SVM Based on Gaussian and Non-Gaussian Double Subspace for Fault Detection

In industrial production processes, the data usually have high-dimensional characteristics. When a support vector machine (SVM) is used for fault detection, it takes a long time to run. For high-dimensional data, principal component analysis (PCA) and independent component analysis (ICA) are very effective dimensionality reduction algorithms. Since the data follow different distributions in industrial processes, PCA is commonly used to process Gaussian-distributed data, and ICA is used to process non-Gaussian distributed data. Addressing these limitations, a novel fault detection method of SVM based on Gaussian and non-Gaussian double subspace (DSSVM) is proposed. The Kolmogorov-Smirnov (KS) test is used to determine the normal distribution characteristics of the process variables in the original data. The process variables are divided into a Gaussian subspace and a non-Gaussian subspace. Fault detection models of the Gaussian subspace are established based on PCA, and models of the non-Gaussian subspace are based on ICA. To reduce the effect of autocorrelation among process variables on the SVM, the delay and time difference input characteristics are introduced into the principal components obtained in the Gaussian subspace and independent components obtained in the non-Gaussian subspace. Finally, the time delay matrix and time difference matrix are combined, and the SVM model is used for fault detection and monitoring. The DSSVM method reduces the data dimensions and eliminates the effect of autocorrelation among process variables on the detection results. The proposed method is applied to a multivariable numerical simulation and the Tennessee-Eastman industrial process. Comparisons with simulation results for the kernel principal component analysis (KPCA), kernel independent component analysis (KICA), SVM and the statistical process monitoring method based on variable distribution characteristics (VDSPM) further verify the effectiveness of the algorithm.

[1]  Chudong Tong,et al.  Nonlinear process monitoring based on decentralized generalized regression neural networks , 2020, Expert Syst. Appl..

[2]  Yi Hu,et al.  Fault Detection and Identification Based on the Neighborhood Standardized Local Outlier Factor Method , 2013, Industrial & Engineering Chemistry Research.

[3]  Sirish L. Shah,et al.  Fault detection and diagnosis in process data using one-class support vector machines , 2009 .

[4]  Xuefeng Yan,et al.  Gaussian and non-Gaussian Double Subspace Statistical Process Monitoring Based on Principal Component Analysis and Independent Component Analysis , 2015 .

[5]  Hazem Nounou,et al.  Reliable Fault Detection and Diagnosis of Large-Scale Nonlinear Uncertain Systems Using Interval Reduced Kernel PLS , 2020, IEEE Access.

[6]  Yi Cao,et al.  Nonlinear Dynamic Process Monitoring Using Canonical Variate Analysis and Kernel Density Estimations , 2010, IEEE Transactions on Industrial Informatics.

[7]  E. F. Vogel,et al.  A plant-wide industrial process control problem , 1993 .

[8]  Chun-Chin Hsu,et al.  Integrating independent component analysis and support vector machine for multivariate process monitoring , 2010, Comput. Ind. Eng..

[9]  Chih-Jen Lin,et al.  A comparison of methods for multiclass support vector machines , 2002, IEEE Trans. Neural Networks.

[10]  Xiaogang Deng,et al.  Modified kernel principal component analysis based on local structure analysis and its application to nonlinear process fault diagnosis , 2013 .

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

[12]  Philippe Dague,et al.  Fault Detection and Isolation of spacecraft thrusters using an extended principal component analysis to interval data , 2017 .

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

[14]  Chih-Jen Lin,et al.  LIBLINEAR: A Library for Large Linear Classification , 2008, J. Mach. Learn. Res..

[15]  Zhiqiang Ge,et al.  Multimode process monitoring based on Bayesian method , 2009 .

[16]  Chih-Chieh Yang,et al.  Multiclass SVM-RFE for product form feature selection , 2008, Expert Syst. Appl..

[17]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[18]  D. Politis,et al.  Statistical Spectral Analysis of Time Series Arising From Stationary Stochastic Processes , 2011 .

[19]  T. J. McAvov,et al.  BASE CONTROL FOR THE TENNESSEE EASTMAN PROBLEM , 2001 .

[20]  Alexander J. Smola,et al.  Support Vector Regression Machines , 1996, NIPS.

[21]  Jie Yu,et al.  A support vector clustering‐based probabilistic method for unsupervised fault detection and classification of complex chemical processes using unlabeled data , 2013 .

[22]  Manabu Kano,et al.  Statistical process monitoring based on dissimilarity of process data , 2002 .

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

[24]  C. Yoo,et al.  Nonlinear process monitoring using kernel principal component analysis , 2004 .

[25]  Yuan Li,et al.  Fault Detection in the Tennessee Eastman Benchmark Process Using Principal Component Difference Based on K-Nearest Neighbors , 2020, IEEE Access.

[26]  Yue-Shi Lee,et al.  Robust and efficient multiclass SVM models for phrase pattern recognition , 2008, Pattern Recognit..

[27]  Hong Wang,et al.  Identity management based on PCA and SVM , 2016, Inf. Syst. Frontiers.

[28]  Ping Zhang,et al.  A comparison study of basic data-driven fault diagnosis and process monitoring methods on the benchmark Tennessee Eastman process , 2012 .

[29]  Fang Wu,et al.  Fault Detection and Diagnosis in Process Data Using Support Vector Machines , 2014, J. Appl. Math..

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

[31]  Arthur K. Kordon,et al.  Fault diagnosis based on Fisher discriminant analysis and support vector machines , 2004, Comput. Chem. Eng..

[32]  Li Zhang,et al.  Nonlinear feature selection using Gaussian kernel SVM-RFE for fault diagnosis , 2018, Applied Intelligence.

[33]  Majdi Mansouri,et al.  Online reduced kernel principal component analysis for process monitoring , 2018 .

[34]  Chudong Tong,et al.  Dynamic statistical process monitoring based on generalized canonical variate analysis , 2020 .

[35]  E. Parzen.  On Estimation of a Probability Density Function and Mode , 1962 .

[36]  Abdessamad Kobi,et al.  Conditional Gaussian Network as PCA for fault detection , 2014 .

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

[38]  Hazem Nounou,et al.  Fault detection using multiscale PCA-based moving window GLRT , 2017 .

[39]  Tiago J. Rato,et al.  Advantage of Using Decorrelated Residuals in Dynamic Principal Component Analysis for Monitoring Large-Scale Systems , 2013 .

[40]  Tao Xu,et al.  A Novel Fault Detection Scheme Based on Difference in Independent Component for Reliable Process Monitoring: Application on the Semiconductor Manufacturing Processes , 2020 .

[41]  Xiao Jie Zhang,et al.  An Integrated Fault Diagnosis Method Based on the ICA-SVM , 2015 .