基于动态特性描述的变量加权型分散式故障检测方法

钟凯; 韩敏; 韩冰

doi:10.16383/j.aas.c180276

基于动态特性描述的变量加权型分散式故障检测方法

doi: 10.16383/j.aas.c180276

钟凯^1,,
韩敏^1,,
韩冰^2,

1.
大连理工大学电子信息与电气工程学部大连 116023
2.
上海船舶运输科学研究所航运技术与安全国家重点实验室上海 200135

基金项目: 国家自然科学基金(61773087)资助

详细信息

作者简介:
钟凯：大连理工大学电子信息与电气工程学部博士研究生. 主要研究方向为工业过程监控, 故障诊断. E-mail: kaizhong0402@ahu.edu.cn

韩敏：大连理工大学电子信息与电气工程学部教授. 主要研究方向为模式识别, 复杂系统建模与分析及时间序列预测. 本文通信作者. E-mail: minhan@dlut.edu.cn

韩冰：航运技术与安全国家重点实验室研究员. 主要研究方向为深海动力定位控制, 船舶动力装置的故障诊断和预测. E-mail: hanbing@sssri.com

计量
- 文章访问数: 699
- HTML全文浏览量: 99
- PDF下载量: 179
- 被引次数: 0
出版历程
- 收稿日期: 2018-05-03
- 录用日期: 2018-12-12
- 刊出日期: 2021-10-13

Dynamic Feature Characterization Based Variable-weighted Decentralized Method for Fault Detection

ZHONG Kai^1
,,
HAN Min^1
,,
HAN Bing^2
,

1.
Faculty of Electronic Information and Electrical Engineering, Dalian University of Technology, Dalian 116023
2.
State Key Laboratory of Navigation and Safety Technology, Shanghai Ship and Shipping Research Institute, Shanghai 200135

Funds: Supported by National Natural Science Foundation of China(61773087)

More Information

Author Bio:
ZHONG Kai　Ph.D. candidate at the Faculty of Electronic Information and Electrical Engineering, Dalian University of Technology. His research interest covers industrial process monitoring and fault diagnosis

HAN Min　Professor at the Faculty of Electronic Information and Electrical Engineering, Dalian University of Technology. Her research interest covers pattern recognition, modeling and analysis of complex system, and time series prediction. Corresponding author of this paper

HAN Bing　Professor at the State Key Laboratory of Navigation and Safety Technology. His research interest covers deep sea dynamic positioning control, and fault diagnosis and prognostic of ship power plant

摘要

摘要: 现代工业生产过程往往具有复杂的动态特性: 不同测量变量间会存在不同的时序相关性, 且变量间的相互影响会反映在不同的采样时刻上. 现有的动态过程监测模型往往不能充分挖掘变量间的动态特性, 其故障检测效果也有待进一步提高. 在此背景下, 本文提出一种基于动态特性描述的变量加权型分散式故障检测方法. 利用最大相关最小冗余(Minimal redundancy maximal relevance, mRMR) 算法更准确地描述动态过程变量间的相关性关系, 并利用该相关性的值对原始增广矩阵进行加权处理, 且不同延迟变量对当前测量值的影响大小就通过权值来体现, 因此能更加全面地刻画该测量值的动态特性. 最后建立一种融合mRMR算法, 贝叶斯推理以及动态主成分分析(Dynamic principal componemt amalysis, DPCA)模型的新的分布式建模策略, 提高了模型的容错能力和泛化能力, 取得了更好的故障检测结果.
- 主成分分析 /
- 故障检测 /
- 最大相关最小冗余 /
- 加权型分散式方法
Abstract: Modern industrial process is often accompanied by complicated dynamic behaviours: different measured variables have different serial correlations, and the interactions among these variables are reflected in different sampling instants. The existing dynamic process monitoring models can not fully excavate the dynamic characteristics among variables, then the corresponding fault detection results need to be further improved. Against this background, this paper proposes the dynamic feature characterization based variable-weighted decentralized method for fault detection, which uses the minimal redundancy maximal relevance (mRMR) algorithm to get more accurate description of the relationship among different measured variables in dynamic process, then the augmented matrix can be weighted according to the mRMR values, and the influences of different delay variables on current measured one are reflected by the weights. Hence, the dynamic characteristics of the measured variable can be described more comprehensively. Finally, a new distributed modeling strategy combining the mRMR algorithm, Bayesian inference and dynamic principal component analysis (DPCA), which improves fault tolerance and generalization ability of the model, thus obtaining better fault detection results.
- Principal component analysis /
- fault detection /
- minimal redundancy maximal relevance (mRMR) /
- weighted decentralized method

HTML全文

图 1 基于mRMR的动态特性描述

Fig. 1 mRMR-based dynamic feature characterization

下载: 全尺寸图片幻灯片

图 2 基于mRMR-WDPCA故障监测的流程图

Fig. 2 Flowchart of mRMR-WDPCA based fault detection

下载: 全尺寸图片幻灯片

图 3 TE过程的结构图

Fig. 3 Structure diagram of the TE process

下载: 全尺寸图片幻灯片

图 4 4种方法的故障平均漏报率

Fig. 4 Average missing alarm rates of the four methods

下载: 全尺寸图片幻灯片

图 5 故障10的过程监控结果

Fig. 5 The monitoring charts of Fault 10

下载: 全尺寸图片幻灯片

图 6 故障16的过程监控结果

Fig. 6 The monitoring charts of Fault 16

下载: 全尺寸图片幻灯片

表 1 TE过程的误报率(%)

Table 1 False alarm rates of TE process (%)

模型	$T^2_s$	${ BIC}_{T^2}/T^2/T^2_d$	${ BIC}_{Q}/Q/Q_r$
DPCA	—	0.63	3.24
DLV	1.00	3.02	3.24
MI-DPCA	—	0.21	1.98
mRMR-WDPCA	—	1.63	2.13

下载: 导出CSV

表 2 TE过程故障漏报率(%)和检测延迟数(个)

Table 2 Missing alarm rates (%) and detection delay (delayed samples) of TE process

故障编号	故障类型	DPCA		DLV		MI-DPCA		mRMR-WDPCA
故障编号	故障类型	$T^2/Q$	检测延迟数	$T^2_{s}/ T^2_{d}/Q_r$	检测延迟数	${BIC}_{T^2} /{BIC}_{Q}$	检测延迟数	${BIC}_{T^2} /{BIC}_{Q}$	检测延迟数
1	阶跃	0.13	0	0.00	0	0.13	0	0.25	0
2	阶跃	1.50	2	1.00	0	1.38	10	1.50	10
4	阶跃	0.00	0	0.00	0	0.00	0	0.00	0
5	阶跃	55.00	0	0.13	0	73.13	0	0.00	0
6	阶跃	0.00	0	0.00	0	0.00	0	0.00	0
7	阶跃	0.00	0	0.00	0	0.00	0	0.00	0
8	随机	2.63	1	6.38	10	2.50	13	1.75	12
10	随机	48.88	18	37.50	7	25.50	24	18.88	2
11	随机	6.00	3	19.00	3	4.63	3	13.50	3
12	随机	0.88	0	9.00	0	0.63	0	0.13	0
13	慢偏移	4.63	35	4.88	26	4.63	39	5.38	41
14	粘滞	0.00	0	0.00	0	0.00	0	0.00	0
16	未知	48.00	10	36.6	39	23.50	11	14.75	7
17	未知	2.25	16	5.13	16	2.13	0	3.38	0
18	未知	9.38	15	9.63	17	9.38	16	9.00	1
19	未知	33.38	0	37.00	10	37.63	1	65.002	2
20	未知	36.38	12	35.13	2	33.38	55	32.50	45
21	恒定故障	49.50	26	49.25	7	42.63	40	47.13	9

下载: 导出CSV

参考文献(28)

[1]	Ge Z Q, Song Z H, Gao F R. Review of recent research on data-based process monitoring. Industrial & Engineering Chemistry Research, 2013, 52(10): 3543-3562
[2]	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. 自动化学报, 2017, 43(3): 349-365 Peng Kai-Xiang, Ma Liang, Zhang Kai. Review of quality-related fault detection and diagnosis techniques for complex industrial processes. Acta Automatica Sinica, 2017, 43(3): 349-365
[3]	Yin S, Ding S X, Xie X, Luo H. A review on basic data-driven approaches for industrial process monitoring. IEEE Transactions on Industrial Electronics, 2014, 61(11): 6418-6428 doi: 10.1109/TIE.2014.2301773
[4]	Jiang Q, Yan X. Nonlinear plant-wide process monitoring using MI-spectral clustering and Bayesian inference-based multiblock KPCA. Journal of Process Control, 2015, 32: 38-50 doi: 10.1016/j.jprocont.2015.04.014
[5]	Jia Q L, Zhang Y. Quality-related fault detection approach based on dynamic kernel partial least squares. Chemical Engineering Research and Design, 2016, 106: 242-252 doi: 10.1016/j.cherd.2015.12.015
[6]	韩敏, 张占奎. 基于加权核独立成分分析的故障检测方法. 控制与决策, 2016, 31(2): 242-248 Han Min, Zhang Zhan-Kui. Fault detection method based on weighted kernel independent component analysis. Control and Decision, 2016, 31(2): 242-248
[7]	Dong Y N, Qin S J. A novel dynamic PCA algorithm for dynamic data modeling and process monitoring. Journal of Process Control, 2018, 67: 1-11 doi: 10.1016/j.jprocont.2017.05.002
[8]	周东华, 胡艳艳. 动态系统的故障诊断技术. 自动化学报, 2009, 35(6): 748-758 Zhou Dong-Hua, Hu Yan-Yan. Fault diagnosis techniques for dynamic systems. Acta Automatica Sinica, 2009, 35(6): 748-758
[9]	Ku W, Storer R H, Georgakis C. Disturbance detection and isolation by dynamic principal component analysis. Chemometrics and Intelligent Laboratory Systems, 1995, 30(1): 179-196 doi: 10.1016/0169-7439(95)00076-3
[10]	Chen J, Liu K C. On-line batch process monitoring using dynamic PCA and dynamic PLS models. Chemical Engineering Science, 2002, 57(1): 63-75 doi: 10.1016/S0009-2509(01)00366-9
[11]	Li W, Qin S J. Consistent dynamic PCA based on errors-in-variables subspace identification. Journal of Process Control, 2001, 11(6): 661-678 doi: 10.1016/S0959-1524(00)00041-X
[12]	Li G, Qin S J, Zhou D H. A new method of dynamic latent-variable modeling for process monitoring. IEEE Transactions on Industrial Electronics, 2014, 61(11): 6438-6445 doi: 10.1109/TIE.2014.2301761
[13]	Zhang Y W, Zhou H, Qin S J. Decentralized fault diagnosis of large-scale processes using multiblock kernel principal component analysis. Acta Automatica Sinica, 2010, 36(4): 593-597
[14]	Ge Z Q, Song Z H. Distributed PCA model for plant-wide process monitoring. Industrial & Engineering Chemistry Research, 2013, 52(5): 1947-1957
[15]	Tong C D, Shi X H. Decentralized monitoring of dynamic processes based on dynamic feature selection and informative fault pattern dissimilarity. IEEE Transactions on Industrial Electronics, 2016, 63(6): 3804-3814 doi: 10.1109/TIE.2016.2530047
[16]	童楚东, 蓝艇, 史旭华. 基于互信息的分散式动态PCA故障检测方法. 化工学报, 2016, 67(10): 4317-4323 Tong Chu-Dong, Lan Ting, Shi Xu-Hua. Fault detection by decentralized dynamic PCA algorithm on mutual information. CIESC Journal, 2016, 67(10): 4317-4323
[17]	Xu C, Zhao S Y, Liu F. Distributed plant-wide process monitoring based on PCA with minimal redundancy maximal relevance. Chemometrics and Intelligent Laboratory Systems, 2017, 169: 53-63 doi: 10.1016/j.chemolab.2017.08.004
[18]	Tong C D, L T, Shi X H. Fault detection and diagnosis of dynamic processes using weighted dynamic decentralized PCA approach. Chemometrics and Intelligent Laboratory Systems, 2017, 161: 34-42 doi: 10.1016/j.chemolab.2016.11.015
[19]	Han M, Zhong K, Qiu T, Han B. Interval type-2 fuzzy neural networks for chaotic time series prediction: A concise overview. IEEE Transactions on Cybernetics, 2019, 49(7): 2720-2731
[20]	Peng H C, Long F H, Ding C. Feature selection based on mutual information: criteria of max-dependency, max-relevance, and min-redundancy. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(8): 1226-1238 doi: 10.1109/TPAMI.2005.159
[21]	Qin S J. Statistical process monitoring: basics and beyond. Journal of Chemometrics, 2003, 17(7-8): 480-502
[22]	Zhong K, Han M, Qiu T, et al. Distributed dynamic process monitoring based on minimal redundancy maximal relevance variable selection and Bayesian inference[J]. IEEE Transactions on Control Systems Technology, 2019, 28(5): 2037-2044
[23]	Downs J J, Vogel E F, A plant-wide industrial process control problem. Computers & Chemical Engineering, 1993, 17(3): 245-255
[24]	曹玉苹, 黄琳哲, 田学民. 一种基于DIOCVA的过程监控方法. 自动化学报, 2015, 41(12): 2072-2080 Cao Yu-Ping, Huang Lin-Zhe, Tian Xue-Min. A process monitoring method using dynamic input-output canonical variate analysis. Acta Automatica Sinica, 2015, 41(12): 2072-2080
[25]	Lee J M, Yoo C K, Lee I B. Statistical process monitoring with independent component analysis. Journal of Process Control, 2004, 14(5): 467-485 doi: 10.1016/j.jprocont.2003.09.004
[26]	Chiang L H, Russell E L, Braatz R D. Fault Detection and Diagnosis in Industrial Systems. London: Springer-Verlag, 2001.
[27]	Li N, Yan W W, Yang Y P. Spatial-statistical local approach for improved manifold-based process monitoring. Industrial & Engineering Chemistry Research, 2015, 54(34): 8509-8519
[28]	Zhao C H, Wang W, Qin Y, Gao F R. Comprehensive subspace decomposition with analysis of between-mode relative changes for multimode process monitoring. Industrial & Engineering Chemistry Research, 2015, 54(12): 3154-3166