123,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
DISTRIBUTED MODEL PREDICTIVE CONTROL FOR PLANT-WIDE SYSTEMS
In this book, experienced researchers gave a thorough explanation of distributed model predictive control (DMPC): its basic concepts, technologies, and implementation in plant-wide systems. Known for its error tolerance, high flexibility, and good dynamic performance, DMPC is a popular topic in the control field and is widely applied in many industries.
To efficiently design DMPC systems, readers will be introduced to several categories of coordinated DMPCs, which are suitable for different control requirements, such as network connectivity, error tolerance, performance of entire closed-loop systems, and calculation of speed. Various real-life industrial applications, theoretical results, and algorithms are provided to illustrate key concepts and methods, as well as to provide solutions to optimize the global performance of plant-wide systems.
- Features system partition methods, coordination strategies, performance analysis, and how to design stabilized DMPC under different coordination strategies.
- Presents useful theories and technologies that can be used in many different industrial fields, examples include metallurgical processes and high-speed transport.
- Reflects the authors’ extensive research in the area, providing a wealth of current and contextual information.
Distributed Model Predictive Control for Plant-Wide Systems is an excellent resource for researchers in control theory for large-scale industrial processes. Advanced students of DMPC and control engineers will also find this as a comprehensive reference text.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 400
Veröffentlichungsjahr: 2017
Ähnliche
Table of Contents
Cover
Title Page
Copyright
Preface
About the Authors
Acknowledgement
List of Figures
List of Tables
Chapter 1: Introduction
1.1 Plant-Wide System
1.2 Control System Structure of the Plant-Wide System
1.3 Predictive Control
1.4 Distributed Predictive Control
1.5 About this Book
Part I: Foundation
Chapter 2: Model Predictive Control
2.1 Introduction
2.2 Dynamic Matrix Control
2.3 Predictive Control with the State Space Model
2.4 Dual Mode Predictive Control
2.5 Conclusion
Chapter 3: Control Structure of Distributed MPC
3.1 Introduction
3.2 Centralized MPC
3.3 Single-Layer Distributed MPC
3.4 Hierarchical Distributed MPC
3.5 Example of the Hierarchical DMPC Structure
3.6 Conclusion
Chapter 4: Structure Model and System Decomposition
4.1 Introduction
4.2 System Mathematic Model
4.3 Structure Model and Structure Controllability
4.4 Related Gain Array Decomposition
4.5 Conclusion
Part II: Unconstrained Distributed Predictive Control
Chapter 5: Local Cost Optimization-based Distributed Model Predictive Control
5.1 Introduction
5.2 Local Cost Optimization-based Distributed Predictive Control
5.3 Distributed MPC Strategy Based on Nash Optimality
5.4 Conclusion
Appendix A. QP problem transformation
Appendix B. Proof of Theorem 5.6
Appendix C. Proof of Theorem 5.11
Chapter 6: Cooperative Distributed Predictive Control
6.1 Introduction
6.2 Noniterative Cooperative DMPC
6.3 Distributed Predictive Control based on Pareto Optimality
6.4 Simulation
6.5 Conclusions
Chapter 7: Networked Distributed Predictive Control with Information Structure Constraints
7.1 Introduction
7.2 Noniterative Networked DMPC
7.3 Networked DMPC with Iterative Algorithm
7.4 Conclusion
Appendix A. Proof of Lemma 7.1
Appendix B. Proof of Lemma 7.2
Appendix C. Proof of Lemma 7.3
Appendix D. Proof of Theorem 7.1
Appendix E. Proof of Theorem 7.2
Appendix F. Derivation of the QP problem (7.52)
Part III: Constraint Distributed Predictive Control
Chapter 8: Local Cost Optimization Based Distributed Predictive Control with Constraints
8.1 Introduction
8.2 Problem Description
8.3 Stabilizing Dual Mode Noncooperative DMPC with Input Constraints
8.4 Analysis
8.5 Example
8.6 Conclusion
Chapter 9: Cooperative Distributed Predictive Control with Constraints
9.1 Introduction
9.2 System Description
9.3 Stabilizing Cooperative DMPC with Input Constraints
9.4 Analysis
9.5 Simulation
9.6 Conclusion
Chapter 10: Networked Distributed Predictive Control with Inputs and Information Structure Constraints
10.1 Introduction
10.2 Problem Description
10.3 Constrained N-DMPC
10.4 Analysis
10.5 Formulations Under Other Coordination Strategies
10.6 Simulation Results
10.7 Conclusions
Part IV: Application
Chapter 11: Hot-Rolled Strip Laminar Cooling Process with Distributed Predictive Control
11.1 Introduction
11.2 Laminar Cooling of Hot-rolled Strip
11.3 Control Strategy of HSLC
11.4 Numerical Experiment
11.5 Experimental Results
11.6 Conclusion
Chapter 12: High-Speed Train Control with Distributed Predictive Control
12.1 Introduction
12.2 System Description
12.3 N-DMPC for High-Speed Trains
12.4 Simulation Results
12.5 Conclusion
Chapter 13: Operation Optimization of Multitype Cooling Source System Based on DMPC
13.1 Introduction
13.2 Structure of Joint Cooling System
13.3 Control Strategy of Joint Cooling System
13.4 Results and Analysis of Simulation
13.5 Conclusion
References
Index
End User License Agreement
Pages
xi
xii
xiii
xv
xvii
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
17
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
39
40
41
42
43
44
45
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
65
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
Guide
Cover
Table of Contents
Begin Reading
List of Tables
Chapter 4: Structure Model and System Decomposition
Table 4.1 The meaning of the value for RGA elements
Chapter 5: Local Cost Optimization-based Distributed Model Predictive Control
Table 5.1 Notations used in this chapter
Chapter 6: Cooperative Distributed Predictive Control
Table 6.1 Notations definition
Chapter 7: Networked Distributed Predictive Control with Information Structure Constraints
Table 7.1 Notations used in this chapter
Table 7.2 The plate parameters and the operating points
Chapter 8: Local Cost Optimization Based Distributed Predictive Control with Constraints
Table 8.1 Notations in this chapter
Table 8.2 Parameters of the LCO-DMPC
Table 8.3 State square errors of the closed-loop system under the control of the centralized MPC (CMPC) and the LCO-DMPC
Chapter 9: Cooperative Distributed Predictive Control with Constraints
Table 9.1 Notations in this chapter
Table 9.2 Parameters of C-DMPC
Table 9.3 State square errors of the closed-loop system under the control of the centralized MPC (CMPC), the LCO-DMPC, and the C-DMPC
Chapter 10: Networked Distributed Predictive Control with Inputs and Information Structure Constraints
Table 10.1 Notations in this chapter
Table 10.2 Parameters of the N-DMPC
Table 10.3 State square errors of the closed-loop system under the control of the centralized MPC(CMPC), the LCO-DMPC, and the N-DMPC
Chapter 11: Hot-Rolled Strip Laminar Cooling Process with Distributed Predictive Control
Table 11.1 Thermal and physical properties of the strip
Table 11.2 Computational burdens of DMPC and centralized MPC
Chapter 12: High-Speed Train Control with Distributed Predictive Control
Table 12.1 Coefficients
Chapter 13: Operation Optimization of Multitype Cooling Source System Based on DMPC
Table 13.1 Power consumption functions of refrigerators under air conditioning operation
Table 13.2 Time-of use power price
Table 13.3 Dynamic parameters of refrigerators under air conditioning operation
Table 13.4 Effect of dynamic optimization
Distributed Model Predictive Control For Plant-Wide Systems
Shaoyuan Li and Yi Zheng
Shanghai Jiao Tong University, China
This edition first published 2015
© 2015 John Wiley & Sons Singapore Pte. Ltd.
Registered office:
John Wiley & Sons Singapore Pte. Ltd., 1 Fusionopolis Walk, #07-01 Solaris South Tower, Singapore 138628.
For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com.
All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as expressly permitted by law, without either the prior written permission of the Publisher, or authorization through payment of the appropriate photocopy fee to the Copyright Clearance Center. Requests for permission should be addressed to the Publisher, John Wiley & Sons Singapore Pte. Ltd., 1 Fusionopolis Walk, #07-01 Solaris South Tower, Singapore 138628, tel: 65-66438000, fax: 65-66438008, email: [email protected].
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Library of Congress Cataloging-in-Publication Data applied for.
A catalogue record for this book is available from the British Library.
ISBN: 9781118921562
Preface
There is a class of complex plant-wide systems which are composed of many physically or geographically divided subsystems. Each subsystem interacts with some so-called neighboring subsystems by their states and inputs. The technical target is to achieve a specific global performance of the entire system.
The classical centralized control solution, which could obtain a good global performance, is often impractical for application to a plant-wide system for computational reasons and lack of error tolerance. When the centralized controller fails or a control component fails, the entire system is out of control and the control integrity cannot be guaranteed.
The distributed (or decentralized) framework, where each subsystem is controlled by an independent controller, has the advantages of error-tolerance, less computational effort, and flexibility to system structure. Thus the distributed control framework is usually adopted in this class of system, in spite of the fact that the dynamic performance of centralized framework is better. Thus, how to improve the global performance under distributed control framework is a valuable problem.
Model predictive control (MPC), as a highly practical control technology with high performance, has been successfully applied to various linear and nonlinear systems in the process industries, and is becoming more widespread. The distributed framework of MPC, distributed MPC (DMPC), is also gradually developed with the development of communication network technologies in process industries that allow the control technologies and methodologies to utilize their potentials for improving control.
For the MPC algorithm applied to the plant-wide systems, the system's architectures can be divided as follows:
Centralized MPC, which is a MIMO system architecture;
Decentralized MPC, one controller-one subsystem, but no information exchange between controllers, and
Distributed MPC, which assumes that each subsystem can exchange information with its neighbor's subset of other subsystems.
Since the centralized MPC is forbidden for the large-scale plant-wide system with hundreds (or thousands) of inputs and outputs variables due to its lesser flexibility, weak error tolerance and the large cost of computation, the distributed framework is usually adopted despite its lower global performance. The schematic of distributed MPC is shown in Figure 1, the whole system is composed by many spatial distributed interconnected sub-systems. Each subsystem is controlled by a subsystem-based MPC and these controllers are interconnected by the network.
Figure 1 The schematic of distributed model predictive control
As mentioned before, how to improve the global performance under distributed control framework is a valuable problem. It is exactly true for the DMPC. There are many DMPC strategies and design methods in the literature, all to different ends. We have done extensive research in this topic for more than 10 years, and have proposed some strategies, e.g., the Nash optimization-based DMPC and the impacted region optimization based DMPC, etc. We found that the DMPC is definitely a useful method for large-scale plant-wide systems. Thus, we decided to write this book.
This book systematically introduces different distributed predictive control methods for plant-wide systems, including system decomposition, classification of distributed predictive control, unconstrained distributed predictive control, and the stabilized distributed predictive control with different coordinating strategies for different purposes, as well as the implementation examples of distributed predictive control. The major new contribution of this book is to show how the distributed MPCs can be coordinated efficiently for different control requirements, namely network connectivity, error tolerance, performance of entire closed-loop system, calculation speed, etc., and how to design distributed MPC. The remaining contents of this book are structured into four parts.
In the first part, we recall the main concepts and some fundamental results of the predictive control for discrete-time linear systems. The system structure model and some decomposition methods to present how to divide the entire system into interacting subsystems according to the specific control requirements is also introduced. Our intent is to provide the necessary background knowledge to understand the rest of the book.
The second part introduces the unconstrained distributed MPCs with different coordination strategies. The simplest and most practical local cost optimization based distributed MPC, Nash optimization based distributed MPC, the cooperative distributed MPC that can obtain very good performance of the entire system but each subsystem-based MPC of which requires the information of the whole system, and the networked distributed MPC with information constraints, which is a tradeoff between the two methods mentioned above. For primary readers, the major ideas and characteristics of distributed MPCs are clearly explained in a simple way without constraints.
The third part focuses on introducing the design of the stabilizing distributed MPCs with constraints for the three types of DMPCs: the local cost optimization based DMPC, the cooperative DMPC, and the networked DMPC with information constraint, respectively. The designed DMPCs can guarantee recursive feasibility and the asymptotic stability of the closed-loop system if the initial feasible solution exists.
In the last part, three practical examples are given to illustrate how to implement the introduced distributed MPC into industrial processes, they are the nonlinear networked DMPC for accelerated cooling processes in heavy plate steel mills, the speed train control with unconstrained networked DMPC, and the hierarchical DMPC for load control of a high building with multicooling resources.
In conclusion, this book tries to give a systematic overview of the latest distributed predictive control technologies to readers. We hope this book can help engineers to design control systems in their daily work or in their new projects. In addition, we believe that this book is fit for the graduate students who are pursuing their master or doctor degree in control theory and control engineering. We will be very pleased if this book is of use to you if you are interested in the control of plant-wide systems or predictive control.
Shaoyuan Li
Yi Zheng
About the Authors
Shaoyuan Li (IEEE Senior Member, 2006) is currently Professor and vice president of the School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University. He is also the discipline response person of Control Theory and Control Engineering, the vice director of Key Laboratory of Ministry of Education, the Vice President of the Chinese Association of Automation. He received his PhD in Computer and System Science from Nankai University in 1997. His research interests include predictive control, intelligent adaptive control, fuzzy intelligent control, and its applications. He has published five books and more than 200 papers in journals/conferences. Prof. Li has worked in the area of distributed model predictive control for more than 13 years.
Yi Zheng (IEEE Member, 2010) is Associate Professor at Shanghai Jiao Tong University and currently works in the School of Electronic Information and Electrical Engineering. He received his PhD in Control Theory and Engineering from Shanghai Jiao Tong University. He was with Shanghai Petrochemical Company, Ltd., Shanghai, from 2000 to 2003, GE-Global Research (Shanghai) from 2010 to 2012, and the University of Alberta from 2014 to 2015. His research interests include smart grids, model predictive control, system identification, and their applications to industrial processes. Zheng has worked in the area of distributed model predictive control for nearly 9 years.
Acknowledgement
This work was supported by the National Nature Science Foundation of China (61233004, 61221003, 61374109, 61304078), the National Basic Research Program of China (973 Program-2013CB035500), and partly sponsored by the International Cooperation Program of Shanghai Science and Technology Commission (12230709600), the Higher Education Research Fund for the Doctoral Program of China (20120073130006, 20110073110018), and the China Postdoctoral Science Foundation (2013M540364).
List of Tables
Table 4.1 The meaning of the value for RGA elements
Table 5.1 Notations used in this chapter
Table 6.1 Notations definition
Table 7.1 Notations used in this chapter
Table 7.2 The plate parameters and the operating points
Table 8.1 Notations in this chapter
Table 8.2 Parameters of the LCO-DMPC
Table 8.3 State square errors of the closed-loop system under the control of the centralized MPC (CMPC) and the LCO-DMPC
Table 9.1 Notations in this chapter
Table 9.2 Parameters of C-DMPC
Table 9.3 State square errors of the closed-loop system under the control of the centralized MPC (CMPC), the LCO-DMPC, and the C-DMPC
Table 10.1 Notations in this chapter
Table 10.2 Parameters of the N-DMPC
Table 10.3 State square errors of the closed-loop system under the control of the centralized MPC(CMPC), the LCO-DMPC, and the N-DMPC
Table 11.1 Thermal and physical properties of the strip
Table 11.2 Computational burdens of DMPC and centralized MPC
Table 12.1 Coefficients
Table 13.1 Power consumption functions of refrigerators under air conditioning operation
Table 13.2 Time-of use power price
Table 13.3 Dynamic parameters of refrigerators under air conditioning operation
Table 13.4 Effect of dynamic optimization
Chapter 1Introduction
1.1 Plant-Wide System
There is a class of systems which are composed of many interacted subsystems' industrial fields. Especially with the development of the advanced technology and the increase in the requirement of products, many new distributed processes have appeared, the processes of producing products have become more and more complex, and the scales of industrial processes have become more and more large. The automation structure for this kind of systems has changed from the traditional centralized automation system to a decentralized and centralized automation system, and then to a distributed automation system.
Correspondingly, the control algorithm and control structure for this kind of system change from centralized control and decentralized control to the distributed control system. The distributed control refers to a control system where each subsystem is controlled by an individual controller, and these controllers communicate with other subsystem-based controllers and are coordinated according to the exchanged information for obtaining good global performance or some special common goals. So far, the distributed control, especially the DMPC, has been studied and are still being studied by many scientists, and many theories and algorithms have been developed. We think it is the right time to introduce the distributed control to more students and engineers.
To make it more clear which kind of system is suitable for distributed control, we give some examples as follows.
1.
Wind power generation farm
In a wind turbine power generation farm, as shown in Figure 1.1, wind turbines are spatially distributed. The output wind flow rate of each wind turbine decreases with increasing generated power. It affects the input wind flow rate of the downstream wind turbines, and then their dynamics. In this way, these wind turbines interact with each other. For the automation system, each wind turbine is controlled by an individual controller. And these controllers are connected by a network (fieldbus) and are able to communicate with each other by the network.
Figure 1.1 The wind farm
2.
Multizone building temperature regulation system
Multizone building temperature regulation systems are a class of typical spatially distributed systems, as shown in Figure 1.2, which are composed of many physically interacted subsystems (rooms or zones) labeled as , respectively. The thermal influences between rooms of the same building occur through internal walls (the internal walls' isolation is weak) and/or door openings. A thermal meter and a heater (or air conditioner) are installed in each zone, which is used to measure and adjust the temperature of the multizone building.
Figure 1.2 The multizone building temperature regulation system
3.
Distributed power network
Power networks are large networks consisting of a large number of components. The dynamics of the power network as a whole are the result of interactions between the individual components. The generators produce power that is injected into the network on the one side, while the loads consume power from the network on the other. If we consider each power plant, load, and station as a subsystem, it is a typical distributed system, whose subsystems interacted with each other and controlled separately.
In addition, since the number of players involved in the generation and distribution of power has increased significantly, in the near future, the number of source nodes of the power distribution network will increase even further as large-scale industrial suppliers and small-scale individual household will also start to feed electricity into the network. As a consequence, the structure of the power distribution network will change into a much more decentralized system with many generating sources and distribution agencies (Figure 1.3).
Figure 1.3 Distributed power generation power network
1.2 Control System Structure of the Plant-Wide System
The control structure is a very general concept. It includes how to schedule the controllers, and the inputs/outputs of each controller. The control system structure of the plant-wide system is shown in Figure 1.4, which is a hierarchical structure. The top layer, denoted as layer 4, is a steady economic optimization layer which is used to optimize the key process parameters, e.g., the product quantity, product quality, feeding material quality, etc. Layer 3 is a real-time optimization layer which dynamically optimizes the set-point of the multivariable layers. This layer considers the dynamic economic performance and efficiency. The slow time variation of the process condition is taken into account in this layer. Below this layer is a multivariable layer which coordinates the interaction between each control loop and gives a set-point for the field control loop. The lowest layer, a field control loop layer, which is not drawn in this figure, is used to regulate the process variable, e.g., the temperature, flow rate, or pressure. In some cases, the multivariable takes some work of the field control loop layer when the control problem is complicated. In this structure, with an increase in the layer level, the information to communicate is deduced, and the computing interval is increased.
Figure 1.4 Hierarchical control system for the plant-wide system
Here, we consider the multivariable control layer. For a plant-wide system, there are many inputs and outputs. With the development of a network, communication technology, and fieldbus product, as well as intelligent meters, the control theory for a multivariable system is developed correspondingly. Many advanced control methods appear in the literature works, and the control structure in a multivariable layer changes from the centralized control to the decentralized control, to the distributed control. In addition, recently, the distributed structures for the real-time dynamic optimization layer and steady-state optimization layer have also appeared in the literature works. The real-time optimization layer and multivariable control loop are combined together in some cases. This is out of the scope of discussion in this book. In the following, three types of control structures, centralized control structure, decentralized control structure, and distributed control structure, in a multivariable control layer are specified to show the advantage of the distributed control framework.
1.2.1 Centralized Control
As shown in Figure 1.5, the centralized multivariable controller gets all the information of the plant-wide system, and then calculates the control law of all the inputs together, and sends the control signals to the actuators via the network. The control structure could achieve the best dynamic performance of the overall closed-loop system. However, since there are hundreds (or thousands) of input and output variables in a large-scale plant-wide system, the computational burden is unavoidably high if all control variables are solved together by a centralized controller during a controller period. In addition, since the information of the whole system is necessary when using centralized control, it requires that the network communication must be robust as the communication load is unavoidably high. Furthermore, under this control structure, if one of the subsystems does not work due to some fault, or we proceed with regular maintenance, the multivariable controller must be stopped, and the control of the whole system is interrupted. Thus, this control structure is not sufficiently flexible. Finally, it can be seen that if any part of the controller, actuators, sensors, networks, or control computer has a fault, the multivariable algorithm will lose its effectiveness, which means a low capability of error tolerance, which will not be expected by either the controller engineer or the owner of factories.
Figure 1.5 Centralized control
1.2.2 Decentralized Control and Hierarchical Coordinated Decentralized Control
Considering the less flexibility, the worse error tolerance, the large computational burden, and the heavy network communication load of centralized control, people decompose the centralized controller into many relevant small-scaled controllers, as shown in Figure 1.6. These controllers work with each other independently even when the corresponding controlled subsystems couple with each other. These classes of multivariable controllers have the advantages of simple structure, less computational burden, better error tolerance, good flexibility, and easy designing and implementation. Since the computation for obtaining the control law of the entire system is distributed to many small-scaled controllers, the computation burden of each controller is dramatically decreased. In addition, if several subsystems or controllers do not work due to some fault, the other controllers are still able to work, which means good error tolerance. Furthermore, if there are some new subsystems required to be appended to or deleted from the existing plant-wide system, it needs to do nothing with the existing controllers, which means good structure flexibility.
Figure 1.6 Decentralized control
However, since there is no communication and coordination among decentralized controllers, the controller performance is destroyed if the coupling among subsystems is strong enough. In order to avoid the degradation of the performance of the global system, one method is to enlarge the scale of each local controller, where several strong coupled subsystems are controlled by one local controller. By using this strategy, the performance of the global system could be guaranteed, but the computational burden of each local controller is increased, and the flexibility of overall system is deduced. This strategy bypasses and does not solve the problem of how to improve the global performance when strong interactions exist among the subsystems each of which is controlled by a separated controller.
Unfortunately, in most cases, strong couplings exist in the plant-wide system. Thus, people add a coordinator to coordinate each subsystem-based controller for improving the global performance of the entire plant-wide system, as shown in Figure 1.7, and is called hierarchical coordinating decentralized control. Through different coordinating algorithms, the global performance of the entire system could be significantly improved if strong interactions exist. However, all local controllers should communicate with the coordinator as the global information is necessary for the coordinator. The centralized structure appeared in the coordinator.
Figure 1.7 Hierarchical coordinated decentralized control
1.2.3 Distributed Control
Recently, with the development of computer technologies, fieldbus, network communicating technologies, and smart meter in process industries, which allows the control technologies and methodologies to utilize their potentials for improving control, the distributed control structure has appeared gradually instead of the centralized and decentralized structure for the plant-wide system. As shown in Figure 1.8, the global system is divided into many interacted subsystems, and each subsystem is controlled by a separate controller; these peer controllers communicate with each other through a network for achieving good global performance or a specifically common goal. This kind of control structure has the advantage of a decentralized control structure, e.g., high flexibility and good error tolerance, and the advantage of a centralized control structure, e.g., good global performance. In a distributed control structure, the most important problem is how to design coordinating strategies for different purposes.
Figure 1.8 Distributed control
Figure 1.9 shows the complete structure of an industrial control system structure with distributed control for the plant-wide system. The multivariable layer in Figure 1.4 is substituted by the distributed control which provides the set-points for the field control loops.
Figure 1.9 Distributed control in the hierarchical control system
1.3 Predictive Control
1.3.1 What is Predictive Control
Model predictive control (MPC), also called receding horizon
