基于滑动窗和多块卷积自编码器的故障检测

doi:10.16182/j.issn1004731x.joss.22-1059

摘要/Abstract

摘要：

为了进一步提升故障检测性能，充分挖掘时序和隐含特征信息，提出一种基于卷积自编码器的故障检测方法。在对原始信息集进行建模的基础上增加了对累计信息与变化率信息的建模，以增强对隐含信息的挖掘；对重构的3个信息集进行滑动窗采样，基于卷积自编码器进行时序特征提取和建模；将卷积自编码器的决策结果进行贝叶斯融合得到统计量，并用核密度估计的方法确定控制限从而进行故障检测。将该方法进行数值仿真并应用于TE过程，仿真结果验证了所提方法的有效性和检测性能。

关键词: 故障检测, 卷积自编码器, 多块建模, 滑动窗, 贝叶斯融合

Abstract:

In order to further improve the fault detection performance and fully mine the timing and hidden feature information, a fault detection method based on convolutional auto encoder is proposed. On the basis of modeling the original information set, the modeling of cumulative information and rate of change information is added to enhance the mining of implicit information; The three reconstructed information sets are sampled by sliding windows, and time series feature extraction and modeling are performed based on convolutional auto encoders.Bayesian fusion of the decision results of the convolutional auto encoder is performed to obtain the statistics, and the control limit is determined by the method of kernel density estimation for fault detection. The method is numerically simulated and applied in TE process, and the simulation results confirm the effectiveness and detection performance.

Key words: fault detection, convolutional autoencoder, multiblock modeling, sliding window, Bayesian fusion

中图分类号:

TP277

牟建鹏,熊伟丽 . 基于滑动窗和多块卷积自编码器的故障检测[J]. 系统仿真学报, 2024, 36(2): 423-435.

Mou Jianpeng,Xiong Weili . Fault Detection Based on Sliding Window and Multiblock Convolutional Autoencoders[J]. Journal of System Simulation, 2024, 36(2): 423-435.

图/表 15

图1

图2

图3

图4

表1

图5

图6

表2

图7

图8

图9

图10

图11

图12

表3

参考文献 29

1	李俨, 杨晨. 基于相对输出信息的多智能体系统分布式故障检测[J]. 控制与决策, 2023, 38(7): 1901-1908.
	Li Yan, Yang Chen. Distributed Fault Detection of Multi-agent System Based on Relative Output Information[J]. Control and Decision, 2023, 38(7): 1901-1908.
2	周东华, 刘洋, 何潇. 闭环系统故障诊断技术综述[J]. 自动化学报, 2013, 39(11): 1933-1943.
	Zhou Donghua, Liu Yang, He Xiao. Review on Fault Diagnosis Techniques for Closed-loop Systems[J]. Acta Automatica Sinica, 2013, 39(11): 1933-1943.
3	高学金, 程琨, 韩华云, 等. 基于中心损失的条件生成式对抗网络的冷水机组故障诊断[J]. 化工学报, 2022, 73(9): 3950-3962.
	Gao Xuejin, Cheng Kun, Han Huayun, et al. Fault Diagnosis of Chillers Using Central Loss Conditional Generative Adversarial Network[J]. CIESC Journal, 2022, 73(9): 3950-3962.
4	Ge Zhiqiang, Song Zhihuan, Gao Furong. Review of Recent Research on Data-based Process Monitoring[J]. Industrial & Engineering Chemistry Research, 2013, 52(10): 3543-3562.
5	魏赟, 李栋. 结合改进卷积神经网络与自编码器的表情识别[J]. 小型微型计算机系统, 2022, 43(2): 387-392.
	Wei Yun, Li Dong. Expression Recognition Based on Improved Convolutional Neural Networks and Self-encoder[J]. Journal of Chinese Computer Systems, 2022, 43(2): 387-392.
6	刘家瑞, 杨国田, 王孝伟. 基于孪生深度神经网络的风电机组故障诊断方法[J]. 系统仿真学报, 2022, 34(11): 2348-2358.
	Liu Jiarui, Yang Guotian, Wang Xiaowei. A Wind Turbine Fault Diagnosis Method Based on Siamese Deep Neural Network[J]. Journal of System Simulation, 2022, 34(11): 2348-2358.
7	杜先君, 巩彬, 余萍, 等. 基于CBAM-CNN的模拟电路故障诊断[J]. 控制与决策, 2022, 37(10): 2609-2618.
	Du Xianjun, Gong Bin, Yu Ping, et al. CBAM-CNN Based Analog Circuit Fault Diagnosis[J]. Control and Decision, 2022, 37(10): 2609-2618.
8	张鹏, 束小曼, 厉雪衣, 等. 基于LSTM的交流电机系统故障诊断方法研究[J]. 电机与控制学报, 2022, 26(3): 109-116.
	Zhang Peng, Shu Xiaoman, Li Xueyi, et al. LSTM-based Fault Diagnosis of AC Electric Machine System[J]. Electric Machines and Control, 2022, 26(3): 109-116.
9	徐林, 郑晓彤, 付博, 等. 基于改进GAN算法的电机轴承故障诊断方法[J]. 东北大学学报(自然科学版), 2019, 40(12): 1679-1684.
	Xu Lin, Zheng Xiaotong, Fu Bo, et al. Fault Diagnosis Method of Motor Bearing Based on Improved GAN Algorithm[J]. Journal of Northeastern University(Natural Science), 2019, 40(12): 1679-1684.
10	刘兴, 余建波. 注意力卷积GRU自编码器及其在工业过程监控的应用[J]. 浙江大学学报(工学版), 2021, 55(9): 1643-1651, 1659.
	Liu Xing, Yu Jianbo. Attention Convolutional GRU-based Autoencoder and Its Application in Industrial Process Monitoring[J]. Journal of Zhejiang University(Engineering Science), 2021, 55(9): 1643-1651, 1659.
11	Rai Khushwant, Hojatpanah Farnam, Firouz Badrkhani Ajaei, et al. Deep Learning for High-impedance Fault Detection: Convolutional Autoencoders[J]. Energies, 2021, 14(12): 3623.
12	Chen Kunjin, Hu Jun, He Jinliang. Detection and Classification of Transmission Line Faults Based on Unsupervised Feature Learning and Convolutional Sparse Autoencoder[J]. IEEE Transactions on Smart Grid, 2018, 9(3): 1748-1758.
13	Yu Jianbo, Liu Xing, Ye L. Convolutional Long Short-term Memory Autoencoder-based Feature Learning for Fault Detection in Industrial Processes[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-15.
14	孙海蓉, 潘子杰, 晏勇. 基于深度卷积自编码网络的小样本光伏热斑识别与定位[J]. 华北电力大学学报(自然科学版), 2021, 48(4): 91-98.
	Sun Hairong, Pan Zijie, Yan Yong. Identification and Location of Photovoltaic Hot Spots with Small Samples Based on Deep Convolutional Autoencoder Network[J]. Journal of North China Electric Power University(Natural Science Edition), 2021, 48(4): 91-98.
15	赵小强, 张潇潇. 基于滑动窗的CVA故障诊断算法[J]. 兰州理工大学学报, 2015, 41(3): 91-95.
	Zhao Xiaoqiang, Zhang Xiaoxiao. CVA Algorithm for Fault Diagnosis Based on Moving Window[J]. Journal of Lanzhou University of Technology, 2015, 41(3): 91-95.
16	肖雄, 王健翔, 张勇军, 等. 一种用于轴承故障诊断的二维卷积神经网络优化方法[J]. 中国电机工程学报, 2019, 39(15): 4558-4567.
	Xiao Xiong, Wang Jianxiang, Zhang Yongjun, et al. A Two-dimensional Convolutional Neural Network Optimization Method for Bearing Fault Diagnosis[J]. Proceedings of the CSEE, 2019, 39(15): 4558-4567.
17	陈含智, 孙蕊, 邱明, 等. 基于自适应噪声方差的卫星定位故障检测法[J]. 北京航空航天大学学报, 2023, 49(2): 406-421.
	Chen Hanzhi, Sun Rui, Qiu Ming, et al. An Adaptive Noise Variance Based Fault Detection Algorithm for GNSS Positioning[J]. Journal of Beijing University of Aeronautics and Astronautics, 2023, 49(2): 406-421.
18	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): 736-749.
	Gu Bingbin, Xiong Weili. Fault Diagnosis Based on PCA Method with Multi-block Information Extraction[J]. CIESC Journal, 2019, 70(2): 736-749.
19	Ge Zhiqiang, Song Zhihuan. Distributed PCA Model for Plant-wide Process Monitoring[J]. Industrial & Engineering Chemistry Research, 2013, 52(5): 1947-1957.
20	何永建. 基于多块PLS的故障检测方法研究[D]. 沈阳: 东北大学, 2014.
	He Yongjian. Research on Fault Detection Method Based on Multiblock PLS[D]. Shenyang: Northeastern University, 2014.
21	朱家桢, 何雨旻, 侍洪波. 基于多块卷积自编码器的故障检测方法[C]//第32届中国过程控制会议(CPCC2021)论文集. 太原: 中国自动化学会, 2021: 1777.
22	郑静, 熊伟丽, 吴晓东. 基于重构误差和多块建模策略的kNN故障监测[J]. 系统仿真学报, 2023, 35(1): 95-109.
	Zheng Jing, Xiong Weili, Wu Xiaodong. kNN Fault Detection Based on Reconstruction Error and Multi-block Modeling Strategy[J]. Journal of System Simulation, 2023, 35(1): 95-109.
23	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.
24	Ge Zhiqiang, Zhang Muguang, Song Zhihuan. Nonlinear Process Monitoring Based on Linear Subspace and Bayesian Inference[J]. Journal of Process Control, 2010, 20(5): 676-688.
25	郭大权, 杨宗圣, 周晓锋, 等. 基于多块信息提取的AUV资源勘查系统故障检测[J]. 控制与决策, 2021, 36(4): 790-800.
	Guo Daquan, Yang Zongsheng, Zhou Xiaofeng, et al. Fault Detection of AUV Resource Exploration System Based on Multi-block Information Extraction[J]. Control and Decision, 2021, 36(4): 790-800.
26	赵杭天. 基于卷积自编码网络的故障在线检测研究[D]. 上海: 中国科学院上海应用物理研究所, 2021.
	Zhao Hangtian. Research on Online Fault Detection Based on Convolutional Auto-encoder Network[D]. Shanghai: Shanghai Institute of Applied Physics, Chinese Academy of Sciences, 2021.
27	Ge Zhiqiang, Song Zhihuan. Multimode Process Monitoring Based on Bayesian Method[J]. Journal of Chemometrics, 2009, 23(12): 636-650.
28	李元, 张昊展, 唐晓初. 基于多模态数据全信息的概率主成分分析故障检测研究[J]. 仪器仪表学报, 2021, 42(2): 75-85.
	Li Yuan, Zhang Haozhan, Tang Xiaochu. Study on Probabilistic Principal Component Analysis Fault Detection Based on Full Information of Multimodal Data[J]. Chinese Journal of Scientific Instrument, 2021, 42(2): 75-85.
29	邓明月, 刘建昌, 许鹏, 等. 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.
	Deng Mingyue, Liu Jianchang, Xu Peng, et al. New Fault Detection and Diagnosis Strategy for Nonlinear Industrial Process Based on KECA[J]. CIESC Journal, 2020, 71(5): 2151-2163.

方法	误报率	报警率
CAE	0.005 0	0.662 5
SW-CAE	0.010 1	0.895 0
MBI-CAE	0.002 5	0.852 5
SW-MBI-CAE	0.005 1	0.942 5

故障编码	MBI-CAE				SW-MBI-CAE
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 2	0.995 0	0.077 5	0.995 0	0.998 7	0.998 7	0.122 0	0.992 5
2	0.985 0	0.983 8	0.013 8	0.985 0	0.998 7	0.998 7	0.018 9	0.982 5
3	0.042 5	0.102 5	0.007 5	0.0737	0.035 2	0.135 8	0.007 5	0.041 3
4	0.972 5	1	0.018 7	1	1	1	0.020 1	1
5	0.255 0	0.882 5	0.071 2	0.828 7	0.254 1	0.932 1	0.084 3	0.833 7
6	1	0.998 8	0.051 2	1	1	1	0.054 1	1
7	1	0.998 8	0.187 5	1	1	1	0.236 5	1
8	0.975 0	0.978 8	0.423 8	0.977 5	0.978 6	0.986 2	0.636 5	0.977 5
9	0.035 0	0.072 5	0.025 0	0.063 8	0.055 3	0.117 0	0.039 0	0.055 0
10	0.450 0	0.600 0	0.095 0	0.611 2	0.508 2	0.667 9	0.150 9	0.625 0
11	0.683 7	0.920 0	0.077 5	0.938 7	0.836 5	0.988 7	0.168 6	0.975 0
12	0.985 0	0.996 2	0.783 7	0.995 0	0.997 5	1	0.953 5	1
13	0.947 5	0.953 7	0.561 3	0.950 0	0.950 9	0.957 1	0.704 4	0.947 5
14	1	0.778 3	0.998 8	1	1	1	1	0.998 8
15	0.071 2	0.170 0	0.015 0	0.135 0	0.088 1	0.223 9	0.039 0	0.165 0
16	0.253 7	0.440 0	0.106 2	0.430 0	0.281 8	0.517 0	0.140 9	0.452 5
17	0.900 0	0.976 2	0.287 5	0.976 2	0.947 2	0.979 9	0.359 7	0.975 0
18	0.9900 0	0.911 3	0.178 7	0.903 8	0.906 9	0.917 0	0.188 7	0.900 0
19	0.155 0	0.112 5	0.418 8	0.463 8	0.654 1	0.293 1	0.903 1	0.856 2
20	0.470 0	0.632 5	0.127 5	0.598 8	0.545 9	0.734 6	0.240 3	0.640 0
21	0.383 7	0.601 2	0.008 8	0.570 0	0.410 1	0.639 0	0.020 1	0.575 0
平均报警率	0.641 0	0.719 2	0.215 9	0.737 9	0.688 0	0.766 0	0.289 9	0.761 5
平均误报率	0.024 4	0.052 5	0.008 0	0.040 8	0.023 9	0.080 0	0.013 9	0.034 2

故障编码	LNSPPCAST^[28]	KECA^[29]	CAE	MBI-CAE	SW-MBI-CAE
平均报警率	0.703 5	0.795 3	0.789 1	0.845 8	0.874 0
1	0.996 9	0.997 5	0.992 5	0.995 0	0.992 5
2	0.988 1	0.981 3	0.982 5	0.985 0	0.982 5
4	0.068 1	1	1	1	1
5	0.513 1	0.246 3	0.258 7	0.828 7	0.833 7
6	0.981 3	1	1	1	1
7	0.251 9	1	1	1	1
8	0.981 2	0.973 8	0.972 5	0.9775	0.977 5
10	0.892 5	0.691 3	0.505 0	0.611 2	0.625 0
11	0.916 3	0.742 5	0.831 2	0.938 7	0.975 0
12	0.670 6	0.991 3	0.995 0	0.995 0	1
13	0.951 9	0.951 3	0.945 0	0.950 0	0.947 5
14	0.986 3	1	0.998 7	1	0.998 8
16	0.063 1	0.858 5	0.280 0	0.430 0	0.452 5
17	0.936 3	0.928 8	0.942 5	0.976 2	0.975 0
18	0.845 0	0.896 3	0.901 2	0.903 8	0.900 0
19	0.619 4	0.157 5	0.650 0	0.463 8	0.856 2
20	0.941 3	0.473 8	0.542 5	0.598 8	0.640 0
21	0.059 4	0.425 0	0.407 5	0.570 0	0.575 0

[1]	郑静, 熊伟丽, 吴晓东. 基于重构误差和多块建模策略的kNN故障监测[J]. 系统仿真学报, 2023, 35(1): 95-109.
[2]	秦婉亭, 老松杨, 汤俊, 卢聪. 基于变分自编码器的飓风轨迹异常检测方法[J]. 系统仿真学报, 2021, 33(9): 2191-2201.
[3]	牛玉广, 王世林, 林忠伟, 李晓明. 基于GPNMF的工业过程故障检测[J]. 系统仿真学报, 2018, 30(2): 521-532.
[4]	李海林, 张斌, 吴德伟, 卢虎. 一种高超声速飞行器组合导航故障检测方法[J]. 系统仿真学报, 2017, 29(8): 1809-1814.
[5]	葛延峰, 梁鹏, 高立群, 翟军昌. 一种滑动加权最小二乘风电功率预测模型[J]. 系统仿真学报, 2016, 28(5): 1031-1037.