基于加权概率CVDA的动态化工系统微小故障检测

doi:10.11949/0438-1157.20220417

摘要/Abstract

摘要：

典型变量差异度分析（CVDA）是近年来提出的一种新型动态过程监控方法，已在微小故障检测领域获得成功应用。针对传统CVDA方法忽视了特征量的概率信息挖掘问题，提出一种基于加权概率CVDA（WPCVDA）的动态化工系统微小故障检测方法。一方面，该方法在基本CVDA模型特征基础上引入Wasserstein距离（WD）度量特征量概率分布的变化，构造概率化的WD特征提高CVDA模型对微小故障的灵敏度；另一方面，进一步考虑不同的WD特征成分携带故障信息的差异性，设计一种自适应权值计算策略，为关键的故障敏感特征成分设置大的权值，突出其在监控统计量中的作用。在一个标准化工过程的验证结果说明，所提出的WPCVDA方法比传统CVDA方法具有更好的微小故障检测性能。

关键词: 微小故障检测, 加权系数, Wasserstein距离, 典型变量差异度分析, 动态过程

Abstract:

Canonical variable dissimilarity analysis (CVDA) is a new dynamic process monitoring method proposed in recent years, which has been successfully applied in the field of incipient fault detection. To solve the problem that traditional CVDA method neglects the probability information mining of features, one novel method based on weighted probability CVDA (WPCVDA) is proposed for incipient fault detection of dynamic chemical system. On the one hand, based on the basic CVDA model features, Wasserstein distance (WD) is introduced to measure the change of feature probability distribution to construct probability-related WD features to improve the sensitivity of CVDA model to incipient faults. On the other hand, further considering the difference of fault information carried by different WD feature components, an adaptive weight calculation strategy is designed to set large weights for key fault-sensitive feature components, so as to highlight their roles in the monitoring statistics. The validation results on a standard chemical process show that the proposed WPCVDA method has better performance than the traditional CVDA method in incipient fault detection.

Key words: incipient fault detection, weighting coefficient, Wasserstein distance, canonical variable dissimilarity analysis, dynamic process

中图分类号:

TP 277

杨明辉, 刘晓月, 邓晓刚, 廖明燕, 侯春望. 基于加权概率CVDA的动态化工系统微小故障检测[J]. 化工学报, 2022, 73(9): 3963-3972.

Minghui YANG, Xiaoyue LIU, Xiaogang DENG, Mingyan LIAO, Chunwang HOU. Incipient fault detection for dynamic chemical processes based on weighted probability CVDA[J]. CIESC Journal, 2022, 73(9): 3963-3972.

图/表 8

图1 基于WPCVDA的过程监控流程图

Fig.1 Flowchart of process monitoring based on WPCVDA

图2 TE过程流程图

Fig.2 TE process flowchart

表1 TE过程的测试故障

Table 1 The testing fault of TE process

故障模式	故障描述
F1	IDV4: 反应器冷却水入口温度阶跃变化
F2	IDV10: 进料温度随机变化
F3	IDV11:反应器冷却水入口温度随机波动
F4	IDV15: 反应器冷却水阀门黏滞
F5, F6	IDV19-20: 未知类型
F7	IDV21: 流4的阀门粘住

图3 三种方法对故障F4的监控图

Fig.3 Monitoring charts for fault F4 by three methods

图4 三种方法对故障F7的监控图

Fig.4 Monitoring charts for fault F7 by three methods

图5 三种方法的差异度特征成分对比分析

Fig.5 The dissimilarity feature components comparison of three methods

表2 四种方法的故障检出率对比/%

Table 2 Comparison of FDRs on four methods/%

Fault	CVDA			CVRA		PCVDA			WPCVDA
Fault	$T 2$	$Q$	$D$	$W D 1$	$W D 2$	$T W D 2$	$Q W D$	$D W D$	$W T W D 2$	$W Q W D$	$W D W D$
F1	81.38	100	99.50	99.25	99.50	97.50	99.88	98.50	98.13	99.88	99.00
F2	89.13	84.75	92.25	95.63	91.00	94.63	97.75	96.88	95.38	97.75	96.88
F3	59.63	75.50	79.13	89.00	91.63	51.25	99.25	93.00	76.25	99.25	96.13
F4	6.38	3.88	20.00	35.13	7.50	17.25	59.88	29.88	17.75	77.38	43.25
F5	71.75	93.00	93.50	99.00	98.88	56.63	99.63	99.50	97.88	99.75	100
F6	88.13	78.38	91.25	90.88	89.88	90.00	91.88	90.25	90.25	91.88	90.63
F7	46.88	38.38	74.75	77.25	45.13	63.13	79.00	80.63	64.50	82.00	80.75
mean	63.32	67.70	78.63	83.73	74.79	67.20	89.61	84.09	77.16	92.55	86.66

表2 四种方法的故障检出率对比/%

Table 2 Comparison of FDRs on four methods/%

Fault	CVDA			CVRA		PCVDA			WPCVDA
Fault	$T 2$	$Q$	$D$	$W D 1$	$W D 2$	$T W D 2$	$Q W D$	$D W D$	$W T W D 2$	$W Q W D$	$W D W D$
F1	81.38	100	99.50	99.25	99.50	97.50	99.88	98.50	98.13	99.88	99.00
F2	89.13	84.75	92.25	95.63	91.00	94.63	97.75	96.88	95.38	97.75	96.88
F3	59.63	75.50	79.13	89.00	91.63	51.25	99.25	93.00	76.25	99.25	96.13
F4	6.38	3.88	20.00	35.13	7.50	17.25	59.88	29.88	17.75	77.38	43.25
F5	71.75	93.00	93.50	99.00	98.88	56.63	99.63	99.50	97.88	99.75	100
F6	88.13	78.38	91.25	90.88	89.88	90.00	91.88	90.25	90.25	91.88	90.63
F7	46.88	38.38	74.75	77.25	45.13	63.13	79.00	80.63	64.50	82.00	80.75
mean	63.32	67.70	78.63	83.73	74.79	67.20	89.61	84.09	77.16	92.55	86.66

表3 四种方法的平均误报率对比/%

Table 3 Comparison of average FARs on four methods /%

FAR	CVDA			CVRA		PCVDA			WPCVDA
FAR	$T 2$	$Q$	$D$	$W D 1$	$W D 2$	$T W D 2$	$Q W D$	$D W D$	$W T W D 2$	$W Q W D$	$W D W D$
mean	0.71	0.63	0.36	0.27	0.36	0	0.27	0	0	0.45	0.09

表3 四种方法的平均误报率对比/%

Table 3 Comparison of average FARs on four methods /%

FAR	CVDA			CVRA		PCVDA			WPCVDA
FAR	$T 2$	$Q$	$D$	$W D 1$	$W D 2$	$T W D 2$	$Q W D$	$D W D$	$W T W D 2$	$W Q W D$	$W D W D$
mean	0.71	0.63	0.36	0.27	0.36	0	0.27	0	0	0.45	0.09

参考文献 32

1	Zhang X C, Jiang D X, Han T, et al. Rotating machinery fault diagnosis for imbalanced data based on fast clustering algorithm and support vector machine[J]. Journal of Sensors, 2017, 2017: 8092691.
2	商亮亮, 陆智林, 文传博, 等. 基于规范变量残差的化工过程微小故障检测与诊断[J]. 控制理论与应用, 2021, 38(8): 1247-1256.
	Shang L L, Lu Z L, Wen C B, et al. Canonical residual based incipient fault detection and diagnosis for chemical process[J]. Control Theory & Applications, 2021, 38(8): 1247-1256.
3	Chen H T, Jiang B, Zhang T Y, et al. Data-driven and deep learning-based detection and diagnosis of incipient faults with application to electrical traction systems[J]. Neurocomputing, 2020, 396: 429-437.
4	Taqvi S A A, Zabiri H, Tufa L D, et al. A review on data-driven learning approaches for fault detection and diagnosis in chemical processes[J]. ChemBioEng Reviews, 2021, 8(3): 239-259.
5	Zhu J L, Ge Z Q, Song Z H, et al. Review and big data perspectives on robust data mining approaches for industrial process modeling with outliers and missing data[J]. Annual Reviews in Control, 2018, 46: 107-133.
6	徐静, 王振雷, 王昕. 基于非线性动态全局局部保留投影算法的化工过程故障检测[J]. 化工学报, 2020, 71(12): 5655-5663.
	Xu J, Wang Z L, Wang X. Fault detection for chemical process based on nonlinear dynamic global-local preserving projections[J]. CIESC Journal, 2020, 71(12): 5655-5663.
7	Liu J H, Qu F M, Hong X W, et al. A small-sample wind turbine fault detection method with synthetic fault data using generative adversarial nets[J]. IEEE Transactions on Industrial Informatics, 2019, 15(7): 3877-3888.
8	胡云鹏, 陈焕新, 周诚, 等. 基于主元分析法的冷水机组传感器故障检测效率分析[J]. 化工学报, 2012, 63: 85-88.
	Hu Y P, Chen H X, Zhou C, et al. Analysis of sensor fault detection in chiller based on PCA method[J]. CIESC Journal, 2012, 63: 85-88.
9	Kano M, Hasebe S, Hashimoto I, et al. A new multivariate statistical process monitoring method using principal component analysis[J]. Computers & Chemical Engineering, 2001, 25(7/8): 1103-1113.
10	Chen Z W, Yang C H, Peng T, et al. A cumulative canonical correlation analysis-based sensor precision degradation detection method[J]. IEEE Transactions on Industrial Electronics, 2019, 66(8): 6321-6330.
11	Liu Y Q, Liu B, Zhao X J, et al. A mixture of variational canonical correlation analysis for nonlinear and quality-relevant process monitoring[J]. IEEE Transactions on Industrial Electronics, 2018, 65(8): 6478-6486.
12	Tong C D, Lan T, Yu H Z, et al. Distributed partial least squares based residual generation for statistical process monitoring[J]. Journal of Process Control, 2019, 75: 77-85.
13	Jia Q L, Zhang Y W. Quality-related fault detection approach based on dynamic kernel partial least squares[J]. Chemical Engineering Research and Design, 2016, 106: 242-252.
14	姚林, 张岩. 基于自适应混合核典型变量分析的工业过程质量相关故障检测[J]. 控制与决策, 2021, 36(4): 801-807.
	Yao L, Zhang Y. Quality-related fault detection for industrial processes based on adaptive mixed kernel canonical variable analysis[J]. Control and Decision, 2021, 36(4): 801-807.
15	Cao Y P, Yu L, Deng X G, et al. Variable sub-region canonical variate analysis for dynamic process monitoring[J]. IEEE Access, 2020, 8: 37775-37789.
16	Pilario K E S, Cao Y. Canonical variate dissimilarity analysis for process incipient fault detection[J]. IEEE Transactions on Industrial Informatics, 2018, 14(12): 5308-5315.
17	Pilario K E S, Cao Y, Shafiee M. Incipient fault detection, diagnosis, and prognosis using canonical variate dissimilarity analysis[J].Computer Aided Chemical Engineering, 2019: 1195-1200.
18	Pilario K E S, Cao Y, Shafiee M. Mixed kernel canonical variate dissimilarity analysis for incipient fault monitoring in nonlinear dynamic processes[J]. Computers & Chemical Engineering, 2019, 123: 143-154.
19	Yu W K, Zhao C H. Low-rank characteristic and temporal correlation analytics for incipient industrial fault detection with missing data[J]. IEEE Transactions on Industrial Informatics, 2021, 17(9): 6337-6346.
20	王宝祥, 潘宏侠. 基于分时段规范变量残差分析的高速自动机动态特性监测[J]. 振动与冲击, 2019, 38(20): 90-96.
	Wang B X, Pan H X. Dynamic performance monitoring of high-speed automata based on phase-partitioned canonical variate dissimilarity analysis[J]. Journal of Vibration and Shock, 2019, 38(20): 90-96.
21	Wu P, Ferrari R M G, Liu Y C, et al. Data-driven incipient fault detection via canonical variate dissimilarity and mixed kernel principal component analysis[J]. IEEE Transactions on Industrial Informatics, 2021, 17(8): 5380-5390.
22	Tang Q, Liu Y, Chai Y, et al. Dynamic process monitoring based on canonical global and local preserving projection analysis[J]. Journal of Process Control, 2021, 106: 221-232.
23	邓晓刚, 田学民. 基于鲁棒规范变量分析的故障诊断方法[J]. 控制与决策, 2008, 23(4): 415-419.
	Deng X G, Tian X M. Fault diagnosis method based on robust canonical variate analysis[J]. Control and Decision, 2008, 23(4): 415-419.
24	Han X J, Jiang J, Xu A D, et al. Fault detection of pneumatic control valves based on canonical variate analysis[J]. IEEE Sensors Journal, 2021, 21(12): 13603-13615.
25	Juricek B C, Seborg D E, Larimore W E. Fault detection using canonical variate analysis[J]. Industrial & Engineering Chemistry Research, 2004, 43(2): 458-474.
26	Odiowei P P, Cao Y. Nonlinear dynamic process monitoring using canonical variate analysis and kernel density estimations[J]. Computer Aided Chemical Engineering, 2009, 27: 1557-1562.
27	Zhang Z, Deng X G. Anomaly detection using improved deep SVDD model with data structure preservation[J]. Pattern Recognition Letters, 2021, 148: 1-6.
28	王晓慧, 王延江, 邓晓刚, 等. 基于加权深度支持向量数据描述的工业过程故障检测[J]. 化工学报, 2021, 72(11): 5707-5716.
	Wang X H, Wang Y J, Deng X G, et al. Industrial process fault detection using weighted deep support vector data description[J]. CIESC Journal, 2021, 72(11): 5707-5716.
29	Ni K Y, Bresson X, Chan T, et al. Local histogram based segmentation using the Wasserstein distance[J]. International Journal of Computer Vision, 2009, 84(1): 97-111.
30	Yang Q S, Yan P K, Zhang Y B, et al. Low-dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss[J]. IEEE Transactions on Medical Imaging, 2018, 37(6): 1348-1357.
31	Oh J H, Pouryahya M, Iyer A, et al. A novel kernel Wasserstein distance on Gaussian measures: an application of identifying dental artifacts in head and neck computed tomography[J]. Computers in Biology and Medicine, 2020, 120: 103731.
32	Deng X G, Deng J W. Incipient fault detection for chemical processes using two-dimensional weighted SLKPCA[J]. Industrial & Engineering Chemistry Research, 2019, 58(6): 2280-2295.

[1]	张国磊, 阙晨宇, 谭袖, 李健, 宋福元, 江正悦. 舰船快速减速过程中主动回汽控制仿真[J]. 化工学报, 2018, 69(S2): 358-364.
[2]	邓晓刚, 邓佳伟, 曹玉苹, 王磊. 基于双层局部KPCA的非线性过程微小故障检测方法[J]. 化工学报, 2018, 69(7): 3092-3100.
[3]	梁超, 王宏, 朱恂, 陈蓉, 丁玉栋, 廖强. 液滴撞击不同浸润性壁面动态过程的数值模拟[J]. 化工学报, 2013, 64(8): 2745-2751.
[4]	兰泽全, 曹欣玉, 饶甦, 赵显桥, 周俊虎, 刘建忠, 赵翔, 岑可法. 电子探针分析炉内沾污结渣动态过程 [J]. 化工学报, 2005, 56(1): 24-29.
[5]	吴建锋; 何小荣; 陈丙珍. 一种用于动态化工过程建模的反馈神经网络新结构 [J]. CIESC Journal, 2002, 53(2): 156-160.
[6]	赵劲松,陈丙珍,沈静珠. 一种用于动态过程故障诊断的子波-S型神经网络 [J]. CIESC Journal, 1997, 48(1): 1-7.
[7]	张树增,吴慧雄,李成岳. 二氧化硫强制动态氧化过程的模型化(Ⅱ)——模型模拟与参数分析 [J]. CIESC Journal, 1995, 46(4): 424-430.