Deep regularized variational autoencoder for intelligent fault diagnosis of rotor–bearing system within entire life-cycle process

作者：

Highlights：

• A deep regularized variational autoencoder (DRVAE) method is proposed.

• Bird swarm algorithm is used to select adaptively the hyper-parameters of DRVAE.

• The regularized techniques are adopted for avoiding the overfitting phenomenon.

• The effectiveness of the proposed method is verified by case studies.

摘要

•A deep regularized variational autoencoder (DRVAE) method is proposed.•Bird swarm algorithm is used to select adaptively the hyper-parameters of DRVAE.•The regularized techniques are adopted for avoiding the overfitting phenomenon.•The effectiveness of the proposed method is verified by case studies.

论文关键词：Deep regularized variational autoencoder,Rotor–bearing system,Intelligent fault diagnosis,Entire life-cycle

论文评审过程：Received 14 February 2021, Revised 10 April 2021, Accepted 10 May 2021, Available online 13 May 2021, Version of Record 18 May 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107142