Ipcnotes

INSTRUMENTATION & PROCESS
CONTROL
Class Notes
Sanjay Dalvi
Assistant Professor
Gharda Institute of Technology
Lavel, Ratnagiri, Maharashtra
Copyright ©2012 by Sanjay Dalvi. All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form
or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as
permitted, without either the prior written permission of the Author, or authorization through payment
of the appropriate per-copy fee to the Author. Requests to the Author for permission should
be addressed to the Sanjay Dalvi,
Limit of Liability/Disclaimer of Warranty: While the author have used their best efforts in
preparing this book, they make no representations 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 for a particular purpose. No warranty may be created or extended by sales
representatives or written sales materials. The advice and strategies contained herein may not be
suitable for your situation. You should consult with a professional where appropriate. The
author shall not be liable for any loss of profit or any other commercial damages, including
but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging-in-Publication Data:
Instrumentation & Process Control / Sanjay Dalvi

“Personal Circulations.”
1. Instrumentation & Process Control. 2. Control
sciences—Research—Process Engineering. I. Dalvi, Sanjay
For Private Circulations only.
CONTENTS
PART I PROCESS CONTROL
1 Concepts of Control System 3

1.1 System 3
1.2 Steady state and Unsteady state 3
1.3 Process Control 3
1.4 Feedback and Feedforward 4
1.5 Transfer function for First Order system 4
1.5.1 Step input to first order system 5
1.5.2 Impulse input to first order system 6
1.5.3 Sinusoidal input to first order system 7
1.6 Transfer function for liquid level in tank 10
1.7 Transfer function for pure capacitive process 12
1.8 First order systems in series 13
1.8.1 Non-interacting 13
1.8.2 Interacting 16
1.9 Second order system 18
1.9.1 Types of second or higher order systems 18
1.9.2 Dynamic behaviour of second order system 18
2 Closed Loop Systems 25

iii
iv CONTENTS
2.1 What is control system? 25

2.2 Feedback 25
2.3 Servo and Regulator Problem 26
2.4 Development of Block Diagram 26
2.4.1 Process 27
2.4.2 Measuring Element 29
2.4.3 Controller & final control element 29
3 Controllers 31
3.1 Proportional Control 31
3.2 ON-OFF Control 33
3.3 Proportional-Integral Control 33
3.4 Proportional-Derivative Control 33
3.5 Proportional-Integral-Derivative Control 34
3.6 Offset 35
3.6.1 Proportional control to first order system 36
3.6.2 PI control to first order system 37
4 Stability 39
4.1 Analysing response of a system 39
4.2 Routh-Hurwitz Criterion 41
4.3 Root-Locus analysis 43
4.4 Frequency response analysis 47
4.4.1 Frequency response of pure capacitive process 48
4.4.2 Frequency response of pure dead time process 49
4.4.3 Frequency response of second order system 49
4.5 Bode Diagram 50
4.5.1 Bode stability criterion 52
4.6 Nyquist plot 56
4.6.1 First order system 56
4.6.2 Second order system 57
4.6.3 Pure dead time 58
4.6.4 Nyquist stability criterion 58
5 Controller Tuning 63
5.1 Time integral performance of controllers 63
5.2 Selection of feedback controllers 64
5.2.1 Rules for selecting controller 65
5.2.2 Controller for different processes 66
5.3 Open-loop controller tuning 67
5.4 Ziegler-Nichols tuning 69
CONTENTS v
6 Introduction to SciCos / XCos 71

6.1 Developing Model 72
6.2 Drawing Blocks 79
6.3 Connecting Blocks 79
6.4 Assigning parameters 80
6.5 Simulating Modelled diagram 81
6.6 Configuring CSCOPE 82
6.7 Configuring PID Control 83
6.8 Super-Blocks 83
6.9 Set Context 84
PART II INSTRUMENTATION
7 Flow measurement 87
7.1 Volumetric flowmeters 87
7.1.1 Rotameter 87
7.1.2 Turbine meter 87
7.1.3 Venturi meter 88
7.1.4 Orifice meter 89
7.1.5 Vortex meter 89
7.1.6 Electromagnetic flowmeter 90
7.1.7 Ultrasonic Flow Equipment 91
7.2 Mass flowmeters 92
7.2.1 Coriolis mass flowmeter 92
7.2.2 Thermal mass flowmeter 92
8 Temperature measurement 95
8.1 Thermometers 95
8.2 Thermocouples 97
8.2.1 Cold junction compensation 98
8.2.2 Types 98
8.3 Resistance Temperature Detector (RTD) 100
8.4 Thermisters 101
8.5 Pyrometers 101
9 Level measurement 105

9.1 Gauge glass 105
9.2 Chain and tape float gauges 105
9.3 Lever and shaft float gauges 105
9.4 Displacer level measuring device 106
9.5 Head-pressure level gauges 106
vi CONTENTS
9.6 Electrical type level gauges and switches 107

9.7 Capacitance probes 108
9.8 Other methods 108
PART III CONTROL STRATEGIES AND APPLICATIONS
10 Control Valve 111

10.1 Actuator 112
10.1.1 Pneumatic Actuators 113
10.1.2 Electric actuators 115
10.1.3 Hydraulic actuators 116
10.2 Positioner 117
10.3 Cage 118
10.4 Flow Characteristics and Valve Selection 118
10.5 Control valve sizing 119
10.5.1 Gas services 119
10.5.2 Liquid Service 121
11 Advanced control strategies 123

11.1 Feedforward Control 123
11.1.1 Comparison of Feedback and Feedforward Control 124
11.2 Ratio Control 126
11.2.1 Applications of Ratio control 126
11.3 Cascade control 127
11.4 Application of Feedforward, Feedback and Cascade control 129
11.5 Control strategy design at basic control level 129
11.6 Determine the different Variables 130
11.6.1 Controlled variable (CVs) 131
11.6.2 Manipulated variables (MVs) 131
11.6.3 Measured variables 131
11.6.4 Disturbance variables (DVs) 132
11.7 Plantwide control design procedure 133
11.8 Some guidelines and recommendation for control system design 137
12 Microprocessor-based controller 139

12.1 Hardware 140
12.2 Distributed Control System (DCS) 140
12.2.1 Historical Review 142
12.2.2 Modes of Computer control 143
12.2.3 Computer Control Networks 143
12.2.4 Small Computer Network 144
12.2.5 Commercial Distributed Control Systems 144
CONTENTS vii
12.2.6 Description of the DCS elements 145

12.2.7 The advantages of DCS systems 146
12.3 Programmable Logic Controllers 147
12.4 Digital control software 147
References 149
A Exercises 151
Exercises 151
PART I
PROCESS CONTROL
CHAPTER 1
CONCEPTS OF CONTROL SYSTEM
1.1 System
System in context to Process control is a set of equipments and devices interacting with
each other. This definition is considerably confusing and do not give exact idea. Therefore
lets consider a heating water bath, where water is heated to desired temperature using
heating element. Now the elements of the complete set-up consist of water tank with or
without agitator, heating element, temperature measuring device and temperature recorder
and/or controller. Combination of all these units is called as system.
1.2 Steady state and Unsteady state
If the process variables do not change with time then system is said to be at steady state.
Steady state models are always easier than unsteady state model to develop. But, unfortu-
nately all process control problems deal with unsteady state operations.
1.3 Process Control
As we discussed in previous section we have to deal with unsteady state operation which
are inherently unstable in nature and requires control. If we talk in simple language process
control subject deals with unsteady processes to make them steady.
Instrumentation & Process Control. 3
By Sanjay Dalvi Copyright © 2014 Sanjay Dalvi
4 CONCEPTS OF CONTROL SYSTEM
1.4 Feedback and Feedforward
When process is required to be controlled, it requires some basic data regarding status
of the process. If data about the status of the system is provided to controller then this
arrangement is called as Feedback control.
Opposite to feedback control, status of input is supplied to controller so that corrective
action can be taken before input disturbs the system. Feed forward arrangement is not
common and obvious.
1.5 Transfer function for First Order system
Before defining what is transfer function lets discuss the model for mercury thermometer.
Assumptions:
i) The resistance offered by the glass and mercury is negligible.
ii) At any instance the mercury assumes a uniform temperature throughout.
iii) The glass wall containing mercury does not expand or contract during the transient
response.
Unsteady state energy balance,
{Input rate} − {Output rate} = {rate of accumulation}
dTth
hA (To − Tth ) − 0 = mCp (1.1)
dt
For steady state conditions
hA (Tos − Tths ) = 0 (1.2)
Subtract equation (1.2) from equation (1.1)
d (Tth − Tths )
hA [(To − Tos ) − (Tth − Tths )] = mCp (1.3)
dt
Let deviation variables,
θo = T o − T o s
θth = Tth − Tths
therefore equation (1.3) becomes,

dθth
hA (θo − θth ) = mCp
dt
dθth
(θo − θth ) = τ (1.4)
dt
mCp
where, τ = hA Taking Laplace Transform of equation (1.4)
θo (s) − θth (s) = τ sθth (s)

θo (s) = θth (s) (τ s + 1)
TRANSFER FUNCTION FOR FIRST ORDER SYSTEM 5
θth (s) 1
= (1.5)
θo (s) τs + 1
The parameter τ is called the time constant of the system and RHS of equation(1.5)is called
the transfer function of the system.
LT of the deviation in thermometer reading
Transfer function =
LT of the deviation in surrounding temperature
Response or output variable
≈
Forcing function or input variable
Y (s)
G (s) =
X (s)
Therefore response of the system,
Y (s) = G (s) X (s)
Above equation can be represented using block diagram,
X(s) G(s) Y(s)
Equation (1.5) represents first-order system as it is derived from first order linear differen-
tial equation.
1.5.1 Step input to first order system

Step input of size A can be represented in mathematical form as given below,
X=0 for t < 0
X=A for t ≥ 0
Transfer function for input for t ≥ 0
A
X (s) = (1.6)
s
From equation (1.5) transfer function for first order system is,
Y (s) 1
= (1.7)
X (s) τs + 1
Put equation (1.6) into (1.7)
A 1
Y (s) =
s τs + 1
RHS of above equation can be simplified using partial fractions.
1 C1 C2
= +
s (τ s + 1) s (τ s + 1)
1 = C1 (τ s + 1) + C2 s
= (C1 τ + C2 ) s + C1
∴ C1 = 1
∴ C1 τ + C2 = 0
∴ C2 = −τ
A A
Y (s) = −
s s + τ1
Inverting above equation,
t
Y (t) = A 1 − e− τ (1.8)
whereas for t < 0 we get Y (t) = 0
Y (t)1.0
A
0 1 2 3 4 5
t
τ
EXAMPLE 1.1
A thermometer having a time constant of 0.1 min is at a steady state temperature of

90°C. At time t = 0, the thermometer is placed in a temperature bath maintained
100°C. Determine the time needed for the thermometer to read 98°C.
Solution:
τ = 0.1 min, xs = 90°C, and A = 10°C
Y (t) = 98 − 90 = 8◦ C
t

Y (t) = A 1 − e− τ
t

8 = 10 1 − e− 0.1
t = 0.161 min
1.5.2 Impulse input to first order system

Unit impulse input can be mathematically represented as,
X (t) = δ (t)
∴ X (s) = 1
1
Y (s) =
τs + 1

1 −t
Y (t) = e τ (1.9)
τ
1.0
τ Y (t)
0 1 2 3 4 5
t
τ
1.5.3 Sinusoidal input to first order system

Sinusoidal input can mathematically be represented as,
X (t) = A sin (ωt)
where ω = radian frequency, rad/time

A = Amplitude
Transfer function of input is,
Aω
X (s) =
s2 + ω2

Aω 1/τ Aω
Y (s) = = (1.10)
s2 + ω 2 s + 1/τ (s2 + ω 2 ) (τ s + 1)
simplifying RHS,
Aω Aω
=
(s2 + ω 2 ) (τ s + 1) (s + ωj) (s − ωj) (τ s + 1)
Aω C1 C2 C3
= + + (1.11)
(s + ωj) (s − ωj) (τ s + 1) (τ s + 1) (s + ωj) (s − ωj)
Multiply equation (1.11) with (s + ωj) and set s = −ωj
−A
C2 =
2j + 2τ ω
−A/2
=
τω + j

−A/2 τ ω − j
=
τω + j τω − j
−A/2 (τ ω − j)
=
τ 2 ω2 + 1
Multiply equation (1.11) with (s − ωj) and set s = ωj
−A/2
C3 =
τω − j

−A/2 τ ω + j
=
τω − j τω + j
−A/2 (τ ω + j)
=
τ 2 ω2 + 1
Multiply equation (1.11) with (τ s + 1) and set s = −1/τ
Aω
C1 =
− τ1 + ωj − τ1 − ωj

Aω Aτ 2 ω
= 1
= 2 2
τ2 + ω2 τ ω +1
Aτ 2 ω

1 A τω − j
Y (s) = 2 −
τ ω+1 τs + 1 2 (τ 2 ω 2 + 1) s + ωj
(1.12)
A τω + j
−
2 (τ 2 ω 2 + 1) s − ωj
Taking inverse Laplace,
Aτ 2 ω

−1 1 Aτ ω t
L = e− τ
τ 2 ω2 + 1 τ s + 1 τ 2 ω2 +1

A τω − j A (τ ω − j) −ωjt
L−1 2 2
= e
2 (τ ω + 1) s + ωj 2 (τ 2 ω 2 + 1)

A τω + j A (τ ω + j) ωjt
L−1 = e
2 (τ 2 ω 2 + 1) s − ωj 2 (τ 2 ω 2 + 1)
therefore response equation becomes,
Aτ ω
Y (t) = e−t/τ
τ 2 ω2 + 1
A
(τ ω − j) e−ωjt + (τ ω + j) eωjt

− 2 2
(1.13)
2 (τ ω + 1)
Using the identity,
e(C1 + jC2 ) t = ec1 t (cos C2 t + j sin C2 t)
here in this case, C1 = 0 and C2 = ω for third term,
eωjt = cos (ωt) + j sin (ωt)
C2 = −ω for second term,
e−ωjt = cos (−ωt) + j sin (−ωt) = cos (ωt) − j sin (ωt)
therefore the square bracket of equation (1.13) is
τ cos (ωt) − τ ωj sin (ωt) − j cos (ωt) − sin (ωt)

+ τ ω cos (ωt) + τ ωj sin (ωt) + j cos (ωt) − sin (ωt)
= 2τ ω cos (ωt) − 2 sin (ωt)
replacing square bracket of equation (1.13) we get,

Aτ ω Aτ ω
Y (t) = e−t/τ − 2 2 cos (ωt)
τ 2 ω2 + 1 (τ ω + 1)
A
+ 2 2 sin (ωt) (1.14)
(τ ω + 1)
Using trigonometric identity,
p cos A + q sin A = r sin (A + θ)

p
where, r = p2 + q 2 and tan θ = p/q here we have,
−Aτ ω
p=
τ 2ω2 + 1
A
q = 2 2
τ ω +1
s
2
(Aτ ω) + A2
r = 2
(τ 2 ω 2 + 1)
A
∴r = √
τ ω2 + 1
2
τ 2 ω2 + 1

−1 −Aτ ω
θ = tan × = tan−1 (−ωτ ) = φ
τ 2 ω2 + 1 A
putting all these in equation (1.14) we get,
Aτ ω A
Y (t) = e−t/τ + √ sin (ωt + φ) (1.15)
τ 2 ω2 + 1 τ 2 ω2 + 1
As t → ∞ the first term of equation (1.15) vanishes
A
Y (t)|s = √ sin (ωt + φ) (1.16)
τ ω2 + 1
2
Period Lag
Input
Output
Figure 1.1 Response of first order system to sinusoidal input
EXAMPLE 1.2
A mercury thermometer having a time constant of 0.1 min is placed in temperature

bath at 100°C and allowed to come to equilibrium with the bath. At time t = 0, the
temperature of the bath begins to vary sinusoidally about its average temperature of
100°C with amplitude of 2°C. If the frequency of oscillation is 10/π cycles/min, what
is the phase lag and lag and amplitude of response?
Solution:
τ = 0.1 min and amplitude A = 2°C
steady state temperature = 100°C
f = 10π cycles/min
ω = 2πf = 2π 10 π = 20 rad/min
Amplitude of the response,
A 2
√ = q = 0.896◦ C
2 2
τ ω +1 2
(0.1) × 400 + 1
π = tan−1 (−ωτ ) = tan−1 (−20 × 0.1) = −63.5◦

∴ Phase lag = 63.5◦
π
As 1 cycle ≡ 360◦ ≡ 10

min
63.5 π
lag = × = 0.0555 min (3.33 sec)
360 10
|φ|
Lag = where, φ in degrees.
360f
1.6 Transfer function for liquid level in tank
Material balance across a tank,

d (V ρ)
Fin ρin − Fout ρout =
dt
TRANSFER FUNCTION FOR LIQUID LEVEL IN TANK 11
Fin
h V
R
Fout
Figure 1.2 Liquid level in tank
for incompressible (liquid) fluid,
ρin = ρout = ρ
dV
= Fin − Fout
dt
as V = Ah and lets consider Fout = h/R
dh h
A = Fin − (1.17)
dt R
At initial steady state i.e. at t = 0
hs
0 = Fins − (1.18)
R
equation (1.17) − (1.18)
d (h − hs ) (h − hs )
A = (Fin − Fins ) − (1.19)
dt R
Using deviation variables as below,
F̄ = Fin − Fins (1.20)

H̄ = h − hs (1.21)
now equation (1.19) can be written as,
dH̄ H̄
A = F̄ −
dt R
Taking Laplace Transform (LT) of above equation,
AsH̄ (s) = F̄ (s) − H̄ (s) /R
H̄ (s) R
= (1.22)
F̄ (s) τs + 1
where, τ = AR Outflow can also be represented in terms of deviation variable as,

H̄
F¯o = Fout − Fouts =
R
taking LT of above equation,
H̄ (s)
F¯o (s) = (1.23)
R
from equation (1.22) and (1.23) we get,
F¯o (s) 1
= (1.24)
F̄ (s) τs + 1
1.7 Transfer function for pure capacitive process
Let consider a storage tank with one inflow stream and one outflow. Liquid withdrawn
from tank is through pump as shown in Fig.1.3. Mass balance across the tank can be given
Fin
h V
Fout
Figure 1.3 Pure capacitive tank
by,
dh
Fin − Fo = A (1.25)
dt
Considering system was at steady state before disturbance (i.e. t ≤ 0). Therefore mass
balance equation at steady state is,
Fin − Fo = 0 (1.26)
Subtracting equation (1.26) from equation (1.25),

H̄
F̄ = A
dt
Taking Laplace Transform of above equation,
F̄ (s) = AsH̄ (s)
H̄ (s) 1
= (1.27)
F̄ (s) As
FIRST ORDER SYSTEMS IN SERIES 13
If input is change by unit step (F (t) = 1), transfer function of input is,
1
F̄ (s) =
s
1
∴ H̄ (s) =
As
t
H̄ (t) =
A
t
h (t) = hs + (1.28)
A
Equation (1.28) is output for the step change shows continuous growth in level and system
is non-regulating. System that have a limited change in output for a sustained change in
input are said to have regulation.
1.8 First order systems in series
When more than one tanks are connected in series then connection between the tanks either
can be interacting or non-interacting. This interaction is nothing but effect of capacitance
of one tank on other.
1.8.1 Non-interacting
h1 Tank – 1
R1 q1
h2 Tank – 2
R2 q2
Figure 1.4 Non-interacting tanks
Mass balance across the tank - 1 and tank - 2 for constant density.
dh1
q − q1 = A1 (1.29)
dt
dh2
q1 − q2 = A2 (1.30)
dt
Flow and head relationship is given by,

h1
q1 = (1.31)
R1
h2
q2 = (1.32)
R2
Put equation (1.31) in (1.29),
h1 dh1
q− = A1 (1.33)
R1 dt
at (t = 0),
h1s
qs −=0 (1.34)
R1
Equation (1.33) – (1.34) and using deviation variables,
Q = q − qs
H1 = h 1 − h 1 s
H1 H1
Q− = A1
R1 dt
Taking Laplace Transform,
1
Q (s) − H1 (s) = A1 sH1 (s)
R1
Rearranging above equation,
H1 (s) R1
= (1.35)
Q (s) τ1 s + 1
where, τ1 = A1 R1 From equation(1.31) we can write,
H1 (s)
Q1 (s) =
R1
Therefore equation(1.35) becomes,
Q1 (s) 1
= (1.36)
Q (s) τ1 s + 1
Similarly, put equation(1.32) in equation(1.30) so we get,
H2 (s) R2
= (1.37)
Q1 (s) τ2 s + 1
where, τ2 = A2 R2 Replace Q1 (s) of equation(1.37) with equation(1.36)

H2 (s) R2 1
= (1.38)
Q (s) τ2 s + 1 τ1 s + 1
or
Q2 (s) 1 1
= (1.39)
Q (s) τ2 s + 1 τ1 s + 1
The overall transfer function is the product of two transfer functions of first order.
Xo X1 X2
G1 (s) G2 (s)
EXAMPLE 1.3
Two non-interacting tanks are connected in series, as shown in Fig. 1.4. The time
constants are τ1 = 1 and τ2 = 0.5; R2 = 1. Sketch the response of the level in Tank
– 2 if a unit-step change is made in the inlet flow rate of Tank – 1.
Solution:
For step change in inlet flow of Tank – 1, Q (s) = 1/s Therefore equation(1.38) is,

1 R2 1
H2 (s) =
s τ2 s + 1 τ1 s + 1
Taking partial fractions of RHS of above equation,
1 C1 C2 C3
= + + (1.40)
s (τ1 s + 1) (τ2 s + 1) s τ2 s + 1 τ1 s + 1
Multiply equation(1.40) with s and set s = 0
∴ C1 = 1
Multiply equation(1.40) with (τ1 s + 1) and set s = − τ11
τ12
∴ C2 = −
τ1 − τ2
Multiply equation(1.40) with (τ2 s + 1) and set s = − τ12
τ22
∴ C3 =
τ1 − τ2
Put values of C1 , C2 and C3 in equation(1.40)

" ! !#
1 τ1 1 τ2 1
H2 (s) = R2 − +
s (τ1 − τ2 ) s + τ11 (τ1 − τ2 ) s + τ12
− τt e− τt2
" !#
τ1 τ2 e 1
H2 (t) = R2 1− −
τ1 − τ2 τ2 τ1
Now, τ2 = 0.5, τ1 = 1 and R2 = 1
H2 (t) = 1 − 2e−t − e−2t

1.0
one tank
two tanks
H2 (t)
0 1 2 3 4 5
t
1.8.2 Interacting
h1
Tank – 1 q1 Tank – 2 h2 q2
R1 R2
Figure 1.5 Interacting tanks
Mass balance across the tank - 1 and tank - 2 for constant density.
dh1
q − q1 = A1 (1.41)
dt
dh2
q1 − q2 = A2 (1.42)
dt
Flow and head relationship is given by,
h1 − h2
q1 = (1.43)
R1
h2
q2 = (1.44)
R2
Using deviation variables equation(1.41), (1.42), (1.43) and (1.44) will be,
dH1
Q − Q1 = A1 (1.45)
dt
dH2
Q1 − Q2 = A2 (1.46)
dt
H1 − H2
Q1 = (1.47)
R1
H2
Q2 = (1.48)
R2
Transforming above equations,
Q (s) − Q1 (s) = A1 sH1 (s) (1.49)

Q1 (s) − Q2 (s) = A2 sH2 (s) (1.50)
R1 Q1 (s) = H1 (s) − H2 (s) (1.51)
R2 Q2 (s) = H2 (s) (1.52)
Replace Q2 (s) of equation(1.50) using equation(1.52)
R2 Q1 (s) − H2 (s) = A2 R2 sH2 (s)
H2 (s) R2
= (1.53)
Q1 (s) τ2 s + 1
where, τ2 = A2 R2
Using equation(1.51) replace H1 (s) of equation(1.49)
Q (s) − Q1 (s) = A1 s [R1 Q1 (s) + H2 (s)]

= τ1 sQ1 (s) + A1 sH2 (s)
Q (s) = (τ1 s + 1) Q1 (s) + A1 sH2 (s)
Using equation(1.53) replace Q1 Q (s) of above equation,

(τ1 s + 1) (τ2 s + 1) + A1 R2 s
Q (s) = H2 (s)
R2
H2 (s) R2
=
Q (s) (τ1 s + 1) (τ2 s + 1) + A1 R2 s
H2 (s) R2
= (1.54)
Q (s) τ1 τ2 s2 + (τ1 + τ2 + A1 R2 ) s + 1
based on outflow,
Q2 (s) H2 (s) Q2 (s)
=
Q (s) Q (s) H2 (s)
Using equation(1.54) and (1.52)
Q2 (s) 1
= 2
(1.55)
Q (s) τ1 τ2 s + (τ1 + τ2 + A1 R2 ) s + 1
1.9 Second order system
A second order system is one whose output, is described by the solution of a second order
differential equation e.g.
d2 y dy
a2 + a1 + a0 y = bf (t) (1.56)
dt2 dt
If a0 6= 0
d2 y dy
τ2 + 2ζτ + y = Kp x (t) (1.57)
dt2 dt
where, τ 2 = a2 /a0 , 2ζτ = a1 /a0 and Kp = b/a0
τ = natural period of oscillation of the system
ζ = damping factor
Kp = steady state or static or simply gain of the system
Equation(1.57) in terms of deviation variables,
d2 Y dY
τ2 + 2ζτ + Y = Kp X (t)
dt2 dt
taking Laplace transform of above equation,
τ 2 s2 Y (s) + 2ζτ sY (s) + Y (s) = Kp X (s)
Y (s) Kp
∴ = 2 2 (1.58)
X (s) (τ s + 2ζτ s + 1)
1.9.1 Types of second or higher order systems

1. Multi-capacity processes: Process that consist of two or more capacities in series
through which material or energy must flow.
2. Inherently second order system: Material possessing inertia is subjected to accelera-

tion.
3. A processing system with its controller.
1.9.2 Dynamic behaviour of second order system

For unit step change in input,
1
X (s) =
s
therefore equation(1.58) becomes,
Kp
Y (s) =
s (τ 2 s2 + 2ζτ s + 1)
Let p1 and p2 are the roots of the denominator of the above equation.
p p
ζ ζ2 − 1 ζ ζ2 − 1
p1 = − + and p2 = − −
τ τ τ τ
SECOND ORDER SYSTEM 19
where p1 and p2 are called poles.
The roots of the polynomial in denominator are called poles.
The roots of the polynomial in numerator are called zeros.
Kp /τ 2
∴ Y (s) = (1.59)
s (s − p1 ) (s − p2 )
Case A : When ζ > 1, two distinct and real poles
Case B : When ζ = 1, two equal poles (multiple poles)
Case C : When ζ < 1, two complex conjugate poles
Case A: Overdamped response, when ζ > 1
1 C1 C2 C3
= + + (1.60)
s (s − p1 ) (s − p2 ) s s − p1 s − p3
multiply equation(1.60) with s and set s = 0
1
C1 =
p1 p2
multiply equation(1.60) with (s − p1 ) and set s = p1
1
c2 =
p1 (p1 − p2 )
multiply equation(1.60) with (s − p2 ) and set s = p2
1
C3 = −
p1 (p1 − p2 )
put C1 , C2 , and C3 in equation(1.60) and then put equation(1.60) into equation(1.59)

Kp 1 1 1
Y (s) = 2 + − (1.61)
τ p1 p2 s p1 (p1 − p2 ) (s − p1 ) p2 (p1 − p2 ) (s − p2 )
inverting above equation
ep1 t ep2 t

Kp 1
Y (t) = 2 + − (1.62)
τ p1 p2 p1 (p1 − p2 ) p2 (p1 − p2 )
Now, using the values of p1 and p2

p ! p !
ζ ζ2 − 1 ζ ζ2 − 1
p1 p2 = − + − −
τ τ τ τ
1
=
τ2 !
p p
ζ ζ2 − 1 ζ ζ2 − 1
p1 − p2 = − + − − −
τ τ τ τ
p
2 ζ2 − 1
=
τ
p
−2ζ ζ 2 − 1 + 2 ζ 2 − 1

p1 (p1 − p2 ) =
p τ2
−2ζ ζ 2 − 1 − 2 ζ 2 − 1

p2 (p1 − p2 ) =
τ2
" t
√ 2 t t
√ 2 t #
e−ζ τ e ζ −1 τ e−ζ τ e− ζ −1 τ
Y (t) = Kp 1 − p + p
2ζ ζ 2 − 1 − 2 (ζ 2 − 1) 2ζ ζ 2 − 1 − 2 (ζ 2 − 1)
Now, let a = 2ζ ζ 2 − 1, b = 2 ζ 2 − 1 , and c = ζ 2 − 1 τt

p p
(a + b) ec − (a − b) e−c

− ζt
Y (t) = Kp 1 − e τ
a2 − b 2
a (ec − e−c ) + b (ec + e−c )

− ζt
= Kp 1 − e τ
a2 − b 2
But,
eα − e−α eα + e−α
sinh α = cosh α =
2 2
therefore,
− ζt 2a sinh c + 2b cosh c
Y (t) = Kp 1 − e τ (1.63)
a2 − b 2
Now replacing a, b, and c,
2
a2 − b2 = 4ζ 2 ζ 2 − 1 − 4 ζ 2 − 1 = 4 ζ 2 − 1

p
4ζ ζ 2 − 1
p
2a sinh c p
2−1
t ζ 2−1
t
= sinh ζ = sinh ζ
a2 − b 2 4 (ζ 2 − 1) τ
p
ζ2 − 1 τ
2

4 ζ −1

2a cosh c p t p t
2 2
= 2
cosh ζ2 − 1 = cosh ζ2 − 1
a −b 4 (ζ − 1) τ τ
( " p p #)
− ζt ζ t t
Y (t) = Kp 1 − e τ p sinh ζ2 − 1 + cosh ζ2 − 1
ζ2 − 1 τ τ
(1.64)
The gain is,
∆ (steady state output)
Kp =
∆ (steady state input)
Y (t)
1
Kp ζ1
=
ζ
ζ2 ζ2 > ζ1 > 1
Figure 1.6 Response curve for overdamped and critically damped system
Case B: Critically damped response, when ζ = 1 Poles will be,

1 1
p1 = − p2 = −
τ τ
therefore transfer function is,
Kp /τ 2 Kp
Y (s) = 2 = 2 (1.65)
s (s + 1/τ ) s (τ s + 1)
taking partial fraction,
1 C1 C2 C3
2 = + + (1.66)
s (τ s + 1) s τ s + 1 (τ s + 1)2
multiply equation(1.66) with s and set s = 0,
C1 = 1
2
multiply equation(1.66) with (τ s + 1) and set s = − τ1 ,
C3 = −τ
put C1 and C3 in equation(1.66),
2
1 = (τ s + 1) C1 + s (τ s + 1) C2 + sC3
= τ 2 C1 + τ C2 s2 + (2τ C1 + C2 + C3 ) s + C1

equating coefficients,
τ 2 C1 + τ C2 = 0
∴ C2 = −τ
therefore equation(1.65) becomes,
" #
1 τ τ
Y (s) = Kp − −
s τ s + 1 (τ s + 1)2
inverting above equation,

t −t/τ −t/τ
Y (t) = Kp 1 − e − e
τ

t −t/τ
= Kp 1 − 1 + e (1.67)
τ
Case C: Underdamped response, when ζ < 1 Partial fractions will be same as in

overdamped case i.e. equation(1.61), therefore we can use equation(1.62).
ep1 t ep2 t

Kp 1
Y (t) = 2 + − (1.62)
τ p1 p2 p1 (p1 − p2 ) p2 (p1 − p2 )
Let, p1 = a + bj and p2 = a − bj therefore,
p 1 p 2 = a2 + b 2
p1 − p2 = 2bj
p1 (p1 − p2 ) = −2b (b − aj)
p2 (p1 − p2 ) = 2b (b + aj)
replace p1 and p2 in equation(1.62),
e(a+bj)t e(a−bj)t

Kp 1
Y (t) = 2 − −
τ a2 + b 2 2b (b − aj) 2b (b + aj)
i
Kp 1 h (a+bj)t (a−bj)t
= 2 2 1 − (b + aj) e + (b − aj) e
τ (a + b2 ) 2b
Using identity,
e(a+bj)t = eat (cos bt + j sin bt)

e(a−bj)t = eat (cos (−bt) + j sin (−bt))
= eat (cos bt − j sin bt)
∴ (b + aj) e(a+bj)t + (b − aj) e(a−bj)t = 2b cos bt − 2a sin bt

eat

Kp
Y (t) = 2 2 1− (b cos bt − a sin bt)
τ (a + b2 ) b
using trigonometric identity,
p cos A + q sin A = r sin (A + θ)

p
where, r = p2 + q 2 and tan θ = p/q
Here we have, p = b, q = −a and A = bt, therefore
p
r = a2 + b 2 and θ = tan−1 (−b/a) = φ
eat p 2

Kp 2
Y (t) = 2 2 1− a + b sin (bt + φ) (1.68)
τ (a + b2 ) b
comparing roots discussed in overdamped section we can write,
p
ζ 1 − ζ2
a=− b=
τ τ
therefore, p
2 1 2 −b 1 − ζ2
a +b = 2 and =
τ a ζ
" #
1 −ζ τt
Y (t) = Kp 1 + p e sin (ωt + φ) (1.69)
1 − ζ2
p
1 − ζ2
where, ω = = radian frequency
τ !
p
1 − ζ 2
φ = tan−1 = phase lag
ζ
Y (t) A
C
Kp ±5%
B
T
trise t
Figure 1.7 Response curve for underdamped system
Characteristics of underdamped response (Refer Fig.1.7)

1. Overshoot: It’s the ratio A/B, where B is the ultimate value of the response and A
is the maximum amount by which the response exceeds its ultimate value.
!
A −πζ
Overshoot = = exp p
B 1 − ζ2
2. Decay ratio: It is the ratio C/A (i.e. the ratio of the amount above the ultimate value
of two successive peaks).
!
C −2πζ
Decay ratio = = exp p = (Overshoot)2
A 1 − ζ2
3. Period of oscillation (T ) : It is time required for one complete oscillation.
p
1 − ζ2
The radian frequency = ω =
τ
As ω = 2πf
2πτ
∴T = p
1 − ζ2
4. Natural period of oscillation: Response will oscillate continuously if ζ = 0.
Kp /τ
G (s) =
s − j τ1 s + j τ1

ωn = 1/τ
Tn = 2πτ
5. Response time: Ultimate value will reach only at t → ∞. Therefore the time needed
to reach ultimate value within ±5% is known as the response time.
6. Rise time: The time required for the response to reach its ultimate value for the first
time.
Questionnaire
1. Explain how order of the system is adjudged?

2. Why the system response become more sluggish when order of the system is high?
3. Why the increasing time constant of first order system makes the system response
slow?
4. Discuss how do you change the time constant a continuous flow liquid storage tank?
5. How does the second order system categorized based on damping factor?
6. Show that pure capacitive system is non-regulating for any type of input? (consider
step, impulse, and ramp)
7. If first order and second order systems are self regulatory then why controller is re-
quired to control these processes?
8. Find out different first and second order systems in process industry?
*******************
CHAPTER 2
CLOSED LOOP SYSTEMS
Generally all controlled systems are closed loop systems. Here we will discuss some of
the commonly encountered chemical processes.
2.1 What is control system?
Let’s consider a simple example of hot water bath. If we have to get hot water at specific
temperature, what we need is heater, temperature measuring device and controller as an
accessory to the water bath, as shown in Fig. 2.1. Temperature of water coming out from
tank will be affected by heat input from heater, temperature and flow rate of inflow water.
Now if we consider that water flow is steady, then outflow temperature is function of inflow
temperature and heat input from heater.
Now system described in Fig. 2.1 can be simplified using block diagram as shown in
Fig. 2.2.
2.2 Feedback
Feedback in control system can be of two types - negative and positive. Negative feedback
is difference between set point (TR ) and measured variable (Tm ). Negative feedback en-
sures that the difference between TR and Tm is used to adjust the control element so that
the tendency is to reduce the error. The decrease in error would cause the controller and
final control element to decrease the load.
26 CLOSED LOOP SYSTEMS
Power
Supply
Temperature
Indicator
Controller
Heater
Temperature
sensor
Figure 2.1 Water heating bath
Comparator Ti , Load
+
TR + ǫ + T
error Contrroller Final Control Element Process
Set point – Controlled
Variable
Tm
Measuring Element
Measured variable
Figure 2.2 Block diagram for hot water bath system
If the signal to the comparator were obtained by adding TR and Tm , we would have a
positive feedback system, which is inherently unstable. Positive feedback would never be
used intentionally in the system as it cause run away.
2.3 Servo and Regulator Problem
Sometimes, instead of load change, set point is required to be changed, this situation is
referred as Servo Problem. Servo Problem is need of the process. For example in heating
water, bath temperature of feed-water may be constant but hot water is required at lower
temperature then controller should be set at lower temperature. Servo problem is rare in
process industry.. Examples are missile and aircraft tracking system.
Regulator problem is common in chemical industry, where load changes and set point
remain constant. Most of the process industry problems falls under this category.
2.4 Development of Block Diagram
Lets consider same example of heating water in bath. From Fig. 2.2 we can prepare block
diagram as shown in Fig. 2.3.
For simplicity we will deal with each block separately. Objective of this system is
controlling the process, therefore first we will discuss process.
DEVELOPMENT OF BLOCK DIAGRAM 27
Ti
+
TR + ǫ + T
Gc (s) GF C (s) Gp (s)
–
Tm
Gm (s)
Figure 2.3 Transfer function bock diagram for heating bath
2.4.1 Process
Heating of water bath at steady flow conditions leaves energy balance for unsteady opera-
tion. General energy balance across the process block shows that two energy input streams
and one energy output stream. Therefore heat accumulation in process can be given as,
dT
q + ṁCp (Ti − To ) − ṁCp (T − To ) = ρCp V (2.1)
dt
At steady state, (t=0),
qs + ṁCp (Tis − To ) − ṁCp (Ts − To ) = 0 (2.2)
Subtract equation (2.2) from equation (2.1),
d (T − Ts )
q − qs + ṁCp [(Ti − Tis ) − (T − Ts )] = ρCp V (2.3)
dt
Using deviation variable concept
T̄i = Ti − Tis
T̄ = T − Ts
Q = q − qs
dT̄
Q + ṁCp T̄i − T̄ = ρCp V (2.4)
dt
Taking Laplace Transform of above equation,

Q (s) + ṁCp T̄i (s) − T̄ (s) = ρCp V sT̄ (s)
1/ṁCp 1
T̄ (s) = Q (s) + T̄i (s) (2.5)
τs + 1 τs + 1
where, τ = ρV /ṁ
Servo Problem: If change in Q (t) only, T̄i (t) = 0

1
T̄ (s) ṁCp
= (2.6)
Q (s) τs + 1
Regulatory Problem : If change in T̄i (t) only, Q (t) = 0
T̄ (s) 1
= (2.7)
T̄i (s) τs + 1
Equation (2.5) can be represented on block diagram as shown in Fig.2.4.
T̄i (s) 1
τs + 1
+
Q(s) 1 + T̄ (s)
ṁCp
τs + 1
Figure 2.4 Block diagram for Process
If we rearrange equation (2.5) as,

1
ṁCp
T̄ (s) = Q (s) + ṁCp T̄i (s) (2.8)
τs + 1
T̄i (s)
ṁCp
P (s) 1 T̄ (s)
ṁCp
τs + 1
+
Q(s) + P (s)
(a) (b)
T̄i (s)
ṁCp
Q(s) + 1 T̄ (s)
+ ṁCp
τs + 1
(c)
Figure 2.5 Alternate Block Diagram for process

Let, Q (s) + ṁCp T̄i (s) = P (s)
1
T̄ (s) ṁCp
∴ =
P (s) τs + 1
this can be represented in terms of block diagram as shown in Fig.2.5(a).

DEVELOPMENT OF BLOCK DIAGRAM 29
But, P (s) can be represented as shown in Fig 2.5(b). Therefore, equation (2.8) can be
represented using block diagram as shown in Fig.2.5(c), and it is same as block diagram
shown in Fig. 2.4.
2.4.2 Measuring Element

As we have discussed in earlier chapter, thermometer is first order system, here we will
consider measuring element exhibit first order dynamic lag.
T Measuring Tm T̄ (s) 1 T̄m (s)

Element τm s + 1
(a) (b)
Figure 2.6 (a) Block diagram, (b) Transfer function block diagram for measuring element
T̄ (s) 1
= (2.9)
T̄m (s) τm s + 1
where, T̄m = Tm − Tms
τm = time constant of measuring element.
2.4.3 Controller & final control element

For convenience, the block representing the controller and the final control element are
combined into one block.
Let assume proportional controller where control action is proportional to error.
q = Kc ǫ + qo (2.10)
where, ǫ = TR − Tm
TR = set point
Kc = proportional sensitivity or gain
qo = heat input when ǫ = 0
At steady state (t=0),
qs = qo (2.11)
Whereas ǫs = 0 and qo is constant.
Subtract equation (2.11) from (2.10)
q − qs = Kc ǫ
Using deviation variables,
Q = Kc ǫ (2.12)
In above equation,
ǫ = ǭ = ǫ − ǫs = (TR − Tm ) − (TRs − Tms )
= (TR − TRs ) − (Tm − Tms )

= T̄R − T̄m
Taking Laplace Transform of above equations,
Q (s) = Kc ǫ (s) (2.13)

ǫ (s) = T̄R (s) − T̄m (s) (2.14)
Equation (2.13) and (2.14)can be represented as shown in Fig.2.7(a) and Fig.2.7(b). There-
fore overall block diagram for controller is as shown in Fig. 2.7(c)
ǫ(s) Q(s) T̄R (s) ǫ(s)

Kc +
–
T̄m (s)
(a) (b)
T̄R (s) ǫ(s) Q(s)

Kc
+
–
T̄m (s)
(c)
Figure 2.7 (a) Representation of equation (2.13), (b) Representation of equation (2.14) on block
diagram, (c) Overall block diagram for Controller
Finally, combining Fig.2.5, Fig.2.6(b) and Fig.2.7(c) we get block diagram for complete
system (see Fig.2.8).
T̄i (s)
ṁCp
T̄R (s) ǫ(s) + 1 T̄ (s)

Q(s) + ṁCp
Kc
+ τs + 1
–
T̄m (s) 1
τm s + 1
Figure 2.8 Transfer function block diagram for heating water bath
CHAPTER 3
CONTROLLERS
In this chapter, we shall present the transfer functions for the controllers frequently used
in industrial processes. Because the transducer and the converter will be lumped together
with the controller for simplicity, the result is that the input will be the measured variable
x (e.g. temperature, level, etc.) and the output will be a pneumatic signal p.
3.1 Proportional Control
The Proportional controller produces an output signal (pressure in the case of a pneumatic
controller, current or voltage for an electronic controller) that is proportional to the error ǫ.
This action may be expressed as
p = K c ǫ + ps (3.1)
where, p = output signal from controller, psig or ma
Kc = gain, or sensitivity
ǫ = error = set point – measured variable
ps = constant
The units of set point and measured variable must be the same, since the error is the
difference between these quantities.
In a controller having adjustable gain, the value of the gain Kc can be varied by moving
a knob in the controller. The value of ps is the value of the output signal when E is zero,
and in most controllers ps can be adjusted to obtain the required output signal when the
control system is at steady state and ǫ = 0.
32 CONTROLLERS
Let consider the deviation variable P = p − ps , therefore equation (3.1) becomes,
P (t) = Kc ǫ (t) (3.2)
Laplace Transform of equation (3.2) gives transfer function of Proportional Control

P (s)
= Kc (3.3)
ǫ (s)
The term proportional band is commonly used among process control engineers in place
of the term gain. Proportional band (pb) is defined as the error (expressed as a percentage
of the range of measured variable) required to move the valve from fully closed to fully
open. A frequently used synonym is bandwidth.
EXAMPLE 3.1
A pneumatic proportional controller is used to control temperature within the range

of 60 to 100°F. The controller is adjusted so that the output pressure goes from 3 psi
(valve fully open) to 15 psi (valve fully closed) as the measured temperature goes for
71 to 75°F with the set point held constant. (a) Find the gain and the proportional
band. (b) assuming the proportional band of the controller is changed to 75%. Find
the gain and the temperature change necessary to cause a valve to go from fully open
to fully closed.
Solution:
(a) Proportional band and Gain:
Range of measured variable (fully open to fully closed) = 60 − 100 = −40°F
Change in measured variable (fully open to fully closed) = 71 − 75 = −4°F
−4
proportional band = × 100 = 10%
−40
∆P 15 − 3
Gain = = = 3 psi/◦ F
∆ǫ 75 − 71
(b) For 75% proportional band(pb):
Change in measured temperature ∆T ,
∆T = pb × (range)
= 0.75 × 40
= 30◦ F
15 − 3
Gain = = 0.4 psi/◦ F
30
From this example, we see that proportional gain corresponds inversely with proportional
band; thus
1
proportional gain ∝
proportional band
Furthermore, the relation between proportional band (pb) in percentage and Kc will be
100
Kc =
pb%
ON-OFF CONTROL 33
3.2 ON-OFF Control
A special case of proportional control is on-off control. If the gain Kc is made very high,
the valve will move from one extreme position to the other if the measured variable deviates
only slightly from the set point. This very sensitive action is called on-off action because
the valve is either fully open (ON) or fully closed (OFF); i.e., the valve acts like a switch.
This is a very simple controller and is exemplified by the thermostat used in a home-heating
system. The bandwidth of an on-off controller is approximately zero.
3.3 Proportional-Integral Control
This mode of control is described by the relationship,
Kc t
Z
p = Kc ǫ + ǫdt + ps (3.4)
τI 0
where, Kc = gain, or sensitivity

τI =integral time, min
ps = constant
The values of Kc and τI may be varied by two knobs in the controller. Integral controller
sum-up all the error values over the period equal to integral time, and corrective action is
proportional to the sum-up value. Therefore smaller the τI faster is the action taken by
integral controller.
Let’s consider unit step change in error, i.e. ǫ = 1. The response of the system can be
given as,
Kc
p (t) = Kc + t + ps (3.5)
τI
Equation (3.5) is linear equation with intercept (Kc + ps ), as shown in Fig.3.1. Using
deviation variable, P = p − ps equation (3.4) will be
Kc t
Z
P (t) = Kc ǫ (t) + ǫ (t) dt
τI 0
Transforming above equation,
Kc
P (s) = Kc ǫ (s) + ǫ (s)
τI s

P (s) 1
= Kc 1 + (3.6)
ǫ (s) τI s
The reciprocal of τI , integral time is also refer as reset rate.
3.4 Proportional-Derivative Control

dǫ
p = Kc ǫ + Kc τD + ps (3.7)
dt
34 CONTROLLERS
error
1
ǫ
0
0
t
Response
Kc
τI
p Kc
ps
0
t
Figure 3.1 Response of PI controller for step change in error
where, Kc = gain
τI = derivative time, min
ps = constant
The values of Kc and τD may be varied by two knobs in the controller. Derivative
control operates on the rate of change of error. Opposite to integral action smaller τD
means slow derivative action. Rate of change of error is calculated during the time period
τD .
Using deviation variable, P = p − ps equation (3.7) will be
dǫ (t)
P (t) = Kc ǫ (t) + Kc τD
dt
P (s) = Kc ǫ (s) + Kc τD sǫ (s)
P (s)
= Kc (1 + τD s) (3.8)
ǫ (s)
3.5 Proportional-Integral-Derivative Control

Kc t dǫ
Z
p = Kc ǫ + ǫdt + Kc τD + ps (3.9)
τI 0 dt
PID controller is three parameter control where separate knobs are available to vary Kc ,
τI , and τD . Now, using deviation variable, P = p − ps equation (3.9) will be
Kc t dǫ (t)
Z
P (t) = Kc ǫ (t) + ǫ (t) dt + Kc τD
τI 0 dt
OFFSET 35
Kc
P (s) = Kc ǫ (s) + ǫ (s) + Kc τD sǫ (s)
τI s

P (s) 1
= Kc 1 + + τD s (3.10)
ǫ (s) τI s
3.6 Offset
The curves of Fig.3.2 show the behaviour of a typical, feedback control system using dif-
ferent kinds of control when it is subjected to a permanent disturbance. The value of the
controlled variable is seen to rise at time zero owing to the disturbance. With no control,
this variable continues to rise to a new steady-state value. With control, after some time
the control system begins to take action to try to maintain the controlled variable close to
the value that existed before the disturbance occurred.
With proportional action only, the control system is able to arrest the rise of the con-
trolled variable and ultimately bring it to rest at a new steady-state value. The difference
between this new steady-state value and the original value is called offset. For the particu-
lar system shown, the offset is seen to be only 22 percent of the ultimate change that would
have been realized for this disturbance in the absence of control.
As shown by the PI curve, the addition of integral action eliminates the offset; the
controlled variable ultimately returns to the original value. This advantage of integral
action is balanced by the disadvantage of a more oscillatory behaviour. The addition of
derivative action to the PI action gives a definite improvement in the response. The rise of
the controlled variable is arrested more quickly, and it is returned rapidly to the original
value with little- or no oscillation.
No Control
Controlled Variable
Propotional
PID PI
time
Figure 3.2 Response of a typical control system showing the effects of various modes of control.
36 CONTROLLERS
P (s)
1/A
X(s) + Y (s)
+ A
Kc
+ τs + 1
–
Figure 3.3 Proportional controller to first order system
3.6.1 Proportional control to first order system

Referring to Figure(3.3) for set point change (i.e. P (s) = 0) transfer function for the
system is,
Y (s) AKc / (τ s + 1)
=
X (s) AKc
1+
(τ s + 1)
AKc A1
= =
τ s + 1 + AKc τ1 s + 1
τ
where, τ1 =
1 + AKc
AKc
A1 =
1 + AKc
for unit step change in set point,
X (s) = 1/s
A1
∴ Y (s) =
s (τ1 s + 1)
The ultimate value of response (Y (∞)) can be found out from above equation using final
value theorem.
Y (∞) = lim Y (t) = lim sY (∞)

t→∞ s→0
A1
= lim s
s→0 s (τ1 s + 1)
= A1
∴ Offset = Set point − Ultimate value of the response

AKc
=1−
1 + AKc
1
=
1 + AKc
OFFSET 37
From above equation it is clear that increase in gain of proportional controller decrease the
offset. Referring to Figure(3.3) for load change (i.e. X (s) = 0) transfer function for the
system is,
Y (s) 1/ (τ s + 1)
=
P (s) AKc
1+
(τ s + 1)
1 A2
= =
τ s + 1 + AKc τ1 s + 1
τ 1
where, τ1 = and A2 =
1 + AKc 1 + AKc
unit step change in load,
P (s) = 1/s
A2
∴ Y (s) =
s (τ1 s + 1)
The ultimate value of response (Y (∞)) can be found out from above equation using
final value theorem.
Y (∞) = lim Y (t) = lim sY (∞)

t→∞ s→0
A2
= lim s
s→0 s (τ1 s + 1)
= A2
∴ Offset = Set point − Ultimate value of the response
1
=0−
1 + AKc
1
=−
1 + AKc
From above equation it is clear that increase in Kc decrease the offset. The negative sign
indicates magnitude of response is higher than set point. If we observed servo as well as
regulatory treatment discussed here, offset is always present for proportional control.
3.6.2 PI control to first order system
P (s)
1/A
X(s) + Y (s)
+ A
Gc
+ τs + 1
–
Figure 3.4 PI controller to first order system

38 CONTROLLERS

1
Gc = Kc 1 +
τI s
Referring to Figure(3.4) for set point change (i.e. P (s) = 0) transfer function for the
system is,

1 Kp
Kc 1 +
Y (s) τ s τs + 1
= I
X (s) 1 Kp
1 + Kc 1 +
τI s τs + 1
Kc Kp (τI s + 1)
=
τI s + 1
τI s (τ s + 1) + Kc Kp τI s (τ s + 1)
τI s (τ s + 1)
Kc Kp (τI s + 1)
=
τI τ s2 + (τI + Kc Kp τI ) s + Kc Kp
with unit step change in set point (i.e. X (s) = 1/s),

1 Kc Kp (τI s + 1)
Y (s) =
s τI τ s2 + (τI + Kc Kp τI ) s + Kc Kp
ultimate value of response,
Y (∞) = lim [sY (s)] = 1

s→0
therefore offset, = X (∞) − Y (∞) = 1 − 1 = 0

Referring to Figure(3.4) for load change (i.e. X (s) = 0) transfer function for the
system is,
1
Y (s) τ s + 1
=
P (s) 1 Kp
1 + Kc 1 +
τI s τs + 1
1
=
Kc Kp
τs + 1 + (τI s + 1)
τI s
τI s
=
with unit step change in load (i.e. P (s) = 1/s),
τI
Y (s) =
ultimate value of response,
Y (∞) = lim [sY (s)] = 0

s→0
therefore offset, = X (∞) − Y (∞) = 0 − 0 = 0
*******************
CHAPTER 4
STABILITY
When response of the system grow after addition of disturbance rather than coming back
to original position then system is said at unstable condition. Mathematically response of
the system can be analysed through roots of the open-loop transfer function.
4.1 Analysing response of a system
Consider a transfer function of a system is given by equation(4.1).
Q (s) Q (s)
G (s) = = (4.1)
P (s) (s − p1 ) (s − p2 ) (s − p3 )m (s − p4 ) (s − p∗4 ) (s − p5 )
Partial fractions of above equation may be,

" #
C1 C2 C31 C32 C3m
G (s) = + + + 2 + ···+ m
s − p1 s − p2 s − p3 (s − p3 ) (s − p3 )
C4 C4∗ C5
+ + ∗ + (4.2)
s − p4 s − p4 s − p5
Location of poles p1 , p2 , p3 , p4 , and p5 are as shown in Fig.(4.1).
Analysis from Figure(4.1)

40 STABILITY
Imaginary axis
b p4
Real axis
b b b b
p1 p5 (p3 ) p2
b p∗4
Figure 4.1 Location of poles on real and imaginary space
1. Real and distinct poles: p1 and p2

Real poles give rise to the term C1 ep1 t and C2 ep2 t . Now p1 is negative therefore
it shows decay and make the system stable. p2 is positive root and it increase the
response and make the system unstable.
2. Multiple real poles: p3

If single pole is repeated multiple times it give rise to term such as,

C32 C33 2 C3m m−1 p3 t
C31 + t+ t + ···+ t e
1! 2! (m − 1)!
for above term if,
(a) p3 ≥ 0 as t → ∞ response increases and system becomes unstable.
(b) p3 < 0 as t → ∞ response decreases and system becomes stable.
3. Complex conjugate poles: p4 and p∗4

Complex conjugates always appear in pairs, i.e. p4 = a + jb and p∗4 = a − jb. This
give rise to the term eat sin (bt + φ). Where, sin (bt + φ) is periodic and the term
‘a’ governs the stability of system.
(a) When a > 0, as t → ∞ response grow and system becomes unstable.
(b) When a < 0, as t → ∞ response decay and system becomes stable.
(c) When a = 0, as t → ∞ response oscillate with constant amplitude.

ROUTH-HURWITZ CRITERION 41
output
output
time time
(a) (b)
output
time
(c)
4. Poles at origin: p5
Poles at origin (p5 = 0) give rise to the term C5 i.e. constant. Therefore, system gives
constant output at any t, and reaches to new steady state level.
EXAMPLE 4.1
Analyse the system whose open-loop transfer function is,

10Kc
G (s) =
s−1
Solution:
The characteristic equation of the open-loop transfer function will be,
1 + G (s) = 0
10Kc
1+ =0
s−1
s − 1 + 10Kc = 0
root of the above equation is p = 1 − 10Kc The gain Kc will always be greater than
zero. For stable system p < 0 condition is required, which is possible if,
1 − 10Kc < 0
−10Kc < −1
1
Kc <
10
4.2 Routh-Hurwitz Criterion
The criterion of stability for closed-loop systems does not require calculation of the actual
values of the roots of the characteristic polynomial. Let expand the characteristic equation
42 STABILITY
into following polynomial form:
1 + Gp Gf Gc Gm = a0 sn + a1 sn−1 + · · · + an−1 s + an = 0
Let a0 be positive. If it is negative, multiply both sides of the equation above by −1.
Test – I: If any of the coefficients a1 , a2 , · · · , an−1 , an is negative, there is at least

one root of the characteristic equation which has positive real part and the corresponding
system is unstable.
Test – II: If all the coefficients a0 , a1 , a2 , · · · , an−1 , an are positive, then from the first
test we cannot conclude anything about the location of the roots. Form the following array
(known as the Routh array)
Row – 1 a0 a2 a4 a6 ···
2 a1 a3 a5 a7 ···
3 A1 A2 A3 ··· ···
4 B1 B2 B3 ··· ···
5 C1 C2 C3 ··· ···
.. .. .. ..
. . . .
n+1 W1 W2 ··· ··· ···
where,
a1 a2 − a0 a3 a1 a4 − a0 a5 a1 a6 − a0 a7
A1 = A2 = A3 =
a1 a1 a1
A1 a3 − a1 A2 A1 a5 − a1 A3
B1 = B2 = ······
A1 A1
B1 A2 − A1 B2 B1 A3 − A1 B3
C1 = C2 = ······
B1 B1
Examine the elements of the first column of array (a0 , a1 , A1 , B1 , C1 , · · · , W1 ).
(a) If any of these element is negative, we have at least one root to the right side of the
imaginary axis and the system is unstable
(b) The number of sign changes in the elements of the first column is equal to the number
of roots to the right of imaginary axis.
(c) If nth row is zero then at least one root is on imaginary axis.
Therefore the system is stable if all the elements in the first order column of the Routh’s
array are positive number.
EXAMPLE 4.2
Check stability of the feedback control system with following characteristic equation.
s3 + 2s2 + (2 + Kc ) s + Kc /τI = 0
ROOT-LOCUS ANALYSIS 43
Solution:
Prepare Routh’s array
Row 1 1 2 + Kc
2 2 Kc /τI
2 (2 + Kc ) − Kc /τI
3 0
2
4 Kc /τI 0
The elements of first column are,
2 (2 + Kc ) − Kc /τI
1, 2, , Kc /τI
2
all are always positive except third. The system is stable if,
Kc
2 (2 + Kc ) >
τI
4.3 Root-Locus analysis
The Routh array is open-loop analysis and do not comment on closed loop response. Routh
Hurwitz criterion requires more mathematical calculation and require to find root of the
characteristic equation for every iteration. Root locus analysis determine the stability char-
acteristics of a closed loop system as the change in controller parameters (like Kc , τI ,
and τD ).
Procedure for plotting root-locus diagram

1. Count number of poles (n) and zeroes (m) present in transfer function.
2. Calculate number of branches of root loci emerge out (n − m).
3. Asymptotes: Loci originate from poles and may end asymptotically. If ‘n’ poles and
‘m’ zeroes are present in function then (n − m) loci approach asymptotically (n − m)
straight lines. The center of gravity of these (n − m) loci is given by,
n
P m
P
pj − zi
j=1 i=1
γ= (4.3)
n−m
These asymptotic lines make angle with real axis,
π (2k + 1)
θk+1 = (4.4)
n−m
4. Break away point: It is a point at which two root loci, emerging from adjacent poles
on real axis intersects and then leave (or enter) the real axis at an angle ±π/2. At
intersection,
m n
X 1 X 1
= (4.5)
i=1
s − z i j=1
s − pj
44 STABILITY
5. Point of intersection on Imaginary axis: Arrange open loop transfer function (G)
in standard form
N
G=K
D
where, K = constant
N = (s − z1 ) (s − z2 ) · · · · · · (s − zm )
D = (s − p1 ) (s − p2 ) · · · · · · (s − pn )
Write characteristic equation,
1+G=0
N
1+K =0
D
D + KN = 0 (4.6)
N
K = −1
D
|N |
K =1 (4.7)
|D|
Using equation(4.6) prepare Routh’s array. Use the Routh’s criterion on nth row.
i.e. if nth row is zero then at least one pair of root lie on imaginary axis. Calculate
intersection at imaginary axis by substituting s = 0 ± jb in characteristic equation.
6. Draw a locus connecting break away point and intersection on imaginary axis and
following branches which are approaching asymptotes.
7. Loci originating from poles end up in adjacent zero if any.
8. For (n − m)≥ 2, the sum of roots of the characteristic equation is constant, real, and
independent of K.
EXAMPLE 4.3
Plot the root locus diagram for the open-loop transfer function,
K
G=
(s + 1) (s + 2) (s + 3)
Solution:
By examining the transfer function we can summarise,
No.of poles, n =3 (−1, −2, −3)
No.of zeroes, m =0
No. of branches =n=3
No.of asymptots =n−m=3
Center of gravity using equation (4.3),

n
P m
P
pj − zi
j=1 i=1 −1 − 2 − 3 − 0
γ= = = −2
n−m 3−0
ROOT-LOCUS ANALYSIS 45
Angles between asymptotes and real axis (eq.(4.4)) is

π (2k + 1)
θk+1 =
n−m
here k = 0, 1, 2
therefore θ = π3 , θ = 3π
3 , and θ = 5π
3
4
Imaginary
3
Real
−4 −3 −2 −1 1
−1
−2
−3
−4
Break away point: One locus will emerge from −3 and follow real axis to asymptote.
Two loci emerge from −1 and −2 towards each other will break away at an angle ± π2
at point(4.5),
1 1 1
0= + +
s − p1 s − p2 s − p3
1 1 1
0= + +
s+1 s+2 s+3
0 = 3s2 + 12s + 11
roots of above quadratic equation are,
s = −2.5773, −1.422
as break away point will be between −1 and −2 break away point is −1.422.
4
Imaginary
1
−1.422
Real
× × ×
−4 −3 −2 −1 1
−1
−2
−3
−4
46 STABILITY
Point of intersection of two loci breaking away from real axis can be found out using
Routh’s array,
D + KN = 0
(s + 1) (s + 2) (s + 3) + K + 1 = 0
s3 + 6s2 + 11s + K + 6 = 0
Routh ’s array,
Row –1 1 11
2 6 K +6
6 × 11 − (K + 6)
3 0
6
4 K +6
If (66 − (K + 6)) /6 = 0 the one root will be on imaginary axis, therefore when
K = 60 loci will intersect imaginary axis. Point of intersection of imaginary axis is
s = a + jb and s = a − jb where a = 0.
If s = jb and K = 60, the characteristic equation becomes,
−jb3 − 6b2 + 11jb + 66 = 0

66 − 6b2 + j 11b − b3 = 0

but we know that real part is zero(i.e. a = 0),
66 − 6b2 = 0
∴ b = 3.32
similarly for s = −jb, b = −3.32 now as we know break away point and point of
intersection with imaginary axis we can draw approximate loci as shown in figure(4.2).
4
Imaginary
1
−1.422
Real
× × ×
−4 −3 −2 −1 1
−1
−2
−3
−4
Figure 4.2 Root locus diagram for example (4.3)

FREQUENCY RESPONSE ANALYSIS 47
4.4 Frequency response analysis
Consider a complex number, C = a + jb where, a is real part of C and b is imaginary part

of C. The modulus or absolute value or magnitude of C is represented by |C| and defined
by p
|C| = a2 + b2 (4.8)
The phase angle or argument of C is represented by C or arg (C) and defined by

−1 b
C = φ = tan (4.9)
a
Imaginary
C = a + jb
b |
|C
φ
a Real
Figure 4.3 Magnitude and direction of imaginary number
From fig.(4.3),
a b
cos φ = sin φ =
|C| |C|
∴ a = |c| cos φ ∴ b = |c| sin φ
therefore the complex number can be written as,
C = |C| cos φ + j|c| sin φ
We know that,
ejx + e−jx ejx − e−jx
cos x = sin x =
2 2j
Then
ejφ + e−jφ ejφ − e−jφ
C = |C| + j|C|
2 2j
jφ −jφ jφ −jφ
e +e +e −e
= |C|
2
jφ
= |C|e (4.10)
Let D = a − jb a conjugate of C then,
|C| = |D| and D = −C
Consider first order system with transfer function,

Kp
G (s) =
τp s + 1
48 STABILITY
put s = jω
Kp
G (jω) =
τp jω + 1
multiply numerator and denominator with (1 − jωτp ),
Kp 1 − jωτp
G (jω) = ×
τp s + 1 1 − jωτp
Kp (1 − jωτp )
=
1 + τp2 ω 2
Kp Kp ωτp
= 2 2
−j
1 + τp ω 1 + τp2 ω 2
Now, G (jω) is complex number with real and imaginary part.
s 2 2
Kp Kp ωτp
|G (jω) | = +
1 + τp2 ω 2 1 + τp2 ω 2
Kp
|G (jω) | = q (4.11)
1 + τp2 ω 2
and
G (jω) = tan−1 (τp ω) (4.12)
If we compare equation (4.11) and (4.12) with response equation(1.16) of first order system
with sinusoidal input (A sin ωt)
A
Y (t) = √ sin (ωt + φ) (1.16)
τ ω2 + 1
2
= A|G (jω) | sin (ωt + φ)
where, φ = tan−1 (−τp ω)

|G (jω) | = Amplitude ratio (AR)
φ = G (jω) = phase shift
Amplitude Ratio (AR) is ratio of amplitude of response to amplitude of input. There-
fore, in-general we can write,
A|G (jω) |
AR =
A
AR = |G (jω) | (4.13)
φ = G (jω) (4.14)
4.4.1 Frequency response of pure capacitive process

As discussed in section(1.7), transfer function of pure capacitive process is given by
Kp
G (s) =
s
Put s = jω,
Kp Kp jω
∴ G (jω) = = ×
jω jω jω
FREQUENCY RESPONSE ANALYSIS 49
p Kp
AR = |G (jω) | = a2 + b 2 = (4.15)
ω
and phase shift,
φ = G (jω) = tan−1 (b/a)

−1 Kp /ω
= tan −
0
= tan−1 (−∞)
φ = −90◦ or − π/2 (4.16)
4.4.2 Frequency response of pure dead time process

The performance equation of dead time process is,
Y (t) = X (t − τd )
Taking Laplace transform of above equation,
L [Y (t)] = L [X (t − τd )]
= e−sτd LX (t)
Y (s) = e−sτd X (s)
G (s) = e−sτd
put s = jω,
G (jω) = e−jτd ω
comparing above equation with equation(4.10) i.e. G (jω) = |G (jω) |e−jφ
AR = 1 (4.17)
φ = −τd ω (4.18)
4.4.3 Frequency response of second order system

Transfer function of second order system as discussed in section(1.9) can be given as,
Kp
G (s) =
τ 2 s2 + 2ζτ s + 1
put s = jω,
1 − τ 2 ω 2 − j2ζτ ω

Kp
G (jω) = ×
1 − τ 2 ω 2 + j2ζτ ω (1 − τ 2 ω 2 ) − j2ζτ ω
Kp 1 − τ 2 ω 2 − j2ζτ ω

=
(1 − τ 2 ω 2 )2 + (2ζτ ω)2
Kp 1 − τ 2 ω 2

2Kp ζτ ω
= 2 2 −j
2 2
(1 − τ ω ) + (2ζτ ω) (1 − τ ω 2 )2 + (2ζτ ω)2
2
50 STABILITY
therefore,
Kp 1 − τ 2 ω 2

a= 2 2
(1 − τ 2 ω 2 ) + (2ζτ ω)
2Kp ζτ ω
b=− 2 2
(1 − τ ω 2 ) + (2ζτ ω)
2
using equation(4.8) and (4.9)

p
AR = a2 + b 2
Kp
AR = q (4.19)
2 2
(1 − τ 2 ω 2 ) + (2ζτ ω)
φ = tan−1 (b/a)

−1 −2ζτ ω
φ = tan (4.20)
1 − τ 2 ω2
4.5 Bode Diagram
It gives the information about change in AR and phase shift with respect to frequency.
Bode diagram consist of two graphs: AR vs. frequency on log-log scale and phase shift vs.
frequency on semi-log scale. Lets analyse few systems.
First order system: As discussed in section(4.4) equations for AR and phase shift is
same,
Kp
AR = q
1 + τp2 ω 2
φ = tan−1 (−τp ω)
Let Kp = 1
1
∴ log AR = − log 1 + τp2 ω 2

2
Let’s calculate asymptotic values,
For ω → 0
1
log AR = − log (1)
2
AR = 1
For ω → ∞
2
1 + τp2 ω 2 ≈ τp2 ω 2

log AR = − log (τp ω)
If we plot above asymptotic values, we get to branches one corresponding to low

frequency (i.e. ω → 0) approaching AR = 1 and another high frequency (ω → 0)
passing through AR = 1 at τp ω = 1.
BODE DIAGRAM 51
The frequency equal to 1/τp is called corner frequency where, two asymptotes inter-
sect.
Deviation of AR at corner frequency is maximum.
Phase shift can be plotted for wide range from ω → 0 to ω → ∞. Phase lag changes
from 0◦ to −90◦ .
1
AR
1
τp
−1
τp ω
0◦
φ
−45◦
−90◦
τp ω
Figure 4.4 Bode digram for first order system
Pure capacitive system: Amplitude ratio and phase shift equations for pure capacitive
system are,
Kp
AR = φ = −90◦
ω
AR
ω=1
−1
ω
0◦
φ
−90◦
−180◦
ω
Figure 4.5 Bode digram for pure capacitive system

52 STABILITY
4.5.1 Bode stability criterion

A control system is unstable if its open-loop frequency response exhibits an AR greater
than unity at the frequency for which the phase shift is −180◦. This frequency is termed
the cross-over frequency (ωco ).
Gain Margin(GM): It is a measure of difference between AR = 1 and AR at cross-over

frequency.
1
GM = (4.21)
ARco
If Gain Margin is less than unity then system is unstable.
Phase Margin (PM): Degree of stability is expressed using phase margin.
P M = 180◦ − phase lag in degrees for which the AR is unity
A negative phase margin indicates an unstable system.

Gain Margin greater than 1.7 and phase margin greater than 30◦ is considered to be safe
for control design.
1 Gain Margin
AR
0◦
Phase Margin
φ
−180◦
−360◦
ω
Figure 4.6 Measuring Gain Margin and Phase Margin on Bode Diagram
EXAMPLE 4.4
A control system using PI control is represented by the block diagram shown in figure
below. The transfer function describing the various blocks are as shown with Kc = 10,
τI = 1 min, K1 = 0.8, and Kv = 0.5. Determine the gain and phase margins.
BODE DIAGRAM 53
U
+ C
+ 1

R + 1
Kc 1 + Kv
τI s (s2 + s + 2) (5s + 2)
–
B
K1 exp (−0.8s)
Solution:
Open-loop transfer function is

1 1
G (s) = Kc 1 + Kv 2 e−0.8s K1
τI s (s + s + 2) (5s + 2)

1 −0.8s
1+ e
s
=
(0.5s2 + 0.5s + 1) (2.5s + 1)

1 1 −0.8s 1
= × × e × 1 +
2.5s + 1 0.5s2 + 0.5s + 1 s
= G1 G2 G3 G4
∴ AR = AR1 AR2 AR3 AR4

log AR = log AR1 + log AR2 + log AR3 + log AR4
and φ = φ1 + φ2 + φ3 + φ4
For composite graph slope of graph is summation of slope of individual graphs.
1) Consider first block, G1
1
G1 =
2.5s + 1
It’s a first order system therefore, amplitude and phase shift will be,
1
AR1 = √ φ1 = tan−1 (−2.5ω)
6.25ω 2 + 1
and corner frequency,
1 1
ωc1 = = = 0.4 rad/min
τ 2.5
First order system has two asymptotes – low frequency asymptote (LFA) and high fre-
quency asymptote (HFA),
when ω → 0 LFA will be AR1 → 1
when ω → ∞ HFA will be AR1 ≈ −1 log (2.5ω)
2) Consider second block, G2
1
G2 =
0.5s2 + 0.5s + 1
√ √
It’s a second order system with τ = 1/ 2 and ζ = 1/ 2 2 therefore, amplitude and
phase shift will be,

1 −1 −0.5ω
AR2 = q φ2 = tan
2 2 1 − 0.5ω 2
(1 − 0.5ω 2 ) + (0.5ω)
54 STABILITY
and corner frequency,

1 √
= 2 = 1.414 rad/min
ωc2 =
τ
Second order system has two asymptotes – low frequency asymptote (LFA) and high
frequency asymptote (HFA),
when ω → 0 LFA will be AR2 → 1
when ω → ∞ HFA will be AR2 ≈ −2 log (0.707ω)
3) Consider third block, G3
G3 = e−0.8s
It’s a pure dead time system therefore, amplitude and phase shift will be,
AR3 = 1 φ3 = −0.8ω
and no corner frequency as pure dead time system has only one asymptote
for ω = 0 to ∞ AR will be AR3 = 1
4) Consider fourth block, G4

1
G4 = 1+
s
It’s a PI controller. Let’s first find the equation for AR and phase shift for PI controller.
Put s = jω in G4 ,
1 1 + jω
G (jω) = 1 + =
jω jω
1 + jω jω
= ×
jω jω
1
=1−j
ω
r
1
AR4 = |G (jω) | = 1 + 2
ω

1
φ = tan−1 −
ω
PI controller has two asymptotes – low frequency asymptote (LFA) and high frequency
asymptote (HFA),
when ω → 0 LFA will be AR4 ≈ −1 log ω
when ω → ∞ HFA will be AR4 → 0
Summarising above calculations we can break complete Bode diagram into four seg-
ments base on corner frequencies as tabulated below.
ω in rad/min Slope of block Slope of Composite curve

1 2 3 4
0 to 0.4 0 0 0 -1 -1
0.4 to 1.0 -1 0 0 -1 -2
1.0 to 1.414 -1 0 0 0 -1
1.414 to ∞ -1 -1 0 0 -2
BODE DIAGRAM 55
From above table we can plot Bode diagram for AR vs. frequency. From overall slope
it is clear that AR value reaches asymptotic values for low as well as high frequency
therefore select definite value for low frequency e.g. 0.01 and for high frequency e.g.
1000 and plot the graph.
100
10
b
AR
b
b
0.1
frequency for
phase margin
0.01 0.1 1 10 100

ω
Similarly we can calculate phase shift for individual block and then for overall.
ωin rad/min φ1 φ2 φ3 φ4 φ
0.01 -1.43 -0.29 -0.01 -89.43 -91.16
0.40 -45.00 -12.30 -0.32 -68.20 -125.82
1.00 -68.20 -45.00 -0.80 -45.00 -159.00
1.41 -74.20 -90.00 -1.13 -35.30 -200.40
10.00 -87.70 -174.20 -8.00 -5.70 -275.60
100.00 -89.80 -179.40 -80.00 -0.57 -349.87
0◦
−90◦
φ
−180◦
−270◦
ωco
−360◦
0.01 0.1 1 10 100
ω
56 STABILITY
4.6 Nyquist plot
A Nyquist plot is an alternate way to represent the frequency response characteristics of a

dynamic system. This is a plot of an imaginary part of G (jω) versus real part of G (jω).
Imaginary
a1 ω→∞ ω=0
φ1 Real
1
R
A
b1
X
ω = ω1
Figure 4.7 Nyquist plot
Referring to Fig.(4.7), if point X is at ω = ω1 then the distance between origin and

point X is AR at ω = ω1 .
G (jω) = a1 + jb1
q
AR1 = a21 + b21 = distance between X and origin

−1 b1
φ1 = tan = angle between real axis and line joining X – origin
a1
The angle φ1 with the real axis is phase shift at ω = ω1 . The Nyquist plot contains the
same information as the pair of Bode plots for the same system.
4.6.1 First order system

As discussed in section(4.4) equations for AR and phase shift is same with Kp = 1,
1
AR = √
1 + τ 2 ω2
φ = tan−1 (−τ ω)
When ω = 0 Amplitude Ratio, AR = 1 and phase shift, φ = 0◦ . Therefore the Nyquist

plot originate on the real axis at a distance from origin equal to 1. When ω → ∞ then
AR → 0 and φ → −90◦ . Therefore the end of the Nyquist plot is at the origin where the
distance from it is zero.
Since for every intermediate frequency 0 < AR < 1 and −90◦ < φ < 0◦ , the Nyquist
plot will be inside a unit circle and will never leave a quadrant as shown in Fig(4.8a).
NYQUIST PLOT 57
Imaginary Imaginary
1 1
ω→∞ ω=0 ω→∞ ω=0

−1 1 Real −1 1 Real
1
≥
ζ
−1 −1 <
1
ζ
(a) (b)
Imaginary
−1 1 Real
−1
(c)
Figure 4.8 Nyquist plot for (a) first order system, (b) second order system, (c) pure dead time
4.6.2 Second order system

1
AR = q
2 2
(1 − τ 2 ω 2 ) + (2ζτ ω)

−2ζτ ω
φ = tan−1
1 − τ 2 ω2
When ω = 0 Amplitude Ratio, AR = 1 and phase shift, φ = 0◦ . Therefore the Nyquist
plot originate on the real axis at a distance from origin equal to 1. When ω → ∞ then
AR → 0 and φ → −180◦. Therefore the end of the Nyquist plot is at the origin where the
distance from it is zero.
58 STABILITY
When ζ ≥ 1 then AR ≤ 1 and the Nyquist plot stays within the unit circle. When ζ < 1
then AR becomes larger than one for a range of frequencies and Nyquist plot crosses the
unit circle as shown in figure(4.8b).
4.6.3 Pure dead time

AR = 1 φ = −τd ω
Refer Fig.(4.8c) for Nyquist plot.
4.6.4 Nyquist stability criterion

If there are any net encirclement of the point (−1, 0) on the Nyquist diagram then, the
system is unstable. Figure (4.9) explains this stability criterion well.
Imaginary
Unstable
−1 1 Real
Stable
Critically
Stable
−1
Figure 4.9 Nyquist stability criterion
EXAMPLE 4.5
Draw a Nyquist plot for the system with open-loop transfer function,

1
1+
s
G=
10s2 + 5s + 1
NYQUIST PLOT 59
Solution:
Put s = jω in transfer function
1 + jω
G (jω) =
jω (−10ω 2 + j5ω + 1)
−5ω 2 + j 10ω 3 − ω

1 + jω
= ×
−5ω 2 − j (10ω 3 − ω) −5ω 2 + j (10ω 3 − ω)
−5ω − j5ω + j 10ω − ω − ω 10ω 3 − ω
2 3 3

= 2
25ω 4 + (10ω 3 − ω)
10ω 4 + 4ω 2 5ω 3 − ω
=− 2 +j 2
25ω 4 + (10ω 3 − ω) 25ω 4 + (10ω 3 − ω)
therefore amplitude ratio and phase shift is,
p
AR = a2 + b 2
v
u 10ω 4 + 4ω 2 2 + 5ω 3 − ω 2
u
=t h
u i2
25ω 4 + (10ω 3 − ω)2
5ω 2 − 1

−1
φ = tan −
10ω 3 + 4ω
Now lets find out asymptotic values of AR and phase shift.

When ω → 0
v
u 2 2
u 4ω + [−ω]2
AR ≈ t h
u i2
2
25ω 4 + (−ω)
s
4ω 4 + ω 2
≈ 2
[25ω 4 + ω 2 ]
s
ω2
≈
[ω 2 ]2
1
≈
ω

−1 1
φ ≈ tan
4ω
60 STABILITY
Therefore AR → ∞ and φ → π/2

When ω → ∞
v
u 10ω 4 2 + 5ω 3 2
u
AR ≈ t h
u i2
2
25ω 4 + (10ω 3 )
s
100ω 8 + 25ω 6
≈ 2
[25ω 4 + 100ω 6 ]
r
100ω 8
≈
10000ω 12
1
≈ 2
ω

−1 1
φ ≈ tan −
2ω
Therefore AR → 0 and φ → −π
Imaginary
−1 1 Real
−1
EXAMPLE 4.6
Draw Nyquist plot for PID controller with gain equal to one.
Solution:
Transfer function for PID Controller is,

1
G = Kc 1 + + τD s
τI s
Put s = jω and simplify the equation to get G (jω) as complex number.
s 2
1
AR = 1 + τD ω −
τI ω

−1 1
φ = tan τD ω −
τI ω
NYQUIST PLOT 61
When ω → 0 AR → ∞ φ → −π/2
When ω → ∞ AR → ∞ φ → +π/2
It means Nyquist curve is passing through/by the unit circle. Near unit circle all the
curves either bend towards origin or repelled from origin. Therefore lets find the
intersection of Nyquist curve with unit circle.
Therefore when AR = 1
" 2 #
1 1
log AR = log 1 + τD ω −
2 τI ω
1
0 = τD ω −
τI ω
r
1
ω=
τI τD
r
−1 τD τD
∴ φ = tan −
τI τI
φ = 0◦
So the Nyquist plot is passing through the point (1,0) as shown in figure.
Imaginary ω→∞
−1 1 Real
−1
ω→0
**************
CHAPTER 5
CONTROLLER TUNING
After the type of feedback controller has been selected, we still have the problem of decid-
ing what values to use for its adjusted parameters. This is known as the controller tuning
problem. There are three general approaches we can use for tuning a controller:
1. Use simple criteria such as the on-quarter decay ratio, minimum settling time mini-
mum error, etc.
2. Use time integral performance criteria.
3. Use semi-empirical rules which have been proven in practice.
5.1 Time integral performance of controllers
Dynamic performance of closed-loop is measured based on the entire response of the pro-
cess. The most often used are:
i) Integral of the square error (ISE)
Z∞
ISE = ǫ2 (t) dt
0

64 CONTROLLER TUNING
ii) Integral of the absolute value of the error (IAE)

Z∞
IAE = |ǫ (t) |dt
0
iii) Integral of the time-weighted absolute value of the error (ITAE)

Z∞
IT AE = t|ǫ (t) |dt
0
The best controller can be selected based on minimum value of ISE, IAE or ITAE.
If large error are present the ISE is better than IAE.
For small errors (ǫ < 1), IAE is better than ISE, as squared quantity will be less for
ǫ < 1.
Error that persist for long time ITAE is better as it amplify the error with t.
ITAE
output
IAE
ISE
time
Figure 5.1 Performance curve for different time integral methods
5.2 Selection of feedback controllers
i) Define an appropriate performance criterion (e.g. ISE, IAE, or ITAE)
ii) Compute the value of the performance criterion using a P or PI or PID controller with
the best settings for the adjusted parameters Kc , τI , and τD .
iii) Select that controller which gives the best value for the performance criterion.
Disadvantages of the process
i) It is very tedious.
SELECTION OF FEEDBACK CONTROLLERS 65
ii) It relies on models (Transfer Functions) for the process, sensor, and final control ele-
ment which may not be known exactly.
iii) It incorporates certain ambiguities as to which is the most appropriate criterion and
what input changes to consider.
Lets revise few characteristics of P, I and D control
Proportional Control
i) Accelerates the response of controlled process.
ii) Produces an offset for all the process except those with term 1s (integrator) in their
transfer function, such as liquid level in a tank or the gas pressure in a vessel.
Integral Control
i) Eliminates any offset
ii) Higher maximum deviations (overshoots).
iii) Produces sluggish, long oscillating responses.
iv) If we increase the gain, Kc to produce faster response, the system becomes more
oscillatory and may be led to instability.
Derivative Control
i) Anticipate future errors and introduces appropriate action.
ii) Introduces stabilizing effect on the closed-loop response of a process.
5.2.1 Rules for selecting controller

1. If possible use simple proportional controller.
when acceptable offset with moderate values of Kc is possible.

1
when the process has an integrating action (i.e. term s in TF)
for example, gas pressure vessel, liquid level control.
2. If a simple P controller is unacceptable, use a PI controller
when the response of seldom process is very fast.

for example, flow control.
3. Use PID controller to increase the speed of the closed-loop response and retain ro-
bustness.
for multi-capacity processes whose response is very sluggish.

for example, temperature and composition control.
5.2.2 Controller for different processes

1. Liquid level control
Liquid level control in distillation column at top or bottom do not require very sharp
control and always controlled in band of acceptable low level and high level based on
design parameters.
LT LT
LC
LC
Use P control Use P control
2. Gas pressure control
LT LC
Use P Control
3. Vapour pressure control
PC
PC
Use PI Controller Use PID Controller

(High speed is required) (High speed and robustness is required)
OPEN-LOOP CONTROLLER TUNING 67
4. Flow control Flow system responds fast and can be controlled by a controller with
slow response. In-fact for stable operation it is good not to use fast controller.
FC
Use PI Control
(Fast system)
5. Temperature control
Response of process seldom is sluggish due to heat transfer between heating media to
process fluid and process fluid to sensor. Therefore to increase the speed of response
‘D’ is required.
TC
TI
Use PID control
6. Composition control
Response of processes is of same type of temperature control. Use PID control.
5.3 Open-loop controller tuning
Consider the control system as shown in Fig. 5.2, which has been opened by disconnecting
the controller form final control element. Introduce a step change of magnitude A in the
variable C which actuates the final control element. In the case of a valve, C is the stem
position. Record the value of output with respect to time. The curve Ym (t) is called the
process reaction curve. Between Ym and C we have the following transfer function.
Ym (s)
GP RC (s) = = Gf (s) Gp (s) Gm (s) (5.1)
C (s)
Cohen and Coon observed that the response of most processing units to an input change,
such as the above, had s sigmoidal shape, which can be adequately approximated by the
C (s)
X (s) + Y (s)
Gc Gf Gp
–
Ym (s)
Gm
Figure 5.2 Block diagram for open loop controller tuning
response of a first-order system with dead time.
Ym (s) Ke−τd s
GP RC (s) = ⋍ (5.2)
C (s) τs + 1
which has three parameters: static gain K, dead time td , and time constant τ . From the
approximate response of Fig.(5.3), it is easy to estimate the values of the three parameters.
Thus,
output at steady state B

K= =
input at steady state A
B
τ=
S
td = intercept on time axis by line of slope S, as shown in Fig.(5.3)
output
S
td time
Figure 5.3 Process Reaction curve
Cohen and Coon used the approximate model and estimated the values of the parameters
K, td , and τ as indicated above. Then they derived expressions for the best controller
settings using load changes and various performance criteria, such as, one quarter decay
ratio, minimum offset, minimum ISE. The results of their analysis are summarized below.
ZIEGLER-NICHOLS TUNING 69
Kc τI τD

1 τ td
P 1+ – –
K td 3τ
1 τ td 30 + 3td /τ
PI 0.9 + td –
K td 12τ 9 + 20td /τ
1 τ 5 td 6 − 2td /τ
PD + – td
K td 4 6τ 22 + 3td /τ
1 τ 4 td 32 + 6td /τ 4
PID + td td
K td 3 4τ 13 + 8td /τ 11 + 2td /τ
The assumption that the first order plus dead time system may be the poor approxi-
mation.
All physical processes encountered in a chemical plant are simple first order or multi-
capacity processes whose response has the general overdamped shape.
The oscillatory overdamped behaviour is produced mainly by the presence of feed-
back controllers.
The gain of PI controller is lower than that of P controller as integral action makes the
system more sensitive, which may lead to instability.
The stabilising effect of derivative control mode allows the use of higher gains in the
PID controller.
5.4 Ziegler-Nichols tuning
Draw a Bode plot for a system (G) without controller as shown in Fig. 5.2. Find cross-over
frequency(ωco) and AR at cross-over frequency (ARco ).
The ultimate gain, Ku is
1
Ku =
ARco
The ultimate period Pu is defined as the period of the sustained cycling that would occur
if a proportional controller with gain Ku were used.
2π
Pu = time/cycle
ωco
Kc τI τD
P 0.5Ku – –
PI 0.45Ku Pu /1.2 –
PID 0.6Ku Pu /2 Pu /8
PD controller tuning is not possible with Ziegler-Nichols method.
******************
CHAPTER 6
INTRODUCTION TO SCICOS / XCOS
Scicos is a graphical dynamical system modeler and simulator. User can create block dia-
grams to model and simulate the dynamics of hybrid dynamical systems (continuous and
discrete time) and compile such models into executable code. It is used for signal pro-
cessing, systems control, queuing systems, and to study physical and biological systems.
It is developed in and distributed with the scientific software package SciLab available
for downloading at http://www.scilab.org . Scicos is available as freeware and
equally competent to Simulink tool of MATLAB, which is an expensive option.
Features of SciCos
Graphically model, compile, and simulate dynamical systems.
Combine continuous and discrete-time behaviors in the same model.
Select model elements from Palettes of standard blocks.
Program new blocks in C, Fortran, or Scilab Language.
Run simulations in batch mode from ScicosLab environment.
Generate C code from Scicos model using a Code Generator.
Run simulations in real time with real devices using Scicos-HIL.
Generate hard real-time control executables with Scicos-RTAI.

72 INTRODUCTION TO SCICOS / XCOS
Simulate digital communications systems with Scicos-ModNum.
Use implicit blocks developed in the Modelica language.
6.1 Developing Model
Scicos contains a graphical editor that can be used to construct block diagram models of
dynamical systems. The blocks can come from various palettes provided in Scicos or can
be user-defined. Different Palettes available are as below
Annotations palette
– Annotations pal : Annotations palette

– TEXT f : Free annotation
Commonly used blocks palette
– Commonlyusedblocks pal : Commonly used blocks palette

– LOGICAL OP : Logical operation
– RELATIONALOP : Relational operation
Continuous time systems palette
– Continuous pal : Continuous time systems palette

– CLINDUMMY f : Dummy
– CLR : Continuous transfer function
– CLSS : Continuous state-space system
– DERIV : Derivative
– INTEGRAL f : Integration
– INTEGRAL m : Integration
– PID : PID regulator
– TCLSS : Continuous linear system with jump
– TIME DELAY : Time delay
– VARIABLE DELAY : Variable delay
Demonstrations blocks palette
– Demonstrationsblocks pal : Demonstrations blocks palette

– AUTOMAT : automata (finite state machine)
– BOUNCE : Balls coordinates generator
– BOUNCEXY : Balls viewer
– BPLATFORM : Balls under a platform viewer
– PDE : 1D PDE block
Discontinuities palette
DEVELOPING MODEL 73
– discontinuities pal : discontinuities palette

– BACKLASH : Backlash
– DEADBAND : Deadband
– HYSTHERESIS : Hystheresis
– RATELIMITER : Rate limiter
– SATURATION : Saturation
Discrete time systems palette
– Discrete pal : Discrete time systems palette

– DELAYV f : Variable delay
– DELAY f : Discrete time delay
– DLR : Discrete transfer function
– DLRADAPT f : Discrete Zero-Pole
– DLSS : Discrete state-space system
– DOLLAR f : Delay operator
– REGISTER : Shift Register
Electrical palette
– Electrical pal : Electrical palette

– CCS : Controllable Modelica current source
– CVS : Controllable Modelica voltage source
– Capacitor : Electrical capacitor
– ConstantVoltage : Electrical DC voltage source
– CurrentSensor : Electrical current sensor
– Diode : Electrical diode
– Ground : Ground (zero potential reference)
– Gyrator : Modelica Gyrator
– IdealTransformer : Ideal Transformer
– Inductor : Electrical inductor
– NMOS : Simple NMOS Transistor
– NPN : NPN transistor
– OpAmp : Ideal opamp (norator-nullator pair)
– PMOS : Simple PMOS Transistor
– PNP : PNP transistor
– PotentialSensor : Potential sensor
– Resistor : Electrical resistor
– SineVoltage : Sine voltage source
– Switch : Non-ideal electrical switch
– VVsourceAC : Variable AC voltage source

– VariableResistor : Electrical variable resistor
– VoltageSensor : Electrical voltage sensor
– VsourceAC : Electrical AC voltage source
Event handling palette
– Events pal : Event handling palette

– ANDBLK : Activation and
– ANDLOG f : Logical and
– CEVENTSCOPE : Activation scope
– CLKFROM : Receives data from a corresponding CLKGOTO
– CLKGOTO : Pass block input to CLKFROM block
– CLKGotoTagVisibility : Define Scope of CLKGOTO tag visibility
– CLKSOMV f : Activation union
– EDGE TRIGGER : EDGE TRIGGER block
– ESELECT f : Synchronous block Event-Select
– EVTDLY c : Event delay
– EVTGEN f : Event generator
– EVTVARDLY : Event variable delay
– Extract Activation : Extract Activation block
– HALT f : Halt
– IFTHEL f : Synchronous block If-Then-Else
– MCLOCK f : MCLOCK f title
– MFCLCK f : MFCLCK f title
– M freq : Multiple Frequencies
– freq div : Frequency division
Implicit palette
– Implicit pal : Implicit palette

– CONSTRAINT c : Constraint
– DIFF f : Derivative
– Integer palette
– Integer pal : Integer palette
– BITCLEAR : Clear a Bit
– BITSET : Set a Bit
– CONVERT : Data Type Conversion
– DFLIPFLOP : D flip-flop
– DLATCH : D latch flip-flop
DEVELOPING MODEL 75
– EXTRACTBITS : Bits Extraction

– INTMUL : Integer matrix multiplication
– JKFLIPFLOP : JK flip-flop
– LOGIC : Combinatorial Logic
– SHIFT : Shift/Rotates Bits
– SRFLIPFLOP : SR flip-flop
Lookup tables palette
– Lookuptables pal : Lookup tables palette
– INTRP2BLK f : 2D interpolation
– INTRPLBLK f : Interpolation
– LOOKUP f : Lookup table
Math operations palette
– Mathoperations pal : Math operations palette
– ABS VALUE : Absolute value
– BIGSOM f : Scalar or vector Addition/Soustraction
– COSBLK f : Cosine
– EXPBLK m : Exponential of a scalar
– GAINBLK f : Gain
– INVBLK : Inverse
– LOGBLK f : Logarithm
– MATMAGPHI : Complex from/to Magnitude and Angle Conversion
– MATZREIM : Complex decomposition/composition
– MAXMIN : Maximum or minimum value of vectors’s elements
– MAX f : Maximum value of a vector’s elements
– MIN f : Minimum value of a vector’s elements
– POWBLK f : Array power
– PRODUCT : Element-wise vector multiplication/division
– PROD f : Element-wise product
– SIGNUM : Sign
– SINBLK f : Sine
– SQRT : Square root
– SUMMATION : Matrix Addition/Subtraction
– SUM f : Addition
– TANBLK f : Tangent
– TrigFun : Trigonometric function
Matrix operation palette
– Matrix pal : Matrix operation palette

– CUMSUM : Cumulative Sum
– EXTRACT : Matrix extractor
– EXTTRI : Triangular or Diagonal Extraction
– MATBKSL : Left matrix division
– MATCATH : Horizontal Concatenation
– MATCATV : Vertical Concatenation
– MATDET : Matrix Determinant
– MATDIAG : Create Diagonal Matrix
– MATDIV : Matrix division
– MATEIG : Matrix Eigenvalues
– MATEXPM : Matrix Exponential
– MATINV : Matrix Inverse
– MATLU : LU Factorization
– MATMUL : Matrix Multiplication
– MATPINV : Matrix PseudoInverse
– MATRESH : Matrix Reshape
– MATSING : SVD Decomposition
– MATSUM : Sum of Matrix’s Elements
– MATTRAN : Matrix Transpose
– MATZCONJ : Conjugate of Matrix’s Elements
– RICC : Riccati Equation
– ROOTCOEF : Polynomial Coefficient Computation
– SUBMAT : Sub-matrix Extraction
Port & Subsystem palette
– Portaction pal : Port & Subsystem palette
– CLKINV f : Input activation port
– CLKOUTV f : Output activation port
– INIMPL f : Implicit Input port
– IN f : Regular Input Port
– OUTIMPL f : Output implicit port
– OUT f : Regular Output Port
Signal processing palette
– Signalprocessing pal : Signal processing palette
– QUANT f : Quantization
– SAMPHOLD m : Sample and hold
DEVELOPING MODEL 77
Signal routing palette

– Signalrouting pal : Signal routing palette
– DEMUX : Demultiplexer
– EXTRACTOR : Extractor
– FROM : FROM Receives data from a corresponding GOTO
– FROMMO : Receives data from a corresponding GOTOMO
– GOTO : GOTO Pass block input to From block
– GOTOMO : Pass block input to FROMMO block
– GotoTagVisibility : Define Scope of GOTO tag visibility
– GotoTagVisibilityMO : Define Scope of GOTOMO tag visibility
– ISELECT m : Iselect
– MUX : Multiplexer
– M SWITCH : Multi-port switch
– NRMSOM f : Merge data
– RELAY f : Relay
– SELECT m : Select
– SWITCH2 m : Switch2
– SWITCH f : Switch
Sinks palette
– Sinks pal : Sinks palette
– AFFICH m : Display
– CANIMXY : y=f(x) animated viewer
– CANIMXY3D : z=f(x,y) animated viewer
– CFSCOPE : Floating point scope
– CMAT3D : Matrix z values 3D viewer
– CMATVIEW : Matrix Colormapped viewer
– CMSCOPE : Multi display scope
– CSCOPE : Single Display Scope
– CSCOPXY : y=f(x) permanent viewer
– CSCOPXY3D : z=f(x,y) permanent viewer
– ENDBLK : END block
– END c : END c block
– TOWS c : Data to Scilab worspace
– TRASH f : Trash block
– WFILE f : Write to file
– WRITEAU f : Write AU sound file
– WRITEC f : Write binary data
Sources palette
– Sources pal : Sources palette
– CLOCK c : Activation clock
– CONST m : Constant
– CURV f : Curve
Counter : Counter
– FROMWSB : Data from Scilab workspace to Xcos
– GENSIN f : Sine wave generator
– GENSQR f : Square wave generator
– Modulo Count : Modulo counter (0 to N counter)
– RAMP : Ramp
– RAND m : Random generator
– READAU f : Read AU sound file
– READC f : Read binary data
– RFILE f : Read from input file
– SAWTOOTH f : Sawtooth generator
– STEP FUNCTION : Step Function
– SampleCLK : Sample Time Clock
– Sigbuilder : Signal creator/generator
– TIME f : Time
– TKSCALE : Adjust value with a graphical widget.
Thermohydraulics palette
– ThermoHydraulics pal : Thermal-Hydraulics toolbox
– Bache : Thermal-hydraulic tank (reservoir)
– PerteDP : Thermal-hydraulic pipe
– PuitsP : Thermal-hydraulic drain (well)
– SourceP : Thermal-hydraulic constant pressure source
– VanneReglante : Thermal-hydraulic control valve
User defined functions palette
– Userdefinedfunctions pal : User defined functions palette
– CBLOCK : New C
– DSUPER : Masked super block
– EXPRESSION : Mathematical expression
– MBLOCK : Modelica generic block
– SUPER f : Super block
– c block : C language
DRAWING BLOCKS 79
– fortran block : Fortran

– generic block3 : Generic block
– scifunc block m : Scilab function block
Zero crossing detection palette

– Zerocrossingdetection pal : Zero crossing detection palette
– GENERAL f : GENERAL f title
– NEGTOPOS f : Threshold negative to positive
– POSTONEG f : Threshold positive to negative
– ZCROSS f : Threshold detection at zero
6.2 Drawing Blocks
Select the required block from Pallet browser. Write click on the block to get submenu.
Click on Add to <name of the file that you have opened or *untitled-xcos if unsaved new
file> (see Fig. 6.1). This operation can be alternatively performed by dragging block icon
from Pallet browser to diagram window.
Figure 6.1 Adding block to drawing pallet
6.3 Connecting Blocks
Block connection is quite user-friendly and can be done easily with mouse operation. Take
cursor near to the output port of the block, as soon as cursor is on the port, it will be
highlighted (Default colour is green) , this indicates the port is ready for the connection.
Left click on the highlighted port and drag the line till input port of downstream block and
release the click. If path between the two ports are not in straight line then the procedure
is slightly different. Left click on output port drag straight till the available straight path.
Release the click where diversion is required (connection mode will remain activated i.e.
thread will remain attached with cursor). Go to till the next diversion and click at new
diversion point (continuous press is not required), by this way it is possible to make “n´’
number of diversions. Finally click on the input port of downstream block the connection
is complete and connection mode will be deactivated. Connection in opposite direction is
also possible method is same only start point is input of downstream block and end point
is output of upstream block.
When connection is required between any port and branch of connection as shown in
Fig. 6.2, then click on port and drag the connection backward, as discussed in previous
paragraph, upto the branch. As the cursor is on the branch, it will be highlighted this indi-
cates that port can be created on branch and is ready for the connection, click on highlighted
branch to connect.
Figure 6.2 Connection of port and branch
6.4 Assigning parameters
Every block connected in diagram will produce proper result unless required parameters
are set as per users process condition. For example if we consider PID block: the gain,
integral time and derivation time is required to be set (see Fig. 6.3) by user to get desired
controlled output from the diagram. Simulation will produce result based on default values,
otherwise.
Parameters of any block can be edited by double clicking on the block or right click on
block to get block menu and click on Block parameter. The hot key is <Ctrl> + <B> on
selected block.
Figure 6.3 Parameter assigning

SIMULATING MODELLED DIAGRAM 81
6.5 Simulating Modelled diagram
Once all blocks are connected and parameters are set control diagram is ready for simula-
tion. If one press “Start” button available on toolbar of diagram window, simulation will
run using default simulation settings. Default setting for stop time is very high and simu-
lation may run for very long time depend on speed of machine and complexity of digram
drawn. This default time can be reset before starting the run. For this click on “Simulation”
menu and select “Setup” submenu, alternatively this can be done by right click option also.
See Fig. 6.4 .
Figure 6.4 Setup menu
Then Change “Final integration time” to required value as shown in Fig. 6.5. Simulation
can be scaled with real time by specifying “Real time scaling”. Other options should be
changed only by advanced user, else default values are sufficient for the most of the control
problems.
Figure 6.5 Changing Final integration time

6.6 Configuring CSCOPE
It is a basic sink block, which represent the variable fed to it on time scale. It has only one
input port but multiple variables can be plot against time by using “MUX” or multiplexer,
see Fig. 6.6. The other input it requires is time ticks, which can be given using “CLOCK c”
block, which is an event handler. For beginners it is advised to change only highlighted
Figure 6.6 Multiple variable input to CSCOPE
data in Fig. 6.7. Whereas to get complete data from beginning of the simulation run to
end on the graph, “Refresh period” should be greater than or equal to “Final integration
time” set for simulation. For details refer block help.
Figure 6.7 CSCOPE parameter settings

CONFIGURING PID CONTROL 83
User should be careful when plotting many variables on same graph, as naming legend
is not accessible until run is complete, therefore it is required to track corresponding graph
manually. The sequence is variable fed to top of the “MUX” is first variable and last is
the bottommost. Colour code for each plot graph is as per sequence specified in parameter
window (see Fig. 6.7) and is “1, 3, 5, 7, 9, 11, 13, 15” unless changed by user.
6.7 Configuring PID Control
It is very easy to set parameters for PID block. See Fig. 6.3 to set gain, integral time and
derivative time. Same block can be used for any type of controller (P, PI, PD or PID) by
assigning zero for the parameter which is not applicable.
6.8 Super-Blocks
This block opens up a new Xcos window for editing a new block diagram. This diagram
describes the internal functions of the super block. See Fig. 6.8.
Figure 6.8 Example of Super-block (in frame) and its internal block diagram
Super block inputs and outputs (regular or event) are designated by special (input or
output) blocks.
Regular input blocks must be numbered from 1 to the number of regular input ports.
Regular input ports of the super block are numbered from the top of the block shape to the
bottom. Regular output ports must be numbered from 1 to the number of regular output
ports. Regular output ports of the super block are numbered from the top of the block shape
to the bottom. Event input blocks must be numbered from 1 to the number of event input
ports. Event input ports of the super block are numbered from the left of the block shape
to the right. Event output ports must be numbered from 1 to the number of event output
ports. Event output ports of the super block are numbered from the left of the block shape
to the right.
Super-block helps to model the system from basic principles and can be used for any
process model including but not limited to distillation column, reactors, heat exchangers,
batch reactors, etc.
6.9 Set Context
When you select this menu item you obtain dialog-box as shown in Fig. 6.9. You can
Figure 6.9 Set Context for simulation
enter here Scilab instructions for defining the symbolic Xcos parameters used in block
definitions or to do whatever you want.
These instructions will be evaluated each time the diagram is loaded. If you change
the value of a symbolic Xcos parameters in the context, all the blocks that contains this
symbolic parameter are updated when you click on OK. For example if you want to set the
value of π and other parameters as shown in Fig. 6.9 .
******************
PART II
INSTRUMENTATION
CHAPTER 7
FLOW MEASUREMENT
7.1 Volumetric flowmeters
7.1.1 Rotameter
Rotameters (also known as variable-area flowmeters) are typically made from a tapered
glass tube that is positioned vertically in the fluid flow (Figure 7.1). A float that is the same
size as the base of the glass tube rides upward in relation to the amount of flow. Because
the tube is larger in diameter at the top of the glass than at the bottom, the float resides
at the point where the differential pressure between the upper and lower surfaces balance
the weight of the float. In most rotameter applications, the flow rate is read directly from a
scale inscribed on the glass; in some cases, an automatic sensing device is used to sense the
level of the float and transmit a flow signal. These transmitting rotameters are often made
from stainless steel or other materials for various fluid applications and higher pressures.
Rotameters may range in size from 1/4 inch to greater then 6 inches. They measure a wider
band of flow (10 to 1) than an orifice plate with an accuracy of 2 percent, and a maximum
operating pressure of 300 psig when constructed of glass. Rotameters are commonly used
for purge flows and levels.
7.1.2 Turbine meter

A turbine meter uses a multi-bladed rotor that is supported by bearings within a pipe sec-
tion perpendicular to the flow (Figure 7.2). Fluid drives the rotor at a velocity that is
88 FLOW MEASUREMENT
Figure 7.1 Rotameter
proportional to the fluid velocity and, consequently, to the overall volume flow rate. A
magnetic coil outside the meter produces an alternating voltage as each blade cuts the coils
magnetic lines of flux. Each pulse, therefore, represents a discrete volume of liquid. Since
the rotor is usually made of stainless steel, it is compatible with many fluids. However, the
bearings, which are necessary to support the rotor and which must allow it to spin freely
at high speeds, require a fairly clean process. Turbine meters are typically available in
pipeline sizes from less than 1/2 inch through 12 inches. They have fast response and good
accuracy.
Figure 7.2 Turbine meter
7.1.3 Venturi meter

Venturi tubes exhibit a very low pressure loss compared to other differential pressure head
meters, but they are also the largest and most costly. They operate by gradually narrow-
ing the diameter of the pipe (Figure 7.3), and measuring the resultant drop in pressure.
An expanding section of the meter then returns the flow to very near its original pressure.
As with the orifice plate, the differential pressure measurement is converted into a corre-
sponding flow rate. Venturi tube applications are generally restricted to those requiring a
low pressure drop and a high accuracy reading. They are widely used in large diameter
pipes such as those found in waste treatment plants because their gradually sloping shape
will allow solids to flow through.
VOLUMETRIC FLOWMETERS 89
Figure 7.3 Ventury meter
7.1.4 Orifice meter

A concentric orifice plate is the simplest and least expensive of the head meters (Figure
7.4). Acting as a primary device, the orifice plate constricts the flow of a fluid to produce
a differential pressure across the plate. The result is a high pressure upstream and a low
pressure downstream that is proportional to the square of the flow velocity. An orifice plate
usually produces a greater overall pressure loss than other primary devices. A practical
advantage of this device is that cost does not increase significantly with pipe size.
Figure 7.4 Orifice meter
7.1.5 Vortex meter

The operating principle of a vortex flowmeter is based on the phenomenon of vortex shed-
ding known as the von Karman effect. As fluid passes a bluff body, it separates and gen-
erates small eddies or vortices that are shed alternately along and behind each side of the
bluff body (Figure 7.5). These vortices cause areas of fluctuating pressure that are detected
by a sensor. The frequency of vortex generation is directly proportional to fluid velocity.
The output of a vortex flowmeter depends on the K-factor. The K-factor relates the
frequency of generated vortices to the fluid velocity. The formula for fluid velocity is as
follows:
Vortex frequency
Fluid Velocity =
K − f actor
The K-factor varies with Reynolds number, but it is virtually constant over a broad flow
range (Figure 7.6). Vortex flowmeters provide highly accurate linear flow rates when op-
erated within this flat region.
90 FLOW MEASUREMENT
Figure 7.5 Vortex meter
Figure 7.6 Relationship between K-factor and Reynolds Number
7.1.6 Electromagnetic flowmeter

The operating principle of magnetic flowmeter system is base upon Faradays Law of elec-
tromagnetic induction, which states that a voltage will be induced in a conductor moving
through a magnetic field. Faradays Law:
E = kBDV
The magnitude of the induced voltage E is directly proportional to the velocity of the con-
ductor V, conductor width D, and the strength of the magnetic field B. Figure 7.7 illustrates
the relationship between the physical components of the magnetic flowmeter and Faradays
Law. Magnetic field coils placed on opposite sides of the pipe generate a magnetic field.
As the conductive process liquid moves through the field with average velocity V, elec-
trodes sense the induced voltage. The width of the conductor is represented by the distance
between electrodes. An insulating liner prevents the signal from shorting to the pipe wall.
The only variable in this application of Faradays law is the velocity of the conductive liquid
V because field strength is controlled constant and electrode spacing is fixed. Therefore,
the output voltage E is directly proportional to liquid velocity, resulting in the linear output
of a magnetic flowmeter.
VOLUMETRIC FLOWMETERS 91
Figure 7.7 Electromagnetic flow meter
7.1.7 Ultrasonic Flow Equipment

Ultrasonic flowmeters use sound waves to determine the flow rate of fluids. Pulses from a
piezoelectric transducer travel through a moving fluid at the speed of sound and provide an
indication of fluid velocity. Two different methods are currently employed to establish this
velocity measurement.
The first ultrasonic meters used a transit-time method, in which two opposing transduc-
ers are mounted so that sound waves travelling between them are at a 45 degree angle to
the direction of flow within a pipe. The speed of sound from the upstream transducer to the
downstream transducer represents the inherent speed of sound plus a contribution due to
the fluid velocity. In a simultaneous measurement in the opposite direction, a value (deter-
mined electronically) is representative of the fluid velocity, which is linearly proportional
to the flow rate. While the transit-time method works well in most fluids, it is essential that
they be free of entrained gas or solids to prevent scattering of the sound waves between
transducers.
Figure 7.8 Ultrasonic flow meter
Another type of ultrasonic meter uses the Doppler effect. This type of ultrasonic meter
uses two tranducer elements as well, but each is mounted in the same case on one side
of the pipe. An ultrasonic sound wave of constant frequency is transmitted into the fluid
by one of the elements. Solids or bubbles within the fluid reflect the sound back to the re-
92 FLOW MEASUREMENT
ceiver element. The Doppler principle states that there will be a shift in apparent frequency
or wavelength when there is relative motion between transmitter and receiver. Within the
Doppler flowmeter, the relative motion of the reflecting bodies suspended within the fluid
tends to compress the sound into a shorter wavelength (high frequency). This new fre-
quency measured at the receiving element is electronically compared with the transmitted
frequency to provide a frequency difference that is directly proportional to the flow veloc-
ity in the pipe. In contrast to the transit-time method, Doppler ultrasonic meters require
entrained gases or suspended solids within the flow to function correctly.
While ultrasonic meters have several advantages, including freedom from obstruction in
the pipe and negligible cost-sensitivity with respect to pipe diameter, their performance is
very dependent on flow conditions. A fair accuracy is attainable with ultrasonic flowmeters
when properly applied to appropriate fluids.
7.2 Mass flowmeters
7.2.1 Coriolis mass flowmeter

The Coriolis meter uses an obstructionless U-shaped tube as a sensor and applies Newton’s
Second Law of Motion to determine flow rate. Inside the sensor housing, the sensor tube
vibrates at its natural frequency (Figure 7.9-a). The sensor tube is driven by an electro-
magnetic drive coil located at the center of the bend in the tube and vibrates similar to that
of a tuning fork.
The fluid flows into the sensor tube and is forced to take on the vertical momentum
of the vibrating tube. When the tube is moving upward during half of its vibration cycle
(Figure 7.9-b), the fluid flowing into the sensor resists being forced upward by pushing
down on the tube.
The fluid flowing out of the sensor has an upward momentum from the motion of the
tube. As it travels around the tube bend, the fluid resists changes in its vertical motion
by pushing up on the tube (Figure 7.9-b). The difference in forces causes the sensor tube
to twist (Figure 7.9-c). When the tube is moving downward during the second half of its
vibration cycle, it twists in the opposite direction. This twisting characteristic is called the
Coriolis effect.
Due to Newton’s Second Law of Motion, the amount of sensor tube twist is directly
proportional to the mass flow rate of the fluid flowing through the tube. Electromagnetic
velocity detectors located on each side of the flow tube measure the velocity of the vibrating
tube. Mass flow is determined by measuring the time difference exhibited by the velocity
detector signals. During zero flow conditions, no tube twist occurs, resulting in no time
difference between the two velocity signals. With flow, a twist occurs with a resulting time
difference between the two velocity signals. This time difference is directly proportional
to mass flow.
7.2.2 Thermal mass flowmeter

This version of a mass flowmeter consists of a flow-tube, an upstream and downstream
temperature sensor and a heat source as illustrated in Figure 7.10. The temperature sensors
are effectively active arms of a Wheatstone bridge. They are mounted equidistant from the
constant-temperature heat source such that for no flow conditions, heat received by each
sensor is the same, and the bridge remains in balance. However, with increasing flow, the
MASS FLOWMETERS 93
Figure 7.9 Coriolis meter
downstream sensor receives progressively more heat than the upstream sensor, causing an
imbalance to occur in the bridge circuit. The temperature difference is proportional to mass
flow rate and an electrical output representing this is developed by the bridge circuit.
Figure 7.10 Thermal mass flow meter
Thermal meters are commonly applied to gas streams only; in fact, to gas streams where
the transfer of heat to and from the stream is a usual element of the metering process.
Measuring this heat transfer supplies data from which a mass flow rate may be calculated.
As mass meters, thermal meters operate independent of density, pressure, and viscosity.
****************
CHAPTER 8
TEMPERATURE MEASUREMENT
A thermometer is a device that measures temperature or temperature gradient using a va-

riety of different principles. A thermometer has two important elements: the temperature
sensor (e.g. the bulb on a mercury thermometer) in which some physical change occurs
with temperature, plus some means of converting this physical change into a numerical
value (e.g. the scale on a mercury thermometer).
8.1 Thermometers
Thermometers may be described as empirical or absolute. Absolute thermometers are

calibrated numerically by the thermodynamic absolute temperature scale. Empirical ther-
mometers are not in general necessarily in exact agreement with absolute thermometers as
to their numerical scale readings, but to qualify as thermometers at all they must agree with
absolute thermometers and with each other in the following way: given any two bodies iso-
lated in their separate respective thermodynamic equilibrium states, all thermometers agree
as to which of the two has the higher temperature, or that the two have equal temperatures.
For any two empirical thermometers, this does not require that the relation between their
numerical scale readings be linear, but it does require that relation to be strictly monotonic.
This is a fundamental character of temperature and thermometers.
As it is customarily stated in textbooks, taken alone, the so-called ’zeroth law of ther-
modynamics’ fails to deliver this information, but the statement of the zeroth law of ther-
modynamics by James Serrin in 1977, though rather mathematically abstract, is more in-
96 TEMPERATURE MEASUREMENT
formative for thermometry: Zeroth Law - There exists a topological line which serves as a
coordinate manifold of material behaviour. The points of the manifold are called ’hotness
levels’, and is called the ’universal hotness manifold’. To this information there needs to
be added a sense of greater hotness; this sense can be had, independently of calorimetry,
of thermodynamics, and of properties of particular materials, from Wien’s displacement
law of thermal radiation: the temperature of a bath of thermal radiation is proportional,
by a universal constant, to the frequency of the maximum of its frequency spectrum; this
frequency is always positive, but can have values that tend to zero.
There are several principles on which empirical thermometers are built, as listed in the
section of this article entitled ’Primary and secondary thermometers’. Several such princi-
ples are essentially based on the constitutive relation between the state of a suitably selected
particular material and its temperature. Only some materials are suitable for this purpose,
and they may be considered as ’thermometric materials’. Radiometric thermometry, in
contrast, can be only very slightly dependent on the constitutive relations of materials. In
a sense then, radiometric thermometry might be thought of as ’universal’. This is because
it rests mainly on a universality character of thermodynamic equilibrium, that it has the
universal property of producing blackbody radiation.
The precision or resolution of a thermometer is simply to what fraction of a degree it is
possible to make a reading. For high temperature work it may only be possible to measure
to the nearest 10◦ C or more. Clinical thermometers and many electronic thermometers are
usually readable to 0.1◦ C. Special instruments can give readings to one thousandth of a
degree. However, this precision does not mean the reading is true or accurate. Thermome-
ters which are calibrated to known fixed points (e.g. 0 and 100◦ C) will be accurate (i.e.
will give a true reading) at those points. Most thermometers are originally calibrated to a
constant-volume gas thermometer. In between a process of interpolation is used, generally
a linear one. This may give significant differences between different types of thermometer
at points far away from the fixed points. For example the expansion of mercury in a glass
thermometer is slightly different from the change in resistance of a platinum resistance of
the thermometer, so these will disagree slightly at around 50◦ C. There may be other causes
due to imperfections in the instrument, e.g. in a liquid-in-glass thermometer if the capillary
tube varies in diameter.
For many purposes reproducibility is important. That is, does the same thermometer
give the same reading for the same temperature (or do replacement or multiple thermome-
ters give the same reading)? Reproducible temperature measurement means that compar-
isons are valid in scientific experiments and industrial processes are consistent. Thus if the
same type of thermometer is calibrated in the same way its readings will be valid even if it
is slightly inaccurate compared to the absolute scale.
An example of a reference thermometer used to check others to industrial standards
would be a platinum resistance thermometer with a digital display to 0.1◦ C (its precision)
which has been calibrated at 5 points against national standards (18, 0, 40, 70, 100◦ C) and
which is certified to an accuracy of ±0.2◦C.
According to British Standards, correctly calibrated, used and maintained liquid-in-
glass thermometers can achieve a measurement uncertainty of 0.01 C in the range 0 to
100 C, and a larger uncertainty outside this range: ±0.05◦ C up to 200 or down to 40◦ C,
±0.2◦ C up to 450 or down to 80◦ C.
THERMOCOUPLES 97
8.2 Thermocouples
A thermocouple consists of two conductors of different materials (usually metal alloys)

that produce a voltage in the vicinity of the point where the two conductors are in contact.
The voltage produced is dependent on, but not necessarily proportional to, the difference
of temperature of the junction to other parts of those conductors. Thermocouples are a
widely used type of temperature sensor for measurement and control and can also be used
to convert a temperature gradient into electricity. Commercial thermocouples are inexpen-
sive, interchangeable, are supplied with standard connectors, and can measure a wide range
of temperatures. In contrast to most other methods of temperature measurement, thermo-
couples are self powered and require no external form of excitation. The main limitation
with thermocouples is accuracy; system errors of less than one degree Celsius (C) can be
difficult to achieve.
Figure 8.1 Thermocouple basic circuit
Any junction of dissimilar metals will produce an electric potential related to temper-
ature. Thermocouples for practical measurement of temperature are junctions of specific
alloys which have a predictable and repeatable relationship between temperature and volt-
age. Different alloys are used for different temperature ranges. Properties such as resis-
tance to corrosion may also be important when choosing a type of thermocouple. Where
the measurement point is far from the measuring instrument, the intermediate connection
can be made by extension wires which are less costly than the materials used to make the
sensor. Thermocouples are usually standardized against a reference temperature of 0 de-
grees Celsius; practical instruments use electronic methods of cold-junction compensation
to adjust for varying temperature at the instrument terminals. Electronic instruments can
also compensate for the varying characteristics of the thermocouple, and so improve the
precision and accuracy of measurements.
When any conductor is subjected to a thermal gradient, it will generate a voltage. This
is now known as the thermoelectric effect or Seebeck effect. Any attempt to measure this
voltage necessarily involves connecting another conductor to the ”hot” end. This additional
conductor will then also experience the temperature gradient, and develop a voltage of its
own which will oppose the original. Fortunately, the magnitude of the effect depends on
the metal in use. Using a dissimilar metal to complete the circuit creates a circuit in which
the two legs generate different voltages, leaving a small difference in voltage available
for measurement. That difference increases with temperature, and is between 1 and 70
microvolts per degree Celsius (µV /◦ C) for standard metal combinations.
The voltage is not generated at the junction of the two metals of the thermocouple but
rather along that portion of the length of the two dissimilar metals that is subjected to
a temperature gradient. Because both lengths of dissimilar metals experience the same
temperature gradient, the end result is a measurement of the difference in temperature
between the thermocouple junction and the reference junction.
8.2.1 Cold junction compensation

Thermocouples measure the temperature difference between two points, not absolute tem-
perature. To measure a single temperature one of the junctions - normally the cold junction
- is maintained at a known reference temperature, and the other junction is at the tempera-
ture to be sensed.
Having a junction of known temperature, while useful for laboratory calibration, is not
convenient for most measurement and control applications. Instead, they incorporate an
artificial cold junction using a thermally sensitive device such as a thermistor or diode
to measure the temperature of the input connections at the instrument, with special care
being taken to minimize any temperature gradient between terminals. Hence, the voltage
from a known cold junction can be simulated, and the appropriate correction applied. This
is known as cold junction compensation. Some integrated circuits are designed for cold
junction temperature compensation for specific thermocouple types.
8.2.2 Types
Certain combinations of alloys have become popular as industry standards. Selection of
the combination is driven by cost, availability, convenience, melting point, chemical prop-
erties, stability, and output. Different types are best suited for different applications. They
are usually selected based on the temperature range and sensitivity needed. Thermocouples
with low sensitivities (B, R, and S types) have correspondingly lower resolutions. Other
selection criteria include the inertness of the thermocouple material, and whether it is mag-
netic or not. Standard thermocouple types are listed below with the positive electrode first,
followed by the negative electrode.
Type K (chromel 90% nickel and 10% chromiumalumel 95% nickel, 2% manganese,
2% aluminium and 1% silicon) is the most common general purpose thermocouple with
a sensitivity of approximately 41 µV /◦ C, chromel positive relative to alumel. It is inex-
pensive, and a wide variety of probes are available in its 200◦C to +1250◦C / -330◦ F to
+2460◦F range. Type K was specified at a time when metallurgy was less advanced than
it is today, and consequently characteristics may vary considerably between samples. One
of the constituent metals, nickel, is magnetic; a characteristic of thermocouples made with
magnetic material is that they undergo a deviation in output when the material reaches its
Curie point; this occurs for type K thermocouples at around 350◦C .
Type E (chromelconstantan) has a high output (68 µV /◦ C) which makes it well suited to
cryogenic use. Additionally, it is non-magnetic. Wide range is -50 to 740◦C and Narrow
range is -110 to 140◦C.
Type J (ironconstantan) has a more restricted range than type K (40 to +750◦ C), but
higher sensitivity of about 55 µV /◦ C. The Curie point of the iron (770◦C) causes an
abrupt change in the characteristic, which determines the upper temperature limit.
THERMOCOUPLES 99
Type N (NicrosilNisil) (nickel-chromium-silicon/nickel-silicon) thermocouples are suit-

able for use between 270◦C and 1300◦C owing to its stability and oxidation resistance.
Sensitivity is about 39 µV /◦ C at 900◦ C, slightly lower compared to type K.
Designed at the Defence Science and Technology Organisation (DSTO), Australia, by
Noel A Burley, type N thermocouples overcome the three principal characteristic types and
causes of thermoelectric instability in the standard base-metal thermoelement materials:
1. A gradual and generally cumulative drift in thermal EMF on long exposure at ele-
vated temperatures. This is observed in all base-metal thermoelement materials and
is mainly due to compositional changes caused by oxidation, carburization or neutron
irradiation that can produce transmutation in nuclear reactor environments. In the case
of type K, manganese and aluminium elements from the KN (negative) wire migrate
to the KP (positive) wire resulting in a down-scale drift due to chemical contamina-
tion. This effect is cumulative and irreversible.
2. A short-term cyclic change in thermal EMF on heating in the temperature range
250650◦C, which occurs in types K, J, T and E thermocouples. This kind of EMF
instability is associated with structural changes like magnetic short range order.
3. A time-independent perturbation in thermal EMF in specific temperature ranges. This
is due to composition-dependent magnetic transformations that perturb the thermal
EMFs in type K thermocouples in the range 25-225◦C, and in type J above 730◦ C.
Nicrosil and Nisil thermocouple alloys show greatly enhanced thermoelectric stability
relative to the other standard base-metal thermocouple alloys because their compositions
substantially reduces the thermoelectric instability described above. This is achieved pri-
marily by increasing component solute concentrations (chromium and silicon) in a base of
nickel above those required to cause a transition from internal to external modes of oxida-
tion, and by selecting solutes (silicon and magnesium) that preferentially oxidize to form a
diffusion-barrier, and hence oxidation inhibiting films.
Platinum types B, R, and S Types B, R, and S thermocouples use platinum or a plat-

inumrhodium alloy for each conductor. These are among the most stable thermocouples,
but have lower sensitivity than other types, approximately 10 µV /◦ C. Type B, R, and S
thermocouples are usually used only for high temperature measurements due to their high
cost and low sensitivity.
Type B thermocouples use a platinumrhodium alloy for each conductor. One conductor
contains 30% rhodium while the other conductor contains 6% rhodium. These thermocou-
ples are suited for use at up to 1800◦C. Type B thermocouples produce the same output at
0◦ C and 42◦ C, limiting their use below about 50◦ C.
Type R thermocouples use a platinumrhodium alloy containing 13% rhodium for one
conductor and pure platinum for the other conductor. Type R thermocouples are used up
to 1600◦ C.
Type S thermocouples are constructed using one wire of 90% Platinum and 10% Rhodium
(the positive or +ve wire) and a second wire of 100% platinum (the negative or -ve wire).
Like type R, type S thermocouples are used up to 1600◦C. In particular, type S is used as
the standard of calibration for the melting point of gold (1064.43◦C).
Type T (copper constantan) thermocouples are suited for measurements in the 200 to
350◦ C range. Often used as a differential measurement since only copper wire touches the
probes. Since both conductors are non-magnetic, there is no Curie point and thus no abrupt
change in characteristics. Type T thermocouples have a sensitivity of about 43 µV /◦ C.
Type C (tungsten 5% rhenium tungsten 26% rhenium) thermocouples are suited for
measurements in the 0◦ C to 2320◦C range. This thermocouple is well-suited for vacuum
furnaces at extremely high temperatures. It must never be used in the presence of oxygen
at temperatures above 260◦ C.
Type M thermocouples use a nickel alloy for each wire. The positive wire (20 Alloy)
contains 18% molybdenum while the negative wire (19 Alloy) contains 0.8% cobalt. These
thermocouples are used in vacuum furnaces for the same reasons as with type C. Upper
temperature is limited to 1400◦C. It is less commonly used than other types.
Chromel-gold/iron In chromel-gold/iron thermocouples, the positive wire is chromel and

the negative wire is gold with a small fraction (0.030.15 atom percent) of iron. It can be
used for cryogenic applications (1.2300 K and even up to 600 K). Both the sensitivity
and the temperature range depends on the iron concentration. The sensitivity is typically
around 15 µV /K at low temperatures and the lowest usable temperature varies between
1.2 and 4.2 K.
8.3 Resistance Temperature Detector (RTD)
All metals are electrical conductors which at all but very low temperatures offer resistance
to the passage of electric current. The electrical resistance exhibited by a conductor is mea-
sured in ohms. The proportional relationship of electrical current and potential difference
is given by Ohm’s law:
R = E/I
where R is resistance in ohms, E is potential difference in volts, and I is current in am-

peres. Different metals show widely different resistivity. The resistance of a conductor is
proportional to its length and inversely proportional to its cross-sectional area, i.e.,
L
R=ρ
A
where R is resistance of the conductor, ρ is resistivity of the material, L is length of the

conductor, and A is cross-sectional area of the conductor. The units of resistivity are ohms-
meter.
The resistivity of a conductor is temperature dependent. The temperature coefficient
of resistivity is positive for metals, that is, the resistance increases with temperature, and
for semiconductors the temperature coefficient is negative. As a general guide at normal
ambient temperatures the coefficient of resistivity of most elemental metals lies in the
region of 0.35 percent to 0.7 percent per ◦ C.
The metals most used for resistance measurement are platinum, nickel, and copper.
These metals have the advantage that they can be manufactured to a high degree of purity
and consequently they can be made with very high reproducibility of resistance character-
istics. Copper has the disadvantage of a low resistivity resulting in inconveniently large
sensing elements and has the further disadvantage of poor resistance to corrosion resulting
in instability of electrical characteristics. The main area of application of copper for resis-
tance thermometers is in electronic instrumentation where it is in a controlled environment
and where an essentially linear temperature characteristic is required.
THERMISTERS 101
Figure 8.2 Resistance Temperature Detector(RTD): circuit diagram
8.4 Thermisters
An alternative to platinum or nickel for resistance thermometer sensing elements is a semi-

conductor composed of mixed metal oxides. The composition of these materials depends
on the particular properties required. Combinations of two or more of the following oxides
are used: cobalt, copper, iron, magnesium, manganese, nickel, tin, titanium, vanadium,
and zinc. Devices made of these materials are called thermistors. They consist of a piece
of the semiconductor to which two connecting wires are attached at opposite sides or ends.
Thermistors have a negative temperature coefficient; that is, as the temperature rises the
electrical resistance of the device falls. This variation of resistance with temperature is
much higher than in the case of metals. Typical resistance values are 10kΩ at 0◦ C and
200Ω at 100◦ C. This very high sensitivity allows measurement or control to a very high
resolution of temperature differences. The accuracy is not as good as for a metallic resis-
tance thermometer owing to the difficulty in controlling the composition of the thermistor
material during manufacture. The resolution differs across the usable span of the devices
due to their non-linearity. With the right choice of device characteristics it is nevertheless
possible to control a temperature to within very close limits: 0.001 ◦ C temperature change
is detectable.
The total range that can be measured with thermistors is from -100◦C to +300◦C. How-
ever, the span cannot be covered by one thermistor type; four or five types are needed.
The physical construction of thermistors covers a wide range. The smallest are encap-
sulated in glass or epoxy beads of 1 – 2.5 mm diameter; bigger ones come as discs 5 –
25 mm diameter or rods 1 – 6 mm diameter and up to 50 mm length. The bigger devices
are able to pass quite high currents and so operate control equipment directly without need
of amplifiers. Thermistors are also available in metal encapsulations like those used for
platinum resistance thermometers.
The big disadvantage of thermistors is that their characteristics are non-linear.
8.5 Pyrometers
A pyrometer is a non-contacting device that intercepts and measures thermal radiation, a

process known as pyrometry. This device can be used to determine the temperature of an
object’s surface.
A pyrometer has an optical system and a detector. The optical system focuses the ther-
mal radiation onto the detector. The output signal of the detector (temperature T) is related
to the thermal radiation or irradiance j∗ of the target object through the StefanBoltzmann
law, the constant of proportionality σ, called the Stefan-Boltzmann constant and the emis-
sivity ε of the object.

j∗ = εσT 4
This output is used to infer the object’s temperature. Thus, there is no need for direct
contact between the pyrometer and the object, as there is with thermocouples and resistance
temperature detectors (RTDs).
Figure 8.3 Pyrometer
The potter Josiah Wedgwood invented the first pyrometer to measure the temperature
in his kilns. Modern pyrometers became available when the first disappearing filament
pyrometer was built by L. Holborn and F. Kurlbaum in 1901. This device superimposed
a thin, heated filament over the object to be measured and relied on the operators eye to
detect when the filament vanished. The object temperature was then read from a scale on
the pyrometer.
The temperature returned by the vanishing filament pyrometer and others of its kind,
called brightness pyrometers, is dependent on the emissivity of the object. With greater
use of brightness pyrometers, it became obvious that problems existed with relying on
knowledge of the value of emissivity. Emissivity was found to change, often drastically,
with surface roughness, bulk and surface composition, and even the temperature itself.
To get around these difficulties, the ratio or two-color pyrometer was developed. They
rely on the fact that Planck’s law, which relates temperature to the intensity of radiation
emitted at individual wavelengths, can be solved for temperature if Plancks statement of
the intensities at two different wavelengths is divided. This solution assumes that the emis-
sivity is the same at both wavelengths and cancels out in the division. This is known as
the gray body assumption. Ratio pyrometers are essentially two brightness pyrometers in
a single instrument. The operational principles of the ratio pyrometers were developed in
the 1920s and 1930s, and they were commercially available in 1939.
As the ratio pyrometer came into popular use, it was determined that many materials,
of which metals are an example, do not have the same emissivity at two wavelengths. For
these materials, the emissivity does not cancel out and the temperature measurement is
in error. The amount of error depends on the emissivities and the wavelengths where the
measurements are taken. Two-color ratio pyrometers cannot measure whether a materials
emissivity is wavelength dependent.
PYROMETERS 103
To more accurately measure the temperature of real objects with unknown or changing
emissivities, multiwavelength pyrometers were envisioned at the US National Institute of
Standards and Technology and described in 1992. Multiwavelength pyrometers use three
or more wavelengths and mathematical manipulation of the results to attempt to achieve
accurate temperature measurement even when the emissivity is unknown, changing, and
different at all wavelengths.
**************
CHAPTER 9
LEVEL MEASUREMENT
9.1 Gauge glass
This is the most commonly used visual process-level device. Gauge glasses are generally
classified as either transparent or reflex types. A transparent gauge glass consists of either
a glass tube or an arrangement of flat glass plates in some type of holder. Since the process
fluid level is viewed directly, the transparent gauge glass is normally used with opaque
fluids. The reflex type has reflecting prisms to aid in viewing transparent fluids. Caution
should be observed when handling and installing these and/or any tempered glass instru-
ment. Scratches or chips can reduce the strength of the glass and cause safety problems.
9.2 Chain and tape float gauges
Used in large, unpressurized storage tanks where the entire full-toempty range must be
measured.
9.3 Lever and shaft float gauges
Used on either unpressurized or pressurized vessels where only a small range of level must
be measured. The range of measurement is determined by the length of the float arm, but
usually is between a few inches and a few feet.
106 LEVEL MEASUREMENT
9.4 Displacer level measuring device
One of the most frequently used level measuring devices is the torque tube displacer. It
is attached to the free end of a torque tube which has elastic properties that permit it to
twist as the displacer tries to float. This slight turning of the free end of the torque tube
is connected to an indicator or transmitter. Torque tube displacement gauges are normally
limited to level spans of ten feet.
Figure 9.1 Level gauges
9.5 Head-pressure level gauges
The true level of a liquid can be determined by dividing the measured hydrostatic head by
the density of the liquid. This method requires a knowledge of the densities of all phases
of the liquid. Some of these methods are: pressure gauge, bubble tube, and differential
pressure measurement. The bubbler is used at vacuum and low pressures and is especially
good for services such as molten sulphur and dirty liquids. In boiling-liquid service, a
condensate trap must be used on the vapour leg. The level of trapped condensate in the
vapour leg will usually be different than the vessel liquid level, requiring compensation of
the transmitter.
ELECTRICAL TYPE LEVEL GAUGES AND SWITCHES 107
Figure 9.2 Head Pressure Level Gauges
9.6 Electrical type level gauges and switches
Two common types of level gauges are the float-magnetic gauge configuration and the
conductive type shown in Fig. 9.3. Slight tension on the tape reel permits the follower
magnet to track the float at the liquid level in the device in Fig. 9.3a. The position of the
reel represents the level and is either connected to an indicating device or a transmitter.
The device shown in Fig. 9.3b illustrates the use of a conductive fluid for high and low
level alarm indication.
Figure 9.3 Electrical Level Gauges/Switches

108 LEVEL MEASUREMENT
9.7 Capacitance probes
A continuous method of level measurement based on electrical properties. This method

uses an electrode placed inside a vessel (or in a protective shell inside the vessel). The
capacitance between the electrode and the wall of the vessel or shell varies as the dielectric
constant varies. The dielectric in this case is the fluid, hence the capacitance varies in
proportion to the liquid level. This capacitance is then measured, and converted to a level
measurement to be indicated or transmitted.
9.8 Other methods
Some of the other level detection methods use principles of:

1. Absorption of nuclear radiation as in a gamma gauge.
2. Absorption of ultrasonic waves.
3. Viscosity of the liquid.
4. Thermal conductivity and expansion.
*****************
PART III
CONTROL STRATEGIES AND

APPLICATIONS
CHAPTER 10
CONTROL VALVE
Process plants consist of hundreds, or even thousands, of control loops all networked to-
gether to produce a product to be offered for sale. Each of these control loops is designed
to keep some important process variable such as pressure, flow, level, temperature, etc.
within a required operating range to ensure the quality of the end product. Each of these
loops receives and internally creates disturbances that detrimentally affect the process vari-
able, and interaction from other loops in the network provides disturbances that influence
the process variable.
To reduce the effect of these load disturbances, sensors and transmitters collect informa-
tion about the process variable and its relationship to some desired set point. A controller
then processes this information and decides what must be done to get the process variable
back to where it should be after a load disturbance occurs. When all the measuring, com-
paring, and calculating are done, some type of final control element must implement the
strategy selected by the controller.
The most common final control element in the process control industries is the control
valve. The control valve manipulates a flowing fluid, such as gas, steam, water, or chemical
compounds, to compensate for the load disturbance and keep the regulated process variable
as close as possible to the desired set point.
Many people who talk about control valves or valves are really referring to a control
valve assembly. The control valve assembly typically consists of the valve body, the inter-
nal trim parts, an actuator to provide the motive power to operate the valve, and a variety of
additional valve accessories, which can include positioners, transducers, supply pressure

112 CONTROL VALVE
regulators, manual operators, snubbers, or limit switches. Other chapters of this handbook
supply more detail about each of these control valve assembly components.
Whether it is called a valve, control valve or a control valve assembly is not as important
as recognizing that the control valve is a critical part of the control loop. It is not accurate
to say that the control valve is the most important part of the loop. It is useful to think of a
control loop as an instrumentation chain. Like any other chain, the whole chain is only as
good as its weakest link. It is important to ensure that the control valve is not the weakest
link.
Figure 10.1 Control valve body – internals
10.1 Actuator
A pneumatic, hydraulic, or electrically powered device that supplies force and motion to
open or close a valve. An actuator, including all the pertinent accessories that make it a
complete operating unit.
A spring, or group of springs, enclosed in the yoke or actuator casing that moves the
actuator stem in a direction opposite to that created by diaphragm pressure. The part that
connects the actuator to the valve stem and transmits motion (force) from the actuator to
the valve.An extension of the piston actuator stem to provide a means of transmitting piston
motion to the valve positioner. The net force from an actuator that is available for actual
positioning of the valve plug.
The methods of valve actuation fall into one of the following categories:
1. Linear
2. Part-turn
3. Multi-turn
Each of the above can be powered by one of the following power sources:
1. Hydraulic
2. Electric
ACTUATOR 113
Figure 10.2 Actuators
3. Pneumatic
When choosing the type of actuator the engineer must consider the following factors:
1. The actuator must deliver enough torque to move and seat the valve with the mini-
mum power supply available and also maintain the required position (open, closed, or
intermediate) under the worst flow conditions that the valve might experience.
2. The actuator must be capable of completing the travel, that is, 90, 180, and multi-turn
at the required speed.
3. The actuator must include a device that prevents excessive forces being applied to the
valve under adverse power supply conditions.
10.1.1 Pneumatic Actuators

Pneumatic actuators are probably the most common type of actuator, because of their cheap
and readily available power source, which is compressed air. Not only are they generally
less expensive than electric or hydraulic alternatives, but they are also less complex in
design and so are easier to maintain.
Pneumatic actuators are particularly suited to part-turn valves with their limited stroke
requirements, however they can also be used for linear closing valves. Pneumatic power
is relatively easy to store and it is suitable for use when power sources are of a limited
capacity or simply not available. Pneumatic actuators can be used on valves of most sizes.
Pneumatic supply is generally compressed dry air, but designs are also available for nat-
ural gas. Compressor and accumulator design is a consideration when designing a network
of pneumatically operated actuators to achieve an efficient system. Valve operating times
are also a significant factor when determining the pneumatic capacity of the supply and the
size of the tubing.
Diaphragm types Diaphragm actuators are usually designed for linear motion, although
it is possible to have rotary motion designs. Pneumatic actuators are usually attached to
114 CONTROL VALVE
the body of the valve by means of a threaded yoke, which can be fitted with a pointer to
indicate the position of the valve.
A flexible plate diaphragm is held between two separate casings that are bolted together
to form two airtight chambers. There are two independent air supplies to the chambers
through which compressed air either can be introduced or can escape. As air is introduced
into one chamber, this causes the diaphragm to flex and this motion either extends or re-
tracts the actuator stem, which is attached to the closure element. The more air that is
supplied the greater the movement of the stem.
In the single-acting design a spring is introduced to the opposite air chamber to assist
the return of the diaphragm when the air pressure is lessened. This spring also acts as a
fail-safe mechanism allowing the valve to return to an open or a closed position, when
there is a power failure.
Direct acting is where the air pressure introduced forces the diaphragm and the attached
valve stem down to the required location. Air failure will cause the stem to be retracted out
of the valve body and open the valve, and make sure it remains open.
Reverse acting is where the air pressure introduced forces the diaphragm and attached
valve stem up to the required position. Any loss of air pressure will allow the stem to move
to the extended position and close the valve. The valve will remain closed until the air
supply is reintroduced. Direct acting for rotary valves is where the air pressure introduced
forces the diaphragm and attached valve stem down and, according to its orientation, the
valve will be either opened or closed. Diaphragm valves are relatively cheap, of a basic
construction, and easy to maintain. They are very well-suited for low thrust requirements
in the lower-pressure piping classes.
Disadvantages of the diaphragm actuators of the linear design are that they can create
height problems on smaller valves in restricted locations and on larger sized valves casings
can become excessive in weight.
Piston type Although not initially the first choice of pneumatic design, the piston ac-
tuator is becoming more commonly used as process control becomes more sophisticated.
This design consists of a cylinder that contains a sliding piston, which is sealed with elas-
tomer O-rings. The actuator cylinder is attached to the valve by a cast yoke that allows for
sufficient room on the stem to accommodate positioners and other accessories.
Piston actuators are usually double acting with an air supply to both chambers and a
positioner to take the signal and supply or bleed air from the desired air chambers. Pistons
generate a linear force that either can be transmitted directly to the actuator stem or has to
be converted to a rotary force for use on part-turn valves. As most of the part-turn valves
require only 90 movement, this can be achieved by one of the following.
1. A scotch yoke actuator (Figures 8-2a and 8-2b) consists of a piston, a connecting
shaft, and a rotary pin.
2. A rack and pinion actuator (Figure 8-3) consists of a single or a double piston that is
coupled with an integral rack that drives the pinion.
3. A trunnion/lever arm actuator normally comprises a simple trunnion mounted cylinder

with a piston that acts directly upon a lever that is attached to the valve shaft.
4. A cam actuator consists of double pistons connected by bars with a cam and shaft
between them.
ACTUATOR 115
Advantages
1. Compressed air is convenient and relatively cheap to supply.
2. A piston is not self-locking and it can be used against a spring so it can be used for
single shot emergency operation.
3. When air is in contact with heat, an increase in pressure caused by thermal expansion
will assist the operation of the actuator.
4. Surplus air can be safely vented to the atmosphere.
5. The mechanical design makes it relatively cheaper than alternatives and easier to
maintain.
Disadvantages
1. Because there is no kinetic energy it is not possible to deliver an initial hammerblow
to unseat a wedge valve.
2. Air is a compressible medium and its ability to maintain a fix position is limited. This
may result in drifting with both gate and butterfly valves.
3. The design of the pneumatic actuator and the fact that it requires soft sealing rings
means that there are temperature limitations and the unit must be equipped with ade-
quate thermal insulation if there is a possibility that the actuator has to function during
a fire.
4. Air motors are susceptible to seizure because of ingress of foreign material, or internal
corrosion if the compressed air has not been sufficiently dried. This will greatly affect
the performance of the actuator.
5. The consumption of air for pneumatic motors will be high and this demand is not
always possible for the larger valves requiring higher torques.
10.1.2 Electric actuators

Modern electric motor powered actuators allow local, remote, and hand operation and they
are available for a variety of types of valves and sizes. They comprise a revisable electric
motor, control box, gearbox, and limit switches. The main advantages of electric actuators
are the smooth operation, stability, very high torque values, and maintained thrust, because
they are not subject to the problems of a compressible fluid.
Motors for multi- and part-turn valves can be either single phase or dc power supplies.
However, the three-phase motor is the most commonly used unit, because it is robust and
easily controlled.
It is imperative that the electric motors used for actuators that will be installed out-
side and thus exposed to environmental conditions are constructed in such a way that no
moisture or condensation can accumulate internally.
Electric motors will supply high speeds that will require a large reduction gear ratio
to convert to a lower, more realistic speed for valve operation. For part-turn valves, even
lower speeds are required, which will require a further reduction.
In the case of electric motor failure, provision can be made for manual handwheel op-
eration. This handwheel is usually engaged by using a lever clutch that will automatically
return to motor driven when the power supply is available.
116 CONTROL VALVE
Advantages
1. Electric motors run at high speeds that can be geared down to suit the required oper-
ating torques.
2. Owing to valve gearing, position stability can be maintained.
3. Electric motors can be integrated with process control systems.
4. Electric motors that have been adequately protected from the environment are ”clean”
and do not generate dirt or moisture.
5. Valves can be easily ”hooked up” to the electric power supply.
6. No surplus energy to be dispersed.
7. The electric supply generates a stiff stroke that does not fluctuate.
8. Compact in design.
Disadvantages
1. Expensive.
2. Complex in design and requires higher level of maintenance.
3. Stay put mode when there is a power failure.
4. Electric cables must be protected in a high temperature environment.
5. Electric power is not easily stored.
6. Electrical equipment must be protected from moisture.
7. Electric motors must be intrinsically safe in hazardous areas.
10.1.3 Hydraulic actuators

Hydraulic actuators are capable of delivering very high torques, and with it the fast stroking
speed necessary to operate larger valves. They offer stiffness in the stroke, because they
are using what is considered to be a non-compressible liquid. Hydraulic actuators are used
for these characteristics when, for certain reasons, pneumatic or electric motors are either
not suitable or not available.
Basically there are four designs for linear, part-turn, and multi-turn valves:
1. Rack and pinion
2. Hydraulic motor, worm, and gear
3. Rotary piston
4. Piston and cylinder
Hydraulic actuators are available as double acting or spring return, similar to pneumatic
actuators and the pressure is supplied by means of a hydraulic pump. This power source can
be supplied either locally by a power pack, or remotely with the hydraulic fluid transferred
POSITIONER 117
through small bore tubing. The pressure of the hydraulic fluid can be varied to supply
different torque outputs. This allows for more flexibility of energy sources and subsequent
outputs than is available with both electric and pneumatic actuators.
The energy output of double-acting actuators is constant, however with the spring return
type there is a loss of energy during the compression of the spring.
Advantages
1. Higher operating forces can be achieved.
2. Hydraulic pressure can be stored, either locally or remotely.
3. The hydraulic fluid is considered to be incompressible and this results in stability

during operation.
4. The hydraulic actuator is compact in size and it can therefore be coated and insulated
economically.
Disadvantages
1. No kinetic energy for hammerblow unseating valves.
2. The cost for providing a hydraulic tubing system to supply a network could be eco-
nomically prohibitive.
3. Remote locations will require long tubing runs and pressure drops that in some cases
will prohibit the use of hydraulic powered actuators.
4. Components for hydraulic actuators require close machining tolerances to maintain

effective seal characteristics. This means that in the event of fire there is a high prob-
ability that these actuators will not operate as required.
5. The hydraulic fluid supply has a relatively high thermal expansion rate and the tubing
system will have to be protected from possible fire. Also because of possible thermal
expansion, it may be necessary to install a relief device to combat over-pressurization.
10.2 Positioner
A position controller (servomechanism) that is mechanically connected to a moving part

of a final control element or its actuator and that automatically adjusts its output to the
actuator to maintain a desired position in proportion to the input signal.
Pneumatically operated valves depend on a positioner to take an input signal from a
process controller and convert it to valve travel. These instruments are available in three
configurations:
1. Pneumatic Positioners: A pneumatic signal (usually 3-15 psig) is supplied to the po-
sitioner. The positioner translates this to a required valve position and supplies the
valve actuator with the required air pressure to move the valve to the correct position.
2. Analog I/P PositionerThis positioner performs the same function as the one above,
but uses electrical current (usually 4-20 mA) instead of air as the input signal.
118 CONTROL VALVE
3. Digital ControllerAlthough this instrument functions very much as the Analog I/P
described above, it differs in that the electronic signal conversion is digital rather than
analog. The digital products cover three categories.
Digital Non-Communicating: A current signal (4-20 mA) is supplied to the po-
sitioner, which both powers the electronics and controls the output.
HART: This is the same as the digital non-communicating but is also capable of
two-way digital communication over the same wires used for the analog signal.
Fieldbus: This type receives digitally based signals and positions the valve using
digital electronic circuitry coupled to mechanical components. An all-digital control
signal replaces the analog control signal. Additionally, two-way digital communica-
tion is possible over the same wires. Fieldbus technologies benefit the end user by
enabling improved control architecture, product capability and reduced wiring.
10.3 Cage
A part of a valve trim that surrounds the closure member and can provide flow characteri-
zation and/or a seating surface. It also provides stability, guiding, balance, and alignment,
and facilitates assembly of other parts of the valve trim. The walls of the cage contain
openings that usually determine the flow characteristic of the control valve. Various cage
styles are shown in Fig.10.3.
Figure 10.3 Characterized Cages for Globe-Style Valve Bodies
10.4 Flow Characteristics and Valve Selection
The flow characteristic of a control valve is the relationship between the flow rate through
the valve and the valve travel as the travel is varied from 0 to 100%. Inherent flow char-
acteristic refers to the characteristic observed during flow with a constant pressure drop
across the valve. Installed flow characteristic refers to the characteristic obtained in ser-
vice when the pressure drop varies with flow and other changes in the system.
Fig.10.4 illustrates typical flow-characteristic curves. The quick-opening flow charac-
teristic provides for maximum change in flow rate at low valve travel with a fairly linear
CONTROL VALVE SIZING 119
Figure 10.4 Control valve Characteristics
relationship. Additional increases in valve travel give sharply reduced changes in flow rate.
When the valve plug nears the wide open position, the change in flow rate approaches zero.
In a control valve, the quick-opening valve plug is used primarily for on-off service; how-
ever, it is also suitable for many applications where a linear valve plug would normally be
specified.
In the equal-percentage flow characteristic, equal increments of valve travel produce
equal percentage changes in the existing flow. The change in flow rate is always propor-
tional to the flow rate just before the change in position is made for a valve plug, disc, or
ball position. When the valve plug, disc, or ball is near its seat and the flow is small, the
change in flow rate will be small; with a large flow, the change in flow rate will be large.
Valves with an equal-percentage flow characteristic are generally used for pressure control
applications. They are also used for other applications where a large percentage of the total
system pressure drop is normally absorbed by the system itself, with only a relatively small
percentage by the control valve. Valves with an equal-percentage characteristic should also
be considered where highly varying pressure drop conditions can be expected.
The modified parabolic-flow characteristic curve falls between the linear and the equal-
percentage curve.
Note: Where detailed process knowledge is lacking, as a rule of thumb, use equal-
percentage characteristics at 70% opening for the valve sizing.
10.5 Control valve sizing
Control valve is proprietary item and need to sized or select from available sizes rather
than designing it. Process engineer should decide the size of control valve to suit process
conditions.
10.5.1 Gas services

Critical flow limitation is a significant problem with sizing valves for gaseous service.
Critical flow is a choked flow condition caused by increasing gas velocity at the vena-
contracta. The vena-contracta is the point of minimum cross-sectional area of the flow
120 CONTROL VALVE
stream which occurs just downstream of the actual physical restriction. When the velocity
at the vena-contracta reaches sonic velocity, additional increases in pressure drop, ∆P , (by
reducing downstream pressure) produce no increase in flow.
In the ISA sizing procedure critical flow limitations are addressed by calculating , the
expansion factor, for utilization within the actual sizing equation.
X
Y = 1−
3Fk Xc
where, Fk = k/1.4 = ratio of specific heats factor, dimensionless

k = ratio of specific heats, dimensionless
X = ratio of pressure drop to absolute inlet pressure (∆P/P1 ), dimensionless
Xc = pressure drop ratio for the subject valve at critical flow, withFk = 1.0
dimensionless
Critical pressure drop, and thus critical flow, is realized when X ≥ Fk Xc . Therefore,
since the flow can’t exceed that produced at the critical pressure drop the value of in the
following sizing equations should never be less than 0.67.
X 1
Y =1− = 1 − = 0.67
3Fk Xc 3
Likewise the value of X in the equations should never exceed Fk Xc .
Sizing Calculation Procedure:

1. Select the appropriate sizing equation based on the stated inlet conditions and units of
measurement. The limitations on and X as discussed above must be observed in all
the sizing equations.
2. Calculate an initial, approximate required Cv based on an assumed Rated Pressure
Drop Ratio Factor, Xc Initial assumptions for the value of Xc can be based on the
general style of valve.
3. From the valve manufacturer’s sizing data select a specific valve type and size such
that the listed Cv is equal to or greater than the calculated Cv . The Xc associated with
the listed Cv should then be used in the chosen sizing equation to calculate a revised,
required Cv . This iteration process continues until the calculated Cv and equals the
manufacturer’s listed Cv .
4. For a new valve selection a valve size is typically chosen such that the maximum,
calculated Cv is close to 75% to 85% of valve travel. This allows for process variabil-
ity while maintaining flow capability. The minimum, calculated Cv should typically
occur at or about 10% of valve travel.
5. Fp is the Piping Geometry Factor. It corrects the sizing equations for the effects of
fittings such as reducers and expanders that are attached to the valve body ends. Fp
values can be determined via test or calculated per the ANSI/ISA S75.01 standard. If
the valve has no such fittings attached, e.g., the nominal value size and nominal pipe
size are the same, then Fp = 1.0.
CONTROL VALVE SIZING 121
10.5.2 Liquid Service

The procedure used to size control valves for liquid service should consider the possibility
of cavitation and flashing since they can limit the capacity and produce physical damage to
the valve. In order to understand the problems more thoroughly, a brief discussion of the
cavitation and flashing process is presented below.
Cavitation: In a control valve, the fluid stream is accelerated as it flows through the
restricted area of the orifice, reaching maximum velocity at the vena contracta. Simultane-
ously, as the velocity increases, an interchange of energy between the velocity and pressure
heads forces a reduction in the pressure.
If the velocity increases sufficiently, the pressure at the vena-contracta will be reduced
to the vapour pressure of the liquid. At this point, vapour cavities or bubbles, the first stage
in cavitation, appear in the fluid stream. Downstream from the vena-contracta, the fluid
stream undergoes a deceleration process resulting in a reversal of the energy interchange
which raises the pressure above the liquid vapour pressure.
The vapour cavities, or bubbles, cannot exist at the increased pressure and are forced to
collapse or implode. These implosions are the final stage in the cavitation process. They
potentially produce noise, vibration, physical damage, and other performance problems.
In order to avoid cavitation completely, the pressures at all points within the valve must re-
main above the vapour pressure of the liquid. Cavitation can occur as the result of changes
in the mean pressures through the valve, but also from localized changes due to flow sep-
arations and other local disturbances that are not indicated by examining just the mean
inlet, vena-contracta, and outlet pressures. Determining when a problem-causing level of
cavitation is present represents a considerable challenge.
p
Non-vaporizing Mass Flow w = 2.73Fp Cv (P1 − P2 ) γ1
r
P1 − P2
Non-vaporizing Volumetric Flow q = 0.0865Fp Cv
Gf
p
Vaporizing Mass Flow w = 2.73FL Cv (P1 − FF Pv ) γ1
r
P1 − FF Pv
Vaporizing Volumetric Flow q = 0.0865FL Cv
Gf
where, Fp = piping geometry factor, dimensionless

P1 = upstream absolute static pressure, measured two nominal pipe diameters
upstream of valve fitting assembly, kPa
P2 = downstream absolute static pressure, measured six nominal pipe diameters
downstream of valve fitting assembly, kPa
FL = liquid pressure recovery factor of a valve without attached fittings,
dimensionless
FF = liquid critical pressure ratio factor, dimensionless
Pv = vapour pressure of liquid at valve inlet temperature, kPa
γ1 = Specific Weight, dimensionless
Gf = Specific Gravity, dimensionless
Table 10.1 Equations for Control Valve Sizing for liquid service
122 CONTROL VALVE
Sizing Calculation Procedure:

1. Select the appropriate sizing equation from Table 10.1 based on the stated inlet con-
ditions and units of measurement.
2. Calculate the Cv required using the equation for non-vaporizing flow.
3. Calculate the Cv using the equation for vaporizing flow. An initial assumed value
of FL can be taken from the manufacturer’s literature. FF the liquid critical pressure
ratio factor, can be found from Fig. 10.5 based on the critical pressure and inlet vapour
pressure for subject liquid. The user must at this point iterate through this calculation
accounting for the variation in FL and valve-rated Cv due to valve style, size, trim,
flow direction, etc.
4. Select the higher of the two calculated Cv ’s as the required Cv .
5. From the valve manufacturer’s sizing data, select a specific valve type and size such
that the listed Cv is equal to or greater than the calculated Cv .
6. See Cavitation and consult the manufacturer’s data for appropriate valve cavitation
operating limits.
Figure 10.5 Critical Pressure Ratios for All Liquids, FF
*****************
CHAPTER 11
ADVANCED CONTROL STRATEGIES
Along with controller tuning, arrangement of controller with process also change the be-
haviour of process. It is very important to establish proper system between measured and
manipulated variables. Different process behave differently as the model governing the
process is different. Therefore it is necessary to analyse the process model before using
particular control strategy.
11.1 Feedforward Control
In feedforward control disturbances are detected and eliminated before it destabilise the
system. Referring to the Fig. 11.1, if ‘D’ is disturbance to be entered in the system and
‘Ysp ’ is set point to the controller then the correction ‘U’ enter in the system at same time
‘D’ enter, which counters the destabilisation effect on system. By this way virtually system
remains always at stable condition. [6]
This is also termed predictive control and can be usefully employed where there are
significant time lags between a load change and its subsequent effect upon a system output
variable.
Feedforward control works exactly opposite to feedback control, where there is al-
lowance of error to occur. Lets consider an example of controller over a boiler. Objec-
tive of control action is to control the level in boiler drum. This can be achieved by using
feedback loop for level control as shown in Fig 11.2.
124 ADVANCED CONTROL STRATEGIES
Figure 11.1 Feedforward and feed back conrol
Figure 11.2 Level control (Feedback) over boiler drum
But the main objective from boiler is to get steam at desired flow rate which can be
controlled using above discussed control mechanism. Therefore a feedforward controller
(see Fig. 11.3)can be placed across boiler drum which measure the flow rate of steam and
manipulates the flow rate of feed water which automatically controls the level in the drum.
Water level in boiler drum is generally not a main objective of the control but it is very
crucial in safety point of view. Drop in level may expose the heating tubes which cause
overheating of tubes. Increase in level floods the boiler and ultimately ends up with no
steam flow. Therefore objective of control changes to multi-variable. This situation can be
handle easily by using combination of feedback and feedforward control as shown in Fig.
11.4.
11.1.1 Comparison of Feedback and Feedforward Control

11.1.1.1 Feedback (FB) Control
Advantages:
Corrective action occurs regardless of the source and type of disturbances.
Requires little knowledge about the process (For example, a process model is not
necessary).
FEEDFORWARD CONTROL 125
Figure 11.3 Steam flow (Feedforward) control across the boiler.
Figure 11.4 Feedforward-Feedback control for boiler drum level
Versatile and robust (Conditions change? May have to re-tune controller).
Disadvantages:
FB control takes no corrective action until a deviation in the controlled variable oc-
curs.
FB control is incapable of correcting a deviation from set point at the time of its
detection.
Theoretically not capable of achieving perfect control.
For frequent and severe disturbances, process may not settle out.
11.1.1.2 Feedforward (FF) Control
Advantages:
Takes corrective action before the process is upset (cf. FB control.)
Theoretically capable of ”perfect control”
Does not affect system stability
Disadvantages:
Disturbance must be measured (capital, operating costs)
Requires more knowledge of the process to be controlled (process model)
Ideal controllers that result in ”perfect control: may be physically unrealizable. Use
practical controllers such as lead-lag units
11.2 Ratio Control
As the name implies, ratio control involves keeping constant the ratio of two or more flow
rates. The flow rate of the ”wild” or uncontrolled stream is measured, and the flow rate of
the manipulated stream is changed to keep the two streams at a constant ratio with each
other. Common examples include holding a constant reflux ratio on a distillation column,
keeping stoichiometric amounts of two reactants being fed into a reactor, and purging off
a fixed percentage of the feed stream to a unit. Ratio control is often part of a feedforward
control structure.[4]
Ratio control is achieved by two alternative schemes, shown in Fig. 11.5 and 11.6. In
the scheme shown in Fig. 11.5, the two flow rates are measured and their ratio is computed
(by the divider). This computed ratio signal is fed into a conventional PI controller as
the process variable (PV) signal. The set point of the ratio controller is the desired ratio.
The output of the controller goes to the valve on the manipulated variable stream, which
changes its flow rate in the correct direction to hold the ratio of the two flows constant.
This computed ratio signal can also be used to trigger an alarm or an interlock.
In the scheme shown in Fig. 11.6, the wild flow is measured and this flow signal is
multiplied by a constant, which is the desired ratio. The output of the multiplier is the set
point of a remote-set how controller on the manipulated variable.
If orifice plates are used as flow sensors, the signals from the differential-pressure trans-
mitters are really the squares of the flow rates, Some instrument engineers prefer to put in
square-root extractors and convert everything to linear flow signals.
11.2.1 Applications of Ratio control

Keep a constant ration between the feed flow rate and the steam in the reboiler of a
distillation column.
Hold constant reflux ratio in a distillation column.

CASCADE CONTROL 127
Figure 11.5 Ratio control – Method I Figure 11.6 Ratio control – Method II
Control the ratio of two reactants entering a reactor at a desired value.

Hold the ratio of two blended streams constant, in order to maintain the composition
of the blend at the desired value.
Hold the ratio of purge stream to the recycle stream constant.
Keep the ratio of fuel/air in a burner at its optimum value.
Maintain the ratio of the liquid flow rate to vapour flow rate in an absorber constant,
in order to achieve the desired composition in the exit vapour stream.
11.3 Cascade control
One of the most useful concepts in advanced control is cascade control. A cascade control
structure has two feedback controllers, with the output of the primary (or master) con-
troller changing the set point of the secondary (or slave) controller. [4] The output of the
secondary goes to the valve, as shown in Fig. 11.8. There are two purposes for cascade
control:
1. to eliminate the effects of some disturbances, and
2. to improve the dynamic performance of the control loop.
To illustrate the disturbance rejection effect, consider the distillation column reboiler
shown in Fig. 11.7. Suppose the steam supply pressure increases. The pressure drop over
the control valve will bc larger, so the steam flow rate will increase. With the single-loop
tcmpcrafure controller, no correction will be made until the higher steam flow rate increases
the vapor boilup and the higher vapor rate begins to raise the temperature on tray 5. Thus,
the whole system is disturbed by a supply steam pressure change. With the cascade control
system, the steam flow controller will immediately see the increase in steam flow and will
Figure 11.7 Distillation column reboiler Figure 11.8 Distillation column reboiler
temperature control - Conventional temperature control - Cascade
pinch back on the steam valve to return the steam flow rate to its setpoint. Thus, the reboiler
and the column are only slightly affected by the steam supply pressure disturbance.
Fig. 11.9 shows another common system where cascade control is used. The reactor
temperature controller is the primary controller; the jacket temperature controller is the
secondary controller. The reactor temperature control is isolated by the cascade system
from disturbances in cooling-water inlet temperature and supply pressure.
Figure 11.9 Controllers for reactor
This system is also a good illustration of the improvement in dynamic performance that
cascade control can provide in some systems. The closed loop time constant of the reactor
temperature will be smaller when the cascade system is used than when the reactor temper-
ature sets the cooling water make-up valve directly. Therefore, performance is improved
by using cascade control.
There are two types of cascade control: series cascade and parallel cascade. The two
examples just discussed are both series cascade systems because the manipulated vari-
APPLICATION OF FEEDFORWARD, FEEDBACK AND CASCADE CONTROL 129
able affects the secondary controlled variable, and then the secondary variable affects the
primary variable. In a parallel cascade system the manipulated variable affects both the
primary and the secondary controlled variables directly. Thus, the two processes are ba-
sically different and result in different dynamic characteristics. We quantify these ideas
later.
11.4 Application of Feedforward, Feedback and Cascade control
A typical example of feedback, feedforward and cascade control is the response of com-
position of the overhead product stream of distillation process to a perturbation in feed
composition, and column temperature. [6]
A typical arrangement is illustrated in Fig. 11.10 where the variation in feed composi-
tion are measured by a suitable composition analyser. The signal from the analyser is fed
directly to the feedforward controller, the output of which is cascaded on to the set point
of the reflux flow controller. Imperfection in feedforward control can often be overcome
by the addition of suitable feedback control loop into action. The reflux flow shown in Fig.
11.10 on the flow control in cascade with the boiling temperature of the liquid at an ap-
propriate point within the column. The inner (or slave) flow controller maintains a steady
reflux flowrate until a change occurs in boiling temperature or in feed composition – either
of which will vary the set point of the flow controller as required.
TC
CC Σ FC
Figure 11.10 Feedforward, feedback and cascade control for distillation column
11.5 Control strategy design at basic control level
The basic control layer is dominated by feed-back loops, eventually provided with feed-
forward actions to cope with disturbances.[3] The basic control layer will include some
cascaded control loops and simple models, such as for de-coupling of interactions, and
heat and mass balance control options. The demand on control, next to hands-off control,
is higher than in the past due to:
more stringent requirements on product quality;
elimination of intermediate storage as lot tanks, check tanks/hoppers and minimiza-

tion of storage, (implementation of JIP, TQC, and FPPP);
high level of process integration;

switchability of process conditions for campaign operations (switchability is defined

as the ease with which the process can be moved from one stationary point to another);
predictive alarming and interlocking actions need to be incorporated to assure safe

operations;
self-optimizing control of units applied for; stand-alone units and at intermediate
time spans between implementation of closed loop steady state operation optimization
set points; and
bases for model predictive control and closed loop process optimization.
To achieve robustness it is vital that any basic model applied for control is preferably
solved by an analytical technique. Iterations and optimization are to be avoided, as these
might fail in finding a solution unless sufficient provisions are made.
Control design requires a controllability analysis to select the most robust control solu-
tion. The result of the analysis may include the need for process modifications to enable
control, reflecting interaction between process design and control design.
A procedure for controllability analysis of a process plant has the following sequential
steps:
1. Define control objectives.
2. Evaluate the open loop stability of the process based on static models.
3. Divide the process into separate process sections; be aware that a process section
must include a recycle stream from another section, as applied for reactors with un-
converted reactants.
4. Determine the degrees of freedom (DOFs).
5. Determine preliminarily the controlled, manipulated, measured and disturbance vari-

ables.
6. Determine feasible pairing options in respect of plant wide control and unit control.
7. Evaluate static interaction of the selected pairing options.
8. Evaluate dynamic interaction of the reduced set of selected pairings. In case of eval-
uation of the controllability during process synthesis this information is input for the
final flowsheet selection.
9. Establish the final pairing and design the controllers.
10. Tune and test the performance of the controller in a dynamic simulation.
11.6 Determine the different Variables
Manipulated and controlled variables are respectively input and output variables which can
be selected by qualitative criteria as formulated.[3]
DETERMINE THE DIFFERENT VARIABLES 131
11.6.1 Controlled variable (CVs)

Selection might preferably follow the guidelines:
Select variables which are not self-regulating; variables which result in a new steady-
state situation without a feed-back control should not be selected. An example of a
self-regulating system is an hydraulic overflow system. As is often applied in decanter
systems, this is termed a simple self-regulating system.
Select variables which exceed the constraint of the process. Safety constraints, as well
the tendency to operate close to its constraints under optimized conditions, requires
that controlled variables are selected which exceed the con- strained values.
Choose variables which directly measure the condition and quality of the process
streams, or that have a strong relation to it. As composition measurements are ex-
pensive and often have long response times, the preference is to select variables that
are easy to measure with short response times, but which have a direct relation to the
property of interest.
Selection of output variables which have a significant interaction with other controlled
variables are preferred. Closed loop performance is improved by stabilizing control
variables which interact significantly.
Select output variables which have favourable static and dynamic responses compared
to other controlled variables.
Variables selected for control should not correlate too closely.
11.6.2 Manipulated variables (MVs)

The selection of manipulated variables (MVs) must follow certain guidelines:
MVs must affect the CV strongly, which means that it requires a large steady- state
gain. This results in rapid responses and less variations in flow of the manipulated
stream, and less potential impact on other units.
Operational range of the manipulated variable is a characteristic which need to be

respected. A bypass valve around a heat exchanger might give a limited controllability
of the exchanger duty.
MVs are preferred which have a direct effect on the CV. A direct effect on a CV results
in fast responses, which are preferred. The removal of inert from a condenser works
rapidly and more effectively than an increase in the cooling flow.
Avoidance of recycling disturbances. Disturbances are at best prevented, but when

they are introduced they should leave the process through its exit streams. When
disturbances are recycled or feed streams are manipulated, they have a larger impact
(see disturbance variables).
11.6.3 Measured variables

Measured variables are to be selected by the following guidelines:
Measured variables must reflect the CVs. Preferably, do not measure the volumetric
flow when you want to control the mass flow. Temperature measurements represent-
ing dual composition in a distillation column require pressure correction.
Measurement should be reliable, robust, and accurate, and have a sufficient opera-
tional range.
Location of the measurement should be sensitive to a disturbance. Pressure mea-

surements should be at the compressor, and not at downstream or upstream units.
Temperature measurements should be installed where variations are observed at first,
such as in a steep temperature transient in a distillation column, or at the front end of
a plug-flow reactor.
Measurements should minimize time delays and time constants. This is determined
by the type of instrument, the location, and any sampling system.
11.6.4 Disturbance variables (DVs)

Disturbance variables (DVs) are selected for testing the robustness of the designed control
configuration, but they must be minimized by appropriate action. Disturbances can be split
in internal disturbances and external disturbances:
External disturbances are:
Weather conditions, as day and night temperature cycle, humidity variations for cool-
ing water systems, rain showers. Atmospheric conditions changes over a longer time
period (daily or weekly or even seasonally are not experienced as a disturbance) are
not experienced as a disturbance for control of the process system which mostly have
hourly or shorter response times.
Feed conditions and compositions.
Utility conditions.
Internal disturbances are:
Economical disturbances which determine capacity, product distribution, internal qual-

ity set points, equipment scheduling. These are grouped under internal disturbances
as the implementation of modified conditions is under operational control.
Operational disturbances are caused by operational activities; for example, when units
or components are taken in or out of operation for cleaning, regeneration, or repair.
Even if these activities are automated, they are still to be recognized as a disturbance.
Control disturbances which cause propagation or even amplification of a disturbance

through the process system, particular so for recycling of disturbances. These are
negatively influenced by; inappropriate control design, and wrong tuning of control
loops.
The impact of both types of disturbances can be handled in different ways they can be
categorized as:
Prevention or limitation of a disturbance is a preferred action, examples are:

PLANTWIDE CONTROL DESIGN PROCEDURE 133
– Design internal utility grids less sensitive to external (site) variations by operating
the internal system with a pressure controller slightly above or below the external
grid level, depending on an import or export situation.
– Inventory with mixing devices will minimize composition or temperature varia-
tions.
– Smoothly ramping of set point changes particular for flow, pressure and tempera-
ture and economically determined set points
– Gradually and smoothly taking units in or out of operation
– Design for back-up supplies like is applicable for heat-integrated systems.
– Correct for any concentration variation of a feed stream. For example hydro-
genation reactors often need to control the mass flow of hydrogen, this means
correction for any impurities is required.
Rejection of disturbance to outside the process is easily applied for utility systems
and inventory feed systems. Examples are:
– Any deviation in utility consumption such as steam is exported to external sup-

plies. Site utility systems are normally designed with short response times to
enable fast compensation of demand variations.
– Reactors experience changes in conversion which have an impact on recycle flows,
and as such varies the demand on feed of reactants. These variations can be ab-
sorbed at the supplier, or in the inventory.
Absorption of disturbances depends on process design and control design. Examples

are:
– Inventory in reactors, column bottoms, reflux drums and surge drums to smooth
variations in flow andjor composition.
– De-coupling of interactions to direct a disturbance outside the process and avoid
recycling or transmitting to other process sections.
These three methods for disturbance handling should be recognized and exploited in
sequential order of the above discussion and in full understanding between process and
control designer. Design modification might be essential to limit the effect of disturbances.
11.7 Plantwide control design procedure
So far in this chapter we have considered some specific features and problems concerning
plantwide control. In this section we present a design procedure that can be used to gener-
ate an effective plantwide control structure. The number of variables that can be controlled
in any plant equals the number of control valves. Most of these “control degrees of free-
dom” must be used to set production rate, control product quality, account for safety and
environmental constraints, control liquid levels, and control gas pressures. Any remaining
degrees of freedom can be used to achieve economic or dynamic objectives.[4]
The method consists of six basic steps. The steps may require some iteration through
the procedure.
1. Count the number of control valves (make sure all are legitimate, i.e., only one valve
in a liquid-filled line). This is the number of control degrees of,freedom.
2. Determine what valve will be used to set production rate. Often this decision is estab-
lished by the design basis of the plant. Production rate is fixed by a feed sent to the
plant. Production rate is fixed by the flow rate of a product stream if a downstream
process can demand an arbitrary flow rate. If neither of these design requirements is
specified, we are free to select the valve that provides smooth and stable production
rate transitions and rejects disturbances. This may be the flow rate of the feed stream
to a separation section, the flow rate of a recycle stream, the flow rate of a catalyst to
a reactor, or a reactor heat removal rate. Controller set points of reactor temperature,
level, or pressure can also be used to change production rate. Dynamic simulations of
alternatives are required to select the best structure.
3. Select the “best” manipulated variables to control each of the product quality vari-
ables and the safety and. environmental variables. These manipulated variables are
selected to give the tightest possible control of these important variables. The dynamic
relationships between controlled and manipulated variables should feature small time
constants and deadtimes, sufficiently large gains, and wide rangeability of manipu-
lated variables. These decisions must consider some other factors, such as the mag-
nitudes of flow rates. For example, in a high-reflux ratio column, the distillate flow
rate should be used to control product quality because the reflux flow rate must be
used to control reflux drum liquid level. Another common example is the control of
temperature and base level in a distillation column with a very small bottoms flow
rate. In most columns, temperature should be controlled by reboiler heat input, and
base level controlled by bottoms flow rate. When the bottoms flow rate is less than
about 20 percent of the rate of liquid entering the column base, these loops should be
reversed (even though this results in “nested” loops).
4. Determine the valves to use for inventory control: all liquid levels (except for surge
volumes in liquid recycle systems) and gas pressures must be controlled. Select the
largest stream to control levels whenever possible. Use proportional only level control
in non-reactive level loops for cascaded units. Fresh feed make-up streams are often-
used to hold levels or pressures when these variables reflect the inventory of specific
components in the process. There should be a flow controller somewhere in all liquid
recycle loops.
5. Make sure that the overall component balances for all chemical components can be
satisfied. Light, heavy, and intermediate inert components must have a way to exit
the system. Reactant components must be consumed in the reaction section or leave
the system as impurities in product streams. Therefore, either reaction rates (temper-
ature, pressure, catalyst addition rate, etc.) must be changed or the flow rates of the
fresh feed make-up streams must be manipulated somehow. Make-ups can be used to
control compositions in the reactor or in recycle streams, or to control inventories that
reflect the amount of the specific components contained in the process. For example,
bring in a gaseous fresh feed to hold the pressure somewhere in the system, or bring
in a liquid fresh feed to hold the level in a reflux drum or column base where the
component is in fairly high concentration (typically in a recycle stream).
6. Use the remaining control valves for either steady-state optimization (minimize en-
ergy, maximize yield, etc.) or to improve dynamic controllability. A common ex-
PLANTWIDE CONTROL DESIGN PROCEDURE 135
ample is controlling purities of recycle streams. Even though these streams .are not
products and do not have any quality specifications, the steady-state and dynamic abil-
ity of the process to handle load and production rate changes is sometimes improved
by controlling recycle purities (perhaps not at precise set points, but sufficiently to
prevent excessive build-up of impurities in the recycles).
EXAMPLE 11.1
MPN is produced by reacting methanol with excess (2 mol%) ACN in isothermal

reactor. It has been found that conversion is 100% with respect to methanol.There is no
purge stream, so excess ACN appears in the MPN product stream. ACN unreacted is
separated from product in distillation column as distillate and recycled back to reactor
along with fresh ACN, as shown in Fig. 11.11. Propose a control system for this plant.
F Cooling
water
ACN
P
F
F
L
F
MeOH
T L
Cooling T
Cooling
water water
return L
supply
Steam
F
MPN product
Figure 11.11 Process Flowsheet for MPN production
Solution: There many be many solutions to this problem, here we will discuss two
options to understand the approach.
Option A:
One possible solution is shown in Fig. 11.12.
1. The production rate objective should be satisfied by the methanol flow rate controller,
since there is essentially total conversion of the methanol to MPN.
2. The excess ACN is obtained by specifying the ACN flow rate to be 2% higher than
the methanol flow rate.
3. Exothermic reactor temperature must be controlled to assure stable reactor operation.
Here, the reactor temperature must be controlled by manipulating cooling water valve.
4. The reactor level (related to residence time) is controlled by manipulating the reactor
outlet flow valve. Notice that there are now only five more loops that can be speci-
fied, since there are a total of nine control valves. The final five loops are associated
with the distillation column; three of these are associated with inventory variables
(distillate receiver level, bottoms level, column pressure).
5. A temperature in the bottom of the column manipulates the steam control valve. Be-
cause of equilibrium relationships, the temperature is related to the product purity.
6. The reflux stream is simply flow controlled. An alternative would be to control a tray
temperature in the top of the column; this would be related to controlling the distil-
late purity. Since the distillate stream is recycled to the reactor, it is not considered
important to maintain the stream composition at a given value.
7. The distillate receiver level is controlled by manipulating the distillate (or recycle)
flow valve.
8. The column pressure is controlled by manipulating the condensate flow. This es-
sentially changes the level of condensate in the overhead condenser, changing the
effective heat transfer area, and therefore the condenser duty.
9. The column bottom level is maintained by manipulating the bottom control valve.
FC
F Cooling
water
ACN PC
P
FC F
FC
F
L LC
F
MeOH
T L LC
TC F
Cooling T TC
Cooling
water water
return L
supply
Steam
LC
F
MPN product
Figure 11.12 Control Option A: for Process Flowsheet of MPN production
The main problem with this solution is that the control structure exhibits a high
sensitivity in the recycle flow rate to small disturbances. This so-called snowball effect.
Luyben et. al. [4] have recommended that at least one control loop in the recycle flow
path be flow controlled in order to minimize the snowball effect.
SOME GUIDELINES AND RECOMMENDATION FOR CONTROL SYSTEM DESIGN 137
Option B:
An alternative solution is shown in Fig. 11.13. Here a flow controller is used to
regulate the mixed stream (recycle + ACN) flow rate by manipulating the distillate
(recycle) flow rate; this allows the control structure to minimize the snowball effect.
The distillate level controller manipulates the fresh ACN flow control valve. The other
seven loops are identical to the previous solution shown in Fig. 11.12.
FC
F Cooling
water
ACN PC
P
FC F
FC
F
L LC
F
MeOH
T L LC
TC F
Cooling T TC
Cooling
water water
return L
supply
Steam
LC
F
MPN product
Figure 11.13 Control Option B: for Process Flowsheet of MPN production
11.8 Some guidelines and recommendation for control system design
Keep the control system as simple as possible. Everyone involved in the process, from
the operators up to the plant manager, should be able to understand the system, at least
conceptually. Use as few pieces of control hardware as possible. Every additional
gadget included in the system is one more item that can fail or . drift. The instrument
salesperson will never tell you this, of course.
Use feedforward control to compensate for large, frequent, and measurable distur-
bances.
Avoid large time lags and deadtimes in feedback loops. Control is improved by keep-
ing the lags and deadtimes inside the loop as small as possible. This means that
sensors should be located close to where the manipulated variable enters the process.
Use proportional-only level controls where the absolute level is not important (surge
tanks and the base of distillation columns) to smooth out disturbances.
Eliminate minor disturbances by using cascade control systems where possible.

Avoid control loop interaction if possible, but if not, make sure the controllers are
tuned to make the entire system stable. Up to this point we have discussed tuning only
single-input, single-output (SISO) control loops. Many chemical engineering systems
are multivariable and inherently interacting, i.e., one control loop affects other control
loops.
A common way to avoid interaction is to tune one loop very tight and the other loop
loose. The performance of the slow loop is thus sacrificed.
Check the control system for potential dynamic problems during abnormal conditions
or at operating conditions that are not the same as the design. The ability of the
control system to work well over a range of conditions is called flexibility. Startup
and Shutdown situations should also be studied. Operation at low throughputs can
also be a problem. Process gains and time constants can change drastically at low
flow rate, and controller retuning may be required. Installation of dual control valves
(one large and one small) may be required.
Avoid saturation of a manipulated variable. A good example of saturation is the level
control of a reflux drum in a distillation column that has a very high reflux ratio.
Avoid “nesting” control loops. Control loops are nested if the operation of the external
loop depends on the operation of the internal loop.
*****************
CHAPTER 12
MICROPROCESSOR-BASED
CONTROLLER
A microprocessor-based controller is essentially a digital computer programmed to per-

form the function of a process controller. For our purpose, the term microprocessor is
synonymous with computer and we could refer to a microprocessor-based controller as a
computer-based controller.[2]
During the past fifty years tremendous development has occurred in process control
hardware. The three phases of development am pneumatic control, electronic control, and
microprocessor-based control. During the 1940s the predominant controller was pneu-
matic, meaning that signals to and from the controller and within the controller mechanism
were air-pressure signals that usually varied from 3 to 15 psig. The development of the
high-gain operational electronic amplifier during the second world war led to the develop-
ment of the electronic controller and also the analog computer. The electronic controller
mimicked the control functions of the pneumatic controller. It also provided some im-
provements, such as accurate and reproducible control parameter settings and reduction in
size of the instruments. In contrast, the pneumatic controller required frequent calibration
of the knobs used to set the various controller parameters. The pneumatic controller had
interaction among the control modes and had inherent lags that became significant at high-
frequency operation. There were frequent debates over the pros and cons of pneumatic
and electronic controllers. For example, the pneumatic controller was rugged, simple to
install, and required little maintenance. Only a source of air pressure was needed to op-
erate the controller. There was initially great concern about the possibility of explosions
with the use of electronic controllers, so the instrument cases for these controllers were

140 MICROPROCESSOR-BASED CONTROLLER
purged with steady streams of air when used in plants producing flammable substances.
The maintenance of electronic controllers also required highly trained technicians.
In the 1960s, the chemical industry made its first attempt at computer process control.
These control systems used large mainframe computers, for which the control programs
had to be written from scratch. The first attempts at computer control were met with mixed
reactions. In the 1970s, there appeared on the market the first generation of digital control
hardware, which was based on the advances in microprocessor-based technology. This
equipment was user friendly and all the software accompanied the hardware. The operator
did not face the problem of writing computer code to implement the control functions; it
was only necessary to learn the instructions needed to configure (set up) the controllers.
12.1 Hardware
In this case, the control algorithm resides as a computer program in the memory of the
computer. The operator communicates with the control system with a keyboard, a monitor,
and a printer. The computer can perform many memory functions than implementation of
the control algorithm as will be discussed later. The recorder of the pneumatic or electronic
system is replaced by a monitor screen on which the transients are shown.[2]
Microprocessor
Measured based
Transducer Transducer Valve To Process
variable
controller
Printer Monitor Keyboard
Figure 12.1 Microprocessor-based controller
In a modem controller both analog and digital signals are processed. The analog signal
is the type that represents a continuous variable that varies over a range of values. The
digital signal is a binary signal that can be represented by two states (on, off, or logic
1, logic 0, etc.). Examples of analog signals am the measurement from a temperature
transmitter or the signal sent to a valve. Examples of digital signals are the output to a
motor, which causes it to be on or off, or the output to an alarm light causing it to be on or
off.
12.2 Distributed Control System (DCS)
So far we have been concerned in this chapter with the operation of a single controller.
Such a controller is referred to as a stand-alone controller because it is not communicating
with other controllers, but only with the one control loop of which it is a part. Present-day
microcomputer-based control systems have the capability of communicating with other
controllers through a network, which is called distributed control. Figure 12.2 shows one
version of the communication linkages that are usually present in a distributed control
system. Each manufacturer of distributed control systems has a different way of organizing
them.
DISTRIBUTED CONTROL SYSTEM (DCS) 141
Figure 12.2 Typical connections in a distributed control system: CP: control processor, AP:
applications processor, WS: workstation, FBM: fieldbus module.
A distributed control system is intended to be used for a large processing facility that
involves as many as fifty to one hundred loops. Examples include a refinery, a brewery, a
power plant, and the like. In Fig. 12.2 , the modules of control equipment that communi-
cate with each other are as follows. [2]
Control processor (CP)
Applications processor (AP)
Workstation (WS)
Fieldbus module (FBM)
The first three of these modules communicate with each other through a nodebus or
“data highway” as it has been called. The fieldbus modules serve as devices that interface
with transducers and valves in the process.
The control processor contains the blocks described earlier (analog input, analog output,
control, linearisation, etc.) that are connected together by software to provide the control
algorithm required for each loop. Communication between the control processor and the
process (a distance away) in the field takes place in the fieldbus module. Two types of
fieldbus modules are available. One type provides a set of analog inputs and a set of analog
outputs that send to and receive from the field continuous signals (4-20 mA). The other
type of module sends to and receives from the field digital signals that often take the form
of switch-contact closures.
The application processor is a microprocessor (or computer) in which the programs (or
software) are stored for performing the many tasks described earlier and for managing the
communication among modules.
The workstation module is connected to a keyboard, a mouse, a monitor, and a printer
for use by process operators to interact with the system. At the workstation, the process
operator can call up on the screen various displays, change set points and controller pa-
rameters, switch from automatic to manual, acknowledge alarms, and perform other tasks
needed to operate a control system consisting of many loops. A control system can also be
configured as an offline task at the workstation. After configuration, the configured control
Figure 12.3 DCS Control room schematic.
system is downloaded to the control processor. If necessary, more than one workstation
can be attached to the nodebus in order to provide communication at several locations in a
plant. If more than one workstation is used, only one of them should have the authority at
a given time to be in charge of the control system.
12.2.1 Historical Review

To fully appreciate and select the current status of affairs in industrial practice it is of
interest to understand the historical perspective on the evolution of control systems im-
plementation philosophy and hardware elements. The evolution concerns the heart of any
control system which is how information flow and decision making advanced.
1. Pneumatic Implementation: In the early implementation of automatic control sys-

tems, information flow was accomplished by pneumatic transmission, and compu-
tation was done by mechanical devices using bellows, spring etc. The pneumatic
controller has high margin for safety since they are explosion proof. However, There
are two fundamental problems associated with pneumatic implementation:
Transmission: the signals transmitted pneumatically (via air pressure) are slow re-
sponding and susceptible to interference.
Calculation: Mechanical computation devices must be relatively simple and tend to
wear out quickly.
2. Electron analog implementation: Electrons are used as the medium of transmission

in his type of implementation mode. Computation devices are still the same as before.
Electrical signals to pressure signals converter (E/P transducers) and vice verse (P/E
transducers) are used to communicate between the mechanical devices and electron
flow. The primary problems associated with electronic analog implementation are:
Transmission: analog signals are susceptible to contamination from stray fields, and
signal quality tends to degrade over long transmission line.
Calculation: the type of computations possible with electronic analog devices is still
limited.
3. Digital Implementation: the transmission medium is still electron, but the signals
are transmitted as binary numbers. Such digital signals are far less sensitive to noise.
The computational devices are digital computers. Digital computers are more flexible
because they are programmable. They are more versatile because there is virtually
no limitation to the complexity of the computations it can carry out. Moreover, it is
possible to carry out computation with a single computing device, or with a network
of such devices.
Many field sensors naturally produce analog voltage or current signals. For this reason
transducers that convert analog signals to digital signals (A/D) and vice verse (D/A) are
used as interface between the analog and digital elements of the modern control system.
With the development of digital implementation systems, which DCS are based on, it is
possible to implement many sophisticated control strategies on a very fast timescale.
12.2.2 Modes of Computer control

Computer control is usually carried out in two modes: supervisory control or direct digital
control. Both are shown in Figure 12.4. Supervisory control involves resetting the set
point for a local controller according to some computer calculation. Direct digital control,
by contrast, requires that all control actions be carried out by the digital computer. Both
modes are in wide use in industrial applications, and both allow incorporating modern
control technologies. Measurements are transmitted to computer and control signals are
sent from computer to control valves at specific time interval known as sampling time. The
latter should be chosen with care.
Figure 12.4 Computer control modes.
12.2.3 Computer Control Networks

The computer control network performs a wide variety of tasks: data acquisition, servicing
of video display units in various laboratories and control rooms, data logging from analyt-
ical laboratories, control of plant processes or pilot plant, etc. The computer network can
be as simple as an array of inexpensive PC’s or it could be a large commercial distributed
control system (DCS).
12.2.4 Small Computer Network

In small processes such as laboratory prototype or pilot plants, the number of control loops
is relatively small. An inexpensive and straightforward way to deal with the systems is to
configure a network of personal computers for data acquisition and control. An example
configuration of a PC network control system is depicted in Figure 12.5. The consists of a
main computer linked directly to the process in two-way channels. Other local computers
are linked to the main computer and are also connected to the process through one-way or
two-way links. Some of these local computers can be interconnected. Each of the local
computers has a video display and a specific function. For example, some local computers
are dedicated for data acquisition only, some for local control only and some other for both
data acquisition and local control. The main computer could have a multiple displays.
All computers operate with a multitasking operating system. They would be normally
configured with local memory, local disk storage, and often have shared disk storage with
a server.
Figure 12.5 PC Network
12.2.5 Commercial Distributed Control Systems

In more complex pilot plants and full-scale plants, the control loops are of the order of
hundreds. For such large processes, the commercial distributed control system is more ap-
propriate. There are many vendors who provide these DCS systems such as Baily, Foxboro,
Honeywell, Rosemont, Yokogawa, etc. In the following only an overview of the role of
DCS is outlined.
Conceptually, the DCS is similar to the simple PC network. However, there are some
differences. First, the hardware and software of the DCS is made more flexible, i.e. easy
to modify and configure, and to be able to handle a large number of loops. Secondly,
the modern DCS are equipped with optimization, high-performance model-building and
control software as options. Therefore, an imaginative engineer who has theoretical back-
ground on modern control systems can quickly configure the DCS network to implement
high performance controllers.
A DCS is then a powerful tool for any large commercial plant. The engineer or operator
can immediately utilize such a system to:
Access a large amount of current information from the data highway.
See trends of past process conditions by calling archival data storage.
Readily install new on-line measurements together with local computers for data ac-
quisition and then use the new data immediately for controlling all loops of the pro-
cess.
Alternate quickly among standard control strategies and readjust controller parame-
ters in software.
A sight full engineer can use the flexibility of the framework to implement his latest
controller design ideas on the host computer or on the main control computer.
In the common DCS architecture, the microcomputer attached to the process are known
as front-end computers and are usually less sophisticated equipment employed for low
level functions. Typically such equipment would acquire process data from the measuring
devices and convert them to standard engineering units. The results at this level are passed
upward to the larger computers that are responsible for more complex operations. These
upper-level computers can be programmed to perform more advanced calculations.
12.2.6 Description of the DCS elements

The typical DCS system shown in Figure 12.5 can consists of one or more of the following
elements:
Local Control Unit (LCU): This is denoted as local computer in Figure 12.5. This unit
can handle 8 to 16 individual PID loops, with 16 to 32 analog input lines, 8 to 16
analog output signals and some a limited number of digital inputs and outputs.
Data Acquisition Unit: This unit may contain 2 to 16 times as many analog input/output
channels as the LCU. Digital (discrete) and analog I/O can be handled. Typically, no
control functions are available.
Batch Sequencing Unit: Typically, this unit contains a number of external events, timing
counters, arbitrary function generators, and internal logic.
Local Display: This device usually provides analog display stations, analog trend recorder,
and sometime video display for readout.
Bulk Memory Unit: This unit is used to store and recall process data. Usually mass stor-
age disks or magnetic tape are used.
General Purpose Computer: This unit is programmed by a customer or third party to

perform sophisticated functions such as optimization, advance control, expert system,
etc.
Central Operator Display: This unit typically will contain one or more consoles for op-
erator communication with the system, and multiple video colour graphics display
units.
Data Highway: A serial digital data transmission link connecting all other components in
the system may consist of coaxial cable. Most commercial DCS allow for redundant
data highway to reduce the risk of data loss.
Local area Network (LAN): Many manufacturers supply a port device to allow connec-
tion to remote devices through a standard local area network.
12.2.7 The advantages of DCS systems

The major advantages of functional hardware distribution are flexibility in system design,
ease of expansion, reliability, and ease of maintenance. A big advantage compared to a
single-computer system is that the user can start out at a low level of investment. Another
obvious advantage of this type of distributed architecture is that complete loss of the data
highway will not cause complete loss of system capability. Often local units can continue
operation with no significant loss of function over moderate or extended periods of time.
Moreover, the DCS network allows different modes of control implementation such as
manual/auto/supervisory/ computer operation for each local control loop. In the manual
mode, the operator manipulates the final control element directly. In the auto mode, the
final control element is manipulated automatically through a low-level controller usually a
PID. The set point for this control loop is entered by the operator. In the supervisory mode,
an advanced digital controller is placed on the top of the low-level controller (Figure 12.4).
The advanced controller sets the set point for the low-level controller. The set point for
the advanced controller can be set either by the operator or a steady state optimization. In
the computer mode, the control system operates in the direct digital mode shown in Figure
12.4.
One of the main goals of using DCS system is allowing the implementation of digital
control algorithms. The benefit of digital control application can include:
Digital systems are more precise.
Digital systems are more flexible. This means that control algorithms can be changed
and control configuration can be modified without having rewiring the system.
Digital system cost less to install and maintain.
Digital data in electronic files are easier to deal with. Operating results can be printed
out, displayed on colour terminals, stored in highly compressed form.
PROGRAMMABLE LOGIC CONTROLLERS 147
12.3 Programmable Logic Controllers
Programmable logic controller (PLC) is another type of digital technology used in process
control. It is exclusively specialized for non-continuous systems such as batch processes
or that contains equipment or control elements that operate discontinuously. It can also
be used for many instants where interlocks are required; for example, a flow control loop
cannot be actuated unless a pump has been turned on. Similarly, during startup or shutdown
of continuous processes many elements must be correctly sequenced; that is, upstream
flows and levels must be established before downstream pumps can be turned on.
The PLC concept is based on designing a sequence of logical decisions to implement
the control for the above mentioned cases. Such a system uses a special purpose computer
called programmable logic controllers because the computer is programmed to execute the
desired Boolean logic and to implement the desired sequencing. In this case, the inputs to
the computer are a set of relay contacts representing the state of various process elements.
Various operator inputs are also provided. The outputs from the computer are a set of
relays energized (activated) by the computer that can turn a pump on or off, activate lights
on a display panel, operate solenoid valve, and so on.
PLCs can handle thousands of digital I/O and hundreds of analog I/O and continuous
PID control. PLC has many features besides the digital system capabilities. However,
PLC lacks the flexibility for expansion and reconfiguration. The operator interface in PLC
systems is also limited. Moreover, programming PLC by a higher-level languages and/or
capability of implementing advanced control algorithms is also limited.
PLCs are not typical in a traditional process plant, but there some operations, such
as sequencing, and interlock operations, that can use the powerful capabilities of a PLC.
They are also quite frequently a cost-effective alternative to DCSs (discussed next) where
sophisticated process control strategies are not needed. Nevertheless, PLCs and DCSs can
be combined in a hybrid system where PLC connected through link to a controller, or
connected directly to network.
12.4 Digital control software
To make the best use of a DCS system, an advance control strategy or supervisory op-
timization can be incorporated in the main host computer. In the past, computer control
projects are written in assembly language, an extremely tedious procedure. Nowadays
most user software is written in higher-level languages such as BASIC, FORTRAN, C etc.
In many cases, the user is able to utilize the template routines supplied by the vendor, and
is required only to duplicate these routines and interconnect them to fit his own application
purposes. Another way is to write his own complete control program and implement it.
Other software in the form of control-oriented programming languages is supplied by
the vendor of process control computers. A simpler approach for the user is to utilize
vendor-supplied firmware or software to avoid writing programs. Currently, most DCS
manufacturers develop their own advance control and optimization software, which can
included in the package as options. Similarly, many control algorithm developers; (DMC,
ASPEN, etc) design a special interface to allow incorporating their own control programs
into most of the commercial DCS network.
***************
REFERENCES
[1] B.W. Bequette. Process Control Modeling, Design and Simulation. Prentice Hall of India
Pvt.Ltd., 1998.
[2] D. F. Coughanowr. Process systems analysis and control. McGraw-Hill, 1991.
[3] J. L. A. Koolen. Design of Simple and Robust Process Plants. Wiley-VCH Verlag GmbH & Co.,
2002.
[4] Luyben W. L. Luyben, M. L. Essentials of Process Control. McGraw-Hill, 1997.
[5] W. L. Luyben. Process Modeling Simulation and Control for Chemical Engineers. McGraw-
Hill, 1990.
[6] Peacock D. G. Richardson, J. F. Coulson & Richardson’s Chemical Engineering, volume 3.
Asian Books Pvt Ltd, New Delhi, 3 edition, 1994.
[7] G. Stephanopoulos. Chemical Process Control. Prentice Hall of India., 1990.

APPENDIX A
EXERCISES
EXERCISES
A.1 Verify whether following system is stable using Routh - Hurwitz criterion.
P (s)
X (s) + Y (s)
Gc G1 + Gp
+
–
Gm

1 1
Gc = 2 1 + G1 =
2s s+1
2 1
Gp = Gm =
3s + 1 0.5s + 1
A.2 Draw a root locus diagram for following open loop transfer function.
K
G (s) =
(s + 1) (s + 2) (s + 3)
A.3 Show that phase lag for pure capacitive system is 90°.
152 EXERCISES
A.4 Find Amplitude Ratio (AR) and phase lag for n-first order capacities in series con-
nected in non-interacting fashion.
A.5 Plot the Bode diagram for the system whose overall transfer function is,
1
G (s) =
(s + 1) (s + 5)
A.6 A thermometer having first-order dynamics with a time constant of 1 min is placed
in a temperature bath at 100oF. After the thermometer reaches steady state, it is suddenly
placed in a bath at 110oF at t = 0 and left there for 1 min, after which it is immediately
returned to the bath at 100o F.
a) Draw a sketch showing the variation of the thermometer reading with time.
b) Calculate the thermometer reading at t = 0.5 min and at t = 2.0 min.
A.7 A thermometer having a time constant of 1 min is initially at 50o C. It is immersed
in a bath maintained at 100o C at t = 0. Determine the temperature reading at t = 1.2 min.
A.8 A jacketed vessel is used to cool a process stream. The volume of liquid in the tank
V and the volume of the coolant in the jacket VJ remain constant. Volumetric flow rate
qF is constant but qJ varies with time. Heat losses from the jacketed vessel are negligible.
Both the tank contents and the jacket contents are well mixed and have significant thermal
capacitances. The thermal capacitances of the tank wall and the jacket wall are negligible.
The over-all heat-transfer coefficient between the tank liquid and the coolant varies with
coolant flow rate as follows:
U = KqJ0.8
Derive a dynamic model for the system. State all the assumptions made in the model.
A.9 Is it possible for a two-tank liquid level system to exhibit the under-damped be-
haviour of a second-order system. Consider both interacting and non-interacting cases.
Support your answer with a proof.
A.10 A particular system has been disturbed with a unit impulse function. The response
of the system in the time domain has been found to be described well by the function
Y (t) = te−t
Determine the unit step response of the system.

A.11 A liquid storage tank has two inlet streams with mass flow rates w1 and w2 an exit
stream with a mass flow rate w3 . The cylindrical tank is 2.5 m tall and 2 m in diameter.
The liquid has a density of 800 kg/m3 . Normal operating procedure is to fill the tank until
the liquid level reaches a nominal value of 1.75 m using constant flow rates: w1 = 120
kg/min, w2 = 100 kg/min and w3 = 200 kg/min. At that point , inlet flow rate w1 adjusted
so that the level remains constant. However, on this particular day, corrosion of the tank
has opened up a hole in the wall at a height of 1 m , producing a leak whose volumetric
flow rate can be approximated by:
√
q4 = 0.025 h − 1 in m3 /min
where the height of the liquid in the tank,h is in meter.

a) If the tank was initially empty, how long did it take for the liquid level to reach
the corrosion point?
EXERCISES 153
b) If the mass flow rates w1 , w2 and w3 are kept constant indefinitely, will the tank
eventually overflow? Justify your answer.
A.12 Consider the measurement of temperature in Blast-furnace using a thermocouple,
whose junction has a mass of m, with heat capacity C. You may assume that only radia-
tive heat-transfer is of importance here. Derive a transfer-function relating the measured
temperature T to the furnace temperature Tf .
A.13 Consider a two-tank Non-interacting system with the following data:
A1 = A2 = 200 cm2
q1s = q2s = qis = 15 lpm
At this steady flow condition the height of water in the first tank is 30 cm. There is a pump
in the outlet of the second tank, which discharges water at a constant rate. Consider the
following disturbances in the inlet to the first tank:
Case I: The inlet flowrate to the first tank increases to 16 lpm at t = 0 as a permanent step
change.
Case II: The same step change in the inlet flowrate occurs, but only for one minute after
which it returns to its original steady-state value.
Case: III One lit of water is added to the first tank suddenly as an impulse disturbance.
In each case, study the dynamic response of the height in tank two, both qualitatively and
quantitatively.
A.14 A two tank non-interacting liquid level system is operating at steady state with
water flowing at 10 lit/min. The cross-sectional areas of the first and second tanks are
A1 = 200 cm2 andA2 = 300 cm2 respectively. The steadystate liquid heights in the
two tanks are h1s = 20 cm and h2s = 15 cm respectively. The outlet resistances can be
assumed to be linear. At some time t = 0, one litre of water is suddenly added to the
first tank. Determine the dynamic response of both the tanks analytically and show the
variation of the liquid heights with respect to time graphically.
A.15 A two tank liquid level system is operating at steady state when a step change is
made in the flow rate to tank 1. The transient response is critically damped, and it takes 1.0
minute for the change in level of the second tank to reach 50 percent of the total change. If
the ratio of the cross-sectional areas of the tanks is A1 /A2 = 2 , calculate the ratio R1 /R2
. Calculate the time-constant for each tank. How long does it take for the level in the first
tank to reach 90 percent of the total change?
A.16 A two tank mixing process is initially operating under steady state conditions with
a volumetric flow rate of 2 m3 /min of a solution. The solute concentration is 1 kg/m3 . The
volume of the first tank is 4 m3 and that of the second tank is 6 m3 . At time t = 0, the
solute concentration in the inlet to the first tank suddenly increases to 2 kg/m3 . How long
does it take for the solute concentration in the outlet from the second tank to reach 1.6
kg/m3 ?
A.17 Consider the following level control system as shown in figure A.1 The cross sec-
tional area of the tank is 225 cm2 and the resistance is 0.005 min/cm2 . The level sensor is
a differential pressure transmitter that gives an output of 4 20 mA for an input variation in
level of 0 60 cm. The control valve is a motorized valve that operates on an input signal
of 4 20 mA, and has a gain of 1.25 lpm/mA. The dynamic lag of the transmitter as well
as that of the valve can be neglected. The controller is a PI controller with a Proportional
Band of 15% and an integral time of 20 sec.
154 EXERCISES
Control Valve
Controller
LT R
Figure A.1 Level controller
a) Derive the transfer functions for each of the individual components that make up
the control system. Calculate the parameters of the transfer functions.
b) Draw a block diagram for the control system. Obtain a closed loop transfer
function for set point changes.
c) Determine the range of values of the gain of the controller for which the system
remains stable.
A.18 A control system consists of the following components:
2
Process : Gp (s) =
0.5s2 + s + 1

3
Controller : Gc (s) = Kc 1 +
s
2
Control Valve : Gv (s) =
0.5s + 1
a) Draw a root locus diagram for the system.

b) Determine the value of Kc needed to obtain a root of the characteristic equation
of the closed-loop response which has an imaginary part 0.75.

Ipcnotes

Uploaded by

Ipcnotes

Uploaded by

INSTRUMENTATION & PROCESS

Library of Congress Cataloging-in-Publication Data:

Instrumentation & Process Control / Sanjay Dalvi

PART I PROCESS CONTROL

1 Concepts of Control System 3

2 Closed Loop Systems 25

2.1 What is control system? 25

6 Introduction to SciCos / XCos 71

9 Level measurement 105

9.6 Electrical type level gauges and switches 107

PART III CONTROL STRATEGIES AND APPLICATIONS

10 Control Valve 111

11 Advanced control strategies 123

12 Microprocessor-based controller 139

12.2.6 Description of the DCS elements 145

CONCEPTS OF CONTROL SYSTEM

1.2 Steady state and Unsteady state

1.3 Process Control

1.4 Feedback and Feedforward

1.5 Transfer function for First Order system

{Input rate} − {Output rate} = {rate of accumulation}

therefore equation (1.3) becomes,

θo (s) − θth (s) = τ sθth (s)

X(s) G(s) Y(s)

1.5.1 Step input to first order system

A thermometer having a time constant of 0.1 min is at a steady state temperature of

1.5.2 Impulse input to first order system

Inverting above equation,

1.5.3 Sinusoidal input to first order system

X (t) = A sin (ωt)

where ω = radian frequency, rad/time

Multiply equation (1.11) with (s + ωj) and set s = −ωj

therefore response equation becomes,

Using the identity,

e(C1 + jC2 ) t = ec1 t (cos C2 t + j sin C2 t)

here in this case, C1 = 0 and C2 = ω for third term,

eωjt = cos (ωt) + j sin (ωt)

C2 = −ω for second term,

e−ωjt = cos (−ωt) + j sin (−ωt) = cos (ωt) − j sin (ωt)

therefore the square bracket of equation (1.13) is

τ cos (ωt) − τ ωj sin (ωt) − j cos (ωt) − sin (ωt)

replacing square bracket of equation (1.13) we get,

p cos A + q sin A = r sin (A + θ)

Figure 1.1 Response of first order system to sinusoidal input

A mercury thermometer having a time constant of 0.1 min is placed in temperature

π = tan−1 (−ωτ ) = tan−1 (−20 × 0.1) = −63.5◦

1.6 Transfer function for liquid level in tank

Material balance across a tank,

Figure 1.2 Liquid level in tank

for incompressible (liquid) fluid,

F̄ = Fin − Fins (1.20)

now equation (1.19) can be written as,

AsH̄ (s) = F̄ (s) − H̄ (s) /R

where, τ = AR Outflow can also be represented in terms of deviation variable as,

1.7 Transfer function for pure capacitive process

Figure 1.3 Pure capacitive tank

Subtracting equation (1.26) from equation (1.25),

F̄ (s) = AsH̄ (s)

1.8 First order systems in series

Figure 1.4 Non-interacting tanks

Flow and head relationship is given by,

Taking partial fractions of RHS of above equation,

Multiply equation(1.40) with s and set s = 0

Multiply equation(1.40) with (τ1 s + 1) and set s = − τ11

Multiply equation(1.40) with (τ2 s + 1) and set s = − τ12

Put values of C1 , C2 and C3 in equation(1.40)

Inverting above equation,

Now, τ2 = 0.5, τ1 = 1 and R2 = 1

H2 (t) = 1 − 2e−t − e−2t

Figure 1.5 Interacting tanks

Flow and head relationship is given by,

Q (s) − Q1 (s) = A1 sH1 (s) (1.49)

Replace Q2 (s) of equation(1.50) using equation(1.52)

R2 Q1 (s) − H2 (s) = A2 R2 sH2 (s)

Q (s) − Q1 (s) = A1 s [R1 Q1 (s) + H2 (s)]

Using equation(1.53) replace Q1 Q (s) of above equation,