Deep multiple auto-encoder with attention mechanism network: A dynamic domain adaptation method for rotary machine fault diagnosis under different working conditions

作者：

Highlights：

• A DMAEAM-DDA method is proposed for fault diagnosis on different working conditions.

• The attention mechanism is introduced into DMAEAM network to improve the adaptability.

• The DMAEAM-DDA method can learn domain invariant-discriminant feature across domains.

• The DMAEAM-DDA method shows outstanding diagnosis performance and strong stability.

摘要

•A DMAEAM-DDA method is proposed for fault diagnosis on different working conditions.•The attention mechanism is introduced into DMAEAM network to improve the adaptability.•The DMAEAM-DDA method can learn domain invariant-discriminant feature across domains.•The DMAEAM-DDA method shows outstanding diagnosis performance and strong stability.

论文关键词：Fault diagnosis,Deep multiple auto-encoder,Attention mechanism,Dynamic domain adaptation,Rotary machine

论文评审过程：Received 23 November 2021, Revised 8 March 2022, Accepted 22 March 2022, Available online 28 March 2022, Version of Record 17 May 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.108639