基于VAE-OCCA的质量相关故障检测方法研究

doi:10.11949/0438-1157.20230058

摘要/Abstract

摘要：

由于闭环反馈系统的存在，并不是所有故障均会导致质量发生恶化。质量变量通常难以获得或具有一定的延迟，传统的无监督方法不能在检测过程是否正常的同时判断故障对质量的影响。典型相关分析(canonical correlation analysis，CCA)是一种经典的有监督方法，可以考虑输入输出间的关系，已被用于质量相关故障检测。然而，过程数据存在着维度高、非线性等问题，流程系统的复杂性使得CCA对于隐藏特征的捕获更具挑战性。提出了一种变分自编码器-正交典型相关分析(variational automatic encoder-orthogonal CCA，VAE-OCCA)方法。首先，利用变分自编码器对输入数据进行无监督自适应学习，实现对高维非线性过程变量的特征提取；进而，基于典型相关分析方法考虑输入输出关系，利用得到的相关系数矩阵进行奇异值分解建立质量相关和质量无关监测统计量；最后，通过工业案例测试说明提出方法的有效性及优越性。

关键词: 故障检测, 过程监测, 典型相关分析, 变分自编码器, 质量相关

Abstract:

Due to the existence of closed-loop feedback system, not all faults will lead to quality deterioration. Quality variables are usually difficult to obtain or have a certain delay. The traditional unsupervised methods cannot judge the impact of faults on quality while detecting whether the process is normal or not. Canonical correlation analysis (CCA), a classical supervised method that can consider the relationship between input and output, has been used for quality-related fault detection. However, process data has problems such as high dimensionality and nonlinearity. The complexity of the system makes CCA more challenging to capture hidden features. This paper proposes a variational automatic encoder-orthogonal canonical correlation analysis (VAE-OCCA) method. First, unsupervised adaptive learning is performed on the input data using a variational automatic encoder to achieve feature extraction for high-dimensional nonlinear process variables. Then, the input-output relationship is considered based on the canonical correlation analysis method, and the obtained correlation coefficient matrix is used to perform singular value decomposition to establish quality-related and quality-independent monitoring statistics. Finally, the effectiveness and superiority of the method proposed in this paper are demonstrated through industrial case tests.

Key words: fault detection, process monitoring, canonical correlation analysis, variational automatic encoder, quality-related

中图分类号:

TP 277

宋冰, 郑城风, 侍洪波, 陶阳, 谭帅. 基于VAE-OCCA的质量相关故障检测方法研究[J]. 化工学报, 2023, 74(4): 1630-1638.

Bing SONG, Chengfeng ZHENG, Hongbo SHI, Yang TAO, Shuai TAN. Research on quality-related fault detection method based on VAE-OCCA[J]. CIESC Journal, 2023, 74(4): 1630-1638.

图/表 14

图1 VAE网络结构示意图

Fig.1 Diagram of VAE network structure

表1 Adam优化器参数

Table 1 Optimizer parameters of Adam

参数	数值
α	0.001
β₁	0.4
β₂	0.999
ε	1×10^-8

图2 基于VAE-OCCA的故障检测流程

Fig.2 Flow chart of fault detection based on VAE-OCCA

图3 故障1中成分G的变化

Fig.3 Change of G in fault 1

图4 VAE-OCCA对故障1的检测结果

Fig.4 Detection result of VAE-OCCA for fault 1

图5 CCA、AE-CCA对故障1的检测结果

Fig.5 Detection result of CCA and AE-CCA for fault 1

图6 故障2中成分G的变化

Fig.6 Change of G in fault 2

图7 VAE-OCCA对故障2的检测结果

Fig.7 Detection result of VAE-OCCA for fault 2

图8 CCA、AE-CCA对故障2的检测结果

Fig.8 Detection result of CCA and AE-CCA for fault 2

表2 故障2在质量相关空间的故障检测率

Table 2 Fault detection rate of fault 2 in quality-related space

方法	FDR/%
CCA	91.750
AE-CCA	7.250
VAE-OCCA	98.125

图9 故障14中成分G的变化

Fig.9 Change of G in fault 14

图10 VAE-OCCA对故障14的检测结果

Fig.10 Detection result of VAE-OCCA for fault 14

图11 CCA、AE-CCA对故障14的检测结果

Fig.11 Detection result of CCA and AE-CCA for fault 14

表3 故障14在质量相关空间的故障误报率

Table 3 False alarm rate of fault 14 in quality-related space

方法	FAR/%
CCA	3.125
AE-CCA	7.375
VAE-OCCA	1.125

参考文献 30

1	Du K Y, Deng X G. Nonlinear industrial fault detection method based on Bayesian randomized principal component analysis[C]//2021 China Automation Congress (CAC). IEEE, 2022: 720-724.
2	Wang F, Wang Q, Nie F P, et al. Efficient tree classifiers for large scale datasets[J]. Neurocomputing, 2018, 284: 70-79.
3	Qin S J. Survey on data-driven industrial process monitoring and diagnosis[J]. Annual Reviews in Control, 2012, 36(2): 220-234.
4	Si Y B, Wang Y Q, Zhou D H. Key-performance-indicator-related process monitoring based on improved kernel partial least squares[J]. IEEE Transactions on Industrial Electronics, 2021, 68(3): 2626-2636.
5	Peng K X, Li Q Q, Zhang K, et al. Quality-related process monitoring for dynamic non-Gaussian batch process with multi-phase using a new data-driven method[J]. Neurocomputing, 2016, 214: 317-328.
6	Peng K X, Zhang K, Li G. Quality-related process monitoring based on total kernel PLS model and its industrial application[J]. Mathematical Problems in Engineering, 2013, 2013: 1-14.
7	Chen Z W, Liu C, Ding S X, et al. A just-in-time-learning-aided canonical correlation analysis method for multimode process monitoring and fault detection[J]. IEEE Transactions on Industrial Electronics, 2021, 68(6): 5259-5270.
8	Chen Z W, Ding S X, Zhang K, et al. Canonical correlation analysis-based fault detection methods with application to alumina evaporation process[J]. Control Engineering Practice, 2016, 46: 51-58.
9	Samuel R T, Cao Y. Kernel canonical variate analysis for nonlinear dynamic process monitoring [J]. IFAC-PapersOnLine, 2015, 48(8): 605-610.
10	Lee J M, Yoo C, Lee I B. Statistical process monitoring with independent component analysis[J]. Journal of Process Control, 2004, 14(5): 467-485.
11	Ge Z Q, Xie L, Kruger U, et al. Local ICA for multivariate statistical fault diagnosis in systems with unknown signal and error distributions[J]. AIChE Journal, 2012, 58(8): 2357-2372.
12	Jiang Q C, Yan X F. Non-Gaussian chemical process monitoring with adaptively weighted independent component analysis and its applications[J]. Journal of Process Control, 2013, 23(9): 1320-1331.
13	Wang Y W, Chen M Y, Zhou D H. Part mutual information based quality-related component analysis for fault detection[C]//2019 CAA Symposium on Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS). IEEE, 2020: 412-416.
14	金雨婷, 侍洪波, 吕晓龙, 等. 基于KVAE-OCCA的质量相关故障检测方法及应用[J]. 控制工程, 2022, 29(2): 348-355.
	Jin Y T, Shi H B, Lyu X L, et al. Quality-related fault detection method and application based on KVAE-OCCA[J]. Control Engineering of China, 2022, 29(2): 348-355.
15	Wang X M, Wu P, Lou S W. Quality-relevant process monitoring based on improved concurrent canonical correlation analysis[C]//2021 IEEE 10th Data Driven Control and Learning Systems Conference (DDCLS). IEEE, 2021: 565-570.
16	Song Y, Wang Y H. A fault detection method for gas pressure regulators based on improved dynamic canonical correlation analysis[C]//2021 33rd Chinese Control and Decision Conference (CCDC). IEEE, 2021: 3623-3627.
17	高学金, 何紫鹤, 高慧慧, 等. 基于联合典型变量矩阵的多阶段发酵过程质量相关故障监测[J]. 化工学报, 2022, 73(3): 1300-1314.
	Gao X J, He Z H, Gao H H, et al. Quality-related fault monitoring of multi-phase fermentation process based on joint canonical variable matrix[J]. CIESC Journal, 2022, 73(3): 1300-1314.
18	Nishimori Y, Akaho S. Learning algorithms utilizing quasi-geodesic flows on the Stiefel manifold[J]. Neurocomputing, 2005, 67: 106-135.
19	董洁, 孙瑞琪, 彭开香, 等. 自动编码器与典型相关分析方法联合驱动的工业过程质量监测[J]. 控制理论与应用, 2019, 36(9): 1493-1500.
	Dong J, Sun R Q, Peng K X, et al. Industrial process quality monitoring method and application joint-driven by automatic encoder and canonical correlation analysis method[J]. Control Theory & Applications, 2019, 36(9): 1493-1500.
20	Song B, Shi H B, Tan S, et al. Multisubspace orthogonal canonical correlation analysis for quality-related plant-wide process monitoring[J]. IEEE Transactions on Industrial Informatics, 2021, 17(9): 6368-6378.
21	Yan X Q, Hu S Z, Mao Y Q, et al. Deep multi-view learning methods: a review[J]. Neurocomputing, 2021, 448: 106-129.
22	Chen K J, Hu J, Zhang Y, et al. Fault location in power distribution systems via deep graph convolutional networks[J]. IEEE Journal on Selected Areas in Communications, 2020, 38(1): 119-131.
23	刘旭婷, 李益国, 孙栓柱, 等. 基于稀疏局部嵌入深度卷积网络的冷水机组故障诊断方法[J]. 化工学报, 2018, 69(12): 5155-5163.
	Liu X T, Li Y G, Sun S Z, et al. Fault diagnosis of chillers using sparsely local embedding deep convolutional neural network[J]. CIESC Journal, 2018, 69(12): 5155-5163.
24	赵佳璐, 任少君, 陈家乐, 等. 基于PRB-SAE算法的非线性系统建模及故障诊断方法[J]. 热能动力工程, 2022, 37(9): 197-205.
	Zhao J L, Ren S J, Chen J L, et al. Nonlinear system modeling and fault diagnosis method based on PRB-SAE algorithm[J]. Journal of Engineering for Thermal Energy and Power, 2022, 37(9): 197-205.
25	Chen H T, Chen Z W, Chai Z, et al. A single-side neural network-aided canonical correlation analysis with applications to fault diagnosis[J]. IEEE Transactions on Cybernetics, 2022, 52(9): 9454-9466.
26	Guo X P, Gao J J, Li Y. Process fault detection based on skew Gaussian distribution transformation and canonical variable analysis method[C]//2019 CAA Symposium on Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS). IEEE, 2020: 121-126.
27	Hou X X, Shen L L, Sun K, et al. Deep feature consistent variational autoencoder[C]//2017 IEEE Winter Conference on Applications of Computer Vision (WACV). IEEE, 2017: 1133-1141.
28	Harmouche J, Delpha C, Diallo D. Incipient fault detection and diagnosis based on Kullback-Leibler divergence using principal component analysis: partⅠ[J]. Signal Processing, 2014, 94: 278-287.
29	Kingma D P, Ba J. Adam: a method for stochastic optimization[EB/OL]. .
30	Zhou D H, Li G, Qin S J. Total projection to latent structures for process monitoring[J]. AIChE Journal, 2010, 56(1): 168-178.

[1]	张逸豪, 王振雷. 基于最大信息系数的分组支持向量数据描述故障检测[J]. 化工学报, 2023, 74(9): 3865-3878.
[2]	邵远哲, 赵忠盖, 刘飞. 基于共同趋势模型的非平稳过程质量相关故障检测方法[J]. 化工学报, 2023, 74(6): 2522-2537.
[3]	王雅琳, 潘雨晴, 刘晨亮. 基于GSA-LSTM动态结构特征提取的间歇过程监测方法[J]. 化工学报, 2022, 73(9): 3994-4002.
[4]	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.
[5]	雍加望, 赵倩倩, 冯能莲. 基于非线性动态模型的质子交换膜燃料电池故障诊断[J]. 化工学报, 2022, 73(9): 3983-3993.
[6]	郭金玉, 王哲, 李元. 基于核熵独立成分分析的故障检测方法[J]. 化工学报, 2022, 73(8): 3647-3658.
[7]	王琨, 侍洪波, 谭帅, 宋冰, 陶阳. 局部时差约束邻域保持嵌入算法在故障检测中的应用[J]. 化工学报, 2022, 73(7): 3109-3119.
[8]	高学金, 何紫鹤, 高慧慧, 齐咏生. 基于联合典型变量矩阵的多阶段发酵过程质量相关故障监测[J]. 化工学报, 2022, 73(3): 1300-1314.
[9]	郭金玉, 李文涛, 李元. 在线压缩KECA的自适应算法在故障检测中的应用[J]. 化工学报, 2021, 72(8): 4227-4238.
[10]	李元, 杨东昇, 赵丽颖, 张成. 层次变分高斯混合模型与主多项式分析的故障检测策略[J]. 化工学报, 2021, 72(3): 1616-1626.
[11]	朱雄卓, 张瀚文, 杨春节. 基于高斯混合模型的MWPCA高炉异常监测算法[J]. 化工学报, 2021, 72(3): 1539-1548.
[12]	王晓慧, 王延江, 邓晓刚, 张政. 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.
[13]	高学金, 刘腾飞, 徐子东, 高慧慧, 于涌川. 基于循环自动编码器的间歇过程故障监测[J]. 化工学报, 2020, 71(7): 3172-3179.
[14]	邓明月, 刘建昌, 许鹏, 谭树彬, 商亮亮. 基于KECA的非线性工业过程故障检测与诊断新方法[J]. 化工学报, 2020, 71(5): 2151-2163.
[15]	韩宇, 李俊芳, 高强, 田宇, 禹国刚. 基于故障判别增强KECA算法的故障检测[J]. 化工学报, 2020, 71(3): 1254-1263.