摘要
在流程工业,尤其在化工生产过程中,软测量技术的发展具有实际意义。这是因为:(1)许多重要变量难于实时测量;(2)对每个生产过程变量设置测量仪表是非常不经济的;(3)在线质量分析仪投资昂贵、难于维护、实时性差。总之,软测量技术将会为生产过程带来极大社会效益和经济效益。
对于高维的复杂生产过程,众多的过程变量之间往往存在严重的多重相关性。为了解决这类病态模型参数估计问题,多元统计投影方法可以通过投影消除过程变量之间的多重相关性并建立起能充分反映关键生产变量的软测量模型。
本文系统和深入地研究了这一方法的若干重要方面。针对现有方法存在的问题,结合工业流化床聚乙烯生产过程的实际应用,提出若干新的算法和解决方法。主要研究内容和成果如下:
1.阐述了软测量技术主要研究内容、常用建模方法、现状和发展情况,着重综述了基于多元统计投影软测量技术的方法和应用发展情况。
2.介绍了基于多元统计投影方法软测量技术的主要数学基础,包括:主元分析、主元回归、偏最小二乘和神经网络,为后续的研究和应用打下理论基础。
3.综合工业流化床聚乙烯生产过程的工艺特点和设计要求,给出了基于上位机的控制系统设计。分析了若干主要过程变量对质量指标(树脂熔融指数和密度)的影响趋势,进行软测量建模变量的挑选和数据准备。
4.提出基于块式递推PLS限定记忆思想的算法,运用这种新方法建立的自适应软测量模型能更好地跟踪过程的变化。完成了对仿真过程和工业流化床聚乙烯生产过程质量指标软测量应用分析。
5.提出将线性PLS模型通过神经网络逼近策略拓展到非线性的PLS-NN方法,构造了基于梯度下降算法的神经网络权值矩阵学习规则。以此算法进行了工业流化床聚乙烯质量指标的软测量建模,以便为生产操作提供指导。
最后,对全文作出总结,归纳了本文解决的问题,指出了基于多元统计投影软测量技术的今后值得关注和深入研究的方向。
In many process control situations, it is often difficult to estimate some important process variables due to the limitation of process techniques or measurement techniques. These variables, which are key indictors of process performance, are normally determined by off-line sample analyses in the laboratory or using an on-line analyzer. Product quality values calculated on-line by using indirect but readily variable measurements (known as soft measurements) have been considered as an efficient method to solve this problem. Soft measurement can be used to obtain a regression model between easily obtained measurements and quality variables using statistical techniques, or neural network, etc.
Soft measurement based on multivariable statistical project method is a kind of data modelling techniques. The paper systematically elaborates this technique from many aspects. Combining the actual application of industrial gas-fluidized bed polyethylene process with the problems which exist in traditional methods, some new algorithms and solutions are brought forward. Contents in this paper are as following:
1. The paper summarizes the definition, content, developing tendency and modeling methods of soft measurement, especially provides complete and historical introduction on multivariate statistical project method.
2. Briefly introduces the academic tools of soft measurement based on multivariate statistical project method, such as principle component analysis, principle component regression, partial least squares and neural network.
3. Analyzing the characters of the gas-fluidized polyethylene process, the paper propose design of control system. In order to build the product quality soft measurement, we discussed the relation of between process variables and product quality variables. Finally, all modeling variables and data are proposed.
4. A new algorithm, finite memory based on recursive PLS, is proposed. The adaptive algorithm is applied to build adaptive soft measurement which have more strong tracking ability and higher precision than traditional model. An application of the method to Slurry-Feed Ceramic Melter and industrial fluidized bed reactor for predicting quality variables is presented to show the effectiveness.
5. By using the universal approximation property of neural networks (NN), a three-layer feedforward neural networks is embedded into the framework of standard PLS (partial least squares) modeling method resulting in a nonlinear PLS-NN model. A gradient descent learning algorithm is employed to train the network.
Finally, the whole thesis is summarized, and some future research areas are highlighted.
引文
1.Liang J, Qian J X. Multivariate Statistical Process Monitoring and Control:Recent Developments and Applications to Chemical Industry[J]. Chinese J. of Chemical Eng., 2003, 11(2):191-203
2.Mcacvoy T J. Contemplative stance for chemical process control[J].Automation,1992,28(2):441-442
3.Van den Bos A. Application of statistical parameter estimation methods to physical measurement[J].J. Phys E: Sci Instrum., 1977,10:753-760
4.Rivier J-M, Bayart M, Thiriet J-M. Intelligent instruments: some modeling approaches[J]. Measurement Control, 1996,29:179-186
5.Hartnett M K, Lightbody G, Irwin G .W. Dynamic inferential estimation using principal component regression(PCR)[J].Chemometrics Intel. Lab. Systems, 1998,40:215-224
6.Russelll, S A, Kesavan P, Lee J H, Ogummnaike B A. Recursive data-based prediction and control of batch product quality [J]. AIChE J., 1998,44:2442-2458
7.Kano M, Miyazaki K, Hasebe S, Hashimoto L. Inferential control system of distillation compositions using dynamic PLS regression [J]. J. Process Control, 2000,10:157-166
8.Kresta J, Marlin T, MacGregor J. Choosing inferential variables for multicomponent distillation using PLS [J]. In: AIChE Annual Meeting, Chicago, 1990
9.MacGregor J F, Skagerberg B, Kiparissides C. Multivariate SPC and property inference applied to low density polyethylene reactor. In: Proceedings of lFAC Symp. ADCHEM'91, 1991,Toulouse,France:663-670
10.Zhang W, Yu J. MIMO soft sensors for estimating product quality with on-line correction [J]. Trans. IchemE, 2000,78A:612-620
11.Qin S J, Yue H, Dunia R. Self-validating inferential sensors with application to air emission monitoring [J].Indus. Eng. Chem. Res., 1997,36:1675-1685
12.Wise B M, Gallagher N B. The process chemometrics approach to process monitoring and fault detection [J]. J. Process Control, 1996,6:329-348
13.Santen A, Koot G L M, Zullo L C. Statistical data analysis of a chemical plant [J]. Computers Chem. Eng. J., 1997,21(suppl.):1123-1129
14.Piovoso M J, Kosanovich K A. Applications of multivariate statistical methods to process monitoring and controller design [J]. International J. Control, 1994,59:743-765
15.Wadi I. Neural network model predicts naphtha cut point[J]. Oil& Gas J., 1996,11:67-70
16.Chen G, McAvoy T J. Process control utilizing data based multivariate statistical models[J]. The Canadian J. Chem. Eng., 1996,74:1010-1024
17.Chen G, McAvoy T J, Pivoso M J. A multivariate statistical controller for on-line quality improvement[J]. J. Process Control, 1998,8:139-149
18.Mejdell T, Skogestad S. Estimation of distillation compositions from multiple temperature measurements using PLS [J]. Indus. Eng. Chem. Res., 1991,30:2543-2555
19.Mejdell T, Skogestad S. Composition estimator in a pilot-plant distillation column using multiple temperatures [J]. Indus. Eng. Chem. Res., 1991,30:2555-2564
20.Yabuki Y, MacGregor J F. Product quality control in semibatch reactors using midcourse corrections policies [J]. In: proceedings of IFAC Symposium on Advanced Control of Chemical Process, Elsevier Science Ltd., 1997,Banff:189-194
21.Baffi G, Martin E B, Morari A J. Nonlinear PLS revisited: The quadratic PLS algorithm [J]. Computers Chem. Eng., 1999,23:395-411
22.Ray W H. Modeling and control of polymerization reactors [J]. In "Dynamics and Control of Chemical Reactors, Distillation Columns and Batch Process",
Oxford: Pergamon Press, 1993:37-46
23.Embirucu M, Lima E L and Pinto J C. A survey of advanced control of polymerization reactors [J]. Polymer Engineering & Science, 1996, 36(2):433-447
24.MacGregor J F, Penlidis A, Hamielec A E. Control of polymerization reactors: a review [J]. Polymer Process Engineering, 1984, 2(2&3):179-206
25.Lines B, Hartlen D, Paquin F D. Polyethylene reactor modeling and control design[J]. Hydrocarbon Processing(International Ed.), 1993,72(6):119-120
26.McAuley K B, MacGregor J F. Nonlinear product property control in industrial gas-phase polyethylene reactors[J]. AIChE J, 1993, 39:855-866
27.Ellis M F, Taylor T W, Jensen K F. On-line molecular weight distribution estimation and control in batch polymerization[J]. AIChE J.,1994,40:445-462
28.Wise B M, A non-linear dynamic model identification challenge problem, Chemom. Intell. Lab. Syst., 1995, 30, 91-96
29.MacGregor J F, Kourti T, Nomikos R Analysis, monitoring and fault diagnosis of industrial processes using multivariate statistical projection methods. Proceedings of 13~(th) IFAC World Congress, San Francisco, USA, 1996
30.McAuley K B, MacGregor J F. A kinetic model for industrial gas-phase ethylene copolymerization[J]. AIChEJ, 1990, 36(6):837-849
31.Geladi P, Kowalski B R. Partial least squares: a tutorial[J]. Analytical Chimica Acta, 1986, 185:1-17
32.Wold S. et al. Nonlinear PLS modeling[J]. Chemometrics and Intelligent Lab. Systems, 1989, 7:53-65
33.Kano M, et al. Inferential Control System of Distillation Compositions Using Dynamic PLS Regression[J]. Journal of Process Control, 2000,10:157-166
34.Helland, et al. Recursive algorithm for partial least squares regression[J]. Chemometrics Intell. Lab. Syst., 1992,14:129-137
35.Qin S J. Recursive PLS algorithms for adaptive data modeling[J]. Computers &
Chemical Engineering, 1998,22(4/5):503-514
36.Wise B M, et al. The process chemometrics approach to process monitoring and fault detection[J], J. Proc. Cont.,1996,16:329-348
37.Wold S, Kettaneh-Wold N, Skagerberg B. Nonlinear PLS Modeling[J], Chemon. Intell. Lab. Syst., 1989, 7:53-56.
38.Wold S. Nonlinear PLS Modeling Ⅱ: Spline Inner Relation(SPL-PLS) [J], Chemon. Intell. Lab. Syst., 1992, 14.
39.Ljung L, Issues in system identification, IEEE Control Syst., 1991, 25-29
40.Qin S J, McAvoy T J. Nonlinear PLS Modeling Using Neural Networks[J]. Computer & Chemical Engineering, 1992, 16:379-392
41.Dayal B S, Macgregor J F. Improved PLS algorithms[J], J. of Chem., 1997,11:73-85
42.Dayal B S, Macgregor J F. Recursive exponentially weighted PLS and its applications to adaptive control and prediction [J], J. Proc. Cont., 1997,3:169-179
43.Wod S, Trygg J, et al. Some recent developments in PLS modeling [J], Chemometrics and Intelligent Laboratory Systems, 2001,58:131-150
44.Xun W, et al. Recursive partial least squares algorithms for monitoring complex industrial processes [J], Control Engineering Practice 11, 2003:613-632
45.Ricker N L, The use of biased least-squares estimaors for parameters in discrete-time pulse response models[J], Ind. Eng. Chem. Res., 1998, 27:343-350
46.李海青 黄志尧 等.软测量技术原理及应用.北京:化学工业出版社,2000
47.王靖岱.连续聚合过程中产品过渡的最优策略研究.杭州:浙江大学硕士学位论文,1999
48.张雪珍.国内外聚乙烯的生产与消费[J].现代化工,2000,10
49.王桂增 王诗宓 等.高等过程控制[M].北京:清华大学出版社,2002,3
50.王树青等.先进控制技术及应用[M].北京:化学工业出版社,2001,7
51.王小慧.实时数据库与DCS通讯接口的设计与实现[J].化工自动化及仪
表,2002,29
52.谭杰,刘圣.聚乙烯生产装置关键控制回路的预测控制[J].化工自功化及仪表,2002,29
53.王跃宣,苏宏业,牟盛静,褚健.组件化软测量软件包的开发与应用[J].化工自动化及仪表,2003,30(3):43-46
54.李胜 等.APP平台下的数据转换与分布式历史数据库的实现[J].工业控制计算机,2001,14
55.夏伯锴等.基于上位机的先进控制与实时优化系统及其应用[J],工业控制计算机,2002,S
56.方崇智,萧德云.过程辨识[M].清华大学出版社,1988.
57.王根平,张玉润,周春晖.制碱过程碳酸化塔中CO_2浓度的软测量[J].浙江大学学报工学版,2000,34(5):561-564.
58.臧春华,郭小萍,王秀丽.基于PCA—改进BP算法的软测量技术[J].仪表技术与传感器.2001,2:26-29.
59.刘瑞兰,苏宏业,褚健.PLS回归软测量方法在催化重整稳定油组分估计中的应用[J].化工自动化及仪表,2002,29(5):44~47
60.冯瑞,张浩然,邵惠鹤。基于SVM的软测量建模[J],信息与控制,2002,31(6),567-571.
61.高倩,苏宏业,刘瑞兰,褚健.基于PLS软测量技术在加氢裂化产品质量控制中的应用[J],上海交通大学学报,2002,36(Sup.):135-137
62.陈曦,钱积新,高福荣.注塑过程熔体充填长度的动态神经网络软测量模型[J],化工自动化及仪表,2002,29(2):43-46
63.易灵芝,朱建林,张林亭,李卫平.软测量的工程化设计[J].自动检测技术,2002,21(6):25-28.
64.杨慧中,张素贞,陶振麟.聚丙烯腈质量指标软测量混合模型的应用[J].化工自动化及仪表,2002,29(6):11-13
65.冯瑞,黎达宇,邵惠鹤.基于实时嵌入式操作系统的软测量软件包[J],自动化仪表,2002,23(9):29-32
66.朱贷,何衍庆,王行愚.软测量技术在PTA生产过程中的应用[J].仪器仪表学报,2001,22(4):271-273
67.李春富,王桂增,徐博文.聚丙烯熔融指数软测量[J].化工自动化及仪表,2002,29(5):22-25
68.江铭虎,袁保宗,林碧琴.特征主元提取与自组织映射的剖析[J].北方交通大学学报,1997,21(5):524-529
69.颜学峰,陈德钊,胡上序.用自适应偏最小二乘回归为药物定量构效关系建模[J],分析化学研究报告,2002,30(5):536—539
70.何梅,李通化,从培盛.用数值遗传算法改进非线性PLS法进行构效关系研究[J],高等学校化学学报,1997,18(6):854—859
71.文忠,俞金寿,华向明.非线性部分最小二乘法在催化剂活性建模中的应用[J].华东理工大学学报,1998,24(6):717—721