Dynamic kernel independent component analysis approach for fault detection and diagnosis

In this paper, a novel kernel independent component analysis method which is named improved DKICA is proposed for dynamic industry processes' fault detection and fault diagnosis. The primary idea of this method is how to obtain an augmented measurement matrix in the data kernel space, the independent component analysis is used, so the dynamic and nonlinear features can be extracted in non-linear non-Gaussian dynamic processes. Furthermore, a contribution plot of nonlinear data is defined for the improved DKICA, with which the root caused for each individual fault can be detected and isolated accurately. Finally, compared with other existing statistical monitoring and fault detection methods, the Tennessee Eastman process proves the improved performance of our proposed method.

[1]  Christos Georgakis,et al.  Plant-wide control of the Tennessee Eastman problem , 1995 .

[2]  P ? ? ? ? ? ? ? % ? ? ? ? , 1991 .

[3]  ChangKyoo Yoo,et al.  Statistical process monitoring with independent component analysis , 2004 .

[4]  Aapo Hyvärinen,et al.  New Approximations of Differential Entropy for Independent Component Analysis and Projection Pursuit , 1997, NIPS.

[5]  Hui Cao,et al.  Dimension reduction method of independent component analysis for process monitoring based on minimum mean square error , 2012 .

[6]  ChangKyoo Yoo,et al.  Statistical monitoring of dynamic processes based on dynamic independent component analysis , 2004 .

[7]  Xue-feng Yan,et al.  Fault Detection and Diagnosis in Chemical Processes Using Sensitive Principal Component Analysis , 2013 .

[8]  Christos Georgakis,et al.  Disturbance detection and isolation by dynamic principal component analysis , 1995 .

[9]  Tao Chen,et al.  A branch and bound method for isolation of faulty variables through missing variable analysis , 2010 .

[10]  Chunhui Zhao,et al.  Nonlinear Batch Process Monitoring Using Phase-Based Kernel-Independent Component Analysis-Principal Component Analysis (KICA-PCA) , 2009 .

[11]  Donghua Zhou,et al.  Generalized Reconstruction-Based Contributions for Output-Relevant Fault Diagnosis With Application to the Tennessee Eastman Process , 2011, IEEE Transactions on Control Systems Technology.

[12]  G.B. Giannakis,et al.  Cumulant-based order determination of non-Gaussian ARMA models , 1990, IEEE Trans. Acoust. Speech Signal Process..

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

[14]  Yingwei Zhang,et al.  Fault Detection and Diagnosis of Nonlinear Processes Using Improved Kernel Independent Component Analysis (KICA) and Support Vector Machine (SVM) , 2008 .

[15]  Carlos F. Alcala,et al.  Reconstruction-based contribution for process monitoring with kernel principal component analysis , 2010, Proceedings of the 2010 American Control Conference.

[16]  In-Beum Lee,et al.  Fault Detection of Non-Linear Processes Using Kernel Independent Component Analysis , 2008 .

[17]  Lv Zhaomin,et al.  Batch Process Monitoring Based on Multisubspace Multiway Principal Component Analysis and Time-Series Bayesian Inference , 2014 .

[18]  Aapo Hyvärinen,et al.  Fast and robust fixed-point algorithms for independent component analysis , 1999, IEEE Trans. Neural Networks.

[19]  Yingwei Zhang,et al.  Enhanced statistical analysis of nonlinear processes using KPCA, KICA and SVM , 2009 .

[20]  Furong Gao,et al.  Batch Process Monitoring in Score Space of Two-Dimensional Dynamic Principal Component Analysis (PCA) , 2007 .

[21]  Richard D. Braatz,et al.  Fault Detection and Diagnosis in Industrial Systems , 2001 .

[22]  Erkki Oja,et al.  Independent component analysis: algorithms and applications , 2000, Neural Networks.