SYSTEMS THEORY – AN INTRODUCTION

1.1 .INTRODUCTION
The word ‘systems’ has been used in many contexts to mean very different things. But in the
present instance of a book on Control Systems, the definition should be clear enough. In a
general sense, a system can be loosely defined as a collection of subsystems or elements
acting in a coordinated way to achieve a specific objective. Taking a more general
perspective, a system can be broadly defined as a combination of complete, predictable
cycles, structures, or processes occurring in a natural manner or may be an artificial
construction created to represent or explain a natural occurrence characterized by system
boundaries and interrelationships of subsystems and is acted on by inputs and results in
outputs from it.

Systems Theory is the interdisciplinary or trans-disciplinary study of the abstract

organization of phenomena, independent of their substance, type, or spatial or
temporal scale of existence. It investigates both the principles common to all complex
entities, and the (usually mathematical) models, which can be used to describe them.

A system in the present context can be characterized by a set of inputs, a set of

outputs and a boundary that separates the system from the environment. Inputs are
those quantities that are acting on the system from the environment and the outputs
are the results of the input acting on the system. Some of the inputs are not easily
understood or cannot be controlled and are not within the capabilities of the operator
to manipulate it to achieve a desired output or end result. These inputs are usually
called disturbances. The boundary that separates the system from the environment can
be a physical one like that of a machine or a geographical area. But in many systems,
this boundary can be purely imaginary or conceptual rather than a physical one,
within the definition of the system. Moreover, the same physical system may be
represented using different inputs and outputs depending on the study that is being
carried out .
disturbances

inputs SYSTEM outputs

environment

Fig.1.1. System – characterization

Consider an example of the familiar power system that supplies electric power to
different consumers on the system. For the electric power-distributing agency, it is
necessary to calculate how much power is required at a given time so that the
generators in the stations can be kept ready to meet the additional power demand as
and when required. Usually, generators require a start-up time of several hours
depending on the type of station. Hydroelectric stations are quickly starting type
whereas thermal stations require considerable to start up.
One way of analyzing the load demand on the power system is to identify this as the
output of a system and the inputs being identified as the causes that ultimately decide
the load demand. For hourly variation of the load, it is convenient to define the inputs
as the temperature, humidity, wind velocity and visibility. As these weather dependent
factors change, the load demand changes significantly. So, a system can be defined as
shown in Fig.1.2.

temperature

humidity
wind velocity
POWER SYSTEM Load demand (MW)
visibility

Fig.1.2. weather dependent model of power system

If on the other hand, the purpose of the analysis or modeling is to determine the
expected installed capacity after a few years, then, a better model of the system can be
with inputs chosen as population (in millions), agricultural outputs and industrial
outputs ( values in millions of rupees ) and the output as the installed capacity or the
load demand in MW( Fig.1.3). The system appears to be different , even though
physically both cases we deal with the same system.

population
agricultural output
POWER SYSTEM Installed capacity
Industrial output (MW)

Fig.1.3. Modeling Power System for Economic Planning

1.2.CLASSIFICATION OF SYSTEMS

Systems can be classified in different ways based on the following.

1. capacity to store energy or otherwise
2. nature of signals
3. input output relations
4. time variation of characteristics
5. temporal and special variation
6. randomness or otherwise

1.2.1.Static and dynamic systems

Systems which are in equilibrium with no significant changes taking place are said to
be static systems. These systems essentially contain only components which can
dissipate or consume energy. These systems can be described in simple
mathematical terms by a set of algebraic equations. Dynamic systems on the other
hand, contains elements or components which can store energy and release it when
allowed to do so. Such systems have the capacity of ‘remembering’ what it had been
subjected to previously, or has some memory built into it. These systems are to be
 described by dynamic equations or differential equations of appropriate type. An
example from electrical engineering for static systems is an electric circuit with only
resistance elements. This circuit cannot store electrical or magnetic energy as a
resistive circuit can only dissipate power as heat which cannot be recovered or stored.
On the other hand, inductive elements can store energy in a magnetic field and release
it when allowed to. Capacitances can store energy in an electric field and can also
release it when a capacitance is given a discharge path. Both these elements, thus,
have the capacity to store electric energy and when describing such elements, static or
algebraic equations are not sufficient. These circuits are to be described by dynamic
equations called differential equations. Similar examples are a flywheel that can store
kinetic energy in moment of inertia and a spring which can store potential energy.

1.2.1.Continuous or discrete:

Systems, which continuously evolve in time, are called continuous systems. Most of
the systems in its natural way are continuous. Discrete or discontinuous systems are
characterized by one or more components, which make some of the variables defined
only at discrete instants of time with no signal in between. The numbering system,
which uses only integers, is obviously discrete. The advent of digital computers and a
variety of digital devices have made the need for the study of discrete systems very
essential. There may be systems which are inherently discrete like the operation of a
radar in which a beam of waves are sent to the prospective target during the scanning
stage and then the radar waits for the reflected waves to identify the target and then
the cycle repeats. However, the introduction of an operation called ‘sampling’ by
which a continuous or analogue signal is sampled at discrete instants of time makes
the system discrete or digital. This is an example for ‘intentional’ discrete systems.
The situation is very similar when a digital computer is part of the system as it can
take only digitally coded signals. There are definite advantages in having a digital
system like reduced power demand, better noise discrimination properties and better
utilization of available resources.

S(t S(kT)

àt à t=kT

Fig.1.3 (a) Continuous signal Fig.1.3 (a) Sampled signal

1.2.2.Linear and or Nonlinear Systems

Systems can also be classified as linear or nonlinear based on the input

output characteristics. Linear systems are those where the superposition principle
holds good. This essentially means that, the response of a system due to several inputs
simultaneously applied can be obtained by adding the responses due to each one of
the inputs separately applied (property called additivity). It will also mean that if the
magnitude of the input is scaled up or down by a constant factor, the corresponding
output also gets multiplied by the same scale factor( homogeneity) .
 Mathematically, if input x1 ( t ) results in output y1 (t ) , and x 2 (t ) results in
output y 2 (t ) then additivity property implies x1 (t ) + x 2 (t ) results in y1 ( t ) + y 2 ( t ) .
Homogeneity implies a x1 (t ) results in a y1 (t ) where a is a real constant. Together,
superposition principle states that :
a x1 (t) + b x2 ( t) results in a y1 (t ) + b y 2 (t )

The validity of the superposition principle is the reason for the generalization
possible in the study of linear systems. However, the nonlinear systems do not satisfy
the superposition principle and consequently generalization of results is not possible n
the case of such systems. Very often, numerical studies based on a specified set of
conditions are required to infer the behaviour of such systems. Nonlinear systems
exhibit special characteristics such as amplitude and frequency jumps in output, limit
cycles, multiple equilibrium points or equilibrium zones and harmonics generation at
the output for a sinusoidal input of constant frequency. It is also well known that all
linear systems are approximations to real systems, which are nonlinear to a certain
extent, the linear approximations being valid for specified range of operation.

1.2.3.Time Invariant and or Time varying Systems

Think of a system whose performance or behaviour depends on the time of

application of the input. If the response depends not only on the nature and type of the
input, but also on the time instant at which the input is applied, a time varying system
results. These changes can take place due to aging and environmental factors or may
even be due to drastic changes in the parameters due to gravitational changes, changes
in payload etc as in the case of space vehicle dynamics. If the parameters can be
assumed to be constant over a wide range of operation, a time- invariant model can be
used to describe the system. Both time invariant and time varying systems can be
described by differential equations The equations for the time invariant systems will
be with constant coefficients and time-varying systems will have coefficients that are
functions of the independent variable. If the inputs are time varying, the output in
most of the cases will be time varying whether the system is time invariant or time
varying. In the case of time varying systems, the time of application of the input also
will be a factor in deciding the nature of the output response.

1.2.4.Lumped and Distributed Systems

There are many natural phenomena that can be described as functions of time
alone. However, there are many which varies according to time and also the place at
which the study is carried out. In such cases, we have to deal with time t and space
(distance) x to be the two independent variables. Familiar examples are the
parameters of a transmission line which are distributed all over the length of the line
in a uniform manner and the load on a concrete beam which is also uniformly
distributed over the length of the beam. While studying a system, usually we are
interested in studying the performance of the system as a function of time, i.e. the
independent variable is taken as time. We tacitly assume that the system is located at a
particular place or there are no variations with respect to spatial location. This may
not be true in all situations. A correct way will be to consider the time and space
(distance from a reference point) as independent variables. When we assume that the
special variation is insignificant or it can be ignored, we have a lumped (parameter)
system. The other, considering both time and space as independent variables are
called ‘distributed’ (parameter) systems. Obviously, when we have two independent
variables, the variation with respect to either or both will be of interest and we need
partial differential equations for describing such systems. The lumped parameter
systems can be effectively described by total differential equations.

1.2.5. Deterministic and stochastic systems

Uncertainty or randomness is present in all systems. If the system behaviour can be

predicted with certainty, the system equations can be solved with ease for different
inputs and exact prediction of the performance under input excitations will be
possible. However, if the characteristics of the system can be described only in terms
of statistical or probabilistic description, stochastic methods will have to be employed
for the analysis. Probability distribution functions, and other statistical measures will
be used to describe the system performance. Typical examples are again, space
applications in which the performance criteria to be satisfied cannot be arrived purely
based on deterministic considerations. Noise properties may not be completely known
and only statistical analysis can be carried in terms of averages, variances etc rather
than exact and precise values.

1.2.6.Adaptive and Non-adaptive systems

Many systems have the property of adapting itself to changes in the

environment, particularly living or biological systems. The sudden changes in the
temperature and other environmental factors trigger almost instantaneous changes in
the biological systems to maintain its existence without adverse effects. These
systems are self-adaptive or have the property of ‘homeostasis’. Non-adaptive
systems cannot perform satisfactorily under changed environmental conditions. In
order to evolve a suitable strategy for controlling non-adaptive systems, it is necessary
to identify the nature of changes taking place in the system due to changes in
environmental factors. This process is called ’identification’ which is an essential
part of control of a non-adaptive system.

1.3. Closed loop and open loop systems

This book is on automatic control systems and the purpose of any control system is to
maintain or control an output in a desired manner by varying the input. In this context,
control systems can be classified into two categories, open loop control systems and
closed loop control systems.

1.3.1. Open loop systems

Open loop control systems are characterized by an open loop controller, which tries
to set the input value such that the output is as close as possible to the desired value.
However, it is important to note that unlike in closed loop systems, there is no
continuous measurement of the output and comparison of the actual output with the
desired output, or in other words, there is no feedback. The success of the control
action greatly depends on the settings of the controller, which may or may not ensure
the output being near the desired value. Mostly, these are manual controllers and an
operator sets the controller with the help of a calibration chart or look up table, which
tabulates the setting for different outputs.Simple examples of open loop control
systems are (a) fan regulator which sets the speed at a value irrespective of whether
the speed of the span is good enough to cool off the room to the desired level. Or a
furnace fitted with a chart ind icating the voltage setting required for different
temperatures inside the furnace.
Controlled
Corrective output
input signal
Open loop Plant being
Controller controlled

Fig.1.4. Open loop control system

Open loop control systems have the advantage of being simple, easy to construct and
cheap, even though the accuracy of the control may not be very good. This will
depend to a large extent on the judgment of the operator for a manually operated
system or the correctness of the calibration chart if a chart is used to set the magnitude
of the corrective signal.

1.3.2. Closed loop systems

Closed loop systems are also called feedback systems; as the actual output is
continuously being measured and fed back to the input side. Very often, a device
called error detector finds the difference between the actual output and the desired
output(called error) and the corrective signal from the controller depends on this error
signal. s there is continuous comparison between the actual output and the desired
output, there is every possibility that they may be close together. This means closed
loop systems are more accurate than closed loop systems. Closed loop systems
however, are more complex and expensive and have tendency to become unstable or
may show oscillatory output.

corrective
signal
Ref input
error
CONTROLLER PLANT Controlled
output

feedback sensor

Fig.1.5. Closed Loop (feed back) system

1.4 Examples of open loop systems

a) A furnace fitted with a calibrated handle to fix the output temperature
b) A fan with a regulator only
c) A blindfolded man reaching for a book on the table
1.5.Examples of closed loop systems
a) A man reaching for a book on the table
b) A thermostat control of fridge (ON-OFF Control)
c) Position Control Systems
d) Closed Loop Voltage Regulators
e) Closed Loop Control of Temperature in furnaces