Improved canonical correlation analysis-based fault detection methods for industrial processes

被引：103

作者：

Chen, Zhiwen ^{[1
]}

Zhang, Kai ^{[1
]}

Ding, Steven X. ^{[1
]}

Shardt, Yuri A. W. ^{[1
]}

Hu, Zhikun ^{[2
]}

机构：

[1] Univ Duisburg Essen, Inst Automat Control & Complex Syst, Bismarckstr 81 BB, D-47057 Duisburg, Germany

[2] Cent S Univ, Sch Phys & Elect, Changsha 410083, Peoples R China

来源：

JOURNAL OF PROCESS CONTROL | 2016年 / 41卷

基金：

中国国家自然科学基金;

关键词：

Canonical correlation analysis; Incipient multiplicative fault detection; Statistical local approach; PRINCIPAL COMPONENT ANALYSIS; PART I; DISTURBANCE DETECTION; DIAGNOSIS; INFORMATION; PERFORMANCE; SYSTEMS;

D O I：

10.1016/j.jprocont.2016.02.006

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Recent research has emphasized the successful application of canonical correlation analysis (CCA) to perform fault detection (FD) in both static and dynamic processes with additive faults. However, dealing with multiplicative faults has not been as successful. Thus, this paper considers the application of CCA to deal with the detection of incipient multiplicative faults in industrial processes. The new approaches incorporate the CCA-based FD with the statistical local approach. It is shown that the methods are effective in detecting incipient multiplicative faults. Experiments using a continuous stirred tank heater and simulations on the Tennessee Eastman process are provided to validate the proposed methods. (C) 2016 Elsevier Ltd. All rights reserved.

引用

页码：26 / 34

页数：9

共 51 条

[1]

[Anonymous], 2016, P DIAGN FAULT TOL CO

[2]

Auret L, 2013, UNSUPERVISED PROCESS

[3] Information criteria for residual generation and fault detection and isolation [J].

Basseville, M .

AUTOMATICA, 1997, 33 (05) :783-803

[4] On-board component fault detection and isolation using the statistical local approach [J].

Basseville, M .

AUTOMATICA, 1998, 34 (11) :1391-1415

[5] SOME THEOREMS ON QUADRATIC FORMS APPLIED IN THE STUDY OF ANALYSIS OF VARIANCE PROBLEMS .1. EFFECT OF INEQUALITY OF VARIANCE IN THE ONE-WAY CLASSIFICATION [J].

BOX, GEP .

ANNALS OF MATHEMATICAL STATISTICS, 1954, 25 (02) :290-302

[6] Canonical correlation analysis-based fault detection methods with application to alumina evaporation process [J].

Chen, Zhiwen ;

Ding, Steven X. ;

Zhang, Kai ;

Li, Zhebin ;

Hu, Zhikun .

CONTROL ENGINEERING PRACTICE, 2016, 46 :51-58

[7] Data-driven design of monitoring and diagnosis systems for dynamic processes: A review of subspace technique based schemes and some recent results [J].

Ding, S. X. .

JOURNAL OF PROCESS CONTROL, 2014, 24 (02) :431-449

[8] Subspace method aided data-driven design of fault detection and isolation systems [J].

Ding, S. X. ;

Zhang, P. ;

Naik, A. ;

Ding, E. L. ;

Huang, B. .

JOURNAL OF PROCESS CONTROL, 2009, 19 (09) :1496-1510

[9]

Ding S. X., 2015, P 9 IFAC S SAFEPROCE, P119

[10]

Ding SX, 2013, ADV IND CONTROL, P1, DOI 10.1007/978-1-4471-4799-2

← 1 2 3 4 5 6 →