A chemical process quality-related fault detection method based on twin-space parallel regression

doi:10.11949/0438-1157.20230877

Abstract

Abstract:

Neighborhood preserving embedding (NPE) is a commonly used unsupervised learning method and has been widely applied in fault detection. However, the features extracted by NPE cannot explain the relationship between input and output data, which limits its application in quality-related fault detection in chemical processes. Moreover, NPE ignores the representation of dynamic information while extracting data manifold structure. To address these issues, this paper proposes a quality-related fault detection method called twin-space parallel regression (TSPR) based on NPE and slow feature analysis (SFA) algorithm, which can simultaneously extract manifold features and velocity information. First, the original process space is divided into serial correlated subspaces and serial correlated subspaces based on mutual information strategy to deal with the differences in time series correlation caused by sensors. Secondly, the proposed neighborhood preserving-slow feature embedding (NP-SFE) algorithm and NPE algorithm are used to extract the effective structural features in two subspaces, and the regression relationship between process variables and quality variables is constructed by using least square regression in both feature subspaces to characterize the change trend of process variables and quality variables. Then, the covariance matrix of the regression coefficients is decomposed to obtain the quality-related subspace and quality-unrelated subspace, and monitoring statistics and control limits are established and estimated respectively. Finally, the proposed method is tested and verified on typical cases to illustrate the effectiveness and rationality of the proposed method.

Key words: neighborhood preserving embedding, slow feature analysis, quality-related, least square regression, fault detection

摘要：

邻域保持嵌入（neighborhood preserving embedding， NPE）是一种常用的无监督学习方法，在故障检测领域得到了广泛应用。由于NPE提取的数据特征无法解释输入数据和输出数据之间的关系，因此在化工过程质量相关故障检测方面存在局限性。另外，NPE在提取数据流形结构时忽略了动态信息的表征。为了解决上述问题，基于NPE和慢特征分析（slow feature analysis，SFA）算法提出了一种名为双子空间并行回归（twin-space parallel regression，TSPR）的质量相关故障检测方法，该方法能够同时提取数据的流形特征和变化速度信息。首先，通过基于互信息的策略将原始过程空间分为序列相关子空间和序列无关子空间，以应对变量在时间序列相关性的差异。其次，在两个子空间中分别应用提出的邻域保持-慢特征嵌入算法（neighborhood preserving-slow feature embedding regression，NP-SFE）和NPE算法提取数据的有效结构特征，并同时用最小二乘回归在两个特征子空间中构建过程变量与质量变量的回归关系。随后，通过对回归系数的协方差矩阵分解，得到质量相关子空间和质量无关子空间，进而在相应子空间建立统计量并估计其控制限。最后，将所提方法在典型案例上进行测试验证，以说明所提方法的有效性和合理性。

关键词: 邻域保持嵌入, 慢特征分析, 质量相关, 最小二乘回归, 故障检测

CLC Number:

TP 277

Bing SONG, Tao GUO, Hongbo SHI, Shuai TAN, Yang TAO, Yuyang MA. A chemical process quality-related fault detection method based on twin-space parallel regression[J]. CIESC Journal, 2023, 74(11): 4600-4610.

宋冰, 郭涛, 侍洪波, 谭帅, 陶阳, 马浴阳. 基于双子空间并行回归的化工过程质量相关故障检测方法[J]. 化工学报, 2023, 74(11): 4600-4610.

Figures/Tables 12

Fig.1 Change in content of component G in fault 14

Table 1 Fault alarm rate in quality-related space

CCA,T²	NPER,T²	TSPR, $B I C T 2$
2.08%	3.02%	0.63%

Table 1 Fault alarm rate in quality-related space

CCA,T²	NPER,T²	TSPR, $B I C T 2$
2.08%	3.02%	0.63%

Table 2 Fault detection rate of fault 14 by PLS，CCA，NPER and TSPR method

PLS,T²	CCA,T²	NPER,T²	TSPR, $B I C T 2$
100%	100%	100%	100%

Table 2 Fault detection rate of fault 14 by PLS，CCA，NPER and TSPR method

PLS,T²	CCA,T²	NPER,T²	TSPR, $B I C T 2$
100%	100%	100%	100%

Fig.2 Detection results of fault 14 by TSPR,PLS,CCA and NPER method

Fig.3 Detection results of fault 14 by TSPR method in each subspace

Fig.4 Change in content of component G in fault 2

Fig.5 Detection results of fault 2 by TSPR,PLS,CCA and NPER method

Table 3 Fault detection rate in quality-related space

PLS,T²	CCA,T²	NPER,T²	TSPR, $B I C T 2$
96.75%	91.63%	95.38%	97.25%

Table 3 Fault detection rate in quality-related space

PLS,T²	CCA,T²	NPER,T²	TSPR, $B I C T 2$
96.75%	91.63%	95.38%	97.25%

Table 4 Fault detection rate of fault 2 by PLS，CCA，NPER and TSPR method

PLS,T²	CCA,T²	NPER,T²	TSPR, $B I C T 2$
98.63%	98.25%	98.25%	98.25%

Table 4 Fault detection rate of fault 2 by PLS，CCA，NPER and TSPR method

PLS,T²	CCA,T²	NPER,T²	TSPR, $B I C T 2$
98.63%	98.25%	98.25%	98.25%

Fig.6 Detection results of fault 2 by TSPR method in each subspace

Fig.7 Change in content of component G in fault 1

Fig.8 Detection results of fault 1 by TSPR,PLS,CCA and NPER method

References 32

1	彭开香, 张传放, 马亮, 等. 面向系统层级的复杂工业过程全息故障诊断[J]. 化工学报, 2019, 70(2): 590-598.
	Peng K X, Zhang C F, Ma L, et al. System-levels-based holographic fault diagnosis for complex industrial processes[J]. CIESC Journal, 2019, 70(2): 590-598.
2	Zhu J L, Ge Z Q, Song Z H, et al. Large-scale plant-wide process modeling and hierarchical monitoring: a distributed Bayesian network approach[J]. Journal of Process Control, 2018, 65: 91-106.
3	赵春晖, 余万科, 高福荣. 非平稳间歇过程数据解析与状态监控: 回顾与展望[J]. 自动化学报, 2020, 46(10): 2072-2091.
	Zhao C H, Yu W K, Gao F R. Data analytics and condition monitoring methods for nonstationary batch processes: current status and future[J]. Acta Automatica Sinica, 2020, 46(10): 2072-2091.
4	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.
5	Dong J, Jiang L Z, Zhang C, et al. A novel quality-related incipient fault detection method based on canonical variate analysis and kullback-leibler divergence for large-scale industrial processes[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-10.
6	周东华, 纪洪泉, 何潇. 高速列车信息控制系统的故障诊断技术[J]. 自动化学报, 2018, 44(7): 1153-1164.
	Zhou D H, Ji H Q, He X. Fault diagnosis techniques for the information control system of high-speed trains[J]. Acta Automatica Sinica, 2018, 44(7): 1153-1164.
7	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.
8	张淑美, 王福利, 谭帅, 等. 多模态过程的全自动离线模态识别方法[J]. 自动化学报, 2016, 42(1): 60-80.
	Zhang S M, Wang F L, Tan S, et al. A fully automatic offline mode identification method for multi-mode processes[J]. Acta Automatica Sinica, 2016, 42(1): 60-80.
9	Ge Z Q, Song Z H. Distributed PCA model for plant-wide process monitoring[J]. Industrial & Engineering Chemistry Research, 2013, 52(5): 1947-1957.
10	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.
11	Zheng Y, Liu Z W, Yang W D, et al. Parallel projection to latent structures for quality-relevant process monitoring[J]. Journal of the Taiwan Institute of Chemical Engineers, 2017, 80: 76-84.
12	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.
13	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.
14	马玉鑫. 流程工业过程故障检测的特征提取方法研究[D]. 上海: 华东理工大学, 2015.
	Ma Y X. Research on feature extraction method of process fault detection in process industry[D]. Shanghai: East China University of Science and Technology, 2015.
15	He X F, Niyogi P. Locality preserving projections[J]. Advances in Neural Information Processing Systems, 2003, 16(16): 153-160.
16	Roweis S T, Saul L K. Nonlinear dimensionality reduction by locally linear embedding[J]. Science, 2000, 290(5500): 2323-2326.
17	Belkin M, Niyogi P. Laplacian eigenmaps and spectral techniques for embedding and clustering[C]//Proceedings of the 14th International Conference on Neural Information Processing Systems: Natural and Synthetic. New York: ACM, 2001: 585-591.
18	He X F, Cai D, Yan S C, et al. Neighborhood preserving embedding[C]//Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1. Piscataway, NJ: IEEE, 2005: 1208-1213.
19	Miao A M, Ge Z Q, Song Z H, et al. Nonlocal structure constrained neighborhood preserving embedding model and its application for fault detection[J]. Chemometrics and Intelligent Laboratory Systems, 2015, 142: 184-196.
20	Song B, Ma Y X, Shi H B. Multimode process monitoring using improved dynamic neighborhood preserving embedding[J]. Chemometrics and Intelligent Laboratory Systems, 2014, 135: 17-30.
21	张妮. 基于流形特征提取的化工过程故障诊断方法研究[D]. 东营: 中国石油大学(华东), 2013.
	Zhang N. Research on fault diagnosis method of chemical process based on manifold feature extraction[D]. Dongying: China University of Petroleum, 2013.
22	苗爱敏, 葛志强, 宋执环, 等. 基于时序扩展的邻域保持嵌入算法及其在故障检测中的应用[J]. 华东理工大学学报(自然科学版), 2014, 40(2): 218-224.
	Miao A M, Ge Z Q, Song Z H, et al. Neighborhood preserving embedding based on temporal extension and its application in fault detection[J]. Journal of East China University of Science and Technology (Natural Science Edition), 2014, 40(2): 218-224.
23	王琨, 侍洪波, 谭帅, 等. 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.
	Wang K, Shi H B, Tan S, et al. Application of local time difference constraint neighborhood preserving embedding algorithm in fault detection[J]. CIESC Journal, 2022, 73(7): 3109-3119.
24	Song B, Song Y M, Jin Y T, et al. Plant-wide process fine-scale monitoring via distributed static magnitude-dynamic difference[J]. IEEE Transactions on Industrial Informatics, 2023, 19(11): 10864-10872.
25	Song B, Shi H B, Tan S, et al. Serial correlated-uncorrelated concurrent space method for process monitoring[J]. Journal of Process Control, 2021, 105: 292-301.
26	郑嘉乐. 复杂工业过程动态表征学习和动静协同的变工况识别方法[D]. 杭州: 浙江大学, 2022.
	Zheng J L. Dynamic representation learning and dynamic-static collaborative variable working condition identification method for complex industrial processes[D]. Hangzhou: Zhejiang University, 2022.
27	Zhang H Y, Li C D, Wei Q L, et al. Fault detection and diagnosis of the air handling unit via combining the feature sparse representation based dynamic SFA and the LSTM network[J]. Energy and Buildings, 2022, 269: 112241.
28	Guo F H, Shang C, Huang B, et al. Monitoring of operating point and process dynamics via probabilistic slow feature analysis[J]. Chemometrics and Intelligent Laboratory Systems, 2016, 151: 115-125.
29	Zhang H Y, Tian X M, Deng X G, et al. Batch process fault detection and identification based on discriminant global preserving kernel slow feature analysis[J]. ISA Transactions, 2018, 79: 108-126.
30	Zhang S M, Zhao C H, Huang B. Simultaneous static and dynamic analysis for fine-scale identification of process operation statuses[J]. IEEE Transactions on Industrial Informatics, 2019, 15(9): 5320-5329.
31	周乐, 沈程凯, 吴超, 等. 深度融合特征提取网络及其在化工过程软测量中的应用[J]. 化工学报, 2022, 73(7): 3156-3165.
	Zhou L, Shen C K, Wu C, et al. Deep fusion feature extraction network and its application in chemical process soft sensing[J]. CIESC Journal, 2022, 73(7): 3156-3165.
32	Zhu Q Q, Liu Q, Qin S J. Concurrent quality and process monitoring with canonical correlation analysis[J]. Journal of Process Control, 2017, 60: 95-103.

[1]	Yihao ZHANG, Zhenlei WANG. Fault detection using grouped support vector data description based on maximum information coefficient [J]. CIESC Journal, 2023, 74(9): 3865-3878.
[2]	Yuanzhe SHAO, Zhonggai ZHAO, Fei LIU. Quality-related non-stationary process fault detection method by common trends model [J]. CIESC Journal, 2023, 74(6): 2522-2537.
[3]	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.
[4]	Jiawang YONG, Qianqian ZHAO, Nenglian FENG. Fault diagnosis of proton exchange membrane fuel cell based on nonlinear dynamic model [J]. CIESC Journal, 2022, 73(9): 3983-3993.
[5]	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.
[6]	Jinyu GUO, Zhe WANG, Yuan LI. Fault detection method based on kernel entropy independent component analysis [J]. CIESC Journal, 2022, 73(8): 3647-3658.
[7]	Kun WANG, Hongbo SHI, Shuai TAN, Bing SONG, Yang TAO. Local time difference constrained neighborhood preserving embedding algorithm for fault detection [J]. CIESC Journal, 2022, 73(7): 3109-3119.
[8]	Jinyu GUO, Wentao LI, Yuan LI. Application of adaptive algorithm of online reduced KECA in fault detection [J]. CIESC Journal, 2021, 72(8): 4227-4238.
[9]	LI Yuan, YANG Dongsheng, ZHAO Liying, ZHANG Cheng. Fault detection using hierarchical variational Gaussian mixture model and principal polynomial analysis [J]. CIESC Journal, 2021, 72(3): 1616-1626.
[10]	Xiaohui WANG, Yanjiang WANG, Xiaogang DENG, Zheng ZHANG. Industrial process fault detection using weighted deep support vector data description [J]. CIESC Journal, 2021, 72(11): 5707-5716.
[11]	Mingyue DENG, Jianchang LIU, Peng XU, Shubin TAN, Liangliang SHANG. New fault detection and diagnosis strategy for nonlinear industrial process based on KECA [J]. CIESC Journal, 2020, 71(5): 2151-2163.
[12]	Yu HAN, Junfang LI, Qiang GAO, Yu TIAN, Guogang YU. Fault detection based on fault discrimination enhanced kernel entropy component analysis algorithm [J]. CIESC Journal, 2020, 71(3): 1254-1263.
[13]	XU Jing,WANG Zhenlei,WANG Xin. Fault detection for chemical process based on nonlinear dynamic global-local preserving projections [J]. CIESC Journal, 2020, 71(12): 5655-5663.
[14]	Zhongjian SUN,Bo YANG,Chu QI,Hongguang LI. An extended logical analysis of data approach to fault detections of industrial hybrid systems [J]. CIESC Journal, 2020, 71(11): 5237-5245.
[15]	Lei YU, Xiaogang DENG, Yuping CAO, Kaiqi LU. Fault detection method of unequal-length batch process based on VGDTW-MCVA [J]. CIESC Journal, 2019, 70(9): 3441-3448.