Decentralized Fault Diagnosis of Large-Scale Processes Using Multiblock Kernel Partial Least Squares

In this paper, a decentralized fault diagnosis approach of complex processes is proposed based on multiblock kernel partial least squares (MBKPLS). To solve the problem posed by nonlinear characteristics, kernel partial least squares (KPLS) approaches have been proposed. In this paper, MBKPLS algorithm is first proposed and applied to monitor large-scale processes. The advantages of MBKPLS are: 1) MBKPLS can capture more useful information between and within blocks compared to partial least squares (PLS); 2) MBKPLS gives nonlinear interpretation compared to MBPLS; 3) Fault diagnosis becomes possible if number of sub-blocks is equal to the number of the variables compared to KPLS. The proposed methods are applied to process monitoring of a continuous annealing process. Application results indicate that the proposed decentralized monitoring scheme effectively captures the complex relations in the process and improves the diagnosis ability tremendously.

[1]  R. Sabatier,et al.  Comparing and predicting sensory profiles from NIRS data: use of the GOMCIA and GOMCIA‐PLS multiblock methods , 2005 .

[2]  Qidi Wu,et al.  Optimal-setting control for complicated industrial processes and its application study , 2004 .

[3]  Per A. Hassel,et al.  Nonlinear partial least squares , 2003 .

[4]  V. E. Vinzi,et al.  PLS regression, PLS path modeling and generalized Procrustean analysis: a combined approach for multiblock analysis , 2005 .

[5]  S. Qin,et al.  Improved nonlinear fault detection technique and statistical analysis , 2008 .

[6]  Julian Morris,et al.  Nonlinear multiscale modelling for fault detection and identification , 2008 .

[7]  Manfred Morari,et al.  PLS/neural networks , 1992 .

[8]  Myrtille Vivien,et al.  A new linear method for analyzing four‐way multiblock tables: STATIS‐4 , 2008 .

[9]  I. E. Frank.  A nonlinear PLS model , 1990 .

[10]  S. Joe Qin,et al.  Statistical process monitoring: basics and beyond , 2003 .

[11]  S. Wold,et al.  Nonlinear PLS modeling , 1989 .

[12]  Anders Berglund,et al.  A serial extension of multiblock PLS , 1999 .

[13]  M. Kramer.  Nonlinear principal component analysis using autoassociative neural networks , 1991 .

[14]  In-Beum Lee,et al.  Fault identification for process monitoring using kernel principal component analysis , 2005 .

[15]  Dong Dong,et al.  Nonlinear principal component analysis-based on principal curves and neural networks , 1994, Proceedings of 1994 American Control Conference - ACC '94.

[16]  Thomas J. McAvoy,et al.  Nonlinear PLS Modeling Using Neural Networks , 1992 .

[17]  Michael J. Piovoso,et al.  On unifying multiblock analysis with application to decentralized process monitoring , 2001 .

[18]  Jia-Yush Yen,et al.  Using MPCA of spectra model for fault detection in a hot strip mill , 2009 .

[19]  A. Smilde,et al.  Deflation in multiblock PLS , 2001 .

[20]  Junghui Chen,et al.  Dynamic process fault monitoring based on neural network and PCA , 2002 .

[21]  R. Rosipal,et al.  Kernel Partial Least Squares for Nonlinear Regression and Discrimination , 2002 .

[22]  S. Joe Qin,et al.  Fault Detection of Nonlinear Processes Using Multiway Kernel Independent Component Analysis , 2007 .

[23]  Age K. Smilde,et al.  Multiway multiblock component and covariates regression models , 2000 .

[24]  S. de Jong,et al.  A framework for sequential multiblock component methods , 2003 .

[25]  S. Wold,et al.  A PLS kernel algorithm for data sets with many variables and fewer objects. Part 1: Theory and algorithm , 1994 .