Nonlinear industrial process fault diagnosis with latent label consistency and sparse Gaussian feature learning

作     者：LI Xian-ling ZHANG Jian-feng ZHAO Chun-hui DING Jin-liang SUN You-xian 李献领;张建峰;赵春晖;丁进良;孙优贤

作者机构：Science and Technology on Thermal Energy and Power LaboratoryWuhan 2nd Ship Design and Research InstituteWuhan 430205China State Key Laboratory of Industrial Control TechnologyCollege of Control Science and EngineeringZhejiang UniversityHangzhou 310027China State Key Laboratory of Synthetical Automation for Process Industries(Northeastern University)Shenyang 110819China 

基　　金：Projects(62125306, 62133003) supported by the National Natural Science Foundation of China Project(TPL2019C03) supported by the Open Fund of Science and Technology on Thermal Energy and Power Laboratory,China Project supported by the Fundamental Research Funds for the Central Universities(Zhejiang University NGICS Platform),China 

出 版 物：《Journal of Central South University》 (中南大学学报（英文版）)

年 卷 期：2022年第29卷第12期

页      码：3956-3973页

摘      要：With the increasing complexity of industrial processes, the high-dimensional industrial data exhibit a strong nonlinearity, bringing considerable challenges to the fault diagnosis of industrial processes. To efficiently extract deep meaningful features that are crucial for fault diagnosis, a sparse Gaussian feature extractor(SGFE) is designed to learn a nonlinear mapping that projects the raw data into the feature space with the fault label dimension. The feature space is described by the one-hot encoding of the fault category label as an orthogonal basis. In this way, the deep sparse Gaussian features related to fault categories can be gradually learned from the raw data by SGFE. In the feature space,the sparse Gaussian(SG) loss function is designed to constrain the distribution of features to multiple sparse multivariate Gaussian distributions. The sparse Gaussian features are linearly separable in the feature space, which is conducive to improving the accuracy of the downstream fault classification task. The feasibility and practical utility of the proposed SGFE are verified by the handwritten digits MNIST benchmark and Tennessee-Eastman(TE) benchmark process,respectively.

主 题 词：nonlinear fault diagnosis multiple multivariate Gaussian distributions sparse Gaussian feature learning Gaussian feature extractor 

学科分类：12[管理学] 1201[管理学-管理科学与工程类] 081104[081104] 08[工学] 0835[0835] 0802[工学-机械学] 0811[工学-水利类] 080201[080201] 0812[工学-测绘类] 

核心收录：

D　O　I：10.1007/s11771-022-5206-3

馆 藏 号：203118029...