kNN Fault Detection Based on Reconstruction Error and Multi-block Modeling Strategy

doi:10.16182/j.issn1004731x.joss.21-0689

Abstract

Abstract:

For the fault monitoring algorithm based on k-nearest neighbor (kNN), the abnormal information that caused the fault is easy to be overwhelmed by the normal operating condition information, which leads to the problem of untimely fault detection and low alarm rate. A kNN fault monitoring method based on reconstruction error is proposed using auto-encoder and multi-block modeling strategy. The method uses the normal working condition data set to train the auto-encoder model, and extracts the reconstruction error based on the model to solve the problem that abnormal information is easy to be overwhelmed. Further considering the fault characteristics such as micro-offset and oscillation, a multi-block modeling strategy is adopted to calculate statistics for each sub-block and merge the detection. Through a numerical example and Tennessee-Eastman (Tennessee-Eastman, TE) process simulation and analysis, the results verify the effectiveness of the proposed method and the improvement of monitoring performance.

Key words: k-nearest neighbor(kNN), reconstruction error, fault detection, information extraction, multi-block modeling

CLC Number:

TP 277

Jing Zheng, Weili Xiong, Xiaodong Wu. kNN Fault Detection Based on Reconstruction Error and Multi-block Modeling Strategy[J]. Journal of System Simulation, 2023, 35(1): 95-109.

Figures/Tables 14

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Table 1

Fig. 7

Table 2

Fig. 8

Fig. 9

Fig. 10

Fig. 11

Table 3

References 28

1	刘强, 卓洁, 郎自强, 等. 数据驱动的工业过程运行监控与自优化研究展望[J]. 自动化学报, 2018, 44(11): 1944-1956.
	Liu Qiang, Zhuo Jie, Lang Ziqiang, et al. Perspectives on Data-driven Operation Monitoring and Self-optimization of Industrial Processes[J]. Acta Automatica Sinica, 2018, 44(11): 1944-1956.
2	Ge Z, Song Z. Review of Recent Research on Data-based Process Monitoring[J]. Industrial & Engineering Chemistry Research (S0888-5885), 2013, 52(10): 3534-3562.
3	彭开香, 马亮, 张凯. 复杂工业过程质量相关的故障检测与诊断技术综述[J]. 自动化学报, 2017, 43(3): 349-365.
	Peng Kaixiang, Ma Liang, Zhang Kai. Review of Quality-related Fault Detection and Diagnosis Techniques for Complex Industrial Processes[J]. Acta Automatica Sinica, 2017, 43(3): 349-365.
4	Jiang Q, Yan X, Zhao W. Fault Detection and Diagnosis in Chemical Processes Using Sensitive Principal Component Analysis[J]. Industrial & Engineering Chemistry Research(S0888-5885), 2013, 52(4): 1635-1644.
5	Yin S, Xie X, Sun W. A Nonlinear Process Monitoring Approach with Locally Weighted Learning of Available Data[J]. IEEE Transactions on Industrial Electronics(S0278-0046), 2016, 64(2): 1507-1516.
6	王文标, 田志远, 汪思源, 等. 交叉分段PCA在锅炉故障诊断中的应用[J]. 信息与控制, 2020, 49(4): 507-512.
	Wang Wenbiao, Tian Zhiyuan, Wang Siyuan, et al. Optimal Operation Method Based on Cross and Piecewise PCA for Industrial Boilers[J]. Information and Control, 2020, 49(4): 507-512.
7	Wang Y, Si Y, Huang B, et al. Survey on the Theoretical Research and Engineering Applications of Multivariate Statistics Process Monitoring Algorithms: 2008-2017[J]. The Canadian Journal of Chemical Engineering(S0008-4034), 2018, 96(10): 2073-2085.
8	Zhou Z, Wen C, Yang C. Fault Detection using Random Projections and k-nearest Neighbor Nule for Semiconductor Manufacturing Processes[J]. IEEE Transactions on Semiconductor Manufacturing (S0894-6507), 2014, 28(1): 70-79.
9	MacGregor J F, Jaeckle C, Kiparissides C, et al. Process Monitoring and Diagnosis by Multiblock PLS Methods[J]. AIChE Journal (S0001-1541), 1994, 40(5): 826-838.
10	Ge Z, Song Z. Distributed PCA Model for Plant-wide Process Monitoring[J]. Industrial & Engineering Chemistry Research (S0888-5885), 2013, 52(5): 1947-1957.
11	Ge Z, Zhang M, Song Z. Nonlinear Process Monitoring Based on Linear Subspace and Bayesian Inference[J]. Journal of Process Control (S0959-1524), 2010, 20(5): 676-688.
12	Jiang Q, Wang B, Yan X. Multiblock Independent Component Analysis Integrated with Hellinger Distance and Bayesian Inference for Non-Gaussian Plant-wide Process Monitoring[J]. Industrial & Engineering Chemistry Research (S0888-5885), 2015, 54(9): 2497-2508.
13	王振雷, 江伟, 王昕. 基于多块MICA-PCA的全流程过程监控方法[J]. 控制与决策, 2018, 33(2): 269-274.
	Wang Zhenlei, Jiang Wei, Wang Xin. Plant-wide Process Monitoring Based on Multiblock MICA-PCA[J]. Control and Decision, 2018, 33(2): 269-274.
14	顾炳斌, 熊伟丽. 基于多块信息提取的PCA故障诊断方法[J]. 化工学报, 2019, 70(2): 316-329.
	Gu Bingbin, Xiong Weili. Fault Diagnosis Based on PCA Method with Multi-block Information Extraction[J]. CIESC Journal, 2019, 70(2): 316-329.
15	He P Q, Wang J. Fault Detection Using the k-nearest Neighbor Rule for Semiconductor Manufacturing Processes[J].IEEE Transactions on Semiconductor Manufacturing (S0894-6507), 2007, 20(4): 345-354.
16	张成, 郭青秀, 冯立伟, 等. 基于局部保持投影-加权k近邻规则的多模态间歇过程故障检测策略[J]. 控制理论与应用, 2019, 36(10): 1682-1689.
	Zhang Chen, Guo Qingxiu, Feng Liwei, et al. Fault Detection Strategy Based on Locality Preserving Projections-weighted k Nearest Neighbors in Multimodal Batch Processes[J]. Control Theory & Applications, 2019, 36(10): 1682-1689.
17	郭金玉, 刘玉超, 李元. 基于局部相对概率密度kNN的多模态过程故障检测[J]. 高校化学工程学报, 2019, 33(1): 159-166.
	Guo Jinyu, Liu Yuchao, Li Yuan. Multimodal Process Fault Detection Based on Local Relative Probability Density kNN[J]. Journal of Chemical Engineering of Chinese Universities, 2019, 33(1): 159-166.
18	刘腾飞. 基于深层自编码器的发酵过程故障监测[D]. 北京: 北京工业大学, 2020: 54-64.
	Liu Tengfei. Fault Monitoring of Fermentation Process Based on Deep Auto-encoder[D]. Beijing: Beijing University of Technology, 2020: 54-64.
19	张成, 高宪文, 徐涛, 等. 基于独立元的k近邻故障检测策略[J]. 控制理论与应用, 2018, 35(6): 805-812.
	Zhang Chen, Gao Xianwen, Xu Tao, et al. Fault Detection Strategy of Independent Component-Based k Nearest Neighbor Rule[J]. Control Theory & Applications, 2018, 35(6): 805-812.
20	马贺贺, 胡益, 侍洪波. 基于距离空间统计量分析的多模态过程无监督故障检测[J]. 化工学报, 2012, 63(3): 873-880.
	Ma Hehe, Hu Yi, Shi Hongbo. Unsupervised Fault Detection for Multimode Processes using Distance Space Statistics Analysis[J]. CIESC Journal, 2012, 63(3): 873-880.
21	Zhu W, Sun W, Romagnoli J. Adaptive k-Nearest-Neighbor Method for Process Monitoring[J]. Industrial & Engineering Chemistry Research (S0888-5885), 2018, 57(7): 2574-2586.
22	Zhou Z, Wen C, Yang C. Fault Isolation Based on k-Nearest Neighbor Rule for Industrial Processes[J]. IEEE Transactions on Industrial Electronics (S0278-0046), 2016, 63(4): 2578-2586.
23	Shen B, Ge Z. Supervised Nonlinear Dynamic System for Soft Sensor Application Aided by Variational Auto-encoder[J]. IEEE Transactions on Instrumentation and Measurement (S0018-9456), 2020, 69(9): 6132-6142.
24	Chen Z, Li W. Multisensor Feature Fusion for Bearing Fault Diagnosis Using Sparse Autoencoder and Deep Belief Network[J]. IEEE Transactions on Instrumentation and Measurement (S0018-9456), 2017, 66(7): 1693-1702.
25	Ge Z, Song Z. Multimode Process Monitoring Based on Bayesian Method[J]. Journal of Chemometrics(S0886-9383), 2009, 23(12): 636-650.
26	Yin S, Ding S X, Haghani A, 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 (S0959-1524), 2012, 22(9): 1567-1581.
27	郭金玉, 王鑫, 李元. 基于集成核局部保持投影算法的故障检测[J]. 信息与控制, 2018, 47(2): 191-199.
	Guo Jinyu, Wang Xin, Li Yuan. Fault Detection Based on Ensemble Kernel Locality Preserving Projections[J]. Information and Control, 2018, 47(2): 191-199.
28	Li N, Yang Y. Ensemble Kernel Principal Component Analysis for Improved Nonlinear Process Monitoring[J]. Industrial & Engineering Chemistry Research (S0888-5885), 2015, 54(1): 318-329.

方法	报警率	误报率
kNN	0.090	0
RE_kNN	0.872	0.04
RE_MBIkNN	0.981	0

故障编码	以原始数据为观测信息				以重构误差为观测信息
故障编码	子块1	子块2	子块3	BIC	子块1	子块2	子块3	BIC
1	0.996	0.998	0.001	0.995	0.996	0.995	0.976	0.995
2	0.983	0.988	0.000	0.984	0.986	0.989	0.978	0.984
3	0.013	0.326	0.001	0.054	0.092	0.195	0.119	0.150
4	0.975	0.999	0.005	0.999	0.999	0.999	0.117	0.999
5	0.260	0.809	0.002	0.381	0.999	0.999	0.336	0.999
6	1	0.999	0.004	0.999	0.999	0.999	0.911	0.999
7	1	0.999	0.007	0.999	0.999	0.999	0.554	0.999
8	0.976	0.993	0.006	0.979	0.979	0.984	0.969	0.976
9	0.020	0.290	0.005	0.051	0.084	0.162	0.135	0.132
10	0.418	0.815	0.000	0.618	0.737	0.860	0.393	0.848
11	0.683	0.941	0.047	0.868	0.758	0.930	0.553	0.913
12	0.989	1.000	0.199	0.993	0.995	1.000	0.980	1.000
13	0.946	0.963	0.031	0.951	0.950	0.958	0.944	0.953
14	1	0.878	0.998	0.998	0.999	0.998	0.999	0.999
15	0.029	0.326	0.004	0.105	0.101	0.220	0.119	0.181
16	0.289	0.820	0.002	0.549	0.729	0.886	0.291	0.893
17	0.919	0.974	0.149	0.966	0.946	0.975	0.864	0.976
18	0.896	0.940	0.041	0.899	0.906	0.914	0.898	0.913
19	0.099	0.412	0.114	0.135	0.669	0.644	0.697	0.779
20	0.495	0.848	0.039	0.642	0.669	0.843	0.452	0.810
21	0.425	0.699	0.001	0.566	0.576	0.664	0.583	0.639
平均报警率	0.639	0.810	0.079	0.701	0.770	0.820	0.613	0.816
平均误报率	0.006	0.159	0.003	0.015	0.046	0.079	0.067	0.059

故障编码	报警率
故障编码	PCA	SVDD	kNN	MBIkNN	RE_kNN	RE_MBIkNN
1	0.999	0.993	0.995	0.995	0.996	0.995
2	0.985	0.983	0.983	0.984	0.986	0.984
3	0.032	0.035	0.013	0.054	0.092	0.150
4	1	0.790	0.974	0.999	0.999	0.999
5	0.253	0.275	0.260	0.381	0.999	0.999
6	1	1	1	0.999	0.999	0.999
7	1	1	1	0.999	0.999	0.999
8	0.976	0.975	0.977	0.979	0.979	0.976
9	0.058	0.029	0.021	0.051	0.084	0.132
10	0.299	0.446	0.418	0.618	0.737	0.848
11	0.754	0.598	0.683	0.868	0.758	0.913
12	0.984	0.987	0.988	0.993	0.995	1
13	0.953	0.944	0.946	0.951	0.950	0.953
14	1	1	1	0.998	0.999	0.999
15	0.037	0.063	0.029	0.105	0.101	0.181
16	0.274	0.284	0.279	0.549	0.729	0.893
17	0.952	0.876	0.919	0.966	0.946	0.976
18	0.901	0.898	0.898	0.899	0.906	0.913
19	0.225	0.045	0.089	0.135	0.669	0.779
20	0.498	0.458	0.485	0.642	0.669	0.810
21	0.513	0.421	0.426	0.566	0.576	0.639
平均报警率	0.652	0.624	0.637	0.701	0.770	0.816
平均误报率	0.004	1.756	0.006	0.015	0.046	0.059

[1]	Xiangping Wu, Lijun Ping, Dongshi Xu. A Visual Analytics of Urban Traffic Events Using Social Media Data [J]. Journal of System Simulation, 2022, 34(5): 1140-1151.
[2]	Ningke Sun, Yan Wang, Zhicheng Ji. Anomaly Detection Method of Electrical Power Consumption Based on Deep Autoencoder [J]. Journal of System Simulation, 2022, 34(12): 2557-2565.
[3]	Zhang Lixia, Feng Fuzhou, Liu Xiangbo, Wang Min. Research on Fault Detection and isolation method for Typical Hydraulic System [J]. Journal of System Simulation, 2018, 30(5): 1818-1825.
[4]	Niu Yuguang, Wang Shilin, Lin Zhongwei, Li Xiaoming. Fault Detection Based on GPNMF for Industrial Process [J]. Journal of System Simulation, 2018, 30(2): 521-532.
[5]	Li Hailin, Zhang Bin, Wu Dewei, Lu Hu. A Way of Integrated Navigation Fault Detection of Near Space Hypersonic Cruising Aircraft [J]. Journal of System Simulation, 2017, 29(8): 1809-1814.
[6]	Li Gang, Xu Pengcheng, Han Longmei. Fault Spatial-temporal Detecting and Diagnosis for Power Grid Based on Wavelet Analysis [J]. Journal of System Simulation, 2015, 27(12): 3018-3024.