Process Control Engineering

Process Control Engineering

P. Sai Krishna
Dept. of Instrumentation Engineering
M.J. College of Engineering & Technology
Hyderabad (A.P).
©Copyright 2020 I.K. International Pvt. Ltd., New Delhi-110002.

This book may not be duplicated in any way without the express written consent of the publisher,
except in the form of brief excerpts or quotations for the purposes of review. The information
contained herein is for the personal use of the reader and may not be incorporated in any commercial
programs, other books, databases, or any kind of software without written consent of the publisher.
Making copies of this book or any portion for any purpose other than your own is a violation of
copyright laws.

Limits of Liability/disclaimer of Warranty: The author and publisher have used their best efforts in
preparing this book. The author make no representation or warranties with respect to the accuracy or
completeness of the contents of this book, and specifically disclaim any implied warranties of
merchantability or fitness of any particular purpose. There are no warranties which extend beyond the
descriptions contained in this paragraph. No warranty may be created or extended by sales
representatives or written sales materials. The accuracy and completeness of the information provided
herein and the opinions stated herein are not guaranteed or warranted to produce any particulars
results, and the advice and strategies contained herein may not be suitable for every individual.
Neither Dreamtech Press nor author shall be liable for any loss of profit or any other commercial
damages, including but not limited to special, incidental, consequential, or other damages.

Trademarks: All brand names and product names used in this book are trademarks, registered
trademarks, or trade names of their respective holders. Dreamtech Press is not associated with any
product or vendor mentioned in this book.

ISBN: 978-93-89633-42-9

EISBN: 978-93-89976-69-4

Edition: 2020
I dedicate this work at the divine feet of my Master and Mentor
“Bhagavan Sri Satya Sai Baba”
 Preface

Process control instrumentation is an experimental science, which directly deals

with the analysis of certain industrial control systems. This book is an attempt to
present certain intricate aspects of this science within the range of
comprehension for undergraduate students pursuing engineering course with
specialization in Instrumentation Technology.
Since this subject comes in the higher semesters of engineering degree
course the students are expected to have the basic knowledge of control systems
engineering, specially the topics related to time domain analysis stability, and
frequency response analysis respectively.
The course work in this textbook is divided into seven chapters dealing in
depth and in a simple and lucid language for easy assimilation and
understanding for the reader.
Chapter 1 introduces general terminology of process control in terms of the
characteristics and dynamics of the process. The mathematical models of any
process or system entirely depend on certain dynamics that govern the system. In
this chapter all the various topics are deal with by taking relevant examples for
better understanding the modeling of the processes. The elements of process
control loop are also discussed exhaustively by citing the examples of level, flow,
and temperature and pressure control systems.
Also piping and instrumentation drawings are also included towards the end
of the chapter, which are based on ISA standards.
Chapter 2 deals with the most important part of any process control system
i.e. controller. It is a device that performs analysis on error and operates the final
control element to achieve the control objective. The various signal transmission
ranges and also the controller parameter range in terms of percentage controller
output with direct and reverse action concepts forms the beginning of the
chapter.
The various controller modes whose mathematical representations are
usually abstract in nature have been dealt with extreme care and simplicity so that
the reader should be able to assimilate the various concepts independently. The
basic discontinuous modes and their extensions that are continuous modes of
control actions have been discussed with examples for better understanding. The
viii 2HAB=?A

mathematical expressions are supplemented by the graphical representations

and real time applications that gives a thorough interpretation to the actual
working of a particular controller mode.
In chapter 3, the effect of control actions are discussed in detail by taking
examples of different types of processes. After completing this chapter the reader
should be able to understand how actually a particular controller mode works, for
example, in a PID controller what are the series of events that take place in
control function.
Examples of single capacitance, single time constant (First order system) and
also multiple time constant process (Second order system) with single input and
multiple inputs are discussed. However the reader is advised to refer and refresh
with the basics of time domain analysis from control systems engineering
specially the topics related to error analysis.
The next chapter 4 deals with the optimum settings of controllers, which
refer to tuning of controllers. The controller should be properly tuned i.e. its
parameters should be optimally adjusted to give better overall performance in a
control loop. One of the classical approaches to tuning is open loop process
reaction curve method, which is limited to processes having self-regulation only.
The more advanced methods approved by ISA standards, which can be
employed to processes with, or with out self-regulation, like Ziegler-Nicols
Tuning etc. are discussed in this chapter. Since the frequency response method is
also discussed the reader should have a clear understanding of BODE PLOTS.
Chapter 5 is dedicated to the implementation of controller configurations.
The implementation of P, P+I, P+I+D, etc. can be done electronically,
pneumatically and hydraulically. The different elements constituting these
various modes and their controller implementation form the subject matter of
this chapter.
Chapter 6 is devoted exclusively to the final control elements. The output of
a controller is fed to a final control element, which manipulates some variable in
a process to ensure that the control objective is achieved. The different types of
actuators, positioners and control valves are discussed in this chapter. The
control valves used in industrial applications generally are either linear, equal
percentage or quick opening. These control valves characteristics are dealt with
initially followed by different types of valves. Also the cavitations and flashing
problems of the valves are also discussed.
Chapter 7 winds up the readers journey of this beautiful subject with
introduction to the other different process control loop configurations. In the
entire text only feedback control loop configuration is considered because
analysis of systems can be better understood with it. However, other
configurations like feedforward, cascade etc. which are modifications of feedback
control are also discussed in this last chapter.
This reference-cum-textbook, I hope will be useful to the reader for clear
understanding of this abstract science. Illustrative examples and a number of
solved problems are included in the text to make the reading self-instructive and
also interesting.
 2HAB=?A ix

Any constructive criticism and suggestions are always welcome from

instructors and students to make the book more informative with relevant
content. This text also covers the syllabi for competitive examinations like GATE
and other industry related examinations like IOC, HPCL, and BARC to name a
few which extensively employ process control instruments and systems.

P. Sai Krishna
 Acknowledgements

I am ever grateful to my parents Shri P. Rajkumar and Smt. (Late) P. Vijaya

Lakshmi without whose unconditional love and support. I wouldn't have been
what I am today. My obeisance to them.
I am also thankful to my wife Smt. P. Sharthi and my son master P. Sai
Vishnu Chandra without whose support the preparation of this manuscript would
have been a distant dream.
Also to all my family members, friends and well wishers whose good wishes of
always chesish.
I extend my gratitude to mangement, Principal, and Vice-Principal and staff
of Instrumentation Engineering department of M. J. College of Engineering and
Technology, Hyderabad for having believed in my strengths and ablities,
extended continuous support in all my endeavours.
Finally, I specially thank all my dearest students spread across length and
brendth of India who have received me and my lectures with uncomparable zeal
and enthusiasm.

Author
 Contents

Preface vii
Acknowledgements xi
1. Introduction 1
1.1 Introduction to Process Control 1
1.2 Process Characteristics 2
1.3 Elements of Process Dynamics 9
1.4 Interacting and Non-interacting Systems 30
1.5 Process Degrees of Freedom 37
1.6 Piping and Instrumentation Drawings 38
2. Modes of Controller Operation 58
2.1 Reverse and Direct Action of Controllers 59
2.2 Modes of Control Actions 59
2.3 Discontinuous Controller Modes 60
2.4 Continuous Controller Modes 64
2.5 Composite Controller Modes 71
3. Effect of Control Actions on Processes 95
3.1 Concept of Closed Loop Automatic Control 95
3.2 Error Equation and Error Terminology 97
3.3 Effect of Two Position Controller 100
3.4 Effect of Proportional Controller Mode 103
3.5 Effect of Integral Controller Mode 114
3.6 Effect of Derivative Controller Mode 120
3.7 Effect of Composite Controller Modes 123
4. Tuning of Controllers 152
4.1 Introduction 152
4.2 Process Reaction Curve Method 154
4.3 Quarter Amplitude Criteria [Cohen-coon Corrections] 156
4.4 Ziegler-Nicol Tuning Method 158
xiv 2HAB=?A

4.5 Fundamental Frequency Response Method of 160

Controller Tuning
4.6 Controller Tuning from Routh Array Method 163
5. Implementation of Control Actions 170
5.1 Introduction to Operational Amplifiers 170
5.2 Electronic Implementation of Controller Modes 172
5.3 Introduction to Pneumatic Elements 184
5.4 Pneumatic Implementation of Controller Modes 188
5.5 Hydraulic Implementation of Controller Modes 193
6. Final Control Elements 198
6.1 Introduction 198
6.2 Introduction to Control Valves 207
6.3 Control Valve Characteristics 211
6.4 Types of Control Valves 214
6.5 Control Valve Sizing and Selection 225
6.6 Cavitation and Flashing Problem in Control Valves 228
7. Process Control System Configurations 229
7.1 Feed Forward and Feed Back Control 229
7.2 Cascade Control Configuration 233
7.3 Ratio Control Configuration 238
7.4 Split Range Control Configuration 24
7.5 Other Types of Control Configurations 240
Index 245
 +0)26-4

Introduction

1.1 INTRODUCTION TO PROCESS CONTROL

The science of Cybernetics or Instrumentation may be defined as an art or
technology of using instruments to measure the physical and chemical
properties and changes of matter. Since it is not possible to measure these
changes directly, certain variables like temperature, pressure, humidity, level
etc. are measured which affect these changes of properties of matter.
Thus, instrumentation predominantly deals with measurement and
control of these variables. It is a science of study of instruments or systems or
their application for the purpose of observation, measurement and control of
non-electrical quantities and parameters.
It may be broadly classified under two Instrumentation
main categories, as shown.
The term “Process” refers to physical or
chemical change of matter or an ensemble of Process Process
environment where a particular variable is Measurement Control
measured and controlled.
Process measurement, therefore, implies “the acquisition of information
that establishes the magnitude of process quantities”.
Process control refers to “Regulation and manipulation of variables
influencing the conduct of a process in such a way as to obtain a product of
desired quality and quantity in an efficient manner”.

Brief History of Automatic Control

Before we delve into the characteristics of process control systems it is
important and instructive to trace the development of automatic control in
brief.
Until the middle of 18th century, automatic control systems were not
developed. The first automatic control system, the flyball governor, to control the
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

2 Process Control Engineering

speed of steam engines was invented by James Watt in the year 1770. After a
long gap of nearly ten decades, Maxwell analysed the dynamics of flyball governor.
During the world war I, positioning of heavy masses like guns and ships
became an absolute necessity.
Minorsky performed the classic work on automatic steering of ships and
positioning of guns on the ship boards.
Servo mechanisms were developed by Hazen during the year 1934. During
world war II, many mathematical, analytical and practical methods were
developed in the field of Automatic Control.

1.2 PROCESS CHARACTERISTICS

To understand the characteristics of process control system, consider a simple
problem where in level of liquid in tank is to measured and controlled i.e. to be
kept at some desired value irrespective of disturbances that tend to deviate it
from the desired value. (Refer Figure 1.1)

qi (Inlet)
Sight tube

Human operator
h
qo
(Control valve) (Outlet)

FIG. 1.1
Consider a tank having inlet with inlet flow rate of liquid (qi) and outlet
with outflow rate (qo). The objective is to maintain the level of the liquid in the
tank at a prescribed height ‘h’ measured from the bottom of the tank.
A control valve is placed at the outlet, so that the outlet flow rate may be
altered and monitored.
To achieve the control objective or to convert it into a control system a
human operator is placed near the control valve as shown. This human
operator has the desired value of level in his mind, also known as set point. He
constantly monitors the sight tube and measures the level of the liquid in the
tank. If the level is above the set point or desired value he performs an
operation on the control valve by opening it till the level reaches the set point
and fixes the control valve at particular position which makes outlet flow rate
equal to inlet flow rate, thus maintaining the level at desired value.
Hence, any changes in inlet flow rate causing the level to change, is
continuously monitored by the human operator, who is performing certain
functions like measuring, comparing, evaluating and operating on the control
valve to ensure that the control objective is met.
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 3

Thus the control system shown in Fig 1.1 is manual control system. In this
context it is important to define “process” as applied to Figure 1.1.
The “process” obviously refers to ensemble of tank, liquid in the tank, the
inlet and outlet. It is an environment where the level of the liquid is measured,
monitored and controlled.
The human operator who is performing the functions of measuring the
level, comparing with the set point, evaluating and opening or closing of
control valve based on results of evaluation is called as Controller.
The control valve is known as Final Control Element.
To change the manual control system into an automatic process control
system, the human operator is replaced with automatic controller.
Let us study and analyse an automatic example of liquid level
measurement and control as shown.

Automatic Level Control System

Set point

Inlet (qi)
Automatic
Controller
LT
h Final control element
qo
(outlet)

FIG. 1.2

Figure1.2 refers to automatic process control system for regulating the

level of the liquid in the tank. Generally, an automatic controller is an
electronic device, and understands only electrical signals. Since its input is
information about level which is non-electrical in nature, it has to be converted
into electrical form so that the automatic controller understands and performs
other related functions.
Therefore, in automatic control systems, measurement of variables is
usually done by devices which are capable of converting non-electrical
quantities into electrical quantities. These devices are known as Transducers.
LT, therefore, is a level transducer. It converts level of liquid in the tank into
proportional electrical signal. This signal is received by controller which
compares with set point and opens or closes the final control element (control
valve) so that the control objective is met. Hence, a transducer may be defined
as a device, which, when actuated with one form of energy is capable of
converting it into any other related form. In automatic process control systems,
measurement constitutes application of transducers.
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

4 Process Control Engineering

Element of Process Control Loop

The description of process control system may be represented in more generic
terms for understanding the sequence of operations involved. A block
diagrammatic approach would provide a more suitable representation.
The various elements constituting the process control loop may be
understood by considering the same liquid level control system shown in
Figure 1.2.

Block diagram of liquid level control system

Controller block

Comparator
Input or
Automatic Final Control Output
Measurand or Process
– Controller Element (Level)
Set point

Measurement

FIG. 1.3

Terminology of Process Control

The various elements of process control loop are shown in Figure 1.3. Apart
from the different blocks showing different elements it is important to get
acquainted with certain terms usually encountered in the language of process
control.
• Process: As defined earlier, it refers to physical or chemical change of
matter or an ensemble of environment where a particular variable is
measured, monitored and controlled. In the level control system,
process refers to ensemble of tank, liquid in the tank, the inlet and
outlet.
• Process variables: They are any variable property of a process. Process
variables may be further classified as follows:
a. Controlled variable b. Manipulating variable
c. Load variable
Controlled variable is that variable in a process, which is measured,
monitored and controlled. The level in the tank is an example of controlled
variable. The controlled variable is also known as dynamic variable.
Manipulating variable is that variable in a process, which is manipulated to
make the controlled variable remain at the set point value. The manipulating
variable in level control system is outlet flow rate.
Load variables in a process refer to set of all other variables that affect the
controlled variable to deviate from set point value. In the example of Fig. 1.2,
apart from outlet flow rate, the changes in inlet flow rate also change the level
in the tank. It constitutes a load variable.
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 5

Process Load
The set of all load variables that cause a change or deviations of controlled
variable from set point is known as process load. Process load excludes the
controlled variable. Any one of the load variable may be made a manipulating
or controlling variable.

Process Equation
In the example shown in Fig. 1.2, a process equation may be written which
describes a process mathematically. The liquid level ‘h’ is a function of changes
in inlet flow rate and outlet flow rate. A simple process equation may be written
as
level (h) = f(qi, qo)

Measurement
It is the determination of the existence (or) the magnitude of a variable i.e.
controlled variable. A transducer is normally employed for the measurement.

Controller
It is an automatic device having an output that varies to regulate a controlled
variable in a specified manner. A controller performs operations such as
comparison, evaluation and operation upon the next stage to meet the
demands of control objectives.
A controller may be a self-contained analog or digital instrument, or it may
be the equivalent of such an instrument in a shared control room.

Final Control System

The device that directly controls the value of manipulated variable in a process
control Loop. Often the final control element is a control valve.

Flow Control System

Set point

4-20 mA Automatic 4-20 mA

Controller
Pressure to Current to
Current P/I I/P Pressure
Converter Converter
3-15 PSI
D.P. Cell 3-15 PSI
Inlet Outlet
Pipe line
flow rate flow rate
(qi) Flow (qo)
Orifice Control valve

FIG. 1.4
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

6 Process Control Engineering

Figure 1.4 shows the physical diagram of Flow Control System. The control
objective is to control the fluid flow rate and obtain constant outlet flow rate,
depending on the set point value. The different load variables of flow control
loop are inlet flow rate (qi) and outlet flow rate (qo).
Since there are only two variables in which outlet flow rate is monitored
and controlled, a simple process equation can be written as
qo = f [qi]

Operation
A orifice plate inserted in the pipe line causes a pressure differential. This
differencial pressure is measured by a differential pressure transmitter (D.P.
Cell) which is a pneumatic device. Since its output is pressure proportional to
the fluid flow rate, it has to be converted into proportional electrical signal if
the controller is an electronic device. Hence, a pressure to current converter is
employed, which converts, pressure into current. The output of the controller,
after evaluation of the measured value by comparing it with set point, is further
converted into a pressure signal so that it can be applied to control valve
actuator, which positions the valve stem to regulate flow rate at the outlet of the
pipe line. It is to be noted that the standard range of 4-20 mA, and the
standard pneumatic signal transmission is 3-15 psi.
Control system evaluation and analysis can be done by considering the
block diagram of above flow control system, in addition to knowing about
elements of flow control loop.

Block Diagram
Controller block
Comparator
Set point Process Flow rate
Automatic I/P Control (Pipe line inlet
– Controller Converter Valve & outlet fluid)

P/I Measurement
Converter (Orifice)

FIG. 1.5

Process constitutes the pipe line, fluid, inlet and outlet. Measurement is
done by orifice plates and electrical signal conversion is done by P to I
converter.
TEMPERATURE CONTROL SYSTEM
A simple temperature control system is showing in Fig.1.6. The control
objective is to maintain the temperature of the liquid at a desired value. The
controlled process variable is therefore TL i.e. temperature of liquid in tank.
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 7

Inlet (qi, Ti)

TA

Hot steam TL
qs , T s qo
(outlet)
Control (T T)
valve

Automatic
Controller

Set point

FIG. 1.6

The different load variables that change the temperature TL are

qi = Inlet flow rate
Ti = Temperature of the liquid of inlet
qs = Steam flow rate
Ts = Temperature of steam
qo = Outlet flow rate
TA = Ambient temperature
Thus, the above load variables constitute process load. Out of these load
variables, the manipulating variable affecting the temperature of liquid in tank
(T L) is chosen to be steam flow rate. A process equation describing the system
may be written as
TL = f (qi, Ti, qs, Ts, qo, TA)
which implies that TL is a function of process load. When all the process load
variables have their nominal values, then the system is said to be having
nominal load and the temperature TL of the liquid is at the nominal set point
value. If anyone of the load variable deviates from the nominal value, then the
temperature of the liquid in the tank also changes.

Operation
The temperature transmitter (TT) which is a temperature transducer measures
the temperature of the liquid in the tank. The measured value is sent to the
controller, which compares it with the set point. After proper evaluation, it
sends a signal to the control valve which then regulates the steam flow rate
depending upon the requirement. The standard signal transmission should be
maintained in all process control applications. The output of the temperature
transducer and that of the controller should always be in standard signal
transmission range of 4-20 mA and 3-15 psi respectively.
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

8 Process Control Engineering

Block Diagram

Controller block
Comparator
Set point Process Output TL
Automatic Control Tank, liquid, inlet
– Controller Valve outlet, steam flow

Temperature
Transmitter
(Transducer)

FIG. 1.7

Pressure Control System

Set point

Controller

Pressure Current to
Sensor I/P
Pressure Converter
Pressure
Vessel Control valve
Gas inflow (qi)
Gas outflow (qo)

FIG. 1.8

Set point Automatic Final control Process Pressure (P)

Controller & Element (Vessel, inlet, outlet
– & gas in the vessel)
I/P converter (Control valve)

Measurement
(Pressure
Transducer)

FIG. 1.9

A pressure control system is shown in Fig. 1.8. The control objective is to

maintain the pressure inside the vessel in accordance with the set point. A
pressure sensor measures the value of pressure inside the vessel, and sends the
signal to the controller. The measured value is compared with the set point and
after proper evaluation the controller output is sent to final control element
through current to pressure converter. The gas outflow rate is then
manipulated to meet the demands of control objectives.
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 9

A process equation may be written

P = f (qi, qo)
where P = pressure of gas inside the vessel.
Process is an ensemble of vessel, inlet, outlet and the gas.

1.3 ELEMENTS OF PROCESS DYNAMICS

In the previous section we have learnt about process characteristics in terms of
different elements describing the complete process control system. In this
section, we learn about the dynamics of process systems.
The mathematical modelling, and theoretical analysis of processes or
process control systems depend on certain dynamics that describe a process.
Therefore, the different elements with which a mathematical model may
be formulated for a process are
a. Resistance Element
b. Capacitance Element
c. Time Constant Element
d. Oscillatory Element
e. Dead Time Element
The mathematical model thus obtained is also known as Transfer function.
An important mathematical tool known as Laplace transforms are employed
for writing transfer functions of physical processes in terms of above
mentioned dynamics.
A transfer function may be defined as ratio of L.T. of output to L.T. of
input under the assumption that all initial conditions are zero i.e. the system is
assumed to be initially relaxed.

Resistance Type Element

Figure 1.10 illustrates the operation of capillary flow system. Two capillary
tubes are inserted at the inlet and outlet sections of the pipe line where the
fluid is flowing. The difference in the head or height (h) between the rise of
fluid in both capillary tubes is because of ‘Resistance’ effect.
Resistance element thus implies any obstruction or constriction which can
be incorporated by external means also.
The flow-head equation for Figure 1.10 may be written as follows.
(head) h ¥ q (Flow rate)
h = Rq ...(1)
Thus if q = input and h = output (1) represents a linear relation between
input and output. Applying L.T to equation (1)
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

10 Process Control Engineering

h

h
Capillary (1) Capillary (2)

Flow rate q fi
of fluid Pipe line

FIG. 1.10
* Pressure drop through pipes and other equipment is the most obvious illustration of
resistance type element.

H(s) = RQ(s)

Q(S) H(S)
R

The resistance elements is also known as proportional element. Electrical

resistance, gas flow resistance thermal resistance constitute other resistance
elements or proportional elements.
In mechanical translational or rotational systems, the spring element
constitutes proportional element.
The turbulent resistance is found from Bernoulli’s equation

q = KA 2 g( h1 - h2 ) ...1(a)
3
q = liquid flow rate (m /s)
K = Flow constant
A = Area of restriction (m2)
g = Acceleration due to gravity (m/s2)
h = head of fluid (m)
Squaring equation 1(a)
q2 = K 2 A 2 2g(h1 – h2)
Þ q.q = K 2 A 2 2g(h1 – h2)

q 2 ( h1 - h2 )
2 2
= ...1(b)
gK A q

h dh
From equation (1) Resistance R = (or) R = [Rate of change of head
q dq
with flow]
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 11

\ Eqn. 1(b) represents resistance for Turbulent flow

dh 2 ( h1 - h2 ) q
R= = =
dq q gK 2 A2

For laminar flow, the poiseville-hagen law states

128 NL 128 VL
h1 – h2 = 4
×q = ×q
πrD g π D4

where
h = head (m)
v = kinematic viscocity (m2/s)
D = Inside distance of pipe (m)
q = Inlet flow rate (m)
m = Absolute viscosity
g = Fluid density (kg/m3)
laminar flow is generally not encountered in industrial applications.
However, the resistance element for laminar flow may be defined as

h dh 128 VL
R= = = ...1(c)
q dq g π D4

Capacitance Element
Capacitance is the ability to store. An illustration of
capacitance element is a tank with inlet as shown in
Figure (1.11). h
The flow of the fluid into the tank is the input and q
the level of the liquid in the tank (h) is the output.
FIG. 1.11
The ability of the tank to store liquid is the
capacitance.
The changes in the level of liquid in the tank (dh/dt) is directly proportional
to inlet flow rate

dh
¥q
dt
dh
Þ C =q ...(2)
dt
where C = capacitance
Applying L.T. to equation (2)
CS H (S) = Q(S)
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

12 Process Control Engineering

Transfer function
H (S) 1
=
Q ( S) CS

Q(S) 1 H(S)
CS

Hence the transfer function represents of capacitance element is (1/CS)

Electrical capacitance, Gas capacitance, thermal capacitance are other forms of
capacitance element symbolising storage.
The response of a capacitance element is obtained by integrating
equation (2).

z dh
dd
dt = z q
c
dt
Level
(h)
(q/c) t

If the inflow is assumed to be constant, then t

q
h= t ...(3)
c

Time Constant Element

A combination of resistance and capacitance elements result in a time constant
process.
Those parts of the process that have the ability to store energy or mass are
termed as capacitance element and those parts that resist the transfer of mass
or energy are termed as resistance, as learnt earlier. Hence, in a process if there
is a combined effect of both these element then it constitutes time constant
process.
Hence in a process if there is a combined effect of both these elements
then it consitutes time constant process.
Consider a liquid level system as shown. The capacitance is the ability of
the tank to store fluid and resistance element is introduced at the outlet in the
form of control valve.
The amount of liquid stored in the tank is directly proportional to net flow
rate.
Inlet (qi)

C R
h
qo
(outlet)
Control
Valve

FIG. 1.12
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 13

Mathematically,
dh
q i – qo µ
dt
(net flow rate) µ (changes in level)
(liquid stored in the tank)
dh
Þ qi – qo = C
dt
dh
qi = C + qo
dt
qo = outlet flow rate = h/R (from equation(1))
dh h
\ qo = C +
dt R
Pre multiplying both sides by R
dh
Rqi = RC +h ...(4)
dt
Applying L.T. to equation (4)
R.Q i (s) = RCS H(S) + H (S)
R.Q i ( s) = H(S) (RCS + 1)
Defining time constant element = RC = T
R.Q i (s) = (TS + 1) H (s)
H(S) R
\ = ...(5)
Q i (S) TS + 1
Equation (5) represents the transfer function of Time Constant Process ‘R’
in the numerator constitutes proportional element.

Qi(S) R H(S)
(TS + 1)

Significance of Time Constant

A time constant element is defined as a product of resistance and capacitance.
The combined effect ‘R’ and ‘C’ in a process can be understood by considering
a sudden change in the input flow rate Q i (S).
A sudden change may be mathematically simulated as step input.
1
Let Q i (S) = (unit step)
S
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

14 Process Control Engineering

R
H(S) = Q i (s)
TS + 1

F 1 - T I (partial fractions)
H(S) =
R
S (TS + 1) GH S 1 + TS JK
=R

F1
H(S) = R G -
I
H S T (S + 1 / T ) JK
T
\

F1 1 I
H(S) = R G -
H S S + 1 / T JK
Applying Inverse L.T.
h(t) = R(1 – e – t /T ) ...(6a)

R(1.0)
R(0.98)
R(0.86)
R(0.63)

t
t = 0 t = 1t t = 2t t = 4t

FIG. 1.13

For t = 1T; h(t) = R (1 – e–1) = R(1 – 0.36) = R(0.64)

t = 2T; h(t) = R(1 – e–2) = R (1 – 0.13) = R(0.84)
t = 4T; h(t) = R (1 – e–4) = R(1 – 0.12) = R (0.98)
As seen from the characteristics, the response of time constant process to a
sudden input or disturbance is sluggish. It exponentially reaches the steady
state value of input. It is because of the combined effect of resistance and
capacitance effect.
A time constant element may be, thus, mathematically defined as “Time
taken by the response to reach 63% of the final value”.
Further, it may be noted that presence of time constant element in a
process, makes the process a first order process.
The transfer function of time constant process can be compared to
standard first order transfer function expressed as
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 15

C(S) K
=
R(S) 1 + TS

where C(S) = Output; R(S) = Input

Oscillatory Element
Another important element, which may not be generally encountered in
ordinary liquid, gas and thermal system is oscillatory element.
This element is a typical characteristic of higher order systems. It can be
proved that the response of second order or higher order processes or systems
exhibit oscillations about the steady state value of input. By using appropriate
damping methods these oscillations can be damped to steady state value of
input. A typical example of second order system is PMMC galvanometer
(permanent magnet moving coil). Its principle of operation is based on
“Principle of suspension galvanometer” which states that a coil of fine wire
suspended between the poles of a permanent magnet with current passing
through it deflects in the magnetic field because of the fundamental property
of electromagnetic force or torque. It continues to deflect until the
electromagnetic torque also known as deflecting torque is counter balanced by
mechanical controlling torque, which is produced due to control springs. At
the equilibrium position when both the torques are equal and opposite in
magnitude, they cancel and the pointer attached to the moving coil, moving
over a calibrated scale, should indicate the magnitude of unknown current
passing through it. But the pointer continues to oscillate about the final value
on account of insufficient friction inherent in the moving parts of the system.
Therefore additional friction is generated by using different damping
methods and this torque is known as damping torque. It is on account of this
damping torque that the pointer rests at the final value indicating the
magnitude of current at the equilibrium position.
Therefore, we conclude that a second order system is characterised in
terms of an important factor called as Damping Factor.
To illustrate the behaviour of a oscillatory element consider a mechanical
translational system as shown.
When a force (F) is applied to this mechanical system, the mass, friction or
damper and spring oppose the applied force. From the principle of
translational mechanics
F = F1 + F2 + F3

dx2 dx
Since F1 = M 2
; F2 = f ; F3 = K x
dt dt

d2 x dx
F=M 2
+f +K x ...(6)
dt dt
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

16 Process Control Engineering

Spring (K) Damper (f)

F3 F2

Mass (M) F1 Displacement (x)

Force (F)

FIG. 1.14

Equation (6) represents a second order differential equation, hence the

system is a second order system.
Applying Laplace transform to equation (6)
F(S) = MS2 X (S) + f S X (S) + K X (S)

X ( S) output 1
\ Transfer function = = ...(7)
F ( S) input MS2 + f S + K

As we are aware that any second order transfer function is characterised in

terms of damping factor “x” (Geta) and Wn (undamped natural frequency
(radians/second)) and is expressed as

C(S) output Wn2

= = 2 ...(8)
R(S) input S + 2 ξ Wn S + Wn2

We can compare equations (7) and (8) and obtain expressions for damping
factor ‘x’and undamped natural frequency ‘W n’ in terms of mass (M) spring (K)
and damper (f).
Rewriting equation (7) as

X ( S) 1/ M
= 2 ...(9)
F ( S) S + ( f / m) S + K / M
Comparing the characteristic equation of equation (8) and (9)
f K
S2 + 2 x W n S + W n2 = S2 + S+
M M
K
Wn2 =
M

K FG rad IJ
Þ Wn =
M H sec K
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 17

f
2xWn =
M

K f
\ 2x =
M M
Squaring both sides

4 ξ2 K f2
=
M M2

f2
x2 =
4K M

f
Þ x=
2 KM

Similarly an R-L-C network also represents second order system exhibiting

oscillatory characteristics.

R LS

V(S) I(S) 1
CS

Applying KVL to the loop

V(S) = I(S)R + I(S)×L (S) + I(S)× 1/CS
I(S) (R + LS + 1/CS)

F LCS 2
+ RCS + 1 I
V(S) = I(S) GH CS JK
I(S) CS
\ Transfer function = 2
V (S) LCS + RCS + 1
Rewriting it as
I( S) (1 / L)S
= 2 ...(10)
V ( S) S + ( R / L)S + (1 / LC)
Comparing characteristic equation of (8) and (10)
S2 + 2 x W n S + W n2 = S2 + (R/L)S + (1/LC)
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

18 Process Control Engineering

1
Wn2 =
LC

1 FG rad IJ
Þ Wn =
LC H sec K
R
2xW nS =
L
Squaring both sides

4 ξ2 R2 R2C
= 2 Þ x2 =
LC L 4L

R C
x=
2 L

To understand oscillatory behaviour of these systems we can consider the

analysis of standard second order transfer function, by subjecting it to a
sudden input or disturbance and observing the response.
Firstly, we analyses the different types responses of second order system
which depend on the roots of the characteristic equation.
The roots of the characteristic equation are S2+ 2 x W n S + W n2 = 0

- b ± b2 - 4 ac
S1, S2 =
2a

- 2 ξ Wn ± 4 ξ2 Wn2 - 4Wn2
\ S1, S2 =
2a

= – x Wn ± Wn ξ2 - 1 ...(11)

Therefore, the discriminant is ξ2 - 1 . This implies that the nature of

roots and hence the corresponding response depend on the damping factor.
Hence, we say, that damping is an important aspect in second order system
analysis. We shall analyse, what values of damping result in oscillatory
behaviour, which is our main concern and derive a relation for oscillatory
response, thus showing the oscillatory behaviour of second order systems.
Case-1: Undamped Case (x = 0)
(There is no damping)
In equation (11) if (x = 0), the roots are lying on ± j W n i.e. imaginary axis
of S-plane.
 5TH PROOF 3-9-09
COM -6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

Introduction 19

t
jwn
Wn(r/s)

value of input
Steady state
–jwn

The response is continuous oscillations about steady state value with

frequency of oscillations ‘W n’ rad/sec i.e. undamped natural frequency.
Case-2: Under Damped Case 0 < x < 1
If we put any value of ‘x’ between 0 and 1 in equation (11) the roots lie in
second and third quadrant of S-plane both real and imaginary parts. The
corresponding nature of response would be

Input and output Nature of response

The response is a Damped Sinusoid. It has damped oscillations as shown

in figure.
Case-3: Critically Damped Case (x = 1)
with (x = 1), equation (11) reduces to repeated roots at –Wn.
Input and output

–Wn

The response is quickly rising exponential without any oscillations.

Case-4: Over Damped Case (x = 1)
The roots for over damped case, by putting any value of (x > 1) in equation
(11) are real and unequal.
The response is very sluggish, rising exponentially to the steady state value
without any oscillations.
From the four types of responses corresponding to different values of
damping factor, oscillatory characteristics are exhibited by underdamped
 5TH PROOF 4-9-09
COM-6\D-DRIVE\I.K. INTERNATIONAL\PROCESS CONTROLL ENGINEERING

20 Process Control Engineering

Input and output

t

response only. Therefore a second order system is typical of oscillatory element

only when dapping factor ‘x’ is less than one.

Oscillatory Response
Consider a second order system with transfer function.

C(S) Wn2
= 2
R(S) S + 2 ξ Wn S + Wn2
Let a sudden input be applied to this system
1
Hence R(S) =
S

1 (S + 2 ξ Wn )
C(S) = – 2
S S + 2 ξ Wn S + Wn2

Rewriting – (S + 2xW n) as
– S – x Wn
+ – x Wn
– S – 2x Wn
and the denominator
S2 + 2 x W nS + Wn2 = (S + x W n)2 – x2 W n2 + W n2
= (S + x W n)2 + Wn2(1 – x2)

Defining Wd = W n 1 - ξ 2 Damped frequency of oscillations (rad/sec)

S2 + 2 x W n S + W n2= (S + xWn )2 + W d2
1 (S + ξWn ) ξWn
\ C(S) = – –
S (S + ξWn )2 + Wd2 (S + ξWn )2 + Wd2
We know that
1
L –1 =1
S
 5TH PROOF 4-9-09

Introduction 21

(S + a )
L –1 = e –a t cos wt
(S + a )2 + W 2

w
L –1 = e –a t sin wt
( s + a )2 + w2

1 (S + ξ Wn ) Wd ξ Wn
\ C(S) = – 2 2
– 2 2
´
S (S + ξ Wn ) + Wd (S + ξ Wn ) + Wd Wd

1 (S + ξ Wn ) Wd ξ Wn
C(S) = – – ´
S (S + ξ Wn )2 + Wd2 (S + ξ Wn )2 + Wd2 Wn 1 - ξ2

1 (S + ξ Wn ) Wd ξ
C(S) = – 2 2
– ´
S (S + ξ Wn ) + Wd (S + ξ Wn )2 + Wd2 1 - ξ2

Applying inverse Laplace transform

ξ
C(t) = 1 – e -ξ Wn t cos W dt – e -ξ Wn t sin W dt ´
1 - ξ2

e - ξ Wn t FG x IJ
C(t) = 1 –
1- ξ 2 H 1 - x 2 cos Wd t +
A
sin Wd t
K
We know that “ Sum of sine wave and cosine wave is another sine wave with
different magnitude and phase angle”

FG B IJ
A Sin wt + B Cos wt =
H
A2 + B2 Sin wt + Tan -1
A K
where A=x

B = 1 - ξ 2 as shown

e - ξ Wn t F 1 - ξ2 I
\ C(t) = 1 – GG
sin W t + Tan -1 JJ ...(12)
H K
d
1 - ξ2 ξ

The equation (12) represents an oscillatory response. It has two parts

(a) Exponential term
(b) Sine term.
 5TH PROOF 4-9-09

22 Process Control Engineering

C(t)

1.0

Frequency of
Damped
Oscillations
} = wd r/s

FIG. 1.15 Oscillatory response of second order system

DEAD TIME ELEMENT

A phenomena often encountered during transfer of mass or energy is
transportation lag. Synonyms for this term are Dead Time (or) Distance
Velocity Lag. Actually, dead time is an inherent characteristic of any system. It
is defined as “The time elapse between the instant system subjected to an input
and its response, beginning to occur” For some systems it is less than micro
seconds.
It’s study as a physical phenomena is in terms of transportation lag.
Consider the following system where hot water is to be passed through a
tube having uniform cross sectional area.

Temperature indicator
T1 T2
Tank

Pipe line

Heating element

FIG. 1.16

Let x(t) be the temperature indicated by “T1” and y(t) be the temperature
indicated by “T2”.Under steady state conditions
x(t) = y(t)
Suppose, if the temperature of water at “T 1” is suddenly changed. This
change would be indicated at T2 only after an elapse of “T” unit of Time by the
fluid to flow from T1 end to T 2 end. It is known as transportation lag. Since the
temperature indicated by “T2” is delayed by “T” units of time
y(t) = x(t – T) ...(13)
 5TH PROOF 4-9-09

Introduction 23

Graphically

Curve - 1

x(t) & y(t)

Curve - 2

t
T

Applying Laplace Transform to equation (13)

Y(S) = e –TS X(S) ...(14)

Y ( S) X(S) Y(S)
= e –T S Þ e –TS
X ( S)

Therefore the representation of dead time or transportation lag is e –TS .

APPROXIMATIONS FOR DEAD TIME EXPRESSION
The dead time transfer function is quite different from other transfer
functions. Lide first order, second order etc. For the purpose of analysis,
several approximations of dead time are considered, if this element is present
in a system.
Non-M inimum phase Approximation:
We know that Y(t) = x(t – T)
Y(s) = e –TS X (S)
Considering X(t – T) and expanding using Taylor Series.
Neglecting the higher order terms
Y(t) = x(t) – T x(t)
Y(S) = X(S) – TS X (S) (Applying L.T.)
Y(S) = X(S) (1 – TS) ...(15)
Comparing (14) and (15)
e –TS » 1 – TS
Thus, dead time as a non-minimum phase function represents a zero in
RHS of S-plane.
FIRST ORDER PADE’ APPROXIMATION
The dead time element is
Y ( S)
F(S) = = e –TS
X ( S)
 5TH PROOF 4-9-09

24 Process Control Engineering

1
Writing e –T S =
eTS
Expanding using Taylor Series

1
e –T S =
FT S I + FT S I
2 2 3 3
(1 + TS) + GH 2! JK GH 3! JK
Keeping only the first two terms
1
e –T S = ...(16)
1 + TS
Equation (16) is a crude approximation of dead time. An improvement in
equation (16) is known as Pade’ approximation.

e -TS /2 1 - TS / 2
e –TS = TS /2
= ...(17)
e 1 + TS / 2
Equation (17) is known as 1st order Pade’ approximation.
SECOND ORDER PADE’ APPROXIMATION

e -TS /2 1 - (TS / 2) + (T 2 S2 / 12)

e –TS = = ...(18)
eTS /2 1 + (TS / 2) + (T 2 S2 / 12)
Equations (15) to (18) are time domain approximations of dead time.
However, the accuracy of transport lag can be evaluated most clearly in
frequency domain.
FREQUENCY DOMAIN ANALYSIS OF DEAD TIME
We known that
e – j q = cos q + j sin q
e – j q = cos q – j sin q
Y ( S)
Therefore, let F(S) = = e –TS
X ( S)
In frequency domain put S = jw
\ F( jw) = e – j w T
e – j w T = cos wT – j sin wT ...(19)
Expressing F(jw) in polar form
F ( jw) = |F(jw)||F ( jw)

|F ( jw)| = ( real part )2 + ( imaginary part )2

 5TH PROOF 4-9-09

Introduction 25

F Imaginary part I
|F(jw)| = tan–1 GH real part JK
\ |e – j w T | = (cos wT )2 + ( - sin wT )2 = 1

F - sin wT I
|e - jwt = + tan–1 GH cos wT JK
= tan–1(– tan wT)
= –wT rad
Since p radians ® 180°
– wT radians ® ?
-wT
\ ´ 180°
π
Þ – 57.3 wT degrees
Therefore we conclude that the magnitude of dead time is unity. It effects
only the phase angle by contributing –57.3 wT degrees

e – j w T = |e – j w T | |e - jwT = |-57.3 wT ...(20)

MATHEMATICAL MODELS FOR SIMPLE SYSTEMS BASED ON

PROCESS DYNAMICS (DYNAMICS OF SIMPLE SYSTEMS)

Liquid Level Systems

The mathematical model of a simple liquid process shown in Fig. (1.11) is
given by equation (5).
In this section we will consider, the liquid level process different load
variables.

Inlet (qi)

C R u
h
qo
(outlet)

FIG. 1.17

qi = inlet flow rate

qo = outlet flow rate
h = level of liquid in tank
 5TH PROOF 4-9-09

26 Process Control Engineering

R = fluid resistance
C = fluid capacitance or hydraulic capacitance
u = down stream head
dh
qi – qo ¥
dt
dh
qi – qo = C
dt
dh
qi = C + qo
dt
Since down stream head is given
h-u
qo =
R
dh h-u
\ qi = C +
dt R
Multiplying both sides by R
dh
Rqi = RC +h–u
dt
dh
Rqi + u = RC +h
dt
Let RC = Time constant = T
dh
Rqi + u = T +h ...(21)
dt
Applying L.T. to equation (21)
Rqi(S) + U(S) = (TS + 1) H(S)
R 1
Þ H(S) = Qi(S) + U(S) ...(22)
TS + 1 TS + 1

U(S) 1
TS + 1

Qi(S) R + H(S)
+
1 + TS

While analysing the behaviour of the system we consider either Qi(S) or

U(S) as input for this process.
Consider the process shown in Figure 1.18.
 5TH PROOF 4-9-09

Introduction 27

In this case u = Inlet flow rate qi u

dh
qi + u – qo = C
dt
C R
h
dh qo
qi + u = C + qo
dt
FIG. 1.18
h
3 qo =
R
dh h
\ qi + u = C +
dt R
Multiplying both sides by ‘R’
dh
Rqi + R.u = RC +h ...(23)
dt
Applying L.T. to equation (23)
RQi(S) + RU(S) = (RCS + 1) H(S)
T = time constant = RC
RQi(S) +RU(S) = (TS + 1) H(S)
R R
H(S) = Q(S) + U(S)
TS + 1 TS + 1

U(S) R
TS + 1

Qi(S) R + H(S)
+
TS + 1

Thermal Systems
Let “hi” be heat input rate and “q ” be the temperature of the liquid at the
outlet.
The mathematical description of the above system can be obtained by
expressing heat input rate “hi” in terms of two quantities.
hi = h1 + h2
h1 = Amount of heat stored in tank
h2 = Amount of heat responsible for temperature “q ”
 5TH PROOF 4-9-09

28 Process Control Engineering

Inlet

Stirrer

q = outlet flow rate

Outlet

hi
(Temperature of
liquid at outlet)

Heat input rate

FIG. 1.19

dq
h1 = MS
dt
where M = Mass of liquid (kg)
S = Specific heat J/kg-°C
Defining thermal capacitance C = MS
dθ
\ h1 = C and h2 = q s q.
dt
q = outlet flow rate (kg/min)
q = temperature of outflowing liquid
Defining Thermal Resistance
1
R=
qs

q
\ h2 =
R
dq θ
hi = c
+
dt R
Multiplying both sides by R
dq
Rhi = Rc +q ...(24)
dt
Applying L.T. to equation (23)
 5TH PROOF 4-9-09

Introduction 29

R.H i(S) = (RCS + 1) q(S)

Let T = RC
R.H i(S) = (TS + 1) q(S)
q(S) R
=
H i ( S) TS + 1

Hi(S) R q (S)
TS + 1

Pressure (or) Gas Systems

R C C = Gas capacitance
Pi Po R = Gas resistance
Gas inflow

Pressure vessel

FIG. 1.20

Rate of gas storage in vessel = Rate of gas inflow

C d Po P - Po
= i
dt R
C d Po P P
+ o = i
dt R R
Multiplying both sides by R
dPo
RC + Po = P i ...(25)
dt
Applying Laplace transform to (25)
Pi(S) = P o(S) (RCS + 1)
T = RC
Pi(s) = Po(S) (TS + 1)
Po (S) 1
=
Pi (S) TS = 1

Pi(S) 1 Po(S)
TS + 1
 5TH PROOF 4-9-09

30 Process Control Engineering

1.4 INTERACTING AND NON-INTERACTING

SYSTEMS
Several systems may be connected in series to study the loading effects and
overall performance of such combination. Two distinct representations to
study the loading effect are:
(a) Interacting system
(b) Non-Interacting System

Non-Interacting System
qi

C1 R1
h1

q1
Tank-1

C2 R2
h2
q0
Tank-2

FIG. 1.21

Consider two tanks (1) and(2) connected as shown in Fig. (1.21). This type
of connection is known as non-interacting system. Any changes in level ‘h’, of
tank-1, will change the level ‘h2’ of tank-2, but the converse, is not possible i.e.
any changes in level ‘h2’ does not affect the level in tank-1. This type of system
is referred to as non-interacting system.
Analysis:
Consider tank-1
dh1
qi – q1 = C 1
dt
dh1
qi = C1 + q1 ...(26)
dt
h1
q1 = ...(27)
R1

dh1 h
\ qi = C1 + 1
dt R1
 5TH PROOF 4-9-09

Introduction 31

dh1
R i q i = R1C1 + h1 ...(28)
dt
Consider tank-2
dh2
q1 – qo = C 2
dt
dh2
q1 = C 2 + qo
dt
h2
qo =
R2

dh2 h
\ q1 = C 2 + 2
dt R2

dh2
R 2q 1 = R 2C 2 + h2
dt
Let T2 = R 2 C 2
dh2
R 2 q 1 = T2 + h2
dt
h1
Since q1 = (from equation (27))
R1

R2 dh
h1 = T 2 2 + h2 ...(29)
R1 dt
Applying L.T. to equations (28) and (29)
R 1 Q i (S) = (T1S + 1) H1(S)
R2
H1(S) = (T2S + 1) H2(S)
R1
The overall input = Qi(S)
The overall output = H2(S)

Qi(S) R1 H1(S) R2/R1 H2(S)

\ T 1S + 1 T 2S + 1

Therefore the overall transfer functin of non-interacting system is

H2 (S) R2
= ...(30)
Q i ( S) (T1S + 1) (T2S + 1)
 5TH PROOF 4-9-09

32 Process Control Engineering

From equation (30), the overall transfer function is the product of transfer
functions of individual tanks.

Interacting System

Inlet
qi

R1 R2
h1 C 1 h2 C 2
q0

Tank-1 q1 Tank-2

FIG. 1.22

Figure 1.22 represents an interacting system. It is so called because any

changes in level ‘h1’ of tank-1 change the level ‘h2’ in tank-2 and vice versa. The
loading effect is more predominant in interacting system rather than in non-
interacting system.
Hence, the term interaction is also sometimes referred to as Loading.
Analysis:
For tank-1
dh1
qi – q1 = C 1 ...(31)
dt
h1 - h2
q1 = ...(32)
R1
For tank-2
dh2
q1 – qo = C 2 ...(33)
dt
h2
qo = ...(34)
R2
Applying Laplace transform to equations (31) to (34)
Qi(S) – Q1(S) = C1S.H1(S) ...(35)
H1(S) - H2 (S)
Q1(S) = ...(36)
R1
Q1(S) – Qo(S) = C2(S).H2(S) ...(37)
H2 (S)
Qo(S) = ...(38)
R2
 5TH PROOF 4-9-09

Introduction 33

Developing a block diagram from equation (35) to (38)

Overall input = Qi(S)
Overall output = H2(S)

H2(S)

Qi(S) 1 H1(S) – 1 Q1(S) 1 H2(S)

– C1S R1 – C2S

Q1(S) Q0(S)
1
R2

Feed back

Simplifying the feedback loop indicated in the block diagram

1 1
C2S C2S
Þ
1 R2C2S + 1
1+
R2C2 S R2C2S

R2
Þ
R2C2S + 1

Qi(S) 1 H1(S) – 1 R2 H2(S)

– C1S R1 R–2C2S + 1

Shifting the take off point after the block.

Feed back

Qi(S) 1 – R2 H2(S)
– C1S R1–R2C2S + 1

R2C2S + 1
R2

Simplifying the feed back loop

R2
R1 R2 C2S + R1
R2
1+
R1 R2C2 S + R1
 5TH PROOF 4-9-09

34 Process Control Engineering

R2
Þ
R1 R2C2S + R1 + + R2

Qi(S) 1 R2 H2(S)
C1S R1R2C2S + –R1+ R2

R2C2S + 1
R2

Simplifying the feed back loop

R2
2
H2 (S) R1C1 R2C2S + R1C1S + R2C1S
=
Q i ( S) R2 R C S +1
1+ ´ 2 2
R1C1 R2C2S2 + R1C1S + R2C1S R2

R2
Þ 2
R1C1 R2C2 S + R1C1S + R2C1S + R2C2S + 1
Therefore the transfer function of interacting system is
H2 (S) R2
= 2
...(39)
Q i ( S) R1C1 R2C2S + S( R1C1 + R2C2 + R2C1 ) + 1
As seen from equation (39), unlike the non-interacting system, the transfer
function of interacting system is not the product of transfer transfer functions
of individual tanks.

Comparisons of Transient Analysis

Let the transfer function of non-interacting system be
H2 (S) 1
= (Assume R2 = 1)
Q i ( S) (T1S + 1) (T2S + 1)
Consider the time constants to be equal. T1 = T 2 = T
H2 (S) 1
\ =
Q i ( S) (TS + 1)2
To understand the transient response, subjecting the system to sudden
input.
1
H2(S) = Qi(S) ´
(TS + 1)2

1
Let Qi(S) = (unit step input)
S
 5TH PROOF 4-9-09

Introduction 35

1
H2(S) =
S(TS + 1)2
Considering the partial fractions
1 A B C
2
= + +
S(TS + 1) S TS + 1 (TS + 1)2

A=1 B = –T C = –T
1 T T
\ H2(S) = – –
S TS + 1 (TS + 1)2

1 1 T
H2(S) = – – 2
S S +1/T T (S + 1 / T )2
Applying Inverse Laplace Transform

t –t/T
h2(t) = 1 – e –t/T – e ...(40)
T
Now consider the transfer function of interacting system
let R1 = R2 = R
C1 = C2 = C and T= RC
H2 (S) R
\ = 2 2
Q i ( S) T S + 3TS + 1
Let the proportional element in the numerator R = 1
H2 (S) 1
= 2 2
Q i ( S) T S + 3TS + 1
The factor of denominator polynomial are

- 3T ± 9T 2 - 4T 2
=
2T 2

- 3T ± 5T 2
=
2T 2

- 3T ± 2.23T
=
2T 2

- 3T ± 2.23T - 3T ± 2.23T
= ,
2T 2 2T 2
 5TH PROOF 4-9-09

36 Process Control Engineering

- 0.777 - 5.23T
= ,
2T 2 2T 2

- 0.385 2.61
= ,
T T
Therefore the factors are

FG S + 0.385IJ FG S + 2.61IJ = 0
H T KH T K
= (TS + 0.385) (TS + 2.61) = 0
= (2.61TS + 1) (0.38TS + 1) = 0
H2 (S) 1
=
Q i ( S) (2.61TS + 1) (0.38TS + 1)
Subjecting the system to step input
1
H2(S) =
S (2.61TS + 1) (0.38TS + 1)

A B C
= + +
S 2.61TS + 1 0.38TS + 1
A=1 B = – 1.17 C = 0.17
1 ( -117)
. 0.17
H2(S) = + +
S 2.61TS + 1 0.38TS + 1
Applying Inverse Laplace Transform

h2(t) = 1 – 1.17 e – t / 2.6 T + 0.17 e– t/0.38 T ...(41)

h2(t)
Steady state value of input
1.0

Non-interacting
system response
(faster)
Interacting system
response (sluggish)

From the characteristics it is clear that the interacting system exhibits a

sluggish response than non-interacting system, thus the loading effect is more
in these systems.
 5TH PROOF 4-9-09

Introduction 37

1.5 PROCESS DEGREES OF FREEDOM

Process degrees of freedom implies assessment of how many independent
controllers can be employed in a control loop configuration, to meet the
requirements of control objective. More number of controllers than required
leads to redundancy and more cost. Hence, the number of independently
acting automatic controllers should not exceed the number of degrees of
freedom.
Let
h v = number of variables in a process
h e = number of defining equations (based on dynamics of
process)
h = number of degrees of freedom
then
h = h v – he ...(42)
Generally, the number of defining equations also refers to transfer
function of process.
For example, in a temperature process number of variables are inlet flow
rate, outlet flow rate, heat input rate. (Rarer Fig. 1.18) and the defining
equation of process is one, thus the number of independent controllers that
can be used without redundancy.
h = h v = he
=3–1=2
Therefore the process degrees of freedom = 2
In chemical processes which involve seperation, distillation and
fractionation i.e. where heterogeneous equilibrium exists the equation (42)
may be modified as
h = h c – hp + 2 ...(43)
h c = number of components
h p = number of phases
Number (2) corresponds to temperature and pressure effects that have to
be necessarily considered for a chemical process.
For isothermal or constant pressure processes
h = hc – hp + 1 ...(44)
Equations (43) and (44) are known as “GIBB’S PHASE RULE”. For
example in a steam boiler the number of phases are two components is one i.e.
liquid and it is not an isothermal process because both temperature and
pressure have to be considered.
 5TH PROOF 4-9-09

38 Process Control Engineering

h = h c, h p + 2
=1–2+2
n=1
Either temperature of steam pressure may be considered an independent
variable.

1.6 PIPING AND INSTRUMENTATION DRAWINGS

Every process control system is represented in terms of piping and
instrumentation drawing, where each instrument is designated by an alpha
numeric code or tag number and the signal type in terms of instrument line
symbols.
The different functions performed by instrument is shown by the
alphanumeric code and the type of the instrument is represented by general
symbols.
Hence, it is important for an instrumentation engineer to decode the
process diagram and identify the functions of each instrument. In this section
the procedure for decoding the process control drawings is attempted.

IDENTIFICATION OF LETTERS
FIRST LETTER SUCCEEDING LETTER
Measured Modifier Readout or Output Modifier
Variable Passive function function
A Analysis — Alarm — —
B Burner — UC UC UC
C UC — — Control —
D UC Differential — — — —
E Voltage — Sensor — —
F Flow rate (fraction) Ratio — — —
G UC — Glass — —
H Hand — — — High
I Current — Indicate — —
J Power Scan — — —
K Time of change Time rate Station — Control —
L level — Light — Low
M UC Momentary — — Middle
Intermediate
N UC — UC UC UC
O UC — Orifice — —
P Pressure — Point Cnnection — —
Q Quantity Integrate — — —
R Radiation — Record — —
S Speed, Frequency Safety — Switch —
 5TH PROOF 4-9-09

Introduction 39

T Temperature — — Transmitter —
U Multi Variable — Multi function Multi function Multi function
V Vabration — — Valve —
W Weight, Force — Well — —
X Unclassified X-axis Unclassified Unclassified Unclassified
Y Event (or) State Y-axis — Relay —
Z Position Z-axis — Driver Actuator, —
FCE
Note: UC = User’s Choice.
INSTRUMENT LINE SYMBOLS
1. Instrument Supply 7. EM or Sonic Signal (Guided)
2. Undefined signal 8. EM or Sonic Signal (Unguided)
3. Pneumatic Signal 9. Internal System Link
4. Electrical Signal (or) 10. Mechanical Link
5. Hydraulic Signal 11. Peneumatic Binary Signal
6. Capillary tube 12. Electrical Binary Signal
(or)

ABBREVIATIONS
AS ® Air Supply HS ® Hydraulic Supply
IA ® Instrument Air NS ® Nitrogen Supply
PA ® Plant Air SS ® Steam supply
ES ® Electric Supply WS ® Water Supply
GS ® Gas Supply
GENERAL INSTRUMENT SYMBOLS
Primary Location Field Mounted Auxiliary Location Not
(Accessible to Operator) not accessible to operator
Discrete
Instruments
Shared Display or
Shared Control
Computer
Function
PLC

Control Valve
Symbol of
with Actuator Two way valve Two way valve
(Fail open) (Fail closed)

Any valve (Fail locked) Three way valve

Four way valve

 5TH PROOF 4-9-09

40 Process Control Engineering

Consider the following example of decoding the information from a

process control drawing.

PIC
Example: 204

Þ General instrument symbol indicating shared display.

PIC 204 Þ Instrument Identification (or) TAG Number

P 204 Þ Loop Identification
204 Þ Loop Number
PIC Þ Functional Identification
P Þ First Letter
IC Þ Succeeding Letters.
From the table
First letter (P) Þ Pressure
Succeeding IC Þ Indicate, control
Therefore PIC implies “Pressure indicator and controller.”
Some times a modifier may be used for first letter. For the same pressure
indicator and controller, if modifier presented as

DIPC D = Modifier for the first from the table

204
D = Differential
Therefore DPIC implies “Differential Pressure Indicator and Controller”.
 5TH PROOF 4-9-09

Introduction 41

SOLVED EXAMPLES
1. Calculate the resistance of a flow metering device whose flow
characteristics are related by q = hn.
Solution: We know that
head ( h )
Resistance (R) =
flow rate ( q )

h Dh dh
R= = = . ...(1)
q Dq dq
Taking “Log” on both sides for the equation q = hn
Log q = Log [hn]
log q = n log h.
Differentiating w.r.t. “q”
d d
Log q = [n Log h]
dq dq

1 1 dh
= n. .
q h dq

dh h
Þ = .
dq nq
Comparing with eqn (1)

h
R= Ans.
nq

2. Obtain an expression for turbulent resistance for fluid flowing in a

pipeline.
Solution: From Bernoulli’s equation

q = KA 2 g( h1 - h2 )
where, K = flow constant, A = Area of restriction (M2).
Squaring b.s.
q2 = K2A2[2g(h1 – h2)].
q.q = K2A22g(h1 – h2).

q 2( h1 - h2 )
2 2
=
K A g q
 5TH PROOF 4-9-09

42 Process Control Engineering

Let h1 – h2 = h

q 2h
\ 2 2
= .
K A .g q

2h
Since R=
q

q
\ R= Ans.
K A2 . g
2

3. Show that the time constant element of fluid flowing in a pipeline

having cross-sectional area ‘A’ is
2HA
t= .
Q
Solution: From Bernoulli’s equation

q = KA 2 g( h1 – h2 )

q2 = K2A22g(h1 – h2).
q.q = K2A22g(h1 – h2).

q 2( h1 – h2 )
2
2
= .
K A g q

Let h1 – h2 = H.
Capacitance C a Area (A)

q
=
2H
= R3 R=
h
.
LM OP
2
2
K A g q q N Q
2H
\ R= [Expressing q = Q].
q

2H
R= .
Q
Multiplying b.s. by C
2HC
RC =
Q

2HA t = RC
t= 3 C = A [ Area ] .
Q
 5TH PROOF 4-9-09

Introduction 43

4. A triangular wier has an equation for flow rate as q = Cv 2 gh 5 .

Find its resistance.
Solution: Taking Log on b.s.

LM L OP OP
2

N MN QQ
Log q2 = Log Cv2 2 gh 5

2 Log q = 2 log Cv + Log (2gh5).

2 Log q = 2 log Cv + log 2g + 5 Log h.
Differentiating w.r.t. q.
2 5 dh
=0+0+ . .
q h dq

dh
Since R= .
dq

2h
R= Ans.
5q

5. A thermometer having time constant response requires 5 minutes to

indicate 98% of the response to sudden input. What is the time
constant?
Solution: The general first order system having transfer function
C( s) K
= has its step response to be
R( s) (1 + Ts)

c(t) = k[1 – e–t/T].

c(t) = 98% = 0.98.
Assuming unit step input
k = 1 & t = 300 secs [5 minutes]
\ 0.98 = [1 – e– 300/T].
Solving for ‘T’ T = 76.68 secs.
6. A temperature sensor modelled as time constant process is suddenly
subjected to a sudden change in temperature from 0°C to 100°C. If it
has a time constant of 5 secs, what temperature will be indicated after 5
secs?
Solution: The first order/time constant process response is
c(t) = k[1 – e– t/T].
 5TH PROOF 4-9-09

44 Process Control Engineering

100°C

0°C
t
5 secs

T = 5 secs t = 5 secs.
c(t) = 100°C [1 – e–5/5]
c(t) = 100°C [1 – 0.37] [e– 1 = 0.37]
c(t) = 63°C.
7. For the process shown in figure calculate the time constant for the
following specifications.

qi

R1
h c

qo

Normal operating head = 5 m.

Outflow = 0.5 m3/s
Area of tank = 10 m2.
Solution: We know that the resistance of fluid flowing in a pipeline

2h
R= = given h = 5 m. q = 0.5 m3/s
q

2´5
\ R= = 20 s/m2.
0.5
Capacitance of the tank = Area = 10 m2.
c = 10 m2
\ Time constant element

20s
T = Rc = ´ 10 m2
2
m

T = 200 secs.
 5TH PROOF 4-9-09

Introduction 45

8. Calculate the time constant of a tank with 3 m head and an outflow of

25 m3/s through a valve. The cross-sectional area of the tank is 5m2.
Solution: The turbulent resistance

2h 2´3
R= = \ h=3m q = 25 m3/s.
q 25
R = 0.24 s/m2.
Since C = A = 5 m2 .
T = Rc = 0.24 s/m2 ´ 5 m2
T = 1.2 secs
T = 1.2 secs.
9. A liquid level process having first order dynamics is suddenly subjected
to a change in level from 2 m – 10 m at time t = 0. If the time constant
of the process is 10 secs, estimate the rise in level after 20 secs.
Solution: From the first order dynamics o/p c(t) = K[1 – e– t/T] when the initial
value of the output is zero.
If the initial value of the output is not zero, then the above equation may be
modified as
c(t) = Ko + [Ki – ko]e– t/T
where, Ki = Initial value.
Ko = Final value.
For liquid level process
output h(t) = ho + [hi – ho] e–t/T.
ho = Final value of level = 10 m
hi = Initial value of level = 2 m.
After t = 20 secs.
h(t) = 10 + [2 – 10] e–20/10
h(t) = 10 + [– 8]e–2
h(t) = 10 + [– 8][0.13].
h(t) = 10 – 1.04 = 8.96 m
\ h = 8.96 m.
10. A temperature sensor having time constant characteristics is subjected
to a sudden change in temperature of 25°C – 200°C. If it has a time
constant of 5 secs, what temperature will be indicated after 8 secs?
 5TH PROOF 4-9-09

46 Process Control Engineering

Solution: T(t) = To + [Ti – To]e– t/T.

Ti = Initial value of temperature = 25°C.
To = Final value of temperature = 200°C.
t = 8 secs and T = 5 secs.
T(t) = 200 + [25 – 200]e– 8/5
= 200 + [– 175] ´ 0.2
200 – 35 = 165°C.
\ T = 165°C.
11. A single degree of freedom system exhibiting oscillatory characteristics
has a mass of 10 ´ 10– 3 kg and stiffness of 1000 n/m.
(a) Calculate the undamped natural frequency.
(b) Damped natural frequency, if the damping factor is adjusted to a
value of 0.6.
Solution: (a) We know that

k 1000
Wn = = = 100 ´ 103
m 10 ´ 10 – 3

W n = 316.22 r/s.
(b) Given x = 0.6.
W n = 316.22 r/s.
\ Damped natural frequency

W d = Wn 1 – x 2

W d = 316.22 1 – (0.6)2

W d = 252.97 r/s.
12. Obtain the time domain specifications of second order under damped
response subjected to step input.
Solution: Mp Tolerance
band
The time domain specifications are: 1.0 2.5%
1. Delay time (td): It is defined as the 0.5
time taken by the system response to
reach 50% of the final value. td
tr
tp
1 + 0.7x ts
td = secs.
Wn
 5TH PROOF 4-9-09

Introduction 47

2. Rise time (tr): It is defined as the time taken by the system response to
reach 100% of the final value.

p–f
LM
tan –1 1 – x 2 OP
tr =
Wd
f=
MN
x.
(radians) .
PQ
3. Peak time (tp): It is defined as time taken by the system response to
reach maximum peck value.
p
tp = (secs).
Wd
4. Setting time (ts): It is defined as time taken by the system response to
reach and settle down within the limits of tolerance band about steady state
value of the input.
4
ts = for 2% Tolerance band
xwn

3
ts = for 5% Tolerance band.
xwn
5. Maximum peak overshoot (mp): It is the maximum value of the
response measured from steady state value.
– xp / 1 – x 2
MP = e
– xp / 1 – x 2
% MP = e ´ 100 .
13. Derive the expression for rise time (tr) for the second order under
damped system subjected to step input.
Solution: We know that under damped response

e – xwn t
c(t) = 1 – sin [wd t + f].
1 – x2

tan –1 1 – x 2
where, f= .
x
At t = tr c(t) = 1

e – xwn tr
1 =1– sin (wd tr + f).
1 – x2
 5TH PROOF 4-9-09

48 Process Control Engineering

e – xwn tr
sin [wd tr + f] = 0.
1 – x2

– xw n tr
e
Since is finite
1 – x2

sin [wd tr + f] = 0.
This is true if
wdtr + f = p

p–f 1 – x2
Þ tr = where, f = tan –1 .
wd x

14. Derive an expression for peak time and maximum peak overshoot
shown in time domain specifications.
Solution: To obtain an expression for peak time
d
c(t) = 0.
dt

d e – xwn t
1– sin [wd t + f] = 0
dt 1 – x2

LM e – xw n t
e – xwn t
OP
=0–
MN 1 – x 2
cos ( w d t + f). wd + sin ( w d t + f)
1 – x2
[– xw n ]
PQ
e – xwn t
Þ [ xwn(sin wdt + f) – wd cos (wdt + f)] = 0
1 – x2

e – xwn t
Since is finite
1 – x2

sin ( wd t + f) w wn 1 - x2
= d =
cos ( wd t + f) ywn x wn

1 – x2
tan [wd t + f] =
x
 5TH PROOF 4-9-09

Introduction 49

1 – x2
Since f = tan –1
x

LM 1 – x2 OP 1 – x2
MN
tan wd t + tan –1
x PQ = x

tan [wd t] = 0
Þ wdt = p

t = tP = p/wd
Substituting t = tP = p/wd in the expression for response

c( t ) t = t = c(t)|max.
p

p
e – xwn ´
wd é p ù
c(t)|max = 1 – sin ê wd + fú .
2
1– x êë wd úû

– xp / 1 – x2
e
c(t)|max = 1 – sin [p + f].
1 – x2
From the under damped pole locations

sin f = 1 – x 2
wn
wn 1 – x2
– xp / 1 – x2
e
\ c(t)|max = 1 – [– sin f] xwn
1 – x2

 
-xp / 1- x 2
e é ù
c(t)|max = 1 - ´ ê- 1 - x2 ú
2 ë û
1- x

– xp / 1 – x 2
c(t)max = 1 + e
MP = c(t)|max – 1

– xp / 1 – x 2
MP = e .
15. A certain oscillatory process has a transfer function.
C( s) K ( s + 1)
= 2 .
R( s) ( s + 2 s + 5)
 5TH PROOF 4-9-09

50 Process Control Engineering

If the steady state value of the response when subjected to unit step
input is 2, find the system gain K.

RS K ( s + 1) UV
T| ( s + 2 s + 5) W|
Solution: C(s) = R(s) 2

let R(s) = 1/5.

K ( s + 1)
C(s) = .
S( s2 + 2 s + 5)
Applying final value theorem.

Lt S. C(s) = 2
s®0

K ( s + 1)
Lt s . =2
s®0 s ( s2 + 2 s + 5)

K
\ = 2 Þ K = 10.
5
16. Obtain the dynamic equations for the mechanical system shown in
figure.

x2

M2
B1 x1

K2 K1 x
M1 F

B2

Solution: The nodal diagram for the above figure is

At note (x)
K1 x1 B1 x2
x
B2

F M1 M2 K2
 5TH PROOF 4-9-09

Introduction 51

F = k1(x – x1).
At node (x1)

d2 x1 d
0 = M1 2
+ B1 (x1 – x2) + K1(x1 – x),
dt dt

d2 x1 d
K1(x – x1) = M1 + B1 (x1 – x2)
dt 2 dt

d2 x1 d
F = M1 + B1 (x1 – x2). ...(1)
dt 2 dt
At node x2

d2 x2 dx d
0 = M2 + B2 2 + B1 (x2 – x1) + k2x2. ...(2)
dt 2 dt dt
17. Find the transfer function of lead compensator network.

R1
I(s)

1
Vi(s) CS R2 V0(s)

Vi(s) = I(s)
LM R + R OP 1

N R cs + 1 Q
1
2

Vi(s) = I(s)
LM R + R + R R CS OP
1 2 1 2

N R CS + 1 Q 1

Vo(s) = I(s)[R2].

Vo ( s) R2 [ R1CS + 1]
\ =
Vi ( s) R1 + R2 + R1 R2CS

R2 ( R1CS + 1)
=
L
R + R M1 +
R R CS O
.

N R + R PQ
1 2
1 2
1 2
 5TH PROOF 4-9-09

52 Process Control Engineering

R2
Let T = R1c ; a = (a < 1)
R1 + R2

Vo ( s) a(1 + Ts)
= .
Vi ( s) 1 + aTs

18. Obtain the transfer function of LAG compensator.

I(s) R1

R2
Vi(s) Vo(s)
1
cs

LM R cs + 1OP 2
N
Vi(s) = I(s) R1 +
cs Q

LM R cs + R cs + 1OP
1 2
Vi(s) = I(s)
N cs Q
LM R cs + 1OP .
2
Vo(s) = I(s)
N cs Q
Vo ( s) R2 cs + 1
=
Vi ( s) R1 cs + R2 cs + 1

R1 + R2
Let b= & T = R2 c
R2

Vo ( s) R2 cs + 1 1 + Ts
= = .
Vi ( s) cs( R1 + R2 ) 1 + bTs
+1
R2

19. Obtain the transfer function of lead-lag compensator.

R1
I(s)

1 R2
Vi(s) c1s Vo(s)
1
c2s
 5TH PROOF 4-9-09

Introduction 53

Solution: Vi(s) = I(s)

LM R + R c s + 1OP
1 2 2

NR c s +1 c s Q
1 1 2

Vo(s) = I(s)
LM R c s + 1OP
2 2

N cs Q
2

Vo ( s) ( R1c1 s + 1) ( R2 c2 s + 1)
\ =
Vi ( s) ( R1c1 s + 1) ( R2 c2 s + 1) + R1c2 s

Vo ( s) ( R1 c1 s + 1) ( R2 c2 s + 1)
=
Vi ( s) R1c1 R2 c2 s 2 + s[ R1c1 + R2 c2 + R1 c2 ] + 1

Vo ( s) a(1 + T1 s) (1 + T2 s)
=
Vi ( s) (1 + aT1 s) (1 + bT2 s)

where, T1 = R1C1 T2 = R2C2

R2 R + R2
a= b= 1 .
R1 + R2 R2

X1( s)
20. Find the transfer function .
F ( s)

K2 x2
x1
K1

M2 M1 F
B2

Solution: The nodal diagram is

x1 K1 x2

B2
F K2
M1 M2
 5TH PROOF 4-9-09

54 Process Control Engineering

At node x1

d2 x1
F= + K1 (x1 – x2).
dt 2
At node x2

M2 d2 x2 B2 dx2
0= + + K2x2 + K1(x2 – x1).
dt 2 dt
Applying Laplace transform
F(s) = X1(s) [M1s2 + K1] – K1X2(s)
0 = X2(s) [M2s2 + B2s + K2 + K1] – K1X 1(s)

K1
\ X 2(s) = . X1(s).
M2 s2 + B2 s + K 2 + K1

2
– K . X1 ( s)
1
\ F(s) = X1(s)[m1s2 + K1]
M2 s2 + B2 s + K2 + K1

LM( M s 2
+ K1 )( M2 s2 + B2 s + K2 + K1 ) – K
2 OP
F(s) = X (s) N Q
1
1
1
M2 s2 + B2 s + K2 + K1

X1( s) M2 s 2 + B2 s + K1 + K2
= .
F ( s) 2
( M1 s 2 + K1 )( M2 s 2 + B2 s + K1 + K2 ) – K
1

QUESTION SET
1. With a neat diagram explain the temperature control system indicating
all the elements of process control loop.
2. Explain the importance of elements of process dynamics in modelling
of processes.
3. What is the significance of time constant in a first order process?
4. Obtain the different approximations of dead time element.
5. Explain clearly how damping affects the nature of response of
oscillatory process.
6. Derive an expression of oscillatory process (under damped) for sudden
changes in input. Mark all the time domain specifications and explain
clearly each one of them?
 5TH PROOF 4-9-09

Introduction 55

7. Obtain the mathematical models for (a) Liquid level system (b)
Thermal system (c) Pneumatic system.
t1( s)
8. A time constant process is shown in fig. Find the transfer function .
Qi( s)

C = Hydraulic
qi
capacitance

c R1
h
qo1

R2

qo2

9. For an interacting system, draw the block diagram schematic and hence
write the mathematical model.

qi

c1 R1 c2 c3 R3
R2
h1 h2 h3
qo

Tank 1 Tank 2 Tank 3

10. For the above interacting system (Prob 9) draw equivalent electrical N/
W and compare the mathematical models.
t13 ( s)
11. Find the transfer function for three tank non-interacting system .
Q i ( s)

qi

c1 R1
h1

c2 R2
h2

R3
h3 c3

qo
 5TH PROOF 4-9-09

56 Process Control Engineering

12. What is meant by process degrees of freedom?

Explain mathematically how it is related to the choice of the number of
independently acting controllers in a control loop.
13. In P and ID symbology the following abbreviations stand for
(a) ES (b) WS (c) AS (d) GS (e) NS
14. Decode the following instrument line symbols.

(a) (b)

(c) (d)

(e)

15. Decode the following piping and instrumentation diagrams.

LC SP FC
104 100

(a)

FT

(b)
TDI
TT TT
103

(c) TIC
TT
101

(d)
PI PI PI
(A) 102 (B)
 5TH PROOF 4-9-09

Introduction 57

Gas Flow

PT
FT TT
108
(e)
FY FIC
111

16. Define the following terms:

(a) Shared Controller (b) Shared Display
(c) Panel Mounted (d) Control Station
(e) Binary
17. Write the dynamic equations and obtain analogies for mechanical and
electrical systems.

R L

f
K
+ C
V i
–

F x

18. A certain process control system is modelled as shown in figure.

R(s) 16
C(s)
– – s(s + 4)

as

What should be the value of ‘a’ such that the response for unit step input
will give an overshoot of 1.5%?
19. A thermometer is subjected to a sudden temperature change of 35°C
® 115°C. What would be the time constant, if it indicates 87.63°C
after 6 secs?
20. For a second order system having transfer function
C( s) 8
= 2
R( s) s + 8 s + 25
Find the time domain specifications.
 +0)26-4

Modes of Controller Operation

In this chapter the different modes of controller operation are discussed. In

any control system, where the dynamic variable has to be maintained at the
desired set point value, it is the controller alone which enables to meet the
requirements of control objective. The controller inputs the result of a
measurement of the controlled or dynamic variable and determines an
appropriate output to the final control element.
Usually final control elements understand certain range of signal
transmission only. Hence the o /p of a controller must be translated to the
range of possible values of final control element. This range is called as
controller parameter range.
In terms of analog signal transmission this range of controller output is
4–20 mA. standard signal.
4 mA corresponds to 0% controller output i.e. minimum and 100%
controller output i.e. maximum would be 20 MA respectively.
In discrete state process control systems, the output will range over all the
states of the n-bit output. Generally, all 0’s correspond to minimum output and
all 1’s represent maximum controller output. The controller output as
percentage of full scale when the output varies between specified limits is given
by the expression

D p - Pmin
%P= ´ 100
Pmax - Pmin

where Dp = Actual controller output.

Pmax, Pmin Þ Standard signal transmission range.
For example in 4–20 mA signal transmission, 4 mA corresponds to 0%
controller 0/p and 20 mA, 100% controller 0/p. If we want to evaluate how
much in terms of percentage would be 12 mA controller output, the above
equation can be used.
 5TH PROOF 4-9-09

Modes of Controllers Operations 59

D p = 12 mA
Pmin = 4 mA, Pmax = 20 mA

12 - 4
\ %P = ´ 100
20 - 4

8
= ´ 100 Þ 50 %
16
Similarly, the standard pneumatic signal transmission range is 3–15 psi.

2.1 REVERSE AND DIRECT ACTION OF

CONTROLLERS
Depending upon the nature of the process and the demands of control
objective a controller has to operate in direct or reverse modes.
A controller is said to operate in reverse action, when an increasing valve
of controlled variable causes a decrease in the controller output.
For example in a themo statically controlled temperature system, the
controller has to turn on the heater when temperature drops below the
thermostat set point.
Similarly controller action would be considered a direct action if the
increasing value of controlled variable results in increased controller output.
An example to this effect is simple level control system. As the level of the
liquid in the tank increases beyond set point [Increasing Controlled Variable]
the controller should open the control valve and allow more fluid to go out (i.e.
the drive signal is also increased).
Hence, the nature of the process determines the direct or reverse action of
the controller.

2.2 MODES OF CONTROL ACTIONS

Generally, the controller is some form of an automatic device or computer,
either analog or digital, pneumatic or electronic, which, using input
measurements, solves certain equations to calculate the proper output. It is
these equations, that describe the modes or actions of controller operation. As
stated earlier, the nature of the process itself and the particular variable to be
controlled determines which mode or modes to be used and the value of
certain controller constants in the mode equations. A process control
technologist should therefore have a clear understanding of the operational
mechanism of each mode together with its advantages and disadvantages.
Then alone he can design a suitable controller for a process control system
keeping in view, the industrial factors and requirements such as process
 5TH PROOF 4-9-09

60 Process Control Engineering

characteristics, dynamics, compatibility, cost analysis, product rate etc. The

controller equations, are written with respect to output of controller denoted
by letter P and input to controller, i.e. error ‘e’.
Thus, the control actions are expressed as P = f (e).
i.e. controller o /p ( p) = function of error signal (e).

Input e
controller P

output

CLASSIFICATION
Modes of
Control Actions

Discontinuous Continuous
Modes of Modes of
Action Action

On / off or Floating Control

Two Position Mode Proportional Integral Derivative Composite
Mode (P-Mode) (I- Mode) (D - Mode) Modes

Singie Speed Multiple Speed

Proportional + Integral Proportional Proportional
(P + I Mode) + +
Derivative Integral +
(P + D Mode) Derivative
(P + I + D
Mode)

2.3 DISCONTINUOUS CONTROLLER MODES

In this section we discuss about the different discontinuous modes of controller
operation.

2.3.1 Two Position Mode

The two position or On /off is the most elementary controller mode. Though
an analytic equation cannot be exactly written, this mode is mathematically
expressed as
P =0% e<0
= 100 % e > 0.
The above representation implies whenever there is error (e > 0 ), the
controller output is 100 % in terms of both ON (or ) OFF states. When there is
no error (e < 0 ) the controller 0 / p cannot be defined.
 5TH PROOF 4-9-09

Modes of Controllers Operations 61

One of the most common examples is in the operation of thermostat. The

heater is turned ON or OFF whenever there is error i.e temperature above or
below set point. When the temperature is at set point (e < 0 ), the controller’s
response is undefined.
Neutral Zone: Practically in designing of two position controller, whenever
the controller output has to change over from 0% ® 100% or vice versa, there
is a differential gap known as Neutral Zone around Zero error point where
virtually no controller output results.
Controller
O/P ON
(P) 100%

OFF
0%

error
(e)
Neutral Zone

FIG. 2.1

This Neutral Zone is purposefully designed above a certain minimum

quantity to prevent excessive cycling. It is an example of desirable hysteresis in
a system. The process under two position control must allow continued
oscillations in the controlled or dynamic variable when the error is zero and
these oscillations are verily the function of Neutral Zone size.

Applications
The main application of two position control are:
(1) Room heating or air conditioning systems
(2) Liquid bath temperature control
(3) Level control in large volume tanks.

Multi position mode

It is the logical extension of two-position control where rather than only two
settings of controller output, namely, 0% (OFF) and 100% (ON), several
intermediate settings are provided.
The most common example is a three position controller

R|100 % e > e2
P= S| 50 % - e1 < e < e2
T 0% e < - e1
 5TH PROOF 4-9-09

62 Process Control Engineering

As long as the error is between e2 and e1 of the set point, the controller stays
at a nominal setting of 50% if the error exceeds the set point by e1 or more, the
output is increased by 100 %. If it is less than the set point by –e1 or more the
controller output is reduced to zero.

2.3.2 Floating Controller Mode

In this control action, the specific controller output is not uniquely determined
by error unlike the previous mode. If the error is zero, the controller output

Controller
0/P
P 100%

50%

0
error
– e1 0 + e1 (e)

FIG. 2.2

floats at whatever setting it was when the error went to zero. And whenever
error occurs the controller output again begins to change.

Single speed floating control mode

In this mode the controller output changes at a fixed rate whenever the error
exceeds the neutral zone. The analytic expression is

dp
= ± KF |e|> Neutral Zone
dt

dp
where = rate of change of controller output,
dt
KF = Rate constant [% / sec]
The specific controller output is obtained by integrating the above
equation.
P = ± KF t + P (0)
where P (0) refers to initial setting of the controller when the error is zero. The
characteristics of single speed floating controller mode is

Illustration of floating action

Consider the graph showing controller output with error, that exceeds. The
Neutral Zone, after certain instants of time as shown in Figure (2.4).
 5TH PROOF 4-9-09

Modes of Controllers Operations 63

dp
dt KF

– KF

error
0 (e)

FIG. 2.3

Between the instants 0 – t1 the error is within the limits of Neutral Zone.
Hence the controller output floats at whatever setting it was when the error
went to zero i.e. P (0). From t1 to t2 error exceeds the neutral zone and the
controller output changes at a fixed rate, ± KF t, depending upon Nature of
Process, i.e. reverse or direct action.
Neutral
Zone
Error t
(e)

Floating
Action
+KFt –KFt
Controller Floating Floating
O/P P(0)
Action Action

t
t1 t2 t3 t4

FIG. 2.4

Multiple speed floating controller mode

In single speed, the rate of change of controller output is fixed at certain levels
of ± KF. In multiple speed control mode, there are several different rates of
controller output depending on the nature of deviation or error.
The analytic expression can thus be modified as

dp
= ± KF i e > ei
dt

The above equation implies, whenever error is in the range of certain

preset values there will be corresponding rate of change of controller output.
 5TH PROOF 4-9-09

64 Process Control Engineering

Application
The single speed or multiple speed floating action can be employed with

Controller
O/p

+ K2
+ K1

– K1

– K2

error
– e2 – e1 0 e1 e2

FIG. 2.5

general flow control system or pressure control system with small capacities
and having self-regulation to some extent with small dead time or Lag.

2.4 CONTINUOUS CONTROLLER MODES

Unlike the discontinuous modes of operation, in continuous modes there is a
smooth relation existing between controller output and error, whenever there
is a deviation of the controlled variable from set point. The controller responds
in a smooth fashion to achieve the control objective.
Some of the continuous modes are natural extension of discontinuous
modes itself. We will discuss about these modes in detail in this section.

2.4.1 Proportional Controller Mode

It is a natural extension of two position mode. In this mode there is a smooth
relation between controller output and error before the output saturates at 0%
[OFF state] or 100% [ON state]. Between these two saturation levels, there is a
band of errors, where every value of error has a unique value of controller
output i.e. there is a one-to-one correspondence existing between controller
output and error.
This range of errors to cover 0 % to100% controller output is known as
“PROPORTIONAL BAND”.
The analytic expression is written within this proportional band which
describes the operation of proportional control action.
P = Kp e + P0 ...(1)
Kp = proportional gain,
P0 = Initial value of controller o/p.
 5TH PROOF 4-9-09

Modes of Controllers Operations 65

The transfer function representation is obtained by applying Laplace

transform to equation (1),
P(s) = Kp. E(s) [Initial value is assumed to be zero]

\ E(S) Kp P(S)

CHARACTERISTICS OF PROPORTIONAL MODE

Figure (2.6) represents the characteristics of P-control action. The
proportional band (P . B) is related to the gain of proportional controller Kp
and is defined

Controller
o/p
100 % Saturation
P 100%

0% Saturation
0%

error
P.B (e)

FIG. 2.6
error
(e)
t
t = t1

Controller
o/p
P
t
t = t1

FIG. 2.7

10 0
as P . B = . Thus by proper choosing of ‘Kp’ the P . B can be changed
Kp
In Figure 2.7, assume a step change in deviation,
Let e = A [step change].
P = Kp e,
Þ P = K p A.
 5TH PROOF 4-9-09

66 Process Control Engineering

Thus if there is error, for every 1% of error a correction of Kp % is added or

subtracted from initial value of controller output ‘P0’ depending on the nature
of error. If the error is zero which implies if the controlled variable is at set
point value the controller output is constant at P0.

Offset
One of the main disadvantage of proportional control action is that it produces
a permanent residual error in the operating point of the controlled variable
when a process load change results in error. This residual error is known as
OFFSET.

Illustration of offset
Consider a first order process having process equation or transfer function

C ( s) 1
= .
R ( s) 1+ TS

R(s) E(s) P(s)

1
kp C(s)
1 + TS
Controller Process

FIG. 2.8

If we assume a sudden change in the input causing a sudden error. The

proportional controller adjusts the final control element proportionately
affecting the dynamics of the process to bring back the output to the set point
value. This is possible if and only if the steady state error is zero.
From the definition of steady state error

S . R ( s)
ess = Lt
s®0
1 + G ( s) H ( s)

A
Let R (s) = [step change]
S

A
s/
s/
ess = Lt
s®0 Kp
1+
1 + TS

A
e ss = ...(2)
1+ Kp
 5TH PROOF 4-9-09

Modes of Controllers Operations 67

This steady state error is not zero but inversely proportional to gain Kp.
Thus, this control action fails to make the error zero whenever the o /p changes
due to changes in input or Load (process) variables. This is known as offset.
From eqn. (2),

1
offset µ ...(3)
Kp

Therefore both P . B and offset are inversely proportional to gain Kp. To

reduce offset Kp should be increased but it reduces the width of P.B and the
control mode changes to simple ON /OFF mode. Hence offset constitutes the
inherent disadvantage of proportional controller mode.

2.4.2 Integral Controller Mode

This mode is an extension of floating control mode. Unlike the previous
discontinuous mode, the rate of change of controller output is not constant at
± KF but is directly proportional to error.

Controller
dp
o/p ae
dt
kF
P
0

– kF

error
o
(e)

FIG. 2.9

The analytic expression may be written as

dp
µe
dt

dp
= KI . e ...(4)
dt

where KI = Integral scaling.

From equation (4), the rate of change of controller o /p is proportional to
error. Thus when error occurs, the controller responds by sending an output at
a rate that depends upon the size of the error and integral scaling KI.
 5TH PROOF 4-9-09

68 Process Control Engineering

For particular error, the output will begin to increase or decrease at a rate
KI % /second for every 1% of error.
The specific controller output is obtained by integrating eqn. (4).

P = KI z e d t + P(0) ...(5)

where P (0) = Initial value of controller o /p when error is zero.

The integral scaling KI is often expressed in terms of integral time “T i”

1
Ti = ...(6)
KI

From (5) and (6),

P=
1
Ti z e d t + P (0) ...(7)

The units of integral time is seconds or minutes.

The transfer function of integral controller mode is obtained by applying
L . T to equation (7)

1
P (s) = . E (s).
Ti s

E(s) 1 P(s)
Ti s

Characteristics of integral action

Consider a step change in deviation
e = A [step charge]

P=
1
Ti z A. d t

Þ P=
LM A OP t
NT Qi

From Figure (2.10), the main disadvantage of integral action is that, its
response to error is slow or sluggish. But however, due to its basic nature of rate
of change of output it eliminates the error and brings back the controlled
variable to the set point value.
 5TH PROOF 4-9-09

Modes of Controllers Operations 69

error
e
A

A
Controller t
Ti
o/p P

t
t = t1

FIG. 2.10
Since the rate of change of controller output can be reset at any instant of
time t by changing “Ti”. This mode is often referred to as “RESET
CONTROLLER MODE”. The integral time “Ti” is also known as Reset time. Let
us see the effect of integral control action on error by taking a first order
process.

(Integral Mode) (Process)

1 P(s) 1
R(s) C(s)
Tis 1 + Ts

FIG. 2.11

S . R ( s)
ess = Lt
s®0
1 + G ( s) H ( s)

A
Let R(s) = [causing step change in deviation],
S

A
s/
ess [error] = Lt /s
s®0
1
1+
b
Ti s 1 + T s g
A
ess [error] = = 0.
1+ ¥
Thus it can be established that integral control action successfully
eliminates the error, but in the process its response is sluggish. The integral
action, due to its basic nature, might introduce oscillations in the controlled
variable about the set point value, whenever it deviates from the set point
value.
 5TH PROOF 4-9-09

70 Process Control Engineering

2.4.3 Derivative Controller Mode

This is the last pure mode of control actions. Unlike the previous two modes,
here the controller output is proportional to rate of change of error.
The analytic expression may be written as
de
Pµ
dt

de
P = KD ...(8)
dt
Defining “T d” derivative time as
Td = KD [derivative gain]
de
P = Td ...(9a)
dt
Applying L .T to equation (9a)
P (s) = Td s . E (s)
E(s) P(s)
Td s

Characteristics
® If the error is step change i.e. sudden, the controller output will be zero
because it responds only to rate of change of error. When e = A (step change)
d
P = Td (A) = 0.
dt
Considering the error to be rate of change i.e. linearly changing w.r.t. time,
e = At [Ramp change in deviation]
d
P = Td [A t]
dt
Þ P = Td A [Controller output]
As seen from Fig. (2.12), the derivative action is anticipating the nature of
error and giving out its output much before the error has occurred. Because of
this action, it is also called as Anticipatory Controller Mode.
Further, since the controller output is proportional to rate of change of
error it is also called as “Rate Controller Mode” and “Td” is called as rate time.
Unlike integral action, the derivative action’s response is much faster. For a
very rapid change of error there could be a large sudden changes of controller
output, which could lead to un-stable condition. This mode is suitable if the
process reaction rate itself is very sluggish, as it considerably improves the
speed of response.
 5TH PROOF 4-9-09

Modes of Controllers Operations 71

(At)
Error
(e)
t

Controller
o/p
P TdA { t
t = t1

FIG. 2.12

2.5 COMPOSITE CONTROLLER MODES

As seen in the previous section, the three individual modes have certain
inherent disadvantages.
To Summarise,
® proportional controller has good speed of response but has a
permanent residual error known as offset.
® Integral controller is capable of eliminating error, but its response
characteristics are sluggish.
® Derivative controller’s output is zero for sudden error [constant] and it
may even drive the process to instability.
Hence to suit the industrial requirements, it is possible to combine these
individual modes to give better overall performance. Such combinations can
improve the speed of response (Transient state behaviour) and /or the steady
state response [i.e. can deal with the error effectively].
We will study the most common composite control actions, that are
normally employed in process industries.

2.5.1 Proportional + Integral Controller Mode

It is the combination of both proportional and integral actions. The analytic
expression describing this mode is

P = Kp e +
Kp
Ti z e d t + P (0) ...(9)

Applying Laplace transform.

Kp
P (s) = Kp E (s) + E (s).
Ti s

|RS LM 1 OP|UV E (s).

|T N
P (s) = K p 1 +
Ti s Q|W
 5TH PROOF 4-9-09

72 Process Control Engineering

Thus, a P +I controller has a transfer function representation as shown in

the block diagram.

E(S) 1 P(S)
Kp 1 +
Tis

Characteristics of 2+ 1 controller mode

To understand the working of proportional action and integral action in this
combination, consider a step change in deviation (error), as shown in Figure
(2.13).
Consider a step change in error at an instant of time t = t1. Initially the
proportional action responds to error. It gives an output that is proportional to
error but it does not make the error zero and floats at the offset condition
which is its inherent characteristic. At this point the integral action takes over
and gives out a smooth rate of change of controller output about the offset
thereby gradually reducing it to zero. Mathematically, the characteristics of
P +I action shown in Figure (2.13) can be represented as follows.

error
(e)
A

t
t = t1

P-only

t
t = t1
I-only

t
t = t1
P + I
n
only A ctio AKp
P +I P = KpA + t
Ti
{

t
t = t1

FIG. 2.13

e = A [step change]

P = K pe +
Kp
Ti z edt.

P = Kp A +
Kp
Ti z Adt.
 5TH PROOF 4-9-09

Modes of Controllers Operations 73

F AK I t
P = Kp A +
 " ! GH T JK i
p

P - Action " "!

I - Action

Therefore, it may be stated that the integral action eliminates the inherent
offset of proportional action. To Illustrate consider a first order process whose
transfer function is

C ( s) 1
=
R ( s) 1 + Ts

P + I controller (Process)

{
{
1
R(s) Kp 1 + 1 C(s)
Tis 1 + Ts

The steady state error is given by the expression

S . R( s)
ess = Lt
S®0
1 + G ( s) H ( s)

A
Since R (s) =
S

LM F 1 I OP LMF 1 I OP
MN GH JK PQ MNGH1 + T JK PQ
G (s)H (s) = K p 1 +
Ti s s

A
s/
Lt /s
ess = S®0
F IF 1 I
1+ Kp GH 1+
1
Ti s JK GH 1 + Ts JK
A A
ess = = =0
e
1 + K p + ¥ (1) j ¥

Thus ess = 0 . Hence it can be mentioned here that a P + I controller

mode is capable of improving the steady state response i.e. minimising the
steady state error in a process. From Figure (2.13) it appear that this control
action also has good speed of response in dealing with the error when ever it
occurs, thus effecting the transient state behaviour of the process. Whether a
particular composite controller mode affects the transient state or speed of
response of the process can be very well verified by considering sinusoidal
deviation.
 5TH PROOF 4-9-09

74 Process Control Engineering

Let there be a sinusoidal deviation (or) error.

e = sin w t.

P = Kp sin w t +
Kp
Ti z sin w t. d t .

F - K I cos w t
P = K p sin w t + GH w T JK p

i
...(10)

We know that the sum of sine wave and co-sine wave is another sine wave
having different magnitude and phase angle

A2 + B2 sin w t + Tan - 1
LM B OP
A sin w t + B cos w t =
N A Q
In equation (10)
- Kp
A = Kp and B =
wTi
therefore,

LM F - K/ I OP
F - K I sin Mw t + Tan GG w T JJ P
p
2
P = eK j + G
H w T JK MM
2

JJ PP
p -1 i
p
i GG K/
H K PQ
p
MN
e K j + FGH w T IJK sin LMNw t - Tan w1T OPQ .
2
2 K p -1
P = p
i i

As seen from eqn. (11) when error (e) is sin w t the phase of the P +I
F 1 I . This implies the overall effect on
controller output lags by Tan–1 GH w T JK i
transient state or speed of response in the process is slowed down due to the
rate of change of controller output of integral mode. Hence this mode can be
optimally used for the improvisation of steady state response characteristics
only in a process.
The characteristics of P + I control action are also comparable with Lag
compensator (or) low pass filter.

2.5.2 Proportional + Derivative Controller Mode

The analytic expressing of this serial or cascaded combination is given by

de
P = K p e + K p Td + P0 ...(11)
dt
 5TH PROOF 4-9-09

Modes of Controllers Operations 75

Applying L . T to get transfer function representation

P(s) = K p E(s) + Kp Td s . E (s)
P(s) = [K p (1+ Td s)] E (s)
E(S) P(S)
Kp (1 + TdS)

Characteristics of P + D controller
We cannot consider a step change in deviation because the derivative action
does not respond to such errors. Hence considering rate of change of error.
At an instant of time t = t1, the error is changing at a continuous rate. As
shown in the figure, the derivative action first sends its output in its in imitable
anticipatory mode which is followed by proportional action. Such a response
would lead to large instability in a process which may result in error increasing
beyond proportions rather than becoming zero. If we consider this control
action for level control system discussed in earlier chapters, when the liquid
level in the tank starts rising beyond the actual set point resulting in an error
shown in Fig (2.14), the derivative mode opens the final control element to its
maximum capacity causing large quantity of outflow, followed by proportional
action further opening the control valve resulting in outflow greater than
inflow. This condition results in large error and High instability. Since this
mode is incapable of improving the steady state behaviour i.e. eliminate steady
state error and inherently exhibits good speed of response, we can use it to
processes having slow response characteristics, to effectively increase their
speed of response. Referring to the figure below

error At
(e)
t
t = t1

P-only
t
t = t1
D-only

t
t = t1
P+D
Action
P+
D { Ac
tion

t
t = t1

FIG. 2.14
 5TH PROOF 4-9-09

76 Process Control Engineering

Controller Process

R(s) Kp(1 + Tds) 1

1 + Ts

A
let R (s) = [ Ramp change in deviation].
S2

A
/s 2
/s
ess = Lt
S®0
K p 1 + Td s 
1+
1 + T s

A
= Lt
S®0
s K p 1 + Td s
s+
1 + T s
A
\ ess = =¥
o
Therefore this mode has no effect on the steady state response. Whether or
not it has any effect on transient state behaviour can be established by
considering sinusoidal deviation and respective analysis
Let e = sin w t

d
P = Kp sin w t + Kp Td sin w t.
dt

P = K p sin w t + w K p Td cos w t.

 
2
P= K p2 + wK p Td . sin [wt + Tan–1wTd]

P= e
K 2p + wK p Td j2
. sin w t + Tan - 1 w Td ...(12)

As seen from equation (12), for sinusoidal deviation the phase of the
controlled variable leads by Tan-1 wTd. Hence a P+D control action can be
employed with processes having slow response characteristics, which can be
conveniently improvised by this mode of control action.
Since the phase of controller output is leading by Tan–1 wTd for sinusoidal
deviation its characteristics are similar to Lead Compensator and High Pass
Filter.
 5TH PROOF 4-9-09

Modes of Controllers Operations 77

2.5.3 Proportional + Integral + Derivative Controller

Mode
[2+ 1 + , Mode]
It is the most versatile controller mode which can be used for treating any type
of errors affecting the processes. It is the combination of all the three modes
and is capable of improving both steady state and transient state responses i.e.
it successfully eliminates the error and also has good speed of response.
The analytic equation can be obtained as follows

P = Kpe +
Kp
Ti z e dt + Kp Td + P(0) ...(13)

The transfer function representation can be obtained by applying Laplace

transform to equation (13).

|RSK LM1 + 1 + T SOP|UV E(s)

T| N T S QW|
P(s) = p d
i

E(S) 1 P(S)
Kp 1 + + TdS
TiS

Characteristics of 2+ 1+ , controller

To understand the working of all the there modes, considering a ramp change
in deviation. Let us assume in the level control system described in chapter-1.
The level of the liquid in the tank is increasing at a constant rate beyond set
point value as the inlet flow rate is also changing at a constant rate. Since the
controller is PID controller, initially when the error begins to occur the
derivative mode opens the control element to its maximum capacity and the
level drastically falls below the set point, Again the level of the liquid starts
increasing from the new value from where it has fallen below set point because
inlet flow rate is changing at a constant rate. At this stage both proportional
and integral actions take over and manipulate the final control element in such
a way that gradually the outlet flow rate and inlet flow rate are made equal.
Thus the error is completely eliminated and the level is made to float at the set
point value. It is important to define the term "Repeats per minute" as applied
to P + I action, which occurs after the derivative action as shown in Fig. 2.15.
The term derives from the observation the integral gain KI, has the effect of
cousing the controller output to change every unit time by proportional mode
amount.
Graphically,
Let e = At

P = Kp At +
Kp
Ti z At dt + Kp T d
d
dt
(At)

P = (KpA) t +
LM A K OP t p 2
+ [A Kp Td] ...(14)
N 2T Q i

The characteristics shown in Fig. 2.15 are governed by dynamics expressed

in equation 14.
 5TH PROOF 4-9-09

78 Process Control Engineering

It’s effect on error can be understood by considering the liquid level

system to be a first order process.
At
e

t
t = t1
I-only
P-only
P

D-only

t
t = t1
tio .
Ac I. D

}
n

P
P.

t
t = t1

FIG. 2.15

S × R( s)
ess = Lt
S®0 1 + G( s) H( s)

A
Let R(s) =
S2

A
S
ess = Lt
S®0 FG 1 IJ FG 1 IJ
1+ Kp 1+
H Ti S
+ Td S
K H 1 + Ts K
A
ess =
L K
0 + M0 + i
+ 0 [1]
OP
N T i Q
(PID controller) Process

1 1
R(s) Kp 1 + + Tds C(s)
Tis 1 + Ts
 5TH PROOF 4-9-09

Modes of Controllers Operations 79

ATi
ess = since Ti << Kp
Kp

ess = 0

The effect of this mode on speed of response or transient state behaviour

can be obtained by considering sinusoidal deviation

let e = sin wt

P = Kp sin wt +
Kp
Ti z sin wt dt + Kp Td
d
dt
sin wt

FG - K IJ cos wt +w K T
p
P = Kp sin wt +
H wT K i
p d cos wt

LM
P = Kp sin wt + w K p Td -
Kp OP cos wt
N wTi Q
2
æ Kp ö é æ 1 öù
P= K p2 + ç wK p Td - ÷ sin ê w t + Tan-1 ç wTd – ú ...(15)
è wTi ø êë è wTi ø÷ úû

Therefore from eqn. (15), that phase of the controller output leads by
FG
Tan–1 wTd -
1 IJ
H wTi K
, which implies it is capable of improving the transient state

behavior of the process.

Therefore the performance of composite controller modes may be
summarised as follows:
(1) P + I controller characteristics are similar to lag compensator or low
pass filter. It is capable of improving only the steady state response
characteristics i.e. the presence of integral action eliminates the offset
of proportional action and also any type of error.
(2) P + D controller characteristics are similar to lead compensator or high
pass filter. It is capable of improving only transient state characteristics
i.e. speed of response of the system. It is best suitable for processes
having very slow response.
(3) P + I + D controller characteristics are similar to LEAD-LAG
compensator. From the pole-zero configuration of lead. Lag
compensator it may be concluded that it exhibits band reject filter
characteristics. It is capable of improving both transient and steady
state response characteristics of the system.
 5TH PROOF 4-9-09

80 Process Control Engineering

SOLVED EXAMPLES
Type-1 Problems
1. Indicate the type of control action and obtain the controller parameters
Kp, T d and Ti for the controller represented as Shown in figure.

E(s)
K + P(s)
+

1
Ts

LM 1 OP E(s)
Solution:
N Ts Q
P(s) = k +

R L 1 OUV E(s).
P(s) = Sk M1 +
T N kTs PQW
It is P + I controller comparing with

P(s) = KP 1 +
LM 1 OP
N Ti s Q
KP = k Ti = kT

E(s) 1
1 + T2s
2. T1s P(s)

RS
P(s) = (1 + T2 s).
1 UV
E(s)
T T1 s W
ì 1 T sü
P(s) = í + 2 ý E(s)
î T1 s T1 s þ

ìï T é T1 1 ù üï
P(s) = í 2 ê1 + . ú ý E(s)
ï T ëê T2 T1 s úû ï
î 1 þ
It is P + I controller.

T2
Kp = Ti = T2
T1
 5TH PROOF 4-9-09

Modes of Controllers Operations 81

E(s) 1 + T1s 1 + T2s 1

3. Ts
P(s)

é T T s2 s [T1 + T2 ] 1 ù
P(s) = ê 1 2 + + ú E(s)
ë Ts Ts Ts û

éT T T +T 1ù
P(s) = ê 1 2 . s + 1 2 + ú E(s)
ë T T Ts û

ìï T + T é T TT T 1 ù üï
E( s). í 1 2 ê . 1 2 .s + 1+ . úý
îï T ë T1 + T2 T T1 + T2 T s û ï
þ

ìï T + T é æ T1T2 ö 1 ù üï
\ P(s) = í 1 2 ê1 + ç T + T ÷ s + (T + T ) s ú ý E(s)
T ë è 1 2ø
îï 1 2 ûþ ï
It is a PID controller comparing with

ì é 1 ùü
P(s) = í K p ê1 + + Td s ú ý E(s)
î ë Ti s ûþ

T1 + T2 T1T2
Kp = Ti = T 1 + T 2 Td =
T T1 + T2

E(s) 1 + T1s 1
4. 1+
T2s
P(s)

é æ 1 öù
P(s) = ê(1 + T1s) ç1 + ÷ ú E(s)
ë è T2sø û

é 1 T sù
P(s) = ê1 + + T1s + 1 ú E(s)
ë T2 s T2 s û

éæ Tö 1 ù
P(s) = êç1 + 1 ÷ + + T1 s ú E(s)
ëè T2 ø T2 s û

éT + T 1 ù
P(s) = ê 1 2 + + T1s ú E(s)
ë T2 T2 s û

ìï T + T é T2 1 T2 ùüï
P(s) = í 1 2
ê1 + . + .T1 s úý E(s)
ï T2 ëê T1 + T2 T2 s T1 + T2 ûúï
î þ
 5TH PROOF 4-9-09

82 Process Control Engineering

ìï T + T é 1 æ T1T2 ö ù üï
\ P(s) = í 1 2 ê1 + (T + T ) s + ç T + T ÷ s ú ý E(s)
ï T2 ë è 1 2 ø ûï
î 1 2 þ
It is a PID controller.
T1 + T2 T1T2
Kp = Ti = T 1 + T 2 Td =
T2 T1 + T2

E(s) T
+
+
T1s P(s)
5. +
T2s

P(s) = [T + T 1s + T 2s]E(s)
= [T + (T1 + T 2]E(s)

ì é æ T + T2 ö ù ü
P(s) = íT ê1 + ç 1 ÷ s ú ý E(s)
î ë è T ø ûþ
It is a P + D controller comparing with
P(s) = [Kp(1 + T d s)]E(s)

T1 + T2
Kp = T Td =
T

E(s) T1s 1
6. 1+
T2s
P(s)

é T sù
P(s) = êT1s + 1 ú E(s)
ë T2 s û

ìT é T2 ùïü
P(s) = ïí 1 ê1 + T . T1 s ú ý E(s)
ï T ë ûï
î 2 1 þ

ìT ü
P(s) = í 1 [1 + T2 s]ý E(s)
T
î 2 þ
It is a P + D controller.
Kp = T1/T2 Td = T 2
 5TH PROOF 4-9-09

Modes of Controllers Operations 83

7.
2
E(s) (1 + Ts) P(s)

P(s) = [1 + T2S2 + 2Ts]E(s)

Put s = d/dt

d2 e de
P = T2 + 2T + e.
dt 2 dt

de
1et = eP.
dt

deP
P = T2
dt
+ 2T . eP + ò e dt .
P

Applying L. T (Laplace transform).

EP ( s)
P(s) = T2. sEP (s) + 2TEP (s) +
s

é 1ù
EP (s) êT 2 . s + 2T + ú .
ë sû

ì é T 1 ùü
P(s) = í2T ê1 + . s + ý EP (s)
î ë 2 2Ts úûþ
It is a PID controller.

\ Kp = 2T Ti = 2T

T
Td =
2

1
8. E(s) T1s P(s)
T2s

éT s ù é T1 ù
P(s) = ê 1 ú E(s) Þ P(s) = ê T ú E(s)
ë T2 s û ë 2û

It is proportional controller

Kp = T1/T2
 5TH PROOF 4-9-09

84 Process Control Engineering

T1
E(s) +
P(s)
9.
+
T2s

P(s) = [T1 + T2 ]E(s).

ì é T ùü
P(s) = íT1 ê1 + 2 . s ú ý E(s)
îï ë T1 û ï þ

It is P + D controller
Kp = T 1 Td = T2 /T1

10. P(s) = [1 + T 1s + T 22 s2]E(s).

d
Put s=
dt

de de p
P = e + T1 + T 22
dt dt

de
let = ep.
dt

de p
P= ò e dt + T e
p 1 p + T 22
dt
It is PID controller.
Applying laplace transform.

é 1 ù
P(s) = êT1 + + T22 s ú E(s)
ë s û

ìï é 1 T 2 ù üï
P(s) = íT1 ê1 + + 2 . s ú ý E(s)
îï ë T1s T1 û þï

T22
Kp = T 1 Ti = T 1 Td =
T1
 5TH PROOF 4-9-09

Modes of Controllers Operations 85

Type-2 Problems
11. In an integral controller with reset time set to 0.6 minutes, what will be the
phase of controller output for sinusoidal deviation?

1
Solution: P=
Ti ò e dt.
Since Ti = 0.6 min. e = sin wt
1
P=
0.6 ò sin wt. dt
- cos wt
P=
w ´ 0.6
Since cos q = sin [90 – q]
– cos q = sin [q – 90°]
sin [ wt - p /2] 1
\ P= Þ sin (wt – p/2)]
0.6w 0.6w
The phase of the controller o/p Lays by p/2 degrees.
12. Determine the phase of the controller output of derivative controller for
sinusoidal deviation?
de
Solution: P = Td Since e = sin wt
dt
d
p = Td (sin wt) = wTd cos wt
dt
P = wTd sin (wt + p/2). [3 sin (90° + q) = cos q]
The phase of the controller o/p leads by p/2.
13. The input error signal for P + I controller is sinusoidal in nature. Prove
that the phase lag is a function of reset time?
Kp
Solution: P = Kpe +
Ti ò e dt
e = sin wt
Kp
p = Kp sin wt +
Ti ò sin wt. dt
é-Kp ù
p = Kp sin wt + ê ú cos wt
ë wTi û
 5TH PROOF 4-9-09

86 Process Control Engineering

A sin wt + B cos wt= A2 + B2 sin [wt + tan– 1 B/A]

é æ - Kp öù
ê 2
ç ú
æ -Kp ö ê wt + tan -1 ç wTi ÷ ú
\ p= K p2 +ç sin ÷
è wTi ÷ø ê èç Kp ø÷ úû
ë

2
æ Kp ö é æ 1 öù
\ P= K p2 +ç ÷ sin ê wt - tan - 1 ç ú
è wTi ø ë è wTi ø÷ û

Thus the phase lag q = tan– 1 (1/wTi) is a function of reset time (T i).
14. A P + I controller subject to sinusoidal deviation, is tuned to T i = 3 min
and Kp = 10. Determine its output when f = 50 Hz.
Kp
Solution: P = Kpe +
Ti ò e dt.
Given e = sin wt; kp = 10; Ti = 3 min
10
P = 10e +
3 ò e dt
( - 3.33)
Þ P = 10 sin wt + cos wt.
w
2
æ 3.33 ö é æ - 3.33/ w ö ù
P = 102 + ç .sin ê wt + tan -1 ç
è w ÷ø ë è 10 ÷ø úû

æ 11 ö é 3.33 ù
P = 102 + ç 2 ÷ .sin ê wt - tan -1
èw ø ë 10 w ûú
w = 2pf = 2 ´ p ´ 50 = 314.15 radians.

11 é 3.33 ù
P= 100 + .sin ê314.15t - tan -1 ú
(314.15)2 ë 10 ´ 314.15 û

P = 10 sin [314.15 + (– 0.06)]

15. What is the phase of the controller output for sinusoidal error for P+D
controller?
de
Solution: P = Kpe + KpT d
dt
e = sin wt
d
P = Kp sin wt + KpTd sin wt
dt
 5TH PROOF 4-9-09

Modes of Controllers Operations 87

P = Kp sin wt + wKpTd cos wt

P= K p2 + (wK pTd )2 . sin [wt + tan– 1 wTd]

The phase leads by tan– 1 (wTd) degrees.

16. A certain controller has a function
1
P=
T2 ò ò edt + P(0).
Prove that the phase of the controller output lags by 180°.
Solution: Applying L.T [Assume Initial Condition to be zero P(0) = 0]

1 E ( s)
P(s) =
T 2 s2
Put s = jw

P ( jw) 1 1 1 1
F( jw) = = = . .
E( jw) ( jw)2 . T 2 T 2 jw jw

1 1 + jw 1 + j 0
F( jw) = .
T 2 0 + jw 0 + jw

1 12 + 02 12 + 02 1
|F( jw)| = . =
T 2 02 + w2 2
0 +w 2 w T22

[tan -1 0/1] [tan -1 0/1]

Ð F( jw) = [0°] .
[tan -1 w /0] [tan -1 w /0]

0° 0°
Þ -1 = = – 180°.
2 T an ¥ 2 ´ 90°
Therefore the phase of the controller 0/p lags by 180°.
17. Show that for a PID controller the phase of its output is a function of rate
time and reset time for sinusoidal deviation?
Kp de
Solution: P = Kpe +
Ti ò e dt + K T p d
dt
e = sin wt
Kp d sin wt
P = Kp sin wt +
Ti ò sin wt + K T p d
dt
 5TH PROOF 4-9-09

88 Process Control Engineering

æ - Kp ö
P = Kp sin wt + ç cos wt + wKpTd cos wt.
è wTi ÷ø

æ Kp ö
P = Kp sin wt + ç wK pTd - cos wt
è wTi ÷ø

é æ Kp öù
2 ê ç wK T - ú
æ Kp ö
ê wt + tan - 1 ç
p d
wTi ÷ ú
P= K p2 + ç wK pTd - .sin ÷
è wTi ÷ø ê çè Kp ÷ø ú
ë û

2
æ Kp ö é æ 1 öù
P= K p2 + ç wK pTd - .sin ê wt + tan -1 ç wTd -
è ÷
wTi ø ë è wTi ø÷ ûú

æ 1 ö
f = tan– 1 ç wTd -
è wTi ø÷

The phase of the controller output is a function of rate time (Td) and reset
time (Ti).
18. A certain P + D controller has Td = Kp = 1. Find the phase of its 0/p
when w = p/3?
Solution: From prob (15). The phase of P + D controller leads by tan– 1(wTd).
\ Given w = p/3 Td = 1
tan–1 (p/3) = 89°.

Type-3 Problems
19. A pneumatic controller is employed to control the valve plug position
between 0 – 10 mm. It the range of the controller output is 3 – 15 psi.
What pressure will bring the valve\plug to a position of 8.5 mm.?
Solution: Since controller final control element operation is linearly
dependent
Y = MX + C Y = valve plug position
0 = M.3 + C X = controller 0/p
10 = M.15 + C
Solving for “M” & “C”
M = 0.833 C = – 2.5.
\ 8.5 = 0.833x – 2.5
8.5 + 2.5 = 0.833x x = 13.20 psi
 5TH PROOF 4-9-09

Modes of Controllers Operations 89

20. In the above problem express 13.20 psi as percentage controller output in
the range of 3 – 15 psi?
Solution: Using the formula
DP - Pmin
%p = ´ 100
Pmax - Pmin

13.20 - 3
= ´ 100
15 - 3
% P = 85%

21. In a 4 – 20 mA current range, it 0% corresponds to 4 mA and 100%

corresponds to 20 mA, and what value of current indicates 42% of output?
Solution: Using the formula
DP - Pmin
%p=
Pmax - Pmin

42 P-4
=
100 20 - 4

42 ´ 16
+ 4 = P.
100
\ P = 10.72 mA

10.27 mA corresponds to 42% of out put.

22. A controller outputs 4-20 mA signal to control motor speed from
140 rpm-600 rpm with Linear dependence
(a) Calculate current corresponding to 310 rpm.
(b) Express part (a) as % of controller output
Solution: (a) Using linear dependence relation
Y = MX + C.
Motor speed ( y) = M[current (x)] + C
\ 140 = 4M + C
600 = 20M + C
Solving for M and C

C = 25 rpm M = 28.75 rpm/mA

\ Y = 28.75(I ) + 25
 5TH PROOF 4-9-09

90 Process Control Engineering

For 300 rpm.

300 = 28.75I + 25
300 - 25
=I
28.75

\ I = 9.91 mA

DP - Pmin
(b) %P = ´ 100
Pmax - Pmin

9.91 - 4
\ %P = ´ 100 = 36.9%
20 - 4

\ % P = 36.9%

23. The controller output in percentage is 54.78%. Determine the

corresponding pressure in the range of 3 – 15 psi?

54.78 DP - Pmin
Solution: =
100 Pmax - Pmin

54.78 DP - 3
\ =
100 15 - 3
\ DP = 9.57 psi

24. In a level control system, the controller output 3 – 15 psi signal to the
final control element to monitor liquid level between 2.5 m – 22.5 m.
What should be the output of controller to maintain level at 19.75 m?
Solution: Using linear dependence equation
Y[level] = M[X (pressure)] + C
2.5 = 3M + C
22.5 = 15M + C
Solving for ‘M’ and C
M = 1.67 m/psi C = 2.5 m
\ 19.75 = 1.67x – 2.5
19.75 + 2.5
\ x= = 13.32 psi
1.67
Therefore the rewired pressure is 13.32 psi
13.32 - 3
%P = ´ 100
15 - 3
% P = 86%
 5TH PROOF 4-9-09

Modes of Controllers Operations 91

25. A liquid level control system converts a displacement of 5 to 10 m into

4 – 20 mA control signal with linear dependence.
A relay, which serves as a two position controller opens or closes the inlet
control valve. The relay closes at 12 mA and opens at 8 mA. What will be
neutral zone in nut ran of the two position controller?
Solution: Using the relation
Y(level) = MX (current) + C
5 = M.4 + C
10 = M.20 + C
Solving for ‘M’ and ‘C’
M = 0.3125 m/mA & C = 3.75 m.
For 12 mA the relay closes
\ Y1 = 0.3125 ´ 12 + 3.75
Y1 = 7.5 m.
The relay opens at 8 mA
Y2 = 0.3125 ´ 8 + 3.75
Y2 = 6.25 m
\ The neutral zone = Y1 – Y2 = 7.5 – 6.25

Neutral zone = 1.25 m

26. In a pressure control system a floating controller with single speed mode
is employed to control pressure in a gas vessel. The controller output
when the process error is within the limits of neutral zone is 35% If the
rate constant KF = +2% per second,
(a) What will be the controller output after 10 seconds.
(b) Find the time when the output saturates
Solution: (a) For floating controller
p = KFt + P(0).
At t = 10 seconds.
P = +2 ´ 10 + 35% = 55%
(b) At p = 100% [output saturation]
100 = +2 ´ t + 55

100 - 55
t= = 22.5 seconds.
2
 5TH PROOF 4-9-09

92 Process Control Engineering

27. In the level control system shown in figure valve ‘A’ is linear one with valve
coefficient of 20 m3/h per percent controller output. In the initial
controller output for inlet flow rate of 300 m2/h is 15%. By how much
percent the valve ‘A’ should be opened/closed if the inlet f low rate
suddenly becomes 600 m3/h. Also calculate offset for setting of
proportional gain (kp) at 10% per %?

Inlet

Controller Set point

[Valve 'A']
h
Outlet

m3
Solution: Let x % be the new controller output; 20 / % ´ x % = 600 m3/h.
h
600
\ x= = 30%
20

x = 30%
Since only proportional gain setting is given the controller is proportional
controller
P = K p e + P0
30% = 10 e + 15%
30 - 15
= e = 1.5%
10
28. Find the response for PI controller with Kp = KI = 1 for the error as
shown in figure for P(0) = 10%

t
0 1 3

Solution: The 0/p equation

P = Kpe + KpKI ò e dt
For 1>t>0 ; e=t
3>t>1 ; e=1
t >3 ; e=0
 5TH PROOF 4-9-09

Modes of Controllers Operations 93

Consider 1>t>0

p = 1.t + 1 ò t + 10%
t2
p=t+ + 10%
2
At t = 1 sec.
1
P=1+ + 10 + 11.5%
2
For 3>t>1

P=1+1 ò 1 dt + 11.5%
P = 1 + t|31 + 11.5%
P = 1 + (3 – 1) + 11.5%
P = 14.5%
For t > 0e = 0
\ P = 14.5%

29. Find the P + D controller output with Kp = KD = 1 for the error shown in
Pwb (28)?

de
Solution: P = Kpe + KpKD
dt
Considering the same initial controller output as in Problem. (28)
For 1>t>0; e=t
d
P=t+ (t) + 10%
dt
At t = 1 sec
P = 1 + 1 + 10% = 12%
For 3>t>1
P = 1 + 0 + 12% = 13%
For t > 3e = 0

\ P = 13%
 5TH PROOF 4-9-09

94 Process Control Engineering

30. Find P + I + D controller output for same settings as in pwb (28) & (29)?
Solution: For 1>t>0 ; e=t
1
t2
P=t+ + 1 + 10%
2 0

At t = 1 sec
P = 1 + 0.5 + 1 + 10% = 12.5%
For 3>t>1; e=1
P = 1 + t°|13 + 0 + 12.5%
1 + (3 – 1) + 12.5%
P = 15.5%
For t >3; e=0
\ P = 15.5%

QUESTION SET
1. What are the types of discontinuous controller modes?
2. Explain the characteristics of two position controller mode?
3. What is floating action? Illustrate with an example?
4. What are the different types of continuous controller modes?
5. Explain the characteristics of
(a) P – controller
(b) I – controller
(c) D – controller
6. Compare and contrast the relative merits and de-merits of above control
actions?
7. For sinusoidal deviation obtain the magnitude and phase angle for P, I
and D actions?
8. What are composite control actions? Explain each one of then?
9. Compare the relative advantages and disadvantages of composite control
actions?
10. Obtain the phase relation for sinusoidal deviation for composite modes?
 !
+0)26-4

Effect of Control Actions

on Processes

3.1 CONCEPT OF CLOSED LOOP AUTOMATIC

CONTROL
This chapter thoroughly deals with the effects of various controller modes on
different types of processes, starting from basic capacitance and time constant
types to complex multiple time constant processes having single and multiple
inputs.
In the last chapter, a comprehensive and easily assimilable analysis of
different modes of controller operation has been presented, with
mathematical, analytical and graphical interpretations of the various actions.
And, since both controller and process together constitute a controlled
system, it is very essential to deal with these controllers affecting the processes
and their characteristics together.
This topic is furthered by considering a process control loop, in feedback
configuration (negative feedback) with the transfer function representations of
various elements constituting the loop.
In analyzing the effects of control actions on processes, the feedback
configuration of process control system is considered. Thus, the automatic
control implies a negative feedback control system whose block diagram
representation is shown in Fig. 3.1.
In Fig. 3.1 the various blocks may be identified as
Gc(s) = Controller function. GP(s) = Process/ plant function.
N(s) = Process load function. R(s) = Input/set point.
u(s) = Load variable (or) Disturbance.
 5TH PROOF 4-9-09

96 Process Control Engineering

U(s)
N(s)

R(s) E(s) +
Gc(s) Gp(s) + C(s)
–

Controller Process

Hm(s) = 1

FIG. 3.1

In any automatic control, analysis of controlled system that is, together

with controller and process is done with respect to changes in set point (or)
input and/or with respect to changes in load variable.
However, in most of the process industries it is the load variables that
disturb the process and its control objective.

Closed Loop Response

Since figure 3.1 represents a multiple input system i.e. R(s), u(s), the closed
loop transfer function when both the input acting simultaneously is obtained
as follows.
The closed loop transfer function when R(s) is acting alone is obtained by
setting u(s) to zero.

R(s) GP(s) GC(s) C(s)

–

C( s) GP ( s)GC ( s)
=
R( s) 1 + GP ( s)GC ( s)

C(s) = R(s)
LM G ( s)G ( s) OP .
P C ...(1)
N1 + G ( s)G ( s) Q
P C

Similarly, the transfer function with respect to load variable u(s) can be
obtained by setting R(s) = 0.

u(s)
N(s) C(s)
+

– [GP(s) GC(s)]
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 97

u(s) 1
N(s) 1 + GP(s) GC(s) C(s)

Therefore,

C ( s) N ( s)
=
u( s) 1 + GP ( s)GC ( s)

C(s) = u(s)
LM N ( s) OP . ...(2)
N1 + G ( s). G ( s) Q
P C

Combining equations (3.1) and (3.2)

C(s) = R(s)
LM G ( s)G ( s) OP + u(s) LM N ( s) OP .
P C
...(3)
N1 + G ( s)G ( s) Q
P C N1 + G ( s)G ( s) QP C

Thus, the transient state analysis of process systems with different

controller modes can be performed either by considering R(s) or u(s) as input
changes from equation (3).
In evaluating the effect of controller on processes another most important
parameter to be considered in addition to transient state behavior (i.e., nature
and speed to response) is its effect on error. Hence, it becomes important to
obtain an expression for error and the related mathematical manipulation to
find the magnitude of error, for changes is set point or load variables.
A particular mode of controller as studied in the previous chapter affects
the process steady state and transient state characteristics.

3.2 ERROR EQUATION AND ERROR TERMINOLOGY

From Fig. 3.1, the expression for error can be obtained as
E(s) = R(s) – C(s).
Since C(s) = E(s)GC(s)GP(s) + u(s)N(s)
E(s) = R(s) – [E(s)GC(s)GP(s) + u(s)N(s)]
E(s)[1 + GC(s)GP(s)] = R(s) – u(s)N(s)
R( s) N ( s)
E(s) = – u(s) . ...(4)
1 + GC ( s)GP ( s) 1 + GC ( s)GP ( s)
By closely observing equations 3 and 4, it is very clearly evident that
controller represented by controller function GC(s) affects the process response
and corresponding error if any, for changes in input/set point R(s) or due to
load variable change/disturbances u(s).
 5TH PROOF 4-9-09

98 Process Control Engineering

Since, magnitude of error described as “ess” is given mathematically as

ess = Lt e(t). ...(5)
t ®¥

Applying final value theorem to Eq. 5

ess = Lt S.E(s). ...(6)
s® o

Substituting Eq. (4) in Eq. (6)

R|SL S. R( s) O – L S.U( s). N ( s) OU| .

V
T|MN1 + G ( s)G ( s) PQ MN1 + G ( s)G ( s) PQW|
ess = Lt ...(7)
s® o C P C P

Thus, from equation (7), it can be stated that magnitude of error, in

principle depends on changes in input or set point R(s) and due to variations
in load variable U(s).

Error Categorization in Process Control

In process instrumentation, the error may be broadly classified into three
types:
® Steady state static error.
® Steady state velocity error.
® Offset

3.2.1 Steady State Static Error

It is defined as the error due to changes in set point or input R(s).
If the controlled variable does not follow exactly, when the input or set
point is suddenly changed from one value to another, then the resulting error
is known as static error.

S. R( s)
ess[Static error] = Lt . ...(8)
s ® o 1 + GC ( s)GP ( s)

In Eq. 8, if the change in set point is sudden in nature [Step change] i.e.,
A
R(s) = [A = Step magnitude]
S

A
s.
s
ess = Lt
s ® o 1 + G ( S )G ( S )
C P

A
ess|static error = . ...(9)
1 + Lt GC ( s)GP ( s)
s®o
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 99

Thus, the magnitude of error is predominantly depending on the function

Lt GC(s) GP(s).
s® o

It is defined as an error constant, known as position error constant

abbreviated as ‘‘ K Posn
  ’’. By properly choosing the mode of controller G C(s), the
error between input and output can be minimized in a particular process.

3.2.2 Steady State Velocity Error

This error results mainly due to steady changes in set point or input R(s) with
respect to time and the response or output attempting to follow the steady
changes in input, and the deviation between them is more phenomenally
termed as velocity error.
Again from equation 7; considering only R(s)
S. R( s)
ess = Lt .
s® o 1 + GC ( s)GP ( s)

A
Let R(s) = [Steadily changing w.r.t time or Ramp change]
s2

A
s.
s2
ess|Velocity error Þ Lt
s ® o 1 + GC ( s)GP ( s)

A
ess|Velocity error Þ . ...(10)
Lt s + Lt s. GC ( s)GP ( s)
s® o s® o

From Eq. (10) it is clear that the magnitude of error for ramp change in
input predominantly depends on.
Lt S. GC(s)GP(s). This error constant is termed as “Velocity error Constant”
s® o
abbreviated as “Kv”.
Similarly, by considering non-linear (or) acceleration type of input, the
error due to such input may be defined as “Steady State acceleration error.”
The non-linear/Acceleration type of input may be considered as a
A
parabolic input where R(s) = .
s3

A
s/ . 3/
\ ess|Acceleration error = Lt s
s® 0 1 + GC ( s)GP ( s)
 5TH PROOF 4-9-09

100 Process Control Engineering

A
ess|Acceleration error = .
Lt s + Lt s2GC ( s)GP ( s)
2
s ®0 s ®0

Defining Acceleration error constant “KA” as Lt s2GC(s)GP(s), the

s®0
magnitude of this error is depending on the magnitude of this error constant.
Hence, it may be generalized as the steady state error, velocity error (or)
acceleration error’s magnitude depends on the respective error constants
K (pos'n ) KV and KA, where

KP = Lt GC(s)GP(s) and KA = Lt S2GC(s)GP(s)

s® o s®0

KV = Lt SGC(s) GP(s).
s®0

And these error constants are further the function of controller i.e., GC(s).
A controller, properly chosen and tuned definitely improves the transient
and steady state responses of the process.

3.2.3 OFFSET
In many of the process industries, the changes in load variables or
disturbances affect the performance of process or plant. This error which
results due to changes in load variables is termed as “offset”. This offset error
should not be confused with that of offset of proportional controller. Here, it is
a special type of error which occurs mainly because of changes in load
variables.
In eqn. 7

ess|offset = Lt
LM – S.U ( s). N ( s) OP . ...(11)
s ®0
N1 + G ( s)G ( s) Q
C P

Thus, offset can be computed directly by using eqn. (11).

To summarize the performance of a controller in process control system
for different types of processes can be very well analyzed using eqns. (3) and
(7), respectively.

3.3 EFFECT OF TWO POSITION CONTROLLER

The analytical expression of a two position controller is reproduced again to
understand its effect on the process.
P = 0% e < 0
P = 100% e > 0.
The above equation implies, that whenever there is error i.e., when the
controlled variable exceeds the set point or less than the set point value the
controller output is 100% in terms of ON (or) OFF. Another important aspect
of this control action is “Neutral Zone”.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 101

This neutral zone often referred to as “differential gap” (or) sometimes

“Hysteresis zone” occurs in the operating point of two position controller
whenever the controller switches between ON and OFF states as the controlled
variable changes with respect to set point i.e., at the occurrence of error.
Consider a single capacitance process monitored by this controller as
shown in Fig. 3.2. The level controller LC turns on the inflow when the level is
below the set point and turns off the control valve when the level of the liquid
in the tank is above the set point value. In this process the level of the liquid in
the tank is the controlled variable.

Control
qi valve
Inlet LC Level Controller

h LT (Level transducer)
u
Outlet

FIG. 3.2

Let qi = Inlet flow rate

u = load variables = outflow
h = level (controlled variable).
c = vessel capacitance.
The process equation (or) transfer function is
dh
qi – u a .
dt
dh
qi – u = c . ...(12)
dt
Applying L.T.
Qi(s) – u(s) = CS .H(s)
1 1
Þ H(s) = .Qi(s) – . u(s).
cs cs

V(s) 1
cs

Qi(s) 1 –
cs + H(s)

FIG. 3.3
 5TH PROOF 4-9-09

102 Process Control Engineering

The control loop may be formulated as

U(s) 1
Q(s)
cs

Set point ON –
1
+ H(s)
R(s) – OFF cs
Controller Process

FIG. 3.4

In Fig. 3.2, when the valve is open i.e., [H(s)(h) < (r)(R(s)] [level less than
set point value], where
h = o/P (level) and r = set point.

dh
qi – u = c [h < r + N]
dt
N = [Positive Limit of Neutral Zone]
When the valve is closed i.e.,
h > r, qi = 0
dh
–u = c [h > r – N] [N = – ve limit of neutral zone]
dt
When the level is exactly at the set point value i.e., when h = r, the controller
oscillates continuously between the ON and OFF states, explained analytically
by integrating the above two equations.

z dh
dt
dt = z qi – u
c
dt

qi – u
h(t) = t + h01 [h01 = integral constant.] ...(13)
c
Similarly,

z dh
dt
. dt = z –u
c
. dt

–u
h(t) = . t + h02 [h02 = Integral constant.] ...(14)
c
From equations 13 and 14, the level rises and falls continuously. within the
limits of neutral zone ±N when the error is zero, i.e., when c = r.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 103

Level
h(t) Limits of
Neutral
+N –u t zone.
qi – u t c
O
c
–N

Time

FIG. 3.5

3.4 EFFECT OF PROPORTIONAL CONTROLLER

MODE
The controller equation (or) transfer function of this mode is
P( s)
= KP ; KP = Proportional Gain.
E( s)
Consider the single capacitance process shown in Fig. 3.2 and its block
diagram representation shown in Fig. 3.4. The ON/OFF control action is
replaced by proportional controller and the control valve is placed at the inlet
flow line.
The inlet flow rate ‘qi’ and load variable ‘u’ are considered as inputs
causing changes in level thereby causing error. Thus, the effect of this
controller on deviation of the controlled variable from set point value [error] is
interpreted in terms of transient state and steady state characteristics either by
considering inlet flow rate changes (qi) or load variable changes (u).

3.4.1 Single Capacitance Process

Controller Process
Qi(s) 1
Kp + H(s)
– cs –

U(s) 1
cs

FIG. 3.6

For multiple input-multiple output linear process control system, the

analysis is done by considering only one input and one output at a time.
 5TH PROOF 4-9-09

104 Process Control Engineering

Qi(s) 1
Kp H(s)
– cs

FIG. 3.7

Let U(s) = 0.
The overall transfer function

KP KP
H( s) cs K P C
= = = . ...(15)
Q ( s) KP cs + K P K
1+ S+ P
cs C
The nature of response with proportional controller can be obtained by
applying inverse laplace transform to eqn. 15.

æ KP ö
K –ç ÷t
\ h(t) = P . e è C ø
. ...(16)
C
Comparing e– (KP/c)t » e– t/T.
1 K C
= P Þ T=
T C KP

The time constant element (T), which indicates the speed of response is
inversely proportional to proportional gain. Thus, as KP ­ , time constant (T)
¯ and speed of response increases.

h(t)
KP
[ K P1 > K P2 ]
c

– KP t
KP
h(t) = e c
c
K

KP = KP2
P

K
=

P1

FIG. 3.8

Similar response characteristics may be obtained even by considering the

load variable u(s) as dominant disturbing factor in a process.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 105

Another most important consideration in discussing the effects of

controller on processes is their effect on error. It may be recalled, that
proportional controller contributes an error called as ‘offset’ at its operating
point.
For single capacitance process, the magnitude of this offset is determined
from the error equation eqn. 8.

S. R( s)
ess|Static error = Lt
s ®0 1 + GC ( s)GP ( s)

1
Gc(s) = KP and GP(s) = .
cs
Since u(s) i.e., outlet flow rate is assumed to be zero.
R(s) = Qi(s) = [continuously changing quantity with respect to time]
A
Thus, Qi(s) = [Ramp change].
s2

A
s.
s2
ess|Static error Þ Lt
s®0 K
1+ P
Cs

A A
Þ Lt = = C .
s®0 s . KP KP
s+
Cs
Thus, the steady state error also called as offset of proportional controller
is inversely proportional to ‘KP’.
An increase in KP will improve the speed of response as well as reduces the
error.
But a very large increase in proportional gain ‘KP’ implies less
proportional band. This converts the operation of proportional mode into an
ON/OFF (two position mode) action.
To summarize, the proportional gain KP, no doubt improves both transient
and steady state responses but at the same time its value should not be large
because the proportional band may become narrow resulting in ON/OFF
characteristics.

3.4.2 Single Time Constant Process

Let us see the effect of proportional controller on single time constant process
as shown in figure.
 5TH PROOF 4-9-09

106 Process Control Engineering

Set point
qi
Inlet
LC Level controller

LT = level transmitter
or LT
level transducer h (Control valve)
q0
Outlet
FIG. 3.9
The process equation/transfer function from chapter (1)

Qi(s) R
H (s )
T = RC
Ts + 1
R = Control valve characteristics
C = Area/capacity of tank (m2).
The control loop with level controller (proportional mode) will be

Qi(s) R H(s)
KP
– Ts + 1

RK P
H ( s) 1 + Ts RK P
= =
Q i ( s) RK P Ts + ( RK P + 1)
1+
1 + Ts

RK P
H ( s) T
= . ...(17)
Q i ( s) (1 + RK P )
S+
T
The transient response is obtained by applying inverse Laplace transform
to eqn. 17.

æ RK P + 1ö
RK P e – èç T ÷ø t
h(t) = . .
T
Comparing

æ RK P + 1ö
–ç
è T ÷ø t
e » e–t/T
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 107

1 RK P + 1 RK P + 1
= =
T T RC

RC
T= .
RK P + 1

The stabilization time which is defined as time taken by the response to

reach 98% of the final value is approximately four times of time constant.

4 RC
Stabilization Time [ts] = 4T = . ...(18)
RK P + 1

The steady state error analysis is obtained by considering the error

equation.

S. R( s)
ess|Static error = Lt
s ® 0 1 + GC ( s)GP ( s)

R
where, Gc (s) = KP ; Gp (s) = and R(s) = Qi(s).
1 + Ts

Let there be sudden change in inlet flow rate (qi) causing the level to
change from the set point value, i.e., Qi(s) = A/s (step change).

A
s.
s
ess|Static error = Lt .
s®0 RK P
1+
1 + TS

A
ess|Static error = . ...(19)
1 + RK P

Eqn. 18 and 19 completely detail the effect of proportional controller on

the single time constant process. There will always be a finite error (offset of
p-controller) which is inversely proportional to gain KP as shown in eqn. 19.
An increase in KP will reduce the stabilization time and error (offset) but
practically its value cannot be very large for the reasons already discussed.

3.4.3 Single Time Constant Process with Multiple

Inputs
Consider a liquid level system having load variable as shown in Fig. 3.10.
 5TH PROOF 4-9-09

108 Process Control Engineering

qi
m[Load variable]

LC T = Time constant = RC

LT
h 'c' 'R'

q0

FIG. 3.10

The process equation/ transfer function is

U(s) R
Ts + 1

+ H(s)
+
Qi(s) R
Ts + 1

Since, the transient analysis is performed either by considering u(s) or

Qi(s), it is similar to the analysis done in the previous case i.e., equations 17 and
19 are valid to this system also.

3.4.4 Effect of Proportional Controller on Multiple

Time Constant Process
A multiple time constant process basically implies a higher order system.
Consider two tanks connected in a non-interacting mode as shown in figure
3.11.

qi

t1 = R1C1 ® for tank (1)

t2 = R2C2 ® for tank (2)
h1 c1 R1
m[Load variable]
Tank (1) q1

h2 c2 R2

q0
Tank (2)

FIG. 3.11
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 109

The process equation/transfer function is

For tank (1)
dh1
qi – q1 = c1 .
dt
dh1
q i = c1 + q1 .
dt
h1
Since q1 = .
R1

dh1 h1
qi = c1 + .
dt R1

dh1
R1qi = R1C1 + h1 .
dt
dh1
R1qi = T1 + h1 . ...(20)
dt
For tank (2)
dh2
qi + u = c2 + qo .
dt

q1 + u = c2
dh2 h2
+
hLM
qo = 2 .
OP
dt R2 R2 N Q
dh2
R2[q1 + u] = R2c2 + h2 .
dt
h1
Since q1 =
R1

R2
LM h
1 OP
+ u = R2 c2
dh2
+ h2 .
NR1 Q dt

R2 dh
h1 + R2u = T 2 2 + h2. ...(21)
R1 dt
Applying L.T to eqn. 20 and 21
R1Qi(s) = [T1s + 1]H1(s)
R1
H1(s) = .Qi (s).
T1 s + 1

R2
.H1(s) + R2u(s) = [T 2s + 1]H2(s)
R1
 5TH PROOF 4-9-09

110 Process Control Engineering

u(s) R2
T2s + 1

Qi(s) R1 H1(s) R2/R1 +

+ H2(s)
T1s + 1 T2s + 1

FIG. 3.12

H2(s) = o/p and Qi(s) and u(s) are input and load variable, respectively.
Since, the overall output of Fig. 3.11 is considered as level in tank (2) i.e.,
‘h2’ a control loop may be formulated either by considering the manipulating
variables to be
qi = inlet flow rate to tank (1)
u = load variable to tank (2)
q0 = outlet flow rate of tank (2) and/or.
q1 = outlet flow rate of tank (1).
The various control schemes are shown in the following figures, 3.13 to
3.15.
qi

h1 R1

q1 Set point

LC
LT
h2 R2

q0

FIG. 3.13

qi

h1 c1 R1
u LC Set point

q1

LT R2
h2 c2
q0

FIG. 3.14
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 111

qi

h1 c1 R1

u LC Set point

q1

LT
h2 R2

q0

FIG. 3.15

Considering closed loop automatic control for the above process.

u(s) R2
T2s + 1

Qi(s) R1 R2/R1 +
KP + H (s )
– T1s + 1 T 2s + 1

Controller Process

FIG. 3.16

Let u(s) = 0.

Qi(s) KPR2
H(s)
– (T1s + 1)(T2s + 1)

K P R2
H ( s) ( 1 1)(T2 s + 1)
T s + K P R2
= =
Q i ( s) K P R2 (T1 s + 1)(T2 s + 1) + K P R2
1+
(T1 s + 1)(T2 s + 1)

H ( s) K P R2
= 2
Q i ( s) T1T2 s + s[T1 + T2 ] + ( K P R2 + 1)
 5TH PROOF 4-9-09

112 Process Control Engineering

K P R2
H ( s) T1T2
Þ ...(22)
Q i ( s) 2 s[T1 + T2 ] K P R2 + 1
s + +
T1T2 T1T2
Comparing the characteristic equation of 22 with std. characteristic
equation of second order system s2 + 2x w ns + w n2 = 0,

T1 + T2 K P R2 + 1
2xwn = ; wn 2 =
T1T2 T1T2

K P R2 + 1
wn = (r/s).
T1T2

K P R2 + 1 T + T2
2x = 1
T1T2 T1T2

F R K + 1I = F T + T I 2
4x 2 GH T T JK GH T T JK
2 P

1 2
1

1 2
2

T1T2 (T1 + T2 ) 2 1
x2 = . .
R2 K P + 1 (T1T2 ) 2 4

(T1 + T2 )2
x= . ...(23)
4T1T2 ( R2 K P + 1)

As seen from the above analysis, the response of the two time constant
process with proportional controller depend on two factors.
(1) Characteristic time, defined as reciprocal of undamped natural
frequency (wn).

1 T1T2
T= = .
wn K P R2 + 1

(2) Damping ratio (x)

(T1 + T2 )2
(x) = .
4T1T2 ( R2 K P + 1)

The proper choice of proportional gain (KP) affects the damping ratio of
the system which may give rise to different types of responses such as over
damped, under damped, and critically damped responses.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 113

1
Since, x a , for smaller values of ‘K P’ the response may be faster
KP
(x = 1) or sluggish (x > 1). As KP is increased, the response exhibits damped
oscillations (x < 1).
Hence, a compromise value of ‘KP’ such that the damping ratio is 0.3 is
generally considered to be satisfactory.
From eqn. 23,

4x 2 ( R2 K P + 1) (T1 + T2 ) 2
=
T1T2 (T1T2 ) 2

(T1 + T2 ) 2
R2KP + 1 =
T1T2 . x 2 .4

T12 + T22 + 2T1T2

R2KP + 1 =
4x 2T1T2

1 é T1 T2 ù
R2KP + 1 = 2 ê
+ + 2ú
4x ë T2 T1 û

ìï 1 é T1 T2 ù üï
Þ R2KP = í 2 ê + + 2ú ý – 1. ...(24)
îï 4x ë T2 T1 û ïþ
The optimum value of proportional gain may be calculated from eqn. 24
for this particular two time constant process.

Error Computation
Error may be computed by considering sudden changes either in the inlet flow
rate qi or the load variable ‘u’.
In Fig. 3.26, let there be a sudden change in inlet flow rate Qi(s)
S . R( s)
ess|Static error = Lt
s®0 1 + Gc ( s)GP ( s)
(offset of proportionnal controller)

S . Q i ( s)
ess|Static error = Lt
s ®0 K P . R2
1+
(T1 s + 1)(T2 s + 1)

1
let Qi(s) = (Sudden change)
S.
 5TH PROOF 4-9-09

114 Process Control Engineering

1s.
s
ess|Static error = Lt
s ®0 K P R2
1+
(T1s + 1) (T2 s + 1)

1
ess|Static error = . ...(25)
1 + K P R2

R2 KP = 0
Error
esc

R2 KP = 1

R2 KP = 10
offset
t

FIG. 3.17

As KP increases, offset considerably gets reduced. However, the value of KP

should not be very large for the reasons already discussed in earlier chapter
and analysis for single time constant process.

3.5 EFFECT OF INTEGRAL CONTROLLER MODE

The controller transfer function is given by
P( s) 1
= (Ti = Integral time or Reset time.)
E( s) Ti s
The process control loop with controller acting on integral mode is shown
in figures 3.18 and 3.19 for single time constant process.
qi = Inlet flow rate
u = downstream head (load variable).
qi
LC
c = capacity of tank (Area) (m2)
R = Resistance Element.

h "c" R u

FIG. 3.18
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 115

V(s) 1
1 + Ts

Qi(s) 1 R + H(s)
Tis 1 + Ts +
–

Controller Process

FIG. 3.19

The effect of integral controller on the transient state behaviour of the

plant or process is obtained either by considering Qi(s) or V(s) as input.
Let V(s) = 0

Qi(s) 1 R
1 + Ts H(s)
– Tis

R
H( s ) Ti s(1 + Ts) R
= Þ
Q i ( s) R 2
TiTs + Ti s + R
1+
Ti s(1 + Ts)

R
H( s ) i .
TT
= . ...(26)
Q i ( s) 1 R
s2 + . s2 +
T TiT .
Comparing the characteristic equation of eqn. 3.26 with standard second
order characteristic equation i.e., s2 + 2x wns + w n2.
1 R
s2 + .s + » s 2 + 2 x wn s + wn 2
T TiT

R
wn = r/s.
TiT .

1 R 1
2x w n = Þ 2x = ...(27)
T TiT T.

4x 2 R 1
Þ = Þ 4x 2RT = Ti
Ti T T 2
 5TH PROOF 4-9-09

116 Process Control Engineering

The integral or reset time Ti is directly proportional to the damping factor,

squared.
Since, the response of the system is dependent on the optimum setting of
damping ratio value, proper tuning of “Ti” enables a satisfactory response of
1
the system. A moderate damping usually about rd is considered to be best
3
1
optimum value. By choosing the value of damping ratio to be , the integral
3
time

1
Ti = 4 ´ ´R´T
(3)2

4 RT
Ti = . ...(28)
9
This setting of Ti results in minimum deviation of the response from the
set point or input value and also avoids encountering excessive oscillations in
the response.
From the error terminology, the integral controller effect can be analyzed
in terms of state error and offset (error due to load variables). The expression
for state error is

S . R( s)
ess|Static error = Lt .
s ® 0 1 + Gc ( s)GP ( s)

S . u( s). N ( s)
And ess|offset = Lt .
s®0 1 + Gc ( s)GP ( s)

Let there be a sudden change in Qi(s).

S . Q i ( s)
ess|Static error = Lt .
s®0 1 + Gc ( s)GP ( s)

1
Qi(s) = [unit step input].
s

1
s.
s
ess|static error = Lt .
s®0 R
1+
Ti s(1 + Ts)

R
where, Gc(s)GP(s) =
Ti s(1 + Ts)
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 117

1
ess|static error = Lt
s®0 R
1+
Ti s(1 + Ts)

1
ess|Static error = = 0.
1+ ¥
Similary, let there be a sudden change in load variable u(s).
1
u(s) = [Unit step change].
s
– S . u( s). N ( s)
ess|offset = Lt .
s®0 1 + Gc ( s)GP ( s)
From the block diagram representation shown in Fig. 3.19.
1 R
N(s) = GP(s) =
1 + Ts 1 + Ts

1
Gc(s) = .
Ti s .

1 1
- s..
s 1 + Ts
ess|offset = Lt
s ®0 R
1+
Ti s(1 + Ts)

1
ess|offset = – = 0.
1+ ¥
The integral controller is very effective in improving the steady state
response characteristics as it completely eliminates the error between input or
set point and controlled variable whether due to variations in the input or set
point or due to load variable changes.
Eqn. 28 may be utilized for obtaining properly tuned value of integral
time ‘Ti’ depending on the dynamics and characteristics of process with which
it is employed.

3.5.1 Effect of Integral Controller on Multiple Time

Constant Process
Consider a two time constant process with controller acting in integral mode,
refer figure 3.13.
The block diagram representation is reproduced as follows.
 5TH PROOF 4-9-09

118 Process Control Engineering

V(s) R2
T2s + 1

Qi(s) 1 R1 R2/R1 +
+ H (s )
– Tis T1s + 1 T 2s + 1

Controller Process

FIG. 3.20

To understand the effect on transient state behaviour of the plant, let

U(s) = 0 [considering only one input at a time].

Qi(s) 1 R2
Tis H(s)
– (T1s + 1)(T2s + 1)

R2
(Ti s)(T1 s + 1)(T2 s + 1)
H( s ) R2
= 1+ 2
Q i ( s) (Ti s)[T1T2 s + s(T1 + T2 ) + 1]

H ( s) R2
= 2
Q i ( s) Ti s[T1T2 s + s[T1 + T2 ] + 1] + R2

H( s ) R2
= 3 2
...(29)
Q i ( s) TiT1T2 s + s Ti (T1 + T2 ) + Ti s + R2
with integral controller, the characteristic equation of the overall system
function has increased to third order polynomial.
Any higher order system’s nature of response is obtained by
approximating it to second order system with respect to dominant pole
analysis. Even in this case also, the integral time T i is a function of damping
ratio (y), which is now called as relative damping factor or damping ratio.
The effect of integral action, whether single time constant process or
multiple time constant process is that it induces oscillations in the nature of
response of the system. The number of oscillations and their peak values can
be monitored by appropriately choosing the value of damping factor, which in
turn depends on the optimum setting of integral or reset time “Ti”. However,
as the order of system is increased with this mode of action, there is definitely
a possibility for the response to exhibit sluggishness. The response of the
system with integral controller for different values of integral or reset time “Ti”
settings are shown below.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 119

Case 1: Case 2:
Ti = 0.75 Ti = 0.25

h (t)
o/p

MP1 MP2
h (t)
o/p
Steady
state
input

tr1 t t
t= o t = o tr2
ts1 ts2

FIG. 3.21

As seen from the above two cases as the value of Ti is reduced, the rise time
decreases and, the number of damped oscillations in the response increases with
maximum peak overshoot shooting up resulting in more settling time.
Note: MP = Maximum peak overshoot
t s = Settling time.
tr = Rise time.
S . Q i ( s)
ess|Static error = Lt
s®0 1 + GC ( s)GP ( s)

1
s.
s
= Lt
s ®0 1 R2
1+ ´
Ti s (T1s + 1) (T2 s + 1)

1
ess|Static error = = 0.
1+ ¥

– S . u( s). N ( s)
ess|offset = Lt
s®0 1 + GC ( s)GP ( s)

1 R2
.
-s.
s T2 s + 1
= Lt
s ®0 1 R2
1+ ´
Ti s (T1s + 1) (T2 s + 1)
 5TH PROOF 4-9-09

120 Process Control Engineering

– R2
ess|offset = = 0.
1+ ¥
Thus, both the error i.e., due to changes in input and load variables are
successfully eliminated by integral action, even for two time constant process.

3.6 EFFECT OF DERIVATIVE CONTROLLER MODE

The transfer function representation of this controller mode is

E(s) T dS P(s)

Let us consider the effect of this mode on single time constant process as
shown in Fig. 3.22.

qi
LC

h "c" 'R' u

FIG. 3.22

The block diagram showing the controller and process is presented below.

V(s) 1
1 + Ts

Qi(s) R + H(s)
Tds +
– 1 + Ts

Controller Process

FIG. 3.23

Considering either Qi(s) or V(s) as dominant input to evaluate the transient

state characteristics.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 121

Qi(s) R
Td s H(s)
– 1 + Ts

RTd s
H ( s) 1 + Ts RTd s
= =
Q i ( s) RTd s 1 + Ts + RTd s
1+
1 + Ts

H ( s) RTd s
=
Q i ( s) s(T + Td R) + 1

RTd s
H ( s) T + Td R
= . ...(30)
Q i ( s) [ s + 1/ T + Td R ]
Consider a sudden change in inlet flow rate ‘qi’ (step change).
1
Let Qi(s) =
s

RTd s
T + Td R 1
H(s) = ´ .
1 s
s+
(T + Td R )

RTd
T + Td R
\ H(s) = .
1
S+
(T + Td R)
Applying final value theorem
Lt h(t) = Lt S . H ( s)
t®¥ s ®0

RTd
S.
T + Td R
\ Lt = 0.
s®0 1
S+
(T + Td R)
Þ h(t) = 0
From the above analysis it is evident that for step change in deviation, the
derivative controller does not respond i.e., its output to the final control
element is zero and hence there is no effect on the output which is level of the
liquid in the tank [h(t) = 0].
 5TH PROOF 4-9-09

122 Process Control Engineering

Let us assume the inlet flow to be changing at some constant rate,

mathematically simulated as ramp change.

F RT s I
H( s)
GH T + T . RJK
d
d
= .
Q i ( s) 1
S+
(T + Td R)

1
Considering Qi =
s2

RTd s
T + Td R 1
H(s) = . .
1 s 2
s +
T + Td R
Again applying final value theorem

RTd
T + Td R
Lt S.H(s) = Lt s . .
s®0 s®0 s
s2 +
T + Td R

RTd
T + Td R.
Þ Lt = RT d.
s®0 1
S+
T + Td R.
This implies that whenever there is a continuous constant change in inlet,
disturbance i.e., load variable or even the set point causing the output i.e.,
liquid level in this example, to also change at a particular rate, the derivative
controller sends a signal to the final control element initiating a sudden action
before the error begins to occur as is illustrated graphically in figure 3.24.

Error
Due to
changes in
inlet flow
rate t
t = t1

Derivative
controller
o/p
} RTd
t

FIG. 3.24
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 123

This type of response results in large errors rather than reducing it to zero
and bringing the controlled variable to the required set point value.
Evaluating error
S . Q i ( s)
e ss|Inlet flow changes = Lt
s ® 0 1 + GC ( s)GP ( s)

1
s.
s2
Þ Lt
s ®0 T s. R
1+ d
1 + Ts

1
= Lt = ¥.
s ®0 s2 RTd
S+
1 + Ts
Infinity error implies, very large error may result due to derivative action.

3.7 EFFECT OF COMPOSITE CONTROLLER MODES

The different combinations of individual continous controller modes are
known as composite controller modes
The three basic combinations, as already studied are reproduced below.
(1) proportional + Integral mode
(2) Proportional + Derivative mode
(3) proportional + Integral + Derivative mode.
The effects of individual modes have been discussed at great length in the
preceding sections of this chapter. From the analysis the various advantages
and disadvantages can be deduced. So to overcome certain short comings of
individual controller actions on processes and to provide better overall control
in terms of speed of response i.e. transient state and action on error i.e. steady
state analysis respectively, the composite controller modes are formulated.
In this section, the effects of there action on processes will be taken up.

3.7.1 Effect of Proportional + Integral Controller

Mode [2 + 1 mode]
Consider a single time constant liquid level process shown in figure 2.28. In
the block diagram representation of closed loop automation the controller
transfer traction is replaced with that of P + I mode function.
 5TH PROOF 4-9-09

124 Process Control Engineering

The analytic equation and transfer function of P + I controller is

Kp
P = Kpe +
Ti ò e dt
E(s) P(s)
KP 1 + 1
Ts i

P ( s) é 1 ù
= Kp ê1 +
E( s) ë Ti s úû

1
1 + Ts

Qi(s) R H(s)
KP 1 + 1 +
+
– Tis 1 + Ts

Controller Process

FIG. 3.25

let u(s) = 0, since analysis is done by considering only one input and output at
active for a transfer function

æ H ( s) ö
The overall transfer function ç can be obtained by considering the
è Q i ( s) ÷ø
block diagram shown in Fig. (3.25).

Qi(s) R H(s)
KP 1 + 1
– Ts i 1 + Ts

é K pTi s + K p ù é R ù
ê úê ú
æ H ( s) ö ë Ti s û ë1 + Ts û
\ çè Q ( s) ÷ø =
i
( K pTi s + K p )( R)
1+
Ti s(1 + Ts)

RK pTi s + RK p
Þ
Ti s(1 + Ts) + RK pTi s + RK p
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 125

RK p [Ti s + 1]
Þ 2
TTs
i + Ti s + RK pTi s + RK p

RK p (1 + Ti s)
TT
i
Þ
s( Ti + RK p Ti ) RK p
s2 + +
Ti T TiT

RK p (1 + Ti s)
H ( s) TiT
Þ s(1 + RK p ) RK p ...(30a)
Q i ( s) 2
s + +
T Ti T

Since the transfer function is of second order comparing with standard

characteristic equation for commenting on the effect of P + I controller on the
overall transient state characteristics. Since for a second order or higher order
system the nature of response in terms of speed, maximum overshoot, settling
time and other performance indices depends on the value of the damping
factor ‘x’ and natural frequency of oscillations in response denoted as “wn”, it
is essential to relate the P + I constants i.e. proportional gain and (kp) and reset
time (Ti) with the above two second order characteristics ‘x’ and ‘wn’
respectively.
Consider the characteristic equation of the transfer function 30a
s[1 + RK p ] RK p
s2 + + =0 ...(31)
T TiT
Comparing with standard second order characteristic equation
s2 + 2xwns + wn2 = 0

RK p RK p
w2n = Þ wn = ...(32)
TiT TiT

The characteristic time is defined as reciprocal of ‘wn’.

1 TiT
\ Tc = = ...(33)
wn RK p

1 + RK p
Now, 2xwn =
T

RK p 1 + RK p
2x =
TiT T
 5TH PROOF 4-9-09

126 Process Control Engineering

4 x2 RK p (1 + RK p )2
=
Ti T T2

Ti (1 + RK p )2
4x2 =
RK p

Ti (1 + RK p )2
x2 =
4 RK p

Ti (1 + RK p )
Þ x= ...(34)
4( RK p )

Equations 33 and 34 relate the parameters ‘K p’ and ‘Ti’ with damping

factor and frequency ‘wn’ we shall consider the effect of adjustment of ‘Kp’ and
Ti on the response after evaluating the error and offset for this process with
P + I controller.
The static error given by equation 8 is
s . R( s)
ess|static error = Lt
s®0 1 + Gc ( s)GP ( s)
Here R(s) = Qi(s)

é 1 ù
Gc(s) = Kp ê1 + ú
ë Tis û

R
GP(s) =
1 + Ts
Assuming step change in input Q i(s).

1
s.
s
ess|static error = Lt
s®0 æ 1 öæ R ö
1 + K p ç1 +
è Ti s ø÷ çè 1 + Ts ÷ø

1
= =0
1 + K p (1 + ¥)( R)

Similarly evaluating offset due to disturbances in load variables.

é - s. u( s). N ( s) ù
ess|offset = Lt ê
s® 0 1 + Gc ( s)GP ( s) ú
ë û
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 127

Here u(s) = load variable

1
N(s) =
1 + Ts

æ 1 ö
Gc(s) = Kp ç1 +
è Ti s ø÷

R
Gp(s) =
1 + Ts

Again considering step change in load variable

u(s) = 1/s

1 1
- s.
.
s 1 + Ts
ess|offset = Lt
s®0 æ 1 öæ R ö
1 + K P ç1 +
è Ti s ÷ø èç 1 + Ts ø÷

-1
ess|offset = =0
1 + K P (1 + ¥)( R)

Hence it may be concluded that in both the cases the error is eliminated
completely. Thus a P + I controller is effective in reducing the error to a
minimum valve let us see the effect of proportional gain and integral time on
the response. For a constant value of integral/Reset time Ti and changing the
value of proportional gain, for higher values of proportional gain the response
exhibits more number of damped oscillations as a result more peak overshoots
which clearly indicates that the damping factor value is between 0 and 1.

Ti = Constant

Curve - 1
Error e

Curve - 2
Curve - 3

O
Curve - 4
Time (t)

FIG. 3.26
 5TH PROOF 4-9-09

128 Process Control Engineering

Curve–1 Indicates the response for a valve of ‘Kp’ between 1 and 10. The
response does not exhibit any oscillation and is a smooth curve decaying to
zero error position.
Curve–2 Is for the valves of proportional gain ‘Kp’ chosen between 10 and
20. Very small oscillations can be observed.
Curve–3 Is the response characteristics for 50 > Kp > 20. The response for
this setting exhibits oscillations in response before reaching the steady state
value.
For higher values of proportional sensitivity there are considerable
overshoots and undershoots in the response before decaying to zero error
position or steady state valve of the input or set point.
Similarly, the effect of reset time on the characteristics of response also can
be analyzed by keeping the proportional gain constant. Figure 3.26 indicates
the nature of response for different valves of ‘Ti’.
Error e

Curve 1

Curve 2
Curve 3
O
Curve 4

Time (t)

FIG. 3.27

For very larger values of Ti (Ti = ¥), the number of oscillations in the
response are very less and the error is also considerable between input and
output as shown in Curve–1.
The response shown in curve (2) is for valves of Ti between 10 and 15.
However when Ti valve is chosen between 0 and 1 [0 < Ti < 1].

3.7.2 Effect of 2 + , Controller mode

From the earlier chapters, the transfer function of Proportional + Derivative
controller is
P(s) = [KP(1 + Tds)]E(s)
To understand the effect of derivative time ‘Td’ and proportional gain “Kp”
let us take an example of multiple time constant process as shown in figure.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 129

Controller
Process
Qi(s) R1 R2/R1 H(s)
KP[1 + Td s]
– T1s + 1 T2s + 1

FIG. 3.28

The overall transfer function

R2 K P (1 + Td s)
H ( s) ( 1s + 1)(T2 s + 1)
T
=
Q i ( s) R K (1 + Td s)
1+ 2 P
(T1 s + 1)(T2 s + 1)

H ( s) R2 K P (1 + Td s)
\ = ...(35)
Q i ( s) (T1 s + 1)(T2 s + 1) + R2 K P (1 + Td s)

The characteristic equation of the transfer function shown above will be

(T1s + 1) (T2s + 1) + R2KP(1 + Td s) = 0
T1T2 s2 + s[T1 + T2 + RKPTd] + (1 + R2KP) = 0

s[T1 + T2 + R2 K P Td ] 1 + R2 K P
Þ s2 + + =0
T1T2 T1T2

1 + R2 K P
wn =
T1T2

T1 + T2 + R2 K P Td
2xwn =
T1T2

1 + R2 K P T + T2 + R2 K P Td
2x = 1
T1T2 T1T2

4x2 (1 + R2 K P ) (T + T2 + R2 K P Td )
= 1
T1T2 (T1T2 )2

(T1 + T2 + R2 K P Td )2
x2 =
4(1 + R2 K P )T1T2
 5TH PROOF 4-9-09

130 Process Control Engineering

(T1 + T2 + R2 K PTd )2
x= ...(36)
4T1T2 (1 + R2 K P )

(2 + K P Td )2
Þ x= [Assuming T1 = T2 = R2 = 1]
4(1 + K P )

For constant value of ‘KP’ say KP = 1

(2 + Td )2
x= ...(37)
8
Td Damping (y)

0 0.707
0.1 0.74
. .
. .
0.5 0.88
. .
. .
1.0 1.16

The table above indicates that an increase in derivate time ‘Td’ increases
the damping ratio ( x), thereby reducing the oscillations in the response.
Hence the addition of a properly tuned derivative action makes the
response of the system faster, in approaching the steady state value of input
Let us see its effect on error
s . R( s)
ess = Lt
s®0 1 + Gc ( s).GP ( s)

1
Assuming R(s) = (Linearly changing input & hence error)
s2

1
s.
s2 1
ess = Lt = =¥
s®0 K P (1 + Td s). R2 0
1+
(T1 s + 1)(T2 s + 1)

The steady state error and/or offset remains when P + D action is used
with processes for obtaining stable operation. However a large proportional
gain may reduce or minimise offset, but practically Large proportional gain
should not be employed because of the reasons already studied in previous
chapters. Therefore it is clear that this system cannot eliminate offset of
proportional controller or any type of steady state error in a process. It is
capable of handling fast process load changes as long as the load change offset
error is acceptable. The transient response for slow processes is also
considerably improved.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 131

3.7.3 Effect of PID Controller Mode

PID scontrol is a very effective action which can be employed for improving
both speed of response and eliminating error between input and output.
This control is effective for any type of process whether it is single or
multiple time constant type and also for different kinds of disturbances
affecting the stability of the system.
The integral action eliminates the offset of proportional action or in other
words the error is completely eliminated, the proportional + Derivative
actions improve the stabilization time i.e., speed of response.
The analytic equation of PID controller is
KP de
P = KP e +
Ti ò e dt + K T P d
dt

ì é 1 ùü
P(s) = í K P ê1 + + Td sú ý E(s).
î ë Ti s ûþ
Consider a single time constant process controlled by PID action

Qi(s) H(s)
1 R
KP(1 + + T d s)
– Tis 1 + Ts

FIG. 3.29

For linearly changing input R(s) = 1/s2

1
s.
s2
ess = Lt
s®0 é æ 1 öùé R ù
1 + ê K P ç1 + + Td s÷ ú ê
ë è Ti s ø û ë 1 + Ts úû

1
= Lt
s®0 é K ùé R ù
s + ê sK P + P + K PTd s2 ú ê ú
ë Ti û ë1 + Ts û

Ti
ess Þ (Since Ti << KP)
RK P
ess » 0.
Thus PID action successfully eliminates the error between input and
output with regard to speed of response Fig. 3.30 indicates the different
characteristics for different values of KP, Ti, Td.
 5TH PROOF 4-9-09

132 Process Control Engineering

KP = 5; Ti = 3 min;
Td = 0 min.
KP = 5; Ti = 3 min;
Td = 0.8 min.

Steady state input

FIG. 3.30

From the above two curves, with addition of derivative gain the systems
stabilization time is considerable reduced thus indicating faster response. The
above curves are based on experimental trials done for processes having single
or multiple time constant characteristics as mathematical analysis of such
systems with PID control is very complicated as far as calculation of responses
for different values of ‘KP’, ‘Td’ and ‘Ti’ are concerned.

Summary
The choice of a particular mode of controller depends on the characteristics of
the process, i.e. whether single capacitance type, single or multiple time
constant type etc.
For processes with more number of storage elements i.e. more time
constants the choice of composite controller modes is best suited. We have
studied in this chapter, the effects of controller modes on different types of
processes. Depending on the mode of controller operation particular process
was chosen for better understanding of the effect of controller mode.
The overall analysis may be summarized as follows.
(1) Proportional controller produces a permanent residual error known as
offset.
(2) Integral controller successfully eliminates offset or error but its overall
speed of response is much slower resulting in large stabilization time.
(3) Proportional + integral controller has no offset but the unstabilizing
influence of integral response is reflected in large stabilization time.
(4) Proportional + Derivative controller cannot eliminate offset or error
how ever due to the presence of derivative action the stabilization time
is considerably reduced thus making the speed of response much faster.
The offset of proportional controller can be made only half of what
would have been without derivative action.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 133

(5) Proportional + Integral + derivative action has smallest stabilization

time thus indicating a very good speed of response and also the offset or
error in eliminated completely because of integral action.
All the above analysis depends on properly tuned values of proportional
gain (KP) Integral or reset time (Ti) and derivative or rate time (Td).
The tuning methods are detailed in the next chapter where different
methods are clearly explained.

SOLVED EXAMPLES
1. For the process shown in Fig. Find
(a) offset
(b) velocity error

1 10 C(s)
R(s) s (s + 1)
+
– +

U(s) 5
(s + 1)

Solution: Using the expressions for offset and velocity error

1 5
. s.
s s +1 5
ess|offset = Lt – Þ =0
s®0 10 1+ ¥
1+
s( s + 1)

1
s.
s2 1
e ss|velocity error = Lt Þ Lt
s®0 10 s®0 s .10
1+ s+
s( s + 1) s ( s + 1)

e 1
ss|velocity error =
= 0.1 units.
10
2. A two vessel process shown in figure has T1 = T2 = 60 secs; R =
1
sec/m2. And another design with T1 = 30 sec s; T2 = 120 sec s; R =
6
1
sec/m2. Which design provides less offset for setting of proportional
6
gain at 10?
 5TH PROOF 4-9-09

134 Process Control Engineering

qi

R1
(C1) q1
h1 u(s) [Disturbance]

Tank 1

R2
(C2)
h2
q0
Tank 2

Solution: Obtaining the process equation for tank (1)

dh1
qi – q1 = C1
dt
dh1 é h1 ù
qi = C1 + q1 ê q1 = R ú
dt ë 1û

dh1
\ R1qi = R1C1 + h1
dt
R1Q1(s) = [T1s + 1]H 1(s) [T 1 = R1C1] ...(1)
For tank (2)

q1 – q1 + u = c2 dh2
dt

q1 = c2 dh2 + q0 – u.
dt
h1 dh
= c2 2 + h2 – R2u.
R1 dt

R2 dh
. h = h1 = R2C2 2 + h2 – R2u.
R1 1 dt

R2
H (s) = [T2s + 1]H2(s) – R2u(s). [T2 = R2C2]
R1 1

R2
H (s) + R2u(s) = [T 2s + 1]H2(s)
R1 1

R2
R1 R2
H2(s) = . H1(s) + u(s)
(T2 s + 1) (T2 s + 1)
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 135

From (1) & (2)

Qi(s) R1 H1(s) R2/R1 H2(s)

T1s + 1 T 2s + 1 +
+

R2
u(s) (T2s + 1)

Process

R2 R2
GP(s) = ; N(s) =
(T1 s + 1)(T2 s + 1) (T2 s + 1)

Case 1: T1 = T 2 = 60 sec s
1
R= sec/m2
6
s . u( s). N ( s)
ess|offset = Lt –
s®0 1 + Gc ( s). GP ( s)

1 1/6
. s.
s (60 s + 1)
ess|offset = Lt – 1/6
s®0
1 + 10.
(60 s + 1)2

Gc(s) = Proportional controller = 10

1
U(s) = Disturbance = (unit step).
s

-1/6 - 0.166
ess|offset = =
10 2.66
1+
6
ess|offset = –0.062 units

Case 2: T1 = 30 sec; T3 = 120 sec s

2
R = 1/6 s/m

1 1/6
. s.
s (120 s + 1)
ess|offset = Lt –
s®0 1/6
1 + 10.
(30 s + 1)(120 s + 1)
 5TH PROOF 4-9-09

136 Process Control Engineering

-1/6
ess|offset = = – 0.062 units.
1 + 10/6

Both designs provide same offset.

3. Calculate
(a) Offset (b) Stabilization time for the process shown in figure.

qi

Set point
LC
h
qo u t

Design specifications
LC = level controller (proportional mode)
KP = 5
A = Area of the tank
= 10 m2
qout = Load variable
= 0.1 step change
qi = Inlet flow rate
= 1 m3/s
Solution: to obtain the Process equation

dh
qi – qout = c [c = capacity of tank].
dt
Applying laplace transform
Q i(s) – Q out(s) = CS. H(S)

1 1
H(s) = . Q i(s) – . Q out ( s)
cs cs
It is a single capacitance process.
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 137

u(s) 1
cs

Qi(s) 1 – H (s )
KP +
cs

Gc(s) = KP = 5 GP(s) = 1/cs N(s) = 1/cs

C = capacity of tand = Area of tank = 10 m2.
s . u( s). N ( s)
ess/offset = Lt –
s®0 1 + Gc ( s) GP ( s)

0.1
u(s) = Q out(s) =
s

0.1 1
s. .
s 10 s
ess /offset = Lt –
s®0 5
1+
10 s

0.1 –0.1
= Lt Þ .
s®0 10 s + 5 5

ess|offset = – 0.02 units.

To find stabilization time considering only one input at a time

Qi(s) 1 H(s)
KP
– cs

H ( s) K P / cs KP KP
= = =
Q i ( s) 1 + K P / cs cs + K P ( s + K P / c)
Applying inverse L.T.
K P – (KP/c)t
h(t) = e .
c
e–[KP /c]t » e– t/T
 5TH PROOF 4-9-09

138 Process Control Engineering

1 K
= P Þ Time Constant T = c/KP
T c
4 c 4 ´ 10
Stabilization time = 4T = =
KP 5
= 8 secs
4. Determine static error and velocity error for the block diagram shown in
figure.

u
N

R G1 G2 +
+ C
–

1 R 1
G1(s) = ; G2(s) = ; N(s) =
Ti s Ts + 1 (Ts + 1)

s . R( s)
Solution: ess|static error Þ Lt
s®0 1 + Gc ( s) GP ( s)
Gc(s) = G1(s) = R/(Ts + 1)
\ R(s) = 1/s (unit step). (Assume)
1
s.
s 1
ess|static error = Lt = =0
s®0 1 R ¥
1+ .
Ti s (Ts + 1)

1
s.
s2
ess|velocity error = Lt
s®0 1 R
1
Ti s 1 + Ts
(R(s) = 1/s2 for velocity error).
1
ess/velocity errors = Lt
s®0 s R
s+ .
Ti s 1 + Ts

Ti
ess/velocity error = .
R
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 139

4 RC
5. For single time constant process prove that stabilization time is ?
1 + RK P
Solution: (Refer text)
6. For the process shown in figure obtain the setting of integral time for
damping ratio 1/3rd?

Design specification
qi
R = 5 s/m2
sp
C = 10 m2
LC

h
q0

Solution: Writing the process equation 3

dh
qi – q0 = c
dt
dh
qi = Rc + q0
dt
q0 = h/R [Resistance (R) at the input line also affects q0 &
also a resistance is provided at the outlet section].
dh h
\ qi = c +
dt R

dh
Rqi = Rc +h
dt
RQ i(s) = [Ts + 1] + H(s)

Gc(s) Gp(s)
Qi(s) 1 R H(s)
– Ti s Ts + 1

H ( s)
Solving for
Q i ( s)
 5TH PROOF 4-9-09

140 Process Control Engineering

R
Ti s(Ts + 1) R
Þ
R TT s2
+ Ti s + R
1+ i
Ti s(Ts + 1)

H ( s) R /TTi
\ =
Q i ( s) 1 R
s2 + s +
T TTi
Comparing the character equation with
s2 + 2xw ns + w2n = 0

R
wn =
TTi

1
2xwn =
T

R 1
2x =
TTi T

4x2 R 1
=
T Ti T2

4x2 R 1
= .
Ti T

Ti = 4x2 R.T.
Given in design specifications
x = 1/3; R = 5; T = RC = 50 sec s.
1
Ti = 4 ´ ´ 5 ´ 50 = 111.11 sec s.
9
Ti = 111.11 sec s. (or)

Ti = 1.85 mins

7. A single capacitance process is controlled by P + I controller. obtain a

relation ship between dynamics of the process and controller
parameters i.e. reset time (T i) and proportional gain KP?
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 141

Solution: The block diagram with controller for single capacitance process is

(Controller) (Process)
Q(s) 1 1 H(s)
KP 1 + +
– Tis cs –

1
u(s) cs

Considering only input at atomic

Q(s) 1 H(s)
KP 1 + 1
– Tis cs

é æ 1 öù 1
ê K P ç1 + T s ÷ ú cs
H ( s) è i øû
= ë
Q ( s) æ 1 ö1
1 + K P ç1 +
è Ti s ø÷ cs

Þ Considering the characteristic equation

KP K
1+ + P2 = 0.
cs Ti cs
Tics2 + KPTis + KP = 0

K P Ti s KP
s2 + + =0
Ti c Ti c

KP K
s2 + .s + P = 0
c Ti c

KP KP
wn = and 2xwn =
Ti c c

KP K
2x = P
Ti ´ c c
 5TH PROOF 4-9-09

142 Process Control Engineering

4x2 K P K P2
= c2
Ti c

4x 2 K
= P
Ti c

Ti K P TK
4x2 = Þ x2 = i P
c 4c

Ti K P
x= or 4cx2 = TiKP
4c

For constant valves of KP and c.

x2 a Ti, which implies the values of Ti must be so chosen or tuned such that
the response is under damped i.e. oscillatory.
8. A two time constant process is shown in figure.
(a) obtain process equation.
(b) with proportional controller, find steady state

qi u

R1
(C1) q1
h1

R2
C2
h2
q0

Solution: (a) For tank (1)

dh1
(qi + u) – q1 = c1
dt
dh1 é h1 ù
qi + u = c 1 + q1 êsince q1 = R ú
dt ë 1û

dh1 h
qi + u = c 1 + 1
dt R1

dh1
R1qi + R1u = R1c1 + h1.
dt
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 143

R1Q i (s) + R1u(s) = (T1s + 1)H1(s) [T1 = R1C1].

For tank (2)
dh2
q1 – q0 = c 2
dt
dh2 é h2 ù
q1 = c 2 + q0 êsince q0 = R ú
dt ë 2û

h1 dh h
= c2 2 + 2
R1 dt R2

R2 dh
h = R2c2 2 + h2
R1 1 dt

R2
H (s) = [T 2 S + 1] [T2 = R2C2] ...(2)
R1 1
Substituting (1) in (2).

R2 ì é R1 ù é R1 ù ü
R1 í ê T s + 1 ú Q i ( s) + ê (T s + 1) ú 4( s)ý = (T2 s + 1)H2(s)
ïî ë 1 û ë 1 û ïþ

é R2 ù é R2 ù
\ H2(s) = ê
(T s + 1)(T s + 1) ú Q i ( s) + ê (T s + 1)(T s + 1) ú u(s)
ë 1 2 û ë 1 2 û

Qi(s) R2
(T1s + 1)(T2s + 1)
+
+ H2(s)
V(s) R2
(T1s + 1)(T2s + 1)

(b) Analysis with proportional controller.

u(s) R2
(T1s + 1)(T2s + 1)

Qi(s) R2 +
+ H2(s)
– (T1s + 1)(T2s + 1)
 5TH PROOF 4-9-09

144 Process Control Engineering

s . R( s)
® ess|static error = Lt
s®0 1 + Gc ( s)GP ( s)

1
let R(s) = (unit step)
s
R2
Gc(s) = KP; GP(s) =
(T1 s + 1)(T2 s + 1)

1
s.
s
ess|static error = Lt
s®0 K P . R2
1+
(T1 s + 1)(T2 s + 1)

1
ess|static error =
1 + K P . R2

s . u( s) N ( s)
® ess|offset = Lt –
s®0 1 + Gc ( s). GP ( s)

1
let u(s) = (unit step)
s
R2
N(s) = .
(T1 s + 1)(T2 s + 1)

1 R2
. s.
s (T1s + 1)(T2 s + 1)
ess|offset = Lt –
s®0 K P R2
1+
(T1 s + 1)(T2 s + 1)

- R2
ess|offset =
1 + K P R2

9. For the process shown in 6 prob (8) show that addition of integral
controller eliminate offset.

æ 1 ö
Solution: Gc(s) = KP ç1 + ÷ Þ P + I controller
è Tisø

u(s) = 1/s (unit step)

R2
N(s) =
(T1 s + 1)(T2 s + 1)
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 145

R2
GP(s) =
(T1 s + 1)(T2 s + 1)

1 R2
- s. .
s (T1 s + 1)(T2 s + 1)
\ ess|offset =
æ 1 ö R2
1 + K P ç1 + ÷
è Ti s ø (T1s + 1)(T2 s + 1)

R2
ess|offset = =0
1 + K P (1 + ¥). R2
10. For a closed loop automatic control from the operational diagram
R1 R1
G1 = Kc; G2 = and N= .
T1 s(T2 s + 1) T1 s(T2 s + 1)
(a) Calculate proportional sensitivity for critical damping?
(b) Calculate static error and offset.
Solution: Given
G1 = Gc(s) = Kc = proportional gain
R1
G2 = GP(s) =
T1 s(T2 s + 1)

R1
N = N(s) = .
T1 s(T2 s + 1)
The block diagram may be formulated as.

u(s) R1
T1s(T2s + 1)

Qi(s) R1 +
KC + H(s)
– (T1s(T2s + 1)

let u(s) = 0.

R1K P
H ( s) T1s(T2 s + 1)
=
Q i ( s) R1K P
1+
T1 s(T2 s + 1)
 5TH PROOF 4-9-09

146 Process Control Engineering

The characteristic equation is

R1K c
1+ =0
T1 s(T2 s + 1)

T1T2 s2 + T1s + R1Kc = 0

1 RK
s2 + .s + 1 c = 0
T2 T1T2

Comparing with s2 + 2xwns + wn2 = 0

R1K c
wn = (r/s)
T1T2

R1K c 1
2x ´ =
T1T2 T2

4x2 R1 K c 1
= 2
T1 T2 T2

T1
4x2R1Kc =
T2

For x = 1 [critical damping]

T1
4R1Kc =
T2

T1
Kc =
4 R1T2

s . R( s)
(b) ess|static error = Lt
s®0 1 + Gc ( s) GP ( s)

let R(s) = 1/s

1
s.
s 1
it Lt = =0
s®0 K c R1 1 + ¥
1+
T1 s(1 + T2 s)
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 147

- s . u( s). N ( s)
® ess|offset = Lt
s ®0 1 + Gc ( s) GP ( s)

1 R1
- s. .
s T1 s(1 + T2 s)
= Lt
s ®0 K c R1
1+
T1 s(1 + T2 s)

- R1
=
T1 s(1 + T2 s) + K c R1

ess|offset = –R1/KcR1 = –1/Kc.

-1
ess|offset =
Kc

11. In a closed loop automatic control system with

R1
Gc(s) = KP(Td s + 1); GP(s) = and
T1 s(T2 s + 1)

R1
N(s) =
T1 s(T2 s + 1)
(a) Calculate KP for damping ratio 1/3rd?
(b) Calculate the period of oscillations?
Solution: Constructing the characteristic equation
1 + Gc(s) GP(s) = 0

K P (Td s + 1). R1
1+ =0
T1 s(T2 s + 1)

T1T 2s2 + T1s + KPR1Td s + KPr1 = 0

T1T 2s2 + s[T 1 + KPR1Td] + KPR1 = 0

s2 +
[T1 + K P R1Td ] + K P R1
=0
T1T2 T1T2

Comparing with s2 + 2xwns + w n2 = 0

K P R1
wn = r/s.
T1T2
 5TH PROOF 4-9-09

148 Process Control Engineering

(b) Since 2pfn = wn.

K P R1
\ 2pfn =
T1T2

K P R1
4p2 f 2n =
T1T2

K P R1
f 2n =
4p2T1T2
Period of oscillation (undamped oscillations)
1 1 4p2T1T2
T= ; T2 = =
fn f n2 K P R1

4 p2T1T2
T= sec s
K P R1

K P R1 T + K P R1Td
(a) 2x = 1
T1T2 T1T2

4x2 . K P R1 (T + K P R1Td )2
= 1
T1T2 (T1T2 ) 2

1
For x=
3

4 (T + K P Td )2
KPR1 = 1
9 T1T2

12. Calculate offset for prob (11)?

é - s . u( s). N ( s) ù
Solution: ess|offset = Lt ê
s ® 0 1 + Gc ( s) GP ( s) ú
ë û

é 1 R1 ù
ê - s . s . T s(T s + 1) ú
1 2
Lt ê ú
s®0 ê K P (Td s + 1). R1 ú
1+
êë T1 s(T2 s + 1) úû

é ( - R1) ù
Lt ê ú
s®0 T s
ë 1 2(T s + 1) + K (
P dT s + 1). R1û

- R1 -1
ess|offset Þ =
K P R1 K P
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 149

13. Prove that for a single capacitance process, integral action is not required
to eliminate static error, when only P-controller is used?
Solution: let Gc(s) = KP = proportional controller
1
GP(s) = [single capacitance process]
cs
s . R( s)
ess|static error = Lt
s®0 1 + Gc ( s) GP ( s)

1
let R(s) =
s

1
s.
s 1
Lt = =0
s ®0 KP 1+ ¥
1+
cs
ess|static error = 0

14. In a closed loop automatic control

If Gc(s) = KP(1 + Td s)

1
GP(s) =
cs
Calculate the system time constant?
Solution: The block diagram representation is

Qi(s) 1 H(s)
KP(1 + Td s)
– cs

H ( s) K P (1 + Td s)/ cs K P (1 + Td s)
= Þ
Q i ( s) K P (1 + Td s) cs + K P (1 + Td s)
1+
cs

K P (1 + Td s) K P (1 + Td s)
Þ
cs + K P + K P Td s cs + K P Td s + K P

K P (1 + Td s)
H ( s) ( c + K P Td )
Þ
Q i ( s) æ KP ö
çè s + c + K T ÷ø
P d
 5TH PROOF 4-9-09

150 Process Control Engineering

Applying Inv. L.T.

K P (1 + Td s)
c + K P Td A
=
æ KP ö æ KP ö
çè s + c + K T ÷ø ç s + c + K T ø÷
è
P d P d

Put s = 0

KP
c + K P Td A
=
KP KP
c + K P Td c + K P Td

KP
\ A=
c + K P Td

æ KP ö
H ( s) çè c + K T ÷ø
P d
\ =
Q i ( s) ( s + K P / c + K P Td )

KP F I
h(t) =
c + K P Td
.e GHKP
c + K PTd
t JK
Comparing with e– t/T
1 KP
=
T c + K P Td

c + K P Td
T=
KP

c
T= + Td
KP

15. Calculate the velocity error for single time constant process with P + I
controller?
Design specifications are
R = 5 s/m2; c = 5 m2; KP = 10
Ti = 2 mins.
s . R( s)
Solution: Ess|velocity error = Lt
s®0 1 + Gc ( s) GP ( s)
 5TH PROOF 4-9-09

Effect of Control Actions on Processes 151

æ 1 ö é 1 ù
Gc(s) = KP ç1 + = 10 ê1 +
è ÷
Ti s ø ë 120s úû

R 5
GP(s) = = [T = Rc]
1 + Ts 1 + 25s

1
s.
ess|velocity error = Lt s2
s®0 æ 1 öæ 5 ö
1 + 10 ç1 + ÷
è 120 s ø çè 1 + 25s ÷ø

1
= Lt
s®0 æ 10 ö 5
s + ç10 s + ÷
è 120 ø 1 + 25s

1
=
é 10 ù
0 + ê0 + [5]
ë 120 úû

12 0
ess|velocity error = = 2.4 units
50

QUESTION SET
1. How are errors classified in process control? Obtain the expressions of
various error representations?
2. Discuss the effect of proportional controller on
(a) Single capacitance process
(b) Time constant process.
3. Why derivative action is not suitable in single mode configuration?
Explain?
4. Discuss the effect of PI action on time constant process?
5. Compare the effects of PI, PD and PID actions on time constant process?
 "
+0)26-4

Tuning of Controllers

4.1 INTRODUCTION
The optimum values of controller parameters Kp, Ti and Td i.e. proportional
gain, integral time and derivative time are intimately related to the dynamics
and characteristics of the process, where a particular dynamic variable is
measured monitored and controlled.
Tuning may be defined as “the optimal adjustment of controller
parameters to achieve satisfactory control”.
The procedure of tuning may depend on elaborate optimization
calculation based on process dynamics and model to trial and error attempt in
achieving the required control objective. Some methods involve semi-
empirical equations, in that they depend on measurements made on the
control system with all its complexities to determine factors used in adjustment
formulas.
The frequency response analysis approach is more analytical and is based
on model or transfer function of the process and the control loop. From the
time response characteristics of second order system subjected to step change
in input or load or set point, the typical criteria for good control is that, there
should be minimum overshoot, rise time and settling time and one-quarter
decay ratio.
One criteria that is often used to evaluate a response of a control system is
the integral of the square of the error with respect to time (ISE)

z
¥
ISE = e2 dt
0

The other two criteria often used in process control are:

z
¥
Integral of the Absolute value of Error (IAE) = |e|dt
0
 5TH PROOF 4-9-09

Tuning of Controllers 153

z
¥
Integral of time Weighted Absolute Error (ITAE) = |e|t × dt
0

However, these criterial may be evolved from specific transfer function

analysis, and are more analytic in nature.
A particular controller mode to be chosen, and the respective tuning of the
related controller parameters using time integral performance criteria require
the shape of the complete closed loop response from time t = 0 till the
response reaches steady state.
ISE is better than IAE, when the controller is required to suppress and
minimise large errors, because the errors are squared and thus contribute
more to the integral value.
The choice of IAE depends on magnitude of errors which are usually
small. To minimise errors that persist for long times, ITAE criteria will tune the
controllers to perform the required operation.
Fig. 4.1 shows the expected responses, when controller parameters are
tuned by ISE or IAE or ITAE performance criteria.

o/p

I TAE
IAE

IsE
t

FIG. 4.1
The various tuning sequences may thus be summarised as follows.

Tuning Methods

Open Loop Configuration Close Loop Configuration

Method Method
1. Ziegler - Nicol Method
The values of critical
Gain and time Period can be
For Processes with Self For Processes without self obtained by
Regulation. Regulation. Time Domain Approach
(a) Process Reaction Curve Tuning not Possible (Routh Array)
Method
Freqency Domain Approach
(Bode plots )
(b) Cohen-Coon
Method 2. Time Integral Performance
Criteria
 5TH PROOF 4-9-09

154 Process Control Engineering

4.2 PROCESS REACTION CURVE METHOD

4.2.1 This method is often referred to as open loop transient response
method, in which the process control loop is ‘opened’ so that no control action
occurs and an open loop transient (disturbance) is introduced by the step
change in the signal to the control valve. The step response is then recorded at
the output of the measuring element.
The response of the system including control valve, process and
measuring element is called the “process reaction curve”.
A typical process reaction curve is “S”–shaped curve as shown in Fig. 4.5
which is also called as sigmoidal shape. However, such response characteristics
are exhibited by first order systems with certain amount of dead time, and
should necessarily be systems with self-regulation.
Tuning of the controller parameters will not be possible of there is no self-
regulation in the process or system.
Consider the liquid level control system shown in Figure 4.2.
The block diagram representation of the above figure is shown in Fig. 4.3.

Inlet Set point

Controller LT = Level Transducer

(Measuring Element)

LT

Outlet

FIG. 4.2

Set point o/p level

Controller Controller Process 'h'
Gc (s) valve Gv(s) Gp (s) c(s)
R (s)

Level Transducer
Hm (s)

FIG. 4.3

Gc(s) = Transfer function of controller

Gv(s) = Transfer function of final control element or control valve
Gp(s) = Transfer function of process
Hm(s) = Transfer function of measuring element (or) level transducer
 5TH PROOF 4-9-09

Tuning of Controllers 155

For the measurement of process reaction curve, the process control loop is
‘opened’ as shown in Fig. 4.4, where the output of the controller is
disconnected from the control valve.

Step change (A/s)

Set point Output
– Gc (s) Gv (s) Gp (s) c(s)
R (s)

Hm (s)

To Recorder
for Measurement
of Process Reaction Curve

FIG. 4.4

A step change of magnitude ‘A’ is introduced at the control valve and the
values of output as time t ® ¥ are recorded at the outup of the measuring
element as shown in Fig. 4.4.
As maintained earlier, the transfer function between control valve, process
and measuring element is approximated to be first order system with dead
time.

K e - td s
Gv(s) Gp(s) Hm(s) = ...(1)
1 + τs

where K = System gain = B/A from Fig. 4.5

td = dead time/Lag (min/secs)
t = Time constant (or) Process reaction time.
The typical sigmodal curve obtained would be as shown in Fig. 4.5
A tangent line is drawn at the inflection point of the curve. The inflection
point is defined as that point on the process reaction curve where the slope
stops increasing and begins to decrease. Wherever the tangent line crosses the
time axis we get,
td = Lag time/Dead time (sec/min)
Also from the curve, we get the process reaction time (or) time constant (t).
Let B = Ultimate value of the output as t ® ¥
B
Slope [s] of the tangent line is given as S=
τ
Defining ‘R’ which is Log ratio and a unit less quantity as
S td
R=
B
 5TH PROOF 4-9-09

156 Process Control Engineering

o/p
B

Tangential
Line
Inflection
Point

t
td t

I/p

FIG. 4.5

where S = slope of tangent line at inflection point (or) It is also called

reaction rate
td = Dead time or Lag
B = Ultimate value of o/p.
4.2.2 Using semi-empirical rules which have been proven in practice, the
controller parameter settings can be obtained as follows,
(1) Proportional only
A
Kp = ...(2)
S × td
where, A = step change at final control elements.
(2) Proportional + Integral mode [P + I mode]
0.9 A
Kp = ; Ti = 3.33 td ...(3)
S × td
(3) Proportional + Integral + Derivative mode [P + I + D]
1.2 A
Kp = ; Ti = 2 t d ; Td = 0.5 td ...(4)
S × td

4.3 QUARTER AMPLITUDE CRITERIA

[COHEN-COON CORRECTIONS]
Cohen-Coon employed the approximate first order with dead time model
given by eqn. 1 and made corrections to the controller parameter settings
 5TH PROOF 4-9-09

Tuning of Controllers 157

obtained from the process reaction curve i.e. eqns. 2 to 4. They derived
expressions for the best controller settings using load changes and various
1
performance criteria such as decay ratio (Quarter amplitude criteria)
4
minimum offset and minimum ISE. Thus, when quarter amplitude criteria,
which is the most favourable performance specification by practising process
control technologists is desired, these c-c-corrections are an absolute necessity.
1
decay ratio implies the ratio between second peak overshoot and first
4
1
peak overshoot from the steady state value of the input should be , in the
4
step response of standard second order system.

y 1
x y x = 4

FIG. 4.6

y 1
A decay ratio = is a reasonable trade off between fast rise time and
x 4
reasonable settling time.

4.3.1 Cohen-coon Corrections for Controller

Parameters Obtained from Process Reaction
Curve
(1) Proportional only

Kp =
A LM1 + S × t OP
d
...(5)
S × td N 3× B Q
(2) Proportional + Integral mode [P + I]

A LM0.9 + R OP ; LM 30 + 3 R OP t
Kp =
S × td N 12 Q Ti =
N9 + 20 R Q d ...(6)
 5TH PROOF 4-9-09

158 Process Control Engineering

(3) Proportional + Integral + Derivative [P + I + D]

A LM R OP
Kp =
S × tdN1.33 +
4 Q
Ti =
LM32 + 6 R OP t
N13 + 8 R Q d

Td =t M
L 4 OP ...(7)
d
N11 + 2 R Q
S × td
where R = log ratio (unit less) = .
B

General Assumptions for C-C-Correction

(1) The corrections given by equations 4.5-4.7 are for first order plus dead
time process whose open loop response characteristics are sigmoidal in
nature.
(2) Most of the single or multiple time constant processes encountered in
process industries exhibit sigmoidal response characteristics when
controller is disconnected from the final control element.
The oscillatory under damped behaviour is produced mainly by the
presence of feedback controllers.
(3) C-C-correction are made to ensure that one quarter delay ratio
performance criteria is satisfied.

4.4 ZIEGLER-NICOL TUNING METHOD

This method of tuning the controller parameters was proposed and developed
by two process control technologists, Ziegler and Nicols, of Taylor Instruments
company, USA.

4.4.1 Preliminary Aspects

It is basically a closed loop tuning method, applicable to any type of system
with or without self-regulation. This technique is also called as “ultimate cycle
method,” because the tuning procedure is based on adjusting the closed loop
until steady oscillations in response occur.

Tuning procedure
(1) The integral and derivative actions are reduced to the minimum effect
after the process reaches steady state at the normal level of operation.
 5TH PROOF 4-9-09

Tuning of Controllers 159

(2) The value of proportional gain K p is gradually increased while

providing periodic small disturbances to the system. If the response
decays, a higher value of ‘Kp’ is selected. This process is repeated till the
response exhibits sustained oscillations about the steady state value of
input.
(3) Note the value of proportional gain at which the response exhibits
continuous and sustained oscillations. This is called as ultimate (or)
critical gain “Kc”. The frequency, and hence the period of oscillations is
ultimate (or) critical period “Tc”.
(4) The optimal values of Kp, Td and Ti can be obtained from critical gain
and period using certain empirical relations, to give better overall
control.
(5) The advantage of Ziegle-Nicol method is that the values of Kc and Tc can
be found out by actually working on the real time process or when an
accurate transfer function of process, control element (valve) and
feedback (measuring element) are known, they can be obtained either
by Time domain method i.e. Routh array analysis or from the frequency
response methods (Bode plots).
(6) The periodic short disturbances to the process control system can be a
step or impulse type change in set point.
(7) The most important precaution one has to take when finding the values
of Kc and Tc is that of limit cycles, which occur due to control value
saturation.
The proportional gain should be gradually increased i.e. in small steps. If
its value is changed suddenly, there is a chance that the valve moves to its
Limits i.e. the valve saturates.
Valve saturation implies that it reaches its limits of 3-15 psi repeatedly.
When this condition occurs, the closed loop response also exhibits sustained
oscillations called as limit cycles which is a non-linear phenomena.
The gain and period of limit cycles are not the actual critical gain ‘Kc’ or
critical period ‘Tc’ and should not be used to determine controller settings.
A process control analyst should therefore have adequate laboratory
experience in identifying these limit cycles that occur in the system.

4.4.2 Z-N Settings for Controller Modes

From the critical gain and period, the settings for Kp, T i and Td are assigned as
follows for different controller modes.
(1) Proportional only
Kp = 0.5 Kc ...(8)
(2) Proportional + Integral only [P + I mode]
Tc
Kp = 0.45 Kc; T i = ...(9)
12
.
 5TH PROOF 4-9-09

160 Process Control Engineering

(3) Proportional + Integral + Derivative (P + I + D)

Tc T
Kp = 0.6 Kc; Ti = ; Td = c ...(10)
2 8

4.5 FUNDAMENTAL FREQUENCY RESPONSE

METHOD OF CONTROLLER TUNING
When any system is subjected to sinusoidal input, the output is also sinusoidal
giving different magnitude and phase angle but same input frequency ‘w’.
Frequency response analysis implies varying ‘w’ from 0 to ¥ and observing
corresponding variations in magnitude and phase angle of the response. The
frequency response analysis can be performed on the approximate or accurate
transfer function of the system by putting s = jw.
Let F(s) = Transfer function
Put s = jw
F(jw) = sinusoidal transfer function
» sinusoidal response.
F(jw) = |F(jw)| |F ( jw ) [Polar form].

|F(jw)| = (Real Part)2 + (Imaginary Part)2

F Imaginary Part I
|F ( jw ) = Tan–1 GH Real Part JK
Bode plots are asymptotic plots that give magnitude and phase
characteristics versus frequency (w) calibrated on log scale. To obtain
asymptotic characteristics of magnitude, its decibel values (db) are considered
where db value is defined as 20 log [F(jw)].
The stability criteria from bode plots is derived from the characteristic
equation 1 + G(s)H(s) = 0 expressed as
G(s)H(s) = – 1 (or) G( jw) H( jw) = – 1+ jo.
We define the following terms to determine the actual stability of the
system.
(1) Gain cross over frequency [MC?]
It is defined as the frequency at which the magnitude of open loop transfer
function G(jw) H( jw) is unity or 0 db.

|G( jw) H(jw)|w = wgc = ( -1)2 + 0 2 = 1

20 log |G(jw)H(jw)| = 20 log 1 = 0 db

 5TH PROOF 4-9-09

Tuning of Controllers 161

(2) Phase cross over frequency [MF?]

It is defined as the frequency at which the phase angle of the open loop
transfer function G(jw) H(jw) is ± 180°
|G( jw ) H( jw )| = ± 180°
w = w pc

(3) Gain Margin [G.M.]

It is the “Allowable gain” i.e. the amount of additional gain that can be
introduced into a system till the system till the system reaches the verge of
instability.
It is obtained at phase cross over frequency (wpc).
Let |G( jw ) H( jw )|w = w pc = X

1
G.M. =
X
1
G.M. [db] = 20 log
X
(4) Phase Margin [P ×M.]
It is the “allowable phase lag” i.e. the amount of additional phase lag that
cab be introduced into a system till the system reaches the verge of instability.
It is obtained at gain cross over frequency (wgc).
Let |G( jw ) H( jw )|w = w gc = f
P×M. = 180° + f.

Determination of G.M. and P.M. from Bode Plots

Consider the following three bode magnitude and phase angle plots shown in
figures 4.7, 4.8 and 4.9.

0db
G.m. = o – (– x)
– Xdb = xdb
f
P.M. P.M. = f – (–180)
– 180°
= 180° + f = + VE

wgc wpc Log w

Stable system
G.M. = P.M. = +VE

FIG. 4.7
 5TH PROOF 4-9-09

162 Process Control Engineering

db 0 – (x) = – x db
0 db

– 180°
180° + f = – ve
f

log w
wpc wgc
Unstable system
G.M. = P.M. = – ve

FIG. 4.8

0 db

– 180°

log w
wgc = wpc
G.M. = P.M. = 0
Marginally stable system

FIG. 4.9

To find gain margin, extend the phase cross over frequency line till it
meets the magnitude characteristics. The distance between 0 db line to the
value in db where wpc line intersects the magnitude plot is the G.M.
Similarly to find phase margin the distance between the point where the
gain cross over frequency intersects the resultant phase angle plot to –180°
reference line gives P.M.

Tuning of Controllers from Bode Plots

The following procedure is adopted for finding the critical or ultimate gain
“Kc” and critical period “Tc” from bode magnitude and phase plots.
(1) Plot Bode magnitude and Phase Plot by taking the transfer function of
plant, control element and measuring element i.e. for Gv(s) Gp(s) Hm(s)
[Refer eqn. 1]. The controller should be excluded, as we have to
determine the optimum settings for it.
 5TH PROOF 4-9-09

Tuning of Controllers 163

(2) From the Bode diagram find gain margin and phase cross over
frequency.
Let G.M. = y db
FG 1 IJ = y db
i.e. 20 log
H xK
1 y
log =
x 20
1
= log–1 (y/20)
x
Thus is the value of critical (or) ultimate gain “Kc”
1
Therefore Kc = = log–1 (y/20).
x
The critical period is,
2p
Tc = [sec/min]
w pc
where wpc = phase cross over frequency
w pc = 2p fc,
1 1
Tc = Þ fc = ,
fc Tc

2p 2p
\ w pc = Þ Tc =
Tc w pc
(3) Use Zeigler Nicol Settings to find Kp, Ti and Td values.

4.6 CONTROLLER TUNING FROM ROUTH ARRAY

METHOD
Routh array is an algorithm, used in time domain to estimate the stability of
the system. Tuning of controllers can be done using this algorithm, where the
values of critical or ultimate again and time period are evaluated from the
array.
The routh array is constructed for the characteristic equation 1 + G(s)
H(s) = 0. Consider a fourth order characteristic equation
P(s) = a0s4 + a1s3 + a2s2 + a3s + a4 = 0
The coefficients of the polynomial are arranged as shown.
 5TH PROOF 4-9-09

164 Process Control Engineering

Rooth Array =
S4 a0 a2 a4
3
S a1 a3 0
2
S b1 b2 0
1
S c1 0 0
0
S d1 0 0

The elements of other rows and column’s b1, b2 etc. are obtained as follows
a1a2 - a0 a3 a1a 4 - a0 ´ 0
b1 = b2 =
a1 a1
b1a3 - a1b2
c1 = c2 = 0
b1
c1b2 - b1 ´ 0
d1 =
c1
After obtaining all the value of other elements, the condition for stability is
there should not be any sign change in first column of Routh array i.e. all the
first column elements should be greater then zero.

Procedure
(1) Construct the characteristic equation
1 + Gc(s) Gv(s) Gp(s) Hm(s) = 0.
With Gc(s) = controller transfer function with only K p [proportional
gain] with derivative and integral actions reduced to minimum effect.
(2) Construct Routh array and evaluate all the elements of the array in
terms of Kp.
(3) Apply, the condition of stability, where all the first column elements
should be greater than zero.
(4) The value of proportional gain Kp which gives marginally stable case,
thus, corresponds to critical or ultimate gain “Kc”.
(5) For this value of proportional gain Kp construct the auxiliary equation
A(s).
(6) Compare the roots of the auxiliary equation A(s) with S = jw and find
‘w’.
2π
Therefore w = wc [critical frequency], wc =
Tc

2π
Þ Tc = (sec/min)
wc
(7) Apply Ziegler-Nicol tuning equations for finding the optimum values of
Kp, T i and Td.
Table 4.1

Type of control action Process reaction Cohen-Coon method Ziegler-Nicols method

curve method (corrections)

A A s.t
Proportional only Kp = Kp = 1+ d Kp = 0.5Kc
LM OP
S .t d S.t d 3B
N Q
A = Step change in input Kc = Critical/utlimate gain
B = Ultimate value of output Tc = Critical/ultimate period
td = Dead time
R = Log ratio

0.9A A R
Proportional + Integral [P + I] only Kp = Kp = 0.9 + Kp = 0.45Kc
LM OP
S ×t d S ×t d N 12 Q
5TH PROOF 4-9-09

Tc
Ti = 3.33 td Ti = d Ti =
LM 30 + 3R OP t
N 9 + 20R Q 1.2

1.2A A L RO
Proportional + Integral + Derivative only [P + I + D] Kp = K =
p 1.33 + P Kp = 0.6Kc
S ×t d d
S ×t MN 4Q

Tc
Ti = 2td i
T =M d Ti =
L 32 + 6R OP t
N 13 + 8R Q 2

Tc
Td = 0.5td Td = M d Td =
L 4 OP t
N11 + 2R Q 8
Tuning of Controllers
165
 5TH PROOF 4-9-09

166 Process Control Engineering

SOLVED EXAMPLES
1. A Process exhibits oscillations for a critical gain setting of 2 in a 30
minute period. Find the settings of Ziegler-Nicol for.
(a) P + I Controller mode
(b) P + I + D Controller mode
Solution:
(a) For P + I mode
Kp = 0.45 Kc. Given Kc = 2
Kp = 0.45 ´ 2 = 0.9 Tc = 30 min

Tc 30
Ti = = = 25 mins.
1.2 1.2
(b) For P + I + D mode
Kp = 0.6 Kc = 0.6 ´ 2 = 1.2

Tc 30
Ti = = = 15 min.
2 2

Tc 30
Td = = = 3.75 min.
8 8
2. In the application of Ziegler-Nicol method oscillations are observed in
the process with proportional band set to 400/01 in the time period of
10 minutes, Find the settings of three mode controller?
Solution: A three mode controller is P I D controller.

100
Since, P.B =
Kp

\ The critical Gain ‘Kc’ is

100 100
Kc = = = 2.5
P. B 40
The critical period ‘Tc’ is given as T c = 10 minutes
\ Kp = 0.6 Kc = 0.6 ´ 2.5
= 1.5

Tc 10
Ti = = = 5 min.
2 2

Tc 10
Td = = = 1.25 min.
8 8
 5TH PROOF 4-9-09

Tuning of Controllers 167

3. For the process shown in figure, determine the optimum settings for
PID controller. (Use Routh Array)

Kp 1 c(s)
s(s + 2) (s + 4)
R(s)

Solution: Let us construct the characteristic equation

1 + Gc(s) Gv(s) Gp(s) H m(s) = 0

Kp
1+ =0
s( s + 2) ( s + 4)

S(S + 2) (S + 4) + Kp = 0
S(S2 + 6S + 8) + Kp = 0
S3 + 6S2 + 8S + Kp = 0

S3 1 8
S2 6 Kp
48 - K p
S1
6
S0 Kp

48 - K p
= 0 Þ Kp = 48
6
\ The critical Gain Kc = 48
Auxillary equation
A(S) = 6 S2 + Kp = 0
6S2 + 48 = 0
S2 = – 8
S = ± j 2.8
Comparing with S = jw
S = ± j2.8 jw
w = 2.8 r/s
The critical period ‘Tc ’ will be

2p 2p
Tc = = = 2.24 secs.
w 2.8
 5TH PROOF 4-9-09

168 Process Control Engineering

The three mode settings are

Kp = 0.6 ´ Kc = 0.6 ´ 48 = 28.8

Tc 2.24
Ti = = = 1.12 secs
2 2

Tc 2.24
Td = = = 0.285 secs.
8 8
A first order process has transfer function
e -2 s
G(S) =
1 + 4s
A Test run is conducted by subjecting to a sudden disturbance of
magnitude.
4. The ultimate value of output stabilized at 3.5. Find the settings of P + I
controller?
Solution: Comparing the transfer function

e -2 s e - td s
=
1 + 4s 1 + Ts

td = 2 secs.
T = 4 secs.
Given A =5
B = 3.5

B 3.5
Slope [S] = = = 0.875
T 4

S td 0.875 ´ 2
Log ratio [R] = = = 0.5
B 3.5
The settings of P + I mode are

0.9 A
Kp = ; Ti = 3.33 td
S td

0.9 ´ 5
Kp = = 10.3
0.875 ´ 2

Ti = 3.33 ´ 2 = 6.66 secs.

 5TH PROOF 4-9-09

Tuning of Controllers 169

6. What is the critical gain and critical period of the system whose
characteristic equation is
S3 + 4S2 + 6S + (4 + K) = 0
Solution:

S3 1 6
S2 4 (4 + K)
24 – (4 + K )
S1 0
4
S0 (4 + K) 0

24 – (4 + K )
=0 Þ K = Kc = 20
4
® The Auxillary equation is
A(S) = 4S2 + (4 + K) = 0
4S2 + (4 + 20) = 0
S2 = – 6

S = ± j 6 jw

w = wc = 6 r/s.

2p 2p
Tc = Þ = 2.61 secs
wc 6
 5
+0)26-4

Implementation of
Control Actions

5.1 INTRODUCTION TO OPERATIONAL

AMPLIFIERS
The operational Amplifier commonly abbreviated as op-Amp was originally
developed and designed for computing mathematical functions such as
addition, subtraction multiplication, etc.
With the addition of suitable external feedback or otherwise components,
its use has been extended for variety of applications such as AC/DC
amplification, filtering, in oscillators, regulators, etc. and also the controllers.
Because of ever-increasing reliability and miniaturization of complex
electronic circuitry of integrated circuits, electronic controllers have replaced
their counterparts i.e., pneumatic and hydraulic controllers gradually and
eventually. In this chapter we will initially see the electronic implementation of
control actions and then proceed to hydraulic and pneumatic controllers.
The schematic symbol is shown in Fig. 5.1. For simplicity the power supply
and other pin connections are omitted.

Inverting V1
input –
A
Non-inverting V + V0
2
input

FIG. 5.1
 5TH PROOF 4-9-09

Implementation of Control Actions 171

The differential inputs are designated by the signs (+) and (–) notations.
These polarity signs refer to the phase relationship between individual input
signals and the output signal.
The output signal is in phase with the signal applied to the input pin
marked “+” and hence is known as “non–inverting input” and anti phase to
the signal applied to the pin marked “–” and is known as “inverting input”.
When the power supply connections are also shown, they are represented
as V + and V – for positive and negative voltages, respectively.
The characteristics of ideal op-Amp are
• Infinite input impedence.
An ideal amplifier can be connected to any signal source with no
loading effects.
• Infinite gain “A”.
• Zero output impedence
An ideal op-Amp is capable of driving any load.
• Infinite bandwidth:
An ideal op .Amp can be made to amplify any signal frequency from 0 to
¥ Hz. without attenuation.
• Infinite common mode rejection ratio so that the output common mode
noise voltage is zero.
• Infinite slew rate so that output voltage changes occur simultaneously
with input voltage changes.
However, practical op. Amps can be made to exhibit some of these
characteristics using a negative feedback arrangement.
The input impedence, output impedence and bandwidth can be brought
close to ideal values by incorporating negative feedback arrangement.
When operated in negative feedback configuration the output of an op.
Amp does not switch between positive and negative saturation levels.
In the implementation of control actions initially the ideal characteristics
are assumed and Laplace transformation of the inputs and impedences is
considered for the case of understanding and analysis.

General Analysis Procedure

Consider the circuit shown in Fig. 5.2.
The non-inverting terminal is grounded and the input is applied to
inverting terminal.

Vin ( s) - V1( s) V ( s) - V0 ( s)
= 1
Z1( s) Z2 ( s)

Since V1(s) » 0 [Virtual Ground].

 5TH PROOF 4-9-09

172 Process Control Engineering

Z2(s)

I(s)

Vin(s) Z1(s) –
V1(s)
I(s) A
+
V0(s)

FIG. 5.2

Vin ( s) - V0 ( s)
=
Z1 ( s) Z2 ( s)

\ V0(s) = –
LM Z ( s) OP .V (s)
2
...(1)
N Z ( s) Q
1
in

Z2 ( s)
The amplifier gain A = .
Z1( s)
By properly choosing the values of z1(s) and z2(s), the op. Amp can be used
for various applications including the controllers variety.

5.2 ELECTRONIC IMPLEMENTATION OF

CONTROLLER MODES
The electronic controllers are designed using op. amp circuits. Initially the
discussion is aimed at discontinuous modes i.e., two position controller,
multiposition controller and floating controller modes respectively furthered
by continuous modes of control actions. Because of the widespread use of
operational amplifiers the emphasis is more on using them for electronic
implementation of controllers.

5.2.1 Implementation of Discontinuous Controllers

As we have already studied in the earlier chapters there are three basic
discontinuous controller modes, namely
• Single mode ON/OFF controller
• Multiposition controller
• Floating controller
 5TH PROOF 4-9-09

Implementation of Control Actions 173

We will now see the electronic implementation of these control actions.

On/off Controller
An on/off controller or two position controller can be implemented in two
designs, namely,
• Electro mechanical design.
• Electronic system design.
They find extensive use in many household applications like air
conditioning systems and heating systems in particular.
The electro mechanical design involves the use of bimetallic strip and
mercury switch.
The inertia of the mercury tends to keep the system either in ON or OFF
position until the temperature changes from the set point value.
This provides the required neutral zone the prevent excessive cycling of
the system.
The electronic implementation of ON/OFF controller with adjustable
neutral zone using op. Amps is shown in Figure 5.3.

R2

Vi R1 R1
R

– R Vo
– Vout =
– 0
+
+
+
+ Vs
Comparator
Summer Inverter

Vref

FIG. 5.3

It consists of three stages, summing amplifier, inverter and comparator.

The operation of the circuit shown in Fig. 5.3 is, If the controller input voltage
“Vi” reaches a value “VH”, the output of the comparator should go to High (or)
ON state thereby indicating some magnitude of the output voltage ‘VO’.
Similarly, if the input voltage ‘‘Vi’’ falls below a value say ‘VL’, the comparator
output should switch to Low (or) OFF state with magnitude of output voltage as
O Volts. The state changes exhibited by the comparator output occur when the
voltage on its input terminal say ‘Vx’ is equal to setpoint or reference voltage
“V ref”.
 5TH PROOF 4-9-09

174 Process Control Engineering

The upper (High) and lower (Low) units are

VH = Vref, and ...(2)

R1
VL = Vref – . VO ...(3)
R2

The width of the neutral between VL and VH can be adjusted by varying the
resistor R2 in Fig. 5.3. The extent of width of neutral zone can be calculated
from the difference of equations (2) and (3). The typical value of inverter
resistance may be in the range of 1 W to 100 K W range.

Vout

(ON) V0

(OFF) O
Vi
VL VH

FIG. 5.4

Multiposition Controller
The implementation of three position controller (multiposition) using two
comparators, a summing amplifier and an inverter is shown in figure 5.5.

Vout

Vout 2

Vout 1

O Vi
Vref (1) Vref (2)
(Set point 1) (Set point 2)

FIG. 5.5
 5TH PROOF 4-9-09

Implementation of Control Actions 175

Figure 5.5 shows the response of three position controller with bias voltage
zero.
The bias voltage is the minimum output voltage when the input voltage Vi
is less than Vref (1) or setpoint – 1.
In the circuit shown in figure 5.6, the comparator output is assumed to be
zero for OFF state and ‘VO’ volts for ON stage. The bias voltage input is
included in the summing amplifier represented as ‘VB’.

VB (Bias Voltage)

R3
Vi – R2 R3
1
Vref (2) +

–
2 – R
–
+ R1
Vref (1) + 0
+ Vout =
Vo
Inverter
Comparators Summer
Amplifier

FIG. 5.6

when Vi < Vref– 1 Vout = VB. U|

Vref –1 < Vi < Vref– 2
R
Vout = VB + 3 .VO
R1
|V ...(4)
R3 R ||
Vi > Vref 2 Vout = VB +
R1
.VO + 3 .VO
R2 W
Floating Controller
The electronic floating controller is implemented by connecting the output of
three position controller to an integrator.
During the interval Vi < Vref– 1
Vi = V1

V1 - V V - Vout
\ = [(V = 0) virtual GND]
R 1/ CS

J
- V1 -8
Vout = Vout = J
RcS 4+ 
 5TH PROOF 4-9-09

176 Process Control Engineering

1
Cs

+ V1
Vi Vref 1 R
O –
Vref 2 Vi V
+
– V1
Vout

Integrator

FIG. 5.7

During the interval Vref– 1 < Vi < Vref– 2 the output will float at previous
setting, on whatever magnitude of the Vout.

During the interval Vi > Vref– 2

Vi = – V1
- V1 - V V - Vout
= (V = 0)
R 1/ CS

V1
- = - Vout . CS
R
t2
V1 V
\ Vout = = 1 t
RC S RC 0
Times t1 and t2 indicate the intervals when the error is in the limits of
neutral zone.

5.2.2 Implementation of Continous Controllers

Proportional Controller
The implementation of proportional controller using op-Amps must satisfy
the equation
P = K p e + P0 ...(5)
where, Kp = Proportional
P0 = Initial value of controller O/P.
Consider a summer realized using op-amps. along with an inverter.
Let Ve = error voltage
Vout = controller output.
 5TH PROOF 4-9-09

Implementation of Control Actions 177

R2 R2
Vo R2

R1
Ve – R
–
+ Vout
+ + Vout

Summer Inverter

FIG. 5.8

Let V = virtual ground.

Ve - V V0 - V V - Vout
+ =
R1 R2 R2
Since V =0
Ve V0 - Vout
+ =
R1 R2 R2
\ O/p of summer is

R2
–Vout = .Ve + V0.
R1
and the output of inverter will yield
R2
+Vout = .Ve + V0. ...(6)
R1

Comparing Eqs. (5) and (6)

R2
Kp = = Proportional Gain.
R1

Integral Controller
The analytic equation of integral controller from previous chapters is
reproduced below.

P=
1
Ti z e dt + P(0). ...(7)

where, Ti = Integral Time

P(0) = controller o/p at t = 0 (Initial value).
The electronic implementation must satisfy the above equation (7).
Consider an integrator circuit along with inverter.
 5TH PROOF 4-9-09

178 Process Control Engineering

1
Cs
R

R
Ve – R
V –
+ Vout
+ + Vout

Integrator
Inverter

FIG. 5.9

Ve - V V - Vout
= (Since V = 0)
R 1/ CS
Ve
–Vout =
RCS
V
Þ +Vout = e (o/p of inverter).
RCS
1
= + Vout =
RC z
Ve dt + V(0)

Thus, equation (8) correlates with that of eqn. (7).

...(8)

The integral time Ti = RC.

V(0) = Controller output at t = 0 (Integral Constant).

Derivative Controller
Practically, Derivative controller mode is not used alone for the reasons
already stated in chapter – 2. However, since this controller is used in
combination with other modes its electronic implementation is discussed here.
The analytic equation of derivative controller is

de
P = Td ...(9)
dt
Thus, to get this kind of response a differentiator along with an inverter is
employed.
The analysis of differentiator yields the equation given by (9).
Ve - V V - Vout
=
1/ CS R
-Vout
Ve.CS =
R
 5TH PROOF 4-9-09

Implementation of Control Actions 179

R
R

1
Cs
Ve – R
V –
+ Vout
+ + Vout

Differentiatior Inverter

FIG. 5.10

Þ –Vout = RCSVe.
dV
–Vout = RC e
dt
dVe
O/p of Inverter will be + Vout = RC ...(10)
dt
Comparing equations (9) and (10), the derivative time Td = RC
Vout = controller output
Ve = Error voltage (Input).
From practical perspective the Fig. 5.10 cannot be used because it tends
to be unstable, i.e. it may exhibit oscillations in the output voltage. The
occurrence of this instability is because the magnitude of the output voltage
increases linearly with frequency. Therefore, when the frequency is infinity, the
output tends to be infinity.
Consider the equation
–Vout = RCSVe
The o/p of the inverter will be
+Vout = RCSVe
Put S = jw
+Vout = jwRCSVe

|Vout| = 02 + ( wRCVe )2
|Vout| = wRCSVe
Since w = 2p f
|Vout| = 2p f RCSVe. ...(11)
Thus, |Vout|af which implies a little high frequency noise will cause large
excursions in the output voltage which is not acceptable in control system
design.
 5TH PROOF 4-9-09

180 Process Control Engineering

A practical derivative controller therefore has to ensure that it provides a

derivative output in the frequencies of practical interest and at higher
frequencies it acts simply as a fixed gain amplifier.
A simple modification of circuit shown in Fig. 5.10 wherein a resistor is
placed in series with the capacitor forms the practical circuit of derivative
mode.

R2

1
R1 Cs
Ve –
V
+ Vout

FIG. 5.11

Ve - V V - Vout
= (Since V = 0).
R1CS + 1 R2
CS
VeCS -Vout
=
R1CS + 1 R2
R2CSVe
= – Vout.
R1CS + 1
Put s = jw
jwR2CVe
–Vout =
jwR1C + 1

wR2C|Ve|
|Vout| =
1 + ( wR1C)2
Since w = 2p f.
2pf R2C|Ve|
|Vout| = ...(12)
1 + (2pf R1C)2
For frequencies in practical range.
2p f R1C << 1
|Vout| = 2p f R2C |Ve|. The circuit acts as a derivative controller.
 5TH PROOF 4-9-09

Implementation of Control Actions 181

For higher frequencies.

2p f R1C >> 1

2p f R2 C |Ve|
|Vout| =
2p f R1 C

R2
|Vout| = |V |. It acts as a fixed gain amplifier. Thus, it can be stated
R1 e
that this modification essentially clamps the gain above some frequency to a
constant value. And it is rest assured that the clamped frequencies are well
above anything that could occur in the actual control system.

Composite Controllers
Proportional + Integral Controller
It is basically the combination of proportional and integral actions,
respectively. The analytic equation of this combination is

P = Kpe +
Kp
Ti z e dt + P(0) ...(13)

The implementation of P + I controller that gives an equation similar to

that shown in equation (13) is as shown in figure 5.12.

1
R2 Cs
R

R1
Ve – R
V –
+ Vout
+ + Vout

PI Controller Inverter

FIG. 5.12

Ve - V V - Vout
=
R1 R2CS + 1
CS
Since V = 0 (Virtual Ground)
Ve ( R2 CS + 1)
= – Vout
R1CS
 5TH PROOF 4-9-09

182 Process Control Engineering

Ve . R2 CS Ve
–Vout = +
R1 CS R1CS

R2 Ve
–Vout = .Ve +
R1 R1CS
The output of the inverter will be
R2Ve Ve
+Vout = +
R1 R1CS

Þ +Vout =
R2
R1
R
.Ve + 2 ×
1
R1 R2C z Ve dt + V(0) ...(14)

Comparison of equations (13) and (14) will indicate that

Vout = Controller output;
Ve = Error voltage
Kp = Proportional gain = R2 /R1
Ti = Integral time = R2C
Proportional + Derivative Controller
The combination has an analytic equation represented as
de
P = Kpe + KpTd ...(15)
dt
The analysis of the circuit shown in Fig. 5.13 should give an output
equation that correlates with equation (15).
R2
R
R1

Ve – R
V –
+ Vout
+ + Vout
1
Cs

P + D Controller Inverter

FIG. 5.13

Ve - V V - Vout
=
R1 R2
R1CS + 1
Since V = 0 (Virtual ground)
 5TH PROOF 4-9-09

Implementation of Control Actions 183

Ve [ R1CS + 1] -Vout
=
R1 R2
R2 R
–Vout = .Ve + 2 .R2CSVe
R1 R1
o/p of inverter will be
R2 R dV
+Vout = .Ve + 2 .R2C e ...(16)
R1 R1 dt
\ Comparing equations (15) and (16)
R2
Kp = ; Td = R2C
R1

Proportional + Integral + Derivative Controller

The most versatile control action is the PID action. The analytic equation is
reproduced below, that determines the action of this controller to errors.

P = Kpe +
Kp
Ti z e dt + KpTd
de
dt
+ P(0) ...(17)

where P(0) = Initial value of controller o/p. There are many modes of
realisation of this control action but with minimum number of op-Amps its
implementation is shown in Fig. 5.14.
1
R2 C2s
R
R1

Ve – R
V –
+ Vout
+ + Vout
1
C1s

PID Controller Inverter

FIG. 5.14

Ve - V V - Vout
=
R1 R2C2S + 1
R1C1S + 1 C2S
(Since V = 0)
Ve [ R1C1S + 1][ R2C2S + 1]
= – Vout
R1C2S
 5TH PROOF 4-9-09

184 Process Control Engineering

Ve
\ –Vout = {[R1C1R2C2S2 + [R1C1 + R2C2]S + 1]}
R1C2 S

\ –Vout = R1C1SVe +
LM C
1
+
R2
V +
VeOP
NC2 R1 e R1C2SQ
C1 R
Since the ratio < < 2 and at the output terminals of inverter, the
C2 R1
above equation reduces to

+Vout =
R2
R1
dV
.Ve + R2C1 e +
dt
1
R1C2 z Vedt + V(0).

Þ +Vout =
R2
R1
R
Ve + 2 .
1
R1 R2 C2 z Vedt +
R2
R1
dV
.R1C1 e + V(0)
dt
...(18)

Comparing equations (17) and (18)

Kp = R2/R1 Ti = R2C2; Td = R1C1 respectively.

5.3 INTRODUCTION TO PNEUMATIC ELEMENTS

In pneumatic systems the working signal is air employed in the pressure range
of 3 – 15 psi.
The standard pneumatic process control signal is derived from regulated
air supply of 20 – 30 psi.
Like how we make use of operational amplifiers (op-amps) in the making
of electronic controllers, the implementation of pneumatic controllers are
based on the principle of certain pneumatic elements.
The basic function of these pneumatic elements is to provide the standard
3 – 15 psi pneumatic pressure signal or sometimes enhance this pressure
signal to meet the demands of next stages that appear after the controller.
Most of the final control elements are either pneumatically or
hydraulically activated devices that operate within the range of 3 – 15 psi.
Hence, the implementation of pneumatic controller should incorporate these
elements, among which the most important signal conversion is from pressure
to mechanical motion and vice versa with working pressure range of 3 – 15 psi
as output. It is provided by FLAPPER/NOZZLE mechanism, which will be
discussed now. Most of the Pneumatic Controllers employ this flapper-Nozzle
Mechanism along with the assembly of bellows as will be discussed later in this
chapter.
 5TH PROOF 4-9-09

Implementation of Control Actions 185

Nozzle and Flapper System

It is also called as Nozzle and BAFFLE system. It consists of a nozzle with an
opening over which a baffle or flapper is pivoted in such a way that the
movement of flapper increases or decreases the gap distance between itself
and nozzle opening as shown in Fig. 5.15.

Flapper
Pivot Flapper motion

Gap

Supply
Pressure
(20 psi) (Restriction)

Signal pressure

FIG. 5.15

A regulated supply of pressure, usually 20 psi is supplied through a small

orifice or restriction about 0.0254 cms internal diameter to the open nozzle.
The nozzle is open at the top end where the gap exists between flapper
and nozzle. The flapper is pivoted in such a way that it can completely close
the gap or may allow air to escape through the gap.
If the flapper moves down and completely covers the gap, the signal
pressure will rise to supply pressure. As the flapper moves away from gap
allowing the air to escape the signal pressure will drop.
Fig. 5.16 shows the characteristics between signal pressure and gap
distance. It is interesting to note that the linear (approximately) characteristics
are recorded between 3 – 15 psi pressure region only, thus providing
maximum sensitivity in the central region.
Therefore, the flapper/nozzle mechanism is designed to operate in the
central region where the slope of the line is greatest. Also the standard
pressure signal of range 3 – 15 psi is obtained as output.

Current to Pressure [I/P] Converters

The I/P converter is one of the most important elements of process control
loop. Generally most of the final control elements are pneumatically driven
devices, hence when electronic controllers are employed in process control
loop it becomes necessary to convert the output of controller which is usually
an electrical signal to pneumatic signal so that the final control element can be
activated. Thus, the current to pressure converters perform this energy
conversion of electrical to pneumatic signals.
 5TH PROOF 4-9-09

186 Process Control Engineering

Signal Pressure

20 psi
(Supply pressure)
15 psi

Linear characteristics

3 psi

Gap Distance

FIG. 5.16

The current to pressure gives a linear way of translating the working

electrical current range of 4 – 20 MA into useable working pressure range of
3 – 15 psi.
The current flowing through the force coil produces a force of attraction,
that tends to pull the flapper down to close the air gap between itself and
nozzle as shown in Fig. 5.17.

Spring
Assembly
Flapper
Pivot

Gap
4-20
N (mA)
o
z
Supply z
Pressure l Coil
(20 psi) e (Force)

3-15 psi
signal

FIG. 5.17

To obtain a linear translation of 4 – 20 MA into 3 – 15 psi, the spring

assembly is adjusted with respect to the flapper and its relative motion towards
the air gap of the nozzle.
 5TH PROOF 4-9-09

Implementation of Control Actions 187

Though there are many designs involved in making of I – P converters but

the basic principle almost always involves the flapper – Nozzle system shown in
Fig. 5.17.
An I/p converter is thus a force balance device in which the coil is
suspended in the field of magnet by a flexure. The relative motion of the
flapper is in direct proportion to the force exerted by current carrying coil.

Pneumatic Pilots
These pneumatic elements are basically used for improvisation of the speed of
pneumatic transmission. One such method is pneumatic amplifier as shown in
Fig. 5.18.

Air Supply

Output
Pressure

Leak

Signal Pressure

FIG. 5.18

The pneumatic amplifying pilot is also known as booster or relay. Its main
function is to raise the air pressure and/or air flow volume by some linearly
proportional amount from the input signal. For the input of 3 – 15 psi, if the booster has
a pressure gain of 100 the output would be 300 – 1500 psi.
In the figure 5.18 if the signal pressure is such that the diaphragm motion moves
the plug in the body block of the booster down the gas leak is reduced and the pressure
in the output line is increased.
Similarly, if the diaphragm moves the plug upwards blocking the air supply then,
the pressures in the input transmission line decreases as there will be considerable air
leak.
A part from power amplification there are also there pilots which are used
for overcoming transmission lags. They are made of double bellows assembly
and are directly inserted in the transmission line.
 5TH PROOF 4-9-09

188 Process Control Engineering

5.4 PNEUMATIC IMPLEMENTATION OF

CONTROLLER MODES
In section 5.2 the electronic implementation of both discontinuous and
continuous control actions has been discussed at great length. However, the
often used control actions in many of the process control industries are
composite controller modes like PI, PD and PID, respectively. Therefore, the
pneumatic implementation of these modes will be discussed in this section.
Unlike electronic controllers, the pneumatic controllers are more bulky and
also more robust with air as transmission medium. These pneumatic
controllers are implemented using the combinations of pneumatic elements
discussed in Section 5.3.

Pneumatic Proportional Controllers

A pneumatic proportional controllers, that employs a power amplifying pilot
for providing larger quantity of air flow, incorporating a negative feedback is
shown in figure 5.19.

Kp La Feedback bellows
rge

all
K p Sm

Flapper

e (Pilot)
Nozzle ‘P’ Output
(Error) Pressure

Air Supply

FIG. 5.19

The principal of operation is as follows:

When the error signal “e” increases the flapper moves toward right side
covering the air gap between itself and nozzle.
As a result the nozzle back pressure increases which in turn increases the
output pressure “P”. The part of the output pressure also flows through the
feedback bellows and repositions the flapper in such a way that the required
controller output ‘P’ is in direct proportion to error (or) actuating signal ‘e’
(P a e).
The flapper is repositioned to the left away from nozzle air gap by the
feed- back bellows till the required output pressures range is obtained.
The different values of “K p” or proportional gain is obtained by
connecting the feedback link to the flapper appropriately. When the feedback
 5TH PROOF 4-9-09

Implementation of Control Actions 189

link is connected in the upper position as shown in Fig. 5.19. The feedback
pressure is small making the proportional gain large (small P.B) thus resulting
in large output pressure change.
Similarly when the feedback link is connected in the lower position, Kp is
small i.e. very large P.B. (Proportional band). By proper design we can choose
the appropriate location for connecting the feedback link to the flapper, so
that the output pressure (P).
P = Kp e.
The effect of variations in temperature, any leakage and supply pressure
causing deviation of the output pressure from the required value is reduced to
minimum because of the negative feedback action of feedback bellows. Also
the P.B and proportional gain may be varied over a wide range of values
(1000 to 1).

Pneumatic Proportional + Integral Controller

The construction of P + I controller is shown in Fig. 5.20. It differs from
proportional controller only in the addition of positive feedback bellows with
restriction ‘R’. The free ends of positive and negative feedback bellows are
rigidly connected with the flapper attached to the bellows rod by an adjustable
link as shown in Fig. 5.20. The combination of capacitance ‘C’ in positive
feedback bellows and resistance ‘R’ i.e. positive feedback restriction provided
by needle valve, gives the required integral action.

(Positive feedback restriction)

R

Positive f.b. igh) Negative f.b.

bellows K p (H
bellows

)
mall
K p (S

Flapper
Pilot
Nozzle Assembly
e Output
Error Pressure
signal ‘P’

Supply
Pressure

FIG. 5.20
 5TH PROOF 4-9-09

190 Process Control Engineering

The operation is as follows:

When the error signal ‘e’ increases it moves the flapper to the right thereby
closing the air gap of nozzle. This results in increased output pressure ‘p’. The
part of increased output pressure also flows through negative f.b. bellows,
which reposition the flapper to the left to get required output pressure. This is
‘p’ action (or) proportional action. Let the pressure in negative feedback
bellows be P(–). The pressure in the positive feedback bellows depends on the
pressure drop across the restriction ‘R’. Thus, by slowly changing ‘R’, the
pressure in the positive feedback bellows can be changed, which will reposition
the flapper by moving it slowly towards right side and steadily increasing the
output pressure ‘P’. Let the pressure in positive feedback bellows be P(+).
The error signal (e) is nothing but the difference between pressures in
negative and positive feedback bellows
e = P(–) – P(+)
Through resistance ‘R’ a flow occurs which is proportional to P(–) – P(+).
Thus, the rate of change of pressure in positive feedback bellows is directly
proportional to P(–) – P(+) i.e. error (e).
dp( + )
R = P(–) – P(+).
dt
dp( + )
or RC = e.
dt
Thus, the integral time adjustment can be done by varying ‘R’.
The value of Kp can be adjusted by connecting the feedback link upwards
or downwards as shown in Fig. 5.20.

Pneumatic Proportional + Derivative Controller

P + D control is obtained by simply placing a restriction ‘R’ in the negative
feedback pressure line. This results in delayed negative feedback and large
output pressure.

Kp (H
igh )
Negative
feedback
bellows
all)
K p (Sm
R
Flapper
Pilot
Nozzle Assembly
error
Output
e Air gap Pressure
‘P’
Supply Pressure
(Air)

FIG. 5.21
 5TH PROOF 4-9-09

Implementation of Control Actions 191

The operation is as follows:

Consider the error signal to be changing linearly with time, it slowly starts
making the flapper move towards right causing the closure of air gap of the
nozzle. The results in steady increase in the output pressure ‘P’. Since part of
this output pressure also flows through the negative feedback line, but because
of the restriction ‘R’ the pressure in the feedback bellows is considerably less.
This results in delayed repositioning of flapper towards left as a result the
output pressure will tend to lead than lag, the error signal ‘e’, which clearly
indicates the derivative action.
Thus, the delayed negative feedback produces the derivative action. The
value of proportional gain ‘Kp’ is adjusted by connecting the feedback link to
the flapper upwards or downwards.
The capacitance of negative feedback bellows and resistance of the needle
valve ‘R’ together constitute the derivative time “Td’’.
\ Td = RC.
By varying ‘R’ ‘Td’ can be changed.

Pneumatic Proportional + Integral + Derivative

Controller
PID action is achieved by adding a restriction in the negative feedback line as
shown in Fig. 5.22.
Positive f.b.
Restriction (R2)

Positive Negative
f.b. bellows f.b. bellows

Negative f.b.
Restriction (R1)
Flapper
Pilot
Nozzle Assembly
Output
error Pressure
(e) (P)
Supply Pressure

FIG. 5.22

The time constant element T 2 = R2C2 corresponding to capacitance of

positive feedback bellow and restriction R2 determines the reset or integral
time “T i”.
 5TH PROOF 4-9-09

192 Process Control Engineering

Similarly, the time constant element T1 = R1C1 corresponding to restric-

tion in negative feedback line (R1) and capacitance of negative feedback
bellows (C1) determines the derivative time ‘Td’.

Thus, Ti = R2C2 and Td = R1C1

The value of proportional gain ‘Kp’ is obtained by adjusting the position of

the connecting link between the flapper and negative f.b. bellows.
When the error signal increases linearly with time causing the flapper to
move towards right side thereby closing the air gap and resulting in an increase
of output pressure, the repositioning of flapper is now dependent on delayed
negative feedback causing a lead in output pressure and the rate of change of
pressures in the positive feedback bellows proportional to the pressures
difference in negative and positive feedback bellows flowing through
resistance ‘R2’.
Initially, the lead in the output pressure indicates the derivative action
followed by integral + proportional actions which tend to reposition the
flapper position to get the desired response. Consider the equation
d
R2 P(+) = P(–) – P(+).
dt
P(–) indicating the pressure in negative feedback bellows, in PID action is
less because of resistance elements ‘R1’ and thus the integral action lags in its
response.
The desired PID characteristic for this arrangement when the error or
actuating signal is linearly changing with time is. As seen in Fig. 5.24, initially
the output pressure (P) leads i.e. it is more as the negative feedback action in

)
C2
2
, R I e
p P+ ons
(K sp
(P) re
Lag
PID
response
R1C1 (Derivative)
t
(Lead)

FIG. 5.23
 5TH PROOF 4-9-09

Implementation of Control Actions 193

repositioning the flapper is delayed because of R1 restriction. However,

because of the action of positive feedback bellows, the rate of change of flapper
position adjustment tends to make the output pressure fall steadily till it
reaches the desired value.

5.5 HYDRAULIC IMPLEMENTATION OF

CONTROLLER MODES
Introduction
In many of the industrial process control applications, hydraulically operated
controllers are employed to deal with certain control elements and other
auxiliary systems which require greater power to handle them.
A hydraulic controller generally consists of two operating units.
(1) A pilot valve to control oil pressure and oil flow output.
(2) A power cylinder and piston to provide required displacement of the
control valve or any other auxiliary system.
The power supply is usually a high pressure oil or in certain instances high
pressure water. Power is supplied to the hydraulic systems with the help of
electrical motor pump as shown in figure 5.25.

Relief Valve To the

Hydraulic
Filter System
Electric
Motor Pump

From
Hydraulic
System

Filter
Sump

FIG. 5.24

The electric motor which is usually of gear type pumps oil from the sump
through filter and passes it to the hydraulic system with certain pressure. A
relief valve is provided in the supply line which when opened sends oil back
into the sump during any untoward emergencies.
The sump collects the return oil from the system and serves to dissipate
heat, which is very important in hydraulic system. Sometime oil coolers are
also used for dissipating the heat.
 5TH PROOF 4-9-09

194 Process Control Engineering

In the event of electrical failure an accumulator is usually connected to the

pump outlet which serves as a oil reservoir as a secondary option, as the motor
pump should run continuously.
An accumulator consists of two portions, of which one portion is filled with
nitrogen gas and the other side filled with oil with suitable check valves for
obtaining the desired oil pressure and speed.
It can supply oil pressure for several minutes or hours depending upon its
design volume to the hydraulic system in the event of sudden electrical power
failure. The required pressure is obtained from a diaphragm separating the
gas section and oil section.

To Hydrautic System

Oil

Compressed Diaphragm
Nitrogen Gas

Accumulator

FIG. 5.25

Hydraulic Proportional Controller

The arrangement of hydraulic proportional controller is shown in Fig. 5.27.
The error or actuating signal is mechanically coupled to the tube attached
between the hydraulic supply system and the power cylinder, which has a
power piston assembly inside it. The operation is as follows:
When the error signal increases, it moves and the connects the tube to the
sump, this results in the movement of the power piston towards left side as the
oil in the power cylinder flows into the sump through the connecting tube,
thereby reducing the oil pressure in it. The feedback link then moves to the
right side which tends to connect the tube to the pump resulting in increased
oil pressure in the power cylinder. The increase in oil pressure in the cylinder
causes the power piston move towards right side at a new position. Therefore,
the position of the power piston is made proportional to the actuating or error
signal “e”.
\ P µ e Þ P = Kp .e
The proportional gain Kp is adjusted by moving the connection of the
feed- back link on the stem where error or actuating signal is applied.
 5TH PROOF 4-9-09

Implementation of Control Actions 195

Feedback Link Connecting

Stem

Oil
Error P
e
Power Power
Cylinder Piston

From To
Pump Sump

FIG. 5.26

Hydraulic Integral Controller

The arrangement to implement a hydraulic integral controller is shown in
figure 5.28. The rate of change of controller output proportional to the error is
obtained by the power piston.

i.e.
dp
dt
µ e ® P = KI z e dt.

Unlike the proportional action, the power cylinder outlets are connected
simultaneously to pump and sump depending upon the magnitude of the
error signal.

Right Side Outlet from Power Cylinder

Left Outlet from Power Cylinder

dP/dt

Oil Oil
Error
Signal
(e)

From To
Pump Sump

FIG. 5.27
 5TH PROOF 4-9-09

196 Process Control Engineering

The operation may be understood as follows:

Depending upon the magnitude of the error signal say the right side
outlet is connected to the pump and the left side outlet is connected to the
sump then oil is drained out from the left side and at the same time filling the
right side. Since on the both sides of power piston either oil is filled or drained
out at the same instant it produces a continuous motion of the power piston
towards left side or right side. This motion of power piston thus represents rate
of change, proportional to actuating or error signal e.
The adjustment of integral time is complex in hydraulic arrangement as it
involves accurate positioning of the connecting link between error signal stem
and connecting tubes of power cylinder.

Hydraulic Proportional + Integral Controller

The implementation of P + I controller has three stages.
(1) The measuring means is the diaphragm which produces a force
proportional to the value of controlled variable.
(2) The set point means is the spring and knob assembly.
(3) A feedback cylinder with a movable piston.
The diaphragm causes a force proportional to the value of the controlled
variable to act against the balance arm. Additionally a force is also exerted by
the set point spring furthered by the feedback spring also on the balance arm.
Reset time (Ti)
Adjustment
Feedback
Cylinder

Set point knob

B
a
l a
a r
n m
c
e

Diaphragm
Jet
Pipe

Power Power
Cylinder Piston
From To
Controlled Pump Sump
Variable
“C”

FIG. 5.28
 5TH PROOF 4-9-09

Implementation of Control Actions 197

Depending upon the magnitude of the error signal which is difference

between the set point and controlled variable, the jet pipe may be connected to
either pump or sump. If the jet pipe connects the power cylinder tube to the
pump it directs high pressure oil into the left side of power cylinder
distribution block. The power piston moves to the right due to the increase in
pressure in the left side, causing oil to flow into the right side of the feedback
cylinder. The feed- back piston moves to the left, increases the force on the
balance arm, returns the jet pipe to a balanced position. This is proportional
action. The integral action is applied by adjusting the integral needle valve
which repositions the jet pipe in such a way till the offset of proportional
action is eliminated. The proportional gain is adjusted by suitable levers
between the feedback spring and the balance arm. The integral time is
adjusted at the needle valve on the feedback piston.
The main sources of difficulty in implementing hydraulic controllers are
dirt and heat. Regularly cleaning the system will help in removing dirt,
particles of metal and oil deposits. Heat causes variations in the viscosity of the
oil and therefore proper cooling methods must be adopted. Usually a large
area sump is used for dissipating the heat.
Another major disadvantage is the portability of hydraulic systems. Since
they are usually bulky in structure, they are not easily portable. They have to be
operated by assigning to one particular position only. Periodic checks for oil
leaks, dirt accumulation in the power cylinders and other accessories and
viscosity of the oil is an absolute necessity for hydraulic systems.

QUESTION SET
1. With next diagrams obtain the implementation of composite control
actions electronically using op-Amps?
2. What are the different types of pneumatic elements used in process
industries? Explain with next diagrams?
3. With a next diagram explain the operation of P to I converter?
4. Explain with next diagrams:
(a) Pneumatic P + I controller
(b) Pneumatic P + I + D controller.
5. Obtain the mechanism of hydraulic implementation of controller
modes? Explain with next diagrams?
 $
+0)26-4

Final Control Elements

6.1 INTRODUCTION
One of the most important components of any process control system is the final
control element. The output of the automatic controller is fed to the final control
element which directly affects the manipulating variable to achieve the desired
control objective. Thus, a final control element may be defined as
“The mechanism which alters the value of the manipulated variable in
response to the output signal from the automatic controller”.
The final control element, abbreviated as “FCE” essentially consists of two
parts
(1) An ACTUATOR, which is primarily used to translate the output signal of
the automatic controller into a position of a member exerting the required
force or displacement or power.
(2) A control valve to adjust the value of the manipulated variable.

6.1.1 Introduction to Actuators and Positioners

An actuator is that portion of a control valve that responds to the applied signal
and causes the motion resulting in modification of fluid flow or related dynamic
variable in the process. It is thus a primary element causing displacement of the
valve stem. It must provide an accurate output position of a member which it is
connected to, proportional to the input signal, in spite of various forces acting
on the output member such as thrust forces caused by weight and unbalanced
fluid pressure, static friction and inertia forces caused by the mass of the moving
parts.
It may be manually positioned device like a lever or hand wheel, positioned
between fully open or fully closed.
 5TH PROOF 4-9-09

Final Control Elements 199

Some of the other actuators are operated by compressed air, hydraulics and
also by electrical means. Thus, the actuator designs depend on the particular
signal with which it is operated upon as enlisted above.

Actuator Types

Electromechanical Actuators
Electro-hydraulic Actuators
Electrical Types

Pneumatic Actuators

The relative merits, demerits and applications of different types of actuators

is discussed in the following Table 6.1.
Table 6.1
Actuator type Merits De-merits Applications

Electromechanical ® High thrust ® Complexity in design. Linear or rotary

Actuators ® High and reliable ® Absence of valves of size 2 -36
stiffness coefficient mechanical fail safe. inches.
powered by ® Bulky structure.
electricity or
pneumatics.
Electro Hydraulic ® High and reliable ® Complexity in design Linear or rotary
Actuators stiffness coefficient ® Bulky structure valves of size 2
® Fast speed ® Specific mounting inches to unlimited
® High thrust and installation sizes.
requirements
® Powered by electrical ® Fluid viscosity check
source
® No pneumatic supply ® Fluid leaks
is required
® Applicable for ® Temperature sensitive.
unlimited valve sizes.
1
Electrical ® Fast and reliable Low thrust Linear valves inch
2
Actuators to 2 inches only.
® Powered by electrical ® Slow speed for
source only. certain applications
®Direct interface with ® Require electrical
computer systems drives for physical
process system
interfacing
® Limited valve sizes.
Pneumatic ® High thrust at low air ® Lack of stiffness Linear and rotary
Actuators pressures 1
valves of inch –
2
24 inches.
Cont.
 5TH PROOF 4-9-09

200 Process Control Engineering

® Longer strokes at ® Instability

higher air pressures
® Mechanical fail safe
option
® Good control with
control devices

6.1.2 Electrical Actuators

Electrical actuators are basically reversible electrical motors provided with
internal worm gear and related accessories to prevent drive direction reversal or
back drives caused by unbalanced loads. Other types include servomotors and
stepper motors. Two phase A.C. Servomotors are available upto 1 hp rating
where-as direct current servomotors can meet higher loads.
Stepper motors which rotate the shaft by a discrete step angle when energised
electrically can also be used as valve actuators. The electromechanical and
electrical types come under the category of electrical actuators.
Electrical motors are devices that accept electrical input and produce a
continuous rotation as output. The motor gear actuator consists of an electrical
motor connected to a gear train, which rotates a stationary drive nut which in
turn drives threaded valve stem. When the valve plug contacts the seat, the
resistance is transmitted to a belleville spring, which at a preset limit interrupts
the motor power circuit. Lever operated valves and dampers are available with
torques of 6.5 to 680 N-m with full stroke speeds of 10 to 60 seconds. For larger
valves, the motor actuator can deliver torques upto 203 to 13,600 N-m with a full
stroke speed of 30 to 300 seconds.
In Fig. 6.1, the structure of rotary output actuator which can deliver an
output speed of one rpm with an input motor speed of 1800 rpm is shown.
It uses a double worm gear reduction with self-locking and thereby
preventing the load from moving downward by back driving the motor.

Geared limit Prime mover

switches

Hammer
blow

Gear drive

Toroue De-clutching
Drive
control handwheel
slfeve

FIG. 6.1
 5TH PROOF 4-9-09

Final Control Elements 201

Reversible
Low-Inertia
AC Motor

Continously
Connected
Handwheel

Motor Drive
Pinion Gears

Load Connection
Linear Motion

FIG. 6.2
In Fig. 6.2, a linear output actuator which uses a worm and a rack and Pinion
arrangement to translate horizontal shaft motor output to vertical linear motion is
shown.
This actuator is designed with a conventional globe valve bonnet for ease of
mounting.
A continuously connected hand wheel, which must rotate the rotor of the
motor, can be used when there is a short stem travel and relatively low force
output. However these electrical actoators employing output gear system for
reduction in speed employ AC motors and/or D.C. motors respectively.
The maximum force output in this case is 680 kgf at about 2.5 mm/minute.
Since the subject matter of electric motors is extensive, the reader is advised
to refer the related texts for more details. In our present context, the application
of these motors is stressed which enables them to be used for actuation purposes.
Apart from worm gears showed in Fig. 6.1, of another most common actuator
employed in industrial application is the spur gear actuator.
The spur gear actuator shown in Fig. 6.3 in contrast to worm gears has very
low power loss. The advantage of spur gear is that it prevents the back driving
torque on the actuator even if the load possesses frictional characteristics.
The spur gear actuator consists of adjustable cams to operate limit and
auxillary switches. A potentiometric or tachogenerator feedback calibrated with
respect to the rotation of the actuator is required for formulating a control circuit.
Actuators of this type must be adaptable to mounting on and operating a variety
 5TH PROOF 4-9-09

202 Process Control Engineering

Output
Shaft Adjustable
Cams

(Optional)
Travel Potentiometer
Limit
Switches Auxiliary
Switches

FIG. 6.3

of quarter turn valves. Hence, the opening of control valve can be controlled by a
limit switch whereas for shutting it is advisable to incorporate a torque limiting
device to sense the closure of the valve.

6.1.3 Pneumatic Actuator

Pneumatic actuation employs compressed air technology, i.e., they operate on
supply of compressed air which must be made available in sufficient quantity and
pressure in accordance with the capacity of the system. The continuous supply of
compressed air is made available through a compressor.
There are basically two types of designs available for pneumatic actuation
(1) Spring and Diaphragm Actuator
(2) Piston Actuator.
In spring and diaphragm actuator, variable air pressure is applied to a
flexible diaphragm to oppose a spring attached to it. The combination of
diaphragm and spring forces acts to balance the fluid forces on the valve.
In pneumatic piston actuator, a combination of fixed and variable air
pressures is applied to a piston in a cylinder to balance the fluid forces on the
valve. A piston actuator is implemented in two ways.
(i) Cushion Loaded: In this type a fixed air pressure known as cushion
pressure is used to balance the fluid forces on the valve.
(ii) Double acting: In this type two opposing variable air pressures are used to
balance the fluid forces on the valve.
 5TH PROOF 4-9-09

Final Control Elements 203

Both the spring and diaphragm actuator and the piston actuator produce
linear motion to move the control valve. They are ideal for use on valves whose
stem motion is linear, i.e., globe valves. For rotary valves such as butterfly valves
some mechanism of linear to rotary motion conversion must be employed.

Spring and Diaphragm Actuators

The basic principle of spring and diaphragm actuator is based on the concept of
pressure as force per unit area. The stem positioning is achieved by a balance of
forces acting on the stem. These forces are due to pressure on the diaphragm,
spring travel and the fluid forces acting on the valve plug. When a net pressure
difference is applied across a diaphragm of surface area ‘A’, then a net force acts
on the diaphragm given by the
F = (P1 – P2) A ...(1)

where, P 1 – P2 = P = Pressure difference (Pa)

A = Effective diaphragm area (m2)

\ Equation 1 can be expressed as F = PA ...(2)

This force (equation 2) exerted by the diaphragm is opposed by spring and
valve plug.
There are two basic designs of Pneumatic actuators using spring and
diaphragm.
® DIRECT ACTING
® REVERSE ACTING
In direct acting actuator, the forces due to pressure on the diaphragm, spring
travel and fluid forces on the valve plug are expressed as
PA – KX – Pv Av = 0

Þ PA = KX + Pv Av ...(3)

A = Effective area of the diaphragm (m2)

Av = Effective inner valve area (m2)

K = Spring constant (N/m)

P = Diaphragm pressure (N/m2)

Pv = Valve pressure drop (N/m2)

X = Stem travel (m)

The direct acting actuator is governed by equation 3. It applies to a push-
down-to-close actuator and valve combination with flow of fluid under the plug.
 5TH PROOF 4-9-09

204 Process Control Engineering

p A

Flow Flow
At PV

AV
(a) (b)
Direct acting actuator Reverse acting actuator
FIG. 6.4
In the reverse acting mode the valve stem is pulled upwards on an increase of
air pressure. Equation 3 is modified as
–PA + KX – PvAv = 0

Þ KX = PA + PvAv ...(4)
The reverse and direct acting spring and diaphragm actuators are shown in
Fig. 6.4 (a) and 6.4 (b) respectively.
Figure 6.5 indicates plot of diaphragm pressure and stem travel neglecting
friction and fluid forces on the valve plug for equation 3.

-1
ve
100% Cu
r e-2
u rv
Closed C

Stem es
travel f orc
plug ce
s
No for
ug
pl
ith
W
0%
Open
Diaphragm
0 3 6 9 12 15 pressure (Psi)

FIG. 6.5
 5TH PROOF 4-9-09

Final Control Elements 205

However, it may be noted that practically many non-linearities due to

variable effective diaphragm areas, valve plug forces, and springs may be
encountered. If we consider these non-linear effects a typical plot of actuator
travel versus diaphragm pressure may not be a straight line as shown in Fig. 6.5.
It is a typical curve as shown in Fig. 6.6.

100%
Ideal
Closed
characteristics

Stem Actual
travel characteristics

0%
Open
Diaphragm
0 3 15 pressure (Psi)

FIG. 6.6

Pneumatic Piston Actuator

Piston actuators comprise pneumatic cylinder and a piston rod moving in the
closed barrel. The movement of the piston is achieved by supplying compressed
air to either side of the piston. Pneumatic cylinders are of two types (a) Single
acting cylinder (b) Double acting cylinder.

Rear end Front end

Piston

Cylinder
Operating air Operating air
pressure inlet pressure inlet

FIG. 6.7 Pneumatic cylinder

A single acting cylinder performs the operating motion in one direction

only as only one operating air inlet is provided to the cylinder. The return
cylinder stroke is achieved by providing an opposing force using a spring
mechanism when compressed air is supplied at the inlet opening, which
provides the thrust motion to the piston pushing it towards the front end or rear
end depending on the operating air pressure inlet provision.
 5TH PROOF 4-9-09

206 Process Control Engineering

Since power is developed in only one direction, heavy control equipment

cannot be attached and the field applications are restricted to embossing,
stretching and riveting, etc.

Cylinder

Piston

Spring

Operating air pressure

FIG. 6.8 Single acting cylinder

The double acting cylinders are equipped with two inlets for the operating
air pressure on both front and rear ends of the cylinder. Thus, the piston motion
is a two way motion. The movement of the piston depends on the quantity of the
compressed air supplied at the front and reverse ends, and the pressure
difference makes the piston travel possible. The throttling of the control valve in
both single and double acting cylinders is achieved by the piston movement or
displacement.

Cylinder

Piston

Inlet Inlet

FIG. 6.9 Double acting cylinder

It may be noted that piston actuators must be properly designed keeping in

view the “fail-safe” option. Air lock system must be employed to force closure of
control valve on air failure. The reliability of these actuators also depend on the
leaks (air leaks) through interconnecting tubes, fittings, etc.

6.1.4 Hydraulic Actuators

The components used in hydraulic actuation is same as that pneumatic piston
actuators. It also consists of a hydraulic cylinder and a piston which provides the
throttling of control valve plug but the difference is, it uses compressed oil
 5TH PROOF 4-9-09

Final Control Elements 207

supply as against compressed air supply of pneumatic actuation. The single and
double acting cylinders discussed in Section 6.13 are also applicable for
hydraulic actuation.
However, the principle of hydraulic actuation is discussed in this section. A
hydraulic actuator converts a small force “Fx” into an amplified force called as
operating force “Fw”

Fx Fw

PH
(A1)
PH(A2) Hydraulic fluid

FIG. 6.10

PH = Hydraulic pressure (Pascals)

A1 = Forcing piston area (m2)
Fx = Applied piston force (N)
Fw = Force of working piston (N)

A2 = Working piston area (m2)

The hydraulic pressure is given by

Fx
PH = ...(5)
A1

The operating or working piston force is given by

Fw = PH ´ A2

A2
\ Fw = Fx × ...(6)
A1

The working piston force can be made very large by changing the area of
forcing piston A1 as seen by equation (6).

6.2 INTRODUCTION TO CONTROL VALVES

The most important final control element so widely used in variety of process
industries is the control valve.
It is basically a variable resistance to the fluid flow, where the resistance and
hence the flow is changed by signal output of the controller.
 5TH PROOF 4-9-09

208 Process Control Engineering

The control valve is divided into two parts

® The valve body
® The trim
The valve body comprises of a housing for mounting the actuator and other
relevant connections for attaching the valve in the fluid flow line.
The trim is enclosed within the body and consists of plug, valve seat and the
valve stem.
The actual control valve unit comes with an actuator. The pneumatic spring
and diaphragm actuator is the most widely used one compared to others because
of its low cost, its relatively high thrust at low air supply pressure and also its
availability with “fail-safe” springs Also this pneumatic actuator is available in
various designs.
The other actuator designs like piston actuators, electromechanical actuators,
etc. can also be used depending on the control valve application and numerous
other factors like longer strokes, high torque, etc.
The basic operation of a control valve can be understood from Fig. 6.11.

Valve stem

Valve plug
Þ Flow Þ Flow
Valve seat

Valve body Pipeline

FIG. 6.11

The pressure on the diaphragm of actuator moves the valve stem which in
turn moves the plug. The amount of resistance offered to the fluid flow depends
on the extent to which the plug covers the valve seat. As the pressure on the valve
stem varies over the normal pneumatic range of 3 – 15 psi, the valve stem motion
also varies from a fraction of inch to several inches, depending on the design of
the valve. Both the direct and reverse acting actuators discussed in Section 6.13,
Fig. 6.4 (a) and (b) can be used for operating the control valve. The direct acting
actuator is also called as air-to-close (or) pressure to close actuator where an
increase in signal pressure above the diaphragm causes the valve stem to move
down. The reverse acting actuator also called as air-to-open (or) pressure to open,
has the air inlet below the diaphragm. An increase in air pressure causes the valve
stem to move upwards. Based on the design of actuators, the valves are also
referred to as AIR - TO - OPEN or AIR - TO - CLOSE valves.
The valve shown in Fig. 6.11 is known as “THE SINGLE SEATED
CONTROL VALVE”. It implies the valve contains one plug with one seating
surface only.
However, to reduce the effort needed to open the valve against full pressure
drop across it, valves are constructed with double seating as shown in Fig. 6.12.
 5TH PROOF 4-9-09

Final Control Elements 209

In double seating, there will be two valve plugs and two seatings. The two
valve plugs are attached to the single valve stem.
The flow pattern through the valve is designed in such a way that pressure
drop across seat ‘A’ tends to open the plug and the pressure drop across seat ‘B’
tends to close the plug. This kind of action tends to reduce the effort needed to
open the valve and consequently less expensive actuator is needed.

Valve stem

A
Þ Flow Þ Flow
B

Flow
profile

FIG. 6.12 Double seated valve

6.2.1 Dynamics of Control Valve

The dynamics of pneumatic spring and diaphragm actuated control valve are
discussed in this section. The position of the stem or equivalently of the valve
plug at the end of the stem will determine the size of the opening for flow and
consequently the flow rate or quantity of flow of fluid. The stem position depends
on the balance of different forces acting on it i.e., upward and downward forces.

Air supply
Fd = PA

Fu = Kx Spring
1
B.dx
Fu = dt Valve stem
2

Valve plug
fi Flow fi Flow

Valve body

FIG. 6.13

The air pressure exerted on the top of diaphragm having cross-sectional area
of “A” m2 causes the diaphragm to deflect downwards. Let ‘p’ be the pressure
signal that opens or closes the valve. This force acts downwards.
Thus, downward force ‘Fd’ is expressed as Fd = P ´ A.
The force exerted by the spring attached to the valve stem is upwards and is
expressed as
 5TH PROOF 4-9-09

210 Process Control Engineering

Fu1 = Kx
where, K = spring constant
x = displacement of the stem.
The downward force ‘Fd’ is further balanced by frictional force exerted
upwards resulting from the proximity of valve stem with valve packing. This
frictional force is expressed as

dx
Fu2 = B ×
dt
where, B = damping coefficient
The third force acting upwards results from the mass of the stem and the
plug and is expressed

m d2 x
Fu3 =
g dt2

where,
M = mass of the stem/plug
g = acceleration due to gravity m2/s
From the principles of translational mechanics
Algebraic sum of downward forces = Algebraic sum of upward forces
Fd = Fu1 + Fu2 + Fu3

dx M d 2 x
PA = Kx + B + ...(7)
dt g dt 2

Equation 7 indicates that the valve stem follows second order dynamics with
pneumatic spring and diaphragm actuator.
Practically the mass of the valve stem and plug are neglected and equation (7)
is approximated to that of first order dynamics.

dx
B + Kx = PA
dt

B dx PA
+x=
K dt K

Let T = time constant element = B/K.

Kp = gain = A/K.
 5TH PROOF 4-9-09

Final Control Elements 211

dx
T× + x = Kp × P. ...(8)
dt

Thus, the stem position ‘x’ for the applied pressure ‘P’ follows first order
dynamics. The presence of time constant element indicates sluggishness of the
response for a given input. But generally most of pneumatic valves respond very
rapidly to changes in air pressure. The valve plug opening or closing is so fast
that even the first order dynamics also may be neglected.
In such a case only a constant gain term will remain which relates the output
from the controller, i.e. air pressure signal to the fluid flow through the valve.
This approximation further makes the control valve to exhibit zero order
dynamics. Equuation 6.8 further reduces to
x = KP × P ...(9)
xµ P ...(10)

6.3 CONTROL VALVE CHARACTERISTICS

The relationship between the flow through the valve and the valve stem position
or lift is called valve characteristic or control valve characteristics, since the
function of a valve is to regulate the flow of the fluid through it with respect to the
applied pressure.
The valve characteristic is assigned with an assumption that the stem
position indicates the extent of valve opening and the pressure difference is
determined by the valve alone. The different types of control valves are classified
in relevance to these characteristics which are in turn dependent on different
shapes of the valve plug and seat as shown in Fig. 6.14. As the stem plug moves
with respect to the seat, the shape of the plug determines the amount of actual
opening of the valve. The inherent valve characteristics may be classified as
(a) Quick opening
(b) Linear
(c) Equal percentage

(a) Quick opening

There are two basic valve configurations prevalent in process industries. They are
(a) Shut off valves
(b) Throttling valves
The quick opening valve is used predominantly for full ON/full OFF control
applications. A relatively small motion of the valve stem produces maximum
possible flow rate through this valve.
Such valves allow 75% of maximum flow for just 25% of stem opening and are
suitable as shut off valves.
 5TH PROOF 4-9-09

212 Process Control Engineering

(b) Linear
In this type of control valve the flow rate varies linearly with stem motion.
Let Q = Flow rate (m2/s)
Qmax = Maximum flow rate (m3/s)
X = Stem position (m)
Xmax = Maximum stem position (m)

Q X
= ...(11)
Q max X max

Equation (11) describes the operation of linear valve, for a constant pressure
drop.

Q
Let Q¢ = where, Q ¢ = fraction of maximum flow.
Q max

X
X¢ = X ¢ = fraction of maximum lift.
X max

We can write a general equation for flow through the valve for constant
pressure drop as
Q¢ = f(X¢)
Sensitivity is defined as

dQ ¢
S= ...(12)
dX ¢

dQ ¢
= K; where K =
For linear valve, equation (12) must be constant. Thus,
dX ¢
any constant, the relationship between flow and valve opening can be
represented as straight line.

(c) Equal percentage

In this type of valve a given percentage change in stem position results in
equivalent change in flow, i.e. equal percentage, at constant pressure drop across
it.
One of the special characteristics of this valve is that it does not shut off the
flow completely in its limit of stem motion. There will be some leakage of
minimum lift to prevent binding and wearing of the plug and seat surfaces.
Hence, to express the range over which an equal percentage valve will follow
the equal percentage characteristic the term rangeability is introduced.
If “Qmin” represents the minimum flow when the stem is at one limit of its travel
and let “Q max” represent maximum flow, i.e., open–valve flow rate, the
RANGEABILITY (R) may be thus defined as
 5TH PROOF 4-9-09

Final Control Elements 213

Q max
R= ...(13)
Q min

The flow rate through equal percentage valve is given in terms of ‘R’ as
Q(Actual flow rate) = Qmin × R(X/Xmax) ...(14)
The valve characteristics for the three types listed above is shown in
Fig. 6.15.

Types of Valve Plugs

Linear Equal percentage

(a) (b)

Quick opening
(c)

FIG. 6.14 Types of valve plugs

(1)
1 ® Quick opening
Flow rate through valve (Q)

2 ® Linear
3 ® Equal percentage
(2)

(3)

Stem Position (X)

FIG. 6.15 Control valve characteristic

In practical applications the quick opening valve may be used as shut off
valve and for throttling purposes equal percentage is best suited.
Linear valve, whose sensitivity is constant is suitable for both complete shut
off and throttle applications.
 5TH PROOF 4-9-09

214 Process Control Engineering

6.4 TYPES OF CONTROL VALVES

Most of the industrial control valves are either shut off valves or throttling valves
depending on their characteristics discussed in Section 6.3. However the types of
control valves may be categorised depending on the material, fluid flow rate,
pressure drop across them and shut off capability.
Majority of control valves are classified under two main categories namely
(a) Sliding Stem Valves
(b) Rotary Valves.

6.4.1 Sliding Stem Control Valves

The most earliest and popular sliding stem valve is the Globe Valve. Characterised
by linear plug movements and actuated by spring/diaphragm and operators. The
main advantages of traditional globe valve design are
• Actuated by spring/diaphragm actuator
• Availability of wide range of valve characteristics.
• Low cavitations and noise.
• Special designs for corrosive abrasive fluids.
• Applicable for high temperature and pressure conditions.
• Relatively small amounts of dead band and hysteresis effects.
• Linear relationship between control signal and valve stem movement.
• Reduced plug mass which leads to less vibrations.
The flow sheet symbols of globe valves are shown in Fig. 6.16.

FC FO
Three way Fail Closed Fail Open
Angle Value

FIG. 6.16

Globe valves are available in variety of configurations as listed below.

• Single ported with characterized plug
• Single ported, cage-guided
• Single ported, split body
 5TH PROOF 4-9-09

Final Control Elements 215

• Double ported, top bottom or skirt guided plug

• Eccentric disc, rotary globe
• Fail open angle valve
• Three way or Y-Type valve.
The sizes of globe valves generally varies between 0.5 to 14 inches [20 to 356
mm].
The maximum size for single ported split body valve is 6 inches [152 mm] and
double ported valve is 16 inches [406 mm].
The fail open angle valve has been made in size a maximum up to 42 inches [1.05 m].
The body material is generally available in iron, forged steel, type 302 to 316
stainless steel, bronze, alloy 20, monel, hastelloy C and sometimes titanium.
The single ported with characterized plug can withstand temperatures in the
range of – 320°F to 1000°F [– 196°C to 538°C]. The design temperatures for other
configurations are as shown below.
For single ported, cage guided ® upto 400°C
For single ported, split body ® 268°C ® 500°C. Special valve configurations
have been designed for operation at over 1200°F (649°C).
The double seated globe valve is an older design and has been replaced with
single seated valve.
In Figure 6.17 top and bottom guided invertible double seated globe valve is
shown. The term “invertible” is used because it can be converted from air-to-open
to air-to-close and vice versa, configurations. This is done by disassembling the
bottom closure flange, bonnet and stem, and by inverting the entire assembly
and reinstalling the same again.

FIG. 6.17
 5TH PROOF 4-9-09

216 Process Control Engineering

Single Seated Globe Valve

Single seated valve configurations is the most widely used pattern for globe
valves.
The advantages of single seated valves over double seated valves are
• Good flow shut off capability
• Less vibrations due to reduced plug mass.
• Have special purpose trims, easy to maintain.
For valve shut off or open, the single seated valve plugs are guided in any one
of the three ways. They are
(a) stem guided
(b) Top-entry, Top-guided
(c) Top-and-bottom guided (Invertible)
However the most commonly used and popular configuration is stem guided
and/or top guided types.
Fig. 6.18 shows typical stem guided globe valve.
Like its double-seated counter part, the top-and-bottom guided design has
certain limitations.

Inline Globe

FIG. 6.18
 5TH PROOF 4-9-09

Final Control Elements 217

FIG. 6.19 Top guided Single Seated Globe Valve

FIG. 6.20 Top & bottom guided invertible Single Seated Globe Valve

The top entry, top guided single seated globe valve has only one body
opening for the bonnet and valve plug mass is reduced, which thereby increases
the trim throttling precisely when compared to the top and bottom guided valve.
When dealing with fluids containing solids, gummy or highly viscous fluids,
and highly corrosive fluids stem guided valve is always a better choice than top-
guided valve because it minimises cavities and ensures trouble free operation.

Cage Valves
One of the most popular variant of single seated globe valve is the cage valve. The
valve plug design and top entry bonnet makes the job of trouble shooting work
very easy, especially when changing the trim or plug. A wide variety of trim or
valve plug designs such as anti-cavitation and reduced-noise trims can be
conveniently installed.
 5TH PROOF 4-9-09

218 Process Control Engineering

There are two basic design configuration available for cage valves.
(a) cage valve with clamped-in seat ring and characterized plug shown in Fig.
6.21
(b) cage valve with unbalanced plug and flow characterized ports shown in
Fig. 6.22.
Very Heavy Guiding:
Two widely-spaced
guides on heavy
stem. plug does not
guide in retainer.
Interchange Ability:
class 600 body used
for class 150, 300 Clamped-in seat ring:
and 600 Ratings valve may be dis-
assembled quickly,
easily by removing
four bonnet bolts

FIG. 6.21

FIG. 6.22

The cage valve design shown in Fig. 6.21 uses cage solely to clamp the seat
ring into the valve body.
 5TH PROOF 4-9-09

Final Control Elements 219

This configuration is usually stem or top guided design and the valve plug
does not contact the cage.
In Fig. 6.22, the cage is used to guide the plug as well as clamp the seat ring
into the body. The cage openings are shaped in such a way so as to provide
desired flow characteristics and the valve plug acts as a curtain to expose the
ports.

Angle Valves
These valves are generally used in piping arrangements which require high
pressure drop services and applications such as in controlling molten metals,
polymer materials, cryogenic fluids and liquid slurries.
The Y-shaped angle valve is a popular design for the above mentioned
applications and is shown in Fig. 6.23.
The valve can be installed in any of the vertical, horizontal or angled piping.
They are usually fitted along with a vacuum jacket, which provides the required
thermal insulation when used in dealing with flow of cryogenic fluids such as
liquid hydrogen. The vacuum jacket is designed with metal bellows to allow for
thermal changes or mechanical tolerances and also has a provision for welding to
the similar jacket around the adjacent piping.

FIG. 6.23 Y-Valve with vacuum jacket

The cooling rate depends on valve design, thickness etc. Since these valves
are used in high pressure applications they are highly susceptible to cavitation
problems (discussed later) and erosion problems.
 5TH PROOF 4-9-09

220 Process Control Engineering

6.4.2 Rotary Valves

There are two basic designs of rotary valves namely
(a) Ball Valves
(b) Butterfly Valves
The advantages of rotary ball or butterfly valves over traditional globe valves
are
(i) Lower cost and weight
(ii) Higher flow capacity
(iii) Tight shut off
(iv) Lower leakages

Ball Valves
A typical ball valve consists of a spherical valve through the valve body. They are
available in three types
(a) conventional or quarter turn pierced ball type
(b) characterized type
(c) Cage type.

Conventional Ball Valves

They are also known as “quarter turn pierced ball valves”. Quarter turn implies
90° rotation of the ball to fully uncover or cover an opening in the valve body, i.e.
to change the valve position from fully open to fully close or vice versa.
The spherical ball plug is pierced in such a way that in one position there is
precise throttling and with quarter turn i.e. 90° rotation there is complete shut
off.
Actuators used for conventional ball valves are similar to those studied in the
previous sections, but with rotating motion.
The flow sheet symbols of conventional ball valves are shown in Fig. 6.24.

Standard Ball Three Way Full Ported

Valve Ball Valve Valve

FIG. 6.24
 5TH PROOF 4-9-09

Final Control Elements 221

The structural details of top entry quarter turn pierced ball valve is shown in
Fig. 6.25.
Stem
Body

Ball
Seal

Top-entry pierced ball valve.

Tea Ports

Angle Ports
Porting arrangements of multiport ball valves.

FIG. 6.25 Top entry pierced conventional ball valve

The size of conventional ball valve varies between ½ inch to 42 inches (12.5
mm to 1.06m) [ANSI class 150 standards]. The design temperature varies
between – 150°C to 315°C, with special designs available up to 1020°C also.
They are usually made of cast or bar stock brass or bronze, carbon steel,
stainless steel, aluminium, monel etc for the body.
For balls, forged naval bronze, carbon steel, plastics, glass, ceramics etc. The
seats are usually made up of teflon Delrin, neoprene etc and related synthetic
materials.
 5TH PROOF 4-9-09

222 Process Control Engineering

Characterized Ball Valves

In these types of ball valves, the ball is modified as V-notched, U-notched or a
parabola ball. Only a portion of the ball is used with its edge contoured or
shaped as V-notched or U-notched designs. A typical characterized ball valve is
shown in Fig. 6.26. The controlling edge of the ball can be notched or contoured
to produce the desired flow characteristics, which are available as designs
mentioned above. The flow characteristics are thus dependent on these
curvatures. The fully open, fully closed and throttling positions for the
characterized ball valves are shown in Fig. 6.27. The segmented ball is available
from 1 inches to 24 inches size with standing a temperature level between –46°C
to 149°C.
The flow sheet symbol of characterized ball valve is shown below
Characterised
Characterised Ball
Ball
Stub
Shaft
Stub
Bearings
Shaft

Bearings
Seal
Rings

Body

FIG. 6.26 Parts of characterized ball valve

Open Throttling Closed

FIG. 6.27 Positions of characterized ball valve

Cage Positioned Ball Valve

In cage type ball valves, the ball is positioned by a cage. The constructional
details of a cage ball valve are;
 5TH PROOF 4-9-09

Final Control Elements 223

It consists of a venturi-ported body with two seat rings, a ball that acts as a
plug (valve plug) or trim. A stem is used for positioning the cage which in turn
position the ball to obtain fully closed or open and/or throttling flow
characteristics. The structural details and various positions enlisted above are
shown in Fig. 6.28.
The cage rolls the ball out of the seat as it is tilted by the stem supporting it and
positions the ball firmly during throttling and lifts it out completely during full opening
position from the flow stream. The cage is designed with four inclined surfaces. The two
surfaces next to the downstream seat lifts the ball out of the seat and roll it over the top
edge of the seat ring thus opening the valve.
During fully open position, the ball rolls down to the centre of the cage to
rest on all four inclined surfaces. The Bernoulli’s effect of the flowing stream
holds it cradled in this position during the rest of the operation. They are
available in sizes ranging from ¼ inches to 14 inches for design pressures limited
to 17 mpa. They can withstand temperatures in the range of – 254°C to 980°C.
The ball is usually made up of stainless steel.

Open Thrcttling Closed

FIG. 6.28 Cage valves (Ball Valve)

The flow sheet symbol cage ball valve is shown below.

Butterfly Valves
The butterfly valve is perhaps the most popular and oldest designs available in
process industries. The earlier designs were particularly used in the low pressure
water works applications. However the modern butterfly valves are used in a wide
variety of fluid applications, those with high pressure drop, tight shut off and
also involving corrosive & erosive characteristics.
The butterfly valve is defined as “A damper or vane or throttle valve in a fluid
pipe line consisting of a disc turning on a diametrical axis”.
The operation of butterfly valve therefore is based on the above definition of
it where it involves only rotating the vane, disc, a flapper or a louver by means of
a shaft to which it is fastened. The rotation is accomplished using Pneumatic,
electric or hydraulic motor drives attached to the shaft by various methods. The
 5TH PROOF 4-9-09

224 Process Control Engineering

fully closed to fully open position is achieved when the disc or vane rotates or
moves through 90° angle. The free area developed by the disc as it moves towards
the full open position provides the throttling operation.
Based on the construction and flow characteristics the butterfly valves are
available in two designs.
(a) Swing through butterfly valves
(b) Shut-off butterfly valves (or) High performance butterfly valves [HPBV].
The vane positions are shown in Fig. 6.29 for fully open, closed and
throttling.

Flow Flow Flow

Fully closed (Damper Throttling Fully open (Damper

perpendicular to fluid parallel to fluid flow)
flow)

FIG. 6.29

Swing-in Through Butterfly Valves

The conventional details of swing through butterfly valves are shown in Fig. 6.30.
The symmetrical disc and the shaft are designed to have certain clearance
between disc and body. The solid ring type body is mounted between pipe
flanges.
Discs are cast in one piece as shown in Figure. The thickness of the disc and
hub along the diameter of the shaft is a function of the maximum pressure drop
and torque required. Swing through butterfly valve design is limited to maximum
rotation of 70° for fully open for standard flow patterns and 60° for heavy flow
patterns with larger diameter shafts.

High Performance Butterfly Valves

The most notable design advancement in the category of tight shut off butterfly
valve abbreviated as HPBV.
Some of the significant features of HPBV are
(a) Reduced operating torque and enhanced tight shut off.
(b) Ability to operate with relatively high pressure drops with excellent
throttling capabilities.
 5TH PROOF 4-9-09

Final Control Elements 225

Shaft Sold Ring

Type Body
Lubrication Packing
Fitting Follower

Packing

Bearings

Disc

FIG. 6.30 Swing-in through butterfly valve

6.5 CONTROL VALVE SIZING AND SELECTION

The flow capacity of a valve is based on an equation relating flow to the square
root of the pressure drop across the valve. In order to specify the size of the valve
in terms of its capacity to provide flow when fully open, the following equation is
used.

DP
q =CV ...(15)
G
CV = factor associated with capacity of valve
(control valve coefficient)
q =design flow rate (gpm)
G =Specific gravity relative to water.
DP =Pressure drop across the valve (Psi)
The size of the control valve depends upon the factor CV i.e. valve coefficient
value. The valve coefficient ‘CV’ is defined as the flow (gpm) of a fluid of unit
specific gravity through a fully open valve, across which a pressure differential of
1.0 psi exists. In equation 15 by putting q = DP Þ G = 1; CV = 1 i.e. the
definition just stated above.
The value of CV varies with the design of the value i.e. shape, size, reynold’s
number etc.
Typical values of ‘CV’ for different valve are shown in the table.
 5TH PROOF 4-9-09

226 Process Control Engineering

Table 6.2
Valve Size (Inches) CV

1
1. 0.3
4

1
2. 3
2
3. 1 14

1
4. 1 35
2
5. 2 55
6. 3 08
7. 4 174
8. 6 400
9. 8 725

6.5.1 Control Valve Selection

The selection of control valve depends upon the valve characteristics, discussed
in Section 6.3.
1. Equal Percentage
In this type of valve equal increments of valve travel produce an equal percentage
of flow through it. It is the most commonly used valve characteristics in valve
control.
Applications
• used in processes where large changes in pressure drop are expected.
• used in process applications where small percentage of the total pressure
drop is permitted by the valve.
• used in temperature and pressure control loops.
• Best suited control for Globe Valves and butterfly valves.
2. Linear
In this type the valve travel and valve stroke are in direct proportion.
Applications
• used in level (liquid) and flow control loops.
• used in systems where the pressure drop across the valve is expected to
remain fairly constant.
• Best suited control for globe valves Ball valves and butterfly valves.
3. Quick opening
In this type of valve a large increase in flow can be observed with small changes in
valve stroke.
 5TH PROOF 4-9-09

Final Control Elements 227

Applications
• used for processes where instant large flow is required.
• used for frequent on/off service.
• Best suited control for globe and ball valves.
Now, let us see a quick comparison between different types of valves.

Globe Valves
The recommended uses are for
1. Throttle service/Flow regulation.
2. Frequent operation.
It can be used for fluids that are liquids, vapours, gases corrosive substances,
and slurries.
Advantages
1. Efficient throttling
2. Accurate flow control
3. Available in multiple ports
Disadvantages
1. High pressure drop
2. Expensive compared to other valves.

Ball Valves
The recommended uses are for Fully open or fully closed, limited throttling
applications.
They can be used for high temperature fluids, most liquids and slurries.
Advantages
1. Low cost
2. High capacity
3. Relatively lower leakages & maintenance.
4. Tight sealing
Disadvantages
1. Poor throttling characteristics
2. Prone to cavitation

Butterfly Valves
The recommended uses are for fully open or fully closed or throttling services,
minimal fluid trapping in line and frequent operation.
They can be used for fluids such as liquids, gases, slurries, and liquids with
suspended solids.
Advantages
1. Low cost and maintenance
2. Low pressure drop
3. Good flow control
 5TH PROOF 4-9-09

228 Process Control Engineering

Disadvantages
1. High pressure required for valve control
2. Prone to cavitation at lower flows

6.6 CAVITATION AND FLASHING PROBLEM IN

CONTROL VALVES
The phenomena of cavitation is related to Bernoulli’s theorem, which describes
the pressure profile as fluid passes through a restriction. As the fluid passes
through a restriction the pressure head is converted into velocity head. That
point where the static pressure is minimum and velocity is maximum is called
vena contracta. At this Vena contracta point when the static pressure head drops
below liquid vapour pressure at that temperature, then vapour bubbles will form
at the down stream of the restriction. As the static pressure recovers to a point
greater than the vapour pressure, the vapour bubbles collapse back into their
liquid phase. The collapse of bubbles produce high energy implosions, known as
cavitation.
These implosions generate noise, fluid shock cells and jets that impinge
upon the trim metal parts causing considerable damage to the valve body parts.
Damage due to cavitation can be overcome using certain special valve trim
designs or to limit the pressure drop by installing control valves in series to
distribute the pressure drop and also to reduce the vena contracta pressure in
each valve.

QUESTION SET
1. Discuss about the spring and diaphragm pneumatic actuator in reverse
and direct acting modes?
2. Explain the principle of hydraulic actuator?
3. Obtain the dynamics of the control valves and the related
approximations?
4. Discuss about the control valve characteristics?
5. Draw the types of valve plugs for different control valves characteristics?
6. Discuss about the different types of sliding stem valves?
7. What are the basic designs available in rotary valves? List their advantages
over conventional globe valves?
8. What is meant by “quarter turn” in ball valve design? Write about the ball
valves generally used in industries?
9. Explain with neat diagrams the vane positions for butterfly valves during
fully open, closed and throttling?
10. Write about
• control valve selection
• sizing of control valves
• cavitation in control valves.
 +0)26-4
%
Process Control System
Configurations

INTRODUCTION
The main objective of a process control system is to monitor and control the
controlled variable in a process and keep it at set point, the desired value, in spite
of disturbances occurring within the process or through any external agencies.
The are various configurations in setting up of process control systems to regulate
and control the dynamic variable in a process. This chapter outlines the
structures and schemes of developing different types of process control loop
configurations. However, the emphasis is more on structural representation than
on mathematical designs and considerations. Also the selection of a particular
type of configuration depends on characteristics and dynamics of the process,
and is usually made by process design engineers.

7.1 FEED FORWARD AND FEED BACK CONTROL

One of the basic and simplest control loop concept as outlined in the previous
chapters, whose mathematical analysis with different modes of controllers was
discussed is the FEEDBACK CONTROL LOOP.
In this feedback control configuration the changes in the controlled variable
are measured through feedback and compared with input or set point and a
corrective action is provided to eliminate the error if any to ensure the control
objective is retained. This implies that the feedback control cannot anticipate
and prevent errors. It can only initiate the corrective action after the error has
occurred.
Figure 7.1 illustrates the structure of feedback control scheme.
 5TH PROOF 4-9-09

230 Process Control Engineering

Manipulated Variable
Controller
Set Point
Load
Variables .
or . Process
. Controlled
Disturbances Variable

FIG. 7.1

In Fig. 7.1, the deviations of controlled variable from set point value due to
variations in load variable or disturbances are first measured through feedback
elements, compared with input or set point at the controller and the extent of
deviation or error is eliminated by converting it into changes in the manipulated
variable and sending back to the process to restore the balance. Consider the
liquid level control system employing feedback control, shown in Fig. 7.2.

Set Point

Inlet Controller

LT
h

Outlet

FIG. 7.2

Whenever, the inlet flow changes inducing error in the plant or process, the
liquid level is brought back to the set point value by the controller after evaluating
upon the magnitude of error.
In contrast to this control loop configuration, FEED FORWARD provides a
more direct solution to control than finding the correct value of the manipulated
variable by trial and error, as observed in the feed back control.
In feed forward control system, the major components of the load variables
or disturbances are entered into the dynamic model to calculate the value of the
manipulated variable required to maintain the controlled variable at the set point
value.
Fig. 7.3 illustrates the structural scheme of feed forward control loop.
Fig. 7.3 illustrates how information flows forward from the load to the
manipulated variable input of the process. When a change in load is sensed, the
manipulated variable is automatically adjusted to the correct value at a rate that
keeps the process continually in balance.
In feed forward control loop the effect of disturbances or load variables are
not seen in the process since the control action starts immediately after a change
in the disturbance has been detected.
 5TH PROOF 4-9-09

Process Control System Configurations 231

Set Point

Controller Manipulated
Variable

Load
Variables
or . Process
. Controlled
Disturbances .
Variable

FIG. 7.3

Let us now consider the same liquid level control system employing feed for-
ward control. The disturbance is the inlet flow rate changes and the manipulated
variable is the outlet flow rate.
The changes in the inlet flow rate are measured, before they affect the
controlled variable (level) and the controller adjusts the outlet flow rate
accordingly to maintain the control objective. As seen in Fig. 7.4 the changes in
the load variable or disturbances are never felt in the process.

P/I
Controller Set Point
Converter
D.P. Cell

Inlet
flow (orifice)
rate

Outlet

FIG. 7.4

7.1.1 Comparison of Feed Forward and

Feedback Control
Feed Forward Feedback

(1) It acts before the effect of disturbance (1) It acts after the disturbance affecting
has been felt in the process. the control objective of the process.
(2) It requires identification of all possible (2) It does not require any identification
disturbances and their measurement. and measurement of load variables or
disturbances.
(3) It is more stable and does not introduce (3) It is relatively less stable and induces
instability when evaluating the closed instability in closed loop response.
loop response.
 5TH PROOF 4-9-09

232 Process Control Engineering

(4) It is a better control configuration for (4) It is unsatisfactory for slow processes
multi capacity or slow processes. as there could be a scope of large
inherent lags.
(5) Feed forward control is a suitable option (5) It is unsuitable for processes having large
for processes having significant dead or significant dead time.
time effects.
(6) Any changes in the process parameters (6) It is insensitive to process parameter
excluding disturbances and manipulated changes.
variables cannot be compensated by
feed forward controller because their
effect cannot be detected.
(7) It requires good knowledge of the (7) It is insensitive to modelling errors.
process model.
(8) It is suitable for processes where load (8) It is not suitable for such processes
or disturbances to the process change as it cannot cope up with rapidly changing
rapidly with time. quantities.
(9) The use of feedback in a feed (9) With feed forward, the feedback cont-
forward system does not detract from roller must only change its output by an
the performance improvement which amount equal to what the feed forward
was gained by feed forward control. system fails to correct.
(10) The feed forward system is more (10) They are comparatively less costly and
costly and requires more engineering require less complicated design methodo-
effort. logies.

7.1.2 Feed Forward and Feedback Control Loops for

Stirred Tank Heater
Consider a stirred tank heater whose control objective is to maintain the
temperature of the water in the tank at desired value. The water in the tank is
heated to the desired temperature value through continous supply of steam. The
load variables or disturbances are inlet flow rate, outlet flow rate, etc. The steam
supply rate is the manipulating variable. Let us see, the feedback and feed
forward control loop configurations to achieve the required control objective.
In Fig. 7.5, whenever there is change in temperature of the water due to
changes in inlet or outlet let flow rates (disturbances), the temperature transmitter
(TT) measures this change and sends the signal to the controller which after
proper evaluation i.e., comparing the measured value with set point opens or
closes the steam flow valve to meet the control objective. This is feedback action.
The effect of disturbance is felt in the process after which corrective action is
provided by the control loop.
Let us see a different control arrangement i.e., feed forward control as shown
in Fig. 7.6. Here the disturbing factors are measured to achieve the control
objective. We observe that feed forward control does not wait until the effect of
disturbances has been felt by the system, but acts appropriately by anticipating
the effects of disturbing elements in the process or system.
In Fig. 7.6, the changes in the inlet flow rate are measured and the feed
forward controller opens or closes the control valve of steam line to keep the
temperature of the water at desired set point value.
 5TH PROOF 4-9-09

Process Control System Configurations 233

Inlet (Stirrer)

TT

Set Point Outlet

Controller

(Control valve)

Steam flow rate

FIG. 7.5 Feedback control system

Orifice Stirrer
Inlet
D.P.
flow
Cell
rate

P/I
Converter
Outlet
Set point Feed forward
controller

(Control valve)

Steam flow

FIG. 7.6 Feed forward control system

Let us now see the generalized block diagram for feedback-feed forward
control loop configuration. This representation is helpful in mathematical
analysis of these control loop configurations.
Fig. 7.7 is the generalised block diagram showing both control actions.

7.2 CASCADE CONTROL CONFIGURATION

In the preceding feedback or feed forward control configurations, measurement
and control of only one variable is effected. In certain processes which require
more than one measurement and one manipulation or more manipulations to
achieve the control objective, more control loops may need to be formulated.
Such a configuration is the cascade control loop.
In cascade control loop configuration we have one manipulated variable and
more than one measurement.
 5TH PROOF 4-9-09

234 Process Control Engineering

R2(s) D(s)
– Hm2(s)

Gc2(s)

R1(s) + +
Gc1(s) + Gp(s) C(s)
–

Hm1(s)

R 1(s) = Set point for feedback control

R 2(s) = Set point for feed forward control
Gc 1 (s) = Feedback controller
Gc 2 (s) = Feed forward controller
G p (s) = Process/Plant
Hm 1(s) = Feedback sensor
Hm 2(s) = Feed forward sensor
D(s) = Disturbances

FIG. 7.7

Consider the liquid level control system employing feedback control action
shown in Fig. 7.2. The manipulating variable is the outlet flow rate directly
affecting the level of the liquid in the tank, which is the controlled variable in the
process.
This inherent interaction between the outlet flow rate and level of the liquid
in the tank is exploited to provide better overall control through cascade
configuration. A cascade control system consists of two control loops, namely
primary and secondary control loops respectively.
The loop that measures the controlled variable i.e., level of the liquid in the
tank (h) is the dominant or primary or master control loop and has its setpoint
supplied by the operator.
The loop that measures the outlet flow rate with the output of the primary or
master controller as its set point is called secondary control loop or slave loop.
Together with these loops exploiting the inherent interaction occurring between
the two control systems is known as cascade control loop.
The major advantage of this configuration is if at all any disturbances
occurring within the secondary loop are corrected by secondary or slave
controller before they can effect the value of the primary controlled output.
 5TH PROOF 4-9-09

Process Control System Configurations 235

Unlike a feedback or feed forward control system, here any disturbances in

the manipulating variable are corrected which results in better overall control of
the actual dynamic or controlled variable.
In a simple feedback loop, there is no corrective action for disturbances felt
in the outlet flow rate but when a control loop is formulated at outlet pipe line the
deviations in the outlet flow rate are never felt in the process.
Fig. 7.8 illustrates the cascade control loop for the liquid level control system
where the controlled variable is level of the liquid in the tank and manipulating
variable is the outlet flow rate. In this configuration there can never be any
possibility of deviations in the manipulating variable affecting the stability of the
process. The dynamics of the secondary loop monitoring, the outlet flow rate are
much faster than those of the primary loop. The lags associated with secondary
loop will be consequently less than that of primary loop. Thus, the disturbances
occurring the secondary loop can be dealt with more effectively without
endangering the stability of the system.

Inlet
Primary (or) Set Point
Master
Controller
LT
Set Point

Secondary (or)
P/I
Slave
Converter
Controller
h
(level) Control
D.P. Valve
Cell

Orifice Outlet
flow rate

FIG. 7.8

Let us now draw a generalized block diagram representing the cascade

control loop. The process is divided into two parts—process – 1 or primary
process has its output the variable we are interested to control i.e., level of the
liquid in the tank. Process – 2 or secondary process has an output that we are not
interested in controlling but which affects the output we want to control i.e.
output flow rate.
Fig. 7.9 shows the general block diagram of cascade control system.
Process-1 comprises the tank, inlet and the liquid in the tank with output as
the controlled variable i.e. level ‘h’.
 Primary Loop

Disturbance Disturbance
236 Process Control Engineering

Primary or Secondary + +
Process Process
Master or Slave + +
Set –2 –1 Output
– Controller – Controller
Point

Measurement
5TH PROOF 4-9-09

Secondary Loop

Measurement

FIG. 7.9 Cascade control system

 5TH PROOF 4-9-09

Process Control System Configurations 237

Process-2 is the outlet pipe line with outlet flow rate as the output i.e.
manipulating variable. This cascade configuration ensures that the stability of the
overall system is never affected due to disturbances in process-1 or process-2
thereby providing better overall control.

7.2.1 Cascade Control for Stirred Tank Heater

Consider the stirred tank heater shown in Fig. 7.5. The cascade control is
obtained by considering two control oops, one for the main process which
includes the tank, the liquid in the tank, inlet and outlet sections wherein the
main controlled variable i.e., temperature of the liquid is measured and the other
control loop across the manipulated variable which is the steam flow, whose
controller will have a set point determined by the master controller of the main
loop.
Thus, process-1 output is the temperature of the liquid which is the actual
controlled variable to be monitored and process-2 output will be steam flow rate.
Any deviations in the manipulating variable are never felt in the actual process or
process-1 because of the cascading effect derived from the interaction between
process-2.
In Fig. 7.10 the primary loop measures the temperature of the liquid in the
stirred tank and the secondary loop measures the steam flow rate.

Inlet Stirrer

Set Point

Primary (or)
Master Controller TT
Set Point
Secondary (or)
Slave Controller
FT
Steam
flow rate (Steam flow line)
Outlet

FIG. 7.10

TT constitutes temperature transmitter which is temperature sensor and FT

is the flow transmitter or flow sensor for measuring flow rate of steam.
The following points summarize the effectiveness of cascade control systems
used in process industries.
• An intermediate process variable that responds both to the manipulated
variable and to some disturbances can be used to achieve more effective
control over the primary process variable.
 5TH PROOF 4-9-09

238 Process Control Engineering

• Cascade loops consist of two or more controllers in series, but have only a
single, independently adjustable set point by the operator to the master or
primary controller.
• The use of secondary loop and slave controller whose set point is
determined from the output of master controller is to act as the first line of
defense against disturbances, preventing these upsets from entering the
primary process.
• The secondary loop must have faster dynamics than the primary loop for
the cascading effect to yield better results.
• Some rules of thumb suggest that slave’s time constant should be under 1/
4th to 1/10th that of the master loop and the slave’s period of oscillations
should be below 1/2 to 1/3 that of the master control loop. Thus, an even
distribution of time constants between master and slave loops is obtained.
The even distribution of time constants between the two loops eliminates the
problems related to instability and saturation in the overall cascade control
configuration.

7.3 RATIO CONTROL CONFIGURATION

Ratio control systems maintain a relationship between two variables to monitor
and regulate/control a third variable. In fact, they represent an elementary form
of feed forward systems.
The major application of ratio control is in chemical industries, used
primarily for blending ingredients into a product or as feed controls to a
chemical reactor.
Ratio control is applied almost exclusively to flows.
Consider maintaining a certain ratio ‘R’ of ingredient ‘X’ to ingredient ‘Y’.

X
Ratio (R) =
Y
The most common method manipulates a flow loop whose set point is
calculated as
X = YR
The set point for the flow controller is generated by an adjustable gain device
known as the ratio station.
A common example is when the ratio of two reactants must be controlled. A
ratio control system should maintain a constant ratio of controlled flow i.e., ‘X’ to
wild flow i.e. ‘Y’.
Consider the Figure 7.11 where reactant ‘Y’ is measured but not regulated
and reactant 'X' is both measured and regulated to provide specified constant
ratio of
 5TH PROOF 4-9-09

Process Control System Configurations 239

X
R= .
Y
The flow rate of reactant ‘Y’ is measured and added, with relevant scaling, to
the measurement of flow rate ‘X’. The controller reacts to the resulting input
signal by adjustment of the control valve in the reactant ‘X’ input line.

Reactant 'Y'

Flow sensor FT
for 'Y' Ratio X
station (or)
¸
Controller or
Set point +

FT

Reactant 'X'

Flow sensor Control

for 'X' valve

FIG. 7.11

Another important consideration in ratio control systems is that of ratio

stations.
The ratio station in figure normally is designed to operate in the gain range
of 0.3 to 3.0. The flow signal ‘Y’, in percentage of scale, is multiplied by the gain
setting to produce a set point for the controller, which has the measured value of
flow signal ‘X’ as one of its input. The true flow ratio must take into account the
scales of the two flow meters. The setting of the ratio station ‘R’ is related to the
true flow ratio by
Flow Y scale
R = (True flow ratio)
Flow X scale
When head flow signals are used, ‘R’ should be a square root scale. When
using a divider as the ratio computing device, with input as linear flow signals,
the scale factor for the divider should be 1/2. This places a ratio of 1.0 or unity at
mid scale.
1 B
R=
2 A
If flow squared signals are used the divider should have a scale factor of 1/3 to
provide a full ratio range. This arrangement places a unity ratio at 0.58 valve on
the square root scale.

1 B2
R2 =
3 A2
 5TH PROOF 4-9-09

240 Process Control Engineering

For example, let the true flow ratio of the additive to gasoline be 2.0 cc/
gallon.
If the additive flow scale is 0 to 1200 cc/min and gasoline flow scale is 0 to
500 gal/min then, the setting of ratio ‘R’ would be

500 gal / min

R = 2.0 cc/gal ´
1200 cc / min

R = 0.833 (unit less)

When a divider is used to calculate flow ratio, then this ratio may be set
remotely as well. For remote setting with linear flow meters the general value of
scaling factor for multiplier (Here multiplier is used as ratio station) is 2.0
X = 2 RY
This setting will ensure a ratio of unity to appear at mid scale of the ratio
input, i.e. when X = Y, R = 50% of the scale.
If squared flow signals are used, a scaling factor of 3.0 provides a ratio range
0 to 1.73.
X2 = 3R2Y2
The reason for using a multiplier in remote setting of ratios is because in
such cases the need is for precision and not rangeability.

7.4 SPLIT RANGE CONTROL CONFIGURATION

The split range control scheme is employed whenever there is measurement and
control of one controlled variable with more than one manipulated variables is to
be done.
Since there is only one controlled output there will be only one control
signal, split into several parts each affecting one of the available manipulated
variables.
Such an action where we can control a single process output by coordinating
the actions of several manipulated variables all of which have some effect on the
controlled output is known as SPLIT RANGE CONTROL CONFIGURATION.
Consider a pressure vessel where the pressure inside the vessel must be
maintained at the desired set point value. It has both inlet and outlet sections
with valves at both sides where in the gas enters through inlet pipeline and exits
through outlet line to meet demands of other processing units external to the
pressure vessel.
To meet the control objective, the two control valves at the inlet and outlet
sections are coordinated as a function of the controllers output signal. Here we
are choosing both gas inlet and outlet flow rates as the manipulating variables to
keep the pressure in the vessel which is the controlled variable at a desired value.
Let us see the split range operation: For various reasons let us assume that the
pressure inside the vessel has increased. This results in an increase of the
 5TH PROOF 4-9-09

Process Control System Configurations 241

Set point

Controller

Pressure vessel

Gas in V1 Pressure
vessel V2 Gas out

FIG. 7.12

controller’s output, which is split into two parts affecting both inlet and outlet
valves simultaneously as shown in Fig. 7.12. Suppose the controller output
increases from 6 psi to 9 psi, valve V2 continues to open while the valve V1
remains completely open, thus resulting in reduction of pressure in the vessel.
For very large increase in the pressure vessel, say if the controller output
increases beyond 9 psi the control valve ‘V2’ is completely opened while valve ‘V1’
starts closing, till the pressure in the vessel reaches the desired value. This
operation of split range control can be clearly understood from Fig. 7.13.

Valve stem
position

Open

V2 V1

Closed Psi
3 6 9 15 Controller output

FIG. 7.13

TABLE SHOWING VALVE POSITIONS

Controller’s output signal (psi) Valve ‘V1’ stem position Valve ‘V2’ stem position
3 psi Open Closed
6 psi Open Partially open
9 psi Open Open
12 psi Partially closed Open
15 psi Closed Open
 5TH PROOF 4-9-09

242 Process Control Engineering

7.5 OTHER TYPES OF CONTROL

CONFIGURATIONS

Selective Control Configuration

In any process control system for each controlled variable there must be at least
one manipulated variable. However, in processes that involve one manipulated
variable and several controlled variables, the system must decide how to share the
manipulated variables. This action is decided and performed by selective control
systems which transfer the control action from one controlled output to another
according to the need.
Switching between controlled variables can be easily and smoothly
accomplished using selective devices called signal selectors. Selective control is
actually the name given to the application of signal selectors in a control strategy.
In many applications selective control is a form of multivariable control where
the selectors facilitate the online modifications of control strategies as a function
of changing operating conditions without disturbing the process.
Selective control strategy enables guarding against exceeding equipment or
operating constraints called over rides.
Over rides represent an application of selective control in which one
controller can take command of a manipulative variable of another controller
when the process tends to exceed its limits or constraints.
Over ride control ensures that interlock tripping conditions are not reached
to protect the system in case any of the equipments exceed their constraints or
limit.
In Fig. 7.14, hard constraint refers to the level where interlock trips. Over
rides come into play at some point before the interlocks and therefore are
referred to as soft constraints.

Interlock trip point

Hard constraint

Soft constraint
Over ride set point

Variable Normal
operating range

time (t)

FIG. 7.14

Consider an example of over ride control to monitor steam pressure in a

boiler and at the same time keep the water level in the boiler drum at half of its
 5TH PROOF 4-9-09

Process Control System Configurations 243

rated capacity. The normal control loop is the pressure control loop on the steam
discharge line. The over ride control loop is formulated at the level side of the
boiler drum and the outputs of normal controller and over ride controller are fed
to low signal selector (LSS).
Whenever the liquid level falls below the allowable limit, the LSS switches the
control action from pressure control to level control and closes the valve on the
discharge line.

Steam discharge line

Steam

PT

LT LC LSS PC
Water
flow
Normal
Steam drum Over ride controller
control loop loop

FIG. 7.15

QUESTION SET
1. Discuss about the merits and demerits of feed forward control.
2. What are the relative advantages and disadvantages of feed forward
control over feed back control configuration?
3. Explain how the inherent interaction existing in a process can be used for
better control using cascade loops.
4. Explain with suitable examples about ratio control systems.
5. What are over rides? Explain the concept of over ride control
configuration.
6. Explain with an example the split range control configuration.
7. Which control configuration is suitable for a process that has more than
one controlled variable and single manipulated variable? Explain.
8. Justify how ratio control can be useful in blending applications.
9. What is selective control loop configuration?
10. It is required to monitor one controlled variable in a process with more
than one manipulated variables. Explain the relevant control loop
configuration.
 Index

A Cohen-Coon corrections (for controller

Angle valves, 219 parameters), 157
Automatic control, 1 Composite controller modes, 71
Automatic level control system, 3 effect of, 123
Composite controllers, 181
B Continuous controller modes, 64
Ball valves, 220, 227 Continuous controllers,
Block diagram, 6 implementation of, 176
Bode plots, Control actions,
determination of, 161 implementation of, 170
Butterfly valves, 223, 227 modes of, 59
swing-in through, 224 Control configurations,
types of, 242
C Control valves, 207
Cage positioned ball valve, 222 cavitation problem in, 228
Cage valves, 217 characteristics of, 211
Capacitance, 11 dynamics of, 209
Cascade control (for stirred tank heater), flashing problem in, 228
237
selection of, 225, 226
Cascade control system, 236
sizing of, 225
Cascade control,
types of, 214
configuration of, 233
Controller modes,
C-C correction,
electronic implementation of, 172
general assumptions for, 158
hydraulic implementation of, 193
Characterized ball valves, 222
pneumatic implementation of, 188
Closed loop automatic control,
Z-N settings for, 159
concept of, 95
Controller operation,
Closed loop response, 96
modes of, 58

Index.p65 245 7/25/2008, 11:16 AM

 246 Index

Controller tuning, 163 General analysis procedure, 171

method of, 160 Globe valves, 227
Controller, 5 GM,
tuning of, 152, 162 determination of, 161
Conventional ball valves, 220
Current to pressure (I/P) converters, 185 H
Cybernetics, 1 High performance butterfly valves, 224
Hydraulic actuators, 206
D Hydraulic integral controller, 195
Dead time element, 22 Hydraulic proportional + integral
Dead time expression, 23 controller, 196
Dead time, Hydraulic proportional controller, 194
frequency domain analysis of, 24
I
Derivative controller mode,
Instrumentation drawings, 38
characteristics of, 70
Instrumentation, 1
effect of, 120
Integral action,
Derivative controller, 178
characteristics of, 68
Diaphragm actuators, 203
Integral controller mode, 67
Direct action of controllers, 59
effect of, 114
Discontinuous controller modes, 60
Integral controller, 177
Discontinuous controllers,
effect of (on multiple time constant
implementation of, 172
process), 117
Double acting cylinder, 206
Interacting system, 30, 32
E
L
Equal percentage, 212
Liquid level control system, 4
Error computation, 113
Liquid level systems, 25
Error equation, 97
Error terminology, 97 M
Measurement, 5
F
Multiposition mode, 61
Feed forward control, 231, 232
Multiposition controller, 174
Feed forward, 229, 231
Multi-speed floating controller mode, 63
Feedback control loop, 229, 232
Feedback control, 229, 231 N
Final control system, 5 Non-interacting system, 30
First order pade approximation, 23 Nozzle and Flapper system, 185
Floating controller mode, 62 Nozzle and Baffle system, 185
Floating controller, 175
Flow control system, 5 O
Fundamentals frequency response, 160 Offset, 66, 100
On/off controller, 173
G

Index.p65 246 7/25/2008, 11:16 AM

 Index 247

Operation, 6 Process reaction curve,

Operational amplifiers, 170 parameters obtained from, 157
Oscillatory element, 15 Proportional + derivative controller
mode, 74
P Proportional + derivative controller, 182
P + D controller mode, Proportional + integral + derivative
effect of, 128 controller mode, 77
P + D controller, Proportional + integral + derivative
characteristics of, 75 controller, 183
P + I + D controller, Proportional + integral controller mode,
71
characteristics of, 77
effect of, 123
P + I controller mode,
Proportional + integral controller, 181
characteristics of, 72
Proportional band, 64
PID controller mode,
Proportional controller (on multiple time
effect of, 131
constant process), 108
Piping, 38
Proportional controller mode, 64
PM,
effect of, 103
determination of, 161
Proportional controller, 176
Pneumatic actuator, 202
Proportional mode,
Pneumatic cylinder, 205
characteristics of, 65
Pneumatic elements, 184
Pneumatic pilots, 187 Q
Pneumatic piston actuator, 205 Quarter amplitude criteria (Cohen-Coon
Pneumatic proportional + derivative corrections), 156
controller, 190 Quick opening, 211
Pneumatic proportional + integral +
derivative controller, 191 R
Pneumatic proportional + integral Ratio control,
controller, 189 configuration of, 238
Pneumatic proportional controllers, 188 Reset controller mode, 69
Pressure (or) gas systems, 29 Resistance type element, 9
Process control loop, Reverse action of controllers, 59
element of, 4 Rotary valves, 220
Process control system, Routh array method, 163
characteristics of, 2
Process control, 1 S
terminology of, 4 Second order pade approximation, 24
Process degrees of freedom, 37 Selective control configuration, 242
Process dynamics, Single acting cylinder, 206
elements of, 9 Single capacitance process, 103
Process equation, 5 Single seated globe valve, 216
Process load, 5 Single speed floating control mode, 62
Process reaction curve method, 154

Index.p65 247 7/25/2008, 11:16 AM

 248 Index

Single time constant process (with Time constant,

multiple inputs), 107 significance of, 13
Single time constant process, 105 Transient analysis, 34
Sliding stem control valves, 214 Two position control,
Split range control, application of, 61
configuration of, 240 Two position controller,
Spring, 203 effect of, 100
Steady state velocity error, 99 Two position mode, 60
Stirred tank heater, 232
Z
T Ziegler-Nicol tuning method, 158
Temperature control system, 6 Z-N settings, 159
Thermal systems, 27
Time constant element, 12

Index.p65 248 7/25/2008, 11:16 AM