1、Basic and AdvancedRegulatory Control:System Design and Application2nd EditionWade04.book Page i Tuesday, July 22, 2008 1:49 PMWade04.book Page ii Tuesday, July 22, 2008 1:49 PMBasic and AdvancedRegulatory Control:System Design and Application2nd EditionBy Harold L. WadeWade04.book Page iii Tuesday,
2、July 22, 2008 1:49 PMNoticeThe information presented in this publication is for the general education of the reader.Because neither the author nor the publisher have any control over the use of the information by thereader, both the author and the publisher disclaim any and all liability of any kind
3、 arising out ofsuch use. The reader is expected to exercise sound professional judgment in using any of the infor-mation presented in a particular application.Additionally, neither the author nor the publisher have investigated or considered the affect ofany patents on the ability of the reader to u
4、se any of the information in a particular application. Thereader is responsible for reviewing any possible patents that may affect any particular use of theinformation presented.Any references to commercial products in the work are cited as examples only. Neither theauthor nor the publisher endorse
5、any referenced commercial product. Any trademarks or trade-names referenced belong to the respective owner of the mark or name. Neither the author nor thepublisher make any representation regarding the availability of any referenced commercial productat any time. The manufacturers instructions on us
6、e of any commercial product must be followed atall times, even if in conflict with the information in this publication.Copyright 2004 by ISA67 Alexander DriveP.O. Box 12277 Research Triangle Park, NC 27709All rights reserved. Printed in the United States of America. 10 9 8 7 6 5 4 ISBN-10: 1-55617-8
7、73-5 ISBN-13: 978-1-55617-873-3 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form orby any means, electronic, mechanical, photocopying, recording or otherwise, without the priorwritten permission of the publisher.Library of Congress Cataloging-in-Public
8、ation DataWade, Harold L.Basic and advanced regulatory control :system design and application/ Harold L. Wade.- 2nd ed.p. cm.ISBN 1-55617-873-5 (pbk.)1. Automatic control. 2. Feedback control systems. I. Title. TJ213.W313 2004629.8-dc222004001322Wade04.book Page iv Tuesday, July 22, 2008 1:49 PMDEDI
9、CATIONTo my mother, who provided an impetus for life-long learning, and to Mary, who has provided love, support,encouragement, and criticism when needed.Wade04.book Page v Tuesday, July 22, 2008 1:49 PMWade04.book Page vi Tuesday, July 22, 2008 1:49 PMABOUT THE AUTHORHarold L. Wade is president of W
10、ade Associates, Inc., a Houston, Texas, firm specializing inprocess control systems consulting and training. He has 45 years experience in applying andinstalling process control systems in such industries as petroleum refining, chemical produc-tion, textiles, and water and waste treatment systems am
11、ong others. He has held technical posi-tions with Honeywell, Foxboro, and Biles new material previously available only in technical journals is presented here. The chapter on model-based control in the first edition has been split into two chapters. Chap-ter 14, devoted primarily to dead-time compen
12、sation, covers Smith predictor control, internalmodel control, and Dahlins algorithm. The other chapter, chapter 15, contains very signifi-cantly expanded coverage of model predictive control. The concluding chapter, which is almost entirely new, covers process control application top-ics that do no
13、t readily fit into any of the other chapters. In addition to cross-limiting control forfired heaters, which was covered in the first edition, these new topics include floating control,techniques for increasing valve rangeability, and time proportioning control. One of the themes of this book is to e
14、mphasize control strategies that are platform indepen-dent. However, since the appearance of the first edition, FOUNDATION Fieldbus (FF),which permits the control strategy to be distributed directly into field devices, has grown inacceptance. The network architecture, communication, and implementati
15、on aspects of FF arebriefly summarized in chapter 5. In this edition, the process control aspects of FF receivegreater coverage. Moreover, the chapters on modifications to feedback control, cascade, ratio,feedforward, and override (chapters 5, 9, 10, 11, and 12) all conclude with an example inwhich
16、that chapters strategy is implemented using FF function blocks. I would like to express gratitude to the many students who, by asking probing questions, haveenabled me to revise and sharpen my presentation and come up with more meaningful exam-ples. In particular, I would like to thank the engineers
17、 at BASFFreeport for encouraging meWade04.book Page xiv Tuesday, July 22, 2008 1:49 PMBASIC AND ADVANCED REGULATORY CONTROL: SYSTEM DESIGN AND APPLICATIONPREFACE XVto develop the controller tuning flow chart, to the staff of the ISA Training Institute for theirsupport during my seminars, and to Adri
18、an and Ivan Susanto in Indonesia and Michael Wangin Taiwan for sponsoring courses and providing me with an opportunity for travel abroad. I would also like to express my thanks to Dr. R. Russell Rhinehart for many helpful commentsand suggestions, to my longtime friend and mentor (and reviewer of thi
19、s book) Greg Shinskey,as well as to John Shaw, Jonas Berge, and Bryan Griffen who have reviewed all or parts of thisbook. And I have special thanks for Susan Colwell, who, through humor and patience, hashelped me endure the arduous task of writing.Wade04.book Page xv Tuesday, July 22, 2008 1:49 PMWa
20、de04.book Page xvi Tuesday, July 22, 2008 1:49 PMCHAPTER 1 1INTRODUCTIONThe term process control implies that there is a process for which there is a desired behaviorand that there is some controlling function that acts to elicit that desired behavior. This broadconcept can embrace everything from s
21、ocietal processes governed by some regulatory controlauthority to automated manufacturing processes. In practically all cases, however, a commonthread is that some measure of the actual process behavior is compared with the desired pro-cess behavior. This feedback action then generates a control pol
22、icy that acts to minimize oreliminate the deviation between desired and actual behavior. We are concerned in this book with a particular segment of automated process controlthatwhich is applied to chemical, refining, pulp and paper, power generation, and similar types ofprocesses. Even within this l
23、imited scope of applications, we will limit the discussion prima-rily to processes that are operated continuously for long periods of time and within a narrowregion of the operating variables. In other words, we exclude such important operating modesas batch processing, start-ups, and grade changes.
24、 Many of the control techniques to be pre-sented here, however, can be adapted to these other modes of operation.For the processes we focus on in this book, the processs behavior is often characterized bymeasured values of such process variables as temperatures, flow rates, pressures, and the like.T
25、he desired behavior, then, is stated to be the set points of those process variables. Until fairlyrecent times, most applications of industrial process control used simple feedback controllersthat regulated the flows, temperatures, and pressures. These controllers required a form ofadjustment called
26、 tuning to match their controlling action to the unique requirements of indi-vidual processes. Occasionally, more advanced forms of control, such as ratio and cascade,could be found; even more rarely one might find a feedforward control loop. As long as mostof the control systems were implemented wi
27、th analog hardware, applications were limited tosimple regulatory control. This was due to the cost of additional components, the additionalinterconnections more advanced control required, the burden of maintenance, and the vulnera-bility to failure of many devices in the control loop. With the adve
28、nt of digital control systems,however, more sophisticated loops became feasible. Advanced regulatory controlwhichincludes the previously mentioned ratio, cascade, and feedforward control as well as additionalforms such as constraint (selector) control and decouplingcould readily be implementedsimply
29、 by configuring software function blocks. With this additional capability, however, a need developed for a systematic approach towardusing it. This is called control strategy design. In order to design a technically successful and1Wade04.book Page 1 Tuesday, July 22, 2008 1:49 PMBASIC AND ADVANCED R
30、EGULATORY CONTROL: SYSTEM DESIGN AND APPLICATION2 CHAPTER 1economically viable control strategy, the control system engineer must be well grounded in thetechniques of feedback control as well as the tools of advanced regulatory control. The requi-site knowledge includes both how to implement and how
31、 to tune. Even before that, however,the control system engineer must be adept at recognizing when to use (and conversely, whennot to use) certain control methods as well as in projecting the expected benefits. Using advanced regulatory control provides many benefits. One of the most important is sim
32、-ply closer control of the process. It will become very clear later in this book that with basicregulatory (i.e., feedback) control, there must be a deviation from set point before controlaction can occur. We will call this the “feedback penalty.” The objective of advanced regula-tory control is for
33、 the control action to be taken by incurring only a minimal feedback penalty.The reduction in feedback penalty may be stated in a variety of ways, such as a reduction ofthe maximum deviation from set point, as a reduction of the standard deviation, or simply as areduction in the amount of off-spec p
34、roduct produced. This reduction in feedback penalty canprovide several forms of economic benefit, such as improvement in product quality, energysavings, increased throughput, or longer equipment life. Process control is but one part of an overall control hierarchy. It extends downward to safetycontr
35、ols and other directly connected process devices and upward to encompass optimizationand even higher levels of business management, such as scheduling, inventory, and asset man-agement (see Figure 1-1). Indeed, corporate profitability may be enhanced more significantlyas a result of these higher-lev
36、el activities than from improved process control per se. However,since each layer of the hierarchy depends upon the proper functioning of the layers beneath it,one of the primary benefits of advanced regulatory control is that it enables the higher levels,such as optimization and enterprise manageme
37、nt and control. SYMBOLSChapter 2 discusses the graphical symbols used in control system documentation. Listedbelow are the mathematical symbols that are used generally throughout the book. Some sym-bols used in this book are used only for the discussion of a particular topic; these symbols aretheref
38、ore defined in that discussion and are not listed here. Chapter 15 uses a unique set ofsymbols that are defined at the beginning of that chapter. The following are the symbols foundthroughout this book:b bias value (manual reset) on proportional-only controller outpute error (deviation between set p
39、oint and process variable)E when capitalized, refers to (Laplace) transform of errorK steady-state gain of first-order lagKCcontroller gain (noninteractive and interactive control algorithms)KDderivative gain (independent gains control algorithm)KIintegral gain (independent gains control algorithm)K
40、Pproportional gain (independent gains control algorithm)Kpprocess gain (change in process variable / change in controller output)m manipulated variable, controller outputWade04.book Page 2 Tuesday, July 22, 2008 1:49 PMBASIC AND ADVANCED REGULATORY CONTROL: SYSTEM DESIGN AND APPLICATIONCHAPTER 1 3M
41、when capitalized, refers to (Laplace) transform of manipulated variablePB proportional bandPI control algorithm with proportional and integral modesPID control algorithm with proportional, integral, and derivative modesPV process variable (see also symbol x)SP set point (see also symbol xSP)TDderiva
42、tive time (noninteractive and interactive control algorithms)TIintegral time (minutes/repeat) (noninteractive and interactive control algorithms)x process variable (see also symbol PV)xSPset point (see also symbol SP)u disturbance variable derivative gain when a derivative filter is used with nonint
43、eractive or interactive control algorithm) dead time first-order lag-time constantFigure 1-1. Overall Process Control and Information System Hierarchy237,0,=$7,21352 if the input signal is decreasing, the outputsignal will be negative.If you are thoroughly comfortable with these intuitive concepts,
44、then you are as familiar withderivatives as you need to be for this book. IntegralsThe mathematician is often interested in the integral of a function. Quite simply, if the functionis represented in graphical form, the integral is simply the “area under the curve.” The controlengineer is also intere
45、sted in the integral. For example, if the function in question is the errorFigure 2-1. Practical Concepts of Differential CalculusGGWGGWGGWGGW,1387 287387,13876,*1$/DEFWade04.book Page 7 Tuesday, July 22, 2008 1:49 PMBASIC AND ADVANCED REGULATORY CONTROL: SYSTEM DESIGN AND APPLICATION8 CHAPTER 2sign
46、al of a control loop, the control engineer might be interested in the area under the curvesince the loop was last switched from manual to automatic or since the last load upset.The mathematician uses analytical forms to express the integral of certain functions. The fol-lowing table gives a list of
47、these functions and their integrals:As with derivatives, the control engineer may have a component (“black box”) inside a control-ler or a line of code inside a computer program that will determine the integral of an incomingsignal. As before, suppose you have such a black box. Suppose as well that
48、the output of theblack box is connected to a meter, specifically a center-zero, bi-deflectional meter as shown inFigure 2-2.To test your intuitive understanding of integrals, suppose you can see the output meter readingbut you do not know the form of the input signal. Cover up the left-hand side of
49、Figure 2-2,observe the meter responses described on the right, and predict what the input signal must be.Then check yourself against the response shown on the left-hand side.Remember, the output signal from the integral unit “black box” will be stationary if and only ifthe input signal is zero. The actual value of the output signal does not matter. If the input signalis positive, the output signal will be increasing; if the input signal is negative, the output signalwill be decreasing. More specifical
