多变量模型预测控制技术应用研究
|
摘要
模型预测控制算法(MPC)自上世纪70年代末问世以来，经过二十多年来的深入研究和发展，其理论和方法日臻完善。如今预测控制策略被认为是工业过程最有前途的先进控制算法，已被许多大公司嵌入到控制系统软件中，在一些大型的过程控制系统中得到应用。但预测控制在许多中小装置的回路还未得到广泛应用，其原因在于该算法与常规的PID控制相比对工程师的专业知识要求较高，计算量也较大。为此，针对过程控制系统的特点，本文研究和开发了低成本预测控制仿真与控制的一体化预测控制软件包并进行了应用研究。
论文主要内容为：
1、介绍了铜冶炼烟气置备硫酸的工艺及对象特性，获取了对象模型。
2、介绍了作为模型预测控制算法之一的动态矩阵控制的原理；重点分析了多输入-多输出动态矩阵控制算法的多步预测原理。
3、以铜冶炼烟气制备硫酸工艺为研究对象，在获得其干吸串酸过程对象模型的基础上，进行了基于多输入-多输出动态矩阵预测控制软件的应用仿真研究，即对控制器中各参数变化、干扰对系统性能的影响做了大量仿真实验研究，表明了其预测控制算法的有效性和适用性。
4、基于Windows操作系统，用VC++语言开发了通用的多输入-多输出动态矩阵控制算法仿真与控制一体化预测控制软件。
5、本论文的研究可为解决工业过程中的多变量控制问题提供一种有效途径，并为进一步如何在工控软件上实现其先进控制算法奠定了基础。
6、最后总结全文并展望了MPC算法的发展方向及未来前景。
Model Predictive Control has gained rapid development both in theory and in industrial applications ever since it came into being in 1970's. Over the past decades, MPC has established itself in industry as an important form of andvanced control. Predictive Control is regarded as the most promising and advanced control algorithm. Many big companies have embedded it into control system. However, predictive control scarcely applied in many medium-sized and pint-sized control systems. Because the predictive algorithm needs much more skilled engineers than that of PID controllers, moreover, predictive algorithm needs more calculations. So aiming at the characters of the process control system, the paper researches and develops the low-cost predictive control software for simulations and real-time controls and proceeds the application study.
Its main contents are as follows:
1. It introduces the sulfuric acid craft of burned copper smoke system and the object character. At the same time, it obtaines the object model.
2. As one MPC algorithm, dynamic matrix control is discussed in detail. And the principals of multi-step predictive dynamic matrix control and MIMO dynamic matrix control are talked in this part.
3. With the burned copper smoke system has the sulfuric acid craft for
research object. Based on acquiring the process object model of the sulfuric acid craft, we proceed the application of the software imitates true research according to MIMO dynamic matrix control. We did much experiment research of the parameter variety and the interference did to the influence of the system function. And it expresses the usefulness of its estimate control calculate
way with the applicability.
4. We uses VC++ to develop the MIMO dynamic matrix control software. The software can in general usely merge simulation and controlling together .
5. The research of this thesis provides a kind of effective way for resolving multiple parameter control in industry process. And it offers the foundation for how to realize the MPC algorithm in industry control software.
6. At last it summaries the text and outlooks the development direction and future foreground of the MPC algorithm.
论文主要内容为：
1、介绍了铜冶炼烟气置备硫酸的工艺及对象特性，获取了对象模型。
2、介绍了作为模型预测控制算法之一的动态矩阵控制的原理；重点分析了多输入-多输出动态矩阵控制算法的多步预测原理。
3、以铜冶炼烟气制备硫酸工艺为研究对象，在获得其干吸串酸过程对象模型的基础上，进行了基于多输入-多输出动态矩阵预测控制软件的应用仿真研究，即对控制器中各参数变化、干扰对系统性能的影响做了大量仿真实验研究，表明了其预测控制算法的有效性和适用性。
4、基于Windows操作系统，用VC++语言开发了通用的多输入-多输出动态矩阵控制算法仿真与控制一体化预测控制软件。
5、本论文的研究可为解决工业过程中的多变量控制问题提供一种有效途径，并为进一步如何在工控软件上实现其先进控制算法奠定了基础。
6、最后总结全文并展望了MPC算法的发展方向及未来前景。
Model Predictive Control has gained rapid development both in theory and in industrial applications ever since it came into being in 1970's. Over the past decades, MPC has established itself in industry as an important form of andvanced control. Predictive Control is regarded as the most promising and advanced control algorithm. Many big companies have embedded it into control system. However, predictive control scarcely applied in many medium-sized and pint-sized control systems. Because the predictive algorithm needs much more skilled engineers than that of PID controllers, moreover, predictive algorithm needs more calculations. So aiming at the characters of the process control system, the paper researches and develops the low-cost predictive control software for simulations and real-time controls and proceeds the application study.
Its main contents are as follows:
1. It introduces the sulfuric acid craft of burned copper smoke system and the object character. At the same time, it obtaines the object model.
2. As one MPC algorithm, dynamic matrix control is discussed in detail. And the principals of multi-step predictive dynamic matrix control and MIMO dynamic matrix control are talked in this part.
3. With the burned copper smoke system has the sulfuric acid craft for
research object. Based on acquiring the process object model of the sulfuric acid craft, we proceed the application of the software imitates true research according to MIMO dynamic matrix control. We did much experiment research of the parameter variety and the interference did to the influence of the system function. And it expresses the usefulness of its estimate control calculate
way with the applicability.
4. We uses VC++ to develop the MIMO dynamic matrix control software. The software can in general usely merge simulation and controlling together .
5. The research of this thesis provides a kind of effective way for resolving multiple parameter control in industry process. And it offers the foundation for how to realize the MPC algorithm in industry control software.
6. At last it summaries the text and outlooks the development direction and future foreground of the MPC algorithm.
引文
[1] 舒迪前，预测控制系统及其应用，机械工业出版社，1996，3
[2] 王树青等，先进控制技术及应用，化学工业出版社，2001，7
[3] 席裕庚，预测控制，国防工业出版社，1993，6
[4] 管善功，模型预测控制算法在应用中的研究和软件包的实现，华北电力大学，2006
[5] 史红梅，冯丽辉，钟伟红等，基于模型的预测控制算法研究，昆明理工大学学报，2002．12
[6] 张宏军等，Visual C++6.0编程案例精解，电子工业出版社，2005．1
[7] 龙腾科技，Visual C++6.0循序渐进教程，北京希望电子出版社，2005．2
[8] 刘永清，硕士学位论文，昆明理工大学，2006
[9] 金以慧，生产过程的先进控制，化工自动化及仪表，1997，24(4)
[10] 金晓明，王树青，荣冈，模型预测控制及其工业应用，化工自动化及仪表，1999，26(5)
[11] 李言俊，张科．系统辨识理论及应用．国防工业出版社，2003．5．
[12] 姜捷李等，OPC规范简介(二)．自动化仪表，2002，23(6)：67-69
[13] 李曼珍，工业过程对象的实用辨识，仪器仪表用户，2005，12(4)：95-96
[14] 王修中等，二阶加滞后连续模型的直接辨识，自动化学报，2001，27(5)：728-730
[15] 戴懿纯，动态矩阵预测控制软件：[硕士学位论文]．广西大学，2004．5
[16] 王树青，金晓明．先进控制技术应用实例，化学工业出版社，2005．2
[17] 朱豫才，过程控制的多变量系统辨识，国防科技大学出版社，2005．8
[18] Nilos. Drakos, An Overview of Industrial Model Predictive Control Technology, Computer Based Learning Unit, University of Leeds.
[19] Ashraf Ai-Ghazzawi, etc. On-line Tuning Strategy for Model Predictive Controllers. Journal of Process Control, 2001, (11): 265-284
[20] Mark J. Willis & Ming T. Tham, Advanced Process Control, Chemical and process engineering
[21] Stephen Wright, Jim Rawlings, Matt Tenny, Optimization Problems in Model Predictive Control, Gabriele Pannocchia University of Wisconsin-Madison, FoCM' 02, Minneapolis, August 6, 2002
[22] Dae Yyook Yang, Model Predictive Control, Dept of Chemical and Biological Engineering of Korea University, Spring 2002
[23] The Cybosoft Company, Comparison of Control Methods
[24] The Cybosoft Company, Current Control Technology
[25] Magnus Akerblad and Anders Hansson, Efficient Solution of Second Order Cone Program For Model Predictive Control, 15th IFAC Triennial World Congress, 2002
[26] DMCplus-Dynamic matrix control software, Aspentech Company;
[27] B. Wayne Bequett, Model Predictive Control, 2000, 9
[28] Seborg, D.A.. A perspective on advanced strategies for process control. Modeling, Identification and Control, 1994, 15(3): 179-189
[29] D.Q. Mayne et al. Constrained model predictive control: Stability and optimality. Automati ca2000,36:789-814
[30] Lee J H ,Yu ZH. Tuning of Model Predictive Controller for Robust Performance. Computer Chem Eng, 1994, 18(1): 15-37
[31] 闫志恒，过程控制系统实时预测控制软件开发与应用，硕士学位论文，上海交通大学，2003
[32] Morari. Model predictive control: past, present and future. Computers & Chemical Engineering, 1999(23):667-682
[33] j. Richalet. Industrial application of model based predictive control. Automatica, 1993,29:1251-1274
[34] Seung Cheol Jeong and PooGyeon Park. Constrained MPC Algorithm for Uncertain Time-VaryingSystems With State-Delay. IEEE Transaction On Automation Control. 2005, 50(2): 257-263
[35] 孙优贤，工业过程控制技术，化学工业出版社，2006
[36] 康红星，预测控制在锅炉燃烧系统中的应用，湖北电力，2001
[37] 梁春燕，谢剑英，预测控制中的若干问题研究，自动化与仪器仪表，1999(4)
[38] 刘妹妹，陈际达，廖力清，给予预测控制理论的位置伺服系统设计，电气传 1997(2)
[39] 闫志恒，李少远，基于PC机平台的预测控制系统仿真软件开发与实现，系统仿真学报，2002
[40] 席裕庚，耿晓军，陈虹(2000)，预测控制性能研究的新进展，控制理论与应用，Vol．17，No．4
[41] 邵雪飞，一类计算机新型控制策略-预测控制的改进研究及仿真设计，1999，5
[42] 徐立鸿，预测控制方法及其应用，上海，同济大学电气系
[43] 濮强国，动态矩阵控制方法的仿真研究，苏州城建环保学院学报．2002
[44] 陈荣保，李劲松，基于动态矩阵算法的液位控制系统，自动化仪表．2001
[45] 孙德敏等，工业过程先进控制及优化软件产业，自动化博览，2003．2
[2] 王树青等，先进控制技术及应用，化学工业出版社，2001，7
[3] 席裕庚，预测控制，国防工业出版社，1993，6
[4] 管善功，模型预测控制算法在应用中的研究和软件包的实现，华北电力大学，2006
[5] 史红梅，冯丽辉，钟伟红等，基于模型的预测控制算法研究，昆明理工大学学报，2002．12
[6] 张宏军等，Visual C++6.0编程案例精解，电子工业出版社，2005．1
[7] 龙腾科技，Visual C++6.0循序渐进教程，北京希望电子出版社，2005．2
[8] 刘永清，硕士学位论文，昆明理工大学，2006
[9] 金以慧，生产过程的先进控制，化工自动化及仪表，1997，24(4)
[10] 金晓明，王树青，荣冈，模型预测控制及其工业应用，化工自动化及仪表，1999，26(5)
[11] 李言俊，张科．系统辨识理论及应用．国防工业出版社，2003．5．
[12] 姜捷李等，OPC规范简介(二)．自动化仪表，2002，23(6)：67-69
[13] 李曼珍，工业过程对象的实用辨识，仪器仪表用户，2005，12(4)：95-96
[14] 王修中等，二阶加滞后连续模型的直接辨识，自动化学报，2001，27(5)：728-730
[15] 戴懿纯，动态矩阵预测控制软件：[硕士学位论文]．广西大学，2004．5
[16] 王树青，金晓明．先进控制技术应用实例，化学工业出版社，2005．2
[17] 朱豫才，过程控制的多变量系统辨识，国防科技大学出版社，2005．8
[18] Nilos. Drakos, An Overview of Industrial Model Predictive Control Technology, Computer Based Learning Unit, University of Leeds.
[19] Ashraf Ai-Ghazzawi, etc. On-line Tuning Strategy for Model Predictive Controllers. Journal of Process Control, 2001, (11): 265-284
[20] Mark J. Willis & Ming T. Tham, Advanced Process Control, Chemical and process engineering
[21] Stephen Wright, Jim Rawlings, Matt Tenny, Optimization Problems in Model Predictive Control, Gabriele Pannocchia University of Wisconsin-Madison, FoCM' 02, Minneapolis, August 6, 2002
[22] Dae Yyook Yang, Model Predictive Control, Dept of Chemical and Biological Engineering of Korea University, Spring 2002
[23] The Cybosoft Company, Comparison of Control Methods
[24] The Cybosoft Company, Current Control Technology
[25] Magnus Akerblad and Anders Hansson, Efficient Solution of Second Order Cone Program For Model Predictive Control, 15th IFAC Triennial World Congress, 2002
[26] DMCplus-Dynamic matrix control software, Aspentech Company;
[27] B. Wayne Bequett, Model Predictive Control, 2000, 9
[28] Seborg, D.A.. A perspective on advanced strategies for process control. Modeling, Identification and Control, 1994, 15(3): 179-189
[29] D.Q. Mayne et al. Constrained model predictive control: Stability and optimality. Automati ca2000,36:789-814
[30] Lee J H ,Yu ZH. Tuning of Model Predictive Controller for Robust Performance. Computer Chem Eng, 1994, 18(1): 15-37
[31] 闫志恒，过程控制系统实时预测控制软件开发与应用，硕士学位论文，上海交通大学，2003
[32] Morari. Model predictive control: past, present and future. Computers & Chemical Engineering, 1999(23):667-682
[33] j. Richalet. Industrial application of model based predictive control. Automatica, 1993,29:1251-1274
[34] Seung Cheol Jeong and PooGyeon Park. Constrained MPC Algorithm for Uncertain Time-VaryingSystems With State-Delay. IEEE Transaction On Automation Control. 2005, 50(2): 257-263
[35] 孙优贤，工业过程控制技术，化学工业出版社，2006
[36] 康红星，预测控制在锅炉燃烧系统中的应用，湖北电力，2001
[37] 梁春燕，谢剑英，预测控制中的若干问题研究，自动化与仪器仪表，1999(4)
[38] 刘妹妹，陈际达，廖力清，给予预测控制理论的位置伺服系统设计，电气传 1997(2)
[39] 闫志恒，李少远，基于PC机平台的预测控制系统仿真软件开发与实现，系统仿真学报，2002
[40] 席裕庚，耿晓军，陈虹(2000)，预测控制性能研究的新进展，控制理论与应用，Vol．17，No．4
[41] 邵雪飞，一类计算机新型控制策略-预测控制的改进研究及仿真设计，1999，5
[42] 徐立鸿，预测控制方法及其应用，上海，同济大学电气系
[43] 濮强国，动态矩阵控制方法的仿真研究，苏州城建环保学院学报．2002
[44] 陈荣保，李劲松，基于动态矩阵算法的液位控制系统，自动化仪表．2001
[45] 孙德敏等，工业过程先进控制及优化软件产业，自动化博览，2003．2