基于邻域保持嵌入的主多项式非线性过程故障检测

doi:10.16183/j.cnki.jsjtu.2020.295

摘要/Abstract

摘要：

针对化工过程的变量数据维数高、非线性的问题,提出基于邻域保持嵌入(NPE)-主多项式分析(PPA) 的过程故障检测算法.应用NPE算法提取高维数据的低维子流形,能够解决传统的线性降维算法不能提取局部结构信息的问题,对维数进行约减.利用PPA法时,使用一组灵活的主多项式分量来描述数据, 能够有效地捕捉过程数据中固有的非线性结构.在降维后的流形空间进行主多项式分析并建立Hotelling’s T²和平方预测误差统计量模型,同时确定控制限以进行故障检测.最后,通过一组非线性数值实例和Tennessee Eastman化工过程数据,将NPE-PPA算法与传统的核主元分析法、PPA法进行对比分析,验证所提算法的有效性及优越性.

关键词: 邻域保持嵌入, 主多项式分析, 非线性过程, 故障检测

Abstract:

Aimed at the problem of high dimension and nonlinearity of variable data in chemical process, a process fault detection algorithm based on neighborhood preserving embedding(NPE )-principal polynomial analysis (PPA) is proposed in this paper. The NPE algorithm is used to extract low dimensional submanifolds of high dimensional data, which overcomes the problem that the traditional linear dimensionality reduction algorithm cannot extract local structure information, so as to reduce the dimensions. The PPA method is used to describe data by a set of flexible principal polynomial components, which can effectively capture the inherent nonlinear structure of process data. The principal polynomial analysis is conducted in the reduced manifold space, and Hotelling’s T² and square prediction error statistical models are established to determine the control limit for fault detection. Finally, compared with the traditional kernel principal component analysis and the PPA method, a group of nonlinear numerical examples and Tennessee Eastman chemical process data experiments are performed to verify the effectiveness and superiority of the NPE-PPA algorithm.

Key words: neighborhood preserving embedding (NPE), principal polynomial analysis (PPA), nonlinear process, fault detection

中图分类号:

TP277

李元, 姚宗禹. 基于邻域保持嵌入的主多项式非线性过程故障检测[J]. 上海交通大学学报, 2021, 55(8): 1001-1008.

LI Yuan, YAO Zongyu. Principal Polynomial Nonlinear Process Fault Detection Based on Neighborhood Preserving Embedding[J]. Journal of Shanghai Jiao Tong University, 2021, 55(8): 1001-1008.

图/表 6

图1

图2

表1

表2

图3

图4

参考文献 25

[1]	SEVERSON K, CHAIWATANODOM P, BRAATZ R D. Perspectives on process monitoring of industrial systems[J]. Annual Reviews in Control, 2016, 42:190-200. doi: 10.1016/j.arcontrol.2016.09.001 URL
[2]	MD NOR N, CHE HASSAN C R, HUSSAIN M A. A review of data-driven fault detection and diagnosis methods: Applications in chemical process systems[J]. Reviews in Chemical Engineering, 2020, 36(4):513-553. doi: 10.1515/revce-2017-0069 URL
[3]	JIANG Q C, YAN X F. Just-in-time reorganized PCA integrated with SVDD for chemical process monitoring[J]. AIChE Journal, 2014, 60(3):949-965. doi: 10.1002/aic.v60.3 URL
[4]	ZHANG X M, LI Y, KANO M. Quality prediction in complex batch processes with just-in-time learning model based on non-Gaussian dissimilarity measure[J]. Industrial and Engineering Chemistry Research, 2015, 54:7694-7705. doi: 10.1021/acs.iecr.5b01425 URL
[5]	LI Y, ZHANG X M. Variable moving windows based non-Gaussian dissimilarity analysis technique for batch processes fault detection and diagnosis[J]. The Canadian Journal of Chemical Engineering, 2015, 93(4):689-707. doi: 10.1002/cjce.v93.4 URL
[6]	KANO M, NAKAGAWA Y. Data-based process monitoring, process control, and quality improvement: Recent developments and applications in steel industry[J]. Computers & Chemical Engineering, 2008, 32(1/2):12-24. doi: 10.1016/j.compchemeng.2007.07.005 URL
[7]	LUO L J, BAO S Y, MAO J F, et al. Monitoring batch processes using sparse parallel factor decomposition[J]. Industrial & Engineering Chemistry Research, 2017, 56(44):12682-12692. doi: 10.1021/acs.iecr.7b02618 URL
[8]	LI Y, ZHANG X M. Diffusion maps based k-nearest-neighbor rule technique for semiconductor manufacturing process fault detection[J]. Chemometrics and Intelligent Laboratory Systems, 2014, 136:47-57. doi: 10.1016/j.chemolab.2014.05.003 URL
[9]	SCHOLKOPF B, SMOLA A, MULLER K R. Nonlinear component analysis as a kernel eigenvalue problem[J]. Neural Computation, 1998, 10(5):1299-1319. doi: 10.1162/089976698300017467 URL
[10]	LEE J M, YOO C, CHOI S W, et al. Nonlinear process monitoring using kernel principal component analysis[J]. Chemical Engineering Science, 2004, 59(1):223-234. doi: 10.1016/j.ces.2003.09.012 URL
[11]	CHOI S W, LEE C, LEE J M, et al. Fault detection and identification of nonlinear processes based on kernel PCA[J]. Chemometrics and Intelligent Laboratory Systems, 2005, 75(1):55-67. doi: 10.1016/j.chemolab.2004.05.001 URL
[12]	BELKIN M, NIYOGI P. Laplacian eigenmaps for dimensionality reduction and data representation[J]. Neural Computation, 2003, 15(6):1373-1396. doi: 10.1162/089976603321780317 URL
[13]	KRAMER M A. Nonlinear principal component analysis using autoassociative neural networks[J]. AIChE Journal, 1991, 37(2):233-243. doi: 10.1002/(ISSN)1547-5905 URL
[14]	MONAHAN A H. Nonlinear principal component analysis by neural networks: Theory and application to the Lorenz system[J]. Journal of Climate, 2000, 13(4):821-835. doi: 10.1175/1520-0442(2000)013<0821:NPCABN>2.0.CO;2 URL
[15]	DONG D, MCAVOY T J. Nonlinear principal component analysis—Based on principal curves and neural networks[J]. Computers & Chemical Engineering, 1996, 20(1):65-78. doi: 10.1016/0098-1354(95)00003-K URL
[16]	张妮, 田学民. 基于等距离映射的非线性动态故障检测方法[J]. 上海交通大学学报, 2011, 45(8):1202-1206.
	ZHANG Ni, TIAN Xuemin. Nonlinear dynamic fault detection method based on isometric mapping[J]. Journal of Shanghai Jiao Tong University, 2011, 45(8):1202-1206.
[17]	LAPARRA V, TUIA D, JIMÉNEZ S, et al. Principal polynomial analysis for remote sensing data processing[C]//2011 IEEE International Geoscience and Remote Sensing Symposium. Vancouver, BC, Canada: IEEE, 2011: 12536130.
[18]	LAPARRA V, TULA D, JIMENEZ S, et al. Nonlinear data description with principal polynomial analysis[C]//2012 IEEE International Workshop on Machine Learning for Signal Processing. Santander, Spain: IEEE, 2012: 13117221.
[19]	ZHANG X M, KANO M, LI Y. Principal polynomial analysis for fault detection and diagnosis of industrial processes[J]. IEEE Access, 2018, 6:52298-52307. doi: 10.1109/ACCESS.2018.2870140 URL
[20]	ZHANG X M, LI Y. Multiway principal polynomial analysis for semiconductor manufacturing process fault detection[J]. Chemometrics and Intelligent Laboratory Systems, 2018, 181:29-35. doi: 10.1016/j.chemolab.2018.08.005 URL
[21]	HE X F, CAI D, YAN S C, et al. Neighborhood preserving embedding[C]//Tenth IEEE International Conference on Computer Vision. Beijing: IEEE, 2005: 1208-1213.
[22]	DOWNS J J, VOGEL E F. A plant-wide industrial process control problem[J]. Computers & Chemical Engineering, 1993, 17(3):245-255. doi: 10.1016/0098-1354(93)80018-I URL
[23]	GE Z Q, SONG Z H. Process monitoring based on independent component analysis-principal component analysis (ICA-PCA) and similarity factors[J]. Industrial & Engineering Chemistry Research, 2007, 46(7):2054-2063. doi: 10.1021/ie061083g URL
[24]	YIN S, DING S X, HAGHANI A, et al. A comparison study of basic data-driven fault diagnosis and process monitoring methods on the benchmark Tennessee Eastman process[J]. Journal of Process Control, 2012, 22(9):1567-1581. doi: 10.1016/j.jprocont.2012.06.009 URL
[25]	JIANG L, SONG Z H, GE Z Q, et al. Robust self-supervised model and its application for fault detection[J]. Industrial & Engineering Chemistry Research, 2017, 56(26):7503-7515. doi: 10.1021/acs.iecr.7b00949 URL

故障编号	性质描述	变化类型
IDV1	物料U/C进料比改变,物料B含量不变	阶跃
IDV2	物料U/C进料比不变,物料B含量改变	阶跃
IDV3	物料D进料温度改变	阶跃
IDV4	反应器冷却入口温度改变	阶跃
IDV5	冷凝器冷却入口温度改变	阶跃
IDV6	物料U进料损失	阶跃
IDV7	物料C压力损失	阶跃
IDV8	物料U、B、C的组成比例改变	随机变量
IDV9	物料D进料温度改变	随机变量
IDV10	物料C进料温度改变	随机变量
IDV11	反应器冷却水入口温度改变	随机变量
IDV12	冷凝器冷却水入口温度改变	随机变量
IDV13	反应动力学参数改变	慢偏移
IDV14	反应器冷却阀门	粘住
IDV15	冷凝器冷却阀门	粘住
IDV16	未知	未知
IDV17	未知	未知
IDV18	未知	未知
IDV19	未知	未知
IDV20	未知	未知
IDV21	物流4阀门固定在恒定位置	恒定位置

故障	KPCA		PPA		NPE-PPA
故障	T²	SPE	T²	SPE	T²	SPE
IDV1	0.99	1	0.99	1	0.98	1
IDV2	0.98	0.98	0.98	0.99	0.98	0.99
IDV3	0.01	0.08	0.12	0.21	0.06	0.20
IDV4	0.09	0.48	0.19	0.97	0.13	0.99
IDV5	0.22	0.30	0.31	0.39	1	1
IDV6	0.91	1	0.99	1	1	1
IDV7	0.99	1	0.48	1	0.71	1
IDV8	0.97	0.98	0.95	0.98	0.85	0.98
IDV9	0.01	0.09	0.15	0.23	0.04	0.16
IDV10	0.08	0.75	0.51	0.67	0.83	0.93
IDV11	0.30	0.54	0.34	0.72	0.13	0.74
IDV12	0.96	0.99	0.97	0.99	0.99	1
IDV13	0.94	0.95	0.94	0.97	0.95	0.96
IDV14	0.94	1	0.83	1	0.89	1
IDV15	0.01	0.10	0.16	0.25	0.07	0.26
IDV16	0.04	0.77	0.39	0.65	0.83	0.95
IDV17	0.69	0.88	0.78	0.94	0.87	0.97
IDV18	0.89	0.89	0.88	0.91	0.90	0.92
IDV19	0	0.45	0.05	0.32	0.60	0.88
IDV20	0.24	0.79	0.43	0.70	0.84	0.92
IDV21	0.29	0.37	0.27	0.50	0.43	0.68