Fault detection based on neural networks and independent component analysis

In this paper a method that integrates neural networks (NN) and independent component analysis (ICA) is used to detect faults in non-linear plants. The neural networks are used to calculate a non-linear and dynamic model of the process in normal operation conditions. ICA is used to monitor and to detect faults in the process using, instead of the measured variables of the process, the residuals calculated as the difference between the process measurements and the output of the networks. This technique has been applied in simulation to a benchmark of the biological wastewater treatment process, a highly non-linear process. In order to prove the advantages of using this monitoring technique called NNICA, a comparison with the classical PCA and ICA methods is carried out in the paper.

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

[2]  Bernhard Schölkopf,et al.  Nonlinear Component Analysis as a Kernel Eigenvalue Problem , 1998, Neural Computation.

[3]  Alberto Ferrer,et al.  Multivariate Statistical Process Control Based on Principal Component Analysis (MSPC-PCA): Some Reflections and a Case Study in an Autobody Assembly Process , 2007 .

[4]  Raghunathan Rengaswamy,et al.  A review of process fault detection and diagnosis: Part III: Process history based methods , 2003, Comput. Chem. Eng..

[5]  Sylvie Gillot,et al.  The COST Simulation Benchmark: Description and Simulator Manual , 2001 .

[6]  Theodora Kourti,et al.  Multivariate SPC Methods for Process and Product Monitoring , 1996 .

[7]  Silvio Simani,et al.  Model-Based Fault Diagnosis Techniques , 2003 .

[8]  Robert M. Pap,et al.  Fault Diagnosis , 1990, Bayesian Networks in Fault Diagnosis.

[9]  Raghunathan Rengaswamy,et al.  A review of process fault detection and diagnosis: Part I: Quantitative model-based methods , 2003, Comput. Chem. Eng..

[10]  Manabu Kano,et al.  Evolution of multivariate statistical process control: application of independent component analysis and external analysis , 2004, Comput. Chem. Eng..

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

[12]  Gail D. Baura,et al.  Nonlinear System Identification , 2002 .

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

[14]  W. Cholewa,et al.  Fault Diagnosis: Models, Artificial Intelligence, Applications , 2004 .

[15]  George W. Irwin,et al.  Improved Nonlinear PCA for Process Monitoring Using Support Vector Data Description , 2011 .

[16]  Seungjin Choi,et al.  Independent Component Analysis , 2009, Handbook of Natural Computing.

[17]  Steven X. Ding,et al.  Model-based Fault Diagnosis Techniques: Design Schemes, Algorithms, and Tools , 2008 .

[18]  Jin Hyun Park,et al.  Fault detection and identification of nonlinear processes based on kernel PCA , 2005 .

[19]  A. J. Morris,et al.  Non-parametric confidence bounds for process performance monitoring charts☆ , 1996 .

[20]  Raghunathan Rengaswamy,et al.  A review of process fault detection and diagnosis: Part II: Qualitative models and search strategies , 2003, Comput. Chem. Eng..

[21]  M. V. Iordache,et al.  Diagnosis and Fault-Tolerant Control , 2007, IEEE Transactions on Automatic Control.

[22]  Theodora Kourti,et al.  Process analysis and abnormal situation detection: from theory to practice , 2002 .