基于增量式等距映射同双重局部密度方法的工业过程故障检测

doi:10.16183/j.cnki.jsjtu.2022.423

上海交通大学学报 ›› 2024, Vol. 58 ›› Issue (4): 525-533.doi: 10.16183/j.cnki.jsjtu.2022.423

基于增量式等距映射同双重局部密度方法的工业过程故障检测

冯立伟¹^,², 孙立文²^,³, 顾欢²^,³, 李元¹()

1.沈阳化工大学理学院,沈阳 110142
2.沈阳化工大学计算机科学与技术学院,沈阳 110142
3.辽宁省化工过程工业智能化技术重点实验室, 沈阳 110142

收稿日期:2022-10-28 修回日期:2022-12-09 接受日期:2022-12-30 出版日期:2024-04-28 发布日期:2024-04-30
通讯作者: 李元,教授,博士生导师; E-mail: li-yuan@mail.tsinghua.edu.cn.
作者简介:冯立伟(1980-),讲师,从事基于数据驱动复杂过程故障监控与诊断研究.
基金资助:
国家自然科学基金(61673279);国家自然科学基金(62273242)

Industrial Process Fault Detection Based on Incremental Isometric Mapping and Double Local Density Method

FENG Liwei¹^,², SUN Liwen²^,³, GU Huan²^,³, LI Yuan¹()

1. College of Science, Shenyang University of Chemical Technology, Shenyang 110142, China
2. College of Computer Science and Technology, Shenyang University of Chemical Technology, Shenyang 110142, China
3. Key Laboratory of Intelligent Technology for Chemical Process Industry of Liaoning Province, Shenyang 110142, China

Received:2022-10-28 Revised:2022-12-09 Accepted:2022-12-30 Online:2024-04-28 Published:2024-04-30

摘要/Abstract

摘要：

针对工业过程的非线性和动态性问题,提出一种基于流形学习下的增量式等距映射(IISOMAP)与双重局部密度(DLD)相结合的故障检测方法(IISOMAP-DLD).利用 IISOMAP 将原始数据映射到低维流形特征子空间和剩余子空间;然后,在两个子空间中分别引入双重局部密度方法构建统计量对过程进行监控;最后,将IISOMAP-DLD方法应用到田纳西-伊斯曼(TE)过程.实验结果表明,IISOMAP-DLD对比其他方法有更高的故障检测率.IISOMAP在保留数据内在特征的同时,解决了过程的非线性问题,而双重局部密度方法可消除过程的动态性.

关键词: 流形学习, 等距映射, 局部密度, 故障检测, 动态性

Abstract:

To address the nonlinearity and dynamics of industrial processes, an incremental isometric mapping (IISOMAP) in combination with double local density (DLD) is proposed as a fault detection method (IISOMAP-DLD) based on stream shape learning. First, IISOMAP is used to map the raw data into a low-dimensional manifold feature subspace and a residual subspace. Then, the double local density method is introduced in the two subspaces respectively to construct statistics to monitor the process. Finally, the IISOMAP-DLD method is applied to the Tennessee-Eastman (TE) process, and the experimental results show that IISOMAP-DLD has a higher fault detection rate than the other methods. IISOMAP preserves the intrinsic characteristics of the data and solves the nonlinear problems of the process, while the double local density method can eliminate the dynamic of the process.

Key words: manifold learning, isometric mapping (ISOMAP), local density, fault detection, dynamic

中图分类号:

TP277

冯立伟, 孙立文, 顾欢, 李元. 基于增量式等距映射同双重局部密度方法的工业过程故障检测[J]. 上海交通大学学报, 2024, 58(4): 525-533.

FENG Liwei, SUN Liwen, GU Huan, LI Yuan. Industrial Process Fault Detection Based on Incremental Isometric Mapping and Double Local Density Method[J]. Journal of Shanghai Jiao Tong University, 2024, 58(4): 525-533.

图/表 8

图1

图2

图3

图4

图5

图6

表1

TE过程中各方法的故障检测率

组别	KPCA		DKPCA		WKNN D/%	KSFA		IISOMAP-DLD
组别	T²/%	SPE/%	T²/%	SPE/%	WKNN D/%	T²/%	$T e 2$ /%	τ/%	τ_e/%
f1	98.86	99.14	98.54	99.43	98.86	99.00	98.57	99.14	99.54
f2	94.01	94.86	92.44	94.15	96.57	97.00	94.86	99.00	99.54
f3	0.43	15.55	0.57	1.41	20.86	25.00	52.29	69.00	19.86
f4	99.71	99.71	100	100	99.86	99.86	99.86	100	99.86
f5	0	30.10	0	0.70	37.71	42.00	57.71	73.71	14.71
f6	99.71	99.71	100	100	99.86	99.86	99.86	100	99.86
f7	99.71	99.71	100	100	99.86	99.86	99.86	100	99.86
f8	86.88	89.87	85.84	89.84	88.43	88.71	88.71	90.00	86.00
f9	3.42	17.26	1.57	3.34	22.00	17.71	47.00	62.14	15.57
f10	85.16	89.30	83.12	84.55	88.14	90.86	87.29	90.43	83.00
f11	96.86	98.43	94.13	98.00	97.86	98.14	92.00	98.71	95.29
f12	36.09	80.88	24.75	46.35	70.86	60.57	73.29	79.71	42.57
f13	94.01	95.01	94.28	94.99	94	94.43	93.71	95.57	93.71
f14	98.57	98.72	98.71	98.86	98.71	98.71	91.43	99.00	98.00
f15	0	3.57	0	0	13.29	7.29	58.86	67.14	7.57
f16	1.28	1.28	1.29	1.00	0.86	0.71	0.71	29.86	16.43
f17	83.17	84.88	82.98	84.55	84.71	84.57	82.43	86.86	83.00
f18	49.07	63.62	49.50	58.08	56.86	55.57	59.86	68.86	62.00
f19	91.87	95.72	89.56	94.42	94.57	96.00	94.43	96.71	95.14
f20	84.88	85.16	84.98	85.41	84.29	84.43	81.86	85.14	83.71
f21	1.43	1.28	1.43	1.00	0.71	0.71	1.41	29.29	15.86

表1

图7

参考文献 30

[1]	RAMAKRISHNA K, MRINMAYEE B, MUDDU M. Kantorovich distance based fault detection scheme for non-linear processes[J]. IEEE Access, 2022, 10: 1051-1067 doi: 10.1109/ACCESS.2021.3138696 URL
[2]	陈法法, 杨晓青, 陈保家, 等. 基于正交邻域保持嵌入与多核相关向量机的滚动轴承早期故障诊断[J]. 计算机集成制造系统, 2018, 24(8): 1946-1954.
	CHEN Fafa, YANG Xiaoqing, CHEN Baojia, et al. Early fault diagnosis of rolling bearing based on orthogonal neighbourhood preserving embedding and multi-kernel relevance vector machine[J]. Computer Integrated Manufacturing Systems, 2018, 24(8): 1946-1954.
[3]	GUO J Y, ZHONG L L, LI Y. Fault detection of multi-mode batch process based on statistics difference LPP[J]. Application Research of Computers, 2019, 36(1): 123-126.
[4]	CAO L J, CHUA K S, CHONG W K, et al. A comparison of PCA, KPCA and ICA for dimensionality reduction in support vector machine[J]. Neurocomputing, 2003, 55(1/2): 321-336. doi: 10.1016/S0925-2312(03)00433-8 URL
[5]	赵小强, 姚红娟. 改进邻域保持嵌入—独立元分析的间歇过程故障检测算法[J]. 计算机集成制造系统, 2021, 27(4): 1062-1071. doi: 10.13196/j.cims.2021.04.010
	ZHAO Xiaoqiang, YAO Hongjuan. Fault detection algorithm of batch process based on improved neighborhood preserving embedding-independent component analysis[J]. Computer Integrated Manufacturing Systems, 2021, 27(4): 1062-1071.
[6]	KONG X Y, LI Q, AN Q S, et al. Quality-related fault detection based on partial least squares score reconstruction[J]. Control Theory and Applications, 2020, 37(11): 2321-2332.
[7]	董顺, 李益国, 孙栓柱, 等. 基于状态空间主成分分析网络的故障检测方法[J]. 化工学报, 2018, 69(8): 237-245.
	DONG Shun, LI Yiguo, SUN Shuanzhu, et al. Fault detection method based on state space principal component analysis network[J]. CIESC Journal, 2018, 69(8): 237-245.
[8]	LEE J, YOO C, LEE I. Fault detection of batch processes using multiway kernel principal component analysis[J]. Computers and Chemical Engineering, 2004, 28(9): 1837-1847. doi: 10.1016/j.compchemeng.2004.02.036 URL
[9]	邓佳伟, 邓晓刚, 曹玉苹, 等. 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. doi: 10.11949/j.issn.0438-1157.20181307
	DENG Jiawei, DENG Xiaogang, CAO Yuping, et al. Incipient fault diagnosis method of nonlinear chemical process based on weighted statistical local KPCA[J]. CIESC Journal, 2019, 70(7): 2594-2605.
[10]	KU W, ROBERT H, CHRISTOS G. Disturbance detection and isolation by dynamic principal component analysis[J]. Chemometrics and Intelligent Laboratory Systems, 1995, 30: 175-196.
[11]	张佳鑫, 罗文嘉, 戴一阳. 基于CTA-DKPCA的化工过程故障诊断[J]. 控制工程, 2021, 28(5): 844-850.
	ZHANG Jiaxin, LUO Wenjia, DAI Yiyang. Chemical process fault diagnosis based on CTA-DKPCA[J]. Control Engineering, 2021, 28(5): 844-850.
[12]	冯立伟, 张成, 李元. 基于PC-WKNN的多工况间歇过程故障检测方法研究[J]. 计算机应用研究, 2018, 35(4): 1130-1134.
	FENG Liwei, ZHANG Cheng, LI Yuan. Research on fault detection method of multi-mode intermittent process based on PC-WKNN[J]. Application Research of Computers, 2018, 35(4): 1130-1134.
[13]	ZHANG C, GAO X W, LI Y, et al. Fault detection strategy based on weighted distance of k nearest neighbors for semiconductor manufacturing processes[J]. IEEE Transactions on Semiconductor Manufacturing, 2018, 32(1): 75-81. doi: 10.1109/TSM.2018.2857818 URL
[14]	WANG Q F, WANG S, WEI B K, et al. Weighted k-NN classification method of bearings fault diagnosis with multi-dimensional sensitive features[J]. IEEE Access, 2021, 9: 45428-45440. doi: 10.1109/ACCESS.2021.3066489 URL
[15]	SHANG C, HUANG B, YANG F, et al. Probabilistic slow feature analysis-based representation learning from massive process data for soft sensor modeling[J]. AIChE Journal, 2015, 61(12): 4126-4139. doi: 10.1002/aic.v61.12 URL
[16]	BOHMER W, GRUNEWALDER S, NICKISCH H, et al. Regularized sparse kernel slow feature analysis[C]// European Conference on Machine Learning and Knowledge Discovery in Databases-Volume Part I. Heidelberg, Germany: Springer, 2011: 235-248.
[17]	卢依容. 基于核慢特征分析算法的故障检测与诊断[D]. 上海: 上海交通大学, 2015.
	LU Yirong. Fault detection and diagnosis based on kernel slow feature analysis algorithm[D]. Shanghai: Shanghai Jiao Tong University, 2015.
[18]	ZHANG H, TIAN X, CAI L. Nonlinear process fault diagnosis using kernel slow feature discriminant analysis[J]. IFAC Papersonline, 2015, 48(21): 607-612. doi: 10.1016/j.ifacol.2015.09.593 URL
[19]	黄健, 杨旭, 陈先中. 基于故障相关慢特征分析的过程监测方法[J]. 高校化学工程学报, 2020, 34(5): 1290-1296.
	HUANG Jian, YANG Xu, CHEN Xianzhong. Process monitoring method based on fault related slow characteristic analysis[J]. Journal of Chemical Engineering of Chinese Universities, 2020, 34(5): 1290-1296.
[20]	TENENBAUM J, SILVA V, LANGFORD J. A global geometric framework for nonlinear dimensionality reduction[J]. Science, 2000, 290(5500): 2319-2323. doi: 10.1126/science.290.5500.2319 pmid: 11125149
[21]	ZHANG Y, LI B W, WANG Z B, et al. Fault diagnosis of rotating machine by isometric feature mapping[J]. Journal of Mechanical Science & Technology, 2013, 27(11): 3215-3221.
[22]	张妮, 田学民, 蔡连芳. 基于RISOMAP的非线性过程故障检测方法[J]. 化工学报, 2013, 64(6): 2125-2130. doi: 10.3969/j.issn.0438-1157.2013.06.031
	ZHANG Ni, TIAN Xuemin, CAI Lianfang. Nonlinear process fault detection method based on RISOMAP[J]. CIESC Journal, 2013, 64(6): 2125-2130.
[23]	COX T, COX M. Multidimensional scaling[J]. Journal of the Royal Statistical Society, 2001, 46(2): 1050-1057.
[24]	NASIR S, HAEWOON N, MIAN I U H, et al. A survey on multidimensional scaling[J]. ACM Computing Surveys (CSUR), 2018, 51(3): 471-496.
[25]	LAW M, JAIN A. Incremental nonlinear dimensionality reduction by manifold learning[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2006, 28(3): 377-391.
[26]	SHANBHAG D N, RAO C R. Handbook of statistics stochastic processes: Theory and methods[M]. Amsterdam,Netherlands: Elsevier, 2001.
[27]	DENG X, TIAN X, CHEN S, et al. Nonlinear process fault diagnosis based on serial principal component analysis[J]. IEEE Transactions on Neural Networks & Learning Systems, 2018, 29(3): 560-572.
[28]	郭金玉, 李文涛, 李元. 在线压缩核主元分析的自适应过程监控[J]. 上海交通大学学报, 2022, 56(10): 1397-1408. doi: 10.16183/j.cnki.jsjtu.2021.084
	GUO Jinyu, LI Wentao, LI Yuan. Adaptive process monitoring based on on-line compressed kernel principal component analysis[J]. Journal of Shanghai Jiao Tong University, 2022, 56(10): 1397-1408.
[29]	ZHONG X, HAN M, QIU T, et al. Fault diagnosis of complex process using sparse kernel local fisher discriminant analysis[J]. IEEE Transactions on Neural Networks & Learning Systems, 2020, 31(5): 1581-1591.
[30]	BATHELT A, RICKER N, JELALI M. Revision of the Tennessee Eastman process model[J]. IFAC PapersOnLine, 2015, 48(8): 309-314.

基于增量式等距映射同双重局部密度方法的工业过程故障检测

Industrial Process Fault Detection Based on Incremental Isometric Mapping and Double Local Density Method

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 30

相关文章 15

编辑推荐

Metrics

本文评价

[1]	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.
[2]	何夏维, 蔡云泽, 严玲玲. 一种合成残差式的反作用轮故障检测方法[J]. 上海交通大学学报, 2021, 55(6): 716-728.
[3]	徐巧宁，艾青林，杜学文，刘毅. 模型-数据联合驱动的船舶舵机电液伺服系统早期故障检测[J]. 上海交通大学学报, 2020, 54(5): 451-464.
[4]	黄健, 杨旭. 基于在线加权慢特征分析的故障检测算法[J]. 上海交通大学学报, 2020, 54(11): 1142-1150.
[5]	郭闻雨，张秀芳，修玉皎，夏唐斌，潘尔顺. 考虑动态性能退化的生产系统预防维护和缓存配置策略[J]. 上海交通大学学报, 2019, 53(9): 1107-1114.
[6]	郭红杰1，徐春玲2，侍洪波1. 基于局部邻域标准化策略的多工况过程故障检测[J]. 上海交通大学学报（自然版）, 2015, 49(06): 868-875.
[7]	柳春,姜斌,张柯,吴云凯. 带扰动的线性系统微小故障早期诊断方法[J]. 上海交通大学学报（自然版）, 2015, 49(06): 889-896.
[8]	丛亚，葛志强，宋执环. 多采样率主元分析的过程故障检测[J]. 上海交通大学学报（自然版）, 2015, 49(06): 762-767.
[9]	刘洋，何潇，周东华. 分布式测量下闭环系统的故障检测[J]. 上海交通大学学报（自然版）, 2015, 49(06): 757-761.
[10]	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（自然版）, 2015, 49(06): 799-805.
[11]	蔡连芳，田学民，张妮. 基于核状态空间ICA的非线性动态过程故障检测方法[J]. 上海交通大学学报（自然版）, 2014, 48(07): 971-976.
[12]	方挺，董冲，胡兴柳，王彦. 航拍图像中绝缘子串的轮廓提取和故障检测[J]. 上海交通大学学报（自然版）, 2013, 47(12): 1818-1822.
[13]	吴文通1，李元祥1，韦邦合2，郑思龙1. 局部测地距离估计的增量等距特征映射算法[J]. 上海交通大学学报（自然版）, 2013, 47(07): 1082-1087.
[14]	王天真1, 2, 高迪驹1, 刘萍1, 汤天浩1 . 基于动态限的非周期非稳定工况故障检测模型[J]. 上海交通大学学报（自然版）, 2012, 46(04): 607-612.
[15]	康嘉, 韩华, 谷波, 邱峰. 制冷空调系统故障检测与诊断中的空气侧换热模型[J]. 上海交通大学学报（自然版）, 2011, 45(09): 1368-1373.