New contributions to non-linear process monitoring through kernel partial least squares

被引:26
作者
Godoy, Jose L. [1 ]
Zumoffen, David A. [2 ]
Vega, Jorge R. [1 ]
Marchetti, Jacinto L. [1 ]
机构
[1] Inst Technol Dev Chem Ind INTEC CONICET UNL, RA-3000 Santa Fe, Argentina
[2] French Argentine Int Ctr Informat & Syst Sci CIFA, Rosario, Santa Fe, Argentina
关键词
KPLS modeling; Fault detection; Fault diagnosis; Prediction risk assessment; Non-linear processes; FAULT-DETECTION; DIAGNOSIS; PREDICTION; REGRESSION; PROJECTION; SELECTION;
D O I
10.1016/j.chemolab.2014.04.001
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The kernel partial least squares (KPLS) method was originally focused on soft-sensor calibration for predicting online quality attributes. In this work, an analysis of the KPLS-based modeling technique and its application to non-linear process monitoring are presented. To this effect, the measurement decomposition, the development of new specific statistics acting on non-overlapped domains, and the contribution analysis are addressed for purposes of fault detection, diagnosis, and prediction risk assessment. Some practical insights for synthesizing the models are also given, which are related to an appropriate order selection and the adoption of the kernel function parameter. A proper combination of scaled statistics allows the definition of an efficient detection index for monitoring a non-linear process. The effectiveness of the proposed methods is confirmed by using simulation examples. (C) 2014 Elsevier B.V. All rights reserved.
引用
收藏
页码:76 / 89
页数:14
相关论文
共 34 条
[1]   Analysis and generalization of fault diagnosis methods for process monitoring [J].
Alcala, Carlos F. ;
Qin, S. Joe .
JOURNAL OF PROCESS CONTROL, 2011, 21 (03) :322-330
[2]   Reconstruction-Based Contribution for Process Monitoring with Kernel Principal Component Analysis [J].
Alcala, Carlos F. ;
Qin, S. Joe .
INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH, 2010, 49 (17) :7849-7857
[3]  
Beardah C.C., 1995, MATLAB ROUTINES KERN
[4]  
Bishop CM., 1995, NEURAL NETWORKS PATT
[5]   Classification of metabolites with kernel-partial least squares (K-PLS) [J].
Embrechts, Mark J. ;
Ekins, Sean .
DRUG METABOLISM AND DISPOSITION, 2007, 35 (03) :325-327
[6]   A fault detection and diagnosis technique for multivariate processes using a PLS-decomposition of the measurement space [J].
Godoy, Jose L. ;
Vega, Jorge R. ;
Marchetti, Jacinto L. .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2013, 128 :25-36
[7]   Multivariate statistical monitoring of an industrial SBR process. Soft-sensor for production and rubber quality [J].
Godoy, Jose L. ;
Minari, Roque J. ;
Vega, Jorge R. ;
Marchetti, Jacinto L. .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2011, 107 (02) :258-268
[8]   Enhanced batch process monitoring using just-in-time-learning based kernel partial least squares [J].
Hu, Yi ;
Ma, Hehe ;
Shi, Hongbo .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2013, 123 :15-27
[9]   On-line batch process monitoring using batch dynamic kernel principal component analysis [J].
Jia, Mingxing ;
Chu, Fei ;
Wang, Fuli ;
Wang, Wei .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2010, 101 (02) :110-122
[10]   A novel multivariate regression approach based on kernel partial least squares with orthogonal signal correction [J].
Kim, K ;
Lee, JM ;
Lee, IB .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2005, 79 (1-2) :22-30