by David W.

Spitzer
Without modern instrumentation control, industry would be at a standstill. Whether the
control hardware consists of local mechanical, pneumatic, or electronic controls, electronic
or pneumatic analog controllers, microprocessor-based controllers, programmable logic controllers, personal computer-based controls, or a distributed control system, a keen awareness
of advanced regulatory control is required by engineers, technicians, educators, managers,
salespersons, and marketing personnel in order that appropriate control technology be applied to industrial processes.
This book is intended to describe advanced regulatory control and its application to continuous processes in a nonmathematical format and in as practical a manner as possible in
order to be of benefit to all skill levels. Manual and regulatory control are described in the
text as a prelude to advanced regulatory control so that their differences might be explored.
It focuses on control but integrates the operation of measurement devices and final control
elements into discussions, because their importance is often the limiting factor in control
system performance.
Readers will find
An explanation of the operation and limitations of feedback control
Illustrations of the integration of cascade and feedforward control into feedback
control loops
A clear approach to design and implement advanced regulatory control
About the Author
David W. Spitzer is a Principal at Spitzer and Boyes, LLC, specializing
in instrumentation and control consulting for industrial users, manufacturers, and representatives world-wide, including serving as an expert
witness. Mr. Spitzer previously worked for Nepera Chemical, Mobay
Chemical, and U.S. Steel, proposing and implementing many control
and energy-saving projects.

An instructor at the ISA Training Center in Research Triangle Park, N.C., throughout
the United States, and internationally in the ISA Short Course Program, Mr. Spitzer is the
author of Industrial Flow Measurement (ISA Press), Variable Speed Drives: Principles and Applications for Energy Cost Savings (ISA Press), Consumer Guides to Coriolis Mass, Differential Pressure,
Magnetic, Ultrasonic, and Vortex Shedding Flowmeters, and Contact and Non-Contact Level Gauges
(Copperhill and Pointer), as well as the Volume Editor of Flow Measurement (ISA Press).
Mr. Spitzer has also written some 150 articles in various technical magazines.
ISBN: 978-1-60650-009-5

90000

www.momentumpress.net

9 781606 500095

Spitzer

Mr. Spitzer is a Life Fellow of the International Society of Automation (ISA) and serves on the American Society of Mechanical Engineers (ASME) Measurement of Fluid Control in Closed Conduits committee (MFC).

Advanced Regulatory Control

Advanced Regulatory Control

Advanced
Regulatory
Control
Applications and Techniques

by David W. Spitzer

Advanced
Regulatory Control:
Applications and Techniques

by David W. Spitzer

MOMENTUM PRESS, LLC, NEW YORK

imo-spitzer-00fm.indd i

9/21/09 11:52 AM

Advanced Regulatory Control: Applications and

Techniques
Copyright Copperhill and Pointer, Inc., 2010
All rights reserved. No part of this publication
may be reproduced, stored in a retrieval system, or
transmitted in any form or by any meanselectronic,
mechanical, photocopy, recording or any other except
for brief quotations, not to exceed 400 words, without
the prior permission of the publisher
First published in 2009 by
Momentum Press, LLC
222 East 46th Street, New York, N.Y. 10017
www.momentumpress.net
ISBN-13: 978-1-60650-124-5 (hard back, case bound)
ISBN-10: 1-60650-124-0
(hard back, case bound)
ISBN-13: 978-1-60650-126-9 (e-book)
ISBN-10: 1-60650-129-1
(e-book)
DOI forthcoming
Interior Design by Scribe, Inc.
First Edition October 2009
10 9 8 7 6 5 4 3 2 1
Printed in the United States of America.

imo-spitzer-00fm.indd ii

9/21/09 11:52 AM

This work is dedicated to my children, Deborah Goldie and

Michael James, in the hope that it will inspire them and others
to continue their education and their pursuit of knowledge; and
to my wife Ida for understanding why I was so late to bed for so
many months.

imo-spitzer-00fm.indd iii

9/21/09 11:52 AM

imo-spitzer-00fm.indd iv

9/21/09 11:52 AM

About the Author

David W. Spitzer is a Principal at Spitzer and Boyes, LLC, specializing in

instrumentation and control consulting for industrial users, manufacturers,
and representatives world-wide, including serving as an expert witness.
Mr. Spitzer previously worked for Nepera Chemical, Mobay Chemical, and
U.S. Steel, proposing and implementing many control and energy-saving
projects.
Mr. Spitzer is a Life Fellow of the International Society of Automation
(ISA) and serves on the American Society of Mechanical Engineers (ASME)
on the Measurement of Fluid Control in Closed Conduits committee (MFC).
An instructor at the ISA Training Center in Research Triangle Park, , N.C.,
throughout the United States, and internationally in the ISA Short Course
Program, Mr. Spitzer is the author of Industrial Flow Measurement (ISA
Press), Variable Speed Drives: Principles and Applications for Energy Cost
Savings (ISA Press), Consumer Guides to Coriolis Mass, Differential Pressure, Magnetic, Ultrasonic, and Vortex Shedding Flowmeters, and Contact
and Non-Contact Level Gauges (Copperhill and Pointer), as well as the Volume Editor of Flow Measurement (ISA Press). Mr. Spitzer has also written
approximately 150 articles in various technical magazines.

imo-spitzer-00fm.indd v

9/21/09 11:52 AM

imo-spitzer-00fm.indd vi

9/21/09 11:52 AM

Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Manual Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3 Field Measurement Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Flow Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Level Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Pressure Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Temperature Measurement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Transmitter Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
For Further Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4 Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Controlling Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Local Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Centralized Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Distributed Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5 Final Control Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Control Valves. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Variable Speed Drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6 Field Equipment Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Control Loop Tuning Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Equipment Adjustments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Sensor Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Transmitter Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Controller Input Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Final Control Element Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7 Regulatory Controller Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Anti-Reset Windup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Bumpless Transfer (Set Point Tracking) . . . . . . . . . . . . . . . . . . . . . . . . 57
Control Dead Band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Controller Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Direct Forward/Reverse Control Action . . . . . . . . . . . . . . . . . . . . . . . . 61
Failure Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Event-Related Controller Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Output Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Remote Set Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Set Point Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Slew Rates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

vii

imo-spitzer-00fm.indd vii

9/21/09 11:52 AM

viii

Contents

Totalizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8 PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Proportional Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Integral Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Derivative Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70
70
71
71
73
75
77
78

9 Controller Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Controller Tuning Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Quarter Amplitude Oscillation Method . . . . . . . . . . . . . . . . . . . . . . . . 82
Trial and Error Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Self-Tuning Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Typical Controller Settings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
For Further Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
10 Regulatory Control Loop Pairing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Regulatory Control Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
11 The Limitations of Regulatory Control . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Control Loop Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Process Upsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Control Loop Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
12 Advanced Regulatory Control Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Mathematical Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Advanced Regulatory Control Process Variables . . . . . . . . . . . . . . . . 105
Feedback Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Cascade Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Feedforward Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Feedforward Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Constraint and Override Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
13 Advanced Regulatory Control Design and Implementation . . . . . . . . 123
Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
14 Applying Advanced Regulatory Control . . . . . . . . . . . . . . . . . . . . . . . 129
Define the Product and Its Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Define the Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Define the Boundaries of the Process . . . . . . . . . . . . . . . . . . . . . . . . . 130
Define Quality Measurements for Evaluation . . . . . . . . . . . . . . . . . . . 130
Define Quality Measurements for On-Line Operation . . . . . . . . . . . . 131
Develop the Control Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Define Instrumentation Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Critique the Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

imo-spitzer-00fm.indd viii

9/21/09 11:52 AM

 Preface

The Challenge of Advanced

Regulatory Control
Whether the instrumentation control hardware consists of local
mechanical, pneumatic, electronic, microprocessor, fieldbus or
wireless controls, microprocessor-based controllers, programmable logic controllers, personal computer-based controls, distributed control systems, or fieldbus controls, a keen awareness of
advanced regulatory control is required by engineers, technicians,
educators, managers, salespersons, and marketing personnel such
that an appropriate control strategy can be applied to industrial
processes.
Make no mistakethe forces of instant gratification are definitely at work in the wonderful world of process control. It
is not uncommon to find support for improved control systems
driven by the aesthetics of pretty pictures, buzz words, the
need to be the first one on the block to have the latest thing, or
the need not to be the only one without one. In these situations,
little attention is truly focused on the quality and sophistication
of control.
The quality and sophistication with which control strategies are
developed and implemented varies with the skill of the process
control engineer. This skill is heightened by a strong control engineering education, exposure to skilled process control engineers,
specialization in the field of process control, and intimate knowledge of the process.
Many persons perform process control engineering on a part-time
basis without the benefit of a process control education, exposure
to skilled practitioners, or understanding of the process. As verified by the observation of numerous control strategy implementations, the typical result of this approach is a control system that
works.

ix

imo-spitzer-00fm.indd ix

9/21/09 11:52 AM

The Challenge of Advanced Regulatory Control

Such a situation can be illustrated by the authors experience with an individual who implemented a distributed
control system and enthusiastically described it, especially
with regard to its flexibility and ease of configuration. After
responding with silence to questions that probed the sophistication of the control strategies, this individual sat in awe as
the operation of a cascade control loop was described. Despite
its outstanding functional capability, this individuals distributed control system implemented less-sophisticated control than is possible with a pneumatic analog control system.
Nonetheless, it cannot be denied that the distributed control
system works, that the plant can boast of having installed
the latest and greatest technology, and that this experienced
expert may influence the decisions of others who are less
knowledgeable than he.
This type of implementation occurs because of a series of
events that typically begins with managements recognition
and support for improved process control. Due to the skill
level of the person selected to perform the control system
engineering function, inappropriate or rudimentary control
strategies may be implemented, which result in the loss of an
opportunity to improve plant efficiency. Operators have little
choice but to accept the control system and work around its
shortcomings, including operating in the manual mode (system
abandonment) because the controls are inconvenient to operate
or do not function properly. However, despite these shortcomings, the operators are usually satisfied because the installation
represents improvement.
The control system that works may be regarded as a success, primarily because management is rarely versed in process
control technology, has difficulty judging the technical quality
of the installation, and is usually forced to draw conclusions
about instrumentation from the opinions of others who are not
skilled control system engineers. It is not uncommon for an
experienced control system engineer to review such an installation and recognize the need for an additional control system
upgrade to improve process performance and personnel safety.
Were the problem solely caused by errors, improved process
control and safety could be achieved with a few more expenditures to correct deficiencies. However, the underlying problem
is that an installation that works is deemed acceptable, regardless of whether it truly is. What is not recognized is that the process may really be poorly controlled and/or unsafe because of
inappropriate control strategies and interlocks.

imo-spitzer-00fm.indd x

9/21/09 11:52 AM

The Challenge of Advanced Regulatory Control

xi

The challenge of providing appropriate controls for industrial

processes is a formidable task due to the many facets, details,
and idiosyncrasies of industrial processes and the instruments
selected to be integrated into their operation. Virtually anyone,
with or without a process control background, can put together
a control system that works with the help of a telephone and
enough suppliers; however, to take advantage of technology to
improve the process through the innovative application of instrumentation requires a skilled, process control system engineer.
A skilled process control system engineer can promote process
improvements based upon specific knowledge of the instrumentation and the process. Such improvements can streamline the
process, provide safe operation, improve efficiency, eliminate
the need for equipment, and reduce installation, operation, and
maintenance costs. Despite claims to the contrary, individuals who have not specialized in process control and have not
exposed themselves to more sophisticated control techniques
have not generally produced innovative installations that
exploit the potential of control technology and open the door to
additional process improvement.
Most industrial control is performed by utilizing a combination
of manual and regulatory control. Despite the installation of
additional field transmitters and advanced control capabilities,
regulatory controls that emulate conventional pneumatic and
electronic analog controllers abound. To become more productive and to improve safety, advanced controls should be appropriately and correctly applied to industrial processes.
Traditionally, for lack of a better definition, advanced control
has been defined as continuous control that was more complex
than regulatory control (containing one measurement device,
one controller, and one final control element). Advanced
regulatory control can be segregated from advanced control
strategies such as model-based control, state variable control,
optimal control, dynamic matrix control, statistical process
control, and the like.
Herein, advanced regulatory control encompasses continuous
controls that are more complex than regulatory controls (as
defined above), but still retain the measurement, the controller,
and the final control element format of regulatory control. The
complexities of advanced regulatory control may be due to multiple measurements, controllers, and/or final control elements as
well as logical manipulation of the control loop.

imo-spitzer-00fm.indd xi

9/21/09 11:52 AM

xii

The Challenge of Advanced Regulatory Control

Advanced regulatory control is but one tool in the process control

system engineers arsenal; however, advanced regulatory control
can and should be applied to the majority of control systems. In
some circles, it is believed that if advanced regulatory control
were properly and appropriately applied, the need to resort to
applying advanced controls would be greatly diminished.
It should be noted that some aspects of advanced regulatory
control have been practiced for decades using mechanical,
pneumatic, and electronic instrumentation. Distributed control
systems, microprocessor-based controllers, and fieldbus control systems have dramatically reduced the cost of performing
advanced regulatory control functions and have simplified their
implementation.
This book is intended to describe advanced regulatory control
and its application to continuous processes in a nonmathematical format and in as practical a manner as possible in order to be
of benefit to all skill levels. Manual and regulatory control are
described in the text as a prelude to advanced regulatory control
so that their differences might be explored.
Discussions of control system hardware are avoided, because
the control function, not its hardware implementation, is important in improving control of the process. This contrasts the
importance of the measurement instruments and final control
elements, which, if not properly selected and installed, can
cause the control system to operate in the GIGO mode, that is,
garbage in, garbage out. In this mode, and despite control
system sophistication, poor process measurements and/or poor
process manipulation methods hamper control of the process. In
many cases, poor process measurements may not show process
instability, which would lead the uninitiated to the erroneous
conclusion that the control is adequate and the instrumentation
is not in need of upgrading.
This text focuses on control but integrates the operation of measurement devices and final control elements into discussions,
because their importance is often the limiting factor in control
system performance.

imo-spitzer-00fm.indd xii

9/21/09 11:52 AM

 Chapter 1

Introduction
People say that the world is getting smaller, in large part due to
technology, yet ironically technology is becoming so specialized
that there should be mulitple persons with distinctly different
job functions performing the same function as one person did
only a few years ago. It would seem logical that a person regularly practicing a particular skill will perform better and more
efficiently than someone who does not. However, as these job
functions migrate farther apart and as companies reduce personnel to become more lean, the ability of one person to adequately
perform all of these job functions is reduced. On the other hand,
organizations that have sufficient personnel for each job function
often have problems interfacing the work products.
The field of instrumentation has not escaped this trend. Ignoring
the specialists that may be employed by larger organizations, such
as in the areas of analyzers, programmable logic controllers, flow
measurement, control valves, and the like, the primary disciplines
responsible for the design and construction of instrumentation projects are typically instrumentation engineers who specify and supervise the installation of field hardware, and process control engineers
who develop control strategies and implement them in software.
This separatist trend has also permeated the literature where one
finds treatises on narrow subjects that are separated by discipline.
For efficient and safe plant operation, the instrumentation and
process control systems should work together. As instrumentation
and process control continue to drift farther apart, problems are
created whereby the disciplines work in a vacuuma situation
that can result in a loss of opportunity. Under these conditions, it
is not uncommon to find instrument engineers who do not understand the concept of object-oriented programming, nor must one
wander far to find process control engineers who do not understand how instruments work and what they measure.
This book attempts to bridge this gap by addressing the control problem in the context of the control system, including the

imo-spitzer-01.indd 1

8/31/09 1:58 PM

Introduction

integration of field instrumentation. For this reason, this book

should be of benefit to engineers, technicians, and educators of
all skill levels.
This book does not teach specific selection and implementation
skills, but rather provides the reader with the underlying technical understanding to implement a control system and highlights
some of the pitfalls. The evolutionary journey flows from manual
control to regulatory control to advanced regulatory control, with
each successive subject building on previously covered material.
Therefore, it is strongly recommended that the text be read in its
entirety to fully grasp the concepts that are presented.

imo-spitzer-01.indd 2

8/31/09 1:58 PM

 Chapter 2

Manual Control
Manual control, in which a minimum of local measurements
and manual throttling valves are required, is becoming increasingly rare in its pure form even though it offers some advantages. The devices required for manual control are inexpensive
relative to automated instrumentation and are easy to specify
and install. Piping requirements are reduced by eliminating
bypass valve arrangements and pneumatically operated equipment, while electrical requirements are reduced by installing
local indication only.
Incorrect installation techniques and inappropriate instrument
selection for manual control can be masked by the lack of process
measurements and the operators ability to compensate for these
hindrances. Manual control is perhaps the most flexible of all
control systems in that changes can be made immediately by simply telling the operator to make them.
Control systems engineering is typically not done for manual systems because there is virtually no control system to design, and
what can be designed is so rudimentary and forgiving that virtually anyone can, and does, perform the function.
Upon closer examination, however, many apparent advantages of manual control can quickly be seen as disadvantages.
Manual control is less expensive to install than other control
systems, but more operators are required to operate the system because a number of adjustments should be made at the
same time in different locations, especially when the safety of
the process is a concern. This approach exposes the operator
to potential hazards because a person is likely to be physically present at the equipment for adjustments to be made. In
addition, the actions of one operator may unknowingly affect
another operators part of the process, and create a process
upset that could cause a potentially unsafe condition (aside
from economic penalties).

imo-spitzer-02.indd 3

9/15/09 10:15 AM

Manual Control

Example 2-1:
Consider the case of an operator who pumps a liquid from one
area of the plant to another. A second operator in the other area
adjusts a valve that modulates the liquid flow into the process.
Seeing that the supply tank level is falling, the first operator may
adjust a valve in the first area that reduces the flow to allow the
level to rise. Meanwhile, the second operator may be in an upset
condition requiring maximum liquid flow. The manual control
strategies of these operators is (at least temporarily) in conflict
and could cause a safety hazard if control of the second process is
lost. Further, this type of incident can occur within the same area
of the plant.
System flexibility may appear to be an advantage; however, it
quickly becomes a disadvantage when one considers that, in a
24-hour-a-day operation, all four operators and their substitutes
should be told of any and every change. Further, it is unlikely
that all four operators perform their normal functions in exactly
the same way, which could result in variations in process conditions and product quality.
From a technical perspective, upgrading manual operations can
provide significant opportunities for process improvements,
capacity increases, and increased productivity. However, if the
system works, specific justification for the expenditures is usually
difficult to find prior to installation, especially since the skilled
process control engineer usually has no choice but to scrap virtually all of the existing instruments because of their age and lack
of versatility.
Control system expenditures are often justified on the basis of
improved quality and increased productivity that are expected
when the process is maintained at a given operating condition. Yield improvements and capacity increases can occur by
observing and adjusting operating conditions, but the amount of
improvement often cannot be predicted prior to installation.
Example 2-2:
The manually controlled boiler shown in Figure 2-1 is operated to control the steam header pressure to the plant. Without
the clutter of the required flame management system on the figure, it can be seen that a minimum of instrumentation has been
installed: two pressure gauges, a sight glass, two throttling valves,
and a damper.

imo-spitzer-02.indd 4

9/15/09 10:15 AM

FIG. 2-1.

Manually controlled boiler (flame management system not shown)

Manual Control

imo-spitzer-02.indd 5

9/15/09 10:15 AM

Manual Control

Steam header pressure (indicated at the top of the boiler) can be

controlled by simultaneously manipulating the fuel valve and
the air damper on the burner. In theory, an operator can perform
this control function, but a logistics problem occurs: how can the
header pressure be observed while the burner is being adjusted?
One solution is to remotely mount the steam header pressure
gauge near the burner.
The operator makes adjustments to the air and fuel flows by
simultaneously adjusting the natural gas valve and the combustion fan damper while visually observing the flame, the natural
gas pressure, and the steam drum pressure. The operator should
then wait for the effect of the firing change and any concurrent
plant load changes on header pressure before making any additional adjustments. Inappropriate adjustment of the air and/or gas
can cause a boiler trip due to a flameout (due to an improper mixture) or high steam pressure (due to overfiring for a given steam
load). In theory it should be possible for an operator to perform
these functions, but in practice the quality with which an operator can perform this function consistently is in question.
While the above is occurring, the boiler drum level is maintained
by another operator who visually observes the level gauge and
adjusts the feedwater valve. The drum level will vary in response
to steam load changes, burner adjustments made by the first operator, and manipulation of the feedwater valve. A third operator
may be similarly adjusting the de-aerator level and temperature
(not shown in Figure 2-1).
The cost to engineer and install this instrumentation system is
low, but the labor cost associated with its operation is high, especially when one considers that most package boiler manufacturers offer automation packages requiring no operator attention
to perform the above functions, while consistently maintaining
better fuel efficiency and a more stable header pressure than the
manual controls.
A manual control system is the most economical instrumentation system to purchase and install, however, the potential for
uneconomical and unsafe operation is high due to the lack of
equipment to monitor the process for human mistakes or equipment failure. The potential economical, safety, and environmental liabilities make manual control an undesirable option for most
industrial processes.

imo-spitzer-02.indd 6

9/15/09 10:15 AM

 Chapter 3

Field Measurement Devices

Instruments are available to continuously measure many process variables, the most common of which are flow, level, pressure, and temperature. Other process variables are routinely
measured, including conductivity, density, pH, speed, position,
current, voltage, and the like. Because a regulatory control loop
uses only one device to measure the process variable, selection of
the appropriate device to achieve the desired measurement with
acceptable accuracy at the proper location is critical to the stable
operation of the process.
In addition to evaluating a measurement instrument with
respect to how well it performs, how long it maintains that performance, and how long it continues to operate without maintenance or failure should also be considered. Instrument failure
and/or continued maintenance can be a symptom of a process
problem, especially when the problem is related to the wetted
parts of the instrument.
While replacement and/or upgrading might solve an immediate
instrument problem, a more detailed analysis of the failure could
lead to the correction of a process problem that, in turn, eliminates the instrument problem. It is not uncommon for properly
designed measurement instruments to operate reliably for years
without failure.
To accompany the subtleties of measurement instrument selection, it is assumed that the reader has a working knowledge of
these instruments. Further insight into measurement may be
obtained from the references listed at the end of this chapter.
It should be noted that selecting an instrument to measure a
process variable may appear to be a relatively straightforward
task, but, in practice, this may turn out to be quite a difficult
exercise.

imo-spitzer-03.indd 7

9/15/09 1:43 PM

Field Measurement Devices

Example 3-1:
Consider measuring the output pressure of a gas pressure regulator. It would appear that a pressure gauge with an appropriate
measurement range would suffice, if its materials of construction
were compatible with the fluid, because the downstream pressure
will be stable at all times. However, it should be recognized that
under no-load conditions, the gas pressure regulator may leak,
causing the measured downstream pressure to eventually become
equal to the upstream pressure. Therefore, the pressure gauge
should also be capable of withstanding the full upstream pressure
without losing calibration or bursting.
Further, if the downstream pipe feeds a large vessel, the dynamics of the reaction in the vessel can affect the pressure downstream of the regulator, causing pressure fluctuations. Such
dynamic conditions (on a seemingly static process) can prematurely wear out the pressure gauge mechanism.
One problem associated with the selection of measurement
devices is the assumption that the process is static. Virtually
all processes are dynamic and should be considered as such.
As illustrated above, often what appears to be an appropriate
measurement technique may not be. Selection of an appropriate measurement device and options can be critical to achieving
proper control.

Flow Measurement
Although routinely applied to industrial processes, flowmeter
installations can often present measurement problems. It may
Flow Measurement Type Measurement Principle
Volumetric
Positive Displacement
Velocity
Magnetic
Oscillatory

Mass
Inferential

Table 3-1.

imo-spitzer-03.indd 8

Turbine
Ultrasonic
Coriolis
Thermal
Differential Pressure
Target
Variable Area

Flow Measurement Methods

9/15/09 1:43 PM

Field Measurement Devices

seem that any flowmeter will perform adequately, however, a

more in-depth examination provides the insight that different
flowmeters measure flow differently.
It should be understood that flowmeters are fundamentally
mechanical devices, even if they appear to be electronic, because
the fluid passes through a flowmeter body. The volumetric fluid
flow (Q) through a conduit (or flowmeter) with cross-sectional
area (A) is related to fluid velocity (v) by

Q=Av

(3-1)

In addition, the fluid mass flow (W) is related to the volumetric

flow and fluid density by

W=Q

(3-2)

All flowmeters measure the fluid flow through the flowmeter

body, but some flowmeters measure volumetric flow (Q) directly,
while others determine volumetric flow by measuring velocity
(v). Mass flowmeters measure mass flow (W) directly, while other
flowmeters are inferential in nature and measure a combination
of mass, velocity, and density.
As shown in Table 3-1, there are four types of flow measurements
that utilize various flowmeter technologies (see Figs. 3-1 through
3-10). Selection of the appropriate flowmeter technology to perform the proper flow measurement for the application is often the
key to controlling the correct process variable.

Example 3-2:
An understanding of the process is essential in the selection of
the best flowmeter type for a given application.

FIG. 3-1.

imo-spitzer-03.indd 9

Oval gear positive displacement flowmeter

9/15/09 1:43 PM

10

Field Measurement Devices

FIG. 3-2.

Magnetic flowmeter (Courtesy ABB)

For example, raw materials feeding a chemical reactor may be

measured best with mass flowmeters because chemical reactions
are mass related. However, the high cost of a mass flowmeter
may not be necessary for all of the raw materials if, for example,
the reaction is unaffected by some excess material that will be
recovered.
A flowmeter that measures volume or velocity may be applicable
to filling a tank. If a tank truck is to be filled to a certain shipping
weight, a mass flowmeter may be appropriate; however, if the
operating fluid density does not vary appreciably, a volumetric or
velocity flowmeter may exhibit acceptable performance.

imo-spitzer-03.indd 10

9/15/09 1:43 PM

Field Measurement Devices

FIG. 3-3.

11

Vortex shedding (oscillatory) flowmeter (Courtesy ABB)

In addition, to achieve optimum accuracy, most flowmeters require

that the fluid pass through the flowmeter body with a known
velocity profile. This velocity profile is usually attained by:
(1) installing the flowmeter to avoid two-phase flow (by keeping
liquid flowmeters free of gas and gas flowmeters free of liquids);
(2) providing an appropriate number of pipe diameters of straight
run upstream and downstream of the flowmeter body; and (3) operating the flowmeter in the hydraulic range in which the it performs

imo-spitzer-03.indd 11

9/15/09 1:43 PM

12

Field Measurement Devices

FIG. 3-4.

Turbine flowmeter

accurately. Straight run requirements vary with flowmeter technology and design; therefore, the use of rules of thumb should be
avoided. It should be noted that improper installation (and lack of
attention to detail) can easily cause a flowmeter to measure inaccurately or fail to operate.
Most flowmeters provide linear output signals; that is, the output
signal is directly proportional to flow. Some flowmeters (most
notably differential pressure flowmeters) have nonlinear output
signals that vary with the square of flow.
It should be noted that the Coriolis mass flowmeter can be
used to simultaneously measure the flow and the density of the
fluid. Conceptually, the density measurement can be used to
infer fluid composition by extrapolating the measured density

imo-spitzer-03.indd 12

9/15/09 1:43 PM

Field Measurement Devices

13

FIG. 3-5. Ultrasonic flowmeter

FIG. 3-6.

Coriolis mass flowmeter

between densities of the pure fluids. In practice, not only do the

densities of the pure fluids vary with temperature, but the fluids
may be miscible in varying degrees at different temperatures.
Despite the complexity of the calculations, inferring fluid composition from the measured density can be an important online
process quality measurement.

imo-spitzer-03.indd 13

9/15/09 1:43 PM

14

Field Measurement Devices

FIG. 3-7.

Thermal flowmeter (Courtesy Kurz Instruments)

Level Measurement
Level measurements are performed on vessels to measure the
inventory of material in vessels. Most level measurement devices
fall into one of three categories: those that measure level, those
that infer level, and those that measure interface level.
Level measuring devices (see Figures 3-11 through 3-16) that
measure level directly include those that sense where the change
of phase occurs in the vessel. The techniques used include
floats, ultrasonics, radar, resistive tapes, capacitance, displacers,
nuclear, and the like. Not all devices are applicable to a given
application because they may be dependent upon the specific
properties and characteristics of the process fluid.
Differential pressure level measurements detect the liquid head
between a tap near the bottom of the tank and the gas space
above the tank contents. This measurement does not sense at the
location of the phase change, but rather is used in conjunction
with the fluid density to infer the level in a vessel. Inaccuracies
can occur because of temperature, pressure, and/or composition
changes that cause variations in the fluid density.
Interface level measurements are usually more difficult to
achieve than the level measurements described above. A differential pressure transmitter or a submerged displacer can
be installed across the interface to sense the overall density

imo-spitzer-03.indd 14

9/15/09 1:43 PM

Field Measurement Devices

FIG. 3-8.

imo-spitzer-03.indd 15

15

Orifice plate (differential pressure) flowmeter

9/15/09 1:43 PM

16

Field Measurement Devices

FIG. 3-9.

Target flowmeter

difference that occurs as the interface level changes. Capacitance technology can be used, if the liquids at the interface have
sufficiently different electrical properties.

Example 3-3:
When the fluid density does not vary appreciably, a differential
pressure level measurement will suffice for measuring the liquid level in a storage tank that is dedicated to the fluid. However, if the tank were to be used for different fluids that have
different densities, the differential pressure level measurement
would be inaccurate, and a technique that measures level
would be appropriate.

Pressure Measurement
Pressure transmitters are typically installed on pipes and
equipment to measure pressure and to infer conditions inside
the equipment (see Figure 3-17). These transmitters are
extremely reliable when properly installed and not abused (by
steaming, for example); however, the pressure tap can plug and
the transmitter can be affected by corrosion. In a given application, suitable materials of construction, the use of diaphragm
seals, and proper installation can eliminate many of these
problems. Although pressure instrument failure and/or

imo-spitzer-03.indd 16

9/15/09 1:43 PM

Field Measurement Devices

FIG. 3-10.

17

Rotameter (variable area) flowmeter (Courtesy ABB)

continued maintenance can be an indication of a process problem (which should be investigated), the problem often lies
with instrument installation and/or selection.

Example 3-4:
Typical of gas service, the pressure transmitter is located
above the pressure tap and valve that are installed on the side

imo-spitzer-03.indd 17

9/15/09 1:43 PM

18

Field Measurement Devices

FIG. 3-11.

Float level transmitter

of a steam pipe. This installation will damage the transmitter

because it allows live steam to enter and overheat the instrument. For hot condensable service, the transmitter should be
installed below the tap (similar to a liquid application) so that a
liquid condensate seal forms to protect the transmitter by isolating it from the live steam.

Temperature Measurement
A number of techniques are used to measure the temperature of
industrial processes. These include many types of thermometers,

imo-spitzer-03.indd 18

9/15/09 1:43 PM

Field Measurement Devices

FIG. 3-12.

19

Ultrasonic level transmitter (Courtesy Gulf Publishing Company)

thermocouples, thermopiles, thermistors, solid-state sensors,

resistance temperature detectors (RTD), and infrared pyrometers
(see Figures 3-18 through 3-20).
A number of temperature scales are in common use and misunderstandings can occur when temperatures are expressed in
degrees without reference to the temperature scale. In process
applications, misunderstandings most often occur between temperatures expressed as degrees Fahrenheit and degrees Celsius,
however, absolute temperatures in the Rankine and Kelvin scales
can also cause confusion.
It should be noted that, due to conduction, convection, and, occasionally, radiation effects, accurate industrial temperature measurement is difficult to achieve. Loss of measurement accuracy
results from a lack of attention to detail and the difficulty in effectively compensating for these effects. In practice, these effects are
often ignored entirely.
Noncontact temperature measurement and fast response time
can be achieved by utilizing infrared pyrometer measurement

imo-spitzer-03.indd 19

9/15/09 1:43 PM

20

Field Measurement Devices

FIG. 3-13.

Resistive tape level transmitter (Courtesy Metritape)

FIG. 3-14(A).

imo-spitzer-03.indd 20

Capacitance level transmitternonconductive fluid

9/15/09 1:43 PM

Field Measurement Devices

21

technology. Measurement inaccuracy can be introduced when the

emissivity of the measured object is not accurately known.

Example 3-5:
It is desired to measure the temperature of the flowing fluid in
a pipe. A thermowell can be inserted into the pipe, and a thermocouple or RTD inserted into the well to measure fluid temperature. The mass of the thermowell and the thermal contact
between the inner wall of the thermowell and the sensor will
limit transient temperature response. When improved response
is necessary, a spring-loaded sensor design or a thermowell with
a built-in thermocouple design may be applied to improve this
response, albeit at higher cost.

Transmitter Technology
Smart transmitters that utilize microprocessor electronics are
more stable and can usually provide more accurate data than
their analog counterparts because of the inherent stability of
digital circuitry and cost-effectiveness of implementing more
sophisticated algorithms. One salient advantage of this technology is the ability to accurately measure the process variable over
small ranges.

FIG. 3-14(B).

imo-spitzer-03.indd 21

Capacitance level transmitterconductive fluid

9/15/09 1:44 PM

22

Field Measurement Devices

FIG. 3-15.

Displacer level transmitter (Courtesy Masoneilan Division of Dresser)

FIG. 3-16(A).

imo-spitzer-03.indd 22

Differential pressure transmitteropen tank

9/15/09 1:44 PM

Field Measurement Devices

FIG. 3-16(B).

23

Differential pressure transmitterclosed tank

FIG. 3-17. Pressure transmitter sensing module (Courtesy Rosemount Division

of Emerson Process Management)

imo-spitzer-03.indd 23

9/15/09 1:44 PM

24

Field Measurement Devices

FIG. 3-18.

Thermocouple

FIG. 3-19.

Resistance temperature detector

In addition, these smart transmitters can provide additional

information to the user through a hand-held configurator or a distributed control system interface (such as transmitter diagnostics
to aid in troubleshooting and to provide maintenance information) and can be electronically re-ranged.
Fieldbus and wireless transmitters utilize microprocessors and
networking technology to transmit the measurement in digital
format. All things being equal, digital signal transmission is
inherently more accurate than analog signal transmission because
it eliminates the conversion of the digital measurement to an
analog signal at the transmitter and the conversion of the analog
signal to a digital measurement at the receiving instrument. In
addition, the effects of analog signal noise during transmission

imo-spitzer-03.indd 24

9/15/09 1:44 PM

Field Measurement Devices

FIG. 3-20.

25

Thermowell installations

are also reduced. This technology can also enable the control system to monitor transmitter diagnostic information.

Summary
The quality with which a measurement is made in the context of
an application represents an extremely important aspect of control loop design. Not only should the measurement be the correct
measurement for the application, but it should also be a faithful
representation of the desired part of the process. If this is not the

imo-spitzer-03.indd 25

9/15/09 1:44 PM

26

Field Measurement Devices

case, a control loop may appear to be functioning properly while

the actual process can be out of control.

For Further Information

1.Gillum, Donald R. Industrial Pressure, Level and Density Measurement.
Research Triangle Park, NC: ISA.
2. Liptak, Bela G. Instrument Engineers Handbook (Vol. 1: Process Measurement and Analysis). Research Triangle Park, NC: ISA/CRC.
3. Miller, Richard W. Flow Measurement Engineering Handbook;
McGraw-Hill.
4. Spitzer, David William. Flow Measurement. Research Triangle Park, NC:
ISA.
5. Spitzer, David William. Industrial Flow Measurement. Research Triangle
Park, NC: ISA.
6. Selection Guide for Temperature Instrumentation. Research Triangle Park,
NC: ISA.

imo-spitzer-03.indd 26

9/15/09 1:44 PM

 Chapter 4

Controllers
Regulatory control is used extensively throughout industry to
control continuous processes. Properly applying this technology can yield control that is vastly superior to manual control.
Improved control can significantly improve yield, increase capacity, increase plant safety, and reduce the number of operators
required to safely operate the plant. The centralization of process
monitoring and process adjustments requires fewer operators and
allows those operators to control the plant more safely without
being physically located in the processing area.
The implementation of a regulatory control system requires
considerably more control system engineering expertise than is
required for a manual system. The cost of implementation is also
considerably higher because better quality instrumentation is
used, additional wiring and piping are necessary, and a control
room is often constructed.
Figure 4-1 illustrates the components of a basic regulatory control
loop. A measurement device senses the process and generates a
signal that is transmitted to the controller. The controller compares
the measurement with the desired set point that has been set by the
operator. The controller output manipulates the final control element, typically a valve. A number of disturbances can upset the
process and cause the measurement to move away from its set point.
The quality of a control system is, among other aspects, determined
by how well the system compensates for these disturbances to safely
maintain the process at its desired operating condition.

Example 4-1:
A regulatory control loop in common use is the cruise control
system installed on some cars (Figure 4-2). The vehicle speed
is sensed by the speedometer (measurement device), which
feeds the speed measurement to the speed controller. When
the desired speed is attained, the operator engages the speed

27

imo-spitzer-04.indd 27

9/15/09 10:45 AM

28

Controllers

FIG. 4-1.

Components of a basic regulatory control loop

FIG. 4-2.

Cruise control loop

control, which also sets the speed set point. The speed controller then attempts to maintain the vehicle at constant speed
by increasing the depression of the gas pedal (final control

imo-spitzer-04.indd 28

9/15/09 10:45 AM

Controllers

29

element) when the speed decreases and decreasing the depression of the gas pedal when speed increases. Note that, while
moving the gas pedal, the final control element manipulates
fuel flow.
This would seem to be a straightforward control system until the
effects of disturbances are considered. What if the road becomes
slippery and control of the car is lost because the speed control
accelerated too much? What if the driver stops suddenly, but
the drivers foot is resting in such a position that it is difficult to
move to the brake pedal quickly? Such disturbances (although
having a small probability of occurring) are not handled well by
this control system, and another control loop (anti-lock brakes,
for example) may be appropriate.

Controlling Instruments
Controlling part of a process entails maintaining a certain
measured parameter called the process variable (PV) at its
desired value or set point (SP). This is achieved by manipulating the process in such a manner that the process variable
reaches the set point. A controller is a device that compares
the process variable to the set point and generates an output
signal (OUT) that manipulates the process in order to make the
process variable equal to its set point. Mechanical, pneumatic,
electronic, microprocessor-based, and fieldbus controllers
are available, each potentially providing progressively more
sophisticated performance.
Most controllers display the set point, the process variable,
and the output; they can be changed from the manual (MAN)
to the automatic (AUTO) mode and vice versa by the operator.
In the manual mode, the operator can directly adjust the controller output, but set point changes made by the operator have
no effect on the controller output. When in the automatic mode,
the controller compares the process variable to the set point and
utilizes its control algorithms to generate the controller output,
but output changes made by the operator have no effect on the
controller output.
Most controllers commonly used in the process industries
contain the features described above (see Figure 4-3) and use
hardware or software devices to change modes and adjust the
controller set point and output. Mechanical and pneumatic controllers may or may not contain all of the above features.

imo-spitzer-04.indd 29

9/15/09 10:45 AM

30

Controllers

FIG. 4-3.

Typical microprocessor-based controller faceplate

Local Control
A local regulatory control loop, as depicted in Figure 4-1, is
located in the field and may indicate and/or record locally.
These local loops typically consist of one or more pieces of
equipment; the number depends upon whether the instrument is designed to combine functions, such as in the case of
a mechanical float valve assembly that measures, controls, and
manipulates the control valve using essentially one piece of
equipment.

imo-spitzer-04.indd 30

9/15/09 10:45 AM

Controllers

31

Loop sophistication (that is, the decision to use mechanical,

pneumatic, and/or electronic devices) is typically dependent
upon the available power source, the desired control loop performance, and cost.

Example 4-2:
Figs. 4-4, 4-5, and 4-6 illustrate progressively more sophisticated local tank temperature control loops. The mechanical
control loop (Figure 4-4) is a self-contained regulating valve that
requires no external power source. Expansion of the fill fluid
in the temperature probe opposes a spring (set by the operator) and mechanically operates the steam valve (manipulating
steam flow), thereby controlling tank temperature. Rudimentary
mechanical adjustments may be available to tune the loop,
but these adjustments are often made at the factory and are not
touched in the field.
Superior control can be obtained by using the system shown in
Figure 4-5, in which the expansion of the fill fluid in the temperature probe is compared to the set point, generating an output signal to the control valve. Mechanical adjustments can be made to
the controller to improve system performance.
Further control improvement can be obtained by using the loop
illustrated in Figure 4-6, in which electronic devices are used for

FIG. 4-4. Local regulatory control loop: tank temperature control with mechanical temperature control valve

imo-spitzer-04.indd 31

9/15/09 10:45 AM

32

Controllers

FIG. 4-5. Local regulatory control loop: tank temperature control with pneumatic controller and control valve

FIG. 4-6. Local regulatory control loop: tank temperature control with electronic controller and pneumatic control valve

their improved accuracy, speed, stability, and flexibility. A wider

range of controller adjustments than those available on local pneumatic controllers can used to further improve system performance.

imo-spitzer-04.indd 32

9/15/09 10:45 AM

Controllers

33

Centralized Control
A centralized regulatory control loop performs the functions
illustrated in Figure 4-1, where the control loop consists of a measurement device, a controller, and a final control element. Signals
from the measurement device and to the final control element are
routed, using pneumatic or electronic technology, to and from a
controller located in a central control room or on a centralized
control panel. A recorder and/or alarm annunciator are often
mounted on the control panel near the controller to record and/or
alarm the process variable.
Analog Signal Transmission neumatic or electronic techniques
can be used to represent the magnitude of an instrument signal
in a continuous manner. The 3-15 psig (0.2-1.0 bar) pneumatic
signal that can be used to represent 0-100 percent signal is
transmitted from one location to another using pneumatic tubing. Pneumatic repeaters may be necessary to reduce the time
required to transmit a signal through long tubing runs. Electronic signals, such as the 4-20 mA signal that can represent
0-100 percent signal, provide virtually instantaneous signal
transmission.
Digital Signal Transmission In a regulatory control loop, digital

signal transmission is functionally equivalent to analog signal transmission (described above) except that the controller
receives its measurement data and/or sends its output signal
as a digital signal; that is, the controller communicates with
the transmitter and/or final control element digitally where the
controller and the communicating device(s) are designed (and
configured) to communicate using a common protocol. Digital
signal transmission using smart transmitters can improve
measurement accuracy and diagnostics and can provide additional maintenance features. Digital communications with the
controller can be implemented using an appropriate controller
or distributed control system with appropriate digital communication hardware.

Example 4-3:
Figure 4-7 illustrates a pneumatic regulatory level control loop
in which the level transmitter and control valve are located in
the field and the controller is located in the control room. In this
loop, the level transmitter transmits a 3-15 psig (0.2-1.0 bar) signal that is proportional to the tank level, to the controller process
variable input. The controller operates the control valve (manipulating flow) by transmitting a 3-15 psig (0.2-1.0 bar) signal to the
control valve actuator.

imo-spitzer-04.indd 33

9/15/09 10:45 AM

34

Controllers

FIG. 4-7.

Centralized analog regulatory control: pneumatic level control loop

FIG. 4-8.

Centralized analog regulatory control: electronic level control loop

Figure 4-8 illustrates the same loop augmented by electronic

instrumentation (which may use analog or digital electronics)
communicating through 4-20 mA signals or digital protocol. The

imo-spitzer-04.indd 34

9/15/09 10:45 AM

Controllers

35

FIG. 4-9. Distributed control system regulatory control: electronic level control loop

electronic signal is converted to a pneumatic signal in the field to

operate the valve.

Distributed Control
The implementation of regulatory control using a distributed control system (DCS), programmable logic controller (PLC), or distributed fieldbus technology is functionally equivalent to the centralized control described above; however, the system is designed
in such a manner that the measurement input, the controller
software, and the control valve output may be in different plant
locations. Controller information can be made available on the
display(s) located in the control room and (potentially) throughout the plant. Control algorithms can be implemented in different
locations such as a centrally or remotely located DCS (Figure 4-9)
or PLC, or in a fieldbus instrument in the field.

Summary
Understanding the fundamental qualitative operation of controllers and the methods by which measurement signals are transmitted to the controller are important. These concepts are especially
important for locating a problem when troubleshooting the control loop.

imo-spitzer-04.indd 35

9/15/09 10:45 AM

imo-spitzer-04.indd 36

9/15/09 10:45 AM

 Chapter 5

Final Control Elements

Final control elements are those devices that manipulate the process. The most common final control element applied to industrial processes is the control valve; however, in some applications, a number of final control element technologies can be used.
These alternate technologies, primarily variable speed drives, can
often provide performance and/or features that are superior to
those of control valves.
IT SHOULD BE CLEARLY NOTED THAT VIRTUALLY ALL FINAL
CONTROL ELEMENTS MANIPULATE FLOW. Varying the opening of a control valve, the opening of a damper, or the speed of
a variable speed drive alters process flow. Heating (cooling) is
effected by manipulating the flow of the heat (cooling) source,
such as steam, hot oil, fuel, electricity, cooling water, brine, and
the like.
This chapter discusses the importance of the final control element
as it relates to the control loop. Further insight into the operation
and application of final control elements may be obtained from
the references.

Control Valves
Control valves are devices that dissipate hydraulic energy in a
controlled manner to achieve a desired flow. The flow through
a piping system without a control valve will be determined by
where the pump curve (which defines pump performance) intersects the system curve (which defines piping head and friction
losses). See Figure 5-1. The flow through the system can be varied
by throttling a control valve, effectively increasing the piping
friction. See Figure 5-2. In this manner, flow can be manipulated
from a minimum of no flow to the maximum flow produced when
the control valve is in its fully open position. Note that a fully
open control valve still throttles the piping system somewhat.

37

imo-spitzer-05.indd 37

9/15/09 10:51 AM

38

Final Control Elements

FIG. 5-1.

Operating point of unthrottled piping system

FIG. 5-2.

imo-spitzer-05.indd 38

Operating point of throttled piping system

9/15/09 10:51 AM

Final Control Elements

39

Control valve selection and sizing constitute a challenging process involving many considerations. The control valve should
be sized to pass the maximum (design) flow at a low enough
pressure drop so as to require no more pressure than the pump
is capable of delivering, yet maintain a high enough pressure
drop to effect good control. Undersized valves will not pass
enough flow, which presents an immediate problem. Oversized
valves usually operate nearly closed resulting in poor control
of flow because small valve position changes result in large
changes in flow.
It is not uncommon to violate one of the above constraints, especially when using high-capacity valves, such as ball or butterfly
valves, that are smaller than the line size. On start-up, construction debris can lodge in a valve and make the valve appear to be
undersized.
Another concern is the hysteresis associated with control valves
that can cause the valve not to operate smoothly. As a result of
the tightening of valve fugitive emission regulations, valve leaks
will increasingly be addressed by adjustment of the packing to
reduce leakage. The problem occurs when the person adjusting
the packing uses the philosophy that if a little is good, more is
better and adjusts the packing extra tight so it is less likely to
leak. What is not realized is that after the adjustment, the valve
may operate sluggishly or not move at all.
Control valves are typically remotely operated using pneumatic diaphragm actuators. A current-to-pneumatic (I/P) converter is used to convert the electronic signal to a pneumatic
signal that is applied to the actuator. The electronic signal can
be varied, which in turn varies the pneumatic signal, which in
turn strokes the valve.
It should be noted that the electronic signal varies the pneumatic signal to the control valve actuator. Valve travel is in
response to the pneumatic signal, and the relationship between
the electronic signal and valve position is not necessarily linear.
Valve position is dependent upon the forces acting on the valve
and the actuator, and the valve may exhibit hysteresis, where
the same signal may locate the valve in different control positions, depending upon the direction of travel, fluid forces, and
friction. Further, an undersized valve actuator can cause control
problems, such as when the valve cannot be opened far enough
to provide the desired flow.

imo-spitzer-05.indd 39

9/15/09 10:51 AM

40

Final Control Elements

Valve positioners are installed on some control valves to

reduce hysteresis and valve position inaccuracy described
above. These devices are self-contained feedback control
loops that measure the actual valve position, detect the
desired valve position, and provide the proper pneumatic
actuator pressure to move the valve to the desired valve position. Thus, the electronic signal and the valve position are linearly related.
The relationship between valve position and valve capacity is
dependent upon the valve trim characteristic (such as linear,
equal percent, or quick opening). In addition, the relationship between valve position and the actual (installed) valve
capacity depends upon the valve characteristic and the actual
hydraulics of the piping system. Selection of a valve trim
characteristic that results in an inappropriate installed valve
charecteristic for a given piping system can cause control
problems.
The control valve should move to its proper failure mode on loss
of power or instrument air. Most valves in industrial applications
fail closed (FC), however, many valves fail open (FO) and or failin-last-position (FLP) depending upon the application.

FIG. 5-3.

imo-spitzer-05.indd 40

Operating point of piping system throttled with a variable speed drive

9/15/09 10:51 AM

Final Control Elements

41

Variable-Speed Drives
Most variable-speed drives vary motor speed, which in turn varies the speed of the load, however, some variable-speed drives
vary only the speed of the load. Significant energy savings can be
achieved using either technology to generate only the amount of
hydraulic energy needed by the process; however, varying motor
speed will typically yield larger savings and exhibit little of the
hysteresis that can plague control valves. In many applications,
the installed cost of variable-speed drives can be less than that of
control valve installations.
Most applications are electronic, variable-frequency drives
applied to centrifugal loads where energy consumption is related
to the cube of motor speed. Therefore, operating the motor at 90
percent of full speed (a modest reduction) will reduce energy
consumption by approximately 27 percent, illustrating the significant magnitude of potential energy savings.
Variable-speed drives cannot provide the tight control valve shutoff capability of a control valve, but flow in the reverse direction
can be reduced by installing a check valve downstream of the
pump or fan. When tight shutoff is required, an on/off valve may
be required.
Variable-speed drive technology can reduce the number of leak
paths in the process, reduce maintenance costs, and provide significant energy savings in many applications. However, this technology is often not applied because: (1) control valves are applied
routinely; (2) many instrument engineers do not understand the
performance of hydraulic equipment; and (3) the instrument engineer typically does not have as much control over the application
of this technology as compared to control valves.

Example 5-1:
Control valves and variable-speed drives can be used to throttle
flow. These systems perform the same function, that is, providing
the desired flow. However, the variable-speed drive system generates only the hydraulic energy necessary to deliver the desired
flow. The full-speed pump will generate excess hydraulic energy
that is dissipated across the control valve to deliver the desired
flow. Varying the speed is more energy efficient and can eliminate
leak paths and reduce hysteresis while subjecting the pump and
piping to lower pressures that can reduce the potential for leaks.
Compare Figures 5-2 and 5-3, and see Figure 5-4.

imo-spitzer-05.indd 41

9/15/09 10:51 AM

Final Control Elements

FIG. 5-4.

Throttling flow by different methods

42

imo-spitzer-05.indd 42

9/15/09 10:51 AM

Final Control Elements

43

Summary
In many applications, lack of attention to detail during control
valve selection can result in potential control problems, potential
safety hazards, and excessive maintenance requirements, which
can compromise the effectiveness of the control system.

References
Borden, Guy Jr., Friedmann, Paul G. Control Valves. Research Triangle Park,
NC: ISA.
Control Valve Handbook. Fisher Controls
Spitzer, David William. Variable Speed DrivesPrinciples and Applications
for Energy Cost Savings. Research Triangle Park, NC: ISA.

imo-spitzer-05.indd 43

9/15/09 10:51 AM

imo-spitzer-05.indd 44

9/15/09 10:51 AM

 Chapter 6

Field Equipment Tuning

Process and disturbance dynamics can vary greatly from process
to process, and adjustments are available to compensate for these
differences. A common misconception is that control loop tuning is the adjustment of controller tuning parameters and that the
controller tuning parameters compensate for any and all process
anomalies. What is often overlooked is the requirement to verify
the appropriateness, the location, the installation, and the adjustment of every device in the control loop, including the process
equipment, field instrumentation, signal tranmission, controller,
and final control element.
Controller tuning parameter adjustment, which is but one facet
of the control loop tuning process, should usually occur after the
other devices in the loop have been verified for appropriate operation. Even with the wide range of controller tuning parameter
adjustments available, an inappropriate or misadjusted instrument in the control loop will cause the control loop to function
in a less than optimal manner, despite the appearance of proper
operation. In many control loops, no amount of controller tuning
parameter adjustment will adequately compensate for inappropriate equipment, instrumentation, or installation. Unfortunately,
it is all too common to start control loop tuning by adjusting
the controller tuning parameters without giving a thought to the
remainder of the devices in the control loop.
It should be understood that many adjustments are available
throughout the control loop and that these adjustments are present in the process equipment, the sensor, the transmitter, the controller, and the final control element. These adjustments may not
be described as such in an instrument instruction manual, especially in the case of field instrumentation, where location and
installation effects can adversely affect performance of an instrument that, in turn, affects the operation of the loop.
Therefore, consideration should be given to all parts of the control loop, because any of these adjustments may be the factor

45

imo-spitzer-06.indd 45

9/15/09 12:03 PM

46

Field Equipment Tuning

that limits the operation of the entire control loop and, perhaps,
the entire control system.
This chapter discusses the importance of the application, installation, and adjustment of field equipment as it relates to the control
loop. Further insight into the various types of field equipment
may be obtained from the references cited in the appropriate
chapters.

Control Loop Tuning Procedure

Control loop tuning should be performed in a logical procedure
to ensure that all devices in the loop operate properly and that
the loop performs satisfactorily in conjunction with the process.
Failure to do so can result in less than optimal performance,
given the investment in equipment. Note that the appropriate
responses of the equipment and the instruments to manipulation
are dependent upon the nature of the process.
PROCEDURE:
1. Prior to start-up, verify that the equipment and each instrument
in the loop is appropriate for the application at hand and that the
instruments intercommunicate through appropriate technology.
2. After the equipment is operating properly during start-up, verify
the sensor, the transmitter, and the signal transmission performance at the input to the controller with the controller in manual
mode. Similarly, check the operation of the final control element
by manipulating the controller output with the controller in
manual mode.
3. With the process operating, adjust the controller tuning parameters for automatic operation.

Equipment Adjustments
Most mechanical equipment requires adjustment for proper operation. Attempts to start up instrumentation on equipment that is
not functioning properly are usually futile. Be sure that equipment is in proper working order before attempting to make instrument adjustments.

Sensor Adjustments
Few sensors have mechanical or electrical adjustments in
the classical sense; however, their location and speed of

imo-spitzer-06.indd 46

9/15/09 12:03 PM

Field Equipment Tuning

47

response can be altered. Sensors should be located WHERE

THE PROCESS VARIABLE IS TO BE CONTROLLED. This
approach may seem trite, but many times the process variable
is measured at an inappropriate location for any of a number
of reasons.

Example 6-1:
The temperature of the contents of a vessel should be measured
in the liquid in the bottom of the vessel, that is, in the location
most indicative of the liquid temperature. However, the thermowell and sensor may actually be installed in an existing well
in the discharge piping to avoid the expense and inconvenience
of welding a nozzle on the vessel. Therefore, the temperature is
not being measured in the desired location. Aside from process
considerations, such as the requirement that the liquid be flowing to continually transport liquid of the proper temperature to
the point of measurement, additional complications may occur
because of improper or inappropriate insulation.
Sensor Damping In addition to appropriately locating the sen-

sor, the sensor may require mechanical adjustments to make

its response to the process more appropriate. When this is not
possible, it may be desirable to employ another measurement
technique.

Example 6-2:
Temperature measurement of a pipe is achieved by installing
a thermowell in a piping tee. Laboratory data confirms that
the temperature measurement is in error. Examination of the
installation reveals that the thermowell with a 3-inch extension is installed in a flanged tee located 12 inches away from
the liquid flow. Therefore, the location of the measurement is
in a stagnant pipe, the temperature of which is not indicative
of the flowing liquid. The measurement will not be responsive
to the process, not because it is performed in an inappropriate location in the process, but rather because the sensor is
insulated from the process. This sluggish response can be remedied by installing a thermowell of appropriate length and, if
additional speed of response is desired, applying a thermowell
with an integral thermocouple.
At times, mechanical damping may be necessary to make the
sensor less sensitive to the process. For example, a block
valve that connects the process to a pressure sensor may be

imo-spitzer-06.indd 47

9/15/09 12:03 PM

48

Field Equipment Tuning

partially closed to dampen the effect of line pressure pulsation

caused by a pump.

Transmitter Adjustments
The basic function of the transmitter in the regulatory control
loop is to MEASURE THE DESIRED PROCESS VARIABLE,
which, in the case of regulatory control, is the process variable
that will be controlled. Due to the nuances of process measurement techniques and transmitter design, measuring the
desired process variable is not as easy or as straightforward as
it may seem.
Transmitter Calibration Range Most transmitters allow adjust-

ment (calibration) of the process zero and process span, which

correspond to the zero and span of the transmitted signal,
respectively. Assuming these adjustments are made to the
correct specifications in a technically correct manner, they
still may be inappropriate with regard to their application to
the process.
The selection of an inappropriate calibration range typically
occurs when the transmitter is calibrated for use under all possible operating conditions, including extreme conditions such as
start-up and emergency shutdown. The use of a large measurement range can cause significant process variations that occur
near normal operating conditions to result in small measurement
signal changes. Therefore, the process can appear to be under
control when the process variable is displayed using the transmitter range, such as in the case of a controller or recorder, potentially masking the importance of the variations.

Example 6-3:
A product in solution can be concentrated by boiling the solution
to remove some or all of its solvent. Temperature is often used to
infer the product composition of the concentrated product and
control heat input. See Figure 6-1.
The composition of a hypothetical product composition as a
function of temperature is:

imo-spitzer-06.indd 48

Temperature (degrees Celsius)

Composition (percent)

150.0
150.5
151.0

45
50
55

9/15/09 12:03 PM

Field Equipment Tuning

FIG. 6-1.

49

Concentrator temperature control

To handle start-up, operating, and over-temperature conditions, it

would be reasonable to specify a temperature transmitter with a calibration of 0-200 degrees Celsius. Therefore, a one degree Celsius
variation in temperature will affect product composition by 10 percent.
However, the one degree variation is only 0.5 percent of the
transmitter range. Examination of this variation on a temperature recorder or indicator with a 0-200 degree Celsius range will
appear to be a straight line, indicative of a process that is under
control, despite the significant composition variation. Further,
measurement accuracy typically deteriorates as the calibrated
range increases, potentially causing process control problems
during normal operation.
Application of a smart transmitter that is capable of accurately measuring the desired temperature range (perhaps 145
to 155 degrees Celsius) could be used to implement the regulatory control loop shown. Notwithstanding more sophisticated
approaches, the smart transmitter approach, even if in addition
to the 0-200 degree Celsius transmitter, would probably be cost
effective due to its ability more accurately to infer composition
and better correct for fluctuations in composition.

imo-spitzer-06.indd 49

9/15/09 12:03 PM

50

Field Equipment Tuning

Transmitter Damping Many transmitters have damping adjustments to reduce the effect of process noise on the output signal. In most applications, the transmitter damping adjustment
is initially set to its default value. During start-up and operation, transmitter operation can be assessed and appropriate
damping adjustments can be made sparingly to increase transmitter responsiveness while maintaining an acceptably low
noise level.

In addition, some flowmeters have a transmitter dropout feature

that sets the flowmeter output to zero when the flow is below a
given percent of full scale (typically 12 percent) to reduce noise
and false signals at low flows.
Signal Transmission For all practical purposes, electronic signal

transmission is instantaneous. However, errors can occur when

wires are improperly connected, the insulating property of the
wire is compromised, when ground loops occur, or when signal
strength is lost because of excessively long wiring. Digital communication is not prone to errors due to loss of signal strength
per se, but a loss of communication will occur if signal strength
is sufficiently low.
It should be noted that pneumatic transmission systems inherently exhibit damping due to the finite time required for the
trasmitter (and repeaters) to supply compressed gas through
the pneumatic tubing system. This effect may significantly
dampen the response of the transmitter system and mask significant process fluctuations. In addition, tubing leaks can
cause inaccurate measurements by shunting some of the signal
(pressure) to atmosphere.

Controller Input Adjustments

The transmitted measurement signal becomes the input to the
control device. The most prevalent input adjustment of a control
device is its damping adjustment. Depending on the technology,
this adjustment may be made by adjusting a valve, a potentiometer, or a filter time constant. Whereas the transmitter damping
adjustment can be used to reduce the noise level while maintaining transmitter responsiveness, the input damping adjustment
can also be used to dampen process transients that are caused by
actual process deviations that may affect the control. It is usually preferable to discover and remedy the cause of the transients
rather than to dampen them out. Therefore, input damping is
typically used sparingly.

imo-spitzer-06.indd 50

9/15/09 12:03 PM

Field Equipment Tuning

51

In most applications, the input damping adjustment should be initially set to its default value. During start-up and operation, control
loop operation can be assessed and appropriate damping adjustments can be made (if necessary) to stabilize the measurement.
Provided that sufficient adjustment is available, either the
transmitter damping adjustment or the input damping adjustment can be made to satisfy both the measurement and control
requirements. Note, however, that such an adjustment (especially if made at the transmitter) may mask important transients that could be caused by instability elsewhere in the process (and subsequently corrected). Individual adjustments that
utilize the objectives described above will (at least) provide a
signal that reflects the actual process variable, with noise filtration performed at the transmitter (probably utilizing a superior
algorithm developed by the transmitter/sensor manufacturer)
that allows transients to be measured by the control system
and analyzed.

Example 6-4:
A pressure transmitter is used to measure the pressure of a gas in
a vessel whose internal pressure is controlled. See Figure 6-2.
Due to the pulsation caused by the gas compressor, the undamped
measurement and recording are inherently noisy, as reflected by a
pressure recording that is a relatively wide band that varies with
time. The transmitter damping can be increased to dampen the

FIG. 6-2.

imo-spitzer-06.indd 51

Gas pressure control

9/15/09 12:03 PM

52

Field Equipment Tuning

measurement (only), thereby reducing the pulsation noise and the

width of the recorded band. The varying pressure that remains
may be caused by process considerations or, perhaps, by valve
hysteresis.
The input damping adjustment is still available if required, but it
is preferable to correct the cause of the variation before increasing
the input damping, which can mask the variation.
It should be noted that had the transmitter or input been
excessively damped, the actual pressure variations may not have
been apparent, and the effects of instability elsewhere may have
been masked. Further, a pneumatic, signal-transmission system,
with its inherent damping effect, may excessively dampen the
measurement system and create the same result.

Final Control Element Adjustments

In most industrial processes, the final control element is a control valve that is used to manipulate flow. Due to reduced costs,
potential energy savings, and other considerations, variable
speed drives are being applied to industrial processes to manipulate flow by varying the amount of mechanical energy put into
the system.
Control Valves In many industrial applications, the pneumatic

spring return actuator provides the mechanical energy to operate the control valve. It should be clearly understood that the
pneumatic signal that is applied to this actuator is nothing more
than thata pneumatic signal applied to the actuator. The value
of this signal is often misunderstood to be equal to the position
of the control valve. Factors in addition to the signal will affect
valve position, such as current-to-pneumatic (I/P) converter accuracy, process conditions (including differential pressure), control
valve hysteresis, and valve adjustments.
The addition of a positioner (which can be considered a valve
adjustment) introduces a local valve position control loop with
the valve signal as its set point. This loop manipulates the pressure to the actuator in such a manner that the valve is moved to
the valve position that is represented by the signal. Utilizing a
positioner, the signal is effectively indicative of valve position,
because the effects of the process conditions and valve hysteresis
are reduced.
Electrically operated control valves are used extensively in industry and exhibit similar characteristics in that they may or may

imo-spitzer-06.indd 52

9/15/09 12:03 PM

Field Equipment Tuning

53

not contain positioning circuitry. It should be noted that without

positioning circuitry, the valve position may change as a result of
process conditions.
Mechanical adjustments can be made to control valves, including
adjustment of the packing, the actuator stroke, the supply pressure, and the like.

Example 6-5:
A liquid flow control loop feeding a tank is operated in the manual mode. (See Figure 6-3). In this case, the manually adjusted
controller output sends a signal to the valve that is representative
of the air signal to the control valve actuator diaphragm.
This flow is known to be difficult to manipulate, because varying
the signal from 30 to 75 percent does not affect the flow. Examination of the valve reveals that the valve packing was tightened
excessively after an apparent packing leak, resulting in excessive
control valve hysteresis.
If the problem persists after making the proper packing adjustment (after repacking, if required), a positioner can be installed to
reduce the effects of hysteresis. If the packing is tightened excessively to reduce leakage, the installation of a control valve that
utilizes a different design may be considered.

FIG. 6-3.

imo-spitzer-06.indd 53

Liquid flow control

9/15/09 12:03 PM

54

Field Equipment Tuning

Variable-Speed Drives A number of variable-speed drive adjustments are available to the user that can be manipulated with
potentiometers or digitally, depending upon the design. Parameters commonly considered for adjustment for control purposes
include the minimum and maximum speeds, ramp-up and rampdown times, and skip frequencies. These adjustments set the
range of operating speeds, the rates at which motor speed can be
increased and decreased, and frequencies at which the drive will
not be allowed to operate. Variable-speed drives exhibit virtually no hysteresis similar to that caused by control valve packing,
because the motor makes small speed adjustments electronically.

Minimum and maximum speeds of liquid positive-displacement

equipment can usually be set at 0 and 100 percent, respectively;
however, gas positive-displacement equipment may be required
to operate above some (non-zero) motor speed to protect the
equipment from internal damage. The minimum speed of centrifugal equipment that overcomes a static head should be set to a
speed at which the static head generated at zero-flow conditions
is somewhat less than the static head that is to be overcome.
The ramp-up and ramp-down times determine the responsiveness of the equipment to a signal change. Factory default settings
are typically used, but some equipment may require more gentle
speed variation in order to protect the equipment and/or not to
overload the drive. Increasing the responsiveness of the motor
should be performed during start-up or during operation when
process conditions warrant and excessive equipment wear will
not occur.
In addition, certain types of equipment require the setting of skip
frequencies that inhibit the drive from operating at those frequencies that can cause damage due to resonance.

Summary
In order for a control loop to function to its fullest, all components of the control loop should be properly adjusted, including
the field equipment. Failure of the field transmitter to faithfully
measure and respond to process conditions compromises the
effectiveness of the entire control loop.

imo-spitzer-06.indd 54

9/15/09 12:03 PM

 Chapter 7

Regulatory Controller
Features
The regulatory controller compares the measurement with the
desired set point that has been established by the operator and
its output manipulates the final control elementtypically a
valve. Before adjusting the controller tuning parameters, a number of controller features should be enabled and adjusted to
ensure that the controller will function properly. Controller tuning performed before setting up the controller is often a frustrating and futile effort.

Anti-Reset Windup
In many processes, upset conditions can cause the controller
output to saturate at one of its extremes. The process is out
of control in the sense that the process variable cannot be
returned to its set point by manipulation of the final control element within the limits of the controller output. The attempt to
bring the process to its set point may require that the controller output remain at an extreme for an extended period of time,
causing it to wind up, that is, to saturate to a value in excess
of its nominal extreme.

Example 7-1:
The average steady-state flows in a condensate system are shown
in Figure 7-1. These flows may be typical of the operation but are
not representative of upset conditions.
Consider the sequence of events initiated when the de-aerator
makeup pump turns off. A few minutes later, the de-aerator
level alarm (not shown) warns the operator of a lower than
normal water level in the de-aerator. The operator determines
the cause of the problem and energizes the backup de-aerator
makeup pump.

55

imo-spitzer-07.indd 55

9/15/09 12:04 PM

56

Regulatory Controller Features

FIG. 7-1.

Condensate return system

Assuming that the tank was half full prior to a 5-minute loss of
pumping, 400 liters of condensate returned to a tank with only
100 liters of usable volume. Due to the rise in level, the makeup
water valve closed shortly after the pump turned off, in an unsuccessful attempt to lower the liquid level in the tank. By the time
the backup pump was energized, more than 300 liters of condensate had overflowed the tank.
During this sequence of events, the rise in tank level caused the
controller output to go to its minimum value in order to close the
valve. Although the minimum signal value is normally considered to be 4.00 mA, the integrating action (reset) of the controller
may cause the actual controller output to be as low as 2.5 mA, the
value at which the controller output electronics saturate.
When the tank level falls after the pump turns on, valve response
is delayed until the controller output integrates from 2.5 mA to
above 4.00 mA, causing a delayed, makeup water response that
can result in a more pronounced drop in tank level and subsequent loss of pump suction.

imo-spitzer-07.indd 56

9/15/09 12:04 PM

Regulatory Controller Features

57

A controller with an anti-reset windup feature would limit the

controller output to between 4.00 mA and 20.00 mA, which
would allow the valve to respond without waiting for the controller to come out of saturation. Similarly, controllers can exhibit
reset windup at full scale with as much as 27 mA.
While this example shows the effects of reset windup, it
also illustrates the potential need for a larger condensate
tank. However, more sophisticated controls and alarms may
allow use of the existing tank or complete elimination of
the tank. The process control engineer can aid in making
this determination.
Reset windup is an inherent feature of many controllers, but it
is most prevalent in pneumatic and analog electronic controllers in which special anti-reset windup circuits can be specified to limit the controller output signal to its nominal values.
Digital controllers and distributed control systems do not
require special circuitry and can perform this function
in software.

Bumpless Transfer (Set Point Tracking)

The controller output that manipulates the final control element can be adjusted manually by the operator or automatically
by the controller, depending upon whether the operator has
selected the manual or the automatic mode of the controller.
During switching between these modes, conditions can arise
that cause improper set points and outputs to be generated.
When the controller is in manual, the operator can directly
manipulate the output. At this time, the set point can be
changed, but these changes will not affect the output.
Although the loop may be stabilized in manual, when the
operator switches the controller to automatic, the controller algorithm adjusts the output in response to the difference
between the set point and the process variable. This can
cause an large upset to the process if the operator fails to set
the set point equal to the process variable before switching
to automatic.
In many applications, it is advantageous to unburden the operator from the inconvenience of having to set the set point equal
to the process variable. This can be accomplished through the
use of a bumpless transfer feature. During the switch from
manual to automatic, this feature causes the set point to equal

imo-spitzer-07.indd 57

9/15/09 12:04 PM

58

Regulatory Controller Features

the process variable before the control algorithm manipulates

the output. As a result, the control loop is switched from manual to automatic without a bump in the output signal. In some
controllers, the set point may follow the process variable in
manual (track the set point), effectively yielding the same result.
It should be noted that while the bump is avoided, the operator
should adjust the set point to the desired value. This, however,
is a function with which the operator should be familiar.

Example 7-2:
Consider the start-up of a drying oven where changing the oven
temperature rapidly can cause the refractory to fail. It is common for the operator to initially raise the firing rate over time
by manipulating the fuel valve with the controller in manual.
When the temperature is about half of the desired operating
temperature, the operator might switch the controller to automatic and let the controller do the work.
This may seem straightforward, but the controller set point may
have been left at the normal operating temperature, and the operator may have failed to set the set point equal to the process variable. In such a case, when the controller is switched to automatic
the set point may be hundreds of degrees higher than the oven
temperature, which will cause the fuel valve to open fully, thus
abruptly raising the oven temperature, and potentially causing
refractory failure.
The addition of the bumpless transfer feature would unburden
the operator and reduce the risk of refractory damage.
It should be noted that, whereas the bumpless transfer feature
would seem advantageous, it is not applicable to all processes.
For example, the start-up procedure of a flow control loop may
be to switch the controller to automatic with the proper set point
and allow the controller to manipulate a valve to achieve the
desired flow. Application of bumpless transfer to this process
would burden the operator with the additional work of setting the
set point to its proper value after switching to automatic, or manually stabilizing the loop near the proper set point before switching to automatic control.

Control Dead Band

When the process variable is near the set point, it is possible
for measurement noise to manipulate the controller output and

imo-spitzer-07.indd 58

9/15/09 12:04 PM

Regulatory Controller Features

59

the final control element. This can cause the process variable
to oscillate near its set point. Some controllers allow a control
dead band to be set around the set point such that control
action stops when the process variable is within this dead
band, the controller algorithm maintains the controller output
at its last value.
This feature can be helpful in reducing the effects of measurement noise on the process, because the noise is ignored when
the process variable is within the dead band. Note that the measurement signal need not be overdamped (as would be done to
reduce oscillations if this feature were not available), and the
controller can be responsive to relatively rapid process changes
outside the controller dead band. In general, the controller dead
band should set as small as practical.
The control dead band feature can also be used to improve
the control of processes that do not necessarily require tight
control. This may seem contradictory, but consider the control
of the condensate tank in Figure 7-1, which serves to absorb
batches of condensate returned after it is accumulated in local
tanks. The condensate tank level rises abruptly when a batch
is returned and falls gradually as the condensate is used while
there is no return.
Tight control would require that the level control valve be
closed (open) when the level is above (below) the setpoint.
Opening the control valve fully when the level is below the set
point (between batches) may put an operational strain on the
makeup water supply.
In this process, there is little need to tightly control level
because the condensate needs of the de-aerator are satisfied as
long as condensate is available; that is, what matters is not tank
level but rather that condensate is available and does not overflow the tank. Therefore, the condensate tank can be allowed
to fluctuate between (say) 40 percent level to ensure sufficient
pump net positive suction head (NPSH), and (say) 70 percent
level to allow volume for returning batches of condensate.
Note that it is not sufficient for condensate to be present in the
tank. Because the condensate may be near its boiling point, a
relatively low tank and high pump NPSH requirement dictates
maintaining the higher condensate tank level to avoid cavitation and subsequent pump damage. This additional requirement
reflects the process requirement that the control of condensate
tank level is not the process, but rather is the means by which

imo-spitzer-07.indd 59

9/15/09 12:04 PM

60

Regulatory Controller Features

the process is managed to allow the condensate be delivered to

the de-aerator.
Recognizing that tight control is not required, a set point of 55
percent and a controller dead band of 10 percent can be utilized
to allow the level to fluctuate with the process while keeping
the level under control, that is, between 40 and 70 percent level.
When in automatic, the controller output will maintain its last
output when the tank level is between 45 and 65 percent. Outside of these limits, the controller will manipulate the control
valve until the level returns within the limits, again holding its
last output.
Eventually, these excursions will cause the controller to reach
an output at which the level does not exceed the controller dead
band, sending a constant controller output to the makeup water
valve even though the batches are returning and the level is fluctuating accordingly. With the exception of an upset condition,
the makeup water flow will be relatively constant and should not
overload the makeup water system.

Controller Range
The measurement range of the process variable determines the
range over which the controller can control. Whereas previous generations of controllers required that the measurement
range and the controller indicator be identical, controllers that
utilize digital electronics allow the controller faceplate to indicate a range different from that of the measurement, as well
as a digital indication of the process variable. The adjustment
of a narrow indication range in which the process normally
operates allows more accurate set point adjustment and easier
detection of actual process deviations.

Example 7-3:
A temperature controller is used to control furnace temperature
at 1000 degrees Celsius. See Figure 7-2. The temperature transmitter range (and hence the controller range) is 0-1500 degrees
Celsius to measure start-up and upset conditions in addition to
the normal operating temperature of approximately 1000 degrees
Celsius.
If the controller indicated 0-1500 degrees Celsius, the loop
would appear to be under control even if process fluctuations
of 50-100 degrees Celsius were present. However, if the controller were adjusted to indicate from 900-1100 degrees Celsius, process fluctuations of 50-100 degrees Celsius could
easily be detected. Further, the digital indicator on the

imo-spitzer-07.indd 60

9/15/09 12:04 PM

Regulatory Controller Features

FIG. 7-2.

61

Simplified furnace chamber temperature control loop

controller faceplate would provide the operator with digital

temperature information at all times, including start-up and
upset conditions.

Direct/Reverse Control Action

In automatic operation, the controller output either increases
to increase the process variable (direct action) or decreases to
increase the process variable (reverse action).

Example 7-7:
Figure 7-3 is a simplified drawing of typical flow controls
(only) for a distillation column reflux tank. Note that the reflux
flow valve fails open, and the take-off flow valve fails closed.
This combination implements the process requirement to
reflux all material in event of an instrument air or instrument
signal loss.
The reflux flow controller is reverse-acting, that is, the controller output decreases (opening the fail-open flow valve) to
increase the process variable. The product take-off flow
controller is direct-acting, that is, the controller output

imo-spitzer-07.indd 61

9/15/09 12:04 PM

62

Regulatory Controller Features

FIG. 7-3.

Simplified reflux and product takeoff flow controls (only).

increases (opening the fail-closed flow valve) to increase the

process variable.

Failure Modes
Despite the many potential controller failure modes, only a
small number of practical failure modes are considered by the
user. The most likely is a controller power failure that affects
the control loop when power is removed and again when it
is restored.
Loss of Power The analog output of a controller has a polar-

ity, in the sense that the minimum controller output signal

may cause the final control element to be at its minimum or
maximum, depending upon the application and the nature of
the final control element. When power to the controller is lost,

imo-spitzer-07.indd 62

9/15/09 12:04 PM

Regulatory Controller Features

63

safety considerations are usually the overriding concerns used

to determine the failure position of the final control element
and, in turn, the polarity of the controller output signal. On
the other hand, safety and economics may be the driving
forces in determining how the controller is to respond after
power is restored.
Whenever possible, the final control element should move to
its failure position when the controller output signal is at its
minimum. Further, fail-safe operation requires that a failure
of the controller signal or the final control element power source
will cause the final control element to move to its (presumably
safe) failure position without the use of external power sources,
such as electricity, instrument air, and the like. When fail-safe
operation is desired, it is important to verify that the final control element moves to its failure position upon the loss of any
external power source.
Even though fail-safe operation may not be required in many
applications, the integration of this concept into the equipment
adds operational security and creates less confusion once it is
understood. In addition, it does not increase cost.
In a fail-safe installation, a loss of power to the controller
results in a loss of the controller output signal, which will cause
the final control element to move to its failure position regardless
of whether other power sources have failed. In addition, failure
of the external power sources that feed the final control element
will cause the final control element to move to its failure position
independent of the controller output signal.
Most electrically operated valves cannot be operated as failsafe because a loss of power or loss of signal will cause the
valve to remain in its last position (FLP), although some actuators will move the valve to a designated failure position on loss
of signal. Similarly, variable speed drives will turn off if power
is lost, but many may be configured to appropriately respond
to a loss of signal. Electrically operated valves are usually not
applied in processes that require fail open (F.O.) or fail closed
(F.C.) fail-safe operation in order to maintain safety. Variable
speed drives inherently stop when power is lost and can be utilized in fail-safe applications that require the final control element (pump) to fail off.
Restoration of Power When power is restored to the controller, the final control element should move to the position that
corresponds to the controllers present output signal. However,
all of the external sources to the final control element may
not be available, and the final control element may remain in

imo-spitzer-07.indd 63

9/15/09 12:04 PM

64

Regulatory Controller Features

its failure position. In either case, the controller is typically

in manual, but may be in automatic if the technology and the
design allow.
In most processes, it would seem to be acceptable for the controller to be in manual, generating a zero output signal after
power is restored. The operator could then manually adjust the
final control element or raise the set point in automatic. This
presents a potential problem in some applications, because
the operator may be faced with the simultaneous startup of
many control loops, in addition to the major process upset that
occurred during the momentary controller power loss. Therefore, after a power failure, it may be easier, safer, and more economical to allow as many control loops as practical to startup
with the controller in automatic with a preset or the last set
point, or in manual with a preset controller output or the output
prior to losing power.
A number of considerations should be weighed in the decision
regarding the controller output and automatic/manual status
after restoration of power. These considerations are dependent
upon the process and should be investigated in a formal hazard review. They include such items as: (1) the consequences
of the final control element overshooting to its maximum;
(2) or not moving immediately after restoration; (3) or not
being in automatic, and the like. After the restoration of power,
the decision to configure the instrument to force the controller
to manual or to return the controller to its mode of operation
prior to power failure depends upon knowledge of the process,
and how the measurement and final control element are integrated into that process.
Other Failure Modes It should be noted that other likely failure

modes may be found by evaluating the controller design and

technology. Examples include the effects of pneumatic tubing
leaks, dirty instrument air, electrical noise, etc.

Event-Related Controller Features

Controllers have traditionally operated in the analog domain,
but controller hardware and software developments have
enabled internal logic functions to influence the controller
in more sophisticated ways than those traditionally achieved
using external logic circuits and relays. The event-driven controller responses described are typical of many digital regulatory controllers, but some digital regulatory controllers use
configuration and programming techniques to allow customization of these features.

imo-spitzer-07.indd 64

9/15/09 12:04 PM

Regulatory Controller Features

65

Enabling Switches Controllers may contain a number of hard-

ware and/or software switches that enable and disable numerous controller functions such as alarm types, bumpless transfer
(set point tracking), hard output limits, output tracking, remote
set point control, and the like. It is not only important to
enable the desired functions but also to verify that they are disabled to ensure that the remaining functions do not interfere
with the desired features.

Example 7-4:
The default of a controller database may enable the bumpless
transfer function. Other desired functions may be enabled for a
specific application, however, if bumpless transfer is not desirable in this application, this switch should be disabled to avoid
operational problems.
Manual/Automatic Switch In some applications, it is desirable

for the controller to switch from automatic to manual and vice

versa as the result of a specific event. This feature could be used
to effectively locate the automatic/manual switch remote from
the controller so that the operator could switch the controller
while directly observing the process. In other applications, specific process conditions could trigger the switch.
Care should be taken to determine whether the condition or a
change in condition causes the controller to switch. This aspect
of the manual/automatic switch should be investigated while
it is appropriately installed to match the requirements of the
process, especially with regard to the ability of the operator to
manually manipulate the final control element after the switch
to automatican ability that may or may not be desirable in a
given application.
Output Preset Value Some applications require that the control-

ler be forced from automatic to manual as well as to a preset

output when a specific condition is present. When the condition
is not present, the controller switches from manual to automatic
and, after the switch, usually (but not always) allows the operator full control of the loop, including the ability to switch the
controller to manual.

Example 7-5:
An air flow control is used to pneumatically transport a powder
(see Figure 7-4). When the feed valve is off, the flow controller
is forced to manual, and the fan is forced to a preset speed to

imo-spitzer-07.indd 65

9/15/09 12:04 PM

66

Regulatory Controller Features

conserve energy. When the feed valve is operating, the controller

is switched to automatic to control the air flow at its set point.
Note that bumpless transfer is not desirable in this application.
Controller responses to the condition and to a change of condition should be investigated to ensure that the responses are
compatible with the process and (preferably) reduce the number
of actions required by the operator.
Output Tracking In some processes, it is desirable for the controller to be forced to manual and have its output forced to the
value of one of its analog inputs that represents a signal from
another source. This feature can be applied to allow the control
system to bypass the controller algorithm and directly manipulate the controller output when a constraint is reached.

Example 7-6:
In Figure 7-5, a gas is fed to the reactor at a flow rate set by
the operator. If the flow rate is excessive, there may not be sufficient time for the gas to react with the liquid in the reactor,
and the reactor pressure can increase due to an accumulation
of unreacted gas. Should this occur, the reactor pressure switch
will sense high pressure, send a contact closure to the flow
controller that will cause the flow controller to cease automatic operation, and output the calculated value present at its
tracking input that could be a function of the reactor pressure.
When operating in this manner, the controller is effectively

FIG. 7-4.

imo-spitzer-07.indd 66

Air flow control.

9/15/09 12:04 PM

Regulatory Controller Features

67

bypassed and its output is manipulated directly by the signal

at its tracking input.

Output Limits
Output limits can be used when it is desirable to limit the controller output to avoid unsafe, unstable, or undesirable operating
conditions. It should be noted that some controllers limit the
output while in all modes of operation; others limit the controller output only when the controller is in automatic operation.
A switch may be available in the controller to enable or modify
these functions.

Example 7-8:
The natural gas flow through the natural gas flow valve of a multifuel boiler cannot sustain combustion during operation under
low-load conditions because the valve closes to less than 25 percent open. To avoid flaming out the boiler under these conditions,

FIG. 7-5.

imo-spitzer-07.indd 67

Reactor gas flow control with output tracking.

9/15/09 12:04 PM

68

Regulatory Controller Features

a natural gas controller low-output limit can be adjusted to

approximately 28 percent. Making this output limit active only
during automatic operation would allow the complete removal
of natural gas when the controller is in manual so as to allow the
burning of other fuels.

Remote Set Point

In many applications, the controller set point is adjusted by the
operator. This need not be the case, however, because some controllers allow external signals to adjust the controller set point
and, in addition, allow the external signal to be scaled, using
ratio and bias adjustments. Enabling the remote set point feature
also enables a remote/local set point switch on the controller
faceplate that allows the operator to switch between the local
(operator-adjusted) set point and the remote set point.

Example 7-9:
Figure 7-6 depicts a flow control loop whose set point is generated by a signal from the computer when the controller is in
remote set point operation. Note that the operator can put the
controller in local control and adjust the set point, or put the
controller in manual control to directly manipulate the controller output.

Set Point Limits

Set point limits can be used to limit the ability of an operator (or
control system) to excessively raise or lower the controller set
point. While the set point limits may be active at all times, the
limits will have no effect on the process when the controller is
operated in manual because the operator can directly manipulate the controller output (independent of the set point).

Example 7-10:
The natural gas flow through the natural gas flow valve of a multifuel boiler can exceed the maximum firing rate of the boiler when
the natural gas flow is more than 75 percent of full scale. To limit
the natural gas flow in automatic operation and reduce the possibility of overfiring the boiler, the natural gas controller high set
point limit can be adjusted to approximately 75 percent of fullscale flow. Note that the operator can further increase the natural
gas flow while the controller is in manual operation.

imo-spitzer-07.indd 68

9/15/09 12:04 PM

Regulatory Controller Features

FIG. 7-6.

69

Flow loop with remote set point

Slew Rates
In some applications, abrupt manipulation of the set point or
output can upset the process, especially in processes that react
slowly to manipulation. To reduce process upsets, slew rate
adjustments may be available to limit the speed with which the
set point to the control algorithm can be changed (regardless of
the abruptness of operator actions) and the speed with which the
controller output is changed by the control algorithm. The output
slew rate may or may not be active when the controller is operating in manual.

Example 7-11:
During start-up, it was noted that when the set point of the
level control loop shown in Figure 7-1 was adjusted abruptly
by 10 percent, the level would oscillate for a period of time
before reaching a stable steady state. It was further noted that
adjusting the level in 1 percent increments produced
negligible oscillation. A set point slew rate adjustment could
be made to this controller to allow the control algorithm to
only gradually be affected by the operators abrupt 10 percent
set point change.

imo-spitzer-07.indd 69

9/15/09 12:04 PM

70

Regulatory Controller Features

Totalizer
Some controllers have the ability to totalize flow. Controller
adjustments should be made to properly scale the totalizer to the
flow signal and allow for its display. This feature may be useful
in some applications, but the inability of the controller to retain
the total during a power failure may preclude its use in some
controllers.

Summary
Regulatory controllers should be configured to reduce hardware
costs and increase safety and reliability. These features should be
thoroughly understood during the control loop design and implementation phases of a project.

imo-spitzer-07.indd 70

9/15/09 12:04 PM

 Chapter 8

PID Control
A regulatory controller is a device that maintains a process
variable at a desired value by manipulating a final control
element, typically a control valve. The controller output that
manipulates the final control element to achieve this end is
generated by a mathematical algorithm that compares the
process variable with the set point. The most common controller algorithm used to determine the controller output is the
PID algorithm.
PID regulatory controllers contain hardware and/or software
that implements proportional, integral (reset), and derivative
(rate) control. The controllers described herein contain all three
control algorithms, but individual controllers may not contain
some algorithms because not all are necessary for all control
loops, and some may even be detrimental to the operation of
certain control loops. Equipment costs of hardware implementations can be reduced by reducing the number of algorithms and
faster algorithm run time is achieved when fewer software algorithms are implemented.
The PID controller generates an error signal by taking the algebraic difference between the process variable and the controller
set point. This error signal, which may be positive or negative, is
used to implement the PID control algorithm.
Note throughout the following discussion that if the error is
zero (that is, the process variable and set point are equal), no
control action is taken, and the controller output will maintain
its last value.

Proportional Control
Proportional control action modifies the controller output by an
amount that is proportional to the error signal (see Fig. 8-1). Before
the step change on the left side of Figure 8-1, the process variable and the set point are shown to be equal in a controller with

71

imo-spitzer-08.indd 71

9/15/09 11:03 AM

72

PID Control

FIG. 8-1.

Response to step process variable changeproportional control

proportional-only control. As such, the controller output remains

at its previous value because no control action is necessary.
When the process variable changes, an error (e) develops, at
which time the proportional controller algorithm modifies the
controller output:

OUT(t + 1) = OUT(t) + (Kp e)

(8-1)

where Kp is the proportional controller gain and t is time. The

illustrated controller response is a step output change. If the controller gain (Kp) were unity, the process variable and the output
step changes would be identical. Because unity controller gain
is not appropriate for all loops, the controller gain (Kp) can be
adjusted to match the process.
Proportional band (PB) is often used to describe the proportional
controller adjustment, where

PB(%) = 100/Kp

(8-2)

The numeric value of proportional band is equal to the error as a

percentage of the process variable range that will result in a 100
percent output change.

Example 8-1:
A proportional controller with a controller gain (Kp) of 2 would
have a proportional band (PB) of 50 percent. An error signal of 50

imo-spitzer-08.indd 72

9/15/09 11:03 AM

PID Control

73

percent of the process variable range will result in a 100 percent

change in controller output.
Note that the above discussion is intended to illustrate the concept of proportional control by assuming a step process variable change that results in a constant error signal and a constant
controller output. In practice, the proportional control algorithm
will adjust the controller output in proportion to the instantanious error signal which will not be constant in the great majority
of applications.

Integral Control
Integral control action modifies the controller output by an
amount that is proportional to the integral of the error signal.
Integration is a function of calculus that can be thought of as taking the area under a curve. In another use of this concept, an integrator is used to totalize flow, which is nothing more than taking the area under the flow measurement curve (see Figure 8-2).
In the left part of Figure 8-3, before the step change, the process
variable and the set point are shown to be equal in a controller
with integral-only control. As such, the output remains at its previous value because no control action is necessary.
When the process variable changes, an error (e) develops,
at which time the integral algorithm modifies the controller
output:

FIG. 8-2.

imo-spitzer-08.indd 73

Flow totalization performed by integration

9/15/09 11:03 AM

74

PID Control

FIG. 8-3.

Response to step process variable changeintegral control (reset)

OUT(t + 1) = OUT(t) + Ki * (e * t)

(8-3)

where Ki is the integral controller constant. The illustrated

controller response is a ramp output change that is proportional to the integral of the error, that is, the area between the
process variable and set point curves. If the integral controller
constant (Ki) were unity, after one unit of time the output value
will have changed from its original value by an amount equal
to the error (e). Because a unity integral controller constant (Ki)
is not appropriate for all loops, the integral controller constant
(Ki) can be adjusted to match the process. The integral controller constant (Ki), often referred to as the reset adjustment, is
typically expressed in terms of minutes per repeat, that is, the
number of minutes necessary for the output to ramp an
amount that is equal to the error when the error is held constant. Note that this adjustment may be expressed as its
inverserepeats per minute.

Example 8-2:
A integral controller with an integral controller constant (Ki)
of 2 minutes per repeat with a constant 10 percent error signal
would ramp its output at a rate of 5 percent per minute, reaching an output signal that is 10 percent higher than the initial
output signal in 2 minutes.

imo-spitzer-08.indd 74

9/15/09 11:03 AM

PID Control

75

Note that the above discussion is intended to illustrate the

concept of integral control by assuming a step process variable
change that results in a constant error signal and a ramping controller output. In practice, the integral control algorithm will
be adjusting the controller output by integrating the instantaneous error signal, which will not be constant in the great
majority of applications.

Derivative Control
Derivative control action modifies the controller output by an
amount that is proportional to the derivative of the error signal.
The derivative, which is a function of calculus used to determine the slope (or rate of change) of a curve, can be loosely
defined as the change in value per unit time.
It should be noted that derivative control is used sparingly
because measurement noise, small upsets, and set point
changes can affect the error signal and potentially have a
large effect on the controller output, inducing possible upsets
and instability.
Before the step change in the left part of Figure 8-4, the process
variable and the set point are shown to be equal in a controller
with derivative-only control. As such, the output remains at its
previous value because no control action is necessary.
When the process variable changes, an error (e) develops, at
which time the derivative controller algorithm modifies the controller output:

FIG. 8-4.

imo-spitzer-08.indd 75

Response to step process variable changederivative control (rate)

9/15/09 11:03 AM

76

PID Control

OUT(t + 1) = OUT(t) + Kd * e / t

(8-4)

where Kd is the derivative controller constant. The illustrated

controller response is a pulse that is proportional to the magnitude of the derivative of the error, that is, the slope of the error
curve. If the derivative controller constant (Kd) were unity and
the slope of the error curve were constant, after one unit of
time the output value will have changed from its original value
by an amount that is equal to the rate of change of the error signal. Because a unity derivative controller constant (Kd) is not
appropriate for all loops, the derivative controller constant (Kd)
can be adjusted to match the process.
The derivative controller constant (Kd), often referred to as
the rate adjustment (referring to an algorithm that utilizes
the rate of change of the input), can be expressed in terms
of minutes, that is, the number of minutes necessary for the
output to change an amount equal to the rate of change of the
error signal when the rate of change of the error signal is held
constant.

Example 8-3:
The process variable of a derivative controller is ramping at a
constant rate of 1 percent per minute. With a derivative controller constant (Kd) of 10 minutes per repeat, the controller
output will change 1 percent in 10 minutes, or 0.1 percent
per minute.
Note that the above example is intended to illustrate the concept of derivative control by assuming that the error is changing at a constant rate. In practice, the derivative control algorithm will be adjusting the controller output by calculating the
rate of change of the instantaneous error signal, which will not
be constant in the great majority of applications.
Derivative on Process Variable In many applications, the deriv-

ative control algorithm using the error signal described above

can adversely affect the controller output when the set point
is adjusted, unless the operator puts the controller in manual
when the change is made or changes the set point slowly.
Some controllers can ignore error changes that are generated
by set point changes and improve the functionality of the controller while reducing the probability of the operator causing
an upset.

imo-spitzer-08.indd 76

9/15/09 11:03 AM

PID Control

FIG. 8-5.

77

Response to step process changePID control

PID Control
PID control infers that the controller algorithm consists of
proportional, integral, and derivative control. Before the step
change on the left side of Figure 8-5, the process variable and
set point are shown to be equal in a PID controller. As such,
the output remains at its previous value because no control
action is necessary.
When the process variable changes, an error (e) develops,
and the output is modified by the proportional, integral, and
derivative control algorithms. As shown in Figure 8-5, when
the step change occurs, the proportional control causes the
output to increase in proportion to the error, while the derivative control causes a spike that is proportional to the rate of
change of the error to occur. The integral control algorithm
exhibits no immediate response, but causes the output to ramp
by an amount that is proportional to the error over time. These
actions can be represented by the summation of the proportional, integral, and derivative control algorithms:

OUT(t + 1) = OUT(t) + Kp * e + Ki * (e * t) + Kd * e / t
(8-5)
The above equation illustrates that the controller output
response to an error is dependent upon the sign and magnitude
of the error, and the tuning constants Kp, Ki, and Kd. Controller

imo-spitzer-08.indd 77

9/15/09 11:03 AM

78

PID Control

tuning consists of adjusting these three variables in such a manner that the process variable stays at its set point under process
operating conditions. The PID velocity (incremental) algorithm
can be expressed as:
PID Controller Implementation The above discussion describing

the PID controller illustrates the operation of a parallel PID controller in which each control algorithm is performed independently and subsequently summed. Therefore, there is no interaction between the three tuning parameters; that is, changing the
value of one parameter has no effect on the other two.
PID controller implementations utilize control algorithms that
create interaction between the tuning parameters to simplify
loop tuning. In addition, the algorithms may be performed
either in series or in parallel due to electronic or computational
limitations. The user should be advised that this can cause different controller output responses from two nonidentical controllers with identical tuning parameter values. In addition,
identical controllers can exhibit the same symptoms when they
are constructed with components that are subject to drift, such
as those components used in mechanical, pneumatic, and analog electronic controllers.

Summary
The PID control algorithm has been the workhorse of industrial
process control and its importance to worldwide industrial
growth during the last half-century should not be understated.
Understanding the operation of this important algorithm and its
nuances can be the key to improving control of the process.

For Further Information

Astrom, Karl J., Hagglund, Tore. Advanced PID Control. Research Triangle
Park, NC: ISA.
Murrill, Paul W. Fundamentals of Process Control Theory. Research Triangle
Park, NC: ISA

imo-spitzer-08.indd 78

9/15/09 11:03 AM

 Chapter 9

Controller Tuning
Control loop tuning is the process of verifying the appropriateness, location, installation, and adjustment of every device in the
control loop, including the equipment, field instrumentation, signal transmitter, controller, and final control element.
Controller tuning, the adjustment of the tuning parameters associated with a controller, is but one facet of control loop tuning.
Before controller tuning is attempted, the equipment should be
adjusted and operational, the controller input should accurately
represent the process variable, and the final control element
should be functioning properly from the controller while in
manual mode. Failure to perform checks to verify the functioning of these items prior to tuning the controller can result in
poor control and can potentially result in equipment damage
and personal injury.
A number of methods can be used to make controller tuning
parameter adjustments. A few of these techniques are discussed
below to convey different perspectives on the subject. Regardless of the method employed, it is important that the person
adjusting the controller carefully observe the control loop
response for a sufficient time before making controller tuning
constant changes.

Controller Tuning Strategies

Before controller tuning constants are set, it is important to
determine the desired control strategy for the control loop. Control loop performance can be measured using a number of techniques, such as the percentage overshoot, time to return to set
point, integrating the error, and integrating the absolute value of
the error. The concepts behind these criteria are valid, but they
are primarily of theoretical value in actual plant operations.
The most important determination is whether the control
loop should be tuned for set point changes, for process

79

imo-spitzer-09.indd 79

9/15/09 11:05 AM

80

Controller Tuning

disturbances, or for reduced overshoot. The desire for stability

(tuning for process disturbances) and operator adjustment (tuning for set point changes) usually precludes tuning for only
one of the above strategies. As such, tuning is often performed
for one strategy, after which the acceptability of the tuning
with regard to the other strategy is verified with compromises
made as necessary. It should be noted that tuning for an inappropriate response can reduce the effectiveness of the control
loop in the process.
Tuning for Set Point Changes Controllers that are tuned for
set point changes react quickly to changes to the controller
set point. Control loop response can be checked by observing
the process variable response to a controller set point change
that is made after the process has been stable for a sufficient
period of time that the only upset condition is an intentional
set point change. The process variable may overshoot the set
point and exhibit damped oscillations before reaching the set
point (see Figure 9-1). Control loops that benefit from this
tuning strategy are usually not subject to significant disturbances from the process.

For example, a flow loop may benefit from tuning for set
point changes when the operator periodically adjusts the
flow based on current process conditions. It should be understood that the control loop should still maintain stability and
handle minor disturbances from the process. Most flow loops,
some temperature loops, and some pressure loops are tuned
for set point changes.

FIG. 9-1.

imo-spitzer-09.indd 80

Control loop response (tuned for set point response)

9/15/09 11:05 AM

Controller Tuning

81

Tuning for Process Disturbances Tuning a controller for disturbances from process upsets involves the examination of control
loop operation when process variable upsets occur and no set
point changes are made. Control loops that benefit from this tuning strategy typically do not require set point changes during
normal operation, but should more often respond to significant
process disturbances. Control loop response can be checked by
observing the process variable during the time when process
upsets are known to occur (see Figure 9-2).

For example, a boiler drum level controller rarely requires set

point changes but rather responds to process load changes.
Controller response could be checked when a large steam user
is turned on (or off) or, perhaps, during starting up. Most level,
many temperature, and some pressure loops exhibit similar attributes and are tuned for process disturbances.
Tuning for Reduced Overshoot Tuning a control loop for reduced

overshoot is desirable when a controller output change can cause

more significant process upsets than the process variable change that
initiated the controller output.
For example, level control of a surge tank is important in order to
maintain material in the tank while not overflowing the tank (see
Figure 9-3). However, a controller tuned for precise level control
may upset other parts of the process by causing abrupt changes in
flow. Some loops of all types are tuned for reduced overshoot.

FIG. 9-2.

imo-spitzer-09.indd 81

Control loop response (tuned for process disturbances)

9/15/09 11:05 AM

82

Controller Tuning

FIG. 9-3. Control loop response (tuned for reduced overshoot to setpoint change)

Quarter Amplitude Oscillation Method

The quarter amplitude oscillation method is based upon a
paper, written in 1942 by J. G. Ziegler and N. B. Nichols,
describing a technique by which the tuning constants can be
determined. Utilizing this method, the process variable will
overshoot and oscillate around the set point with each subsequent oscillation being approximately 25 percent of the amplitude of the previous oscillation.
This closed-loop technique requires that the controller gain be
increased in automatic until the loop steadily oscillates about
its set point while the integral and derivative algorithms are
disabled. The period of oscillation in minutes (Tu), while operating in this manner, and the ultimate gain (Ku) are used to calculate the proportional, integral, and derivative controller tuning
parameters using the following relationships:

Kp (gain)
Ti (minutes)
Td (minutes)

PI

PID

0.5 Ku

0.45 Ku
Tu/1.2

0.6 Ku
Tu/2
Tu/8

The advantage of this technique is that once the period and

ultimate gain are known, the controller tuning parameters can

imo-spitzer-09.indd 82

9/15/09 11:05 AM

Controller Tuning

83

be readily calculated. The disadvantage is that the control

loop is allowed to oscillate and may be driven to instability in
order to determine these parameters, potentially compromising
process safety and economic constraints. In addition, this technique can be time consuming and frustrating when the period
is long or when variations in the operation of the process
impair the ability to obtain accurate data.
Appropriate parameters can also be determined using the
open-loop method by introducing an abrupt output change and
observing the response while the control loop is in manual.
The dead time (the time during which the measurement does
not react after manipulating the final control element) and
the time constant (the time required for the measurement to
change 63.2 percent of its response to a step manipulation
of the final control element) can be used to calculate tuning
parameters for quarter amplitude decay using the following
open-loop relationships.
P

Kp (gain)
Ti (minutes)
Td (minutes)

1TC/(K*DT)

PI

PID

0.9TC/(K*DT)
3.33DT

1.2TC/(K*DT)
2.0DT
0.5DT

where TC is the process time constant (minutes), DT is the process dead time (minutes), and K is the open-loop process gain
(the ratio of the change in process variable to the controller output change that caused the change).
Alternate Method This method, described by G. K. McMillan,

is similar to the Ziegler-Nichols open-loop method but does not

drive the control loop into oscillation. The integral and derivative algorithms are disabled, and the gain is increased until each
successive oscillation peak in the response to a step set point
change is 25 percent of the previous peak. The gain (Kg) and
period (Tg)(minutes) of the closed-loop oscillations can be determined and used to calculate the tuning parameters:

p
*

K
Ki
Kd

P
Kg/1.2

PI

PID

Kg/2.4
Tg

Kg/1.2
Tg/2**
Tg/10**

* Multiply by 0.8 for slow loops

** Double for loops with dead band (from valve hysteresis)

imo-spitzer-09.indd 83

9/15/09 11:05 AM

84

Controller Tuning

Trial and Error Method

A survey of field practice would show this method to be much
more commonly used than any of the above methods. There is
no formula or formal methodology per se, because each individual has a personal proprietary method. However, most of
these methods include increasing the gain with proportional
control only until the loop is unstable, then increasing the
integral. Subsequent responses are observed with adjustments
made by feel.
As noted above, the controller may be tuned for process disturbances, set point changes, or reduced overshoot. When using the
trial and error method, it is important to challenge the control
loop with appropriate upset conditions and tune the controller to
produce the desired response.

Self-Tuning Method
Some controllers contain algorithms that enable the controller to calculate controller tuning constants based upon natural or, more commonly, induced process perturbations. Some
algorithms continually update the controller tuning constants
so that the current controller tuning reflects process response
changes. Process response changes may be caused by such
influences as mechanical sluggishness, mechanical wear, aging
of parts, and the like.
Other algorithms calculate suggested controller tuning
constants that the user can enter into the controller. The
approach of maintaining fixed controller tuning constants is
based on the premise that when the process dynamics remain
fixed, fixed controller tuning constants are satisfactory. This
approach also reduces the risk of the generation and insertion
of inappropriate controller tuning constants caused by process upsets and interaction.

Typical Controller Settings

Due to the uniqueness of each process, there are no universal
tuning constants for all controllers. However, G. K. McMillan suggests the following typical controller settings:

imo-spitzer-09.indd 84

9/15/09 11:05 AM

Controller Tuning

Flow
Liquid Pressure
Gas Pressure
Level
Temperature
Chromatograph

Proportional
(gain)
(% PB)
0.2-1.0
100-500
0.2-1.0
100-500
2-100
1.0-50
2-100
1.0-50
1.0-10
10-100
0.125-0.50 200-800

85

Integral
Derivative
(repeats/minute) (minutes)
10-50
none
10-50
none
0.02-1.0
0.02-0.10
0.01-0.10
0.01-0.05
0.02-0.05
0.5-2.0
0.01-0.10
none

Summary
Controller tuning compensates for the dynamics of the control
loop, enabling a process variable to be stabilized in the face of
potentially destabilizing factors. To improve control of the process, elimination of potentially destabilizing factors that may
be part of the control loop (such as an excessively damped measurement, an inaccurate instrument, a sluggish control valve,
and the like) should be performed prior to controller tuning.
While there are guidelines for tuning controllers, controller tuning is not an exact science and, as such, may require many trial
and error adjustments before the control loop functions satisfactorily, especially when controllers are applied to control loops
that have long time constants and long response times.

For Further Information

Corripio, Armando B., Tuning of Industrial Control Systems. Research Triangle Park, NC: ISA.
McMillan, Gregory K., Good Tuning: A Pocket Guide. Research Triangle
Park, NC: ISA.

imo-spitzer-09.indd 85

9/15/09 11:05 AM

imo-spitzer-09.indd 86

9/15/09 11:05 AM

 Chapter 10

Regulatory Control Loop

Pairing
Implementing each individual facet of instrumentation control
loop design correctly is critical to the proper operation of the control loop. The field measurement device should accurately measure the process variable and the final control element (typically
a control valve) should be capable of adequately manipulating the
process. The controller should contain the appropriate control
algorithms and features to control the process at hand. All devices
in the loop should be properly interconnected and calibrated.
However, implementing each individual facet of instrumentation
control loop design correctly is no guarantee that the control loop
will function properly in the actual process, even if the control
loop is operating as designed.
Previous chapters focused on the function of each individual
part of a control loop and the importance of implementing each
function correctly. This chapter focuses on the need to select the
proper process variable to control and the proper final control
element to manipulate.

Regulatory Control Strategy

A regulatory control loop is defined herein as a control loop that
contains one measurement device, one controller, and one final
control element. Flow, level, pressure, and temperature measurements comprise the majority of industrial measurements. PID
controllers dominate the industrial controllers, while the majority
of final control elements used in industry are control valves.
Proper operation of the control loop infers that each part of the
control loop function properly, both individually and in conjunction with the control loop, and that the measurement instrument
properly represents the process variable.

87

imo-spitzer-10.indd 87

9/15/09 11:11 AM

88

Regulatory Control Loop Pairing

In addition, the control loop should implement the an appropriate control strategy for the process at hand. For regulatory control
loops, the proper process variable should be selected to feed the
controller that manipulates the appropriate final control element.
This selection process is called pairing.
In practice, determining the regulatory control loop strategy
and pairing the process variables with final control elements
is typically based on experience. It is essential that the person
determining the control strategy have intimate knowledge of the
process. It should be remembered that THE GREAT MAJORITY
OF FINAL CONTROL ELEMENTS MANIPULATE FLOW, even
though it may be desirable to control a process variable other
than flow. Further, the manipulated final control element should
be capable of adequately affecting the process variable that is to
be controlled.
Also, manipulation of what appears to be the appropriate final
control element to control the desired variable can have an undesirable effect on another part of the process. When this occurs,
more sophisticated techniques may be required to properly control the process.
Flow Control One of the most straightforward applications of regulatory control is that of controlling flow, where flow is the only
process variable of concern. In the application shown in Figure
10-1A, the flow leaving a tank is controlled by measuring the flow
and manipulating the valve in line with the flowmeter, or perhaps

FIG. 10-1(A). Regulatory control of flow. Regulatory control loop with control valve.

imo-spitzer-10.indd 88

9/15/09 11:11 AM

Regulatory Control Loop Pairing

89

FIG. 10-1(B). Regulatory control of flow. Regulatory control loop with

variable-speed drive.

as shown in Figure 10-1B by adjusting a variable-speed drive

rotating the pump in series with the flowmeter. In the case of the
control valve, if the flow increases (decreases), the flow controller
throttles the valve more closed (open) to bring the flow down (up)
to the set point.
Level Control If it were desirable to control the level in a tank
because the tank could overflow or its level could become too low
to maintain pump suction, a level controller could manipulate a
makeup control valve or a discharge control valve.

Selecting the proper control valve to manipulate depends on

the process and the effects of manipulating the upstream and/
or downstream operations (see Figure 10-2A). Note that when
manipulating the makeup flow, an increase (decrease) in tank
level causes the level controller to throttle the makeup valve
more closed (open) to bring the level down (up) to its setpoint.
Similarly, the level controller in Figure 10-2B that manipulates
the discharge flow will throttle the discharge valve open (closed)
to bring the level down (up) to its set point by manipulating the
flow out of the tank. If the tank level were paired with the wrong
valve, a process upset could occur elsewhere in the process, even
though the level control loop is functioning properly.
In many installations, flowmeters are installed to monitor flows
that may be important as illustrated in Figures 10-2A and 10-2B.

imo-spitzer-10.indd 89

9/15/09 11:11 AM

90

imo-spitzer-10.indd 90

Regulatory Control Loop Pairing

FIG. 10-2(A).

Regulatory control of level. Regulatory level control loop controlling inlet flow.

FIG. 10-2(B).
charge flow.

Regulatory control of level. Regulatory level control loop controlling dis-

9/15/09 11:11 AM

Regulatory Control Loop Pairing

91

Pressure Control If it were desirable to control the pressure in a

vessel, a pressure controller could manipulate the inlet valve or
the discharge control valve.

FIG. 10-3(A). Regulatory control of pressure. Regulatory gas pressure control loop controlling inlet flow.

FIG. 10-3(B). Regulatory control of pressure. Regulatory gas pressure control loop controlling discharge flow.

imo-spitzer-10.indd 91

9/15/09 11:11 AM

92

Regulatory Control Loop Pairing

Selecting the proper valve for manipulation depends upon the

process and the effects of manipulation on the upstream and/
or downstream operations. Note that an increase (decrease)
in tank pressure will cause the pressure controller to throttle
the inlet valve in Figure 10-3A more closed (open) to bring the
pressure down (up) to its set point by manipulating the flow
to the tank. Similarly, the pressure controller that controls the
discharge valve will throttle the discharge valve in Figure 10-3B
open (closed) to bring the pressure down (up) to its set point by
manipulating the flow out of the vessel. If the pressure process
variable were paired with a control valve that did not appreciably affect the pressure, this pressure control loop would not
function properly.
Flowmeters are usually not installed to monitor flows in pressure
control applications unless the flows are unusually important.
Temperature Control

If it were desirable to control the temperature of the discharge

flow of a heat exchanger, a temperature controller could manipulate the heating valve or the cooling control valve.
Note that an increase (decrease) in the discharge temperature
causes the heating temperature controller in Figure 10-4A to
throttle the steam valve more closed (open) to bring the temperature down (up) to its set point by manipulating the steam flow
to the heat exchanger. Similarly, the temperature controller that
controls the cooling water valve in Figure 10-4B throttles the discharge valve open (closed) to bring the temperature down (up) to
its set point by manipulating the cooling water flow.
Flowmeters are usually not installed to monitor utility flows
unless the flows are large enough to be significant or otherwise
important.

Summary
Regulatory control (as defined above) is a collection of individual
control loops that are limited to one process variable and one
manipulated variable. Most regulatory control loops manipulate
flow to control the desired process variable.
Should the control strategy and pairing be inappropriate, it may
not be possible to bring the process under control. Further, seemingly appropriate strategies and pairings may upset the process.

imo-spitzer-10.indd 92

9/15/09 11:11 AM

Regulatory Control Loop Pairing

93

FIG. 10-4(A). Regulatory control of temperature. Regulatory temperature control loop (heating with steam).

FIG. 10-4(B). Regulatory control of temperature. Regulatory temperature

control loop (cooling water).

imo-spitzer-10.indd 93

9/15/09 11:11 AM

imo-spitzer-10.indd 94

9/15/09 11:11 AM

 Chapter 11

The Limitations of
Regulatory Control
Regulatory control has been available for more than 50 years, and
its single input-single output nature allows it to be easily understood and applied with apparent success. This success is usually
evaluated using such criteria as: (1) the flow was not excessive;
(2) the tank did not overflow; (3) the pressure relief valve did not
operate; and (4) the product did not boil or freeze.
While these expectations may have been reasonable decades ago,
they are unrealistically low given the control tools available. It
should be noted that the relative simplicity of regulatory control
and its criteria for success have created a large cadre of people
with low expectations and the belief that regulatory control is all
that will ever be required.
Having presented and described regulatory control in the pristine
environment of steady-state operation that represents reality to
so many, the problems associated with some regulatory control
loops under dynamic conditions can be explored.

Control Loop Response

Control loops can be roughly divided into loops that respond
relatively quickly (fast loops) and those loops that respond sluggishly (slow loops).
Most flow control loops are fast loops: when the control valve
(or variable speed drive) is manipulated, the flowmeter usually
senses the change almost immediately. Some gas flow control
loops can have damped responses due to long pipe lengths, fluctuating gas pressure, large pipe size, and flowmeter technology.
Despite this sluggishness, even these flow loops are usually the
fastest loops in the process and can be considered to be fast. See
Figure 11-1.

95

imo-spitzer-11.indd 95

9/15/09 11:14 AM

96

The Limitations of Regulatory Control

FIG. 11-1. Flow control loop with fast response damped by pipe length and flowmeter
technology

Most pressure control loops are also fast loops, especially where
the controlled volume and flow are small compared to the control
valve capacity. In many applications, pressure loops are slower
than the flow loops, but react quickly enough to be considered
fast. See Figure 11-2. However, some pressure loops can be considerably slower, such as when the control valve capacity is small
compared to the volume and/or process flow.
Level control loops are typically slow loops because deviation
from set point usually causes a small change in flow relative to

FIG. 11-2.

imo-spitzer-11.indd 96

Gas pressure control loop

9/15/09 11:14 AM

The Limitations of Regulatory Control

97

tank volume. It is not uncommon for properly tuned level loops

to require tens of minutes to stabilize. The exception is the case
in which the control valve capacity is large compared to the volume of the tank, a situation can be indicative of an undersized
tank, an oversized control valve, an unnecessary tank, or an inappropriate control strategy.
Temperature control loops are usually slow loops, especially
when a heat exchanger is involved. See Figure 11-3. It is not
uncommon for heat exchangers to exhibit process dead times of
tens of minutes (that is, the time necessary for the process discharge temperature measurement to start sensing the effect of
a heating or cooling medium change). In addition, the process
time constant (that is, the time necessary for the process variable to change 63.2 percent of its change after the process dead
time) can be in minutes for heat exchangers of significant size
relative to process flow.
Other processes, such as furnace temperature control, can have
shorter dead times and time constants and generally exhibit
a significantly faster response than described above for heat
exchangers. Despite this improvement, temperature loops are

FIG. 11-3.

imo-spitzer-11.indd 97

Regulatory temperature control loop (heating with steam)

9/15/09 11:14 AM

98

The Limitations of Regulatory Control

usually sufficiently slow compared to other control loops in the

process that they can be considered (relatively) slow loops.
Fast and Slow Loop Tuning

Barring uncommon idiosyncracies, fast control loops can typically be tuned satisfactorily using the proportional and integral
control algorithms, because the process variable responds to
final control element changes almost immediately. As a result,
the effect of controller tuning adjustments can be observed in
seconds or minutes, and necessary additional controller tuning
adjustments can be quickly analyzed.
Slow control loops can represent a formidable tuning challenge. Consider regulatory temperature control of the process
output of a heat exchanger using steam as the heating medium
(see Figure 11-3).
The apparent control strategy is to control the process fluid
temperature at a desired temperature. As such, the heat
exchanger process outlet temperature is measured and controlled by manipulating the steam valve. Using this pairing,
it would seem reasonable that, if the temperature increased
(decreased), the controller would sense the deviation and
throttle the steam valve more closed (open) to maintain the
temperature at the desired set point. This reasoning implicitly
applies to the steady state, that is, the steam valve will be more
closed (open) when the temperature is returned to set point.
However, this approach completely ignores the actual control
loop dynamics.
Assuming that the temperature is stabilized at its set point,
changing the set point will move the steam valve, but will not
affect the measurement during its dead time (typically minutes) to allow time for heat transfer and time for the heated
fluid to reach the location of the temperature measurement (see
Figure 11-4). During the dead time, the lack of process variable
response maintains the deviation between the set point and
process variable, which causes the integral controller algorithm
to continue to move the steam valve in what appears to be the
proper direction. When the heated fluid reaches the measurement location, the measurement changes relatively rapidly,
causing the steam valve to relatively abruptly move in the
opposite direction.
Because a properly tuned temperature controller may take
tens of minutes or even hours to stabilize, it is not uncommon
to make tuning changes twice a dayonce when arriving at

imo-spitzer-11.indd 98

9/15/09 11:14 AM

The Limitations of Regulatory Control

FIG. 11-4.

99

Regulatory temperature control loop response

work and again prior to leaving work. To complicate the issue,

in addition to the proportional and integral algorithms used
for fast loops, the derivative algorithm should be used in many
applications. In order to maintain stability, these loops are usually slow and sluggish to react to either set point or process disturbances. The difficulty and frustration of tuning these loops
and the poor dynamic loop performance are often not recognized by anyone except the individual performing the tuning.

Process Upsets
Virtually all regulatory control loops manipulate a process flow.
Set point changes and process disturbances to slow loops may
cause flow changes that have a small effect on the process variable but a large effect on the operating process.
The bottoms level of distillation column A (see Figure 11-5) is
controlled by manipulating the bottoms control valve that throttles the flow of liquid leaving the column to directly feed distillation column B. In this application, the set point is rarely changed,
so the control loop should be tuned for process disturbances that
can (and will) occur. Because small changes in the control valve
position have a negligible effect on the level of column A, the
level control loop is often tuned in such a manner that relatively
small level disturbances will cause significant movement of the
control valve, in turn causing relatively large flow fluctuations,
even in so-called steady state. Even though the level loop for column A is functioning properly, these large fluctuations can have a
significant effect on the operation of distillation column B.

imo-spitzer-11.indd 99

9/15/09 11:14 AM

100

The Limitations of Regulatory Control

FIG. 11-5.

Regulatory level control loop (distillation column bottoms)

Control Loop Interaction

Few processes are a collection of totally independent entities that
can be satisfactorily controlled by a number of regulatory control
loops that contain one measurement, one controller, and one final
control element. As illustrated in the level control loop above,
the operation of one control loop (level) may have a pronounced
effect another (flow).
In the control loop shown in Figure 11-6, the cooling water
supply and return headers are undersized so the pressure difference between the supply and return headers will vary with
the cooling water flow rate. Increasing the cooling water flow
in Process A will cause the cooling water flow through the Process B heat exchanger to be reduced. The effects of this reduction will not be measured until some time later, after which
the Process B heat exchanger cooling water will be increased,
in turn reducing the cooling water to Process A, and so on. In
this application, what appear to be unrelated regulatory control
loops on different processes can interact with one another to
upset the process.

imo-spitzer-11.indd 100

9/15/09 11:14 AM

The Limitations of Regulatory Control

FIG. 11-6.

101

Regulatory temperature control loops with interaction

Summary
It is not uncommon to be exposed to lengthy analyses about
how each regulatory control loop stabilizes its respective process variable while implying (but not stating) that the process
is at steady state. Real processes exhibit dynamic considerations that are absent from such analyses. How well these

imo-spitzer-11.indd 101

9/15/09 11:14 AM

102

The Limitations of Regulatory Control

considerations are compensated for by the control system often

determines the quality of control and, in some applications,
the ability to control. Of significant concern is when slow
loops measure process variables that react slowly but manipulate process variables that react quickly.
The ability of regulatory control loops to compensate for process
dynamics, process upsets, and control loop interaction is limited
by the single process variable and single manipulated variable
nature of the regulatory control loop. It is estimated that over 50
percent of existing regulatory control loops could benefit from
some form of more sophisticated control.

imo-spitzer-11.indd 102

9/15/09 11:14 AM

 Chapter 12

Advanced Regulatory
Control Tools
Advanced regulatory control tools include mathematical functions that are used in conjunction with feedback, feedforward,
cascade control, feedback trim, dynamic compensation, and/or
feedforward action to implement advanced regulatory control
strategies. The intent of these more sophisticated control strategies is to manipulate the final control element to obtain superior
control more intelligently than would otherwise be possible
using regulatory control.
Some of these concepts are not new and have been around for
decades, but the availability of digital controllers, programmable logic controllers, distributed control systems, and fieldbus control systems have reduced their cost and simplified
their implementation. However, other tools have become more
sophisticated, for example, the controller set point may track
the process variable and freeze its output to avoid upsets when
a bad measurement is detected.

Mathematical Functions
Mathematical functions represent a set of powerful tools that
can be used to significantly improve control of the process when
applying advanced regulatory control strategies. Some of the typical tools that are available in digital controllers, programmable
logic controllers, distributed control systems, and fieldbus control
systems are reviewed below.
Calculation Calculation capabilities may be limited to preset
formulas, but it usually is possible to write algebraic equations
that will perform calculations using real-time measurements.
Available functions typically include addition, subtraction, multiplication, division, ratios, exponentials, logarithms, trigonometric functions, and the like.

103

imo-spitzer-12.indd 103

9/15/09 11:38 AM

104

Advanced Regulatory Control Tools

Characterization A characterizer defines a fixed relationship

between its input signal and its output signal that is usually
defined in tables of corresponding input and output data from
which the characterizer extrapolates the output signal corresponding to a given input. Linearization of an input signal
from a look-up table is a typical application of the characterization function.
Dead Time Algorithm The dead time algorithm is a dynamic

compensator that is used to delay a signal by a specified period

of time and is primarily used where the effects of the signal on
the process would be premature and, as such, should be delayed
(see Figure 12-1).
Dead time is often applied in the feedforward signal path in conjunction with a lead/lag algorithm and in processes that contain
transportation time delays.
Lead/Lag Algorithm The lead/lag algorithm is a dynamic compensator that is almost exclusively used in the feedforward
signal path. The lead algorithm is used to immediately affect
the process as a result of a feedforward variable change. The lag
algorithm is used to dampen the effect of a signal on the process. (See Figure 12-2).

A combination of lead and lag can be applied to a signal to obtain

an immediate response followed by a damped response over
time. In some applications, failure to use a lead/lag algorithm to
damp the response can trigger process upsets, as can maladjustment of the lead/lag algorithm.

FIG. 12-1.

imo-spitzer-12.indd 104

Dead time.

9/15/09 11:38 AM

Advanced Regulatory Control Tools

FIG. 12-2.

105

Lead/lag algorithm

Logic Functions Control systems can utilize field measurements

to determine when a given process condition exists and take
action accordingly. For example, when the output of one controller is at its maximum, another controller could be activated
to control another process variable in automatic operation with
a preset set point. The ability of some systems to perform logic
functions is virtually limitless.
Ramp Function The ramp function can be used to increase
(decrease) a signal at a constant rate between its starting and ending points. This function may be desirable when used to reduce
abrupt changes that may upset the process, such as slowly opening a control valve to prevent damage from a large flow increase
resulting from faster operation of the control valve.
Signal Limiter Signal limiters are used to clamp a signal between
(high and/or low) limits.
Signal Selection The function of a signal selector is to select one

input (as its output) based upon its signal selection criteria, such
as high signal, low signal, average signal, and the like. This function is often used in combustion controls to select the hightest
(lowest) of the desired and required air (gas) flows to ensure sufficient air to combust the fuel.

Advanced Regulatory Control Process Variables

Mathematical tools available within the realm of advanced
regulatory control can be used to calculate process variables

imo-spitzer-12.indd 105

9/15/09 11:38 AM

106

Advanced Regulatory Control Tools

that are more accurate than the raw measurement and more
representative of the process. Based upon a number of process
measurements, it is possible to calculate process variables that
have physical significance but can not be measured directly
and may not physically exist. In addition, mathematical functions can be used to correct for processes that are mathematically understood.
Determining which process variables are of importance and
should be controlled in a given application requires an intimate knowledge of the process, an understanding of the appropriate measurement techniques, and experience. It is important
to get as close to the process as possible, measure what is to
be controlled, understand what is being measured, and implement those measurements with the understanding that the
measurement does not stop at the transmitter. Both assessing
the reliability of the selected instruments and the probability
of success are based on experience.
Flow Measurement Flow measurement presents a number of

challenges that are often not addressed, such as the type of

flow measurement required for a specific application. Flowmeters are available to measure mass, volume, velocity, and
dynamic head, not all of which may be suitable for the process
at hand.
If it is desired to measure and control the mass flow of a liquid with varying specific gravity, it would seem logical that
a mass flowmeter should be used. However, it is not uncommon to find a flowmeter that measures dynamic head scaled
to mass units (which it does not measure). Moreover, this
measurement is misleading and may not accurately represent
the process.
When a flowmeter is not available to measure the desired
type of flow, it may still be possible to improve the measurement by mathematically compensating for fluid properties.
The equations for the installed flowmeter should be obtained
and examined to determine how the fluid properties affect
the flow measurement. An equation can often be implemented to correct the raw flow measurement for the fluid
properties that are measurable (or available) to obtain a more
accurate flow measurement.
Pressure and Temperature Compensation The effects of pressure and temperature are far more pronounced on gas flow
measurements than on liquid flow measurements.

imo-spitzer-12.indd 106

9/15/09 11:38 AM

Advanced Regulatory Control Tools

107

Nonetheless, the number of gas flow control loops that do not

compensate for pressure and temperature is astounding. For
example, at ambient temperatures, a 3-degree Celsius gas temperature change affects the volumetric flow by approximately
1 percent, while a pressure change of 0.3 bar (4.5 psi) at 2.0
bar (29 0 psi) will affect volumetric flow by approximately 10
percent. If the raw flow measurement is not corrected for these
and other process variations, the instrumentation will not be
controlling the proper parameter, despite the apparent control
of the (incorrect) raw flow measurement signal. This scenario,
and worse, is not uncommon in actual processes where maintaining a constant known flow is extremely important to the
safe and economical operation of the process.
In some applications, the Ideal Gas Law can be used to correct the
raw flow signal by multiplying the raw volumetric flow signal by

(P Tref) / (Pref T)
(P Tref) / (Pref T)

(12-1) (velocity flowmeters)

(12-2) (head flowmeters)

where P and T are the absolute pressure and temperature, respectively, and the subscript ref denotes the reference conditions to
which the flowmeter was calibrated. When available, more definitive process and fluid information should be used.
Level Measurement Many level measurement devices measure

either actual level or the head of liquid at a known tank location.

Note that the measurement of the actual level does not measure
the mass of the liquid in the tank. On the other hand, head measurements that imply liquid level can be significantly affected by
variations in liquid composition and temperature that cause the
specific gravity of the liquid to change. With sufficient process
information, an equation can be written to compensate for these
specific gravity changes. However, putting another liquid with a
different specific gravity in the vessel can cause significant measurement error when the transmitter is not recalibrated or when
the calculation is not altered.
Fortunately, most industrial level control loops need not control
the level precisely, although the operation of some processes may
be improved by doing so.
Pressure Measurement Process pressures that can be measured

include absolute pressure, gauge pressure, and differential

pressure. The application of each type of transmitter may

imo-spitzer-12.indd 107

9/15/09 11:38 AM

108

Advanced Regulatory Control Tools

seem obvious, but there are applications in which this is not

the case.
For example, it would seem that the differential pressure
across an atmospheric distillation column would be measured
with a differential pressure transmitter, but condensation can
occur in the impulse lines that can affect the measurement.
Using capillary tubing with diaphragm seals would resolve
this problem, but the taps may be so far apart that the maximum capillary length can be exceeded. Two gauge pressure
transmitters could be applied and their signals subtracted
to obtain the differential pressure, while making the pressure information available. However, for chemical reasons,
the absolute column pressure may be of interest, and because
atmospheric pressure changes will not be measured by the
gauge pressure transmitter, two absolute pressure transmitters
should be applied. In addition, these transmitters should be
sufficiently accurate to correctly measure the differential pressure with the desired accuracy.
Temperature Measurement Temperature measurements can be

measured using a number of devices, such as thermocouples,

resistance temperature detector (RTDs), and infrared pyrometers. It is important that the temperature-sensing technique
used be appropriate for the application and properly located.
In addition, appropriate signal conditioning should be performed to achieve accurate temperature measurement in the
range of interest.
Composition Measurement Many techniques can be used to measure composition, but, regardless of the technique, the analyzer
should sample the process at the correct location. In addition,
when the analyzer is not in situ, it should be sufficiently close to
the process to avoid exhibiting excessive transport delay, which
adds to the analysis time.
Heat Flow Many processes would benefit from the control of heat
flow, a process variable that typically cannot be measured but can
be calculated using appropriate process measurements.

In a multi-fuel furnace in which the waste fuel flow rates vary,

it is common to control furnace temperature by adjusting the
purchased fuel with a temperature controller. Superior control
could be obtained by stabilizing the total amount of fuel that
enters the furnace. The total fuel entering the furnace can be calculated by summing each fuel flow multiplied by its respective
heating value, which can be measured or estimated:

imo-spitzer-12.indd 108

9/15/09 11:38 AM

Advanced Regulatory Control Tools

Total Fuel Content = (Q1 * HV1) + (Q2 * HV2) + ...

109

(12-3)

where Qn is the nth fuel flow and HVn is the heating value of the
nth fuel in like units.
In a flow stream, such as would flow through a heat exchanger,
the heat gained or lost can be calculated by the following:

Heat Gain (Loss) = m * cp * t

(12-4)

where m is the fluid mass flow, cp is the specific heat of the fluid,
and t is the temperature difference between the inlet and outlet
streams. Be aware that process dynamics, especially prevalent in
equipment exhibiting long dead times and time constants, such
as heat exchangers, can adversely affect the accuracy of this calculation during transients.

Feedback Control
A feedback controller compares the process variable (measured in real time) with the set point and manipulates the
final control element based upon the controller tuning adjustments. In the absence of any other control strategy, this type
of feedback control is regulatory control. Regulatory control
and its limitations have been discussed at length and will not
be discussed here. It should be noted, however, that virtually all control loops utilize some form of feedback control in
their control strategies.
Feedback Control with Variable Gain In some applications, the
process variable changes nonlinearly with the manipulated variable. For example, when controlling pH by manipulating reagent
flow, the effect of increasing reagent flow by large amounts when
the pH is relatively low or high is a small pH change (see Figure
12-3). When the pH is near 7, small changes in reagent flow can
drastically affect the pH. Were a PID controller applied to this
process, the reagent flow would change based upon the error,
regardless of whether or not the pH was near 7. Therefore, tuning the controller near (far from) a pH of 7 would result in an
extremely slow (fast) pH response when the pH is far from (near)
a pH of 7.

To compensate for the changing process gain, some controllers

allow the characterization of the gain, so that the pH controller

imo-spitzer-12.indd 109

9/15/09 11:38 AM

110

Advanced Regulatory Control Tools

FIG. 12-3.

Relationship between reagent flow and pH

makes substantial reagent adjustments when the pH is far from 7

and small reagent adjustments when the pH is near 7.
Fuzzy Logic Control The controllers described heretofore have

precise algorithms that produce a precise output to a given set

of input conditions. The so-called fuzzy logic controller produces an output that is dependent upon the determination of
the relative value of the process variable, such as warm, hot, and
very hot. In some applications, these controllers can respond to
process upsets and process changes more quickly with less overshoot than can a PID controller. Fuzzy logic algorithms and their
respective adjustments are proprietary and often self-adjusting.

Cascade Control
A cascade control loop consists of two feedback control loops,
where each control loop has its own process variable measurement, but only one control loop manipulates the final control element. Significantly improved control can often be achieved using
cascade control, however two process measurements and two
controllers are required for implementation.
In a regulatory control loop, the regulatory controller manipulates the final control element directly (typically varying flow).

imo-spitzer-12.indd 110

9/15/09 11:38 AM

Advanced Regulatory Control Tools

111

In a corresponding cascade control loop, the output of the

master controller uses the corresponding regulatory controller
process variable as its process variable to set the set point of
a slave controller. Based upon this setpoint, the slave controller manipulates the corresponding regulatory controller output
using the measurement of the parameter that the final control
element varies (typically flow) as its process variable. The slave
control loop, typically a relatively fast flow or pressure control
loop, manipulates the final control element to whatever position is necessary to maintain its desired set point. As such, the
nonlinearities and hysteresis associated with the valve actuator,
process conditions, piping hydraulics, and the like are effectively linearized because the master controller output sets the
slave set point, not the output to final control element as in the
case of regulatory control.
Despite the existence of a relatively slow process variable and
the relatively fast manipulation of flow by the final control element, regulatory control implies that the manipulation of both
process variables be accomplished with one controller and,
hence, one set of tuning constants. Cascade control overcomes
this limitation through the use of two control loops, thereby
allowing the use of two sets of tuning constants. Using cascade
control, the relatively fast loop can be independently tuned for
improved control of the fast process variable. Similarly, after
the fast control loop is tuned, the slow control loop tuning can
be adjusted for improved control because its tuning adjustments do not have to deal with the dynamics of the fast control loop. In addition, each incremental output of the master
controller will produce a nearly identical change in the slave
controller process variable, making the master control loop
easier to tune than its regulatory counterpart that compensates
for the valve and process dynamics of the slave process variable. In other words, using individual controllers for the fast
and slow control loops separates the tuning of the fast and
slow loops into two easier tasks, noting that each respective
controller in the cascade control loop is a regulatory controller that is more directly manipulating the process variable it is
attempting to control.
Using a cascade control structure is especially beneficial where
changes in the slave controller process variable can upset the
process. Because the slave controller quickly returns its process
variable to its set point despite nonlinearities and disturbances,
the master controller output tends not produce much overshoot.
As such, the master controller can improve the control of its
process variable while producing smaller fluctuations in the
slave process variable as compared to regulatory control.

imo-spitzer-12.indd 111

9/15/09 11:38 AM

112

Advanced Regulatory Control Tools

It is important that the slave control loop be faster than the

master control loop. When the speed of the master and slave
controllers are similar, the slave control loop may not have sufficient time to reach its set point before the master controller
output changes the slave set point. If this is the case, the control
loops will interact and fight each other. Detuning the master
controller for a slower response can often alleviate this problem.
The cascade distillation column bottoms level control shown
in Figure 12-4 consists of a flow (slave) control loop, the
remote set point (RSP) of which is set by the output of the
level (master) control loop. The flow loop is a fast loop that
can be tuned for set point changes with minimal overshoot. It
is reasonable to expect that the flow would reach and maintain
its set point within a few seconds after a set point change. In
addition, even though the flow control loop is tuned for set
point changes, the flow control loop will react to process disturbances affecting flow much more rapidly than would a corresponding regulatory level control loop that would have to
wait until the effect of the flow disturbance was measured by
the level transmitter.

FIG. 12-4.

imo-spitzer-12.indd 112

Cascade level control loop (distillation column bottoms)

9/15/09 11:38 AM

Advanced Regulatory Control Tools

113

The level control loop is significantly slower than the flow

control loop and should be tuned for process disturbances
(after the flow control loop is tuned) because its set point is
not changed often by the operator. The cascade control strategy
reduces flow variations because the level controller adjusts
flow instead of valve position, as was the case of the corresponding regulatory control loop. In this manner, flow need
only be varied by a small amount to bring the level back to its
set point, as compared to its regulatory counterpart that moved
a valve and incremented flow by an unknown amount while
measuring its effect on the level.
The advantage of using cascade control is the ability to better
control the master controller process variable by separately and
more easily tuning the fast and slow control loops while, at the
same time reducing fluctuations of the fast process variable,
which in turn can reduce process upsets elsewhere in the process. It should be noted that cascade control can reduce or eliminate the need for the installation of a surge tank between distillation columnsa practice that is not all that uncommon when
regulatory control is applied.

Feedforward Control
Feedforward control is often used in conjunction with feedback
control to adjust a controller set point or controller output in
response to the change in a related process measurement before
the effect of the process change is measured by the controller process variable measurement instrument. The intent of feedforward
control is to achieve better control of the process variable by having the control system respond to measured disturbances prior
to the effect of the disturbance being measured at the controllers
process variable input, thereby reducing the impact of the disturbance on the process variable and, hence, reduce the effect of the
disturbance on the process.
Pure feedforward control produces an open-loop response to a
disturbance and implies prior quantitative process knowledge,
such as a definitive mathematical relationship between process
inputs and process outputs. In most applications, pure feedforward control can be difficult to implement for varying loads and
dynamics, so feedforward control is typically used in conjunction
with feedback control to trim the process.
Additive Feedforward Control In the application shown in Fig-

ure 12-5, additive feedforward control was applied because in

imo-spitzer-12.indd 113

9/15/09 11:38 AM

114

Advanced Regulatory Control Tools

FIG. 12-5. Boiler drum level control loop with feedforward, additive feedback, and cascade control

the steady state, an increment of the feedforward variable (steam

flow) results in a corresponding increment of the manipulated
variable (feedwater flow).
The boiler drum level control loop shown consists of a level
control loop cascaded to a flow control loop. Due to the fundamental feedback nature of this cascade control loop (that is, the
level controller will not react until the level changes), abrupt
variations in steam demand may cause the drum level to vary
significantly. Recognizing that the magnitude of steam flow
variations can be related to subsequent level changes, the steam
flow measurement can be multiplied by a constant and added

imo-spitzer-12.indd 114

9/15/09 11:38 AM

Advanced Regulatory Control Tools

115

to the level controller output (flow controller remote set point

input) in an attempt to correct for the impending level change
before the level change occurs.
Note that if cascade control had not been applied, the feedforward signal would adjust the control valve directly instead of
adjusting the flow set point. If this were the case, a given multiplier may work well at some steaming rates but not at other
steaming rates.
Multiplicative Feedforward Control Multiplicative feedforward
control can be applied when, in the steady state, an increment
of the feedforward variable does not increment the manipulated
variable by a certain amount but rather applies to the entire
manipulated variable stream. In a well-defined process, feedforward control could be applied in the absence of all other
controls in an attempt to manipulate the process based upon
process measurements.

An example of multiplicative feedforward control is the ratio

control of two flows, only one of which can be manipulated. In
the application shown in Figure 12-6, the remote set point of the
manipulated flow controller is set in a ratio to the measured wild

FIG. 12-6.

imo-spitzer-12.indd 115

Multiplicative feedforward ratio control loop

9/15/09 11:38 AM

116

Advanced Regulatory Control Tools

flow. The objective of this control strategy is to provide composition control of the mixed stream where the nature eof the process
is multiplicative (not additive).
Multiplicative Feedforward Control with Feedback Trim While

the use of pure feedforward control can be practiced, such as in

some combustion control systems, their nature is such that they
cannot compensate for unknown process disturbances, such as
composition changes, temperature fluctuations, pressure fluctuations, and the like. Therefore, feedforward control is commonly
applied in conjunction with multiplicative feedback trim control
to provide the relatively quick feedforward response to measured
disturbances in conjunction with a slower feedback response to
unmeasured disturbances.
The simplified boiler combustion control system shown in Figure 12-7 that does not show the cross-ties necessary to lead firing increases (decreases) with air (fuel), illustrates individual
combustion air and fuel flow measurements and controllers. The
boiler firing rate is changed by adjusting the fuel controller set
point. Disregarding the oxygen trim signal, combustion air flow is
controlled as a pure feedforward (multiplicative) ratio of the fuel
flow measurement.
The purpose of this process is to combust the raw materials (air
and fuel) to release heat that is recovered to produce steam. The
process produces two products: useful steam and a stack gas
(waste product). For increased efficiency, it is desirable to recover
as much heat as practical from the stack gas. One way to reduce
heat losses out the stack is to reduce the amount of excess combustion air that is fed to the burnera control strategy that is
not addressed in the pure feedforward control strategy described
above.
By installing a oxygen analyzer in the stack, the stack gas oxygen
content can be measured and fed to the stack oxygen controller.
The output of the stack oxygen controller could be multiplied
by the fuel measurement to generate the combustion air flow set
point. In practice, this output is usually used to trim the combustion air flow within prescribed amounts, typically by about 10
percent. In this manner, small unmeasured process disturbances
(such as the heating value of the fuel, the oxygen content of the
combustion air, and the like) can be compensated for using feedback trim. Note that this control strategy applies feedforward control of combustion air in conjunction with multiplicative feedback trim from the excess oxygen controller.

imo-spitzer-12.indd 116

9/15/09 11:38 AM

FIG. 12-7.

Simplified boiler combustion controls with multiplicative feedforward control and feedback trim

Advanced Regulatory Control Tools

imo-spitzer-12.indd 117

117

9/15/09 11:38 AM

118

Advanced Regulatory Control Tools

FIG. 12-8 Heat exchanger controls with multiplicative feedforward control, feedback trim,
and dynamic compensation

Feedforward Control with Dynamic Compensation The feedfor-

ward boiler combustion controls described above react relatively

quickly to measured disturbances. In some processes, such as a
heat exchanger temperature control, quick response to process
measurements can cause process upsets due to the sluggish process. See Figure 12-8.
Disregarding the lead/lag and assuming steady state operation,
the fluid flow measurement, fluid inlet temperature measurement, and the fluid outlet temperature set point can be used to
calculate the steady-state amount of steam required to raise the
fluid exit temperature to its set point. In the steady state, the
calculated steam flow can be used as the steam flow controller
set point. The temperature trim controller can be used to trim
the calculated steam flow set point (using feedback) to take

imo-spitzer-12.indd 118

9/15/09 11:38 AM

Advanced Regulatory Control Tools

119

heat losses, steam header pressure variations, and the like into
account.
It should be noted that, although measurements occur relatively
quickly, the process can exhibit a delayed dynamic response to
flow changes so it is not necessarily desirable to immediately
change the steam flow controller set point based upon current measurement values. Because changes to fluid flow would
abruptly (multiplicatively) affect steam flow, the lead/lag can be
used to modify the fluid flow signal being fed to the calculation to
temper the response of the control system to fluid flow changes.

Feedforward Action
Feedforward action is defined herein to be the utilization of process interaction to provide feedforward control without feedforward measurements per se. This differs from feedforward control
where process measurements allow the control system to respond
to process disturbances before the effects of the measured disturbances affect the controlled variable. Applying feedforward
action requires an in-depth (but not necessarily mathematical)
understanding of the process, process equipment, and their
dynamics. Depending on the application, implementation of
feedforward action could be described as the judicious selection
of the process measurements in a cascade control strategy, or perhaps the careful selection of the final control element.
The relatively simple heat exchanger control application of Figure
12-9 implements a deceptively complex control strategy. The temperature measurement and controller are relatively straightforward.
However, steam is manipulated with a remotely-set pressure regulator that effectively implements a cascade control strategy by varying steam pressure based upon the temperature controller output
instead of the usual handwheel or set screw of a typical pressure
regulator. If the inlet fluid temperature decreases (increases) or the
fluid flow increases (decreases), more (less) steam will condense,
causing the steam pressure to fall and causing the pressure regulator to more open (closed) to compensate for the load. Movement of
the pressure regulator occurs almost immediately after the process
change is sensed by the steam in the heat exchanger. This reduces
the effect of the process change on the outlet temperature as well
as effectively reducing the effect of steam header pressure variations on the fluid outlet temperature. The steam pressure measurement is provided for diagnostic purposes.
In another application, the purpose of the makeup air flow control shown in Figure 12-10, is to provide a constant air flow to

imo-spitzer-12.indd 119

9/15/09 11:38 AM

120

Advanced Regulatory Control Tools

FIG. 12-9.

Heat exchanger controls with feedforward action

the downstream load such as an incinerator. This control strategy

appears to be similar to the feedforward ratio control described
above, except that the makeup air set point is the difference
between the desired outlet air flow and the wild air flow
(instead of a ratio). Applying a variable speed drive to vary fan
speed to manipulate makeup air flow is an equipment selection
that was no accident.
If the controlled air is manipulated using controlling a valve or
damper, wild air flow changes would have a relatively small
effect on makeup air flow during transients. Maintaining a relatively stable makeup air flow might seem to be a virtue, however
the outlet air flow would be directly (and perhaps significantly)
affected by wild air flow changes. Therefore, wild flow
variations would result in makeup air flow controller set point
changes that would result in movement of the valve or damper
until sufficient time elapses such that the air flow can stabilize at
its new set point.
Manipulating the makeup air flow using a variable speed drive
tends to operate the fan on the relatively flat portion of its curve.
Therefore, varying downstream pressure changes that result
from wild flow variations tend to cause the makeup air flow to

imo-spitzer-12.indd 120

9/15/09 11:38 AM

Advanced Regulatory Control Tools

FIG. 12-10.

121

Makeup air flow control with feedforward action

substantially compensate for these wild flow variations with

little change in fan speed. Because of this attribute, this control
strategy will better control the total flow in the face of wild
flow variations because the feedforward action of moving on the
fan curve occurs quickly through the process without the use of
measurements. In addition, variable speed drive fan operation is
more energy efficient as compared to using a valve or damper.

Constraint and Override Control

The control loop may be allowed to manipulate the final control
element freely, as long as a particular constraint is within limits.
However, when the constraint moves outside these limits, a controller may become active and assume control of the final control
element by either directly manipulating the final control element
or overriding the control loop, perhaps using a signal selector.

imo-spitzer-12.indd 121

9/15/09 11:38 AM

122

Advanced Regulatory Control Tools

Summary
Advanced regulatory control provide a powerful set of tools
to improve control of the process. Seemingly straight-forward,
these tools can be combined to produce results that are strikingly superior to the results achievable using regulatory control.

imo-spitzer-12.indd 122

9/15/09 11:38 AM

 Chapter 13

Advanced Regulatory Control

Design and Implementation
Advanced regulatory control should be designed from the top
down and implemented from the bottom up. Failure to use this
overall implementation strategy can reduce the effectiveness of
the installed control system.

Design
Top-down design (see Figure 13-1) implies that the detailed
specifications of a particular instrument (bottom level, containing details) cannot be implemented until the process control
goals and control strategies (top level, containing more general
ideas) are developed and finalized. Similarly, a number of intermediate steps, which may overlap somewhat between the top
and bottom, should be performed, starting from the most general
and constantly moving to the more specific.
Define the Product and Its Quality Examination of a process

flow sheet or process and instrumentation diagrams (P&IDs) will

usually result in a list of flows, levels, pressures, and temperatures to be controlled in order to manage the process. Implementation based upon this list will result in an installation that

FIG. 13-1.

Top-down design

123

imo-spitzer-13.indd 123

9/15/09 11:45 AM

124

Advanced Regulatory Control Design and Implementation

does not streamline the process, make the process more efficient, nor exploit available control technology to improve control of the process and its overall performance.
Before attempting to instrument the process, the output of
the process should be identified and its quality defined. THE
PURPOSE OF THE PROCESS SHOULD BE CLEARLY UNDERSTOOD. Quality may include a number of factors that include
peak flow, average flow, minimum flow, flow stability, composition, color, texture, temperature, pressure, and the like.
For example, a 99 percent pure chemical produced at a nearly
constant flow rate with less than 0.5 percent of a given
impurity could be the specification of a chemical intermediate produced by the process to be controlled, even though the
intermediate is used in the next process and not sold. On the
other hand, maintaining the level of a tank perfectly stable
is not a product, despite its potentially critical nature in the
process.
Define the Raw Materials Having defined the product of the

process, the raw materials used to make the product should

be defined and their quality quantified. Raw materials in a
given process can include chemicals, materials, utilities, process equipment, labor, and other items needed to produce
the product.
For example, utilities such as steam are often taken for granted.
However, if the steam generation and distribution systems are
not stiff, soot blows, rain on uninsulated pipes, and the like
can reduce steam pressure and upset the process. Such knowledge could trigger the consideration of pressure compensation for
steam flow control measurements.
Define the Boundaries of the Process Because control problems

can stem from any of these raw materials, the boundaries of the
project should include systems that affect all of the raw materials, even if improvements to these systems will not be currently
funded. This approach is preferred to designing the control system to attempt to hide a real problem that can be more difficult
and costly to correct later. Often, subtle changes in the operation
of a process can have dramatic positive or negative effects on
upstream, downstream, and related processes.
For example, investigation may show that the boiler steam
pressure controller may be improperly adjusted or defective.

imo-spitzer-13.indd 124

9/15/09 11:45 AM

Advanced Regulatory Control Design and Implementation

125

In the absence of this information, installing pressure compensation may alleviate the problem at hand, but the underlying
problem may adversely affect other processes. Treating the
problem does not preclude pressure compensation because
even with a constant pressure steam supply, other considerations, such as piping losses and the effects of multiple users,
can affect the process when temperature compensation is not
applied. In general, improved process performance can be
achieved when variations in raw material uniformity and quality are reduced.
Define Quality Measurements for Evaluation Recognizing that

product quality and capacity are important, which measurements to make to determine quality and where to make those
measurements should be defined for evaluation. Likewise,
measurements of in-process material should be evaluated. By
performing this step, control of the process to achieve final
product quality can be broken up into smaller processes that
are often easier to address.
For example, producing 99 percent pure material may consist
of a reaction that produces 90 percent pure material, followed
by a concentration process that produces 99 percent pure material. This problem may now be broken into two simpler problems: appropriate assays may show that the reacted material
varies between 80 and 95 percent purity, and the product varies between 98 and 99.5 percent purity. These components can
now be addressed individually, recognizing that reducing the
variations in composition from the reactor (using an in-process
quality measurement) may obviate the need for upgrading the
concentration process.
Define Quality Measurements for On-Line Operation When the
evaluation of the data is complete, on-line process quality measurements can be located at various locations in the process,
if possible and practical. These measurements are important,
because the purpose of the control system is to stabilize these
parameters. In addition, these measurements can be viewed in
real time so as to allow timely correction that can significantly
improve the operation of the process.
Develop the Control Strategy The control strategy should

be developed to stabilize and control the on-line quality measurements, utilizing the regulatory and advanced regulatory
control tools available. The strategy should involve measurement and control as close as possible to the process to reduce

imo-spitzer-13.indd 125

9/15/09 11:45 AM

126

Advanced Regulatory Control Design and Implementation

dead time. For example, if vapor flow is important and reacts

quickly to process changes, the flow should be measured
instead of waiting tens of minutes for a downstream condensation tank level to rise. Measurements such as these transform
slow loops into faster loops that can react more quickly to process upsets, thus making the process easier to control.
Define the process variables that should be managedincluding calculated measurements based upon more than one process variable. Note that despite the relatively large number of
process flow measurements, most flow measurements are not
controlled per se but rather used in cascade control strategies,
used in calculations, or indicated. The controlled process variable such as flow (sometimes), level, pressure, temperature or
composition can usually be paired in a cascade control strategy
with an appropriate manipulated variabletypically a flow.
Understand that cascading slow loops into fast loops will
tend to stabilize both the controlled variable and the manipulated variable. Improved process stabilization may require that
virtually all flows be controlled as the slave in a cascade control loop.
Fast loops should manipulate and control one process variable
only. For example, if it is desired to control the sum of two
flows where only one can be manipulated, the flow controller
should manipulate the flow directly, and the flow controller set
point should be calculated by subtracting the wild flow from
the desired flow. In other words, do not build additional process dynamics into process variables that will be detrimental
to loop tuning.
Define Instrumentation Hardware The function and location
of field instruments performing the measurements needed to
implement the control strategy should be defined. The control
instrumentation should also be defined. This information can be
used to estimate the project cost.
Critique the Process When the control system has been developed, the control system and process should be examined to
determine where improvements can be made. For example, the
process engineer may include equipment that requires additional instrumentation to manage, but which may not seem necessary given the availability of recently introduced instrumentation. Location of a piece of equipment on the next higher platform may provide sufficient head for gravity flow, eliminating

imo-spitzer-13.indd 126

9/15/09 11:45 AM

Advanced Regulatory Control Design and Implementation

127

the need for a pump while allowing more space on the lower
platform for maintenance. The implementation of some measurements may require that the process be modified.
Proposed improvements should be evaluated for their technical
merit to avoid missing an opportunity for improvement or causing costly changes during implementation.

Implementation
Bottom-up implementation (see Figure 13-2) implies that the
details regarding the overall process strategy (top-level control)
cannot be implemented until the details regarding the process
equipment and field instrumentation (bottom-level measurement and control) are developed and finalized. Similarly, a
number of intermediate steps that may overlap somewhat
between the bottom and the top should be performed, starting
from the specific and constantly moving to the more general.
Implementation should be based upon the information developed during the design phase of the project.
Provide Equipment Data Information such as openings on tanks,
equipment nozzle data, and the like, are required by the equipment manufacturer in order to fabricate the equipment. Due to
the long lead times involved as compared to instrumentation,
this information should be provided to the appropriate party relatively early in the project.
Specify Hardware The field equipment and control hardware

should be specified to provide timely delivery and allow sufficient time for software development. Field equipment should be
specified to measure the desired process measurement.

FIG. 13-2.

imo-spitzer-13.indd 127

Bottom-up implementation

9/15/09 11:45 AM

128

Advanced Regulatory Control Design and Implementation

Develop Software Software development to implement the control strategy should proceed from the bottom up. This should
start with the definition of inputs/outputs and progress to fast
control loops, slow control loops, and beyond. Care should be
taken to make sure that the operator interface is easy to use and
will achieve bumpless, balanceless transfer between the various
controller operating modes, such as manual, automatic, cascade,
hold, tracking, and the like.
Installation Careful attention should be paid to the installation
details to ensure that the instrument is installed correctly and
measures the desired process variable at the proper location without introducing delay.
Start-up Start-up should proceed from the bottom up by verifying

the process equipment, field instrumentation, fast control loops,

slow control loops, and beyond. Fast loops are usually tuned for
set point changes; slow loops with manual setpoints are usually
tuned for process disturbances.

Summary
The implementation of a successful advanced regulatory control
system entails top-down design and bottom-up implementation,
that is, design should start from the general and migrate to the
details, whereas implementation should start with the details and
migrate to the general. Violation of these guidelines can destine
projects to failure before project activities commence.

imo-spitzer-13.indd 128

9/15/09 11:45 AM

Chapter 14

Applying Advanced
Regulatory Control
To illustrate the application of advanced regulatory control, the
stripping process in Figure 14-1 is considered. The intent of this
exercise is to illustrate the implementation of the design process
proposed in the previous chapter. Due to the numerous process

FIG. 14-1.

Stripping process

129

imo-spitzer-14.indd 129

9/15/09 1:39 PM

130

Applying Advanced Regulatory Control

constraints and idiosyncracies, the proposed control system

should not be construed as a viable or an appropriate way to control a particular process.
The process consists of a vessel under vacuum (provided by a
steam jet) in which the liquid is boiled, using a thermosyphon
reboiler. The vapor is condensed in a water-cooled condenser that
allows volatile material to exit the system via the vent. A pump is
used to transfer the liquid to another part of the process.
The top-down design procedure can be used to develop a control
strategy.

Define the Product and Its Quality

The purpose of this process is to produce a liquid that contains
only small amounts of volatile material. The quantity of the material processed should approximately equal the flow of liquid from
the upstream process, which contains small amounts of volatiles.
In addition, the process produces a waste stream that goes to
vent, warm cooling water, and steam condensate. It is desired that
the amount of nonvolatile material in the vent steam be reduced.

Define the Raw Materials

The raw materials of this process are the liquid with unknown
amounts of volatile material, cooling water, steam to the reboiler,
and steam to the jet. Steam pressure can vary by about 10 percent
of the absolute steam pressure. Cooling water temperature will
vary with time of day and season; cooling water flow can vary
because of occasional manual balancing by the operator.

Define the Boundaries of the Process

The process is ostensibly limited to the process shown on the
sketch; however, the upstream liquid process, downstream liquid process, cooling water supply, cooling water return, steam
supplies, and condensate return systems will not be considered
off limits.

Define Quality Measurements for Evaluation

The quality of the product is determined by the quantity of volatile
materials in the liquid leaving the process. Sampling this liquid
and analyzing for volatile material is a measure of product quality.

imo-spitzer-14.indd 130

9/15/09 1:39 PM

Applying Advanced Regulatory Control

131

Define Quality Measurements for On-Line Operation

Installing an on-line analyzer system is expensive and deemed
impractical in this application. However, it is known that if the
liquid in the vessel is maintained above a certain temperature at
a given vacuum, the liquid exiting the process will contain a sufficiently low amount of volatile materials.
The temperature of the liquid in the vessel could be used as an
indication of the on-line quality of the liquid. This approach
would seem reasonable if the vacuum were maintained below
an certain vacuum. However, the vacuum can vary significantly,
so the temperature measurement (which is intended to infer
composition) could be mathematically corrected for changing
vacuum, that could be implemented using an absolute pressure
transmitter. In this manner, composition can be inferred more
accurately than by using temperature alone to infer composition
(See Figure 14-2).

Develop the Control Strategy

The control strategy illustrated in Figure. 14-2 can be developed
starting with the loops closest to the process. Because virtually
all controllers that operate final control elements manipulate flow
in order to control another parameter, the application of flow controls should be considered first, followed by the management of
volumes, and finally, the control of composition.
As was previously stated, most manipulated variables are flows,
which in this application include:
liquid flow to the next process,
steam flow to the jet, and
steam flow to the reboiler.
The cooling water flow to the condenser is an additional, manipulated variable that is ignored because the cooling water will be
manually balanced.
It is assumed that the liquid flow into the process cannot be
manipulated, but the control of the upstream process should be
examined to determine whether improvements can be made to
steady this flow.
Variables that should be controlled and/or managed include:

imo-spitzer-14.indd 131

9/15/09 1:39 PM

FIG. 14-2.

Stripping process with fast control loops and controlled variable measurements

132

imo-spitzer-14.indd 132

Applying Advanced Regulatory Control

9/15/09 1:39 PM

Applying Advanced Regulatory Control

133

vessel liquid level,

vessel vacuum, and
liquid composition in the vessel.
The condenser outlet temperature is ignored, because it is
assumed that the operator will balance the system with the
proper amount of cooling water.
Ideally, manipulating one variable will affect only one controlled variable. However, analysis of the process will reveal
that this is not the case in this application. For example, consider that a steam flow increase to the reboiler will increase the
temperature of the liquid, causing more boiling, which in turn
increases the absolute pressure in the vessel and lowers the liquid level.
To stabilize and linearize the manipulated variables, it would
initially appear that flow control loops should be applied to
all flows that can be manipulated because these loops are fast
process variables that will react significantly faster than absolute pressure or composition loops. However, steam jet suction is determined by steam pressure, which reacts quickly to
valve manipulation and can be considered a fast control loop.
In this application, steam pressure will be controlled instead
of steam flow.
Vessel temperature should be measured to infer composition.
The incoming liquid flow should be measured even though it
cannot be controlled.
Vessel Liquid Level As alluded to above, the liquid level in the

vessel can be affected by changing any of the manipulated variables, but it is most directly affected by the liquid flow leaving
the vessel. This level should be controlled at a level slightly
above the top of the reboiler tube sheet in order to utilize the
full heat exchanger surface and maintain steady recirculation.
To control the level and stabilize the flow to the next process,
a level control loop cascading level into flow would seem
appropriate (see Figure. 14-3). It is also appropriate to investigate the control loop (if any) for the liquid that feeds this
process, because its stabilization (perhaps using cascade control) might reduce the fluctuation of the liquid flow leaving
the vessel.

imo-spitzer-14.indd 133

9/15/09 1:39 PM

134

imo-spitzer-14.indd 134

Applying Advanced Regulatory Control

FIG. 14-3.

Vessel liquid level control

FIG. 14-4.

Vessel vacuum control

9/15/09 1:39 PM

Applying Advanced Regulatory Control

135

Vessel Vacuum Vessel vacuum is indirectly affected by reboiler

steam flow, somewhat more affected by liquid flow into the vessel (which cannot be manipulated in this application), but much
more directly affected by adjusting the vacuum source. An
absolute pressure controller cascaded to the steam jet pressure
controller can be applied to control vacuum in the vessel. See
Figure. 14-4.
The steam pressure control loop can be more economically
and more simply implemented using a remotely set pressure
regulator.
Liquid Composition in the Vessel Based upon process knowl-

edge, the measured liquid temperature in the vessel can be

corrected to a reference absolute pressure by calculating the
temperature at the reference absolute pressure condition that
would have the same percent volatile material as the actual
liquid at the measured temperature and absolute pressure. In
this manner, controlling the corrected temperature can control
the liquid composition in the vessel and compensate for vessel
pressure fluctuations.
Vessel temperature is most affected by steam to the reboiler,
therefore, a composition controller cascaded to reboiler steam
flow can be applied (see Figure. 14-5). Pressure compensation
that could be applied to the steam flow measurement to compensate for steam header pressure disturbances is not shown.
Feedforward Control Feedforward control may be needed to
allow the control loops to react to load changes, that is, changes
to the liquid flow entering the vessel. If not corrected, load
changes can change the vacuum, liquid temperature, and liquid
level. Therefore, in the steady state, both steam flow and the
liquid flow leaving the vessel should be increased (decreased)
when load increases (decreases).

The cascade absolute pressure control loop will likely react relatively quickly to a load change through its feedback algorithm
because the load change will tend to quickly affect the vessel
pressure. The outputs of the level controller and the corrected
temperature controller may require additive feedforward. However, if the reboiler steam control loop were replaced with a
control loop that used feedforward action (using a remotely set
pressure regulator), the corrected temperature controller should
not require feedforward control.

imo-spitzer-14.indd 135

9/15/09 1:39 PM

136

FIG. 14-5.

Applying Advanced Regulatory Control

Liquid temperature control

Dynamic Compensation As mentioned above, the absolute

pressure control loop should be relatively responsive to pressure changes and should not require dynamic compensation.
However, the response of the corrected temperature controller
may benefit from dynamic compensation of the feedforward
signal (when feedforward action is not used). Dynamic compensation of the level control loop feedforward signal may be
required (see Figure. 14-6).
Logic and Data Manipulation A controller can effectively lose
control of the process when its output is at its minimum or maximum value. Override control can be used in some applications
to enable other control loops to attempt to keep the process
under control. For example, if the reboiler steam control valve
is completely open, a corrected temperature override controller
can be activated to generate the set point of the steam jet pressure controller in an attempt to maintain liquid composition by
increasing steam jet pressure.

Constraint control can also be used to meet certain control

objectives by relaxing other control objectives. For example, if

imo-spitzer-14.indd 136

9/15/09 1:39 PM

FIG. 14-6. Stripping process with feedforward action, feedforward control, dynamic compensation, override control, constraint control, and
liquid fed to the inlet of the reboiler

Applying Advanced Regulatory Control

imo-spitzer-14.indd 137

137

9/15/09 1:39 PM

138

Applying Advanced Regulatory Control

the liquid could not be heated above a certain temperature

without appreciable degradation, a temperature constraint controller may become active and set the reboiler steam flow to
maintain the maximum allowable temperature, at the expense of
the composition control objective. In an attempt to compensate
for the relaxation of the composition control objective, process
conditions may also activate the corrected temperature override
controller.

Define Instrumentation Hardware

Knowledge of the process and the control strategy developed
above can be used to define the instrumentation necessary to
implement the control strategy. In this application, field instrumentation suitable for hazardous electrical locations and an
industrial environment is necessary. The controllers or control
system should be capable of implementing the above cited algorithms and provide flexibility for future changes.

Critique the Process

As stated above, the upstream process should be examined to
determine if the incoming flow can be stabilized or, better yet,
manipulated in an attempt to stabilize the flow leaving this
process.
A surge tank might be considered between this process and the
downstream process. If the liquid flow leaving this process can be
adequately stabilized, the surge tank may be eliminated or, better
yet, not installed.
Introduction of the liquid at the reboiler inlet may dampen the
effect of load changes on vessel pressure and temperature. This
would be especially beneficial when feedforward action is used,
because the reboiler will modulate the steam valve quickly in
response to the load change before the liquid reaches the vessel.

Summary
As can be seen from the above analysis, control system design
takes many facets of the operation into account. In addition, even
the experts will not agree upon one correct control system design
for a process. However, proper application of advanced regulatory
control can alleviate the need for more sophisticated techniques
in all but the most demanding control applications.

imo-spitzer-14.indd 138

9/15/09 1:39 PM

Index

Additive feedforward control ........................................................................113

Advanced regulatory control .................................................................123, 129
Tools ...........................................................................................................103
Analog electronic controllers...........................................................................57
Analog signal transmission ..............................................................................33
Anti-reset windup ............................................................................................55
Automatic (AUTO) mode .................................................................................29
Bias adjustments...............................................................................................68
Bottom-up implementation............................................................................127
Bumpless transfer .......................................................................................57, 66
Calculation......................................................................................................103
Calibration ........................................................................................................48
Capacitance level transmitter ....................................................................20, 21
Capacitance technology ...................................................................................16
Cascade control ......................................................................................110113
Characterization .............................................................................................104
Check valve.......................................................................................................41
Closed-loop technique .....................................................................................82
Composition measurement ............................................................................108
Constraint .......................................................................................................121
Control
Additive feedforward ................................................................................113
Advanced regulatory .........................................................................123, 129
Cascade ..............................................................................................110113
Derivative ...............................................................................................75, 77
Feedforward ...............................................................................113, 116, 135
Flow .............................................................................................................88
Integral ...................................................................................................73, 77
Level .............................................................................................................89
Manual ...........................................................................................................3
Proportional ...........................................................................................71, 77
Ratio ...........................................................................................................115
Regulatory ....................................................................................................27
Control action ...................................................................................................61
Direct ............................................................................................................61
Reverse .........................................................................................................61
Control algorithm .......................................................................................29, 66
Control dead band ............................................................................................58
Control loop
Dynamics .....................................................................................................98
Interaction ..........................................................................................100, 112
Response ......................................................................................................95
Tuning ....................................................................................................46, 79
Control panel ....................................................................................................33
Control room.....................................................................................................33
Control strategy ........................................................................................87, 131
Control valve ................................................................30, 33, 37, 39, 41, 52, 53
Failure mode ................................................................................................40
Shutoff..........................................................................................................41
Sizing ...........................................................................................................39

139

imo-spitzer-15in.indd 139

9/2/09 12:24 PM

140

Index
Controller ..........................................................................................................27
Analog electronic.........................................................................................57
Digital ...........................................................................................................60
Faceplate ......................................................................................................60
Gain ..............................................................................................................72
Indicator .......................................................................................................60
Input adjustments ........................................................................................50
Master.................................................................................................111, 112
Output ..........................................................................................................62
PID regulatory ..............................................................................................71
Range ............................................................................................................60
Settings.........................................................................................................84
Strategies ......................................................................................................79
Tuning ..............................................................................................45, 55, 79
Coriolis mass flowmeter .............................................................................12, 13
Corrosion ..........................................................................................................16
Current-to-pneumatic (I/P) converter ..............................................................39
Dampening..................................................................................................47, 50
Damper..............................................................................................................37
Dead time ............................................................................................97, 98, 104
Density measurement .......................................................................................12
Derivative control .......................................................................................75, 77
Derivative controller constant .........................................................................76
Derivativeon process variable ..........................................................................76
Detuning .........................................................................................................112
Diaphragm actuators ........................................................................................39
Diaphragm seals ...............................................................................................16
Differential pressure flowmeters......................................................................12
Differential pressure level measurements .......................................................14
Differential pressure transmitter......................................................................14
Closed tank ..................................................................................................23
Open tank.....................................................................................................22
Digital communication...............................................................................33, 50
Digital controller ..............................................................................................60
Digital protocol .................................................................................................34
Digital signal transmission...............................................................................33
Direct/Reverse control action ..........................................................................61
Direct ............................................................................................................61
Reverse .........................................................................................................61
Displacer level transmitter ...............................................................................22
Distributed control system ...............................................................................35
Disturbances .....................................................................................................27
Dynamic compensation..................................................................................135
Electrically operated valves .............................................................................63
Enabling switches.............................................................................................65
Energy savings ..................................................................................................41
Equipment adjustments ...................................................................................46
Error signal .......................................................................................................71
Fail closed (F.C.) ...............................................................................................63
Fail open (F.O.) .................................................................................................63

imo-spitzer-15in.indd 140

9/2/09 12:24 PM

Index

141

Fail-safe.............................................................................................................63
Failure modes ...................................................................................................62
Loss of power ...............................................................................................62
Fast and slow loop tuning................................................................................98
Fast control loops .............................................................................................98
Fast loops ..........................................................................................................95
Feedbackc ontrol ............................................................................................109
Feedback trim .........................................................................................113, 116
Feedforward action ........................................................................................119
Feedforward control .......................................................................113, 116, 135
With dynamic compensation ....................................................................118
Field equipment tuning ...................................................................................45
Final control element ...........................................................................27, 37, 62
Adjustments .................................................................................................52
Float level transmitter ......................................................................................18
Flow control .....................................................................................................88
Flow control loop .......................................................................................53, 95
Flow measurement .....................................................................................8, 106
Flowmeter ...........................................................................................8, 9, 11, 50
Coriolis mass..........................................................................................12, 13
Differential pressure ....................................................................................12
Magnetic.......................................................................................................10
Orifice plate (differential pressure).............................................................15
Oval gear positive displacement ...................................................................9
Rotameter (variable area).............................................................................17
Target ............................................................................................................16
Thermal ........................................................................................................14
Ultrasonic.....................................................................................................13
Vortex shedding (oscillatory) ......................................................................11
Flowmeter type...................................................................................................8
Volumetric .....................................................................................................8
Positive displacement ............................................................................8
Velocity ..........................................................................................................8
Magnetic .................................................................................................8
Oscillatory ..............................................................................................8
Turbine ....................................................................................................8
Ultrasonic ...............................................................................................8
Mass ...............................................................................................................8
Coriolis....................................................................................................8
Hydraulic Wheatstone Bridge ................................................................8
Thermal...................................................................................................8
Inferential.......................................................................................................8
Differential pressure ...............................................................................8
Target ......................................................................................................8
Variable area ...........................................................................................8
Fluid mass flow ..................................................................................................9
Fugitive emission regulations ..........................................................................39
Fuzzy logic control .........................................................................................110
Heat exchanger .................................................................................................97
Heat flow .........................................................................................................108
Hysteresis........................................................................................39, 41, 5254
I/P converter .....................................................................................................52

imo-spitzer-15in.indd 141

9/2/09 12:24 PM

142

Index
Infrared pyrometer ...........................................................................................19
Input damping adjustment.........................................................................51, 52
Installed valve characteristic ...........................................................................40
Integral control ............................................................................................73,77
Integral controller constant ..............................................................................74
Interact ............................................................................................................112
Interface level measurements ..........................................................................14
Lead/lag ..........................................................................................................118
Lead/lag algorithm .........................................................................................104
Level control loops ...........................................................................................96
Level control .....................................................................................................89
Level measurements .................................................................................14, 107
Floats ............................................................................................................14
Ultrasonics ...................................................................................................14
Radar ............................................................................................................14
Resistive tapes .............................................................................................14
Capacitance ..................................................................................................14
Level transmitter ..............................................................................................33
Limitations of regulatory control .....................................................................95
Logic and data manipulation .................................................................105, 136
Magnetic flowmeter ..........................................................................................10
Manual/automatic switch ................................................................................65
Manual control ...................................................................................................3
Manual (MAN) mode .......................................................................................29
Master controller ....................................................................................111, 112
Mathematical functions .................................................................................103
McMillan, G.K ............................................................................................83, 83
Measurement ....................................................................................................27
Measurement range ..........................................................................................60
Mechanical adjustments .................................................................................4 7
Multiplicative feedforward control ...............................................................115
With feedback trim ....................................................................................116
Nichols, N.B......................................................................................................82
Noise filtration ..................................................................................................51
On-line operation ...........................................................................................131
Orifice plate (differential pressure) flowmeter ................................................15
Output
Limits ...........................................................................................................67
Preset valve ..................................................................................................65
Slew rate ......................................................................................................69
Tracking .......................................................................................................66
Oval gear positive displacement flowmeter ......................................................9
Override control .............................................................................................121
Packing..............................................................................................................39
Pairing .........................................................................................................87, 88
Period of oscillation .........................................................................................82
PID (proportional, integral, derivative) ...........................................................71
Algorithm .....................................................................................................71

imo-spitzer-15in.indd 142

9/2/09 12:24 PM

Index

143

Control ...................................................................................................71, 77
Controller implementation..........................................................................78
Regulatory controllers .................................................................................71
Pneumatic controllers ......................................................................................57
Pneumatic signal ..............................................................................................52
Pneumatic transmission ...................................................................................50
Positioner ....................................................................................................52, 53
Pressure
And temperature compensation ...............................................................106
Control .........................................................................................................91
Measurement .......................................................................................16, 107
Transmitter sensing module ........................................................................23
Transmitters .................................................................................................16
Process disturbance ..........................................................................................79
Process noise ....................................................................................................50
Process upsets...................................................................................................99
Proportional band (PB).....................................................................................72
Proportional control ...................................................................................71, 77
Protocol .............................................................................................................33
Pump curve.......................................................................................................37
Quarter amplitude oscillation method ............................................................82
Ramp function ................................................................................................105
Rate ...................................................................................................................76
Ratio adjustments .............................................................................................68
Ratio control ...................................................................................................115
Regulatory control ............................................................................................27
Limitations of...............................................................................................95
Loop .......................................................................................................27, 30
Tools, advanced .........................................................................................103
Regulatory controller........................................................................................71
Remote set point .......................................................................................68, 112
Reset ..................................................................................................................74
Resistance temperature detector ......................................................................24
Resistive tape level transmitter........................................................................20
Reynolds number .............................................................................................11
Rotameter (variable area) flowmeter ................................................................17
Self-contained regulating valve .......................................................................31
Self-tuning method...........................................................................................84
Sensor adjustments ..........................................................................................46
Setpoint (SP) ...............................................................................................27, 29
Changes ........................................................................................................79
Limits ...........................................................................................................68
Tracking .......................................................................................................57
Signal limiter ..................................................................................................105
Signal selection ..............................................................................................105
Signal transmission ..........................................................................................50
Slave controller ......................................................................................111, 112
Slew rates..........................................................................................................69
Adjustments .................................................................................................69
Slow control loops .....................................................................................95, 98

imo-spitzer-15in.indd 143

9/2/09 12:24 PM

144

Index
Smart transmitters ............................................................................21, 33, 49
Straight run .......................................................................................................11
Submerged displacer ........................................................................................14
System curve ....................................................................................................37
Target flowmeter ...............................................................................................16
Temperature control .........................................................................................92
Temperature control loops ...............................................................................97
Temperature measurement.................................................................18, 19, 108
Thermometers ..............................................................................................19
Thermocouples ............................................................................................19
Thermopiles .................................................................................................19
Thermistors ..................................................................................................19
Resistance temperature detectors (RTD).....................................................19
Infrared pyrometers .....................................................................................19
Temperature scales ...........................................................................................19
Thermal flowmeter ...........................................................................................14
Thermocouple ..................................................................................................24
Thermowell installations .................................................................................25
Time constants..................................................................................................97
Top-down design ............................................................................................123
Totalizer ............................................................................................................70
Tracking input ..................................................................................................67
Transmitter adjustments...................................................................................48
Transmitter dropout feature .............................................................................50
Trial and error method .....................................................................................84
Tuning
Control loop ...........................................................................................46, 79
Controller .........................................................................................45, 55, 79
Fast and slow loop.......................................................................................98
Field equipment ..........................................................................................45
For process disturbances .............................................................................81
For reduced overshoot.................................................................................81
For set point changes...................................................................................80
Self-tuning method ......................................................................................84
Turbine flowmeter ............................................................................................12
Two-phase flow.................................................................................................11
Ultimate gain ....................................................................................................82
Ultrasonic flowmeter ........................................................................................13
Ultrasonic level transmitter .............................................................................19
Valve .................................................................................................................27
Actuator .......................................................................................................39
Capacity .......................................................................................................40
Check............................................................................................................41
Control .....................................................................30, 33, 37, 39, 41, 52, 53
Output preset ...............................................................................................65
Position ........................................................................................................40
Positioners ...................................................................................................40
Self-contained regulating ............................................................................31
Trim characteristic .......................................................................................40
Variable frequency drives ................................................................................41

imo-spitzer-15in.indd 144

9/2/09 12:24 PM

Index

145

Variable gain ...................................................................................................109

Variable speed drives .......................................................37, 41, 52, 54, 63, 120
Volumetric fluid flow .........................................................................................9
Vortex shedding (oscillatory) flowmeter .........................................................11
Ziegler, J.G ........................................................................................................82

imo-spitzer-15in.indd 145

9/2/09 12:24 PM

by David W. Spitzer
Without modern instrumentation control, industry would be at a standstill. Whether the
control hardware consists of local mechanical, pneumatic, or electronic controls, electronic
or pneumatic analog controllers, microprocessor-based controllers, programmable logic controllers, personal computer-based controls, or a distributed control system, a keen awareness
of advanced regulatory control is required by engineers, technicians, educators, managers,
salespersons, and marketing personnel in order that appropriate control technology be applied to industrial processes.
This book is intended to describe advanced regulatory control and its application to continuous processes in a nonmathematical format and in as practical a manner as possible in
order to be of benefit to all skill levels. Manual and regulatory control are described in the
text as a prelude to advanced regulatory control so that their differences might be explored.
It focuses on control but integrates the operation of measurement devices and final control
elements into discussions, because their importance is often the limiting factor in control
system performance.
Readers will find
An explanation of the operation and limitations of feedback control
Illustrations of the integration of cascade and feedforward control into feedback
control loops
A clear approach to design and implement advanced regulatory control
About the Author
David W. Spitzer is a Principal at Spitzer and Boyes, LLC, specializing
in instrumentation and control consulting for industrial users, manufacturers, and representatives world-wide, including serving as an expert
witness. Mr. Spitzer previously worked for Nepera Chemical, Mobay
Chemical, and U.S. Steel, proposing and implementing many control
and energy-saving projects.

An instructor at the ISA Training Center in Research Triangle Park, N.C., throughout
the United States, and internationally in the ISA Short Course Program, Mr. Spitzer is the
author of Industrial Flow Measurement (ISA Press), Variable Speed Drives: Principles and Applications for Energy Cost Savings (ISA Press), Consumer Guides to Coriolis Mass, Differential Pressure,
Magnetic, Ultrasonic, and Vortex Shedding Flowmeters, and Contact and Non-Contact Level Gauges
(Copperhill and Pointer), as well as the Volume Editor of Flow Measurement (ISA Press).
Mr. Spitzer has also written some 150 articles in various technical magazines.
ISBN: 978-1-60650-009-5

90000

www.momentumpress.net

9 781606 500095

Spitzer

Mr. Spitzer is a Life Fellow of the International Society of Automation (ISA) and serves on the American Society of Mechanical Engineers (ASME) Measurement of Fluid Control in Closed Conduits committee (MFC).

Advanced Regulatory Control

Advanced Regulatory Control

Advanced
Regulatory
Control
Applications and Techniques

by David W. Spitzer