APPENDIX

APPENDIX
B

APPENDIX
C

APPENDIX
D

APPENDIX D.II

MATLAB lab Coding

Please refer Matlab and Control System Tool Box Help for
assistance.
Some helpful instructions are given below:
The transfer function of a linear, time-invariant, differential equation
system is defined as the ratio of the Laplace transform of the output
variable to the Laplace transform of the input variable, with all initial
conditions assumed to be zero. Although the transfer function can be used
only for linear systems, it yields more intuitive information than the
differential equation. The characteristic equation is obtained by setting
the denominator polynomials of the transfer function to zero. The roots of
the denominator are the system poles, and the roots of the numerator are
the system zeros. The system transfer function can then be specified to
within a constant by specifying the system poles and zeros. The constant,
usually denoted by K, is the system gain factor.
The transfer function model enables us to change system parameters
and rapidly sense the effect of these changes on the system response.
The transfer function is also useful in modeling the interconnection of
subsystems by forming a block diagram representation. The time
response of a system is obtained by the inverse transform of the s-domain
response. This usually requires expansion of the rational function using
partial fractions.
The Control System Toolbox function sys = tf(num, den) creates a
continuous-time transfer function. The output sys is a tf object. For SISO
models, num and den are row vectors listing the numerator and
denominator coefficients in descending powers of s. For example, the
commands
num=[1 4]; den=[1 2 10];
sys=tf(num, den)
results in

Transfer function=

s+ 4
s +2 s +10
2

1. Polynomial Roots and Characteristic Polynomial

If p is a row vector containing the coefficients of a polynomial, roots (p)
returns a column vector whose elements are the roots of the polynomial. If
r is a column vector containing the roots of a polynomial, poly(r) returns
a row vector whose elements are the coefficients of the polynomial.
2. Poles and Zeros of a Transfer Function
tf2zp finds the zeros, poles and gains of a transfer function
3. Partial-Fraction Expansion
[r,p,k] = residue[b,a] finds the residues, poles, and direct terms of a
partial fraction expansion of the ratio of two polynomials. [num, den] =
residue(res, poles, K) converts the partial fraction expansion back to
the polynomial P(s)=Q(s).
4. The Time Response of Second-Order Systems
Assessing the time-domain performance of closed-loop system models is
important because control systems are inherently time-domain systems.
The performance of dynamic systems in the time domain can be defined
in terms of the time response to standard test inputs. One very common
input to control systems is the step function. If the response to a step
input is known, it is mathematically possible to compute the response to
any input. Another input of major importance is the sinusoidal function. A
sinusoidal steady-state output is obtained when an asymptotically stable
linear system is subjected to a sinusoidal input. Thus, if we know the
response of a linear time invariant system to sinusoids of all frequencies,
we have a complete description of the system.
A function called timespec(num, den) is written which obtains the
time-domain performance specifications, tp, tr and ts. num and den are

the numerator and denominator of the system closed-loop transfer

function. Given a transfer function of a closed-loop control system, the
Control System Toolbox function step(num, den) produces the step
response plot with the time vector automatically determined.
5. Frequency Response of Systems
The frequency response of a system is defined as the steady-state
response of the system to a sinusoidal input signal.
The frequency response method and the root-locus method are simply
two different ways of applying the same basic principles of analysis. These
methods supplement each other, and in many practical design problems,
both techniques are employed. One advantage of the frequency response
method is that the transfer function of a system can be determined
experimentally by frequency response tests. Furthermore, the design of a
system in the frequency domain provides the designer with control over
the system bandwidth and over the effect of noise and disturbance on the
system response.
i.

Bode Plot

The Bode plot consists of two graphs plotted on semi-log paper with
linear vertical scales and logarithmic horizontal scales. The first graph is a
plot of the magnitude of a frequency response function
decibels versus the logarithm of

G( j)

in

, the frequency. The second graph of

a Bode plot shows the phase function

()

versus the logaritm of .

The logarithmic representation is useful in that it shows both the low- and
high-frequency characteristics of the transfer function in one diagram.
Furthermore, the frequency response of a system may be approximated
by a series of straight line segments.
Given a transfer function of a system, the Control System Toolbox
function bode(num, den) produces the frequency response plot with the
frequency vector automatically determined.
However, the function frqspec(w, mag) is developed which will return
r , M p , and

based

on the functional values of w and mag.

ii.

Polar Plot

A polar plot, also called the Nyquist plot, is a graph of

ReG ( j)

with varying from

ImG ( j)

versus

+ . The polar plot may be directly

graphed from sinusoidal steady-state measurements on the components

of the open-loop transfer function. The MATLAB Control System Toolbox
function [Re, Im]= nyquist(num, den, ) returns the real and
imaginary parts of a transfer function for the specified range of
frequencies.
6. Stability
For a system to be usable it must be stable. A linear time-invariant
system is stable if every bounded input produces a bounded output. We
call this characteristic stability. The response is bounded if the poles of the
closed loop system are in the left-hand portion of the s-plane. Thus, a
necessary and sufficient condition for a feedback system to be stable is
that all the poles of the system transfer function have negative real parts.
The stability of a linear time-invariant system may be checked by using
the Control System Toolbox function impulse to obtain the impulse
response of the system. The system is stable if its impulse response
approaches zero as time approaches infinity. One way to determine the
stability of a system is by simulation. The function lsim can be used to
observe the output for typical inputs. This is particularly useful for
nonlinear systems. Alternatively, the MATLAB function roots can be
utilized to obtain the roots of the characteristic equations. In the classical
control theory, several techniques have been developed requiring little
computation for stability analysis. One of these techniques is the RouthHurwitz criterion.
A function called routh(a) is written that forms the Routh array and
determines if any roots have positive real parts. a is a row vector
containing the coefficients of the characteristic equation.

7. Steady-State Error and System Type

In addition to being stable, a control system is also expected to meet a
specified performance requirement when it is commanded by a set-point
change or disturbed by an external force. The performance of the control
system is judged not only by the transient response, but also by steadystate error. The steady-state error is the error as the transient response
decays leaving only the continuous response. High loop gains, in addition
to sensitivity reduction, will also reduce the steady-state error. The steady
state error for a control system is classified according to its response
characteristics to a polynomial input. A system may have no steady-state
error to a step input, but the same system may exhibit nonzero steadystate error to a ramp input. This depends on the type of the open-loop
transfer function.
Two functions, errorzp(z,p,k) and errortf(num, den), are written for
computation of system steady-state error due to typical inputs, namely
unit step, unit ramp, and unit parabolic. errorzp(z,p,k) finds the steadystate error when the system is represented by the zeros, poles, and gain.
z is a column vector containing the transfer function zeros, p is a column
vector containing the poles, and k is the gain. If the numerator power, m,
is less than the denominator power, n, then there are n-m zeros at infinity
and vector z must be padded with (n-m) inf. errortf(num, den) finds the
steady-state error when the transfer function is expressed as the ratio of
two polynomials.
8. Root-locus
The root locus method enables us to find the closed-loop poles from the
open-loop poles for all the values of the gain of the open-loop transfer
function. The root-locus of a system is a plot of the roots of the system
characteristic equation as the gain factor K is varied. Therefore, the
designer can select a suitable gain factor to achieve the desired
performance criteria. If the required performance cannot be achieved, a

controller can be added to the system to alter the root-locus in the

required manner.
Given a transfer function of an open-loop control system, the Control
System Toolbox function rlocus(num, den ) produces a root-locus plot
with the gain vector automatically determined.
rlocus function is accurate, and we use it to obtain the root-locus. The
command rlocfind finds the root locus gains for a given set of roots. [K,
Poles] = rlocfind(num, den) is used for interactive gain selection from
the root locus plot of a system generated by rlocus. rlocfind puts up a
crosshair cursor in the graphics window which is used to select a pole
location on an existing root locus. The root locus gain associated with this
point is returned in K and all the system poles for this gain are returned in
POLES. A good knowledge of the characteristics of the root loci offers
insights into the effects of adding poles and zeros to the system transfer
function. It is important to know how to construct the root locus by hand,
so we can design a simple system and be able to understand and develop
the computer-generated loci. The Control System Toolbox function rlocus
(num, den, K) can be used to obtain the closed-loop root-locus for
varying K.
9. Root-Locus Design
The design specifications considered here are limited to those dealing
with system accuracy and time-domain performance specifications. These
performance specifications can be defined in terms of the desirable
location of the dominant closed-loop poles.
The root-locus can be used to determine the value of the loop gain K,
which results in a satisfactory closed-loop behavior. This is called the
proportional controller and provides gradual response to deviations from
the set point. There are practical limits as to how large the gain can be
made. In fact, very high gains lead to instabilities. If the root-locus plot is
such that the desired performance cannot be achieved by the adjustment
of the gain, then it is necessary to reshape the root-loci by adding the

additional controller Gc(s) to the open-loop transfer function. Gc(s) must

be chosen so that the root-locus will pass through the proper region of the
s-plane. The proportional controller (P) has no sense of time, and its action
is determined by the present value of the error. An appropriate controller
must make corrections based on the past and future values. This can be
accomplished by combining proportional with integral action (PI) or
proportional with derivative action (PD).
There is also a proportional-plus-integral-plus-derivative controller (PID).
Gc ( s ) =K p +

KI
+ KD s
s

The ideal integral and differential compensators require the use of

active amplifiers.
Other compensators which can be realized with only passive network
elements are lead, lag, and lead-lag compensators. A first-order
compensator having a single zero and pole in its transfer function is
Gc ( s ) =

K c (s+ z 0 )
s+ p 0

Several functions are developed for the root-locus design. These are
Function

Controller

[numopen, denopen, denclsd] = pcomp(num, den, _)

Proportional
[numopen, denopen, denclsd] = phlead(num, den, s1)
Phase-Lead
[numopen, denopen, denclsd] = phlag(num, den, _)
Phase-Lag
[numopen, denopen, denclsd] = pdcomp(num, den, s1)

PD

[numopen, denopen, denclsd] = picomp(num, den, s1)

PI

[numopen, denopen, denclsd] = pidcomp(num, den, s1) PID

Alternatively, the function [numopen, denopen, denclsd] =
rldesign(num, den, s1) allows the user to select any of the above

controller designs.

s 1= + j

is a desired pole of the closed-loop transfer

function, except for the pcomp and phlag controllers where

, the

damping ratio of the dominant poles, is substituted for s1. num and den
are row vectors of polynomial coefficients of the uncompensated openloop plant transfer function. The function phlead(num, den, s1) may
also be used to design phase-lag controllers. To do this, the desired pole
location s1 must be assumed slightly to the right of the uncompensated
pole position. The function returns the open-loop and closed-loop
numerators and denominators of the compensated system transfer
function.
i.

Gain Factor Compensation or P Controller

The proportional controller is a pure gain controller. The design is

accomplished by choosing a value of K0 which results in a satisfactory
transient response. The developed function [numopen; denopen; denclsd]
= rldesign(num; den; ) displays six options for root-locus design. For a
proportional controller, option 1 must be selected. This option calls upon
the function [numopen, denopen, denclsd] = pcomp(num, den, ).
The function computes the required gain K0 that will result in a desirable
damping ratio in the step response.
ii.

Phase-Lead Design

The function [numopen, denopen, denclsd] =phlead (num, den,

s1) is developed for the phase-lead controller design. num and den are
the numerator and denominator of the polynomial coefficients of the open
loop plant transfer function and s1 is the desired dominant closed-loop
pole. The user enters the compensator DC gain. The function obtains the
controller transfer function and roots of the compensated characteristic
equation. Also, the function returns the
open-loop and closed-loop numerators and denominators of the
compensated system
transfer function.

iii.

Phase-Lag Design

The function [numopen, denopen, denclsd] = phlag(num, den, )

is developed for the phase-lag controller design, based on the above
approximate criteria. num and den are the numerator and denominator
of the polynomial coefficients of the open loop plant transfer function, and
is the desired damping ratio of the dominant closed loop poles. The user
is prompted to enter the gain K to realize the steady-state error and the
compensator zero, z0. The function obtains the controller transfer function
and roots of the compensated characteristic equation. Also, the function
returns the open-loop and closed-loop numerators and denominators of
the compensated system transfer function.
A lag-lead controller may be obtained by appropriately combining a lag
and a lead network in series.
iv.

PID Design

One of the most common controllers available commercially is the PID

controller. Different processes are suited to different combinations of
proportional, integral, and derivative control. The control engineers task
is to adjust the three gain factors to arrive at an acceptable degree of
error reduction simultaneously with acceptable dynamic response.
Based on the above equations, three functions called [numopen,
denopen, denclsd] = pdcomp(num, den, s1), [numopen, denopen,
denclsd] = picomp(num, den, s1) and [numopen, denopen,
denclsd] = pidcomp(num, den, s1) are written for the PID controller
design. num and den are the numerator and denominator of the
polynomial coefficients of the open-loop plant transfer function and s1 is
the desired dominant closed-loop pole. The function obtains the controller
transfer function and roots of the compensated characteristic equation.
Also, the function returns the open-loop and closed-loop numerators and
denominators of the compensated system transfer function.

a. PD Controller
Here both the error and its derivative are used for control, and the
compensator transfer function is

G c ( s ) =K p + K D s=K D s+

Kp
KD

From above, it can be seen that the PD controller is equivalent to the

addition of a simple zero at

Kp
KD

( )

s=

to the open-loop transfer function

which improves the transient response. From a different point of view, the
PD controller may also be used to improve the steady-state error, because
it anticipates large errors and attempts corrective action before they
occur. The function [numopen, denopen, denclsd] = rldesign(num,
den, s1) with option 4 is used for the PD controller design.
b. PI Controller
The integral of the error as well as the error itself is used for control, and
the compensator transfer function is
G c ( s ) =K p +

K I K p ( s+ K I /K p )
=
s
s

The PI controller is common in process control or regulating systems.

Integral control bases its corrective action on the cumulative error
integrated over time. The controller increases the type of system by 1 and
is used to reduce the steady-state errors. The function [numopen,
denopen, denclsd] = rldesign(num, den, s1) with option 5 is used for
the PI controller design.
c. PID Controller
The PID controller is used to improve the dynamic response as well as to
reduce or eliminate the steady-state error. The function [numopen,
denopen, denclsd] = rldesign(num, den, s1) with option 6 is used for
the PID controller design.