Paper Info
Title
Deep Principal Component Analysis Based on Layerwise Feature Extraction and Its Application to Nonlinear Process Monitoring
Abstract
In order to deeply exploit intrinsic data feature information hidden among the process data, an improved kernel principal component analysis (KPCA) method is proposed, which is referred to as deep principal component analysis (DePCA). Specifically, motivated by the deep learning strategy, we design a hierarchical statistical model structure to extract multilayer data features, including both the linear and nonlinear principal components. To reduce the computation complexity in nonlinear feature extraction, the feature-samplesu0027 selection technique is applied to build the sparse kernel model for DePCA. To integrate the monitoring statistics at each feature layer, Bayesian inference is used to transform the monitoring statistics into fault probabilities, and then, two probability-based DePCA monitoring statistics are constructed by weighting the fault probabilities at all the feature layers. Two case studies involving a simulated nonlinear system and the benchmark Tennessee Eastman process demonstrate the superior fault detection performance of the proposed DePCA method over the traditional KPCA-based methods.
Year
DOI
Venue
2019
10.1109/tcst.2018.2865413
IEEE Transactions on Control Systems and Technology
Keywords
Field
DocType
Monitoring,Principal component analysis,Feature extraction,Kernel,Eigenvalues and eigenfunctions,Computational modeling,Machine learning
Kernel (linear algebra),Bayesian inference,Pattern recognition,Fault detection and isolation,Feature extraction,Control engineering,Kernel principal component analysis,Statistical model,Artificial intelligence,Deep learning,Mathematics,Principal component analysis
Journal
Volume
Issue
ISSN
27
6
1063-6536
Citations 
PageRank 
References 
3
0.38
0
Authors
4
Name
Order
Citations
PageRank
Deng Xiaogang111517.49
Xuemin Tian2717.54
Sheng Chen31035111.98
Chris J. Harris470066.65