4.8 Article

Data-Driven Fault Diagnosis Using Deep Canonical Variate Analysis and Fisher Discriminant Analysis

期刊

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
卷 17, 期 5, 页码 3324-3334

出版社

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TII.2020.3030179

关键词

Fault diagnosis; Correlation; Kernel; Generators; Informatics; Neural networks; Principal component analysis; Bayesian classifier; canonical variate analysis (CVA); deep neural networks (DNN); fault diagnosis; Fisher discriminant analysis (FDA); residual generator

资金

  1. National Natural Science Foundation of China [61703371]
  2. Social Development Project of Zhejiang Provincial Public Technology Research [LGF19F030004, TII-20-3359]

向作者/读者索取更多资源

A novel data-driven fault diagnosis method was proposed, combining deep learning and statistical analysis to achieve better fault classification results through deep neural networks and analysis of feature space.
In this article, a novel data-driven fault diagnosis method by combining deep canonical variate analysis and Fisher discriminant analysis (DCVA-FDA) is proposed for complex industrial processes. Inspired by the recently developed deep canonical correlation analysis, a new nonlinear canonical variate analysis (CVA) called DCVA is first developed by incorporating deep neural networks into CVA. Based on DCVA, a residual generator is designed for the fault diagnosis process. FDA is applied in the feature space spanned by residual vectors. Then, a Bayesian inference classifier is performed in the reduced dimensional space of FDA to label the class of process data. A continuous stirred-tank reactor and an industrial benchmark of the Tennessee Eastman process are carried out to test the performance of DCVA-FDA fault diagnosis. The experimental results demonstrate that the proposed DCVA-FDA fault diagnosis is able to significantly improve the fault diagnosis performance when compared to other methods also examined in this article.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.8
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据