n
overshoot
maximum
= = c max =
e 'C//T=c
5
(696) (697)
maximum
overshoot
lOOer*^ 1
282
TimeDomain Analysis
of Control
Systems
Chap. 6
0.4
0.6
0.8
ratio f
1.0
1.2
Damping
damping
Note that
response
is
maximum
per cent
is
maximum
shown
in Fig. 612.
From
consideration, the
maximum
occurs are
all
For the delay time, rise time, and settling time, however, the relationships are not so simple. It would be difficult to determine the exact expressions for these quantities. For instance, for the delay time, we would have to set c(t) = 0.5 in Eq. (684) and solve for /. An easier way would be to plot cotd versus as shown in Fig. 613. Then, over the range of < < 1 .0 it is possible to approximate the curve by a straight line,
exactly expressed in terms of
and
co.
^~l+0.7
Thus the delay time
is
(698)
U
For a wider range of
,
+ 07C
1
(699)
co
id
_1 = + 0.6 + 0.15
CO
is
(6100)
For the
directly
rise
time
tr ,
which
10 per cent to 90 per cent of its final value, the exact values can again be obtained
Fig. 610.
The
plot of
cot r
versus
is
shown
in
by a
0.8
+
COn
2.5
o<<
(6101)
better approximation
may be
Sec. 6.5
283
^
H'rf
= l+0.7f
<%) _
R (s)
0.5
<
s2+2$us +
2 wn
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
f
Fig. 613.
system.
5.0
"i
4.0
2.0
I
L
0.8
J_
1.0
1.2
0.2
0.4
0.6
?
Fig. 614.
Normalized
rise
then
,_
~ +
l
lK+lAC 2
co
(6102)
clear that the expression for the
From
it is
<f<
shown
in Fig. 615.
284
Chap. 6
Fig. 615.
<
<
1).
From
the figure
it is
same
CcOnl.
result is
used. Therefore,
1.05
(6103)
Solving for
cot s
from the
last equation,
1
we have
(6104)
<oj,= 4ln[0.05yiC]
For small values of f, Eq. (6104)
CO n t s
is
simplified to
= y  c<
o<c<i
rise time,
(6105)
or
s
(6106)
Now
time,
it is
and
settling
seen that small values of f would yield short rise time and short delay time. However, a fast settling time requires a large value for . Therefore, a compromise in the value of should be made when all these criteria are to be
satisfactorily
met
consideration on
maximum
overshoot, a generally accepted range of damping ratio for satisfactory allaround performance is between 0.5 and 0.8.
6.6
Time Response of
a Positional Control
System
we shall study the timedomain performance of a control system whose objective is to control the position of a load that has viscous friction and inertia. The schematic diagram of the system is shown in Fig. 616.
In this section
Sec. 6.6
285
dc motor
dc
amplifier
;
if
= constant
*
Error detector
t ^Q09
form the error detector with sensitivity The error detector sends a signal to the dc amplifier which is proportionafto *e dtfference between the angular positions of the reference input shaft anS he C UtPUt f he dC ""***" is USed to contro1 the armature of a dcmot ,! dc motor. TH current The the field of the dc motor is held constant The parameters of the system are given as follows:
set
of potentiometers
Gain of dc amplifier
Resistance of armature of motor Inductance of armature of motor
K = 1/57.3 volt/deg = A (variable) * = 5Q L = negligible Jm = 10" lbftsec* Bm = negligible BL = 0.1 lbftsec = 0.11bftsec = NJN 1 = TV K = 0.5 lbft/amp
s
volt/rad
/z,
The
first
is
This
cause
1.
Error detector:
9eQ)
e(t)
(6107) (6108)
2.
DC amplifier:
ea {t)
= Ae(t)
(6109)
286
TimeDomain Analysis
of Control
Systems
Chap. 6
3.
Armaturecontrolled dc motor:
LjMl =  RJ {t) +
a
e a (t)
e b (t)
(61 10)
e{t)
= K co m {t)
b
(6111)
where
b is
the back
TJf)
Km
(t)
(6
112 )
jj^ = Bme co m +
where Jme and
TJt)
(6113)
Bme
coefficients seen
by the motor,
3
and
(6114)
(61
1
= Jm + Bme = Bm
Jme
4.
n 2 JL
\
10 3 lbftsec 2
5)
Output:
^
0c(t)
t
.
= =
co m {t)
(6116) (6117)
n8Jt)
,
The value of the back emf constant, Kb is not given originally, but a definite relationship exists between Kb and K In the British unit system, K, is given in lbft/amp and the units of the back emf constant are volts/rad/sec. With these units, Kb is related to K, through a constant ratio. The mechanical power
developed in the motor armature
Pit)
is
e b {t)i a {t)
watts
(6118)
=
Substituting Eqs. (6111)
^e
b (t)it)
hp
and
we have
(6119)
P(f)
= ^TJfyoJt) = ^Tjfajt)
hp
Also,
it is
known
that
Pit)
hp
(6120)
A,
or
= 2a = /40
1.36*;
0.737A.
(6121)
K =
b
(6122)
Thus, given K, to be 0.5 lbft/amp, K is found to be 0.68 volt/rad/sec. Using Eqs. (6107) through (6116), the state equations of the system are
written in the matrix
form
as follows:
Sec. 6.6
Time Response
dia {ty
of a Positional Control
System
287
K
=
J me
nAK,
r us)
[AK.1
dt
dco m (t)
dt
Bme
'me
co m (t)
0,(0
(6123)
dBJf)
L
dt
J
is
[0(O
is
Fig. 617.
9<a (0+)A
9 B m (0
+)/s
"AK.IL.
Fig. 617. State
in Fig. 616.
When
assumed to be
obtained as
0,(0
M
where
K AK n
s
t
*A.4l + V>U +
?.
W) + K K
b
lS
+ K.AK,n
(6 " 124)
negligible
^ me
~n
^ SeC
The
is
denned
as
0.(0
0,(00,(0
is
(6125)
The openloop
G(s)
= JsW =
K,AK n
K ^meS{l +
t.j)(1
r me s)
+KK
b
(6126)
lS
transition equations of the system can be obtained from the state diagram by use of the gain formula, in the usual fashion. However, the main objective of this problem is to demonstrate the behavior
The
state
288
TimeDomain Analysis
of Control
Systems
Chap. 6
the positional control system with respect to the system parameters, and sufficient to assume that all the initial states of the system are zero.
it
is
Since
(6124)
is
La
is
negligible, t
0,
simplified to
OJs)
d,{s)
R.Jm
y + (K K
t
+ R aB m .)s +
K,AK n
l
(&l27) v
of the second order; thus it can be written in the standard form of Eq. (681). The natural undamped frequency of the system is
Equation (6127)
is
AK,n w = V K m i R.J
s
"
(6128) v
The damping
ratio
is
(6129)
When
the values of the system parameters are substituted into Eq. (6127), the
JA G
5A
34.5s
VJs)s 1 +
undamped frequency and
the
5A
(6130) ' l
The natural
damping
= C =
The
s
2
31.6rad/sec
0.546
is
(6131) (6132)
34.5s
1000
(6133)
17.25 j26.5
(6134)
u s (t) rad;
^=
or
fl
'
1
W
1
1000
+ 3^7+
17
(6 " ,35)
1000) J
t ii
rA =
4
9 <,o.5 4 6<r
^ n rn.R17m
ir.
tan"
1.53)
=
normalized time
the gain
1.2e
"'
sin (26.4?
is
(6136)
236.8)
It is interesting to
when
the value of
the gain
A is varied. Note that in Eq. (6128) and in A increases the natural undamped frequency oj> but decreases the damping ratio . For A = 1500, = 0.2, and = 86.2 rad/sec. The output response to a unit step displacement input for A = 1500 is plotted as shown in
Sec. 6.6
Time Response
of a Positional Control
System
289
Bc {t)
Fig. 618.
in Fig. 616
the input
when
Fig. 618. In this case the overshoot is very high, although the rise time and delay time are apparently reduced. The settling time is also shortened by the increase in A, although the responses in Fig. 618 do not show this because the time axis is normalized by .. In fact, in Fig. 618 the step response for A 1500 seems to take a longer time to reach the steady state than that for A 200 This is distorted by the normalized time axis; for A 1500, co 86 2 rad/sec as compared to a> 31.6 rad/sec for A 200.
low value, 13.5, the damping ratio and the are f 2.1 and 8.22 rad/sec, respectively . fcince is greater than one, the step response corresponding to A 13 5 is overdamped. Table 62 gives the comparison of the time responses of the system for the three different values of A used.
natural
When A
is set
at a relatively
undamped frequency
m =
Table 62
Comparison of Transient Response of a SecondOrder Positional Control System When the Gain Varies
Damping
Gain
Maximum
(On
Ratio C
2.1
Overshoot
Tt
0.325
Tr
1.02
Ts
1.51
'max
13.5
200 1500
0.546
0.2
0.043 0.013
0.064
0.015
0.174
0.17
0.119
0.037
0.52
290
TimeDomain Analysis
of Control
Systems
Chap. 6
When
the value of
is set
at 13.5
characteristic
and
listed as follows:
s u s2
A =
13.5
,4=200
as
jj 2
= =
2.1, 32.4
17.25 y26.
shown
These roots, together with the ones for A = 1500, are located in the splane in Fig. 619. In general, for any value of A, the roots of the characterequation are given by
Sl
teristic
,s 2
17.25
1 j*/U9020A
(6137)
A
.splane
= 1500
/86
^
<
2.1 = 135
.4
= 200
;26.5
A<0
,4
= 13.5
A =
A<0
/26.5
Root
loci
A = 1500 A
/86
Fig. 619.
A between
lie
and the system is overdamped. For values of A greater than 59.5, the roots are complex conjugate; the real parts of the roots are equal to 17.25 and are not affected by the value of A. Therefore, as A approaches infinity, the damping factor of the secondorder system is always 1 When A varies continuously between co and o, the equal to 17.25 sec two roots of the characteristic equation trace out two continuous loci in the splane, as shown in Fig. 619. In general, these are called the root loci of
on the negative
real axis in the splane,
.
the characteristic equation or of the control system. In control system studies a rootlocus diagram can be very useful for analysis and design once the relation
and the
is
all
When A
positive,
eC
' '
Time Response
of a Positional Control
System
291
which corresponds to a time response that increases monotonically with time and the system is said to be unstable. The dynamic characteristics of the transient response as determined from the rootlocus diagram of Fig. 619 are summarized
as follows
Amplifier Gain
System Dynamics
<A<
A
59.5
Two Two
Overdamped
Critically
(
>
1)
59.5
damped
(
roots
59.5
= n
oo
Two complex
Two
Underdamped
<
1)
oo
positive
Unstable system
(f
< 0)
Since the system under consideration is of type 1, the steadystate error of is zero for all positive values of A, when the input is a step function in other words, for a step input, the positional error constant K, is to be used Substituting the openloop transfer function G(s) for the system and H(s) 1 into Eq. (621), we have
the system
K v
j>
= lim
"i? 5(s
is
5A
34.5)
(6
'
138 )
= T+X =
( 6 139)
The
Time Response to a
Ramp input
input 0,(0
is
When
tem
a unit
ramp function
described by
0,(0
JB'
UV +
2Cto.fi
is
From
written
U0 = ~ S + aWl'C^
'
'
Sin [(a
^TZ=T
2 '
(6H1)
where
he nSeS f r P re ' f, Fig. 620 Notice that in this case the steadystate error of the system As seen from Eq. (6141), the last term will decay to zero as
nT
. (6 142)
shown
in
infinity.
nonzero time approaches Therefore, the steadystate response of the system due to a unit ramp
292
TimeDomain Analysis
of Control Systems
Chap. 6
4.0
e(A = 200)
3.0
ea (A
= 1500) s
e ss
A
0,(0 = tu s (t)
2.0
= 1500
(A = 13.5)
A A
= 13.5
= 200
Fig. 620.
in Fig. 616
when
input
is
the input
a unit
ramp
function.
lim 6 c {t)
It is
lim
_2f
<w.
(6143)
ramp input
is
34.5
5A
(6144)
which
is
a constant.
error to a
A more systematic method of determining the steadystate input is to use the velocity error constant K From Eq. (626),
v.
ramp
K,
lim sG(s)H(s)
lim
^3 = ^
(6145)
34.5
(6146)
result
of Eq. (6144).
Equation (6146) shows that the steadystate error is inversely proportional if we choose to improve the steadystate accuracy of the system by increasing the forward gain, the transient response becomes
oscillatory.
more
This phenomenon
if
is
rather typical in
is
all
higherorder systems,
may
become
unstable.
Time Response
of a Positional Control
System
293
was shown
is
dc motor
constant To
(6124)
is
is
now
0.02 sec.
The closedloop
now
armature time by Eq y 4
'
&)
9 r (s)
or
*(*)
__
U.005<1
0.05,4
0.02j)(1
+ 2s) +
+
250,4
34s
0.05,4
250,4
js
0M
The openloop
50.5**
1725^
(6
"
148 )
transfer function
is
G (A {S)
The
flcfr)
~W)~
+
50.5s 2
250,4
s(s>
50.5,
is
1725)
(6
"
149 )
\725s
250A
is
(6150)
It is
now
G)
ris)
=
(1
0.48 j)(l
0.0298j
0.000616* 2 )
The
at
characteristic equation has a real root at s 2.08 and two complex roots 24.2 j 32.2. For a unit step function input the output response is
characteristic root at s 2.08. The response due to the last term of Eq. (6152) decreases to zero very rapidly with the increase of t. Comparing Eq. (6152) with Eq. (6136) we see that the damping factor for the secondorder system (L 0) is 17 25 whereas for the thirdorder system (L. 0.1) the output response is governed by the exponential term, which has an equivalent damping factor of 2.08 Thus the thirdorder system will have a slower rise time. This result is expected, since the presence of inductance will slow down the buildup of the armature current thus slowing down the torque development of the motor. However, higher inductance will cause a higher overshoot in the step response, as shown in Fig 621
0.0667e* sin(32.2< 1.88) (6152) In Eq. (61 52), since the time constant of the pure exponential term is more than 10 times greater than that of the damped sinusoid term, the transient response of 8t) is predominantly governed by the
1
fl e
(0
= 
I.06e""
With
La 
0.1
H, and A
factored as
is
(*+50.5)(j*
1725)
(6153)
two imaginary roots at s = ;41 5 The response corresponding to these imaginary roots is an undamped sinusoid. The
characteristic equation has
Thus the
l.O
1
1 1 1 i
1.4
/~\
/L
\
"
=0A
1.2
^=0.01^.^^
1.0
0.8
v/
i
^
i
0.6
0.4
0.2
0.1
0.2
0.3
0.4
0.5
0.6
Time (second)
Fig. 621.
A =
200.
splane
A
A= A=
13.5
= 348
/41.5
<**
A
25.25
= 13.5
2.08
A =0
/41.5 = 348
Fig. 622.
is
Root
294
Sec. 6.7
295
From this illustrative example we have learned that a secondorder system always stable as long as the loop gain is finite and positive; third and higherorder systems may become unstable if the loop gain becomes high.
is
frequency of the sinusoidal response is 41.5 rad/sec. When the roots of a characequation lie on the imaginary axis of the splane, such as in the present situation, the linear system is said to be on the verge of instability. Figure 622 shows the root loci of the characteristic equation of Eq (6150) when A is varied from zero to infinity. For all values of A greater than 348 the two complex roots are found in the righthalf of the *plane, and, with time the step response of the thirdorder system will increase without bound.
teristic
6.7
Time Response of
The
control systems considered thus far in this chapter are all of the proportional type, in that the system develops a correcting effort that is
to the
has the limitation or disadvantage that it is often difficult to find a proper forward path gain so that the steadystate and the transient responses satisfy their respective requirements Often, in practice, a single gain parameter is seldom sufficient to meet the design requirements on two performance criteria.
to perform other operations, in addition to the proportional signal. In terms of signal processing, we may perform a time derivative of the actuating signal. Figure 623 shows the block diagram
control,
It is logical
magnitude of the actuating signal only. The illustrative example given in Section 6.6 shows that a proportional type of control system
proportional
on the actuating
Qs)
Fig. 623.
of a typical secondorder feedback control system with derivative control added to the proportional control. In this case the constant x d represents the amount of derivative control used in proportion to the ordinary proportional control The openloop transfer function of the system is now
E(s)s(s
Analytically, Eq. (6154)
2tco)
<6
"
154 )
i/ T,
296
TimeDomain Analysis
of Control
Systems
Chap. 6
de(t)
dt
(c)
Fig. 624.
Waveforms of c(/), e(t), and de(t)ldt showing the effect of derivaStep response, (b) Error signal, (c) Time rate of change of
error signal.
The
effect
control system can be investigated by referring to Fig. 624. Let us assume that
shown
in Fig. 624(a).
The corresponding error signal e(t) and the time derivative of e{t) are as shown in Fig. 624(b) and (c), respectively. Notice that under the assumed case the step response exhibits a high peak overshoot. For a system that is driven by a motor of some kind, this large overshoot is due to the excessive amount of torque < t < t u during which the error developed by the motor in the time interval signal is positive. For the time interval f < t < t 3 the error signal is negative, and the corresponding motor torque is negative. This negative torque tends to reverse the direction of motion of the output, thus causing c(t) to undershoot during t 3 < t < t 5 During the time interval t 3 < t < t 5 the motor torque is again positive, thus tending to reduce the undershoot in the response caused by the negative torque in the previous interval. Since the system is assumed to be stable, the error amplitude is reduced with each oscillation, and the output
,
eventually
is
we can
Sec 6 7
'
297
The
in the interval
interval
t,
< < is too large, and (2) the retarding torque in the time < < is inadequate. Therefore, in order to reduce the overshoot
/
/,
t t2
approach is, to decrease the amount of positive correcting torque and to increase the retarding torque. Similarly, in the time interval t 2 t r 4 , the negative corrective torque should be reduced, and the
< <
is
now
be increased in
The
cisely the
by the system of
compensation
is proportional r d de(t)/dt. In other words, in addition to the error signal, a signal that is proportional to the time rate of change of error is applied to the motor. As shown in Fig. 624(c), for t r the time derivative of e(t) is negative; this will reduce the original torque developed due to e(f) alone. For
that the proportional type of control system whose signals are described in Fig. 624 is now modified so that the torque developed by the motor
< <
r, r2 both e?(r) and de(t)/dt are negative, which means that the negative retarding torque developed will be greater than that of the proportional case. Therefore, all these effects will result in a smaller overshoot. It is easy to see that e(/) and de(t)Jdt have opposite signs in the time interval r2 t t3 therefore, the negative torque that originally contributes to the undershoot
,
<t <
< <
is
reduced
also!
Since de(t)jdt represents the slope of e{i), the derivative control is essentially an anticipatory type of control. Normally, in a linear system, if the slope of e(t) or c(f) due to a step input is large, a high overshoot will subsequently occur.
The derivative control measures the instantaneous slope of e(t), predicts the large overshoot ahead of time, and makes a proper correcting effort before the overshoot actually occurs.
It is apparent that the derivative control will affect the steadystate error of a system only if the steadystate error varies with time. If the
steadystate error
constant with respect to time, the time derivative of this error is zero, and the derivative control has no effect on the steadystate error. But if the
is
of a system
to de(t)jdt,
steadystate error increases with time, a torque is again developed in proportion which will reduce the magnitude of the error. Consider that the positional control system of Fig. 616 is
modified by
replacing the dc amplifier with derivative control so that the openloop transfer function is now
ee (s) 9e (s)
_ 
saq.
s(s
+ td s) + 34.5)
(6
" 155
Figure 625 illustrates the unit step responses of the closedloop system 13.5 and r d 0.01, 0.1, and 1.0. Notice that in the case of the low value for A, an increasing value of t has the effect of slowing d down the response and the damping is increased. Figure 626 illustrates the unit step responses when A 1500. In this case the damping is noticeably improved by
when A
tive control.
When
When
xd
0, Fig.
A =
14 per cent.
%d
the deriva618 shows that the overshoot is 52 per cent for 0.01, the overshoot is reduced to approximately
,4
= 13.5
1.0
Td =
O.OlX^s.
^=0.1 A = 13.5
0.8
// //
Tj=1.0
0.6
A=
13.5
0.4 ~
II
0.2
1.0
2.0
3.0
4.0
(Seconds)
Fig. 625. Step responses of the positional control system of Fig. 616 with
derivative control;
A =
13.5.
^=0.01
0.02
0.03
0.04
(Seconds)
0.05
0.08
Fig. 626. Step responses of the positional control system of Fig. 616 with
derivative control
A =
500.
298
Sec 6 7

299
response
is
slow in reaching
response for r d
1.0.
The effect of derivative control on the transient response of a feedback control system can also be studied by referring to the openloop transfer function of Eq. (6154). The corresponding closedloop transfer function of the
system
is
*(*)
(2fa
+
is
xi(ol)s
+ oil
(6_156)
The
(2c0
+ tco1;)s + =
2
(6157)
Notice that the derivative control has the effect of increasing the coefficient of the s term by the quantity Td coJ. This means that the damping of the system is increased. Using the values as represented by Eq. (6155), the characteristic equation becomes
s
2
(34.5
5Ard)s
5A
=
=
(6158)
Figure 627 shows the loci of the roots of Eq. (6158) when A 13.5 and x d is varied from to oo. The improvement on the system damping due to the derivative control is illustrated by the root loci of Fig. 628 with A set at 1500. Notice that for small values of r, the roots of the characteristic equation are still complex, although they are farther away from the imaginary axis than those when
Td
0.
For
become
real.
/w
iplane
'
Td
rd =0 *
32.4
rd
fc
=Q **
Td ='
2.1
Fig. 627.
Root loci of the characteristic equation of positional control system with derivative control,* 2 + (34.5 5Ax d)s 5A 0,A = 13.5.
It
by the scheme shown in Fig. 623 generally improves the damping of a feedback control system, no considerations have been given here on the practicability of the configuration. In practice, however, the transfer function of (1 + xd s) cannot be physically realized by passive elements. Furthermore, derivative control has
the characteristics of a highpass
filter which tends to propagate noise and disturbances through the system. Practical controllers which are used to improve
300
TimeDomain Analysis
of Control
Systems
Chap. 6
/86
splane
*d
J<

/86
Td
=0
Root loci of the characteristic equation of positional control system with derivative control; s 2 + (34.5 + 5Azd)s + 5 A = 0, <4 1500.
Fig. 628.
more complex
transfer func
6.8
Time Response of
The counterpart of
derivative control
is
signal in proportion to the time integral of the error. Figure 629 illustrates the
The
signal
instantaneous error signal, the other, to the time integral of the error.
The
parameter
the system
K
is
is
C(s)_
E(s)
(oljs
s (s
2
+ Kj + 2Ca>)
or
s(j
(6159)
R(s)
>l
.
y
*1
HgH
i
,
C(s)
+ 2?<o)
[ Jo
dr
Integral contro
Fig. 629.
integral
com rol.
Sec. 6.8
301
tem by
One obvious effect of the integral control is that it increases the order of 1 More important is that it increases the type of the
.
the sys
system by
There
improved by an order of 1. P U S Constant the inte S ^ cm the system will nowral contrl educes it to zero. In the case of Eq (6159), have a zero steadystate error when the input is a ramp function. However, because the system is now of the third order it tends to be less stable than the original secondorder
'
of the original system without integral control is In other words, if the steadystate error to a given
system. Tn fact,
if
ot the system
is
may
be unstable.
Let us consider the positional control system of Fig. 616 again, with L integral control, the openloop transfer function of the system be
ec {s)
0e(s)
(6
" 16
The
characteristic equation
j3
is
34.5s 2
5As
5AK,
(6161)
of the value of K, on the transient behavior of the system may be investigated by plotting the roots of Eq. (6161) as a function of A. Figure 630
effect
1
The
OO t t
4 ;co
splane
A
splane
A =0 34.5
A =oo

,A
=0
=
A
AT,

A =0
A=0
34.5 +
1
(a)
(b)
(a) Root loci of the characteristic equation of a feedback control system with integral control, s * + 34.5*2 + SAs + 5AKi = 0, K t < 34.5. (b) Root loci of the characteristic equation of a feedback control system with integral control, + 34.5*2 5As =
Fig. 630.
+ 5AK
0,
Ki
34.5.
302
TimeDomain Analysis
of Control
Systems
Chap. 6
splane
A =
.4=0
34.5
\
*i
134.5 +
AT!
T
ooll
(c)
Root
loci
34.5.S 2
5As
5AK\ =
0,
K >
x
34.5.
when A
34.5, the
varies
lies
between
and
When
K =
t
values of A
becomes second order and the two roots lie on the yeaaxis for all between and oo, and the system is on the verge of instability. When
the value of
all
is
unstable for
values of A.
As an alternative, we can fix the value of A and show the effect of varying on the roots of the characteristic equation. Figure 63 1 shows the root loci K < oo. of Eq. (6161) with A = 13.5 and
t
<
To
verify the analytical results obtained in the preceding sections, Fig. 632
x
As predicted by the
is
shows the step responses of the system with A = 13.5 and K = 1, 10, and 34.5. root loci of Figs. 630 and 631, the response for K = 34.5
x
a pure sinusoid.
6.9
The philosophy of using the derivative of the actuating signal to improve the damping of a system can be applied to the output signal to achieve a similar
splane
K,
AT,
=0
K
32.4
Fig. 631. Root loci of the characteristic equation of a feedback control system with integral control, + 34.5*2 + 5 As + 5AK
= 0, A =
13.5.
1_
ATj
= M).
=34.5
1.0
2.0
3.0
4.0
5.0
Time (seconds)
La =
Fig. 632. Step responses of the positional control system of Fig 616 with and integral control, 13.5.
A=
303
304
TimeDomain Analysis
of Control Systems
Chap. 6
In other words, the derivative of the output signal is fed back and compared with the reference input. In practice, if the output variable is mechanical displacement, a tachometer may be used which converts the mechanical displacement into an electrical signal that is proportional to the derivative of the
effect.
displacement. Figure 633 shows the block diagram of a secondorder system with a secondary path that feeds back the derivative of the output. The transfer
is
R(s)
>, <),/\ ?)
w2
s(.s
^9
t
i
*
C(s)
+ 2?oo)
K,s
The closedloop
is
<M
R(s)
s*
col
(2{a>
K,col)s
+ col =
(6162)
and the
characteristic equation
s
1
is
(2co
Kfi>l)s
col
(6163)
(6163) with Eq. (6157), which is the characteristic equation of the secondorder system with derivative control, we see that the two equations are of the same form. In fact, they are identical if % d is interchanged with
Comparing Eq.
K,.
Therefore,
we conclude
should be noted
that the closedloop transfer function of Eq. (6162) does not have a zero, and thus the responses of the two systems will not be identical even if K,
equals x d
The openloop
is
E(s)
s(s
2co B
(6164)
K,col)
The system
is
is still
of type
1,
not altered. However, for a unit ramp function input, the steadystate error to
Sec. 6.10
305
the system of Fig. 623 which has the derivative control of the system in Fig. 633 is (2 K,(a)/co
is 2Clco,
whereas that
n.
6.10
One of the
design techniques in
is
dynamics and feedback from the output variable, flexibility can be gained by feeding back some or all of the state variables to control the
process. In the system with tachometer feedback, shown in Fig. 633, if we decompose the process in the forward path by direct decomposition, we can show that the system is actually equivalent to having state feedback. Figure 634(a) shows the state diagram of the process
C() E(s)
=
s(s
+ 2Cca)
(6165)
which
ically accessible,
x t and x2 are physfeed back these variables through individual gains, as shown in Fig. 634(b), to give closedloop control of the process. The closedloop transfer function of the system in Fig. 6 34(b) is
we may
C(s)
is
decomposed by
col
s*
(2Co>,
+ g )s +T>
t
(6 " 166)
CO
O
c
2f"
(a)
Si
(b)
a secondorder system by
state feedback.
306
TimeDomain Analysis
of Control
Systems
Chap. 6
Comparing
if
two transfer functions would be identical and g 2 = K,col. In fact, in selecting the feedback gains g^ and g 2 in order to have zero steadystate error for a step input, g should equal col. The value of g 2 is selected to satisfy the damping requirements of the system. The significance of this illustration is that, if all the state variables are available for feedback, we can achieve at least the same or better control with state feedback as with dynamic controllers and feeding back only the output. Note that the system with tachometer feedback shown in Fig. 633 has only the output variable available for feedback. If we regard the output as one of the states, it is fed back through a unity gain. The second state needed for feedback is actually "reconstructed" by the dynamics of the tachometer. In other words, the tachometer acts as an "observer" (see Chapter 11), which recovers the state variable from the output variable. In modern control theory, certain types of design algorithm, such as the
back, Eq. (6162),
we
col
Since the
eigenvalues (or the poles of the closedloop transfer function) of a linear system
directly control the transient response of the system,
it
would be
desirable
if
the
designer
is
tions. It is
shown
in
Chapter 11 that
gives
if
a system
is
an
illustration
on how the
feedback and
Example
63
is
= i^ = E(s)
s 2 (s
2 ,
,, 1)
(6167)
Figure 635(a) shows the state diagram of G(s), and Fig. 635(b) shows the state diagram with feedback from all three states. The closedloop transfer function of the sys
tem
is
W)
Let us assume that
step function,
=
*
+ (g +
3
I)* 2
+ Sis + gl
(6 " 168)
we desire to have zero steadystate error with the input being a unit and in addition, two of the closedloop poles must be at s = 1 +j and s = 1 j. The steadystate requirement fixes the value of gi at 20, and only #2 and  3 need to be determined from the eigenvalue location.
The
characteristic equation of the system
s
3
is
(g 3
Ds
+ Sis +
g2
20
=
and
j)(s
/)(>
+ a)
we
(6169)
Equating
get
=22
g3
11
and the
pole
is
third pole
is
at s
10.
damping
ratio of 0.707,
and the
third
of these poles, the system acts like a secondorder system. Figure 636 shows that the unit step response has an overshoot of 4 per cent.
20
oe
O
O
(a)
(b)
by
state feedback.
Example
63.
307
308
TimeDomain Analysis
of Control Systems
Chap. 6
REFERENCES
Time Domain Analysis
1.
O. L. R. Jacobs, "The Damping Ratio of an Optimal Control System," Trans. Automatic Control, Vol. AC10, pp. 473476, Oct. 1965.
IEEE
2.
G. A. Jones, "On the Step Response of a Class of ThirdOrder Linear Systems," IEEE Trans. Automatic Control, Vol. AC12, p. 341, June 1967.
R. A. Monzingo,
trol,
3.
"On Approximating
IEEE
in Multiinput Controllable
Linear Sys
5.
C. Willems and S. K. Mitter, "Controllability, Observability, Pole Allocation, and State Reconstruction," IEEE Trans. Automatic Control, Vol. AC16, pp.
582595, Dec. 1971.
PROBLEMS
6.1.
For each
(b)
(c)
(d)
6.2.
^ 0.707, co !> 2 rad/sec (positive damping) < C < 0.707 CO < 2 rad/sec (positive damping) C ^ 5, 2 < <o < 4 rad/sec (positive damping) 0.5 < < 0.707, co < 2 rad/sec (positive and negative damping)
velocity,
The openloop
G(s)
50
(1
,
(b) G(s)
(c)
G(s)
6.3.
ramp
Problem 6.2, determine the steadystate error input, and an acceleration input t 2 /2.
6.4.
is
riA
500
Evaluate the error series for the system. Determine the steadystate error of the system when the following inputs are applied:
Chap. 6
Problems / 309
(a) r{t)
(b) r(t)
= ,(r)/2 = (1 + It + t*)u,0)
f 2
that the steadystate error obtained from the error series is equal to the inverse Laplace transform of E(s) with the terms generated by the poles of E(s)IR(s) discarded.
6.5.
Show
In Problem
/
 0,
for co
6.4, if a sinusoidal input #(/) = sin cot is applied to the system at determine the steadystate error of the system by use of the error series
5 rad/sec.
What
when
r(t) is
sinusoidal?
6.6.
shown in Fig. P66 with a twoaxis control system. Sketch the reference input of each of the twoaxis systems as a function of time.
= 4.5 sec
7 sec
12
IS
6.7.
machinetool contouring control system is to cut a perfect circular arc with a twoaxis control system. Determine the reference inputs of the two systems that will accomplish this.
6.8.
A step motor gives a single step response shown in Fig. P68 after a pulse
exci
Step position
Step position
0.005 sec
Figure P68.
310
TimeDomain Analysis
of Control Systems
Chap. 6
tation
is
applied.
Find a
model the
motor
6.9.
vectoring.
The attitude control of the missile shown in Fig. P51 1 is accomplished by thrust The transfer function between the thrust angle 5 and the angle of
can be represented by
attack
(refer to
trol
Problem
is
5.1 1)
where
The
attitude con
system
attitude
Or
P~
.
^c^ J
5
.
Gp (s)
Attitude sensor Ks
Figure P69.
(a)
is
in operation
(b)
(c)
Determine the performance of the overall system with respect to the relative values of K, Ks and a. Consider that both loops are in operation. Determine the minimum values oiK, and K, in terms of A' and a so that the missile will not tumble. It is desired that the closedloop system should have a damping ratio of and a natural undamped frequency of w. Find the values of K, and K, in terms of , co, a, and K.
,
= 0).
6.10.
The
control
is
glXi
giX 2
where gi and g 2 are real constants. (a) Find the locus in the gi versus g 2 plane on which the overall system has a natural undamped frequency of ^/2" rad/sec.
(b)
in the gi versus
g 2 plane on which
damping
(c)
= 0.707
and
co
= V/T rad/sec.
6.11.
= Ax + Bw
Chap. 6
Problems / 311
where
1
B
1
2
The input u
satisfies
is
Ax + B =
6.12.
Repeat Problem
6.11
when
1
0"
1
B
shown
2
6.13.
3
is
in Fig. P613
represented by
by
U(s)
L
Js*
^r\
+
e Gc (s)
U
L/Js 2
8o
Desired output
*
Figure P613.
312
TimeDomain Analysis
of Control Systems
Chap. 6
G c (s), and
(s) is
(b)
(c)
With Gc (s) = 1, determine the response of the system, 9 a (t), when the input 6 r {t) is a unit step function. Assume zero initial conditions. Discuss the effects of L and / on this response. Let Gc (s) = (1 + Td s), L = 10, and J = 1000. Determine the value of Td so that the system is critically damped. It is desired to obtain an output response with no overshoot (see Fig. P613); the response time may not be minimum. Let Gc (s) = 1, and the system is controlled through the input 6 r (t), which is chosen to be of the form shown. Determine the values of k and f so that the desired output is obtained. Use the same values of L and / as given before.
i
6.14.
is
shown
in Fig. P614.
R(s)
A V
.
C(s)
.
Amplifiei
gain
Motor I
s
2
as
+o
Figure P614.
(a)
For 10, determine the values of a and b to give an overshoot of 16 per cent and a time constant of 0.1 sec of the system response to a unit step
input.
K=
Time constant
is
is
decreased
how
does
it
affect the
of the system?
(c)
on
6.15.
The parameters of
below:
shown
1
in Fig.
JL
Jm
ftlb/rad/sec 2
BL = Bm = Rf =
Lf R
t
= motor inertia
motor viscous
generator
total
friction
field resistance
50
5
= =
henry
K Kg = generator constant
La =
(a)
0.812 ftlb/A
200 volts/amp
negligible
K=
L (,O
when 0,(0
is
a unit step
Chap. 6
Problems
313
A/VWr
WW*
Amplifier
gain
Rf
'f
,
Ik
~T~ "
eb
'/
= :onstan1
'
+
J
(
e f
Lf
(g\ eg
iflZ>
Motor
to load gear
ratio
i 50
Load
to potentiometer
gearratio =1/1
Figure P615.
(c)
(d)
(a)
(a)
and (b) for K = 60.7. and (b) for K = 50. How does the
for the servo system
steadystate response
6.16.
shown
in Fig.
P616
motor
Demodulator
ed
'f
~ constant
dc amplifier
60 Hz
ac
o
Synchros
Figure P616.
314
TimeDomain Analysis
of Control
Systems
Chap. 6
K = sensitivity of error detector JL = load inertia BL = load viscous friction Jm = motor inertia
s
volt/rad
0.05 ftlb/rad/sec 2
0.005 ftlb/rad/sec
0.05 ftlb/rad/sec 2
negligible
1
= motor viscous friction Ki = motor torque constant L = motor armature inductance Ra = motor armature resistance Kd = gain of demodulator Gear ratio n = Ni/N 2
B;
(a)
ftlb/amp
negligible
ion
1
volt/volt
1
1:
Write the characteristic equation of the system and determine the value of the dc amplifier gain A for a critically damped system.
(b)
For a
value of
what should be the minimum 9 L (t) will follow the reference input with a positional error not exceeding 0.0186 rad? With this gain setting, evaluate the output response 9 L (t).
unit
= tu {t),
s
6.17.
In the feedback control system shown in Fig. P617, the sensitivity of the synchro error detector is 1 V/deg. After the entire system is set up, the transfer function of the twophase motor is determined experimentally as
Km
s(l
+ Tm s)
where
Km
10 voltssec, and
Tm =
0.1 sec.
/
/
2phase
servomot
ec
=eL
,e m
=iooei
Figure P617.
(a) If the
is
to be driven in
its
the
minimum
value of gain
of the amplifier in
order that the deviation between output and input positions will not exceed
(b)
when the steady state is reached ? The gain of the amplifier is given by A = 35. Determine C and the undamped natural frequency of the system.
the
damping
ratio
Chap 6
'
Problems
315
(c)
The
amplifier
is
of the amplifier
written
e 2 {t)
Ae(t)
+ ATd d ^is
where
A 35. Determine the value of Td so that the damping ratio per cent. Repeat part (a) with the modified amplifier.
40
7
Stability of Control
Systems
7.1
Introduction
It
system
was shown in Chapter 6 that the transient response of a linear feedback control is governed by the location of the roots of the characteristic equation.
may be
regarded as
a way that the corresponding system will perform according to the prescribed
specifications.
We shall learn in
Chapter
1 1
must
Among
the
is that the system must be However, there are various ways of defining stability, especially when we include all types of systems, such as linear and nonlinear. In any case, the notion of stability is used to distinguish two classes of systems: useful and useless. From a practical standpoint, a stable system may be useful, whereas an unstable system is useless. In essence, the analysis and design of linear control systems is centered on stability studies. The notion of stability may be expanded to include absolute stability and relative stability. Absolute stability refers to the condition of stable or unstable. Once the system is found to be stable, it is of interest to determine how stable it is, and the degree of stability is measured by relative stability. Parameters such as overshoot and damping ratio used in relation to the transient response in Chapter 6 provide indications of the relative stability of a linear timeinvariant system in the time domain.
316
Sec 7
'
'
Stability. Characteristic
317
In this chapter
we
is
are concerned with the subject of absolute stability of simply a yes or no proposition.
From the illustrative examples of Chapter 6 we may summarize the relation between the transient response and the characteristic equation roots as follows:
1.
all the roots of the characteristic equation are found in the half of the jplane, the system responses due to the initial conditions will decrease to zero as time approaches left
When
infinity.
2.
one or more pairs of simple roots are located on the imaginary axis of the jplane, but there are no roots in the right half of the jplane, the responses due to initial conditions will be undamped
If
sinusoidal oscillations.
3.
one or more roots are found in the right half of the jplane, the responses will increase in magnitude as time increases.
If
sponses.
In linear system theory, the last two categories are usually defined as Note that the responses referred to in the above conditions are due to initial conditions only, and they are often called the zeroinput re*
unstable conditions.
7.2
We can show from a more rigorous approach that the zeroinput stability of a linear timeinvariant system is determined by the location of the roots of the characteristic equation or the behavior of the state transition matrix <h(t) Let a linear timeinvariant system be described by the state equation
x(0
where x(t)is the
satisfies the
= Ax(t) + Bu(r)
the input vector. For zero input x(t) equation x(0 Ax(f ) and is defined
(7
.jx
state vector
and u
homogeneous
state
equd&num state of the system. The zeroinput stability is defined as follows If the zeromput response x(t\ subject to finite initial state
as the
x(t
is
),
returns to the
tzzjzsr
A
I
is
as
ms
said to bistable
* k
*"
there
is
the foregoing definition may be statedsaid to be stable {zero input) if for any finite a positive number [which depends x(l)] sZhthat
>
\\x(t)\[<M
for all
t>t
rj_ 2
and
(2)
lira
t">3

x(r)

(73)
318
Stability of Control
Systems
Chap. 7
where
\ \
x(t)

represents the
norm* of the
l*(OII
=r
11/ (74)
_(=1
The condition
(>
as represented
(73) states
infinity.
any by the norm of the vector x(r) must be bounded. Equation that the system must reach its equilibrium state as time approaches
stated in Eq. (72) implies that the transition of state for
is
The
illustrated
by the secondorder case shown in Fig. 71. The state trajectory represents the transition of x(t) for t > t from a finite initial state x(r ). As shown in the figure, x(t ) is represented by a point that is the tip of the vector obtained from the forms the upper bound vector sum *i(r ) and x 2 (t ). A cylinder with radius for the trajectory points for all t > t and as t approaches infinity, the system
reaches
its
0.
*2<>o)
Next we
shall
show
that the definition of stability of linear timeinvariant same conclusion on the condition of the roots
of the characteristic equation. For the zeroinput condition, the state transition equation of the system is
x(f)
<J>(r
(.W,)
(75)
where {t t ) is the state transition matrix. Taking the norm on both sides of Eq. (75) gives
x(OII
IIW'o)x('o)l
x is
(76)
The norm
number.
of a vector
is
always a real
Sec. 7.3
319
An important property
which
of the
norm of a
vector
is
lx(0<<K''o)lx(f
is
)[
(77)
analogous to the relation between lengths of vectors. Then the condition in Eq. (72) requires that lx(f ) be finite \\tft Thus, if x(/ ) is finite as postulated, <j>(t /) must also be finite for / t Similarly, Eq. (73) leads to the condition that
1

 OH
>
lim

Mtt) =
')
(7. 8 )
or
! *'& i,j
=
<f>(t
(79)
n, where 1, 2, (t , /) is the yth element of <f>u In Eq. (442) the state transition matrix is written
.
. .
 /).
(7 .10)
<(>(0
 [(iIA)']
1
Since si A is the characteristic equation of the system, Eq (711) implies that the time response of ft/) is governed by the roots of the characteristic equation. Thus the condition in Eq. (78) requires that the roots of the

characteristic equation
must
all
have negative
real parts.
7.3
systems given in the for the zeroinput condition, we can show that, in general the stability condition for this class of systems is independent of the inputs An alternative way of defining stability of linear timeinvariant
Although the
preceding section
systems
is
system
as follows
is
bounded for any bounded input be the output and r(t) the input of a linear system
K0l<JV<oo

for/>/
for
/
(7 . 12)
t
< M < oo
>
(7_ 13 )
subjected to a unit steo function input. In this case the amplitude of the input is bounded but the amplitude of the output is not, since an impulse is
However, there are a few exceptions to the foregoing definition. gives rise to an impulse response at / = t when it is
A differentiator
have
a'n infinite
known
to
However, since a differentiator and useful systems, they are defined as stable systems and are exceptions to the stability condition defined above. shall show that the bounded inputbounded output definition of stability again leads to the requirement that the roots of the characteristic equation be located in the left half of the jplane an integrator are
all
amplitude. Also, when a unit step function the output is an unbounded ramp function.
is
We
320
Stability of Control
Systems
Chap. 7
Let us express the inputoutput relation of a linear system by the convolution integral
c(t)
= T r{t  z)g(r) dr
J
(714)
where g(r) is the impulse response of the system. Taking the absolute value on both sides of Eq.
\c(t)\
(714),
we
get
(715)
= \r r{tx)g{x)dx
I
J is
written
rfr
\c(t)\<
I""
r(f
 t)1
*<t)

(716)
Now
if r{i) is
" < Jo N
f
g(z) dx

= N f" JO
g(r) dx

(717)
Therefore,
if c{t) is
to be a
bounded output,
N
or
[\g{T)\dT<M<oo
I
(718)
rg(T)^T<
J o
P<oo
(719)
(719)
is
g(t),
evaluated from
to
oo,
must be
shall now show that the requirement on the impulse response for can be linked to the restrictions on the characteristic equation roots. By definition, the transfer function G{s) of the system and the impulse response g{i) are related through the Laplace transform integral
stability
We
G(s)=
f g(t)e'"dt
left
(720)
\G(s)\<r\g(t)\\e"\dt
I
(721)
The
e~"
is

roots of the characteristic equation are the poles of G(s), and when s oo. Also, s a +jco; the absolute value of
e~"
\.
(722)
one or more roots of the characteristic equation are in the right half or 0, and thus \e~'"\<N= 1. Thus on the imaginary axis of the splane, a
If
>
Sec. 7.4
Methods
Systems
321
Eq. (722)
is
written
for Re(s)
<f N\g(t)\dt = j\ g
Q
(t)\dt
(723)
> 0.
Since Eq. (723) contradicts the stability criterion given in Eq (719) we conclude that for the system to be stable, the roots of the characteristic equation must all he inside the left half of the splane. The discussions conducted in the preceding sections lead to the conclusions that the stability of linear timeinvariant systems can be determined by checking whether any roots of the characteristic equation are in the right half or on the imaginary axis of the *plane. The regions of stability and instability in the splane are illustrated in Fig. 72. The imaginary axis, excluding the origin, is included in the unstable region.
]<x>
v
Stable
splane
region
^/ \
/
Unstable
region
Stable
region
^ l Oy
K
'*
Unstable
region
\/
Fig. 72. Stable
in the .splane.
7.4
Methods of Determining
Systems
Although the
stability
may
by finding the roots of the characteristic equation, these criteria are difficult to implement in practice. For instance, the impulse response is obtained by taking the inverse Laplace transform of the transfer function, which is not always a simple taska similar process is required to evaluate the state transition matrix Mt) The solving of the roots of a highorder polynomial can only be carried out' by a digital computer. In practice, therefore, the stability analysis of a linear system
investigating the
be checked by
322
Stability of Control
Systems
Chap. 7
is
seldom carried out by working with the impulse response or the state transiby finding the exact location of the roots of the characteristic
equation. In general,
to apply
we
and which can provide answers to stability or instability, without excessive computations. The methods outlined below are frequently used for
the stability studies of linear timeinvariant systems.
1.
RouthHurwitz criterion an algebraic method that provides information on the absolute stability of a linear timeinvariant system. The criterion tests whether any roots of the characteristic equation
1
:
lie
The number of
roots that
lie
on
and
Nyquist criterion 6
between the number of poles and zeros of the closed loop transfer function by observing the behavior of the Nyquist plot of the loop transfer function. The poles of the closed
on the
difference
loop transfer function are the roots of the characteristic equation. This method requires that
we know
8):
roots
lie
when
When
closedloop system
4.
unstable.
the loop transfer used to determine the stability of the may be closedloop system. However, the method can be used only if G(s)
5.
H(s) has no poles and zeros in the righthalf splane. Lyapunov's stability criterion: a method of determining the stability of nonlinear systems, although it can also be applied to linear systems. The stability of the system is determined by checking on the
properties of the Lyapunov function of the system.
7.5
RouthHurwitz Criterion
1"
The RouthHurwitz
criterion represents a
location
of zeros of a polynomial with constant real coefficients with respect to the left half and the right half of the jplane, without actually solving for the
zeros.
The method can be applied to systems with and with multiple inputs and outputs, as well as
single input
and output
single or multipleloop
systems.
of the form
F(s)
s"
a,5"'
a z s"' 2
an _
an
(724)
where
all
Sec. 7.5
RouthHurwitz
Criterion /
323
it is
In order that there be no roots of the last equation with positive real parts necessary but not sufficient that*
1.
2.
All the coefficients of the polynomial have the same sign. None of the coefficients vanishes.
How
with
halfofthejplane.
the left half
The necessary and sufficient condition that all the roots of Eq. (724) of the splane is that the polynomial's Hurwitz determinants, 1>2 n, must be all positive. The Hurwitz determinants of Eq. (724) are given by
Dk
D =
t
ai
D2 =
0
a3 a2
a,
a3 a2
*3
do
a\
o3 a2
a,
a5
2l
a4
22 023
a3
a2
"t
24
25
(725)
where the
replaced by zeros.
At first glance the application of the Hurwitz determinants may seem to be formidable for highorder polynomials, because of the labor involved in evaluating the determinants in Eq. (725). Fortunately,
mtoa
the rule
was
simplified
by Routh
work with
in
e/S
C baSfC
kWS f a
'
gebra
'
obse
~=
all
roots
at a time
Zl
taken 3 at a time
~
AH
the ratios
=(!)"
products of
all
roots
at least
must be
positive
p&rt.
real
324
Stability of Control
Systems
Chap. 7
The
first,
first
RouthHurwitz criterion is to first row consists of the coefficients, and the second row consists of the second, coefficients, as shown in the following tabulation
ao
a\
az
a$
a\
as
a& an
as
09
... ...
The next
step
is
shown
is
s" 55 s*
ao
a\
ai a3
^
<H a5
a6
a\a 2
a\
a az
a\a*
a\
aoa; a\a6
=B =
n
06
aia *
a\
i
S3
s2
Aai
a\B
c
_ E
Aai
xO A
= a6 x0 _
BC  AD c ED Cat, E Fa 6  E x
A Ca 6  A x
CxO/ixO =
c
F
r.
a6
tabulation or the
The array of numbers and operations given above is known as the Routh Routh array. The column of ^s on the left side is used for Once Routh tabulation has been completed, the last step in the Routhis to investigate the signs of the numbers in the first column tabulation. The following conclusions are drawn The roots of the polythe
:
identification purpose.
Hurwitz criterion
of the nomial are all in the left half of the splane if all the elements of the first column of the Routh tabulation are of the same sign. If there are changes of signs in the elements of the first column, the number of sign changes indicates the number of
roots with positive real parts.
The reason for the foregoing conclusion is simple, based on the requirements on the Hurwitz determinants. The relations between the elements in the first column of the Routh tabulation and the Hurwitz determinants are:
a
a,
=a =D
a
t
D
D
ES
3
'a
Sec 7 5
' 
RouthHurwitz
Criterion /
325
Therefore, if all the Hurwitz determinants are positive, the elements in the first column would also be of the same sign. The following two examples illustrate the application of the RouthHurwitz
criterion to simple problems.
Example
71
 2)(s +
1)0
3)
= s  4s 2 + s +
3
(726)
which has two negative coefficients. Thus, from the necessary condition, we know without applying the RouthHurwitz test that the equation has at least one root with positive real parts. But for the purpose of illustrating the RouthHurwitz criterion, the Routh tabulation is formed as follows:
s3
Sign change
*2
4
(4)W(6)(1)
Sign change
=25
=
6
Q Q
(25X6)
(4)(0)
Since there are two sign changes in the first column of the tabulation, the polynomial has two roots located in the right half of the jplane. This agrees
result, as
= 3.
with the known Eq. (726) clearly shows that the two righthalf plane roots are ats 2 and
Example
72
+ s + 3s + 5s +
3
10
(727)
same
the necessary condition for not having roots in the right half of the 5plane or on the imaginary axis. However, the sufficient condition must still be checked. The Routh tabulation is made as follows
satisfies
2
1
10
5
2)(5)
Sign change
)(
Sign change
V
10
=7
=
643
10
(7)(5)_dX10)
first
Special Cases
The two illustrative examples given above are designed so that the RouthHurwitz criterion can be carried out without any complications. However depending upon the equation to be tested, the following
difficulties
may
occur
326
Stability of Control
Systems
Chap. 7
occasionally
when
Routh
test:
1.
The
first
is
zero,
The elements
first
all zero.
In the
in the next
row
situation
may be
is
become infinite, and the Routh test breaks down. This corrected by multiplying the equation by the factor (s a),
where a
Example
73
that at least
1)
2
2
2)
is
= s* zero,
3s
(728)
we know from
determine
how many
is located in the right half of the jplane. To of the roots are in the righthalf plane, we carry out the Routh
tabulation as follows
.$3
3
2
s2 Sl
first
first
element of the
row
is infinite.
To
the factor
one's
(s
+
1
a),
mind
is 1.
choose a to be
where a is However, for reasons that will become apparent or 2. Let a = 3; then Eq. (728) becomes
multiply both sides of Eq. (728) by an arbitrary number. The simplest number that enters
later,
we
we do not
(729)
(s
l)\s
+ 2)(s +
3)
 si +
is
3s 3
3s 2 7s +
=0
The Routh
3 7
2
3
Sign change
?
3
+ 7_ ~
~
6
18
Sign change
( jX7)
2Q
Since there are two changes in sign in the first column of the Routh tabulation, the equation has two roots in the right half of the fplane.
As an
alternative to the
remedy of the situation described above, we may Routh tabulation by an arbitrary small positive
*If
one chooses to use a negative number for a, the (s + a) term will contribute a root in and this root must be taken into account when interpreting the
Routh
tabulation.
.:
Sec 7 5

RouthHurwitz
Criterion
327
number
we may
;
test. For instance, for the equation replace the zero element in the second row of the
Sl
Sign change
Since e is postulated to be a small positive number, ( 3e 2)/e approaches 2/e, which is a negative number; thus the first column of the last tabulation has two sign changes. This agrees with the result obtained earlier.
On
the
other hand,
we may assume
e to be negative,
and we can
is still two, but they are between the first three rows. In the second special case, when all the elements in one row of the Routh tabulation are zeros, it indicates that one or more of the following conditions
may
exist
2.
3.
Pairs of real roots with opposite signs. Pairs of imaginary roots. Pairs of complexconjugate roots forming symmetry about the origin of the splane.
The equation that is formed by using the coefficients of the row just above row of zeros is called the auxiliary equation. The order of the auxiliary equation is always even, and it indicates the number of root pairs that are equal in magnitude but opposite in sign. For instance, if the auxiliary equation is of the second order, there are two equal and opposite roots. For a fourthorder auxiliary equation, there must be two pairs of equal and opposite roots. All these roots of equal magnitude can be obtained by solving the auxiliary equation. When a row of zeros appear in the Routh tabulation, again the test breaks down. The test may be carried on by performing the following remedies:
the
1.
2.
3.
derivative of the auxiliary equation with respect to s. Replace the row of zeros with the coefficients of the resultant equation obtained by taking the derivative of the auxiliary equation. Carry on the Routh test in the usual manner with the newly
Take the
formed
tabulation.
Example
74
Consider the same equation, Eq. (728), which is used in Example 73. In multiplying this equation by a factor (s + a), logically the first
number
that
comes
into one's
(s
=
2
I.
Multiplying
l) 2 (s
2)(s
1)
= j* + ^ _
_^+
(7 . 30)
328
Stability of Control
Systems
Chap
The Routh
tabulation
is
made
as follows:
s4 s3
V ^
1 1
ax 3) ax
j
10
3
.
, 2
^_2=0
Since the s 1
row contains
1,
all zeros,
the
Routh
test
terminates prematurely.
The
has a root
(2).
at s
by the factor
is,
1).
To remedy
this situation,
we form
the
row
This makes the new equation fit special case the auxiliary equation using the coefficients preceding the row of zeros. Thus the auxiliary
A(s)
2s 2
+2=0
s gives
(731)
4.s
is
(732)
Now,
the
row of zeros
in the
Routh
tabulation
(732),
s>
3 1
2
(coefficients of auxiliary equations)
Sign change
s2 i1
2 4
2
[coefficients of dA(s)jds]
Sign change
s
Since the preceding tabulation has two sign changes, the equation of Eq. (728) has two roots in the righthalf plane. By solving the roots of the auxiliary equation in Eq. (731), we have
s1
or
= 1
These are also the roots of the equation in Eq. (730). It should be remembered that the roots of the auxiliary equation are also roots of the original equation, which is under the Routh test.
Example
75
As
(733)
which
s
is
known
to have
signs at i
= 2
and
1
1 1
2 3 3
7 4 4
4
Sec 7

RouthHurwitz Criterion
329
Since a
row of
coefficients
zeros appears prematurely, we form the auxiliary equation using the of the s* row. The auxiliary equation is
A(s)
= s* is
3s 2
(734)
The
jf = 4s 6s =
3
(735)
in the
from which the coefficients 4 and 6 are used to replace the row of zeros tabulation. The new Routh tabulation is
Routh
1 1
2 3 3 6 4
7 4 4
[coefficients of
4
Sign change
1.5
16.7
^^]
4
Since there is one change in sign in the first column of the new Routh tabulation, Eq. (733) has one root in the right half of the jplane. The equal, but opposite roots that caused the allzero row to occur are solved from the auxiliary equation. From Eq. (734) these roots are found to be
1
frequent use of the RouthHurwitz criterion is for a quick check of the and the simple design of a linear feedback control system. For example, the control system with integral control shown in Fig. 629 has the characteristic equation
stability
s3
34.5s 2
7500s
7500/^,
(736)
The RouthHurwitz
criterion
may
is
be used to determine the range of AT, for stable. The Routh tabulation of Eq. (736) is
7500
7500ATi
j2
,.,
34.5
258,750
7500A,
341
s
7500ATi
to be stable, all the coefficients in the first column of the last tabulation should be of the same sign. This leads to the following conditions:
258,750
7500/sT,
343
7500*,
> >
(737)
7_ 38)
From
we have
AT,
< 34.5
(7.39)
330
Stability of Control
Systems
Cn ap 7
in Eq. (738) gives
Ki
Therefore, the condition for stability
>
that A^ must satisfy
34.5
x
(740)
is
<K <
Example
76
(741)
3Ks 2
(K
2)s
+4=
is
(742)
stable.
It is
The Routh
tabula
s3 s2
(K
+
4
2)
3K
so
4
is
From
K>
and from the
s 1 row, the condition of stability
is
3K + 6K  4 >
2
or
K<
Thus
2.528
or
K>
it
0.528
is
When
the conditions of
K>
and
K>
one.
more stringent
K must satisfy
K>
The requirement of K
It
0.528
<
2.528
is
disregarded since
K cannot be negative.
is
valid only
if
algebraic and
all
one of the coefficients of the characteristic equation is a complex number, or if the equation contains exponential functions of s, such as in the case of a system
with time delays, the RouthHurwitz criterion cannot be applied.
Another limitation of the RouthHurwitz criterion is that it offers information only on the absolute stability of the system. If a system is found to be stable by the Routh's test, one still does not know how good the system is
in other words,
how
On
if
the system
is
shown
to be unstable, the
RouthHurwitz
criterion gives
no indication of how
7.6
Thus
far,
stability
The
Sec. 7.6
Nyquist Criterion
331
on a
digital
finding subroutine.
2.
The
These two methods are not very useful for design purposes. It is necessary to devise stability criteria that will enable the analyst to determine the relative stability of the system. It would be desirable if the criterion indicates how the stability of the system might be improved.
The Nyquist criterion possesses the following features that for the analysis as well as the design of control systems
1
make
it
desirable
2.
3.
provides the same amount of information on the absolute stability of a control system as the RouthHurwitz criterion. In addition to absolute system stability, the Nyquist criterion indicates the degree of stability of a stable system and gives an indication of how the system stability may be improved, if needed. It gives information on the frequencydomain response of
It
the system.
4. It
5.
It
can be used for a stability study of systems with time delay. can be modified for nonlinear systems.
formulate the Nyquist criterion by first using modern control state equations. Let a linear control system with a single input be represented by the following set of equations:
notation, that
is,
We may
x(?)
block diagram that gives the transfer relations of the system is shown in Fig. 73. This is known as a closedloop system with unity feedback. The openloop transfer function of the system is defined as
G{s)
= H(sl ~ A) "'I*
(7 . 46)
When
the feedback
is
G(s)
>
D
C(s)
R(s)
"V*?
(si
A) IB
Fig. 73.
back.
332
Stability of Control
Systems
Chap. 7
*~ C(s)
Fig. 74.
back.
74.
The closedloop
CO)
R(s)
1
G(s)
G(s)H(s)
(747)
byFO). Then
F(s)
It is
= +
1
G(s)H(s)
= + D(sl  A)'B#(s)
1
(748)
easy to see that the zeros of F(s) are the roots of the characteristic equation of the system, if there is no polezero cancellation in G(s)H(s).
Let us assume that G(s) and H(s) are rational functions; that is, G(s)H(s) a quotient of two polynomials with constant coefficients. In general, F(s) can be written
is
+ z,)(j + z ) (2 + z m ) (749) + pMs + p )...(s + p where z,(i = 1,2, ... ,m) &ndp k (k = n) are either real or in complex2, conjugate pairs. Then the roots of the characteristic equations are s = z
F(s)
K(s
s J (s
n)
1,
t ,
z 2
zm
to be stable,
it is
of these roots has a positive real part. There is no particular restriction on the location of the poles of F(s), which are at s , p. 0, It is p l7 p z important to note that the poles of F(s) are the same as those of G(s)H(s). If
lies in
still be of F(s) are found in the left half of the ,?plane. This is a very important feature of a feedback control system. In preceding chapters it has been pointed out that a high forwardpath gain generally reduces the steady
state error
it is
may
result in
an
made
stable
by proper design. Before embarking on the fundamentals of the Nyquist criterion, it is essential to summarize on the polezero relationships with respect to the system
functions.
Sec. 7.6
Nyquist Criterion
333
1.
Identification of poles
and zeros:
= = =
G(s)H(s)
transfer
2.
The
function G(s)H(s).
3.
For a closedloop system to be stable, there is no restriction on the and zeros of the loop transfer function G(s) H(s), but the poles of the closedloop transfer function must all be
located in the
left half
of the
.splane.
Encircled.
the path.
since
A point is said to be encircled by a closed path if it is found inside For example, point A in Fig. 75 is encircled by the closed path T, is found inside the closed path, whereas point B is not encircled. In
it.
the application of the Nyquist criterion, the closed path also has a direction
associated with
As shown
all
is
said to be encircled
by
When
considering
we can
say that
334
Stability of Control
Systems
Chap. 7
said to be enclosed by a closed path if it is when the path is traversed in the prescribed found to lie to the left of the path direction. Putting it another way, enclosure is equivalent to counterclockwise encirclement. For instance, the shaded regions shown in Fig. 76(a) and (b) are considered to be enclosed by the closed path I\ In other words, point A in Fig. 76(a) is enclosed by I\ but point A in Fig. 76(b) is not. However, in Fig. 76(b), point B and all the points in the region outside T are considered to be enclosed.
Enclosed.
point or region
is
(a)
(b)
and
is
enclosed
by T.
(b) Point
is
B is enclosed
by the locus T.
Number of
encirclement
and
enclosure.
When
point
is
encircled or
number
to
N may be
be.
assigned to the
number of
en
The value of N may be determined any arbitrary point s on the closed path r and
may
follow the path in the prescribed direction until it returns to the starting point. The total net number of revolutions traversed by this vector is N. For example, point A in Fig. 77(a) is encircled once by T, and point B is
(a)
(b)
enclosure.
Sec. 7.6
encircled twice,
all in
once; point
is
enclosed twice.
Principle of the
Argument
The Nyquist criterion was originated as an engineering application of the wellknown principle of the argument in complex variable theory. The principle is stated as follows, in a heuristic manner. Let F(s) be a singlevalued rational function that is analytic everywhere in a specific region except at a finite number of points in the splane. For each point at which F(s) is analytic in the specified
region in the splane, there
is
Suppose that a continuous closed path r, is arbitrarily chosen in the splane, as shown in Fig. 78(a). If all the points on I\ are in the specified region
,,
](X>
ly'ImF
splane
F(s)plane
*
ReF
(a)
(b)
Correspond
ing locus
TF
in the
FO)plane.
in
which F(s)
is
is
F(s)plane
rF mapped by the function F(s) into the shown in Fig. 78(b). If, corresponding to
located in the F(s)plane, then as the r,
point s
locus
F(s
) is
is
and then returning to jj after going through all the points on the shown in Fig. 78(a)], the corresponding F P locus will start from point F(sj) and go through points F(s 2 ) and F(s 3 ), which correspond to s 2 and s 3 respectively, and return to the starting point, F(Ji). The direction of traverse of r F may be either clockwise or counterclockwise; that is, in the same direction or the opposite direction as that of r s depending on the particular function
(clockwise)
locus [as
F(s).
rP
is
shown, for
illustration purposes, to
be
counterclockwise.
It
the F(j)plane
should be pointed out that, although the mapping from the jplane to is one to one for a rational function F(s), the reverse process is
336
Stability of Control
Systems
Chap. 7
which is analytic in the finite .splane except at the points s = 0, 1, and 2. For each value of s in the finite .splane other than the three points 0, 1 and 2, there is only one corresponding point in the F(.s)plane. However, for each
,
maps
Eq. (750) as
(751)
s{s+l)(s
The
F(s)
left side
is
+ 2)=^
of Eq. (751)
is
chosen to be a constant.
principle of the
is
The
argument can be
that
is
chosen
in the splane
s;
mapped
by
many
the
number of
the zeros
and
the
IV
N= Z  P
where
(752)
N=
Z=
P
origin
made by
the F(s)plane
rF
encircled by the splane locus
T
T
in the
encircled
by the
.splane locus
in the
can be positive (Z > P), zero (Z In general, These three situations will now be described.
1.
= P),
or negative (Z
< P).
N>
wise),
(Z
> P).
is
a positive integer. In this case the F(s)plane locus will times in the same direction
as that of 1%.
2.
/V
=
<
(Z
TV
(Z
< P).
FF
direction
from that of T
A convenient way of determining N with respect to the origin (or any other
point) of the F(s) plane
is
to
draw a
line
in
any direction to
Sec. 7.6
Nyquist Criterion
337
F(s)plane
F(s)p\ane
N=2
F(s )plane
N=
Fig. 79.
N in the F(i)plane.
infinity; the
number of net
magnitude of N. Figure 79 gives several examples of this method of determining N. It is assumed that the T s locus has a counterclockwise sense. A rigorous proof of the principle of the argument is not given here. The
following illustration
ciple.
may be considered as a heuristic explanation of the prinLet us consider the function F(s) given by
F(s)
+ />i)C* + Pi)
(753)
where
K is a positive
\F(s)\/F(s)
+ Pi
+p
Us
2
Is
+ Pi js + p
(754)
2)
338
Stability of Control
Systems
Chap. 7
xplane
ReF
of
(a)
(753).
5^
at s
st
given by
F( Sl )
zt)
+ />i)C*i + Pi)
x t
(755)
The factor s + z can be represented graphically by the vector drawn from z, to s Similar vectors can be defined for {s + p,) and (s + Pi)Thus F(s ) is represented by the vectors drawn from the given poles and zero to the point s u as shown in Fig. 710(a). Now, if the point s moves along the
t x
x
.
it
returns to the
Sec. 7.6
if there were any) that are not encircled by T s when Si completes one round trip are zero; whereas the vector (j, + z,) drawn from the zero at z u which is encircled by I\, generates a positive angle (counterclockwise sense) of 2n rad. Then, in Eq. (754), the net angle or argument of F(s) as the point Si travels around T s once is equal to 2n, which means that the corresponding F(s) plot must go around the origin 2n radians or one revolution in a counterclockwise direction, as shown in Fig. 710(b). This is why only the poles and zeros of F(s), which are inside the JT, path in the jplane, would contribute to the value of N of Eq. (752). Since poles of F(s) correspond to negative phase angles and zeros correspond to positive phase angles, the value of depends only on the difference between Z and P.
zeros
Z=
Thus
F=0
1
N = Z P=
which means that the F(y)plane locus should encircle the origin once in the same direction as the splane locus. It should be kept in mind that Z and P refer only to the zeros and poles, respectively, of F(s) that are encircled by r and not the total number of zeros and poles of F(s). In general, if there are more zeros than poles of F(s), which are encircled by the jplane locus r, in a prescribed direction, the net angle traversed by the .F(j)plane locus as the splane locus is traversed once is equal to
2n(Z
in the
P) = 2nN
(756)
This equation implies that the F(s)plane locus will encircle the origin
N times
same direction as that of r,. Conversely, if N more poles than zeros are encircled by T s in a given prescribed direction, N in Eq. (756) will be negative, and the F(s)p\ane locus must encircle the origin times in the opposite direction
,
to that of
IV
all
A
is
summary of
Table 71
Summary
Principle of the
Argument
F(s)Plane Locus
N=ZP
Sense of the
sPlane Locus
Number of Encirclements
of the Origin
Direction of
Encirclement
N>
N <0
Af
Clockwise Counterclockwise
N N
Clockwise
Counterclockwise Counterclockwise Clockwise
Clockwise
Counterclockwise
Clockwise Counterclockwise
No encirclement No encirclement
340
Stability of Control
Systems
Chap. 7
Nyquist Path
At
many
if
is
taken to be one
Of
course, as an alternative,
I\ can be chosen to encircle the entire lefthalf splane, as the solution is a relative one. Figure 711 illustrates a T s locus, with a counterclockwise sense, which encircles the entire right half of the splane. This path is often called the
Nyquist path.
Since the Nyquist path
the small semicircles
that the path should go
must not pass through any shown along the^'co axis in Fig. 711
singularity of F{s),
apparent that
splane
Poles of F(s)
Sec. 7.6
if
lies inside
it
will
be
For the convenience of analysis, the Nyquist path is divided into a miniof three sections. The exact number of sections depends upon how many of those small semicircles are necessary on the imaginary axis. For the situation illustrated in Fig. 71 1, a total of eight sections needs to be denned. The order + of numbering these sections is entirely arbitrary. The notations j co\, j co~i, +j0 + j0 ja>U and jcoj are used to identify the starting and ending points of
mum
Section 1 from s +j to jcot along they'co axis. Section 2: from +jcot to +jcoT along the small semicircle around
:
=j bi
section 3
Section 4: Section 5
:
from 5 = jco~[ to +J0 + along the jco axis. from +j0 + to y'0 + along the small semicircle around s = from s = j0 + to year along the jco axis (mirror image
of section
3).
Section 6: from j
s
jco~[
to
j cot along
ja>i (mirror image of section 2). j along the jco axis (mirror Section 7: from s jco\ to s
Section 8: from
1).
to
+jo along
the
semicircle
of
The Nyquist criterion is a direct application of the principle of the argument when the jplane locus is the Nyquist path. In principle, once the Nyquist path
is specified,
the F(s)
G(s)H(s) locus
when
j takes
and
investigating the behavior of the F(s) plot with respect to the origin of the
is
However, since usually the approach is to construct the Nyquist plot of G(s)H(s), and the same result of F(s) can be determined from the behavior of the G(s)H(s) plot with respect to the (l,jO) point in the G(s)H(s)F(s)plane. This
called the Nyquist plot of F(s).
given, a simpler
function G(s)H(s)
is
plane. This
is
l,y'0)
if
We
are inter
two types of
stability
stability of the
that G(s)H(s) has poles only in the left half of the jplane.
342
Stability of Control
Systems
Chap. 7
With
two
sets
this
stability
problem,
it is
necessary to define
of N, Z, and P, as follows
G(s)H(s)
number of zeros of G(s)H(s) that are encircled by the Nyquist path, or in the right half of the jplane
of poles of G(s)H(s) that are encircled by the Nyquist path, or in the right half of the jplane
= =
number of zeros of 1 G(s)H(s) that are encircled by the Nyquist path, or in the right half of the jplane
Nyquist path, or in the right half of the jplane
Several facts
at this point
Pi
(757)
since G(s)H(s)
and
Z_!
but openloop stability requires that
(758)
P =
The crux of
the matter
is
(759)
is
determined by the
stability studies
principle of the
argument for
summarized as follows
1.
is
defined according
The Nyquist plot of G(s)H(s) is constructed. The values of N and JV_, are determined by observing the behavior of the Nyquist plot of G(s)H(s) with respect to the origin and the
( \,j0)
4.
point, respectively.
1
(if it is
not already
N =Z P
if
(760)
,
Z
[
is is
given.
Once
P
t
is
determined, P_
=P
[Eq. (757)],
and
Z_
determined from
N. =
Since
it
Z_,
/>_!
(761)
must
(762)
P_,
Sec. 7.6
may
loop system to be
point as
the splane,
stable, the
many times as
and
it is
In general,
number of poles of G(s)H(s) that are in the right half of must be made in the clockwise direction. not always necessary to construct the Nyquist plot which
the
fact, a great
problems can be solved simply by sketching the part of G(s)H(s) that corresponds to the +ycoaxis of the splane. The condition when this simplified procedure is
possible
is
when P
0; that
is,
of the
.yplane.
Z_,
(763)
which means that the Nyquist plot of G(s)H(s) can encircle the ( l,j0) point only in the counterclockwise direction, since N_ in Eq. (763) can only be zero
i
is
equivalent to enclosure,
we
stability
critical
whether N.
only the
.yplane.
if we are interested only in we need not sketch the entire Nyquist plot for G(s)H(s), portion from s = +yoo to s = along the imaginary axis of the
is
zero,
The Nyquist
P_ case may be stated: If the function half of the splane, for the closedloop system to
t
be stable, the Nyquist plot of G(s)H(s) must not enclose the critical point ( 1J0). Furthermore, if we are not interested in the number of roots of the characteristic
stability,
equation that are in the righthalf plane, but only the closedloop only the G(s)H(s) plot that corresponds to the positive imaginary axis of the jplane is necessary. Figure 712 illustrates how the s /oo to s
Im
GH
G(s)#(s>plane
ReGH
Fig. 712.
= jco to s
0,
enclosed.
344
Stability of Control
Systems
Chap. 7
may
critical
point
( l.y'O)
is
enclosed.
7.7
Example
77
W = rh)
It is
(7  64)
where
Pa =
P_i
= 0.
To determine
see
if it
corresponds to j
=/co
to s
encloses the
1,7*0)
we
shall construct
is
shown
in Fig. 713.
splane
Example
77.
it is
0.
shown
in Fig. 713.
may be
e">
(765)
where { 0) and 9 denote the magnitude and phase of the phasor, respectively. As the Nyquist path is traversed from +j0 + to jO* along section 2, the phasor of
Sec. 7.7
345
i /
Im
GH
GCs)i/(s)plane
splane
80 counterclockwise
rotation
*~
*
ReGH
(a)
co
(b)
Eq. (765) rotates in the clockwise direction through 180. Also, in going from
to
+j0 +
to The corresponding Nyquist plot of , 9 G(s)H(s) can be determined simply by substituting Eq. (765) into Eq. (764). Thus
varies
0.
y'0 +
from +90
90 through
G(s)H(s)
s=ee )o
K
ee">(ee">
+ a)
(766)
Since e
0,
is
simplified to
G(s)H{s)
K
aeeJ<>
ooe je
(767)
which indicates that all points on the Nyquist plot of G(s)H(s) that correspond to section 2 of the Nyquist path have an infinite magnitude, and the corresponding phase is opposite that of the jplane locus. Since the phase of the Nyquist path varies from +90 to 90 in the clockwise direction, the minus sign in the phase relation of Eq. (767) indicates that the corresponding G(s)H(s) plot should have a phase that varies from 90 to +90 in the counterclockwise direction, as shown in Fig. 714(b).
In general,
plots, the
lim , s ^o s(s
K + a)
.
lim
#
=
is
(768)
inversely propor
As
is
1 80, the corresponding G(s)H(s) plot is traced out by a phasor with an infinite magnitude, 180 in the opposite or counterclockwise direction. It can be concluded that, in general, if the limit of
magnitude, from
+/0 + to
346
Stability of Control
Systems
Chap. 7
lim Jfo"
is
(769)
traced out
by a phasor of
direction
if
infinitesimally small
is
magnitude n x 180 degrees in the clockwise used, and by a phasor of infinite magnitude n x 180 in the
if
counterclockwise direction
is
used.
The
in Fig. 713
isolated, as
The
points
on the
semicircle
5
may
Rei*
plane
Re/*
l
Im
GH
G(,s)i/(.s)plane
180
counterclockwise
rotation
360 clockwise
rotation
ReG
Radius
(b)
(b)
Nyquist plot
where
oo. Substituting
G(s)H(s)
_
Re j*
K ~ RW* = Oei
1*
(771)
which implies that the behavior of the G{s)H{s) plot at infinite frequency is described by a phasor with infinitesimally small magnitude which rotates around the origin 2 x 1 80
360 in the clockwise direction. Thus the G(s)H(s) plot that corresponds to section
4 of the Nyquist path is sketched as shown in Fig. 71 5(b). Now to complete the Nyquist plot of the transfer function of Eq. (764) we must consider sections 1 and 3. Section 1 is usually constructed by substituting s = jca into Eq. (764) and solving for the possible crossing points on the real axis of the G(s)H(s)
Sec. 7.7
347
becomes
G(jOi)H(jOi)
K
jco(jco
+ a)
(772)
which
is
rationalized
G(jco)HUco)
(773)
The
intersect of
G(jco)HQco) on the
real axis is
found from
Im G(jCO)H(jco)
which gives
plane
is
=
co*
Kaco + a 2 C0 2
Ka
co(cb 2
+a
2 )
(774)
co
co. This
means that
on the real
Nyquist criterion is not concerned with the exact shape of the G(s)H(s) locus but only the number of encirclements, it is not necessary to obtain an exact plot of the locus. The complete Nyquist plot of the function of Eq. (764) is now sketched in Fig. 716 by connecting the terminal points of the loci
i /
Im
GH
G(s)H(s)p\ane
ReGH
Fig. 716.
K/ls(s
a)].
348
Stability of Control
Systems
Chap. 7
and
4,
finite
It is of interest to check all the pertinent data that can be obtained from the = iVi = 0. By inspection of Eq. (764), Z = Nyquist plot of Fig. 716. First, andP = 0, which satisfy Eq. (760). SinceP = Pu Eq. (761) leads to
Z_!
Therefore, the closedloop system
is
Ni
+P_!
=0
is
(775)
stable.
as
+K=
whose roots will always lie in the left half of the iplane for positive a and K. Figure 716 also shows that for this problem it is necessary only to sketch the
portion of G(s)H(s) that corresponds to section
that the
1
It is
all
apparent
positive
( 1, y'O)
values of K.
Example
78
Consider that a control system with single feedback loop has the loop
transfer function
<*>" =
The
characteristic equation of the system
s
2
WTTY
(7 " 77)
is
(1
+ K)s  K =
all
(778)
positive K.
Section 2.
e ie
lim G(s)H(s)
= lim
ooe^< + )
(779)
This means that the Nyquist plot of G{s)H{s) which corresponds to section 2 of the Nyquist path is traced by a phasor with infinite magnitude. This phasor starts at an
+90 and ends at 90 and goes around the origin of the G(s)H(s)p\ane counterclockwise a total of 180.
angle of
Section
4.
Re'*
lim G{s)H{s)
= lim =
0e~'*
(780)
Thus
the Nyquist plot of G(s)H(s) corresponding to section 4 of the Nyquist path goes around the origin 180 in the clockwise direction with zero magnitude. Section 1. s =jCO:
G(j<d)H(jco)
JooUm
1}
o4
+W
A
a>(co*
1)
(781)
we have
(782)
real axis.
= 1
rad/sec
at
Then
(783)
GUl)H(jl)
Based on the information gathered
=K
complete Nyquist
Sec. 7.7
349
lm
GH
G(s)H(s)
plane
AL, =
ReGH
Fig. 717.
K(s
l)/[s(s
1)].
plot of G(s)H(s)
is
shown
in Fig. 717.
Z =1
P =p_,
Figure 717 indicates that Ao Figure 717 also gives A/_j =
=o
agreement with the Nyquist
criterion.
=
1.
1,
which
is
in
Then
Z_, =AT_,
+P_, =
( 7 _84)
which means that the closedloop system is unstable, since the characteristic equation has one root in the right half of the 5plane. The Nyquist plot of Fig. 717 further
indicates that the system cannot be stabilized by changing only the value of A'.
Example
79
Consider the control system shown in Fig. 718. It is desired to determine the range of for which the system is stable. The openloop
is
C(s)
Grrt
10*(j
2)
(785)
on the jco
axis, the
Nyquist path
350
Stability of Control
Systems
Chap. 7
R(s)
N
E( S )
K(s + 2)
Y^
.S
10
2
(.5
C() C
+ 3)
>
Fig. 718.
79.
.splane
Fig. 719.
(785).
can consist of only three sections, as shown Nyquist plot of G(s) is outlined as follows
Section 4.
s
The construction of
the
Re 1 *
lim G(s)
10K
0e j2 *
(786)
As
the phasor for section 4 of the Nyquist path is traversed from 90 to +90 through 180 counterclockwise, Eq. (786) indicates that the corresponding Nyquist plot of G(s) is traced by a phasor of practically zero length from +180 to 180
through a
Section
=jco:
GUco) (l03co*)jcoi
Rationalizing, the last equation becomes
l0K(ja>
2)
(787)
G(jco)
10 AT[2(10
3CQ 2 )
co* +7CQ(10
3co 2 ) +./2CQ 3 ]
(10
3co 2 ) 2
co 6
(788)
Sec. 7.7
351
rad/sec
and
co
= VlO rad/sec
which correspond to the frequencies at the intersects on the real axis of the G(s)plane. In this case it is necessary to determine the intersect of the G(s) plot on the imaginary axis. Setting the real part of G(jG>) to zero in Eq. (788), we have
co 4
6co 2
20
(789)
which gives
co
1.54 rad/sec
C(/0)
2 AT
and
C(yVlO)
and the
intersect
= K
on the imaginary
axis
is
G(j^/T) =j]0/TK/3
With the information gathered
of Eq. (785)
is
in the
sketched as shown
in Fig. 720.
1
preceding steps, the Nyquist plot for G(s) The information on the imaginary axis
be determined without actually plotting
may
../ImC
G(s)plane
Fig. 720.
\QK(s
2)/(s 3
3s 2
10).
352
Stability of Control
Systems
Chap. 7
A = 2.
.splane,
Since Eq.
this
right half
of the
Z = 0;
means
P =
2.
Thus P_
=2. Now,
t
we have
(790)
t
JV_,
= Z_  P_t = Z_, 
Thus for the closedloop system to be stable, Z_i = 0, which requires that JV_ = 2. With reference to Fig. 720, the stability criterion requires that the ( 1,;0) point must
be encircled twice in the clockwise direction. In other words, the be to the right of the crossover point at K. Thus, for stability,
critical
point should
K>
The reader can
easily verify this solution
(791)
criterion
It
is
much
simpler to use in stability problems such as the one stated in this illustrative
example, in general the Nyquist criterion leads to a more versatile solution, which also includes information on the relative stability of the system.
7.8
it
informative to illustrate
added to a typical loop transfer function G{s)H{s). This investigation will also be helpful to gain further insight on the quick sketch of the Nyquist locus of a
are
given function.
G(s)H(s)
T1
(792)
Ti
is
for
< ca <
a semicircle, as
shown
ImGH
G(s)H(s)p\ane
CO
co
ReGH
Ik
Fig. 721.
Kj(\
+ T lS
).
Sec. 7.8
353
Addition of poles at s
0.
is
added to the
G(s)H(s)
K
is
(793)
The effect of adding this pole is that the phase of G(jco)H(ja>) 90 at both zero and infinite frequencies.
In other words, the Nyquist locus of G(j(o)H(joo)
that of Fig. 721 at co
is
reduced by
rotated by
90 from
and
a,
oo, as
shown
Im
GH
G(s)H(s)phne
*ReGH
Fig. 722.
K/[s(l
Tis)].
magnitude of G(jco)H(j(o) at
becomes infinite. In general, adding a pole co of multiplicity j at s to the transfer function of Eq. (792) will give the following properties to the Nyquist locus of G(s)H(s)
lim IG(j<B)H(jco)
(j
Z
1)
(794)
lim /GUco)H(jco)
co~.0
= j* =
= K
(795) (796)
(797)
limGO'a>)#C/e>)
\im\G{jm)H{j(o)\
G(s)H(s)
(798)
and
G(s)H{s)
In view of these illustrations
will affect the stability adversely,
it is
s\\
K +T
s)
(799)
at s
354
Stability of Control
Systems
Chap. 7
G(s)#Cs)plane
ReGH
= K/[s 2 (l + TlS
)].
(2)
G(s)H(s)
K/[sHl
Tis)].
When
a pole at s
l/T2
is
+ of G(s)H(s) at co =
(1
G{s)H(s)
K
7V)(1
is
+T
(7100)
2 s)
lim GUco)H(joi)
=K
(7101)
at co
is
lim G(jco)HUco)
<.a
lim
K = ~"
2
2.0)
/180
(7102)
m >o 1 \1
Im
GH
G(s)H(s)p\ane
*
Re Gtf
Kj[{\
+ TlS)(l + T2 s)].
(2)
G(s)H(s)
#/[(!
Tis)(l
+ T2 s)(\ + T
3 s)].
Sec. 7.8
355
Thus the
Eq. (792)
is
effect
of adding a pole at
= \/T
to shift the
quency, as shown in
phase of the Nyquist locus by 90 at infinite Fig. 724. This figure also shows the Nyquist locus of
G(s)H(s)
(1
7V)(1
K + 7(1 + T
(7103)
3 s)
These examples show the adverse effects on stability that result from the addition of poles to the loop transfer function.
Addition of zeros.
derivative control
stable.
It was pointed out in Chapter 6 that the effect of the on a closedloop control system is to make the system more In terms of the Nyquist plot, this stabilization effect is easily shown,
by 90
at co
+ 7 to = oo.
is
given
by
^>^
It
,( i
+ ri(i + r
2 ,)
(7
104>
can be shown that the closedloop system is stable for < (7, + T2 )/ T2 Suppose that a zero at s \\Td is added to the transfer function of Eq. (7104), such as with a derivative control. Then
T,
.
<K
G(s)H(s)
s(l
K{\
T,5)(l
+ 7 + TV)
(7105)
are sketched as
The Nyquist loci of the two transfer functions of Eqs. (7104) and (7105) shown in Fig. 725. The effect of the zero in Eq. (7105) is to add
,
.
Im
GH
G(s)H(s)p\ane
T T2 K{T + T2 )
t
*
KeGH
<(&. T)
(2)
co
Fig. 725.
Nyquist
G(s)H(s)
G(s)H(.s)
K(l
+ Tds)/[s(l +
TisXl
(2)
356
Stability of Control
Systems
Chap. 7
=
is
= 0. + T )/T T to
2
1
K(T + T^IT^il + Td
plane.
which
7.9
Stability of Multiloop
Systems
The
stability analyses conducted in the preceding sections are all centered toward systems with a single feedback loop, with the exception of the RouthHurwitz criterion, which apparently can be applied to systems of any conis
known
We
shall
now
illus
how
is
back loops.
In principle, all feedback systems with single input and output can be reduced to the basic system configuration of Fig. 74. Then it would seem that
the Nyquist criterion can be applied in a straightforward
lent
manner
to the equiva
loop transfer function G(s)H(s). However, owing to the multiloop nature of the original system, the poles and/or the zeros of G(s)H(s) may be unknown,
to the
means of a specific example. Figure 726 gives the block diagram of a control system with two loops. The transfer function of each
block
is
it is
G(s)
(s
(7106)
5]
The
stability
H(s) = 5
Fig. 726.
Sec. 7.9
Stability of Multiloop
Systems
357
locus of G(s), except that the poles of G(s) are not entirely known.
the construction of the entire Nyquist locus of G(s),
in
To avoid
we can
two stages, as there are two feedback loops. First, consider only the inner whose loop transfer function is G 2 (s)H(s). We shall first sketch the Nyquist locus of G z (s)H(s) for co < oo. The property of the G 2 (s)H(s) plot with respect to the ( l,jO) point gives an indication of the number of zeros of 1 + G 2 (s)H(s) that are in the right half of the Jplane. Having found this information,
loop,
<
we then proceed
co
<
<
G 2 (s)H(s)
5 s{s
+
,.
1)(j
(7107)
2)
j\mG 2 H
G 2 (sW(s)plane
ReG 2 //
Fig. 727.
Nyquist plot of
G 2 (s)H(s) =
Ms +
i)(s
2)].
Since the
itself,
( l,y'0)
point
1
is
is
stable
by
we
be
only have to investigate the crossover point of the G(s) locus with respect to the
( l.v'O)
range of
K for
stability is
is
Now,
726.
let
+2 s+
1
(7108)
Gl{s)
His)
~ sis + =5
K
IX*
(7109)
2)
(7110)
368
Stability of Control
Systems
Chap. 7
jlmG
G(i)plane
*
ReG
Fig.
[K(s
2)]/{U
10)[s(s
+
l)(j
2)
5]}.
In this case
we cannot use
the
method
parameter
is
However, we may
The openloop
G(S )
_ W~
G.(j)Ga(j)
1
+G +
2 (s)H(.
s
(s
(7111)
I0)[s(s
l)(j
+ 2) +
5tf]
as a gain factor of G(s), Nyquist locus of G(s)/K. However, we can write the characteristic equation of the overall system as
it
Since the
would be of no
s(s
10)(j
1)(*
2)
+s+2+
5K(s
10)
(7112)
K as
we
do not
contain K.
We have
1
+ s(s +
is
of the form
(7 " 113)
of the charac
by sketching the Nyquist locus of G 3 (s). However, the poles of G 3 (s) are not known, since the denominator of G 3 (s) is not in factored form. The zeros of the polynominal s(s + 10)(s + l)(s + 2) f s + 2 may be studied by investigating the Nyquist plot of still another function G 4 (j), which we create as follows
equation
investigated
may be
Figure 729 shows that the Nyquist locus of G 4 (s) intersects the real axis ( 1, y"0) point. Thus all the poles of G 3 (s) are in the left half
of the 5plane. The Nyquist plot of G 3 (s) is sketched as shown in Fig. 730. Since
Sec. 7.9
Stability of Multiloop
Systems
359
,.
/Im C 4
G 4 (s )plane
0.009
ReC,
r
\
G4 (s) =
/
(s
2)/[s(s
10)(j
1)(*
2)].
,.
Im G 3 (s)
C 3 0?)plane
*
ReG,
(s)
[5K(s
+s+
W)]J[ s (s
I0){s
l)(s
+ 2)
2].
<K<
on the
real axis
is
at
stability
10.
In this section we have investigated the application of the Nyquist criterion to multiloop control systems. For analysis problems, the stability of the system can be investigated by applying the Nyquist criterion in a systematic fashion from the inner loop toward the outer loops. For design problems when a system parameter to be determined for stability, it is sometimes necessary to start with the characteristic equation, which may be written in the form
Kh
P{s)
KQ{s)
Q
is
(7 _ 115)
where P(s) and Q(s) are polynomials. The stability of the system sketching the Nyquist plot of the equivalent openloop transfer
studied by
function
(7116)
<*)3gj!
360
Stability of Control
Systems
Chap. 7
Thus we have
locus represents a
also indicated a
method of applying
The
is still
written
*IA
the RouthHurwitz criterion
is
=
manner outlined
(7117)
We can also
is
apply
in Eqs. (7115)
and
On
independent
all
we can apply
other
7.10
Systems with time delays and their modeling have been discussed in Section 5.12. In general, closedloop systems with time delays in the loops will be subject
to
more
stability
Td is modeled by the
no longer have constant coefficients. Therefore, the RouthHurwitz criterion is not applicable. However, we shall show in the following that the Nyquist criterion is readily applicable to a system with a pure time
tion of the system will
delay.
Let us consider that the loop transfer function of a feedback control sysis represented by
G(s)H(s)
^e T G
's
(s)H, (s)
(7 1
8)
G^H^s) is a rational function with constant coefficients; Ta is the pure time delay in seconds. Whether the time delay occurs in the forward path or the
where
feedback path of the system
is
In principle, the stability of the closedloop system can be investigated by sketching the Nyquist locus of G(s)H(s) and then observing its behavior with
reference to the
is that it rotates the phasor by an angle of coTd radians in the clockwise direction. The amplitude of G^jcoJH^jco) is not affected by the time delay, since the magnitude of e~' mT' is unity for all frequencies. In control systems the magnitude of GiO'cw)^iO'co) usually approaches zero as co approaches infinity. Thus the Nyquist locus of the transfer function of Eq. (7118) will usually spiral toward the origin as co approaches infinity, and there are infinite number of intersects on the negative real axis of the G(s)H(s)ip\a.ae. For the closedloop system to be stable, all the intersects of the G(jco)H(jco) locus with the real axis must occur to the right of the (IJ0) point.
The
effect
GyfJo^Hiijco) at each co
Sec. 7.10
361
Im
GH
G(s)#(s)plane
^ReGH
T,=4
0.16
Fig. 731. Nyquist plots of G(s)H(s)
e T "/[s(s
l)(.s
2)].
G(s)H(s)
for several values of
e''G^H^s)
=
s(s
IX*
(7119)
2)
Td
It is
system
is
stable
when
is
Td is zero,
riorates as
Td
increases.
Td =
point.
sec.
This
The system is on the verge of becoming unstable when shown with the Nyquist plot passing through the ( l,y0)
Unlike the rational function case, the analytical solution of the crossover is not trivial, since the equations that govern the crossings are no longer algebraic. For instance, the loop transfer
may be rationalized in the usual manner by multiplying numerator and denominator by the complex conjugate of the denominator.
result
is
< cos <>>
The
GUcomjco)
T
<
ffi
ff i  Lt
(
/CQ(2
~ G)2)1
<
7  120)
362
Stability of Control
Systems
Chap. 7
The condition
for crossings
3co
2
on the
coTd
real axis
co(2
.
of the G(s)H(s)p\ane
2
is
sin
Td
co
cos coTd
=
1
which
is
for
all
frequencies, the
crossover problem
is
Bode
plot
Since the timedelay term affects only the phase but not the magnitude of G(jco)
H(jco), the phase of the latter
is
angle of
coTd
locus intersects the negative real axis. In general, analysis and design problems
more
it
Bode
diagram.
If the time delay is small,
is
e~
T"
Ts+
d
Tjs*
2!
3!
(7121)
Figure 732 shows the Nyquist plots of the transfer function of Eq. (7119)
Im
GH
G(s)//(i)plane
KeGH
Twoterm
approximation
Fig. 732.
Sec. 7.11
Stability of Nonlinear
Systems Popov's
Criterion /
363
with
Td =
0.8 sec,
and
G(s)H(s)
=
s(s
l7>
+
l)(s
(7122)
2)
which
is
the result of truncating the series of Eq. (7121) after two terms
7.11
Stability of Nonlinear
Systems Popov's
Criterion 16
was mentioned in Section 7.1 that the stability analysis of nonlinear systems is a complex subject, and unlike the linear timeinvariant case, no single method can be applied to all nonlinear systems. Although the major emphasis of this book is on linear systems, we shall show in the following that a certain class of nonlinear systems can be studied with stability criteria that are quite
It
When
a system
is
nonlinear,
it
and therefore there are no eigenvalues to speak of. The RouthHurwitz criterion becomes useless in this case. The kind of stability that we are concerned with here is asymptotic stability, which means that the equilibrium state x e is asymptotically stable if every motion starting at an initial state x will converge
to
x e as time approaches
Popov's
infinity.
shown
in Fig. 733.
The
nonlinearity is described
Fig. 733.
lie
as
shown
Many nonlinear control systems in practice can be modeled by the block diagram and nonlinear characteristic of Figs. 733 and 734. For instance, Fig 735 illustrates several commontype nonlinearities, which are encountered often in control systems that fit the characteristic specified by Fig. 734. Popov's stability theorem is based on the following assumptions:
1.
The
2.
is described by the transfer function which has more poles than zeros, and there are no cancellations of poles and zeros. The nonlinear characteristic is bound by k and k 2 as shown in Fig.
G(s),
734; or
(7123)
Me)
Fig. 734.
Nonlinear characteristics.
N(e)
1%)
/'
/
k,
=0
_z_
D
/
/ /
/
/
/
/
i
=0
/
(b)
7
(c) Saturation
/
k,
=0
zone
common
tems, (a) Ideal relay, (b) Relay with dead zone, (c) Saturation. Id) Saturation with
dead zone.
364
Sec. 7.11
Stability of Nonlinear
Systems
Popov's Criterion
The
365
The theorem
loop system
is
is
closed
asymptotically stable if the Nyquist plot of G(jco) does not intersect or enclose the circle that is described by
*i
~t~
k2
=
_
*."
(7124)
LK^Ki
where x and y denote the real and imaginary coordinates of the G(ja>)plane, respectively.
It
is sufficient
but
is
not
system
is
stable,
if
On
violated,
mean
unstable.
The
Fig. 736.
Im G
G(/oj)plane
ReG
It is interesting to
observe that
if
( 1
y'O)
point,
are with
A:,
0.
x
For
stability,
(7125)
line.
366
Stability of Control
Systems
Chap. 7
Consider that the block diagram shown in Fig. 737 represents a feedback control system that has a saturation element in the forward path. The system could be a position control system using a motor as an actuator. The saturation is in the power amplifier of the motor controller, with representing the gain of the linear portion of the amplifier characteristic.
1
Fig. 737.
The
of the system
is
G(s)
=
s(s
1)0
(7126)
2)
It can be shown that if the saturation characteristic were absent, the closedloop < 6. It is desired to determine the value of so system would be stable for that the nonlinear system is assured of stability. Notice that the Popov criterion is of such a nature that for the saturation nonlinearity, the result of the criterion is inde
<K
pendent of the level of saturation, which is in a way a qualitativeness of the test. The Nyquist locus of GO' CO) is sketched as shown in Fig. 738. For stability of the closedloop system, the G(joi) locus must not intersect the vertical line, which passes
the
Re GU<o)
The frequency
at
9co2+{2
is
_ (0
2 2 )
(7127)
which Re G(Jco)
is
maximum
[9co
2
determined from
2
)] 2 2 2 ) ]
d Re
which gives
G(jco)
_ <a[18 + 4(2  co
~~
dm
(O
(2
 o>
(7128)
= 0. Thus
Max. Re
G(j<x>)
= f
(7129)
is
stable for
K<
which
is
(7130)
Chap. 7
References
367
/'
Im G
ReC
Popov
REFERENCES
RouthHurwitz Criterion
1.
E.
J.
& Co.
2.
6,
Part
II,
Macmillan
N. N. Puri and C. N. Weygandt, "Second Method of Liapunov and Routh's Canonical Form,"/. Franklin Inst., Vol. 76, pp. 365384, Nov. 1963.
G. V.
trol,
3.
S. S.
IEEE
4.
V.
Criterion
and
IEEE
V. Krishnamurthi, "Gain Margin of Conditionally Stable Systems from Routh's Stability Criterion," IEEE Trans. Automatic Control, Vol. AC17, pp. 551552,
Aug. 1972.
368
Stability or Control
Systems
Chap. 7
Nyquist
J.,
7.
C. H. Hoffman,
teristic
I.
8.
C. H. Hoffman,
"How To Check Linear Systems Stability: Roots by Algebra," Control Engineering, pp. 8488, Feb. 1965
Locating the
9.
C. H. Hoffman, "How To Check Linear Systems Stability: III. Locating the Roots Graphically," Control Engineering, pp. 7178, June 1965.
10.
M. R. Stoji6 and D. D. Siuak, "Generalization of Hurwitz, Nyquist, and Mikhailov Stability Criteria," IEEE Trans. Automatic Control, Vol. AC10, pp.
250254, July 1965.
11.
R.
Part I,"
12.
W. Brockett and J. L. Willems, "Frequency Domain Stability Criteria IEEE Trans. Automatic Control, Vol. AC10, pp. 255261, July 1965. W. Brockett and J. L. Willems, "Frequency Domain Stability CriteriaIEEE Trans. Automatic Control, Vol. AC10, pp. 407413, Oct. 1965.
p. 317, April 1966.
R.
Part II,"
13.
14.
IEEE
Trans.
ACU,
T. R. Natesan,
rion,"
"A Supplement
Crite
IEEE
Popov's Criterion
16.
V.
M. Popov, "Absolute Stability of Nonlinear Systems of Automatic Control," Automation and Remote Control, Vol. 22, pp. 961978, Aug. 1961.
17.
Z. V. Rekasuis,
Element,"
18.
"A Stability Criterion for Feedback Systems with One Nonlinear IEEE Trans. Automatic Control, Vol. AC9, pp. 4650, Jan. 1964. "A
"On
Generalization of the
C. A. Desoer,
Popov
Criterion,"
IEEE
Trans. Auto
C. Pincura,
Popov
IEEE
AC12, pp.
Y. S. Cho and K. S. Narendia, "An OffAxis Circle Criterion for the Stability of Feedback Systems with a Monotonic Nonlinearity," IEEE Trans. Automatic Control, Vol. AC13, No. 4, pp. 413416, Aug. 1968.
J.
21.
B.
Moore, "A
IEEE
No.
1,
C. A. Desoer,
Control, Vol.
"An Extension
IEEE
Trans. Automatic
23.
E. Y.
No.
3,
Chap. 7
Problems
369
24.
A. C. Tsoi and H. M. Power, "Equivalent Predictions of the Circle Criterion and an Optimum Quadratic Form for a SecondOrder System," IEEE Trans. Automatic Control, Vol. AC17, pp. 565566, Aug. 1972.
C. E. Zimmerman and G. J. Thaler, "Application of the Popov Criterion to Design of Nonlinear Systems," IEEE Trans. Automatic Control, Vol. AC16, pp. 7679, Feb. 1971.
25.
26.
H. M. Power and A. C. Tsoi, "Improving the Predictions of the Circle Criterion by Combining Quadratic Forms," IEEE Trans. Automatic Control, Vol. AC18,
pp. 6567, Feb. 1973.
PROBLEMS
7.1.
By means of
the
RouthHurwitz
criterion,
determine the
stability
of systems
that have the following characteristic equations. In each case, determine the
(d)
(e)
16s
16
=
that correspond to a stable
7.2.
The
(b) s*
(c)
s3
+ + +
20Ks 3
15
7.3.
given as
G(S)
It is
= s(l+7s)
lie
desired that
left
all
in the
region to the
of the line s
= a.
is obtained, but also that the system has a minimum amount of damping. Extend the RouthHurwitz criterion to this case, and determine the values of K and T required so that there are no roots to the right of the line s = a.
system
7.4.
The loop
is
given by
^^ s(i+
1)
rsxi
h2s)
The parameters K and Tmay be represented in a plane with K as the horizontal axis and T as the vertical axis. Determine the region in which the closedloop
system
7.5.
is
stable.
The openloop
is
given by
GM =
y '
K{s
s 3 {s
370
/ Stability of Control
Systems
Chap. 7
What
7.6.
A controlled process
x2
The
control
is
= Xi 3X2 = 5X\ + u
= ^1*1 + glXl
and gz are real constants. Determine the region in the gz versus ^i t plane in which the overall system is stable.
where g
7.7.
Given a
is
equations
(t)
= Ax(?) + B(?)
(T
1
where
1
o
A=
.0
B =
_1
state feedback so that
3
2
is
implemented by
u(t)
where
= Gx(t) G = [gi gz
gi\ with
g u g 2 and g 3 equal
,
to
real constants.
so that the
overall system
7.8.
= Ax +
1
B, where
on
A=
Consider that state feedback state feedback?
7.9.
2
3_
B =
_1
Is the
may
be implemented.
system stabilizable by
functions, sketch the Nyquist diagrams that correspond to the entire Nyquist path. In each case check the values of N, P, and with respect to the origin in the Gflplane. Determine the values of N, P, and Z with respect to the 1 point, and determine if the closedloop system is stable. Specify in which case it is necessary to sketch only the Nyquist plot for Co = to oo (section 1) on the Nyquist path to investigate the stability of the
closedloop system.
(a)
G(s)H(s)
+ Q J)(1 + Q2s) +
+
Q
5(
(b) G(s)H(s)
25s){l
( \ <CJ
r<\n<\ U(S)H(S)
100(1
]i)(1
+ Q5s) + s)
0.25s)
+ a5j)(1 + Q8s)
(d) G(s)H(s)
= 5(1 + oa7 5 )(
1
Chap. 7
Problems
371
(e)
G(s)H(s)
G(s)H(s)
=
j(1
10
+ 02s)(s 
1)
(f)
ffi+ffi
of the systems.
7.10.
Sketch Nyquist diagrams for the following loop transfer functions. Sketch only the portion that is necessary to determine the stability of the closedloop system.
Determine the
(a)
stability
G(s)H(s)
= =
100
s(s 2
+2s +
50
2)0
4)
1)
(b) G(s)H(s)
s(s
(c)
+
s
2)(s 2
G(s)H(s)
1
 0.2s
7.11.
Figure P71 1 shows the entire Nyquist plots of the loop gains G(s)H(s) of some feedback control systems. It is known that in each case, the zeros of G(s)H(s)
/
Im
oo cj
0
= +
Arrows on diagram
correspond to defined sense of the Nyquist
path
"".^
G//plane
j Re
372
Stability of Control
Systems
Chap. 7
co
= 
+
.
"""
/Im ~~ " ^^
/
(
W *Ju
(1./0)
X /
/?*
(///plane
\

= =
/**'
Re
=+
1
co
/
'
CO= +
(c)
are
all
located in the
left
is,
Z=
the right half of the splane. State the stability of the openloop systems. State
stable ;
if
The
given by
5Ks
+ (2K +
3)s
10
Apply the Nyquist criterion to determine the values of K for a stable closedloop system. Check the answer by means of the RouthHurwitz criterion.
7.13.
The Nyquist
criterion
was
sketching the Nyquist plot of G(s)H(s) that corresponds to the Nyquist path, it is possible to tell whether the system's characteristic equation has roots in the right half of the iplane. (a) Define a new Nyquist path in the jplane that may be used to ensure that all the complex roots of the characteristic equation have damping ratios
of a closedloop system.
By
may
left
7.14.
in Fig. P714.
is
stable.
R(s)
* as)
Figure P714.
Chap. 7
Problems
373
(b)
all lie
to the
of the Re(s)
7.15.
shown
in Fig. P715.
*,(*)
*C.
(s)
R 2 (s)
*~C 2 (s)
Figure P715
The
is
given by
K
(j
IX*
2)
and
(a)
K is
K so
is
asymptotically stable.
(b)
7.16.
Figure P71 6 shows the block diagram of a control system in which the amplifier
u
Amplifier
G(s)
Amplifier
characteristic
G(s)s(l
+0.1
,s)(l
+0.2s)
Figure P716.
374
Stability of Control
Systems
Chap. 7
maximum
value of
will
be absolutely
stable.
Figure P717 shows a control system that has a synchro control transformer as an error transducer. The output of the synchro is proportional to the sine of the
input shaft position. Assuming that the synchro operation
is
limited
to
n/1
<9 < nil, the inputoutput characteristic of the device may be repree
shown
in Fig. P717.
K so
is
absolutely stable.
Y% J
Synchro
control
u
G(s)
transformer
G(s) =
s(\
K
+0.5s)(l +.s)
Synchro
10
characteristic
Figure P717.
8
Root Locus Techniques
8.1
Introduction
it is
a given range. Since the characteristic equation plays an important role in the
linear systems,
an important problem
in linear control
when a certain system parameter examples of Chapter 6 already have illustrated the importance of the root loci in the study of linear control systems. The root locus technique is not confined to the inclusive study of control
equation, or simply, the root loci,
systems.
also be used to
In general, the root locus problem for one variable parameter can be defined by referring to equations of the following form
F(s)
s"
ai s">
+...
a n _ lS
an
where
cients
K is
oo
and
oo.
The
coeffi
b m are assumed to be fixed. These coefficients can be real or complex, although our main interest here is in real coefficients.
!,...,
x ,
...
bm _
t ,
The root
in this chapter.
Although the
loci
when
K varies between
co and
375
376
Chap. 8
oo are generally referred to as the root loci in control system literature, the follow
Root
loci: the
is,
when
A'
assumes positive
values; that
2.
< K<
oo.
Complementary root
loci: the
is,
when
oo <
K<
0.
Root contours:
loci of roots
varies.
The complete
and the
complementary root
8.2
is
the characteristic equation of a linear control system that has the closedloop
C(s) ~~ R(s)
_
1
G(s) G(s)H(s)
,
v
2>
The
G(s)H(s)
(83)
The reader should have a special understanding of the relationship between G(s)H(s) = 0, and the function the characteristic equation, the equation 1 these relationships are very simple to understand, one can G(s)H(s). Although
intri
we
do not contain K, we
get
,
K(s m
s
(S
'
Comparing Eqs.
lished
:
(83)
and
(84)
we
can be estab
G(s)H(s)
s"
aiS"
where G(s)H(s) is known as the loop transfer function of the control system. Since we have mentioned that the root locus technique is not limited to control systems, in general, given Eq. (81), we can regard G(s)H(s) of Eq. (85) as the loop transfer function of an equivalent control system. The control system
is
it
its
characteristic equation.
Later
we
from Eq.
(81) is a
many
Sec. 8.2
377
name, we
K as
We
can
now
define the complete root loci as the loci of the points in the
is
varied between
Now we
satisfied.
oo
(83) is
G(s)H(s)
= KG^H^s)
(86)
where
G^H^s) no longer contains the variable parameter A'. Then Eq. (83) is
written
G^H^s)^^
To
satisfy this equation, the following conditions
(87)
(?,(*)#,(*)
= pL
l)w
co
<
tf
<
oo
(88)
/G,(j)ff,(j)
K>
(89)
K<0
and then
the values
(810)
where k
0,
1, 2,
(all integers).
In practice, the complete root loci are constructed by finding all points in
the splane that satisfy Eqs. (89)
loci are
and
(810),
of
K along
the
(88).
is
The construction of the root loci some of the rules of construction are
arrived at analytically.
is
The
(85)
starting point
must
first
be
G(*\H(*\
W WK + ~ (s+
(s
Z 'X J
(811)
where the poles and zeros of G(s)H(s) are real or complexconjugate numbers. Using Eq. (811), the conditions stated in Eqs. (88), (89), and (810)
become
m
\G (s)H {s)\=i
l 1
= rlT
I*
co<tf<co
(812)
and
/GMH^s) =
'=i
H +z
ls
^ ls+
j=i
Pj
(813)
= We shall
first
(2k
+
is
l)7i
<K<
.
oo
a rational function of
s.
378
Chap. 8
/G^HM = 2 +  S
/s
z,
/s
i=i
j
(814)
=
for
A:
2kn
oo<K<0
and
(814)
It
0, 1, 2, was mentioned
may
construction of the complete root loci in the splane. In other words, Eq. (813)
implies that for any positive value of AT, a point
(e.g., Si) in
the splane
is
on the root loci if the difference between the sums of the angles of the vectors drawn from the zeros and the poles of G(s)H(s) to s, is an odd multiple of 180. Similarly, for negative values of A", Eq. (814) shows that any point on the complementary root loci must satisfy the condition that the difference between the sums of the angles of the vectors drawn from the zeros and the poles to
the point
is
0.
To
root
illustrate the
loci, let
G(s)H(s)
=
s(s
(815)
The locations of the poles and the zero of G(s)H(s) are arbitrarily assumed, as shown in Fig. 8rl. Next, we select an arbitrary point, s in the .yplane and draw vectors directing from the poles and the zero of G(s)H(s) to the point j,. If j,
t ,
splane
[K(s
zi)]/[s(s
U+p
+ p 2)
3 )].
Sec. 8.2
379
is
<K<
oo)
it
G(s)H(s)],
[remember that the root loci must satisfy the following two
.\_?j_IiL jllkl+/2ll*l+/j
L
l*
1)tt
fc
(816)
(813),
2 )
0,
1,2,...
(817)
to be a point
loci
( oo
<K
<
0), it
must
satisfy
/j
(/i.
M+
J?2
/Ji
+
,
;>, )
= Ikn
(818)
for&
= 0, 1, 2, As shown in Fig.
become
tl
ri ,
0,,,
d P2 and
0,,,
vectors measured with the positive real axis as zero reference. Equations (817)
and
(818)
(0 Pl
+e +d
p!
P3 )
= =
(2k
l)n
0<K<oo
(819)
and
0.i
~ (K +
d pi )
2kn
co<K<0
is
(820)
respectively.
If s t is found to satisfy either Eq. (819) or Eq. (820), Eq. (816) determine the value of AT at the point. Rewriting Eq. (816), we have
used to
ff
kilbi
\
+ Pi\\s, + p s + z
i i
I
(8
_ 21
^x
zx
is
z to
= B^
(822)
complementary root
following two steps
depends on whether Sx is on the root loci or the Consequently, given the polezero configuration of G(s)H(s), the construction of the complete root locus diagram involves the
sign of K, of course,
loci.
The
1.
search for
all
the
*,
2.
and (810). The determination of the values of K at points on the root the complementary root loci by use of Eq. (88).
and
From
the basic principles of the root locus plot discussed thus far,
all
it
may
380
Chap. 8
and
(810)
is
a very tedious task. However, aided with the properties of the root
are going to assemble in the next section, the actual sketching of
loci that
we
most cases
is
with some experience on the part of the analyst, the root loci can be sketched by
following through the "rules" of construction.
since the Spirule
the Spirule, can also be used to help plot the root locus diagram. However,
is
can be used
effectively
set
we
already
know
x
when we
general knowledge
we select a trial point and test it in Eqs. (813) not close to any point on the root loci, the search procedure can be frustrating, even with the aid of a Spirule. " Digital and analog computer programs 32 34 can be prepared for the plotting
of the location of this point; then
(814). If the trial point is
and
~ 29
be used
if the poles and zeros of G(s)H(s) are not known a priori. The material presented in this chapter will emphasize the principle of construction of the root loci, since one must obtain a thorough understanding of
before
successfully.
8.3
The following
from the
relation between
be regarded only as an aid to the construction of the root loci and the complementary root loci, as they do not give the exact plots.
K=
Points
Theorem
ofG(s)H(s).
Proof:
81.
The
K=
From
Eq. (812),
m
GMH^s) =
I
^
J=i
= r^
(823)
As
s
G^H^s)
or of G(s)H(s); that
the sign of
is,
approaches
(j =
1, 2,
n). It is
K has
no bearing
in Eq. (823).
Sec. 8.3
381
Example
81
2)(j
3)
K(s
1)
(824)
When K =
(824)
0,
AT (the
= 0, j = 2, and s = 3. These G(s)H (s) if we divide both sides of Eq. Golden Rule) and establish the relationK(s
ship
1
G(s)H(s)
+
1)
1)
s(s
2)C$
+
3)
=
3)
(825)
Thus
G(s)H(s)
s(s
K(s
2)(j
(826)
The
three
points
loci are as
shown
in Fig. 82.
.splane
;cj
K=
K=Q
K=
which
K=
loci
of s(s
3)
K(s
1)
2)
0.
K = oo
Points
82.
Theorem
The
K = oo
zeros of G{s)H(s).
Proof: Referring again to Eq. (823), as K approaches co, the equation approaches zero. This corresponds to i approaching the zeros of G{s)H{s); or "> ^ * {i= s approaching z t '' 1,2, ...,m).
'
Example
82
+ 2)(s + 3) + K(s + 1) = (827) It apparent that when K is very large, the equation can be approximated by K(s+l) = (828) which has the root s = 1. Notice that this is also the zero of G(s)H(s) in Eq. (826). Therefore, Fig. 83 shows the point j = 1 at which K = oo. However, G(s)H(s) in
is
infinity,
382
Chap. 8
splane
/to
A>
2
other K = points at infinity
Two
which
K=
<*>
loci
of s(s
2)
3)
K(s
1)
0.
the total
co,
if
the poles
and zeros
at infinity are
1
K = co points are at j =
and
co.
Number of Branches on
values between
loci
A branch of the complete root loci is the locus of one root when K takes on oo and oo. Since the number of branches of the complete root
must equal the number of roots of the equation, the following theorem
results
Theorem
83.
the greater of n
loci
of Eq. (81)
is
equal to
Example
83
is
three, since
roots,
s, it
The proof of the first statement is selfevident, since, for real coeffimust be real or in complexconjugate pairs.
The reasoning behind the second statement is also simple, since if the poles and zeros of G(s)H(s) are symmetrical to an axis other than the real axis in the jplane, we can regard this axis of symmetry as if it were the real axis of a new complex plane obtained through a linear transformation.
Sec. 8.3
383
Example
84
1)(*
2)
+K=
(830)
Dividing both sides of the equation by the terms that do not contain G(s)H(s)
K leads
to
(831)
=
s(s
K
+
l)(.s
2)
loci
shown
in Fig. 84.
Notice that
loci are
complete root
.splane
K<0
< +
11
)!
I
/
/ K<0
A
Fig. 84.
Axis of
symmetry
Root
loci of s(s
l)(s
2)
+K=
0,
of symmetry.
Example
85
When
is
symmetrical with
respect to a point in the jplane, the complete root loci will also be symmetrical to that point. This is illustrated by the root locus plot of
s(s
2)(.s
+ +/X* +
1
~j)
+K=
(832)
as
shown
in Fig. 85.
384
Chap. 8
splane
Fig. 85.
Root
loci of s(s
2)(s
+y")0
y)
A: == 0,
showing
co)
The
properties of the complete root loci near infinity in the jplane are
important, since
the jplane.
m\ of the
loci will
approach
infinity in
Theorem 85. For large values ofs, the root loci for straight lines or asymptotes with angles given by
ek
.
= (?KDz n m K<
n
1.
m\ 1* and n and m are defined in Eq. (81). ,\n where k = 0, 1, 2, 0, the angles of the asymptotes are For the complementary root loci,
.
.
9,
2kn
m
=
(834)
where k
0, 1, 2,
,\n
m\
According

0,
m\ asymptotes
1, 2, we need
However,
to assign only
values to k.
Sec. 8.3
385
Equations (833) and (834) imply that the asymptotes of the complementary root loci are linear extensions of the asymptotes of the root loci, and vice versa. When the asymptotes of one type of root loci are determined, those of the other type are found without further calculation.
Proof:
b.s"
...
bm .
bm
We
then have
S"
j"
(835)
Carrying out the fraction of the left side of Eq. (835) by the process of long division, and for large s neglecting all but the first two terms, we have
s"
+
1
(a l
 b )s n m
l
^ K = (*)"<is
(836)
or
a
'
'
(837)
The
factor
[1
(a,
b
1
)/sy /l
" m>
in Eq. (837)
sion,
and Eq.
(837)
becomes
+ (
m)s
t"vl/(nm) = (*)
(838)
Again,
if
only the
first
two terms
we
get
'
Now
let s
+ jco,
ai
is
written
o +jco
for
bi m
and
cos
(2k
n
+ l)n m +
cos
sin
(2k n
+ \)n m
2kn m ^ n
(840)
<K<
oo
oo,
1/( "~ m)
I
2kn
n
sm
(841)
for
<
A:<0, and&
real
Equating the
for
0< K<
we
have,
oo,
a
and
^i
~ Ki/<m) cos(2k +
,
\)n
(842)
<a^*'
Solving for AT I/( " m) from Eq. (843),
/(
>sin
(2 *
m
t
+1) *
(843)
we have o
a n
(844)
386
Chap. 8
or
CO
~ tan
2k
n
ZUZi) + V* m (g + n m I \
line in the jplane,
(845)
is
of
form
co
= M{a  a
we have
(846)
where
and a
is
From
M = tan?^iirc n m
0, 1, 2,
. .
.
(847)
\n
m\
1,
and
a, M
b,
m
Note that these properties of the asymptotes are
only.
Similarly,
(848)
from Eq.
(841)
we can show
loci(oo
< K<G),
M = tan ^n m
k
(849)
as in
0, 1, 2,
\n
m\
1,
Eq. (848)
is
obtained for
Eqs. (833) and (834), have been proved. This proof also provided a byproduct, which is the intersect of the asymptotes with the real axis of the
jplane,
and therefore
m\ asymptotes of the Theorem 86. (a) The intersection of the 2\n plete root loci lies on the real axis of the splane. (b) The intersection of the asymptotes is given by
com
~
where a u b u
Proof:
n,
fc
'
g m
'
(850)
and
The proof of
that the complete root loci are symmetrical to the real axis. The proof of (b) is a consequence of Eq. (848). Furthermore, a function G(s)H(s) as in Eq. (85), Eq. (850) may be written as
b_
if
we
define
xZL a A
_ finite poles of G(s)H(s)  finite zeros of G(s)H(s) ~ number of finite poles of G(s)H(s) number of finite zeros of G(s)H(s)
(8
' 51
Sec. 8.3
since
a,
= = b =
t
roots of s"
finite
+
+
a,s"' x
+
l
+
.
a_,s
a.
=
(852)
poles of G(s)H(s)
b,s m
~
sum
of the roots of s m
finite
b^^s
bm
=
(853)
sum of the
zeros of G(s)H(s)
Since the poles and zeros are either real or complexconjugate pairs, the imagi
nary parts always cancel each other. Thus in Eq. (851) the terms in the summations
may
respectively.
It
is
complementary root
Example
86
4)(s 2
2s
2)
K(s
1)
(854)
<*'>*
The polezero
+
is
w VL
shown
+ 2)
From
<8
55 >
configuration of G(s)H(s)
in Fig. 86.
on the construction of the complete root loci described so far, the following information concerning the root loci and the complementary root loci of Eq. (854) is obtained
1
K=
The
:
K=
=
0, s
points
of
2.
= 4, s = 1 +j\, and s = 1 yl. K = co The K = co points on the complete root loci are at the zeros of G(s)H(s): s = 1, s = co, j = co, and s =
G(s)H(s): s
<x>.
3.
4.
5.
of the fourth order, there are four complete root loci. The complete root loci are symmetrical to the real axis. For large values of s, the root loci are asymptotic to straight lines with angles measured from the real axis Root loci {K 0), Eq. (833)
is
:
>
k
k
= = =
l
^
3
60
d,=~ =
92
900 =^ =
180
20O
A:
=0
=
=
1
=^ = o
120
6^^ = =^ =
2
24O
388
Chap. 8
splane
\
loci
\
Fig. 86.
loci of s(s
4)0 2
r
Is
2)
K(s
0.
The asymptotes of the complementary root loci may be obtained by extending the asymptotes of the root
6. loci.
The
<r,
six
loci intersect at
= S finite poles
S finite zeros
of G(s)H(s)
(856)
 (o4i
+yi l yi)(i)
41
_ _a
3
as
shown
in Fig. 86.
Example
87
are
The asymptotes of the complete root loci for several shown in Fig. 87.
different equations
{K >
Theorem
87. (d)
Root
loci:
On a
may
<
Asymptotes of
/CO
splane
iplane
Asymptote of
root locus
_~(Pi + P 2 +P3) ,
4
"l
="Pi/2
/V
I
45
/
/
\
\
\ \
I
/
/
G(s)H(s) =
s(s
K
+p
x
G(s)H(s) =
)
splane
\
'
Asymptotes of complementary
\
/
'
/CO
/"
splane
root loci
^ Asymptotes of
root loci
'(Pi
/ /
/
root loci
\
i\
\
/
\ vX/\
\
1
i
'
/\
/
\
45
+P2+P3)^*7V^
X
\
o,
r
45
c
/
/ /
\
\
/
/
/
\
I
G(s)H(s)=
s
2
l
K
(s+p )(s+p 2 )(s+ p 3 )
Fig. 87.
G(s)H(s)
i 2(i
v >
K(s + zi)
+p
)( J
+ p 2 )( s + p 3
loci.
389
390
Chap. 8
we can state that complementary root loci will be found in sections on the real axis not occupied by the root loci. In all cases the complex poles and zeros ofG{s)H{s) do not affect the existence properties of the root loci on the real axis.
Proof:
1
The proof of
the theorem
t
is
At any point (e.g., s ) on the real axis, the angles of the vectors drawn from the complexconjugate poles and zeros of G(s)H(s) add up to
be zero. Therefore, the only contribution to the angular relations in
Eqs. (813) and (814)
is from the real poles and zeros of G(s)H(s). and zeros of G(s)H(s) that lie to the right of the may contribute to Eqs. (813) and (814), since real poles
2.
Only the
point
St
real poles
3.
and zeros that lie to the left of the point contribute zero degrees. Each real pole of G(s)H(s) to the right of the point j, contributes 180 and each zero to the right of the point contributes 180 to Eqs. (813) and (814).
t
The last observation shows that for s to be a point on the root loci, there must be an odd number of poles and zeros of G(s)H(s) to the right of s u and for Si to be a point on the complementary root loci the total number of poles and
splane
Complementary
root loci
Root
loci
on the
real axis.
Sec. 8.3
391
complete
loci
root loci
Example
88
on the
two
complementary root
loci.
Angles of Departure (from Poles) and the Angles of Arrival (at Zeros) of the Complete Root Loci
{arrival)
ofG(s)H(s) denotes the behavior of the root loci near that pole (zero). For the root
(K >
0) these angles
For
instance, in
desired to deter
<*><
AT
Fig. 89.
Complete root
loci of s(s
3)0 2
2s
2)
+ K=
to illus
392
Chap. 8
which the root locus leaves the pole at 1 +jl. Notice that is measured with respect to the real axis. Let us assume that s is a point on the root locus leaving the pole at 1 + j 1 and is very near the pole. Then s must satisfy Eq. (813). Thus
at
the
unknown
l
angle 9 2
/GjsjHjSj)
Since s,
is
(0,
62
+ + + j\,
3
0)
(2k
1)180
(857)
drawn becomes
(858)
92
90
26.6)
(2k
We can
obtained for
all
92
431.6
of G(s)H(s)
determined,
which
is
the
same
as 71.6.
is
same point
differs
from
this
now
be used.
The points where the complete root loci intersect the imaginary axis of the and the corresponding values of K, may be determined by means of the RouthHurwitz criterion. For complex situations with multiple intersections, the critical values of K and co may be more easily determined approximately from the Bode diagram.
splane,
Example
89
loci
of the equation
2)
3)(s 2
+2s +
+K=
(859)
The root loci intersect the /(Waxis at two conjugate points. Applying the RouthHurwitz criterion to Eq. (859), we have, by solving the auxiliary equation, Kc = 8.16 and co c = 1.095 rad/sec.
are
in Fig. 89.
drawn
Breakaway Points (Saddle Points) on the Complete Root Loci Breakaway points or saddle points on the root loci of an equation correspond to multipleorder roots of the equation. Figure 8 10(a) illustrates a case in which two branches of the root loci meet at the breakaway point on the real axis and then depart from the axis in opposite directions. In this case the breakaway point represents a double root of the equation to which the root loci belong. Figure 8 10(b) shows another common situation where a breakaway
point
may
occur.
may involve more than two root loci. Figure where the breakaway point represents a fourthorder
A root locus diagram can, of course, have more than one breakaway point. Moreover, the breakaway points need not always be on the real axis. However, because of the conjugate symmetry of the root loci, the breakaway points must either be real or in complexconjugate pairs.
Sec. 8.3
393
splane
splane
K=Q
K
point
<
e
Breakaway
point
Breakaway
(a)
(b)
Fig. 810.
/'co
splane
breakaway point.
in Fig. 810(a)
Because of the symmetry of the root loci, it is easy to see that the root loci and (b) break away at 180 apart, whereas in Fig. 811 the four < oo) root loci depart with angles 90 apart. In general, if n root loci ( oo <
394
Chap. 8
Several graphical and analytical methods are available for the determination
of the location of the breakaway points. be the most general are presented below.
Two
analytical
Method
1.
KG
Theorem 88. The breakaway points on the complete root (s)H (s) = must satisfy
x
loci
of 1
dG^sJH^s)
Proof:
(86Q)
KP(s)
(861)
Then Eq.
(85)
may
be written
GW" =
(5)
w
AK, Eq.
(861)
* (8 62)
If
we
is
varied by an increment
becomes
(863)
Q(s)
(K +AK)P(s)
=
we have
KP(s),
AKF(s)
=
S)
(865)
where
F(S)
Since the denominator of F(s) a breakaway point of n
is
Q(s)+ KP(s)
same
(*<>V
the
which corresponds to
where
=
t
(F*i? = (0r
<
8  67 >
is
a constant.
1+^ =
from which we obtain
Taking the have
limit
(868)
on both
sides
of the
last
equation as
AK approaches
zero,
we
lim
Ajco
* =
As
^=
as
it is
(870) v /
We have shown that at a breakaway point on the root loci, dKjds is zero.*
*The quantity (dsjs)l(dKIK)
is
proved that
at the break
away
Sec. 8.3
395
Now,
or
must
also satisfy
(871)
+ KG^H^s) =
it is
is
equivalent to
ds
It is important to point out that the condition for the breakaway point given by Eq. (873) is necessary but not sufficient. In other words, all breakaway points must satisfy Eq. (873), but not all solutions of Eq. (873) are breakaway points. To be a breakaway point, the solution of Eq. (873) must also satisfy Eq. (871);
or,
Eq. (873) must be a factor ofEq. (871) for some real K. In general, the following conclusions can be made with regard to the solu
All real solutions of Eq. (873) are breakaway points on the root
loci
( oo
<K<
if
of the splane
are
is
The complexconjugate
points only
solutions
of Eq.
(873)
breakaway
cause difficulty in the effective use of Eq. (873), since the other properties of the root loci are usually sufficient to provide information
and
error.
Example 810
Consider that
equation
it is
s(s
2)
K(s
4)
(874)
Based on some of the theorems on root loci, the root loci of Eq. (874) are easily shown in Fig. 812. It can be proven that the complex part of the loci is described by a circle. The two breakaway points are all on the real axis, one between and 2 and the other between 4 and co.
sketched, as
of Eq. (874) by
s(s
2) (the
C,(5)ff,(5)
= g^ +
2(s
dG&Wiis)
ds
= *fr +
s2
2)
s 2 (s
or
Ss
+ l)(s + 4) = + 2) =
2
(877)
loci are at s
we
396
Chap. 8
splane
Fig. 812.
Root
loci
of s(s
2)
K(s
4)
0.
Note also
that the
all
on
Example
2s
+ K(s +
K(s
s2
2)
=
(878)
(878)
by s
2s
2,
2)
+2s +
(879)
Based on the poles and zeros of G(s)H(s), the complete root loci of Eq. (878) shown in Fig. 813. The diagram shows that both the root loci and the complementary root loci possess a breakaway point. These breakaway points are determined from
are sketched as
<iWgi()
ds
...
d
ds s 2
s
2
(s
+
2s
2)
2
(880)
2)
Sec. 8.3
397
/CO
xplane
Fig. 813.
Root
loci of s 2
2s
K(s
2)
0.
or
s2
4s
+ =
= =
is
(881)
Upon
and
loci,
0.586
3.414
whereas s
812
0.586
is
Example
+ 4)(s 2 + 4s +
20)
+K= K =
20)
(882)
Dividing both sides of Eq. (882) by the terms that do not contain K,
1
we have
(883)
G(s)H(s)
s(s
4)(s 2
+4s +
Since the poles of G(s)H(s) are symmetrical about the axes a 2 and ca in the iplane, the complete root loci of the equation are also symmetrical with respect to these two axes.
=
get
to
s,
we
2
ddJsW^s)
ds
=
+
4s 3
[s(s
(8 " 84)
or
s3
6s 2
18s
20
(885)
398
Chap. 8
xplane
Fig. 814.
Complete root
loci of s(s
4)0 2
4s
20)
+K
=
0.
Because of the symmetry of the poles of G(s)H(s), one of the breakaway points is determined to be at s = 2. The other two breakaway points are found by solving Eq. (885) using this information; they are s = 2 +j2.45 and s = 2
easily
j2A5.
Example
813
In this example
we
shall
show
do not
necessarily represent
loci of the
2s
+ 2) + K =
loci
(886)
are
shown
breakaway point
have any
+ KGMH^s) =
+ s(s 2 +
K
2s
+ 2)
(887)
Sec. 8.3
399
Fig. 815.
Complete root
loci
of s(s*
2s
2)
+K=
0.
(860),
we have
1
ds
ds s{s 2
3s 2
+25 +
2
=
2)
(888)
which gives
4s
(889)
The roots of Eq. (889) are s = 0.677 +./0.471 and* = 0.677 yO.471. These two roots do not represent breakaway points on the root loci, since they do not satisfy Eq. (886) for any real values of K. Another way of stating this is that Eq. (889) is not a
factor of Eq. (886) for any real K.
Method
loci
2.
An
was introduced by Remec. 17 The method is derived from the theory of equations, 11 and the proofs of its necessary and sufficient condioo)
( co
<K<
400
Chap. 8
The breakawaypoint algorithm using a tabulation Routh tabulation for stability study is described below:
1.
Let the equation for which the root loci are desired be written as
F(s)
=A
.
s"
A,si ,
A n _ lS
+
Am
(890)
2.
A n _ A are constant coefficients and the variable where A ,A parameter K is considered to be contained in the coefficients. Obtain the function F'(s) which is the derivative of F(s) with respect to s. Let F'(s) be of the form
t ,
.
.
B. 2 s
,
B_,
(891)
etc.
3.
Arrange the
coefficients of F(s)
and
F'(s) in
Ap
Bo
4.
Al Si
A2
... ...
An 1
U_i
An
B2
Form
numbers obtained by the indicated not a Routh tabulation, although the crossmultiplication operation is the same as in the Routh tabulation. The illustrated tabulation is shown for a fourthorder
the following array of
operations. Notice that this
is
is,
for n
4.
s*
s*
,
Ap
Ai
Bi
BoAi
Bo
A2 B2
At
Bo
BiAo
n _ BpAi BjAg
Bo
Bi
r _ B A 3 B 3 Ap
Bo
r _
B, Bq A* Bi
_
_
Am
Ho
Cp
j3
2 s
B2
n Do
BqC\
To
BiCo
n
'
BqCi
B 2 Cp
To
n Ul
_ BpC B
3
To
n3
_B ~~
Ei
Bo
OCp
ft
ft
S2
s2
E =
DqBi
Dp
DiBp
= D B2  D 2 Bq
Dp
Di
= D B3  D 3B
Dp
=0
Dp
D2 F2 =0
,i
,i
j>
F = D Ei ~ DjE G = F Di D Fl
<>
Fi
= DoEi ~ DlE
Fo
Gi= F D * F* D 'o
Ft
F
^oG ff0==
l
sp
F Go
1
1.
The
, ri) terms assigned to each s 1 (J 0, 1, 2, tion are used for reference purposes.
.
row of the
tabula
Sec. 8.3
401
2.
The
1.
s' terms repeat for three consecutive rows for/ There is only one s" row and one s" row.
(n
1),
3.
If
we
same
row of the
If F(s) has multipleorder roots, which means that the root loci will have breakaway points, a row of the tabulation shown above will contain all zero
The multipleorder
roots,
which are the breakaway points, are obtained by by using the row of coefficients just preceding the
row of zeros.
Let us use the following numerical examples to illustrate the application
of the tabulation method of finding the breakaway points.
Example 814
F(s)
(s
1) C$
2)
s*
6i 3
13s 2
12s
(892)
We have stated the problem in such a way that the equation is known to have two double roots at s = 1 and s = 2. We are merely going to demonstrate the properties
s,
we have
(893)
= 4s +
l&s
26s
12
=
in the following
is
made
13
manner:
6 18
12
4
(4)(6)
s3
S3
4 4
_i
4
(D(18)
(4X13)
4
18
26
(D(26) _
13
12
(4X12)
0X12)
4 26
4
12
S2
S2
_3
18
i
12
S2
jl
6 _i
_3
"i
a row of zero elements in the tabulation before the tabulation process equation F(s) has multipleorder roots. The equation that needs to be solved is formed with the coefficients taken from the row just above the row of zeros. The order of the equation is given by the power of s in the reference column. Therefore, we have
Since there
is
is
completed,
\s 2
or
s2
 Is +
3s
(894)
The
= 1
and
+2= s = 2,
(895)
multiple order.
402
Chap. 8
Example
815
Let us consider a root locus problem, that is, the determination of the breakaway points on the root loci. The equation, Eq. (874), considered in
Example 810
will
is
rewritten
(896)
F(s)
=j +
2
(2
K)s
+ 4K =
Then
F'C?)
2s
(2
+ K) =
(897)
The following
tabulation
is
made
1
J2
2 2
2
sl si
+K +K
4K
4K
+K
2 2
2
+K
< 2 +4
*> 2
to contain
all
we
4K (2
or
4
JQ 2
(898)
K +
2
12^:
4=
(899)
Solving for
and
K= 11.656
When
contain
in the
all
two
values, the s
row of
row preceding
The equation
2s
thus formed
Now substituting K =
0.344 and
(8100)
(8100) in turn,
we find
the
loci at
= =
1.172
6.828
K = 0.344
and
s
It is
K= 11.656
apparent that these answers agree with those obtained in Example 810. Furthermore, a byproduct of the tabulation method is that the values of K at the breakaway points are also determined. In fact, the values of K are always determined first before the breakaway points are found.
Example 816
Now
(882)
written as
K= (8101) The root loci of Eq. (8101) have three breakaway points at s = 2, 2 + J2.45, and 2 y'2.45. Let us now determine the breakaway points by the tabulation method.
F(s)
Ss 3
= s* +
+36s 2 +80s +
Sec. 8.3
403
We
have
F'( s )
= 4s + 24s +
3
12s
18s
+ 80 =
+
20
(8102)
6s 1
(8103)
is
made
36 80 20
6 18 6
18
2
1
60
18
isT
#
20
6 2 18
24
40
20
 K40
6
24
#40
10 #40
20
#40
elements
We would like to interrupt the tabulation at this point to remark that the
in the first
group can be made zero by setting K = 100. Therefore, F(s) = 100. The breakaway points are found from the has multipleorder roots when
row of the
s1
equation
6s 2
24s
+
2
(K
 40) =
+
10
K=
100
(8104)
or
s
4s
=
s
(8105)
= 2 +y'2.45 and which are the solutions of Eq. (8105). In order to complete the tabulation, we must now consider that
Therefore, the two breakaway points occur at s
the coefficients in the subsequent
fact, since
= 2 y'2.45,
100, so that
K^
finite. In 100 are already determined, should there be any additional multipleorder roots, they would have to occur at different values of K than 100. Therefore, in the remaining three rows of tabulation it is implied that
row
(the second
row of
K=
K ^
100.
Resuming
the tabulation,
we have
12
10
_ #40
6
#40 #40 20
K^ 100
K=
KNow
the only
64
row
that can be
is
all
zero
is
the s row,
64. Therefore,
found by substituting
10
K=
^)+(
s
40
')=
(8106)
which gives
= 2
to factor out the
An
factor 10
alternative
(#
common
the tabulation
is
as follows
404
Chap. 8
12
1
K40
2
K64
Therefore, the
same
results,
K=
64 and s
= 2,
show
are obtained.
Example
817
In this example
813, does not
we
shall
diagram
in Fig. 815,
points.
Equation (886)
is
written
F(s)
s3
+ 2s +
2
2s
+K=
2
(8107) (8108)
Then
F'(s)
3s 2
4s
The following
tabulation
is
made
K
2
K
2
K%
1
 21K
K
*(*)(')
It is
last
we
set
2\(k)(i^)=0
which
is
(8109)
simplified to
811s: 2
real values of
(8110)
is that F{s) does not have any multipleorder roots, or the root have any breakaway points.
do not
The condition
sary but not
breakaway point
is
neces
sufficient.
The method
breakaway points as roots of Eq. (860). For higherorder systems with a large number of breakaway points, the amount of work involved in solving a higherorder equation in Eq. (860) may be
excessive.
2.
The tabulation method gives a necessary and sufficient condition for the breakaway points. In general, the procedure still involves
the solving of the roots of an equation, but the order of the equation
may be
method
lower than that involved in the first method. The tabulation at the breakaway points. also gives the values of
Sec. 8.3
405
These methods also represent ways of solving for the multipleorder root
of an equation.
Calculation of
K on
Once the root loci have been constructed, the values of K at any point s^ on the loci can be determined by use of the denning equation of Eq. (88); that
is,
r
If G,(j)#i(j) is of the
iwwi
1),
<!M1,)
is
form shown
in Eq. (81
Eq. (812)
written
1*1
ni*i+/j ^
(8112)
or
li
product of lengths of vectors drawn from the poles of G^H^s) to s product of lengths of vectors drawn from the zeros of dls^H^s) to s t
(8113)
Usually,
is
method
if the root loci are already drawn accurately, the graphical more convenient. For example, the root loci of the equation
are
shown
in Fig. 816.
(8114)
is
given by
(8115)
*=^
t
.
where A and B are the lengths of the vectors drawn from the poles of G(s)H(s) = K{s + 2)/(s 2 + 2s + 2) to the point j, and C is the length of the vector drawn from the zero of G{s)H(s) to s In the illustrated case s is on the root
t
loci,
so
K is positive.
sign.
If j, is
loci,
K should
have a negative
The value of at the point where the root loci intersect the imaginary axis can also be found by the method just described. However, the RouthHurwitz
criterion usually represents a
more
direct
method of computing
this critical
value of K.
The eleven rules on the construction of root locus diagrams described above should be regarded only as important properties of the root loci. Remember that earlier it was pointed out [Eq. (811)] that the usefulness of most of these rules of construction depends on first writing Eq. (81) in the form
(s
+ Pl )(s +p
)...(s+
p)
K(s
z t )(s
z2 )
(s
zm )
(8116)
Then, except for extremely complex cases, these rules are usually adequate for the analyst to make a reasonably accurate sketch of the root loci just short of plotting them point by point. In complicated situations, one has to rely on a computer as a more practical means of constructing the root loci.
406
Chap. 8
.splane
method of finding
the values of
serves as an illustration
Example
818
loci
(8117)
constructed. Using
the rules of construction, the following properties of the root loci are determined:
1.
j1,
The K = points on the complete root loci are at s = 0, 5, 6, 1 + and 1 j 1 Notice that these points are the poles of G(s)H(s), where
.
G(s)H(s)
2.
=
s(s
K(s
3)
5)(s
6)(s 2
+2s +
(8118)
2)
3, oo, oo,
3. 4.
co points on the complete root loci are at s which are the zeros of G(s)H(s). There are five separate branches on the complete root loci. The complete root loci are symmetrical with respect to the
The
oo, oo,
K=
real axis of
the splane.
5.
The
angles of the asymptotes of the root loci at infinity are given by [Eq.
(833)]
6k
for
= SJ3 = Q!LTE n m 5 1
Thus the four root
o<*<co
approach
(8119)
infinity in the
= 0,
1, 2, 3.
loci that
jplane as
K approaches
angles
Sec. 8.3
407
of 45, 45, 135, and 135, respectively. The angles of the asymptotes of the complementary root loci at infinity are given by [Eq. (834)]
ek
Therefore, as
2kn
2kn
1
co
<K<0
(8120)
K approaches
approach
6.
infinity
co, four complementary root loci should along asymptotes with angles of 0, 90, 180, and 270.
is
The
(8121)
The
results
from these
;w
splane
n
+K
K=
6
K=
AT= +
oo\ y'45A
3' \~
2
K=
_1
= X
AT>
\\ K 2.5 \
/l
5)(s
6)(s 2
2s
2)
K(s
3)
0.
408
Chap. 8
In general, the rules on the asymptotes do not indicate on which side of the asymptote the root locus will lie. Therefore, the asymptotes indi
7.
more than the behavior of the root loci as s < oo. The complete root loci portions can be correctly sketched as shown in Fig. 817 only if the entire solution to the root loci problem is known. Complete root loci on the real axis There are root loci (0 K < oo)
cate nothing
:
<
on the
and
loci
=
s
3,
s ==
( oo
between
< K< = 3
and
co
splane
Complementary
root loci
K=
x
Fig. 818.
+
8.
2)
loci
on the
5)0
6)0 2
2s
the pole at
Angles of departure: The angle of departure, 9, of the root locus leaving 1 + / 1 is determined using Eq. (813). If s is a point on the
t
1 +jl, and
St is
very close to
1
+jl
as
shown
in
+3 
(/j,
Is,
+yl +
/s t
1
+5 +
=
/.
sv
+
iplane
;1)
(Ik
1)180
(8122)
Fig. 819.
(s
6)0 2
loci
of s(s
5)
Sec. 8.3
409
or
26.6
(135
90
14
11.4
+ 9) s;
(2k
1)180
(8123)
for
A:
= 0,
1, 2,
Therefore,
9 =* 43.8
(8124)
Similarly, Eq. (814) is used to determine the angle of arrival of the complementary root locus arriving at the point 1 +jl. If this angle is
designated as
9',
it is
is,
9'
9.
180
43.8
136.2
is
(8125)
The
determined by
rewritten
(8126)
13s*
54s 3
is
82s 2
(60
+ K)s + 3K =
13
54 82
60
+K
3K
47.7
65.6 65.6
.s
60
0.769ST
0.212A:
3K
3940 105A:0.163*:2
0.212K
3K
For Eq. same
sign.
no roots in the right half of the splane, the column of the Routh tabulation should be of the Therefore, the following inequalities must be satisfied:
(8126) to have
3940
or or
K< K<
309
35
K>0
Hence
if
of Eq. (8126) will stay in the left half of the jplane and 35, which means that the root loci of Eq. (8126) cross the imaginary axis when K = 35 and K = 0. The coordinate at the crossover point on the imaginary axis that corresponds to K = 35 is determined from the auxiliary equation
all
the roots
K lies between
A(s)
Substituting
(65.6
 0.212A> + 3K =
2
(8130)
K=
we have
105
(8131)
which
yields s
/1.34
10.
trial
loci, and the point lies between the two poles of G(s)H(s) at s = 5 and 6. In this case, since there is only one breakaway point for this fifthorder system, the value of at the point is obtained from the s row of the breakawaypoint tabulation, which would contain a total of 14 rows if carried out. In this case
410
Chap. 8
it is
since
we know
s3
error,
Applying dK/ds
= +
13.55*
142s 1
123s
45
= =
5.53.
After a few trialanderror calculations, the root of the last equation that
is
found to be
From
is
the information obtained in these 10 steps, the complete root locus diagram
In this section we have described 1 1 important properties of the root loci. These properties have been regarded as rules when they are used in aiding the
Fig.
820.
Complete root
loci
of s(s
5)(s
6)0 2
2s
2)
K(s
3)
0.
Sec. 8.3
Of course,
minor propit is
of the root loci which are not mentioned here. However, in general,
1 1
found that these sketch of the complete root For easy reference, the
them.
Table
81.
Table 81
K=
points
The
K=
points
the poles of
infinity.)
2.
on the complete root loci are G(s)H (s). (The poles include those
at
at
K=
oo
points
The
K = oo
points
loci are
3.
Number
root loci
of separate
The
loci is
4.
Symmetry of root
loci
loci
Asymptotes of root
loci as i
For
large values of
s,
(K
>
0) are
>
oo
and
for the
complementary root
a Ok
loci
(K
<
0)
2kn
where
6.
A:
0,\,2,...,\n
m\
\.
Intersection of the
(a)
The
on
on
asymptotes
(centroids)
(b)
The point of
the real axis
is
given by (for
all
values of
K)
2 real parts of
a
7.
_ 2 real parts
of
poles of G(s)H(s)
zeros of G(s)H(s)
Root
loci
on
the
real axis
a given section on the real axis in the splane, in the section only if the total number of real poles and real zeros of G(s)H(s) to the right of the section is odd. If the
root loci are found for
On
K>
total
number of real poles and zeros to the right of a given section is even, complementary root loci
(^< 0)
8.
Angles of departure
and
arrival
The angle of departure of the root locus (K 0) from a pole or the angle of arrival at a zero of
G(s)H(s) can be determined by assuming a point si that is on the root locus associated with the pole, or
zero,
>
and which
is
412
Chap. 8
and applying
/GfriV/fri)
the equation
 1=1
=
(2k
Ai
z,
 2 =
J
Ai
l
+ P;
1, 2,
...
\)n
0,
The
determined from
z,,/s
,
9.
Intersection of the
of the
may
imaginary axis
criterion.
The Bode
10.
plot of G(s)H(s)
may
also be used.
loci
Breakaway points
(saddle points)
The breakaway
or dG(s)H(s)/ds
points
0,
These are necessary conditions only. Alternatively, the breakaway points are determined from a tabulation using the coefficients of the characteristic equations F(s) = and F'(s) = 0. The conditions are necessary and sufficient.
0.
1 1
Calculation of the
values of
The
absolute value of
is
1
K at any point s
on the corn
K on the
root loci
\G(si)H{ Sl )\
product of lengths of vectors drawn
s\
8.4
Although the root locus diagram is primarily intended for the portrayal of the trajectory of roots of a polynomial when a parameter, K, varies, the technique
may
when
all
The
principle
is
best illustrated
by an example. Con
F(s)
s3
3s 2
45
20
=
first
(8132)
To
convert Eq.
Since Eq. (8132) does not have a variable parameter K, the step of conversion is generally not unique. In other words,
(8132) into the
(84).
form of Eq.
we can regard any one of the four coefficients of the polynomial as the parameter number of these coefficients as K will
of the root
loci.
In this case
it is
not
difficult to
see that
it is
more
3s 2
Thus Eq.
(8132) leads to
Sec. 8.4
413
+
which
is
A* + 3) with K = 4.
(8133)
Furthermore,
it is
apparent that
must also satisfy Eq. (8133). Now the problem of solving Eq. (8132) becomes that of a root locus problem, based on the polezero
the roots of Eq. (8132)
configuration of
_ bWW  K + s\s +
G(s)H(s)
(s
5)
(8134)
3)
The desired roots are then found by setting K = 4. From a logistic standpoint, we have embedded a specific problem in a more general one by first solving for
the solution to the general problem.
The root locus diagram based on the function G(s)H(s) of Eq. (8134) for K is shown in Fig. 821. When K = 4, the real root of Eq. (8132) lies between 3 and 5, while the other two roots are complex with positive real
positive
splane
K 5
=
K=
*
X
K=
K=
Fig. 821.
4s
20
3s 2
0.
414
Chap. 8
A few trialanderror steps 3.495. Thus the complex roots s = 0.2475 ./2.381.
parts.
are found to be at s
0.2475
+ y2.381
and
Example
819
5s 3
2s 2
+s+
10
=
first
(8135)
Since this
its
is
a fourthorder polynomial,
it is
roots will be
more
difficult
by the
first
than that of the last example. Let us two terms of the equation; we have
2(s 2
s*(s
+ 0.5j + + 5)
5)
"
(8_136)
or G(s)H(s)
^afr+y
5>
K=2
(8137)
The root locus diagram based on the poles and zeros of G(s)H(s) is constructed as 0. However, from this root locus diagram it is not shown in Fig. 822 for oo > K
>
splane
Fig. 822.
5s 3
2s 2
+s+
= 0.
Sec. 8.4
415
clear
(8135) by
the
first
+ 10 + 5s +
+ 10) + 5s +
=
2)
(8138)
The root
locus plot of
G(s)H(s)
K(s
s 2 (s 2
2)
K=
(8139)
with co
The purpose of constructing the second root two root locus diagrams, since the roots of Eq. (8135) must be found at the same points on both diagrams. Comparing the two root locus diagrams, we may conclude that if Eq. (8135) has real roots, there must be two, and they must lie between 0.44 and 4.56 on the real axis. Once this range has been established, the real roots can be found by trial and error from
is
> K>
shown
in Fig. 823.
locus diagram
is
Eq. (8135).
(8135) by (s
One
real root
is
found to be
3.425.S 2
at s
= +
1.575),
we have
s
3
the remainder,
3.394.T
6.346
=
right of s
(8140)
Now we
repeat the procedure to find the roots of Eq. (8140). Since one of the
we can reason
that
it
must be to the
4.56 on
/CJ
splane
Fig. 823.
Root
10
2s 2
+s+
5s 3
0.
416
Chap. 8
this
3.3940?
s 2 (s
1.87)
3.425)
G(s)H(s)
K(s
s 2 (s
 1.87) + 3.425)
3.394
(8141)
loci
co
< K<
of Eq. (8140) with the G(s)H(s) of Eq. (8141) are 0. Investigation of this complementary root locus
'
/w
splane
K<0
K<0
Fig. 824.
3.425.S 2
3.3945
6.346
0.
diagram and the root loci of Fig. 824 reveals that the real root must lie between 3.425 and 4.56. To find the real root of Eq. (8140), which is at s = 4.493, is a simple matter. The two complex roots are subsequently found to be at 5 = 0.534 + j 1.057 and s = 0.534 / 1.057. Thus the roots of Eq. (8135) are now all determined:
5 s
s
.y
= = = =
1.575
4.493
0.534 0.534
+/ 1.057
y 1.057
In summarizing, we have used two examples to illustrate the application of the root locus method to the solution of the roots of a highorder equation,
F(s)
(8142)
Sec. 8.5
Some
a polynomial in
417
where F(s)
is
s.
is
form
+ GO) 
Then, G(s)H(s) = Q(s)/P(s), or G(s)H(s) = P(s)/Q(s). The selection of the polynomials P(s) and Q{s) from F(s) is more or less arbitrary, although in general the orders of P(s) and Q{s) should be close so that the root locus problem is made
many circumstances it may be desirable to use more than one choice of division of F{s) into P{s) and Q(s). This will usually provide additional information on the location of some of the roots of the equation, so that the trialanderror procedure can be simplified. Example 819 illustrates
as simple as possible. In
how
this is
done.
In essence, the
method of root
is
that of utilizing
some
of the roots. The basic method of finding the exact roots is still cutandtry. For highorder equations, and for equations only with complex roots, the root
locus
still
method of root
finding
may
is
8.5
aspects of the root locus techniques is that for most consystems with moderate complexity, the analyst or designer may conduct a quick study of the system in the splane by making a sketch of the root loci using some or all of the rules of construction. In general, it is not necessary to
trol
make an exact plot of the root loci. Therefore, time may be saved by skipping some of the rules, and the sketching of the root locus diagram becomes an art
that depends to
In this section
loci
some extent on the experience of the analyst. we shall present some of the important properties of the root
in the
and Zeros
In Chapter 6 the effects of the derivative and integral control were illustrated
From
we may
when
may
tion G(s)H(s) in the left half of the splane has the effect of pushing the original
it is
make
a precise
we can
by several
examples.
Let us consider the function
G(s)H(s)
K
,
,
>
(8144)
The zeros of
825(a).
These root
G(s)H(s) are represented by the root locus diagram of Fig. loci are constructed based on the poles of G(s)H(s) at s
OO
splane
splane
K=Q
l
K=
a
2
<
K
i
(a)
(b)
(c)
(d)
Fig. 825.
that
show
G(s)H(s).
418
Sec. 8.5
Some
419
and
= a. Now
let
us introduce a pole at
b so that
G(s)H(s)
=
s(s
~
K
a)(s
b)
b>
(8 1 45)
Figure 825(b) shows that the additional pole causes the complex part of the
The angles of the asympchanged from 90 to 60. The breakaway point is also moved to the right. For instance, if a = and b = 2, the breakaway point is moved from 0.5 to 0.422 on the real axis. If G(s)H(s) represents the loop transfer funcroot loci to bend toward the right half of the splane.
totes are
l
tion of a feedback control system, the system with the root loci in Fig. 825(b)
may become
unstable
if
the value of
loci
K exceeds
of Fig. 825(a) is always stable. Figure 825(c) shows the root loci when another pole is added to G(s)H{s) at s = c.
The system
is
now
loci are
moved
45.
The angles of
two
loci are
even more restricted. Figure 825(d) illustrates that the addition of a pair of complexconjugate poles to the original twopole configuration will result in
we may draw a general conclusion that the addition of poles to the function G(s)H(s) has the effect of moving the root loci toward the right half of the .yplane.
a similar effect. Therefore,
moving
Addition of zeros. Adding zeros to the function G(s)H(s) has the effect of the root loci toward the left half of the jplane. For instance, Fig. 826(a)
/CJ
OO
i
/CO
splane
6 =
<
is
5plane
K = ^
/
/
/
o
\
\
\
\
\
\
\ a/2
1
a:
= o
*
T
/
*
)
tf
K=0

K=
a
K=
o
a/2
~b
/
'
'
K =
b = <*>^
(a)
%
i
(b)
Fig. 826.
G(.s)H(s).
that
show
420
Chap. 8
71
JCO
7 K/
f
<
splane
4
1
J
AT
K=Q K =
b
u
\
\
A'
\ \
w A
(c)
\\
\
that
show
at 5
= b
is
>
if
and form a
circle.
Therefore,
is improved by the addition of the zero. Figure 826(b) illustrates that a similar effect will result if a pair of complexconjugate zeros is added to the function of Eq. (8144). Figure 826(c) shows the root locus diagram when a zero at v = c is added to the transfer function of Eq. (8145).
Effects of
It was mentioned earlier that the construction of the root locus diagram depends greatly on the understanding of the principle of the technique rather
than just the rigid rules of construction. In this section we show that in all cases the study of the effects of the movement of the poles and zeros of G{s)H(s) on the root loci is an important and useful subject. Again, the best way to
illustrate the subject is to
II
o
c
"E.
u,
o
II
:
e c
:=;
E
.a S
">
+
^ "
II II
f}
II
"
O ,
o , aS O Q
^
S3
N
M
as a> 00
*,
II
o
Q
.fa tt
421
o
II
422
Sec. 8.5
Some
423
/CJ
5plane
K=
K=
(e)a=
Example 820
^(s
which
is
a)
K(s
1
b)
(8146)
easily
G(s)H(s)
= 0,
with
(8147)
G(s)H(s)
Let us set b
=
*^
=
= 1 and investigate the root loci of Eq. (8146) for several values of a. Figure 827(a) illustrates the root loci of Eq. (8146) with a = 10 and b 1. The two breakaway points are found at s = 2.5 and 6.5. It can be shown that for arbitrary a the nonzero breakaway points are given by
a
1
^Vo 2
10a
(8148)
When
at j
= 9, Eq. (8148) indicates that the breakaway points converge to one point 3, and the root locus diagram becomes that of Fig. 827(b). It is interesting to note that a change of the pole from 10 to 9 equals a considerable change to the root loci. For values of a less than 9, the values of s as given by Eq. (8148) no longer satisfy the equation in Eq. (8146), which means that there are no finite, nonzero, breakaway points. Figure 827(c) illustrates this case with a = 8. As the pole at s a is
a

424
Chap. 8
farther to the right, the complex portion of the root loci is pushed farther toward the righthalf plane. When a = b, the pole ats = a and the zero at b cancel each other, and the root loci degenerate into a secondorder one and lie on the imaginary axis. These two cases are shown in Fig. 827(d) and (e), respectively.
moved
Example
821
2s
1
which
is
+ +
a)
K(s
G(s)H(s)
+ 2) = = 0, with +
2
(8149)
(8150) l a) The objective is to study the complete root loci (co < K < co) for various values of a{> 0). As a start, let a = so that the poles of G(s)H(s) are at s = 0, 1, and 1.
s{
G(s)H(s)
this
By
setting dG(s)H(s),lds
breakaway points are found at * = 0.38, 1, and  2.618. As the value of a is increased from unity, the two double poles of G(s)H{s) at s = 1 will move vertically up and down. The sketch of the root loci is governed mainly by the knowledge of the breakaway points. We can show that dG(s)H(s)lds
leads to
3
.y
4s 2
As
(8151)
As the value of a increases, the breakaway points at s = 0.38 and s = 2.618 move to the left, whereas the breakaway point at s = 1 moves toward the right.
Figure 828(b) shows the complete root
loci
with a
1.12; that
is,
Since the real parts of the poles and zeros of G(s)H(s) are not affected by the value
of
a,
is
1.12 are at s
always at the origin of the .jplane. The break0.493, 0.857, and 2.65. These are obtained
(8151).
By
shown
when
1.185,
it
can be
and s = 1 converge to a point. The root loci for this situation are sketched as shown in Fig. 828(c). When a is greater than 1.185, Eq. (8151) yields one real root and two complexconjugate roots. Although complex breakaway points do occur quite often in root loci, we can easily show in the present case that these complex roots do not satisfy the original equation of Eq. (8149) for any real K. Thus the root loci have only one breakaway point, as shown in Fig. 828(d) for a = 3. The transition between the cases in Fig. 828(c) and (d) should be apparent.
that the
lie
between s
8.6
The root
a controller that
is
it is
when
these poles
^
'
I!
&
""
ft
C o
XI
ll
"3
o
C
(1)
+
r
J3 **
fc
+
r*i
II II
.S
3 *
3 O '4
o
&5
<3
& ^
90 .
00
11 11
v^2II
as a;
DC
H. "b
425
>t
t
*
if
3
8
*
k o
II
./
H ^
+
2 * 7
8
/^"^
(
""
*
o
x
t^r
T
7
^
I
l
\
1
\
a) to
2'
/
en
1
v.: \^y
k
e
8
1
426
Sec. 8.6
427
In Section 8.5 the root locus diagrams of equations with two variable parameters are studied by assigning different values to one of the parameters. In this section the multiparameter problem is investigated through a more systematic method of embedding. When more than one parameter varies continuously from
It will
oo to oo, the loci of the root are referred to as the root contours. be shown that the same conditions and rules of the root loci are still applicable to the construction of the root contours. The principle of the root contours can be illustrated by considering the
equation
Q(s)
t
K.PAs)
+KP
2
2 (s)
(8153)
where K and K2 are the variable parameters and Q(s), P^s), and P2 (s) are polynomials of s. The first step involves the setting of one of the parameters equal to zero. Let us set K2 equal to zero. Then Eq. (8153) becomes
Q(s)
+ K^is) =
(8154)
loci of this equation may be obtained by dividing both sides of the equation (the Golden Rule) by Q(s). Thus
The root
Kx
Z^=
1
(8155)
or
1
G,(j)ff,(s)
(8156)
The
G^H^s) =
Next,
^^
We
have
(8157)
we
2,
(8153), with
or
1
+G
(s)H2 (s)
=
upon the
(8159)
poles and
^'^ ^HW)
is
(8  160)
G 2 (s)H2 (s)
are identical to
Thus the root contours of the original equation must all start {K 2 = 0) at the points that lie on the root loci of Eq. (8156). This is the reason why one root contour problem is considered to be embedded in another. The same procedure may be extended to more than two
variable parameters.
Example 822
+ K s 2 + K iS + K =0
2
x
(8161)
lie
where
KA
and
between
and
oo.
428
Chap. 8
As a
which
first
step
we
let
K =
2
0; Eq. (8161)
.y
becomes
K^s
Ki
=0
(8162)
is
converted to
(8163)
The root
loci of
as
shown
in Fig. 829(a).
Next, we let 2 vary between zero and infinity while holding Ki at a constant nonzero value. Dividing both sides of Eq. (8161) by the terms that do not contain K2 we have
,
K s*
2
K,s
+K
2
(8165)
t
(8161)
when
K
1
varies
may
G 2 (s)H2 (s)
=
s
KiS
K,
(8166)
iu
f
AT,
splane
K, =
AT,
=0
0.5
(a)
Root contours
for s 3
+ K2 s 2 + K
+K =
t
0,
K2 =
Sec. 8.6
Root Contour
MultipleParameter Variation
429
splane
K2
/
=0(0.4+/1.74)
K, = 2.56
(b)
Root contours
for s 3
\
K2S i + K iS + K
K2
The
varies,
K =
t
constant.
zeros of
G 2 (s)H2 (s)
are at s
0, 0;
G^H^s)
829(a).
which have been found on the contours of Fig. 829(a). Thus for fixed A"i the root contours when K2 varies must all emanate from the root contours of Fig.
Example
823
K +
Ts)(s z
2s
(8167)
1)
It is
K and T as
+
Ts)(s
2
The
written
2s
2)
+K=
(8168)
430
Chap. 8
First,
we
shall set
T equal
this
to zero.
s(s
2
The
characteristic equation
becomes
(8169)
2s
equation when
+ 2) + K = K varies are
rf . s(s 2
.l.
Gi(s)Hds)
as
A, + ^ + 2s
2)
..
17 )
shown
in Fig. 830(a).
fU)
splane
/CO
splane
+/1
*>%_
/t = o / K=\0
K=
4
K=
^K
]
K=
<
T= oo
k=io
T=
o"t = ~
K=
T=
oo
K=4 K=
\\
(a)
a:=io
\ T=
\
(b)
Root
2s
+
2s
2)
+K=
2)]/s(s 2
0. (b)
figuration of
G 2 (s)H2 (s) =
[Ts 2 (s 2
2s
2)
K].
4
G^tfaCs)
Ts
s(s 2
(s
2s
+ 2s + 2) + 2) + K
(8171)
from the polezero on the root contours are at the poles of G 2 (s)H 2 (s), which are the points on the root loci of Eq. (8169), as shown in Fig. 830(b) for K = 10. The T = oo points on the root contours are at the zeros of G 2 (s)H 2 (s), and these are at s = 0, 0, 1 +j\, and 1 y'l. The root contours for the system are sketched in Figs. 831, 832, and 833 for three different values of K; when K 0.5 and T = 0.5, the characteristic equation has a quadruple root at s 1.
varies are constructed
when T
0,
configuration of
the points
Example 824
As an example
G(s)H(s), consider
G(s)H(s)
K(l
s(s
+
1)(*
Ts)
+ 2)
(8172)
The problem may be regarded as a study of the effect of derivative control, in Section 6.7, on the root locations of the characteristic equation. The characteristic equation of the system is
s(s
as discussed
l)(s
2)
K(l
+Ts)=0
T=
in
(8173)
Let us
first
Eq. (8173)
Sec. 8.6
Root Contour
MultipleParameter Variation
431
splane
Fig. 831.
Root contours
for s(l
sT)(s 2
2s
2)
+K=
0;
K>
4.
splane
/
r=i
o<r
1
i splane
/
r^oo
/
o
/ T=0
T^KX>
'o
v" r =
O^T
^\.
T=0
r>
CO
rK
7^0^
^ MJ
r=0
jT^oo
T'*
f
oo
r>=c
\
+
X
\
\
+
2s
Fig. 833.
\
\
Root contours
0; for
,j(l
Fig. 832.
Root contours
for s(l
sT){s 2
2)
+ K = 0; K = 0.5.
yields
*(j
\ + sT)(s 2 + 2s
2)
+K=
K<
0.5.
l)(s
2)
+K= + =
2)
(8174)
which leads to
s(s
l)(s
(8175)
The root
ration of
loci
of Eq. (8174) are sketched in Fig. 834, based on the polezero configu
When T varies
infinity,
we
+ G 2 (s)H2 (s) =
+ s(s +
TKs
l)(s
+ 2)+K
(8177)
432
Chap. 8
/CO
splane
K=
*AT
K=0
K
K=0
*
K=
\',K
= 6
Fig. 834.
Root
1)0
2)
+K=
0.
/OJ
/
/
5plane
,*K =20
AT
= 20
iHt
+H^
1
3.85
\
X
0.425/2.235
K=
20
\\
+
1)0
G 2 (s)H2 (s) =
TKs/[s(s
+ 2)
+
s(s
K],
K=
20.
The points that correspond to T = on the root contours are at the roots of + 1)0 + 2) + K = 0, whose loci are sketched as shown in Fig. 834. If we choose K = 20, the polezero configuration of G (s)H2 (s) is shown in Fig. 835. The root contours of Eq. (8173) for < T < co are sketched in Fig. 836 for three values of
2
Sec. 8.6
433
splane
\a:=20, r=o
Fig. 836.
l)(s
2)
+K+
is
KTs
0.
3.85
0.425
1
+ 0.425 =
1.5
(8178)
is
is always at s = 1.5 because the sum always equal to 3, regardless of the value of K, and the
sum
is
zero.
434
Cna _
The root contours shown in Fig. 836 verify the wellknown fact that the derivative control generally improves the relative stability of the closedloop system by moving the characteristic equation roots toward the left in the splane. The root contours also clearly indicate an important characteristic of the derivative control in that the bandwidth of the system is increased. In certain cases the contribution to the increase in bandwidth by increasing the value of T far exceeds the improvement made on the
relative stability of the system.
bilized for all values of
As shown
K = 20,
largest
the system
is
sta
However, the
damping
ratio that
8.7
is
we
shall
may be
written
(8179)
KP{s)e' T '
s.
=
An
(8180)
where
G.fr)^) =
Thus, similar to the development in Section
the following conditions must be
gg
8.2, in
(8181)
met simultaneously:
where
point s
Note that the condition for any on the complete root loci is given in Eqs. (8183) and (8184), which differ from the conditions of Eqs. (89) and (810) by the term coT. When 0, Eqs. (8183) and (8184) revert to Eqs. (89) and (810). Since co is a variable in the jplane, the angular conditions of Eqs. (8183) and (8184) are no longer constant in the splane but depend upon the point at which a root of Eq. (8179) may lie. Viewing the problem from
s
=a = 5] in
 oo < K <
coT
co
(81 82)
K>0 K<0
(8183) (8184)
T=
another standpoint,
it is recognized that if T 0, given a value of K, there are only n points in the splane that will satisfy either Eq. (8183) or Eq. (8184), for all possible values of k, where n is the highest order of P(s) and Q{s). How
ever, for
I?t0,
co,
may
Sec. 8.7
435
takes
be more than n points which satisfy the angular conditions in the splane, as k on all possible integral values. In fact, there are an infinite number of
these points, since Eq. (8179), which
infinite
is
transcendental,
is
known
to
have an
is
that
many
no longer valid for the present case. It is of interest to investigate how some of the rules of construction given in Section 8.3 may be modified to apply
delay are
to the timedelay case.
K=
Points
Theorem
89.
The
K=
and a
= .
Proof:
Equation (8182)
is
repeated,
e^IG^)/^)^^
Thus,
if
(8185)
G^H^s),
or a, which
is
approaches co.
The
K co
Theorem
Points
810.
The
K = points on the complete root loci of Eq. (8180) (s) and a = co.
t
Proof:
evident.
Number of Branches on
since
loci are
but with
infinite
(8186)
where
is
436
Chap. 8
= m=
n
poles of
zeros of
G^H^s) G^H^s)
Table 82
K=
= N= = N= = N= =
0,
Asymptotes
K =
Asymptotes
>o
Odd
Even
N = even integers
2, 4, odd integers 1, 3, 5,...
.
.
N odd integers
.
<o
Odd
Even
odd integers 1, 3, 5,
even integers
0,
...
2, 4,
= 1, :t3, 5,... = odd integers = 1, .b3, 5,... N = even integers = 0, 2, 4, N = even integers = 0, 2, t4,..
iV
. .
:
> oo on the root loci, K either approaches zero or oo, Theorems 89 and 810 show that the asymptotes are at a = oo (K = oo) and cr = oo (K = 0). The intersections of the asymptotes with the ./coaxis and the conditions given in Table 82 are arrived at by use of Eqs. (8183) and
Proof:
Since as s
(8184).
stated in
The property of the root loci of Eq. (8179) on the real axis is the same as Theorem 87, because on the real axis, co = 0, the angular conditions of Eqs. (8183) and (8184) revert to those of Eqs. (89) and (810), respectively.
Intersection of the
Since Eq. (8179) is not an algebraic equation of s, the intersection of its with the imaginary axis cannot be determined by use of the RouthHurwitz criterion. The determination of all the points of intersection of the root loci with the^'cDaxis is a difficult task, since the root loci have an infinite number of
loci
branches. However,
we
shall
show
sections nearest the real axis are of interest for stability studies.
Breakaway Points
Theorem
must
satisfy
812.
loci
ofEq. (8179)
Sec. 8.7
437
dG
(.s)H 1 (s)e
T'
_
Theorem
(8187)
ds
Proof:
is
similar to that of
88.
st
on
is
l*l
= Tr77^T771T Gl(Sl)lC5l)l
(8
"
188)
where
<r
is
Example
825
+ Ke T *=0
s,
(8189)
It is
desired to construct the complete root loci of this equation for a fixed value of T.
we
get
+ s
Kp~ Ts
=
y
K=
at s
(8190)
which
is
CPifj) =
The following
1.
(8191)
The
K=
From Theorem
89,
and
at
co.
82,
we have
approaches
co
at co
= 0,
=
at
2a/T,
K < 0: K approaches
571 IT,
2.
. .
zero as
a approaches co
at co
n/T,
371/7",
The K = o points: From Theorem 810, Theorem 811 and Table 82, we have
K=
oo
a
at
=
ft)
co.
Using
K>0: K
approaches
. .
+oo
as
approaches
+co
nlT,
37t/r,
K< 0:
A"
approaches
.
co as a approaches +oo at co =
points,
0,
.1njT,
47r/r;
0,
K=
oo
is
loci are
837.
The notation of
for the
loci,
and 0~
complementary root
loci
(K > 0) occupy the negative real axis, and (K 0) occupy the positive real axis.
the
complemen
<
intersections of the root loci with the jco axis are relatively easy to
this
determine for
Since
simple problem.
G i(s)Hi(s)
0, for
any point
si
on
438
Chap. 8
/CO
splane
^K
j4ir/T
K^oo
K~>+
oo
0^K
+
Pn/T
I2v/T
^K
K^
oo
<*>
0+K
+
MT
K^+
K^
>
^K
dzi^.
jn/T
oo
+
Q^K
+
+
J2*IT
a:*
oo
0^K
+ <K
miT
j4*/T
K* +oo
o
Fig. 837.
loci
of the equation s
Ke~ Ts
0.
/GifoWifri)
and for any point
s^
axis,
(8192)
on the negative j CO
/gi(Ji)tfifri)
=4r
loci
(8193)
Thus, for
axis (co
K>
0),
0,
on they CO
>
~ = (2k +
k
l)7t
+coT
(8194)
= 0, 1,
2,
The values of co
n_ 5n_ 9n_
(8195)
Sec. 8.7
439
For
K > 0,
and
<a
< 0,
~
=
. .
(2k
l)w
+ coT
(8196)
and for k
= 0,
1, 2,
CO
= jj, ~2j,
~2j;,
...
Similarly, for
K < 0,
y = 2ta +cor
y=
The crossover
the last
2A:7t
co>0
co
(8198)
+ cor
<
k
(8199)
= 0,
1, 2,
into
two equations.
We have
(8200)
G>=f,2y,^,..5.
Breakaway points: The breakaway points on the complete root determined by the use of Eq. (8187). Thus
loci are
dG (s)H
i
(s)eTs
d^\
ds\
s
}
ds
(8 . 201)
or
Te~ Ts s
from which we have
e~ T '(Ts
e~ T =
at s
(8202)
1)
(8203)
is
IjT.
The
values of
Eq. (8188).
K at the crossover point on they'd) axis are found by using = 0on they co axis, we have Since
<7
where COc is a crossover point. Based on the properties accumulated above, the complete root
are sketched as
loci
of Eq. (8189)
shown
in Fig. 838.
infinite number of roots, and therenumber of branches, from the system analysis standpoint, only the branches that lie between njT < co < n\T are of interest. We shall refer to these as the primary branches. One reason is that the critical value of K at the crossover point on this portion of the root loci is equal to 71/27', whereas the critical
value of
fore,
K at the next branch at co = 5nj2T 571/2T, which much greater. ThereK = 7C/Tis the critical value for stability. Another reason for labeling the primary that for any value of K less than the critical value of branches as the dominant loci
is
is
is
7t/2r, the
all
440
Chap. 8
splane
+ <K
+ <K
Fig. 838.
Complete root
loci for j
Ke~ Ts
0.
Therefore, the transient response of the system, which has Eq. (8189) as its characteristic equation, is predominantly controlled by the roots on the primary branches.
Example 826
As a
loci
slightly
tem shown
G(s)H(s)
Ke~ T
s(s
1)
The
characteristic equation
is
Sec. 8.7
441
r(t)
^ 9
.
K
s{s+
l)
cU)