Process Control

Jean-Pierre Corriou

Process Control
Theory and Applications

Second Edition

123
Jean-Pierre Corriou
LRGP-CNRS-ENSIC
University of Lorraine
Nancy Cedex
France

ISBN 978-3-319-61142-6 ISBN 978-3-319-61143-3 (eBook)

DOI 10.1007/978-3-319-61143-3

Library of Congress Control Number: 2017944293

MATLAB® is a registered trademark of The MathWorks, Inc., 1 Apple Hill Drive, Natick, MA
01760-2098, USA, [Link]

1st edition: © Springer-Verlag London 2004

2nd edition: © Springer International Publishing AG 2018, corrected publication October 2017
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publi-
cation does not imply, even in the absence of a specific statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
The author dedicates this book to his wife
Cécile, his children Alain and Gilles, who
showed so much patience and understanding
in the face of his being monopolized during
the realization of this book. He also dedicates
this book to the future generation, his
grandson and granddaughters Alexandre,
Anne and Lucie, who still have to trace their
own trajectory.
Preface

Organization of the Book

This book has been conceived to progressively introduce concepts of increasing

difficulty and allow learning of theories and control methods in a way which is not
too brutal. It contains different levels of reading (Fig. 1). In particular, the majority
of the first part can be undertaken by students beginning in control or by technicians
and engineers coming from the industrial world, having up to that point only
practical contact with control and a desire to improve their knowledge.

Part I Part II

Continuous Time Multivariable Control

Models by Transfer Function Matrices

Control

Part III Part IV

Discrete Time Discrete Time

Discrete Control
Identification
Optimal Control
Predictive Control

Part V Part VI
Process Control
Nonlinear Control
Reactors
Nonlinear Observers Distillation Columns
Process Examples

Fig. 1 General organization of the book

vii
viii Preface

The subsequent parts need a minimum previous knowledge in control. They also
allow the use of often higher-performance techniques. Without pretending to be
exhaustive, this book proposes a wide range of identification and control methods
applicable to processes and is accompanied by similar typical examples that provide
comparison elements.
This book does not pretend to compete with theoretical control books specialized
in one subject or another, e.g. identification, signal processing, multivariable con-
trol, robust control or nonlinear control. However, the readers whoever they are,
undergraduate or graduate students, engineers, researchers, professors, will find
numerous references and statements that enable understanding and application in
their own domain of a large number of these concepts, taking their inspiration from
the cases treated in the present book (Table 0.1).
Several controls are examined under different angles:
• single-input single-output internal model control in continuous and discrete time
and multivariable internal model control in discrete time,
• pole-placement control in continuous and discrete time,
• single-input single-output linear quadratic control by continuous transfer func-
tion and multivariable state-space linear quadratic control in continuous and
discrete time and
• single-input single-output generalized predictive control and multivariable
model predictive control, possibly with observer, linear or nonlinear.
The consideration of the same problem by different approaches thus enables a
thorough examination.
Examples are most often taken from the chemical engineering domain and
include in general chemical, biological and polymerization reactors, heat
exchangers, a catalytic cracking reactor (FCC) and distillation columns. These
examples are detailed, even at the numerical level, so that the used reasoning can be
verified and taken again by the reader. Simulations have been realized by means of
MATLAB® and Fortran77 codes.
Part I concerns single-input single-output continuous-time control. The chosen
presentation of single-input single-output feedback linear process control is clas-
sical and voluntarily simple for most of the part. It presents the pedagogical
advantage of decomposing the approach to a control problem, introduces a given
number of important notions and, according to our opinion, facilitates the under-
standing of discrete-time control, called digital control, and of nonlinear control.
Also, it is close, in its conception, to a large part of industrial practice, at least in the
domain of chemical engineering. Continuous PID control is abundantly treated, but
without exclusivity. The main types of dynamic models met in chemical engi-
neering are commented on, and system models are presented as well in state space
as transfer functions (Chap. 1). Control is first simply related to the PID controller
(Chap. 2). Stability is presented both for linear and for nonlinear systems. Thus, the
stability of a polymerization reactor is detailed by displaying the multiple stationary
states in relation to the physical behaviour of the reactor (Chap. 3). The design of
Preface ix

Table 0.1 Contents of the chapters

Part I
Dynamic Modelling, Open-Loop Dynamics Chapter 1
PID Linear Feedback Control Chapter 2
Linear and Nonlinear Stability Analysis Chapter 3
Design of Feedback PID Controllers, Pole-Placement Chapter 4
Linear Quadratic Control, Internal Model Control Chapter 5
Frequency Analysis, Robustness
Improved Controllers, Smith Predictor Chapter 6
Cascade, Feedforward Control
State Representation, Controllability, Observability Chapter 7
Realizations and Model Reduction
Part II
Multivariable Control by Transfer Function Matrices Chapter 8
Part III
Discrete-Time Generalities Chapter 9
Signal Processing
Identification Principles Chapter 10
Identification Models Chapter 11
Identification Algorithms Chapter 12
Part IV
Discrete Pole-Placement Chapter 13
Discrete PID
Discrete Internal Model Control Chapter 14
Optimal Control
Continuous LQ and LQG Control Chapter 15
Discrete LQ and LQG Control
SISO Generalized Predictive Control
MIMO Model Based Predictive Control Chapter 16
Part V
Nonlinear Control Chapter 17
Nonlinear Observers, Statistical Estimators Chapter 18
Part VI
Reactors Chapter 19
Distillation Columns Chapter 20
Process Examples, Benchmarks Chapter 21

controllers first deals with PID and then is broadened to internal model control,
which is very important industrially, pole-placement control and linear quadratic
control by means of continuous transfer functions (Chap. 4). Frequency analysis
begins classically by the analysis in Bode and Nyquist representations, but is then
extended to robustness and sensitivity functions (Chap. 5). The improvements of
controllers including time delay compensation, cascade control and feedforward
control are reviewed with application examples in industrial processes (Chap. 6).
x Preface

The first part finishes with the concepts of state representation for linear systems,
controllability and observability (Chap. 7). Some more difficult parts of different
chapters such as robustness, pole-placement control or linear quadratic control can
be tackled in a subsequent reading.
Part II consists of only one chapter which deals with multivariable control by
either continuous or discrete transfer function matrix. This choice was made because
of the relatively common practice of system representation chosen in process con-
trol. The chapter essentially presents general concepts for the design of a multi-
variable control system. Other types of multivariable control are treated in specific
chapters in other parts of the book: linear quadratic control and Gaussian linear
quadratic control in Chap. 14, model predictive control in Chap. 16 and nonlinear
multivariable control in Chap. 17. In fact, Part III can be studied before Part II.
Part III begins by considering signal processing whose general concepts are
necessary in identification and control. Then, the general aspects of digital control
and sampling are treated, and discrete transfer functions are introduced (Chap. 9).
The remainder of Part III is devoted to discrete-time identification. First, different
model types are presented and the principles of identification are explained (Chap.
10), and then, different types of models are presented (Chap. 11). Lastly, the main
algorithms of parametric identification are detailed with many indications on usage
precautions (Chap. 12). The parametric identification of a chemical reactor is
presented. Identification is treated as single-input single-output except with respect
to the Kalman filter, which can be applied to multi-input multi-output systems.
In Part IV, several classical types of digital control are studied. Chapter 13
describes pole-placement, digital PID and discrete internal model control as
single-input single-output control with application to the same chemical reactor. In
Chap. 14, optimal control is considered in the general framework of dynamic
optimization applicable to nonlinear continuous or discrete systems; it includes
general methods such as variational methods, Euler, Hamilton–Jacobi, Pontryagin
and dynamic programming. Linear quadratic control and linear quadratic Gaussian
control are presented in direct relation to optimal control both for continuous and
for discrete state-space system descriptions. An application of multivariable linear
quadratic Gaussian control to an extractive distillation column with two inputs and
two outputs is presented. Two types of predictive control are studied. Chapter 15
concerns generalized predictive control for single-input single-output systems
represented by their discrete transfer function with application to the previously
mentioned chemical reactor. Chapter 16 is devoted to model predictive control
applicable to large multivariable systems known by transfer functions or state-space
models. Furthermore, model predictive control is popular in industry as it allows
constraints to be taken into account. Two multivariable applications for a catalytic
cracking reactor (FCC) are shown.
Part V concerns nonlinear control presented through differential geometry
(Chap. 17) and state observers (Chap. 18). These are recent developments in control
and are potentially very powerful. To facilitate its approach, several concepts are
analysed from a linear point of view, and then, nonlinear control for a single-input
single-output system is studied with input-state and input-output linearization.
Preface xi

Nonlinear multivariable control is just outlined. State estimation is necessary in

nonlinear control. Chapter 18 on observers does not concern the linear Kalman filter
described in Part III, but considers nonlinear observers including the extended
Kalman filter and the high-gain observer, as well as statistical estimators.
Part VI considers two important classes of chemical processes: reactors and
distillation columns. In the previous parts, linear identification and linear control are
applied to the chemical reactor described in detail in Chap. 19. In this first chapter
of Part VI, the use of geometric nonlinear control, based on the knowledge model
of the process or coupled with a state observer, is explained for a chemical reactor
and a biological reactor. It must be noticed that this is the same simulated chemical
reactor that was used for identification of a linear model and several discrete linear
control methods. Chapter 20 sweeps the control methods used in distillation since
the 1970s until our epoch that is marked by the extensive use of model predictive
control and the start of industrial use of nonlinear control. Chapter 21 describes
different processes and benchmarks that can be used as more or less complicated
examples to test various control strategies.

Acknowledgements

Through the teaching that the author gives at the National School of Chemical
Industries (ENSIC, Nancy), the contacts that he maintains with many control
specialists and his knowledge of process industries, he realizes the difficulty in
sharing control concepts with engineers that have a chemistry or chemical engi-
neering background. Indeed, engineers responsible for a plant must master many
disciplines, e.g. chemistry, chemical engineering, fluid mechanics and heat transfer,
which allow them to understand the behaviour of their process or to explain the
malfunctions that they observe. If this book is useful to students of process control
who have originated from either chemical engineering or control, engineers in the
plant or doctorate students beginning in process control or who want to have an
overview of process control, the author’s objective will have been reached.
The author wishes to thank his many colleagues of ENSIC and the Chemical
Engineering Sciences Laboratory (LSGC), now entitled Reaction and Process
Engineering Laboratory who have encouraged him in his work, in particular ENSIC
Director A. Storck and LSGC Director D. Tondeur. He thanks his closest col-
leagues of the research group TASC (Treatment and Acquisition of chemical
information, process Simulation and Control) of LSGC, his doctorate students, in
particular in order of seniority K. Abidi, Z.L. Wang, S. Lucena, C. Gentric, M. Ben
Thabet, A. Maidi, I.D. Gil, A. Boum, who took interest in nonlinear control of
chemical and biological processes and who shared difficulties and satisfactions, and
also his other doctorate students in the hope that they did not suffer too much during
the long writing. He does not forget the staff for their pleasant collaboration climate
and the Computing Department staff who helped him in particular in Linux
installation. All world developers of Latex and Linux are gratefully acknowledged.
xii Preface

The author also wishes to thank his colleagues who read and criticized some
parts of the book: J. Ragot, Professor at the National Polytechnic Institute of
Lorraine, and P. Sibille, Associate Professor at Nancy I University, both members
of CRAN, Nancy, France; G. Thomas, Professor at Ecole Centrale of Lyon,
France; S. Rohani, Professor at Western University, Ontario, Canada; and U. Volk,
Engineer at Shell Godorf, Germany. Let M. Alamir, Director of research at
GIPSA-lab, Grenoble, France, who accepted the heavy task of reading the whole
book, finds here the expression of the author’s deep gratitude.

Nancy, France Jean-Pierre Corriou

April 2017
Contents

Part I Continuous-Time Control

1 Dynamic Modelling of Chemical Processes . . . . . . . . . . . . . . . . . . . . 3
1.1 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Applications of Process Control . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Process Description from the Control Engineer’s Viewpoint . . . 5
1.4 Model Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 State-Space Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.6 Examples of Models in Chemical Engineering . . . . . . . . . . . . . . 10
1.6.1 Lumped-Parameter Systems . . . . . . . . . . . . . . . . . . . . . 10
1.6.2 Distributed-Parameter Systems . . . . . . . . . . . . . . . . . . 19
1.6.3 Degrees of Freedom . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.7 Process Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.8 Order of a System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.9 Laplace Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.9.1 Linearization and Deviation Variables . . . . . . . . . . . . . 31
1.9.2 Some Important Properties of Laplace
Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.9.3 Transfer Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.9.4 Poles and Zeros of a Transfer Function . . . . . . . . . . . . 45
1.9.5 Qualitative Analysis of a System Response . . . . . . . . . 45
1.10 Linear Systems in State Space . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.10.1 General Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1.10.2 Analog Representation . . . . . . . . . . . . . . . . . . . . . . . . . 50
1.11 Dynamic Behaviour of Simple Processes . . . . . . . . . . . . . . . . . . 53
1.11.1 First-Order Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
1.11.2 Integrating Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
1.11.3 Second-Order Systems . . . . . . . . . . . . . . . . . . . . . . . . . 56
1.11.4 Higher-Order Systems . . . . . . . . . . . . . . . . . . . . . . . . . 62
1.11.5 Process Identification in the Continuous Domain. . . . . 68
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

xiii
xiv Contents

2 Linear Feedback Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

2.1 Design of a Feedback Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
2.1.1 Block Diagram of the Feedback Loop . . . . . . . . . . . . . 77
2.1.2 General Types of Controllers . . . . . . . . . . . . . . . . . . . . 79
2.1.3 Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2.1.4 Transmission Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.1.5 Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.2 Block Diagrams, Signal-Flow Graphs, Calculation Rules . . . . . . 86
2.3 Dynamics of Feedback-Controlled Processes . . . . . . . . . . . . . . . 95
2.3.1 Study of Different Actions . . . . . . . . . . . . . . . . . . . . . . 98
2.3.2 Influence of Proportional Action . . . . . . . . . . . . . . . . . 99
2.3.3 Influence of Integral Action . . . . . . . . . . . . . . . . . . . . . 104
2.3.4 Influence of Derivative Action. . . . . . . . . . . . . . . . . . . 107
2.3.5 Summary of Controllers Characteristics . . . . . . . . . . . . 113
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3 Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 117
3.1 Case of a System Defined by Its Transfer Function . . . . . . .... 117
3.2 State-Space Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 118
3.2.1 General Analysis for a Continuous Nonlinear
System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 118
3.2.2 Case of a Linear Continuous System . . . . . . . . . . .... 123
3.2.3 Case of a Nonlinear Continuous System:
The Polymerization Reactor . . . . . . . . . . . . . . . . . . . . . 126
3.2.4 State-Space Analysis of a Linear System . . . . . . . . . . . 131
3.3 Stability Analysis of Feedback Systems . . . . . . . . . . . . . . . . . . . 132
3.3.1 Routh–Hurwitz Criterion . . . . . . . . . . . . . . . . . . . . . . . 134
3.3.2 Root Locus Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 136
3.3.3 Frequency Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4 Design of Feedback Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.1 Performance Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.2 Transient Response Characteristics . . . . . . . . . . . . . . . . . . . . . . . 145
4.3 Performance Criteria for Design . . . . . . . . . . . . . . . . . . . . . . . . . 146
4.4 Choice of PID Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.4.1 General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.4.2 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.5 PID Controller Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
4.5.1 Tuning by Trial and Error . . . . . . . . . . . . . . . . . . . . . . 150
4.5.2 Sustained Oscillation Method . . . . . . . . . . . . . . . . . . . 151
4.5.3 Relay Oscillation Method . . . . . . . . . . . . . . . . . . . . . . 152
4.5.4 Process Reaction Curve Method . . . . . . . . . . . . . . . . . 157
Contents xv

4.5.5 Tuning Rule of Tavakoli and Fleming for PI

Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 160
4.5.6 Robust Tuning Rule for PID Controllers . . . . . . . .... 160
4.6 PID Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 161
4.6.1 PID Controller with Derivative Action
on the Measured Output . . . . . . . . . . . . . . . . . . . . . . . 162
4.6.2 Use of a Reference Trajectory . . . . . . . . . . . . . . . . . . . 162
4.6.3 Discretized PID Controller. . . . . . . . . . . . . . . . . . . . . . 164
4.6.4 Anti-Windup Controller . . . . . . . . . . . . . . . . . . . . . . . . 165
4.6.5 PID Control by On Off Action . . . . . . . . . . . . . . . . . 167
4.6.6 PH Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.7 Direct Synthesis Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.8 Internal Model Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.9 Pole-Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
4.9.1 Robustness of Pole-Placement Control . . . . . . . . . . . . 188
4.9.2 Unitary Feedback Controller . . . . . . . . . . . . . . . . . . . . 190
4.10 Linear Quadratic Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
4.10.1 Regulation Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . 192
4.10.2 Tracking Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
5 Frequency Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
5.1 Response of a Linear System to a Sinusoidal Input . . . . . . . . . . 199
5.1.1 Case of a First-Order Process . . . . . . . . . . . . . . . . . . . 199
5.1.2 Note on Complex Numbers . . . . . . . . . . . . . . . . . . . . . 201
5.1.3 Case of Any Linear Process . . . . . . . . . . . . . . . . . . . . 202
5.1.4 Case of Linear Systems in Series . . . . . . . . . . . . . . . . 204
5.2 Graphical Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
5.2.1 Bode Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
5.2.2 nth-Order System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
5.2.3 Nyquist Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
5.2.4 nth-Order System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
5.2.5 Black Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
5.3 Characterization of a System by Frequency Analysis . . . . . . . . . 213
5.4 Frequency Response of Feedback Controllers . . . . . . . . . . . . . . . 213
5.4.1 Proportional Controller . . . . . . . . . . . . . . . . . . . . . . . . 213
5.4.2 Proportional-Integral Controller . . . . . . . . . . . . . . . . . . 214
5.4.3 Ideal Proportional-Derivative Controller . . . . . . . . . . . 215
5.4.4 Proportional-Integral-Derivative Controller . . . . . . . . . 216
5.5 Bode Stability Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
5.6 Gain and Phase Margins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.6.1 Gain Margin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.6.2 Phase Margin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
xvi Contents

5.7 Nyquist Stability Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

5.8 Closed-Loop Frequency Response . . . . . . . . . . . . . . . . . . . . . . . 234
5.9 Internal Model Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.10 Robustness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.11 Summary for Controller Design . . . . . . . . . . . . . . . . . . . . . . . . . 257
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
6 Improvement of Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
6.1 Compensation of Time Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
6.2 Inverse Response Compensation . . . . . . . . . . . . . . . . . . . . . . . . . 263
6.3 Cascade Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
6.4 Selective Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
6.5 Split-Range Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
6.6 Feedforward Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
6.6.1 Generalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
6.6.2 Application in Distillation . . . . . . . . . . . . . . . . . . . . . . 274
6.6.3 Synthesis of a Feedforward Controller . . . . . . . . . . . . . 275
6.6.4 Realization of a Feedforward Controller . . . . . . . . . . . 278
6.6.5 Feedforward and Feedback Control . . . . . . . . . . . . . . . 280
6.7 Ratio Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
7 State Representation, Controllability and Observability . . . . . . . . . . 285
7.1 State Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
7.1.1 Monovariable System . . . . . . . . . . . . . . . . . . . . . . . . . 285
7.1.2 Multivariable System . . . . . . . . . . . . . . . . . . . . . . . . . . 287
7.2 Controllability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
7.3 Observability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
7.4 Realizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
7.5 Remark on Controllability and Observability
in Discrete Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 300
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 301

Part II Multivariable Control

8 Multivariable Control by Transfer Function Matrix . . . . . . . . .... 305
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 305
8.2 Representation of a Multivariable Process by Transfer
Function Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
8.3 Stability Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
8.3.1 Smith-McMillan Form . . . . . . . . . . . . . . . . . . . . . . . . . 308
8.3.2 Poles and Zeros of a Transfer Function Matrix . . . . . . 308
8.3.3 Generalized Nyquist Criterion . . . . . . . . . . . . . . . . . . . 309
8.3.4 Characteristic Loci . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
8.3.5 Gershgorin Circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
8.3.6 Niederlinski Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Contents xvii

8.4 Interaction and Decoupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312

8.4.1 Decoupling for a 2  2 System . . . . . . . . . . . . . . . . . . 313
8.4.2 Disturbance Rejection . . . . . . . . . . . . . . . . . . . . . . . . . 317
8.4.3 Singular Value Decomposition . . . . . . . . . . . . . . . . . . 317
8.4.4 Relative Gain Array. . . . . . . . . . . . . . . . . . . . . . . . . . . 318
8.4.5 Gershgorin Circles and Interaction. . . . . . . . . . . . . . . . 325
8.5 Multivariable Robustness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
8.6 Robustness Study of a 2  2 Distillation Column . . . . . . . . . . . . 330
8.6.1 Simplified Decoupling Analysis. . . . . . . . . . . . . . . . . . 331
8.6.2 Ideal Decoupling Analysis . . . . . . . . . . . . . . . . . . . . . . 332
8.6.3 One-Way Decoupling Analysis . . . . . . . . . . . . . . . . . . 332
8.6.4 Comparison of the Three Previous Decouplings . . . . . 333
8.7 Synthesis of a Multivariable Controller. . . . . . . . . . . . . . . . . . . . 333
8.7.1 Controller Tuning by the Largest
Modulus Method . . . . . . . . . . . . . . . . . . . . . . . . . .... 334
8.7.2 Controller Tuning by the Characteristic
Loci Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 334
8.8 Discrete Multivariable Internal Model Control . . . . . . . . . . .... 335
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 337

Part III Discrete-Time Identification

9 Discrete-Time Generalities and Basic Signal Processing . . . . . . . . . . 341
9.1 Fourier Transformation and Signal Processing . . . . . . . . . . . . . . 341
9.1.1 Continuous Fourier Transform . . . . . . . . . . . . . . . . . . . 342
9.1.2 Discrete Fourier Transform . . . . . . . . . . . . . . . . . . . . . 348
9.1.3 Stochastic Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
9.1.4 Stochastic Stationary Signals . . . . . . . . . . . . . . . . . . . . 354
9.1.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
9.2 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
9.2.1 D/A and A/D Conversions. . . . . . . . . . . . . . . . . . . . . . 356
9.2.2 Choice of Sampling Period . . . . . . . . . . . . . . . . . . . . . 358
9.3 Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
9.3.1 First-Order Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
9.3.2 Second-Order Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . 366
9.3.3 Moving Average Filter . . . . . . . . . . . . . . . . . . . . . . . . 367
9.3.4 Fast Transient Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
9.4 Discrete-Time and Finite-Differences Models . . . . . . . . . . . . . . . 368
9.5 Different Discrete Representations of a System. . . . . . . . . . . . . . 370
9.5.1 Discrete Representation: z-Transform . . . . . . . . . . . . . 370
9.5.2 Conversion of a Continuous Description
in Discrete Time . . . . . . . . . . . . . . . . . . . . ......... 392
9.5.3 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 396
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 399
xviii Contents

10 Identification Principles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401

10.1 System Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
10.1.1 System Without Disturbance . . . . . . . . . . . . . . . . . . . . 401
10.1.2 Disturbance Representation . . . . . . . . . . . . . . . . . . . . . 402
10.2 Nonparametric Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
10.2.1 Frequency Identification . . . . . . . . . . . . . . . . . . . . . . . 404
10.2.2 Identification by Correlation Analysis . . . . . . . . . . . . . 405
10.2.3 Spectral Identification . . . . . . . . . . . . . . . . . . . . . . . . . 406
10.3 Parametric Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
10.3.1 Prediction Principles . . . . . . . . . . . . . . . . . . . . . . . . . . 410
10.3.2 One-Step Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
10.3.3 p-Step Predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
11 Models and Methods for Parametric Identification. . . . . . . . . . . . . . 419
11.1 Model Structure for Parametric Identification . . . . . . . . . . . . . . . 419
11.1.1 Linear Models of Transfer Functions . . . . . . . . . . . . . . 419
11.1.2 Models for Estimation in State Space . . . . . . . . . . . . . 431
11.2 Models of Time-Varying Linear Systems . . . . . . . . . . . . . . . . . . 440
11.3 Linearization of Nonlinear Time-Varying Models . . . . . . . . . . . . 440
11.4 Principles of Parametric Estimation . . . . . . . . . . . . . . . . . . . . . . 441
11.4.1 Minimization of Prediction Errors . . . . . . . . . . . . . . . . 441
11.4.2 Linear Regression and Least Squares . . . . . . . . . . . . . 443
11.4.3 Maximum Likelihood Method . . . . . . . . . . . . . . . . . . . 447
11.4.4 Correlation of Prediction Errors with Past Data . . . . . . 450
11.4.5 Instrumental Variable Method . . . . . . . . . . . . . . . . . . . 451
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
12 Parametric Estimation Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
12.1 Linear Regression and Least Squares . . . . . . . . . . . . . . . . . . . . . 455
12.2 Gradient Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
12.2.1 Gradient Method Based on a Priori Error . . . . . . . . . . 458
12.2.2 Gradient Method Based on a Posteriori Error . . . . . . . 462
12.3 Recursive Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
12.3.1 Simple Recursive Least Squares . . . . . . . . . . . . . . . . . 464
12.3.2 Recursive Extended Least Squares . . . . . . . . . . . . . . . 473
12.3.3 Recursive Generalized Least Squares. . . . . . . . . . . . . . 474
12.3.4 Recursive Maximum Likelihood . . . . . . . . . . . . . . . . . 475
12.3.5 Recursive Prediction Error Method . . . . . . . . . . . . . . . 476
12.3.6 Instrumental Variable Method . . . . . . . . . . . . . . . . . . . 479
12.3.7 Output Error Method . . . . . . . . . . . . . . . . . . . . . . . . . . 480
12.4 Algorithm Robustification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
12.5 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Contents xix

12.6 Input Sequences for Identification . . . . . . . . . . . . . . . . . . . . . . . . 484

12.6.1 Pseudo-Random Binary Sequence . . . . . . . . . . . . . . . . 484
12.6.2 Other Sequences for Identification . . . . . . . . . . . . . . . . 487
12.7 Identification Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
12.7.1 Academic Example of a Second-Order System . . . . . . 495
12.7.2 Identification of a Simulated Chemical Reactor . . . . . . 501
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503

Part IV Discrete Time Control

13 Digital Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
13.1 Pole-Placement Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
13.1.1 Influence of Pole Position . . . . . . . . . . . . . . . . . . . . . . 507
13.1.2 Control Synthesis by Pole-Placement . . . . . . . . . . . . . 508
13.1.3 Relation Between Pole-Placement and State
Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
13.1.4 General Pole-Placement Design . . . . . . . . . . . . . . . . . . 519
13.1.5 Digital PID Controller . . . . . . . . . . . . . . . . . . . . . . . . . 528
13.2 Discrete Internal Model Control . . . . . . . . . . . . . . . . . . . . . . . . . 530
13.3 Generalities in Adaptive Control . . . . . . . . . . . . . . . . . . . . . . . . . 535
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 538
14 Optimal Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
14.2 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540
14.3 Variational Method in the Mathematical Framework . . . . . . . . . 543
14.3.1 Variation of the Criterion . . . . . . . . . . . . . . . . . . . . . . 544
14.3.2 Variational Problem Without Constraints, Fixed
Boundaries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 545
14.3.3 Variational Problem with Constraints,
General Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
14.3.4 Hamilton–Jacobi Equation . . . . . . . . . . . . . . . . . . . . . . 549
14.4 Optimal Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
14.4.1 Variational Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 552
14.4.2 Variation of the Criterion . . . . . . . . . . . . . . . . . . . . . . 552
14.4.3 Euler Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
14.4.4 Weierstrass Condition and Hamiltonian
Maximization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
14.4.5 Hamilton–Jacobi Conditions and Equation . . . . . . . . . 558
14.4.6 Maximum Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . 562
14.4.7 Singular Arcs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564
14.4.8 Numerical Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573
14.5 Dynamic Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 578
14.5.1 Classical Dynamic Programming . . . . . . . . . . . . . . . . . 578
14.5.2 Hamilton–Jacobi–Bellman Equation . . . . . . . . . . . . . . 583
xx Contents

14.6 Linear Quadratic Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585

14.6.1 Continuous-Time Linear Quadratic Control . . . . . . . . . 585
14.6.2 Linear Quadratic Gaussian Control . . . . . . . . . . . . . . . 596
14.6.3 Discrete-Time Linear Quadratic Control . . . . . . . . . . . 600
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 606
15 Generalized Predictive Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
15.1 Interest in Generalized Predictive Control . . . . . . . . . . . . . . . . . . 611
15.2 Brief Overview of Predictive Control Evolution . . . . . . . . . . . . . 612
15.3 Simple Generalized Predictive Control . . . . . . . . . . . . . . . . . . . . 613
15.3.1 Theoretical Presentation . . . . . . . . . . . . . . . . . . . . . . . . 613
15.3.2 Numerical Example: Generalized Predictive Control
of a Chemical Reactor . . . . . . . . . . . . . . . . . . . . . . . . . 617
15.3.3 GPC Seen as a Pole-Placement . . . . . . . . . . . . . . . . . . 619
15.4 Generalized Predictive Control with Multiple
Reference Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621
15.4.1 Theoretical Presentation . . . . . . . . . . . . . . . . . . . . . . . . 621
15.4.2 Numerical Example: Generalized Predictive Control
with Performance Model of a Chemical Reactor . . . . . 624
15.5 Partial State Reference Model Control . . . . . . . . . . . . . . . . . . . . 625
15.6 Generalized Predictive Control of a Chemical Reactor . . . . . . . . 626
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 629
16 Model Predictive Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631
16.1 A General View of Model Predictive Control . . . . . . . . . . . . . . . 631
16.2 Linear Model Predictive Control . . . . . . . . . . . . . . . . . . . . . . . . . 636
16.2.1 In the Absence of Constraints . . . . . . . . . . . . . . . . . . . 636
16.2.2 In the Presence of Constraints . . . . . . . . . . . . . . . . . . . 637
16.2.3 Short Description of IDCOM . . . . . . . . . . . . . . . . . . . 637
16.2.4 Dynamic Matrix Control (DMC) . . . . . . . . . . . . . . . . . 638
16.2.5 Quadratic Dynamic Matrix Control (QDMC) . . . . . . . 646
16.2.6 State-Space Formulation of Dynamic
Matrix Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651
16.2.7 State-Space Linear Model Predictive Control as
OBMPC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 653
16.2.8 State-Space Linear Model Predictive Control as
General Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . 657
16.3 Nonlinear Model Predictive Control . . . . . . . . . . . . . . . . . . . . . . 658
16.3.1 Nonlinear Quadratic Dynamic Matrix Control
(NLQDMC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 658
16.3.2 Other Approaches of Nonlinear Model Predictive
Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660
Contents xxi

16.4 Model Predictive Control of a FCC . . . . . . . . . . . . . . . . . . . . . . 664

16.4.1 FCC Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664
16.4.2 FCC Simulation and Control . . . . . . . . . . . . . . . . . . . . 670
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 674

Part V Nonlinear Control

17 Nonlinear Geometric Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681
17.1 Some Linear Notions Useful in Nonlinear Control . . . . . . . . . . . 682
17.1.1 Influence of a Coordinate Change
in Linear Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682
17.1.2 Relative Degree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683
17.1.3 Normal Form and Relative Degree . . . . . . . . . . . . . . . 684
17.1.4 Zero Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
17.1.5 Static State Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . 687
17.1.6 Pole-Placement by Static State Feedback . . . . . . . . . . . 688
17.1.7 Input–Output Pole-Placement . . . . . . . . . . . . . . . . . . . 690
17.2 Monovariable Nonlinear Control. . . . . . . . . . . . . . . . . . . . . . . . . 691
17.2.1 Some Notions of Differential Geometry. . . . . . . . . . . . 691
17.2.2 Relative Degree of a Monovariable Nonlinear
System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693
17.2.3 Frobenius Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 694
17.2.4 Coordinates Change . . . . . . . . . . . . . . . . . . . . . . . . . . . 696
17.2.5 Normal Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 697
17.2.6 Controllability and Observability . . . . . . . . . . . . . . . . . 698
17.2.7 Principle of Feedback Linearization . . . . . . . . . . . . . . . 699
17.2.8 Exact Input-State Linearization for a System
of Relative Degree Equal to n . . . . . . . . . . . . . . . . . . . 700
17.2.9 Input–Output Linearization of a System
with Relative Degree r Lower than or Equal to n . . . . 703
17.2.10 Zero Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 704
17.2.11 Asymptotic Stability . . . . . . . . . . . . . . . . . . . . . . . . . . 706
17.2.12 Tracking of a Reference Trajectory . . . . . . . . . . . . . . . 709
17.2.13 Decoupling with Respect to a Disturbance . . . . . . . . . 710
17.2.14 Case of Nonminimum-Phase Systems . . . . . . . . . . . . . 711
17.2.15 Globally Linearizing Control . . . . . . . . . . . . . . . . . . . . 712
17.3 Multivariable Nonlinear Control . . . . . . . . . . . . . . . . . . . . . . . . . 713
17.3.1 Relative Degree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713
17.3.2 Coordinate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . 714
17.3.3 Normal Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715
17.3.4 Zero Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 716
17.3.5 Exact Linearization by State Feedback
and Diffeomorphism . . . . . . . . . . . . . . . . . . . . . . . . . . 716
17.3.6 Nonlinear Control Perfectly Decoupled by Static
State Feedback. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717
xxii Contents

17.3.7 Obtaining a Relative Degree

by Dynamic Extension . . . . . . . . . . . . . . . . . . . . . . . . 719
17.3.8 Nonlinear Adaptive Control . . . . . . . . . . . . . . . . . . . . . 720
17.4 Applications of Nonlinear Geometric Control . . . . . . . . . . . . . . . 720
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723
18 State Observers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725
18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725
18.1.1 Indirect Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 726
18.1.2 Observer Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 726
18.2 Parameter Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 727
18.3 Statistical Estimation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 728
18.3.1 About the Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 728
18.3.2 Principal Component Analysis . . . . . . . . . . . . . . . . . . . 729
18.3.3 Partial Least Squares . . . . . . . . . . . . . . . . . . . . . . . . . . 731
18.4 Observers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 732
18.4.1 Luenberger Observer . . . . . . . . . . . . . . . . . . . . . . . . . . 732
18.4.2 Linear Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . . . . 736
18.4.3 Extended Kalman Filter (EKF) in Continuous-
Discrete Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 739
18.4.4 Unscented Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . 741
18.4.5 Particle Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 744
18.4.6 Ensemble Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . 750
18.4.7 Globally Linearizing Observer . . . . . . . . . . . . . . . . . . . 752
18.4.8 High-Gain Observer. . . . . . . . . . . . . . . . . . . . . . . . . . . 753
18.4.9 Moving Horizon State Estimation . . . . . . . . . . . . . . . . 757
18.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763

Part VI Applications to Processes

19 Nonlinear Control of Reactors with State Estimation . . . . . . . . . . . . 769
19.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 769
19.2 Chemical Reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 769
19.2.1 Model of the Chemical Reactor . . . . . . . . . . . . . . . . . . 770
19.2.2 Control Problem Setting . . . . . . . . . . . . . . . . . . . . . . . 772
19.2.3 Control Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 773
19.2.4 State Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775
19.2.5 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 777
19.3 Biological Reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 781
19.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 781
19.3.2 Dynamic Model of the Biological Reactor . . . . . . . . . 782
19.3.3 Synthesis of the Nonlinear Control Law . . . . . . . . . . . 784
19.3.4 Simulation Conditions . . . . . . . . . . . . . . . . . . . . . . . . . 787
Contents xxiii

19.3.5 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 788

19.3.6 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 790
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 790
20 Distillation Column Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
20.1 Generalities for Distillation Columns Behaviour . . . . . . . . . . . . . 793
20.2 Dynamic Model of the Distillation Column . . . . . . . . . . . . . . . . 796
20.3 Generalities for Distillation Column Control . . . . . . . . . . . . . . . . 800
20.4 Different Types of Distillation Column Control . . . . . . . . . . . . . 802
20.4.1 Single-Input Single-Output Control . . . . . . . . . . . . . . . 802
20.4.2 Dual Decoupling Control . . . . . . . . . . . . . . . . . . . . . . . 802
20.4.3 The Column as a 5  5 System . . . . . . . . . . . . . . . . . . 804
20.4.4 Linear Digital Control . . . . . . . . . . . . . . . . . . . . . . . . . 807
20.4.5 Model Predictive Control. . . . . . . . . . . . . . . . . . . . . . . 809
20.4.6 Bilinear Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
20.4.7 Nonlinear Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
20.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817
21 Examples and Benchmarks of Typical Processes . . . . . . . . . . . . . . . 821
21.1 Single-Input Single-Output Processes . . . . . . . . . . . . . . . . . . . . . 821
21.1.1 Description by Transfer Functions . . . . . . . . . . . . . . . . 821
21.1.2 Description by a Linear State-Space Model . . . . . . . . . 822
21.1.3 Description by a State-Space Knowledge Model . . . . . 822
21.2 Multivariable Processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 839
21.2.1 Matrices of Continuous Transfer Functions . . . . . . . . . 839
21.2.2 Description by a Linear State-Space Model . . . . . . . . . 842
21.2.3 Description by State-Space Knowledge Models . . . . . . 844
21.2.4 State-Space Knowledge Models for Simulation
and Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 847
21.2.5 Continuous State-Space Models as Benchmarks . .... 849
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 851
Erratum to: Process Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E1
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 853