面向多模式多元未知分布的协方差过程监控

doi:10.3969/j.issn.1004-132X.2023.22.009

摘要/Abstract

摘要： 针对工业过程数据的多模式、高维性和非正态性等挑战，设计了一种考虑模式过渡约束的基于协方差检验的多模式在线监测方法。首先利用交叉验证的线性收缩估计方法对待监测数据的协方差矩阵进行估计，然后基于估计的协方差矩阵计算稀疏主特征值检验统计量，进而基于稀疏主特征值检验统计量设计具有滑动窗口的指数加权移动平均（EWMA）控制图——MSPEWMA控制图，结合模式过渡约束得到最终的检验统计量。通过Monte Carlo模拟研究了MSPEWMA控制图在不同条件（变量维度、漂移大小、可控样本大小、过渡约束参数和观测数据分布）下的性能，结果表明，与基于协方差检验的其他控制图相比，新提出的控制图在大漂移和非正态条件下具有更好的监测效果。最后利用风力发电机组真实SCADA数据证明了所提方法的有效性。

关键词: 高维度, 多模式, 分布未知, 协方差估计, 多元指数加权移动平均

Abstract: Aiming at the challenges of multi-mode, high dimension and nonnormality of industrial process data, a multi-mode online monitoring method was designed herein based on covariance tests considering time constraints. Firstly, the covariance matrix of the data to be detected was estimated using a cross-validated linear shrinkage estimation method. Secondly, the sparse principal eigenvalue test statistic was calculated based on the estimated covariance matrix. Subsequently, an EWMA（exponentially weighted moving-average） control chart with sliding window was designed based on the sparse principal eigenvalue test statistic: the MSPEWMA control chart, and the final test statistic was obtained by combining the mode transition constraints. Through Monte Carlo simulation, the performance of the MSPEWMA control chart under different conditions(variable dimension, drift size, control sample size, transition constraint parameters, and observation data distribution) was investigated. The results show that the newly proposed control chart has better monitoring results under large drift and non-normal conditions compared with other control charts based on covariance tests. Finally, the effectiveness of the method was demonstrated using real SCADA data of wind turbines.

Key words: high dimension, multi-mode, unknown distribution, covariance estimation, multivariate exponentially weighted moving-average

中图分类号:

TP277

赵宇, 李艳婷, 吴振宇, 周笛, 胡洁. 面向多模式多元未知分布的协方差过程监控[J]. 中国机械工程, 2023, 34(22): 2721-2736，2757.

ZHAO Yu, LI Yanting, WU Zhenyu, ZHOU Di, HU Jie. Covariance Process Monitoring Scheme for Multi-mode Multivariate Unknown Distributions[J]. China Mechanical Engineering, 2023, 34(22): 2721-2736，2757.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.cmemo.org.cn/CN/10.3969/j.issn.1004-132X.2023.22.009

http://www.cmemo.org.cn/CN/Y2023/V34/I22/2721

参考文献

［1］李静,刘坚,李蓉.方差监测的自适应累积和车身质量控制图研究［J］.中国机械工程,2013,24(14):1979-1983.
LI Jing, LIU Jian, LI Rong. Research on Adaptive Accumulation of Variance Monitoring and Car Body Quality Control Chart［J］. China Mechanical Engineering,2013,24(14):1979-1983.
［2］WANG R, EDGAR T F, BALDEA M, et al. A Geometric Method for Batch Data Visualization, Process Monitoring and Fault Detection［J］. Journal of Process Control, 2018, 67:197-205.［3］赵春晖,余万科,髙福荣.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 Chunhui, YU Wanke, GAO Furong. Data Analytics and Condition Monitoring Methods for Non-stationary Batch Processes-current Status and Future［J］. Acta Automatica Sinica, 2020, 46(10) :2072-2091.
［4］赵春华,汪成康,华露,等.基于融合特征约减和支持向量机的控制图模式识别［J］.中国机械工程, 2017,28(8):930-935.
ZHAO Chunhua, WANG Chengkang, HUA Lu, et al. Control Chart Pattern Recognition Based on Fusion Feature Reduction and Support Vector Machine［J］. China Mechanical Engineering, 2017,28(8):930-935.
［5］张建，胡小锋，张亚辉. 基于自步学习的刀具加工过程监测数据异常检测方法［J］.上海交通大学学报, 2023, 57(10)：1346-1354.
ZHANG Jian, HU Xiaofeng, ZHANG Yanhui. Abnormal Detection Method of Tool Machining Monitoring Dat a Based on Self-paced Learning［J］. Journal of Shanghai Jiao Tong University, 2023, 57(10)：1346-1354.
［6］黄健，杨旭. 基于在线加权慢特征分析的故障检测算法［J］. 上海交通大学学报, 2020, 54(11):1142-1150.
HUANG Jian, YANG Xu. Online Weighted Slow Feature Analysis Based Fault Detection Algorithm［J］. Journal of Shanghai Jiao Tong University, 2020, 54(11):1142-1150.
［7］UEDA R M, AGOSTINO I, SOUZA A M. Analysis and Perspectives on Multivariate Statistical Process Control Charts Used in the Industrial Sector:a Systematic Literature Review［J］. Management and Production Engineering Review,2022,13(2):48-60.
［8］SHANG J, ZHOU D, CHEN M, et al. Incipient Sensor Fault Diagnosis in Multimode Processes Using Conditionally Independent Bayesian Learning Based Recursive Transformed Component Statistical Analysis［J］. Journal of Process Control, 2019, 77:7-19.
［9］YAN Dandan, ZHANG Shuai, JUNG U K. A Variable-selection Control Chart via Penalized Likelihood and Gaussian Mixture Model for Multimodal and High-dimensional Processes［J］. Quality and Reliability Engineering International,2019, 35(4):1263-1275.
［10］JIANG Q, YAN X. Multimode Process Monitoring Using Variational Bayesian Inference and Canonical Correlation Analysis［J］. IEEE Transactions on Automation Science and Engineering, 2019, 16(4):1814-1824.
［11］MA H, YI H, SHI H. A Novel Local Neighborhood Standardization Strategy and Its Application in Fault Detection of Multimode Processes［J］. Chemometrics and Intelligent Laboratory Systems, 2012, 118:287-300.
［12］LYU Feiya, WEN Chenglin, LIU Meiqin. Representation Learning Based Adaptive Multimode Process Monitoring［J］. Chemometrics and Intelligent Laboratory Systems, 2018, 181:95-104.
［13］LI Shuai, ZHOU Xiaofeng, SHI Haibo. Multimode Processes Monitoring Based on Hierarchical Mode Division and Subspace Decomposition［J］. The Canadian Journal of Chemical Engineering, 2018, 96:2420- 2430.
［14］郭金玉，刘玉超，李元. 基于概率密度PCA的多模态过程故障检测［J］.计算机应用研究, 2019, 36(5):1396-1399.
GUO Jinyu, LIU Yuchao, LI Yuan. PCA Based on Probability Density for Fault Detection of Multimodal Processes［J］. Application Research of Computers,2019, 36(5):1396-1399.
［15］CHEN Yushan, LIN Yuan, ZHENG Tianyu. AnImproved JITL Method for Soft Sensing of Multimodal Industrial Processes for Search Efficiency［J］. Journal of Physics:Conference Series, 2021, 1952:022036.
［16］CAO Yue, MAGBOOL J N, HUANG Biao, et al. Multimodal Process Monitoring Based on Variational Bayesian PCA and Kullback-Leibler Divergence between Mixture Models［J］. Chemometrics and Intelligent Laboratory Systems,2021, 210:104230.
［17］QUINONES-GRUEIRO M, PRIETO-MORENO A, VERDE C, et al. Data-driven Monitoring of Multimode Continuous Processes :a Review［J］. Chemometrics and Intelligent Laboratory Systems,2019,189:56-71.
［18］SAMMAKNEJAD N, ZHAO Y, HUANG B. A Review of the Expectation Maximization Algorithm in Data-driven Process Identification［J］. Journal of Process Control, 2019, 73:123-136.
［19］JI H, HUANG K, ZHOU D. Incipient Sensor Fault Isolation Based on Augmented Mahalanobis Distance［J］. Control Engineering Practice, 2019, 86:144-154.
［20］JIANG Qingchao, YAN Xuefeng. Multimode Process Monitoring Using Variational Bayesian Inference and Canonical Correlation Analysis［J］. IEEE Transactions on Automation Science and Engineering, 2019, 16(4):1814-1824.
［21］CHEN Shutian, JIANG Qingchao, YAN Xuefeng. Multimodal Process Monitoring Based on Transition-constrained Gaussian Mixture Model［J］. Chinese Journal of Chemical Engineering, 2020, 28:3070-3078.
［22］CAO Yue, MAGBOOL J N,HUANG Biao, et al. No-delay Multimodal Process Monitoring Using Kullback-Leibler Divergence-based Statisticsin Probabilistic Mixture Models［J］. IEEE Transactions on Automation Science and Engineering, 2022, 20(1):167-178.
［23］LI J, CHEN S X. Two Sample Tests for High-dimensional Covariance Matrices［J］. The Annals of Statistics, 2012, 40(2):908-940.
［24］CAI T, LIU W, XIA Y. Two-sample Covariance Matrix Testing and Support Recovery in High-Dimensional and Sparse Settings［J］. Journal of the American Statistical Association, 2013, 108(501):265-277.
［25］ZHU L, LEI J, DEVLIN B, ROEDER K. Testing High-dimensional Covariance Matrices, with Application to Detecting Schizophrenia Risk Genes［J］. The Annals of Applied Statistics, 2017, 11(3):1810-1831.
［26］KIM J, ABDELLA G M, KIM S, et al. Control Charts for Variability Monitoring in High-Dimensional Processes［J］. Computers & Industrial Engineering, 2019, 130:309-316.
［27］ALFARO J L, ORTEGA J F. A New Multivariate Variability Control Chart Based on a Covariance Matrix Combination［J］. Applied Stochastic Models in Business and Industry, 2019, 35(3):823-836.
［28］NING Xiu, LI Pingke. A Simulation Comparison of Some Distance-based EWMA Control Charts for Monitoring the Covariance Matrix with Individual Observations［J］. Quality and Reliability Engineering International, 2020:36(1):50-67.
［29］EBADI M, CHENOURI S, LIN DKJ, STEINER S H. Statistical Monitoring of the Covariance Matrix in Multivariate Processes:a Literature Review［J］. Journal of Quality Technology, 2022, 54(3):269-289.
［30］WANG Yang, WAN Yiming, ZHANG Hong, et al. Robust Decomposition of Kernel Function-based Nonlinear Robust Multimode Process Monitoring［J］. IEEE Transactions on Instrumentation and Measurement, 2023, 72:3511711.
［31］YANG Guangren, LIU Yiming, PAN Guangming. Weighted Covariance Matrix Estimation［J］. Computational Statistics & Data Analysis, 2019, 139:82-98.

[1]	范新宇, 王鹏, 殷杰, 朱严兵. 新型复合式电磁直线执行器的多模式协调控制[J]. 中国机械工程, 2023, 34(03): 292-299.
[2]	赵宇, 李艳婷. 面向小样本高维数据的秩检验非参控制图[J]. 中国机械工程, 2022, 33(09): 1104-1114.
[3]	寇发荣;魏冬冬;梁津;田蕾. 一种馈能型混合悬架的多模式协调控制[J]. 中国机械工程, 2018, 29(11): 1356-1363.
[4]	赵勇, 秦大同, 武文辉. 考虑失效相关的盾构机刀盘驱动多级行星传动系统的可靠性模型 [J]. 中国机械工程, 2011, 22(5): 522-527.