Nonlinear Process Monitoring Based on Improved Kernel ICA

An industrial process often presents a large number of measured variables, which are usually driven by fewer nonlinear essential variables. An improved kernel independent component analysis based on particle swarm optimization (PSO-KICA) is presented to extract these essential variables from the process recorded variables in the KPCA feature space. Process faults can be detected more efficiently by monitoring the independent components. To monitor the non-Gaussian distributed independent components obtained by PSO-KICA, the empirical control limit is employed. The proposed approach is illustrated by the application to the nonisothermal CSTR process

[1]  Aapo Hyvärinen,et al.  Icasso: software for investigating the reliability of ICA estimates by clustering and visualization , 2003, 2003 IEEE XIII Workshop on Neural Networks for Signal Processing (IEEE Cat. No.03TH8718).

[2]  X. Wang,et al.  Dimension reduction of process dynamic trends using independent component analysis , 2002 .

[3]  Manabu Kano,et al.  Monitoring independent components for fault detection , 2003 .

[4]  Terrence J. Sejnowski,et al.  Blind separation and blind deconvolution: an information-theoretic approach , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.

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

[6]  James Kennedy,et al.  Particle swarm optimization , 2002, Proceedings of ICNN'95 - International Conference on Neural Networks.

[7]  G. Baudat,et al.  Feature vector selection and projection using kernels , 2003, Neurocomputing.

[8]  Seongkyu Yoon,et al.  Fault diagnosis with multivariate statistical models part I: using steady state fault signatures , 2001 .