\)
Advanced
PID Control
Karl J. Astrom Tore Hdgglund
Department  of Automatic  Control
Lund  Institute  of Technology
Lund  Uniuersity
z\
Jlsnr
,?
Systems,
ISA-The 
Instrumentalion, 
and Automation Society
Copyright @ 2006 by 
ISA -  Instrumentation, Systems, and Automation Society
67 Alexander Drive
P.O. Box 12277
Research Triangle Park, NC 27709
All rights reserved.
Printed in the United States of America.
1 0 9 8 7 6 5 4 3 2
ISBN  r-55617-942-r
No part of this work may be reproduced, stored in a retrieval system, or transmitted in
any form or by any means, electronic, mechanical, photocopying, recording or
otherwise, without  the prior written  permission of the publisher.
Notice
The information presented in this publication is for the general education of the
reader. Because neither the author nor the publisher has any control over the use of
the information by the reader, both the author and the publisher disclaim any and all
liability of any kind arising out of such use. The reader is expected to exercise sound
professional judgment in using any of the information presented in a particular
application.
Additionally, neither the author nor the publisher have investigated or considered
the effect of any patents on the ability  of the reader to use any of the information in a
particular application. The reader is responsible for reviewing any possible patents
that  may affect any particular  use of the information presented.
Any references to commercial products in the work are cited as examples only.
Neither the author nor the publisher endorses any referenced commercial product. Any
trademarks or tradenames referenced belong to the respective owner of the mark or
name. Neither the author nor the publisher makes any representation regarding the
availability  of any referenced commercial product at any time. The manufacturer's
instructions on use of any commercial product must be followed at all times, even if in
conflict with  the information in this publication.
Library  of Congress  Cataloging-in-Publication  Data
Astrtjm, Karl  J. (Karl Johan), lg34-
Advanced PID control / Karl Johan Astrom and Tore Hagglund.
p. cm.
Includes bibliographical references and index.
ISBN 1-55617-942-1 (pbk.)
1. PID controllers. I. Hrigglund, Tore. II. Title.
TJ223.P55A85 2006
629.8'3--dc22
20050r4664
Preface
The PID  controller  is the  most common solution to practical  control problems.
Although  controllers  with  proportional  and integral  action have been used from
the  time  when  windmills  and  steam engines were the  dominant  technologies,
the  current  form  of the  PID  controller  emerged with  the  pneumatic  controllers
in  the  1930s. The  controllers  have been implemented  in  many  different  ways
using mechanical, pneumatic,  electronic, and computer technolory. The  devel-
opment accelerated when the microprocessor implementations  appeared in  the
1980s. One  reason was  that  the  computer  implementations  made  it  possible
to add features like  auto-tuning  and  diagnostics, which  are very  beneficial for
users. From  an  engineering  perspective, it  is  particularly  interesting  to  ana-
lyze  what  happened  at  the  technology shifts,  when  some important  features
were rediscovered and others were added.
This  book has  grown  out  of  more  than  25  years  of  development  of  auto-
tuners  for  PID  controllers  in  close collaboration  with  industry.  Through  this
work,  we have been exposed to  a large number  of real industrial  control prob-
lems. We have benefited much from  participating  in  development, commission-
ing,  and troubleshooting  of industrial  controllers. The  practical  work  has also
inspired  research.
This  book is the  last  part  of a trilogy.  The  first  book, Automatic  Tfu.ning of
PID  Controllers,  1988, which  had  6 chapters, gave a  short  description  of our
early experiences with  development of relay auto-tuners. The second book, PID
Controllers:  Theory, Design,, and  Tfu.ning, 1995, which  has 7 chapters,  grew  out
of the need for a broader coverage of many  aspects of PID  control. In  particular,
it  reviews  many  design methods  for  PID  controllers  that  we  investigated  in
connection with  our  work  on auto-tuners.
The  knowledge about PID  control  in  1995 still  was not  satisfactory for  de-
sign of auto-tuners.  One drawback  was that  the  user  had  to  provide  the  con-
troller  with  design  choices. It  is  particularly  difficult  for  a  user  to  assess if
dynamics is  dead-time or lag  dominated.  This  question stimulated  further  re-
search. Because of the drastic increase in  computing power, it  was also possible
to use design algorithms  that  require  more computations.
Thning  and design of PID  controllers have traditionally  been based on spe-
cial techniques. Robust control was a major  development of control theory that
matured  in  the  late  1990s, resulting  in  powerful  design methods based on ro-
bust  loop shaping. This  stimulated  us to initiate  a research program  to  adapt
Preface
these methods  to  PID  control.  At  the  same time,  it  seemed natural  to  bring
PID  control  closer to  the  mainstream  ideas in  control.  When  working  with  in-
dustrial  auto-tuners,  we  also  saw  a  great  need  to  include  diagnostics  in  the
controller,  because it  is  no  use to  tune  a  controller  if  the  process has  severe
malfunctions.  The  present  book, Aduanced  PID  Control,  is  the  result  of this
effort.
With  a total  of 13 chapters,  this  new book substantially  expands on some of
the  topics  covered in  the  previous  versions  and  provides  several  new  chapters
that  deal with  controller  design, feedforward  design, replacement  of the  Ziegler-
Nichols tuning  rules, predictive control, loop and performanee asqeqqrnnnr ?-i
interaction.  At  this  point  in  our  book  trilogy,  we  assume that  the  reader  rb
highly  familiar  with  control  theory.
Our  research  has  given  a  deeper understanding  of  the  trade-offs  between
load  disturbance  attenuation,  injection  of  measurement  noise, and  set-point
response. We have also been able to answer questions like:  Should a controller
be tuned  for  response to  load  disturbances  or  set points?  What  information 
is
required  to  design  a  PID  controller?  When  can  derivative  action  give  signif-
icant  improvements?  When  are  more  complicated  controllers  justified?  When
is  it  justified  to  develop more  accurate process models? With  the  knowledge
developed, it  is now possible to design auto-tuners  that  can make these assess-
ments  autonomously. In  addition,  we have developed new  simple  methods for
designing PID  controllers.
As  an  example  of  the  insight  gained  we  can  mention  that  control  theory
tells  that  it  is  not  necessary to  make  a  compromise between  tuning  for  load
disturbance  response and  set-point response. Both  requirements  can be satis-
fied by using  a controller  with  two  degrees of freedom, which  combines feedback
and  feedforward.  The  feedback gains  should  be chosen to satisfy  requirements
on disturbance  attenuation  and robustness. The  desired response to  set-point
changes can then be obtained by proper use of feedforward. Set-point weighting
is a simple form  of feedforward for PID  control. In  some cases, it  is justified  to
use more  elaborate feedforward.  For  this  reason, we have included  a chapter
on controller  design and another  chapter  on feedforward  in  the  new book.
The robustness analysis also shows the  advantage of having  low  controller
gain  at  high  frequency,  high  frequency  roll-off.  This  can  be  accomplished  by
flltering  the  process output  by  a  second order  filter.  Based on the  insight  ob-
tained,  we  recommend extended use of set-point weighting  or more  advanced
feedforward.  We  also recommend  that  the  process output  is  fiItered  using  a
second order filter.
We  would  like  to  thank  many  people who  have  given  knowledge,  insight,
and inspiration.  Our  interest  in  PID  control  was inspired  by Axel  Westrenius
and Mike  Somerville of Eurotherm  in  the  early  1980s. We have learned much
from  working  with  students;  particular 
thanks  are  due to  Lars  Gdran  Elfgren
(Eurotherm),  Giiran  Grtinhammar  (LTH),  Ari  Ingimundarson  (UPC),  Oskar
Nordin  (Volvo), Helene Panagopoulos (Volvo), Per Persson (Volvo), Mikael  Pe-
tersson (ABB),  Ola  Slattke  (ABB),  and Anders  Wall6n  (Ericsson Mobile  Plat-
forms), who continue to give us valuable insight  even if  they  are now pursuing
careers in  industry.
We  are  very  grateful  to  Sune  Larsson  and  Lars  Beath,  formerly  of  NAF
vi
Controls,  with  whom  we  developed the  first  industrial  relay  auto-tuner.  The
company  NAF  Controls  was  merged  several  times  and  is  now  part  of  ABB,
where  we have enjoyed interactions  with  Gdran Arinder,  Alf  Isaksson, Per Erik
Mod6n, Lars  Pernebo, and Thomas Vonheim. We have shared the joy  and chal-
lenges in  moving  techniques  for  auto-tuning  and  diagnostics  into  a wide  range
of industrial  products. Many  stimulating  discussions with  our colleagues Anton
Cervin  (LTH),  Sebastian Dormido  (UNED),  Gty  Dumont  (UBC),  Chang Chieh
Hang  (NUS),  Karl  Henrik  Johansson  (KTH),  Birgitta  Kristiansson  (CTH),
Bengt  Lennartsson  (CTH),  Manfred  Morari  (ETH),  Dale Seborg (UCSB),  Sig-
(LTH),  and Karl-Erik  Arz6n (LTH)
urd  Skogestad (NTNU),  Bjiirn  Wittenmark 
from  academia  are  also highly  appreciated.
Our  friends  in  industry  Bill  Bialkowski,  Terry  Blevins,  Greg  McMillan,
and  Willy  Wojsznis  from  Emerson,  Edgar  Bristol,  Sigifredo  Niflo,  and  Greg
Shinskey from  Foxboro, Bdrje Eriksson  (M-real),  Krister  Forsman  (Perstorp),
Ken  Goff (Leeds and Northrup),  Niklas  Karlsson  (Evolution  Robotics), Joseph
Lu  (Honeywell),  Tor  Steinar  Schei (Cybernetica),  Stefan  Rtrnnbiick  (Optima-
tion),  have  generously  shared  their  knowledge  and  insight  with  us.  We  are
particularly  grateful  to  Peter  Hansen,  formerly  of Foxboro, who  read  the  com-
plete  manuscript  and  gave us very  good feedback.
We are very  grateful  to Leif  Andersson  who made the  layout  of the  text  and
gave much  assistance with  TbX, Agneta  T\rszyriski  who translated  much  of the
text  to  I4rIbX, and  Eva  Dagnegfird  who  drew  several  of the  figures.
Finally,  we  would  like  to  thank  the  Swedish  Research Council  (VR),  the
Swedish Agency for  Innovation  Systems (VINNOVA),  and the  Swedish Foun-
dation for Strategic Research (SSF) who have supported our research for many
years.
Kanr, JoHaN AsrnOnt
Tonn HAccluNn
Department  of Automatic  Control
Lund  Institute  of Technology
Box 118, SE-22I00  Lund,  Sweden
karl-j  ohan. astrom@control . Ith.  se
tore  . hagglund@control . lth.  se
vu
Contents
1 .
2.
3.
4.
Technology Changes and Knowledge Thansfer
Summary
.
Introduction  .
Static Models
Feature-Based Models
Tlpical  Process Models
.  :  .  :  .  :  .
Introductio" 
Feedback
Simple Forms  of Feedback
Introduction
1.1 
1.2 
1.3 
1.4  How  the  PID  Controller  Developed
1.5 
1.6  Outline  of the  Contents of the  Book  .
I.7 
1.8  Notes and References
Process  Models 
2.1 
2.2 
2.3  Dynamic Models
2.4 
2.5 
2.6  Models for  Disturbances  .
2.7 
How  to Obtain  the  Models
2.8  Model  Reduction
2.9 
2.t0  Notes and References
PID  Control 
3.1 
3.2 
3.3 
3.4 
3.5 
3.6  When Can PID  Control Be Used?
3.7 
3.8  Notes and References
Controller  Design
4.1 
4.2 
4.3 
4.4 
Introduction  .
A  Rich Variety  of Control  Problems
Feedback Fundamentals  .
Stability
Introduction  .
The PID  Controller  .
Filtering  the  Derivative
Set-Point Weighting
Integrator  Windup
Summary
.
Summary
1
1
2
3
5
7
8
10
10
t2
t2
13
T4
23
28
44
47
56
6 1
62
64
64
64
73
74
76
87
92
93
95
95
96
96
r02
ix
Contents
6.
7.
Classical Specifications
Introduction  .
Improved  Set-Point Response .
Set-Point Weighting
Fast  Set-Point Response .
Summary
Notes and References
Closed-Loop Poles and Zeros
The  Sensitivity  Functions  .
Robustness to Process Variations
Introduction  .
Ziegler-Nichols and Related Methods
Rule-Based Empirical  T\rning.
Pole Placement
Lambda  T\rning  .
Algebraic  Design
4.5 
4.6 
4.7 
4.8  Quantifying  the  Requirements
4.9 
4.10  Summary
A.LI  Notes and References
Feedforward  Design
5.1 
5.2 
5.3 
5.4  Neutral  Feedforward  .
5.5 
5.6  DisturbanceAttenuation
5.7 
5.8 
PID  Design
6.1 
6.2 
6.3 
6.4 
6.5 
6.6 
6.7  Optimization  Methods
Robust Loop Shaping
6.8 
6.9 
Summary
6.10  Notes and References
A  Ziegler-Nichols  Replacement
7.1 
7.2 
7.3 
7.4 
7.5 
7.6 
7.7 
7.8  Measurement  Noise and Filtering
7.9  Detuning .
7.10  Summary
7.1I  Notes and References
Predictive  Control
Introduction .
8.1 
8.2  The Smith Predictor
8.3  Analysis of Smith Predictor Control .
8.4  The PPI Controller .
8.5  Predictors for Integrating Processes
8.6  Model Predictive Control
8.7  Summarv
Introduction  .
The Test Batch
PI  Control  .
PID  Control
Frequency Response Methods
PID  Control  Based on Second-Order Model
Comparison of the  Methods  .
109
1 1 1
1 1 8
122
128
136
137
139
139
139
I45
I46
150
154
156
157
158
158
159
169
t74
186
189
196
206
221
222
225
225
226
228
230
238
242
247
251
253
262
265
266
266
267
271
279
283
285
291
Introduction  .
Integrated  Ttrning  and Diagnosis
Introduction  .
Interaction  of Simple  Loops .
.
Notes and References
Supervision  of Adaptive  Controllers
Iterative  Feedback Ttrning
Commercial  Products
Summary
8.8 
Automatic  Thning  and  Adaptation
Introduction  .
9.1 
Process Knowledge  .
9.2 
Adaptive  Techniques .
9.3 
9.4  Model-Based Methods
9.5  Rule-Based Methods
9.6 
9.7 
9.8 
9.9 
9.10  Notes and References
Loop  and  Performance  Assessment
10.1 
I0.2  Valves
10.3  Loop Assessment
10.4  Performance Assessment
10.5 
10.6  Summary
10.7  Notes and References
Interaction
11.1 
LI.2 
11.3  Decoupling
1I.4  Parallel  Systems
11.5  Summary
11.6  Notes and References
Control  Paradigms
12.1 
Introduction 
I2.2  Bottom-Up  and Top-Down Approaches
12.3  Repetitive  Control
12.4  Cascade Control
I2.5  Mid-Range  and Split-Range  Control  .
12.6  Nonlinear  Elements
I2.7  Neural  Network  Control  .
I2.8  Fuzzy  Control  .
I2.9  System Structuring
12.10  Summary
t2.II  Notes and References
Implementation
13.1 
I3.2  Analog  Implementations
13.3  Computer  Implementations  .
13.4  Velocity  Algorithms
13.5  OperationalAspects
13.6  Controller  Outputs  .
L3.7  Summarv
Introduction  .
10.
l l .
12.
13.
Contents
292
293
293
294
295
298
302
304
3 1 3
3 1 6
327
327
329
329
329
334
336
344
345
346
347
347
347
354
360
364
365
366
366
367
368
373
378
3 8 1
389
392
398
404
406
407
407
408
4r2
420
423
427
43r
xi