Incipient fault diagnosis method of nonlinear chemical process based on weighted statistical local KPCA

doi:10.11949/j.issn.0438-1157.20181307

Abstract

Abstract:

The traditional local kernel principal component analysis (SLKPCA) does not consider the difference of samples when constructing the improved residual, so that the fault sample information is easily covered by other samples. This paper proposes a new fault diagnosis method of nonlinear chemical process based on weighted statistical local kernel principal component analysis (WSLKPCA). Firstly, the score vectors and the eigenvalues are obtained using KPCA and the residual function is constructed. Then, a weighting strategy based on the distance between the test sample and the training sample is designed to construct the weighting improved residual, which assigns larger weights to samples with strong incipient fault information to enhance the impact of fault samples. Finally, the contribution graph is constructed based on the weighted mutual information between the measured variables and monitoring statistics to identify the fault source variables. Simulation results on continuous stirred tank reactor and TE process show that the proposed method can effectively detect incipient faults, and has better fault recognition performance.

Key words: chemical process, incipient fault, kernel principal component analysis, statistical local approach, fault diagnosis

摘要：

传统统计局部核主元分析（statistical local kernel principal component analysis, SLKPCA）在构造改进残差时未考虑样本的差异性，使得故障样本信息易于被其他样本所掩盖，针对该问题，提出一种基于加权统计局部核主元分析（weighted statistical local kernel principal component analysis, WSLKPCA）的非线性化工过程微小故障诊断方法。该方法首先利用KPCA获取过程的得分向量和特征值并构建初始残差。然后设计了一种基于测试样本与训练样本之间距离的加权策略构建加权改进残差，对含有较强微小故障信息的样本赋予较大权值，以增强故障样本的影响。最后，采用基于测量变量与监控统计量之间的加权互信息构建贡献图以识别故障源变量。在连续搅拌反应釜和田纳西伊斯曼（Tennessee Eastman, TE）化工过程上的仿真结果表明,所提方法具有良好的微小故障检测与识别性能。

关键词: 化工过程, 微小故障, 核主元分析, 统计局部方法, 故障诊断

CLC Number:

TP 277

Jiawei DENG, Xiaogang DENG, Yuping CAO, Xiaoling ZHANG. Incipient fault diagnosis method of nonlinear chemical process based on weighted statistical local KPCA[J]. CIESC Journal, 2019, 70(7): 2594-2605.

邓佳伟, 邓晓刚, 曹玉苹, 张晓玲. 基于加权统计局部核主元分析的非线性化工过程微小故障诊断方法[J]. 化工学报, 2019, 70(7): 2594-2605.

Figures/Tables 17

References 37

1	ZhangS, ZhaoC. Stationarity test and Bayesian monitoring strategy for fault detection in nonlinear multimode processes[J]. Chemometrics and Intelligent Laboratory Systems, 2017, 168: 45-61.
2	韩敏, 张占奎. 基于改进核主成分分析的故障检测与诊断方法[J]. 化工学报, 2015, 66(6): 2139-2149.
	HanM, ZhangZ K. Fault detection and diagnosis method based on modified kernel principal component analysis[J]. Journal of Chemical Industry and Engineering, 2015, 66(6): 2139-2149.
3	GeZ, SongZ, DingS X, et al. Data mining and analytics in the process industry: the role of machine learning[J]. IEEE Access, 2017; 5: 20590-20616.
4	王磊, 邓晓刚, 徐莹, 等. 基于变量子域PCA的故障检测方法[J]. 化工学报, 2016, 67(10): 4300-4308.
	WangL, DengX G, XuY, et al. Fault detection method based on variable sub-region PCA[J]. CIESC Journal, 2016, 67(10): 4300-4308.
5	DengX, TianX, ChenS, et al. Fault discriminant enhanced kernel principal component analysis incorporating prior fault information for monitoring nonlinear processes[J]. Chemom. Intell. Lab. Syst., 2017, 162: 21-34.
6	文成林, 吕菲亚, 包哲静, 等. 基于数据驱动的微小故障诊断方法综述[J]. 自动化学报, 2016, 42(9): 1285-1299.
	WenC L, LyuF Y, BaoZ J, et al. A review of data driven-based incipient fault diagnosis[J]. Acta Automatica Sinica, 2016, 42(9): 1285-1299.
7	李娟, 周东华, 司小胜, 等. 微小故障诊断方法综述[J]. 控制理论与应用, 2012, 29(12): 1517-1529.
	LiJ, ZhouD H, SiX S, et al. Review of incipient fault diagnosis methods[J]. Control Theory and Applications, 2012, 29(12): 1517-1529.
8	ShenY, StevenX D, AdelH, 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.
9	李元, 宋玉成. MWMPCA在过程故障检测与诊断中的应用[J]. 沈阳化工大学学报, 2010, 24(2): 170-174.
	LiY, SongY C. Application of MWMPCA in process fault detection and diagnosis[J]. Journal of Shenyang University of Chemical Technology 2010, 24(2): 170-174.
10	王政, 孙锦程, 王迎春, 等. 基于复杂网络理论的符号有向图(SDG)化工故障诊断[J]. 化工进展, 2016, 35(5): 1344-1352.
	WangZ, SunJ C, WangY C, et al. Research on chemical process signed directed graph(SDG) fault diagnosis based on complex network[J]. Chemical Industry and Engineering Progress, 2016, 35(5): 1344-1352.
11	郑鑫, 田学民, 张汉元. 基于动态稀疏保局投影的故障检测方法[J]. 化工学报, 2016, 67(3): 833-838.
	ZhengX, TianX M, ZhangH Y. Fault detection method based on dynamic sparse locality preserving projections[J]. CIESC Journal, 2016, 67(3): 833-838.
12	WoldS. Exponentially weighted moving principal components analysis and projections to latent structures[J]. Chemometrics & Intelligent Laboratory Systems, 1994, 23(1): 149-161.
13	HarmoucheJ, DelphaC, DialloD. Incipient fault detection and diagnosis based on Kullback-Leibler divergence using principal component analysis: Part I[J]. Signal Processing, 2014, 94(1): 278-287.
14	ChenH, JiangB, LuN. An improved incipient fault detection method based on Kullback-Leibler divergence[J]. ISA Transaction, 2018, 79: 127-136.
15	GeW, WangJ, ZhouJ, et al. Incipient fault detection based on fault extraction and residual evaluation[J]. Industrial & Engineering Chemistry Research, 2015, 54(14): 3664-3677.
16	JiH, HeX, ShangJ, et al. Exponential smoothing reconstruction approach for incipient fault isolation[J]. Ind. Eng. Chem. Res., 2018, 57(18), 6353-6363.
17	尚骏, 陈茂银, 周东华. 基于变元统计分析的微小故障检测[J]. 上海交通大学学报, 2015, 49(6): 799-805.
	ShangJ, ChenM Y, ZhouD H. Incipient fault detection using transformed component statistical analysis[J]. Journal of Shanghai Jiaotong University, 2015, 49(6): 799-805.
18	ChenH, JiangB, LuN, et al. Deep PCA based real-time incipient fault detection and diagnosis methodology for electrical drive in high-speed trains[J]. IEEE Trans. Veh. Technol., 2018, 67(6): 4819-4830.
19	ZhangS, ZhaoC, GaoF. Incipient fault detection for multiphase batch processes with limited batches[J]. IEEE Transactions on Control Systems Technology, 2019, 27(1): 1-15.
20	ZhaoC, GaoF. A sparse dissimilarity analysis algorithm for incipient fault isolation with no priori fault information[J]. Control Engineering Practice, 2017, 65: 70-82.
21	ZhaoB, ShiH, SongB, et al. Quality weakly related fault detection based on weighted Dual-Step feature extraction[J]. IEEE Access, 2019, PP(99): 1-1.
22	GeZ, YangC, SongZ. Improved kernel PCA-based monitoring approach for nonlinear processes[J]. Chemical Engineering Science, 2009, 64(9): 2245-2255.
23	吴希军. 基于主元分析方法的空调系统传感器故障诊断研究[D]. 秦皇岛: 燕山大学, 2005.
	WuX J . Study of the sensor fault diagnosis in air conditioning system based on PCA[D]. Qinhuangdao: Yanshan University, 2005.
24	张成, 李元. 基于统计模量分析间歇过程故障检测方法研究[J]. 仪器仪表学报, 2013, 34(9): 2103-2110.
	ZhangC, LiY. Study on the fault detection method in batch process based on statistical pattern analysis[J]. Chinese Journal of Scientific Instrument, 2013, 34(9): 2103-2110.
25	LeeJ M, YooC K, LeeI B. Fault detection of batch processes using multiway kernel principal component analysis[J]. Computers & Chemical Engineering, 2004, 28(9): 1837-1847.
26	赵忠盖, 刘飞. 基于稀疏核主元分析的在线非线性过程监控[J]. 化工学报, 2008, 59(7): 1773-1777.
	ZhaoZ G, LiuF. On-line nonlinear process monitoring based on sparse kernel principal component analysis[J]. Journal of Chemical Industry and Engineering(China), 2008, 59(7): 1773-1777.
27	KrugeraU. Improved principal component monitoring using the local approach[J]. Automatica, 2007, 43(9): 1532-1542.
28	胡金海, 谢寿生, 骆广琦, 等. 一种基于贡献率图的KPCA故障识别方法[J]. 系统工程与电子技术, 2008, 30(3): 572-576.
	HuJ H, XieS S, LuoG Q, et al. Fault identification method of kernel principal component analysis based on contribution plots and its application[J]. Systems Engineering and Electronics, 2008, 30(3): 572-576.
29	ChenH M, VarshneyP K, AroraM K. Performance of mutual information similarity measure for registration of multitemporal remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2003, 41(11): 2445-2454.
30	KeysD, KholikovS, PevtsovA A. Application of mutual information methods in time distance helioseismology[J]. Solar Physics, 2015, 290(3): 659-671.
31	QianW, ShuW. Mutual information criterion for feature selection from incomplete data[J]. Neurocomputing, 2015, 168: 210-220.
32	HaeriM A, EbadzadehM M. Estimation of mutual information by the fuzzy histogram[J]. Fuzzy Optimization and Decision Making. 2014, 13(3): 287-318.
33	周东华, 李钢, 李元. 数据驱动的工业过程故障诊断技术: 基于主元分析与偏最小二乘的方法[M]. 北京: 科学出版社, 2011.
	ZhouD H, LiG, LiY. Data Driven Industrial Process Fault Diagnosis Technology: Method Based on Principal Component Analysis and Partial Least Squares[M]. Beijing: Science Press, 2011.
34	DengX, TianX. Multimode process fault detection using local neighborhood similarity analysis[J]. Chinese Journal of Chemical Engineering, 2014, 22(11/12): 1260-1267.
35	DengX, WangL. Modified kernel principal component analysis using double-weighted local outlier factor and its application to nonlinear process monitoring[J]. ISA Transactions, 2018, 72: 218-228.
36	LiN, YangY. Ensemble kernel principal component analysis for improved nonlinear process monitoring[J]. Industrial & Engineering Chemistry Research, 2015, 54(1): 318-329.
37	TongC, SongY, YanX. Distributed statistical process monitoring based on four-subspace construction and Bayesian inference[J]. Industrial & Engineering Chemistry Research, 2013, 52(29): 9897-9907.

变量	说明
C_a	反应物A从反应釜流出时的浓度
T	反应釜的温度
T_c	夹套出口冷却剂的温度
h	反应釜液位高度
Q	反应釜流出物料的浓度
Q_f	进料A的流量
Q_af T_f	反应釜进料A的浓度进料A的温度
Q_c	夹套内冷却剂的流量
T_cf	夹套入口冷却剂温度

变量	说明
C_a	反应物A从反应釜流出时的浓度
T	反应釜的温度
T_c	夹套出口冷却剂的温度
h	反应釜液位高度
Q	反应釜流出物料的浓度
Q_f	进料A的流量
Q_af T_f	反应釜进料A的浓度进料A的温度
Q_c	夹套内冷却剂的流量
T_cf	夹套入口冷却剂温度

故障	KPCA		SLKPCA		WSLKPCA
故障	T²	SPE	T²	SPE	T²	SPE
1	7.14	0.71	58.71	25.71	60.29	54
2	2.71	1.71	51.57	61.29	56	62.57
3	1.29	3.57	7.14	60.86	15.71	61
4	1	0.71	25.29	22.57	53.71	37.29
5	1.43	3.71	5.86	90.86	16.14	90.86
平均	2.71	2.00	29.71	52.26	40.37	61.14

故障	KPCA		SLKPCA		WSLKPCA
故障	T²	SPE	T²	SPE	T²	SPE
1	7.14	0.71	58.71	25.71	60.29	54
2	2.71	1.71	51.57	61.29	56	62.57
3	1.29	3.57	7.14	60.86	15.71	61
4	1	0.71	25.29	22.57	53.71	37.29
5	1.43	3.71	5.86	90.86	16.14	90.86
平均	2.71	2.00	29.71	52.26	40.37	61.14

故障	KPCA		SLKPCA		WSLKPCA
故障	T²	SPE	T²	SPE	T²	SPE
6	99.86	99.86	99.86	99.86	99.86	99.86
7	99.86	99.86	99.71	100	99.71	100
平均	99.86	99.86	99.79	99.93	99.79	99.93