Rolling bearing fault diagnosis using optimal ensemble deep transfer network

作者：

Highlights：

• OEDTN integrates parameter transfer learning, domain adaptation, ensemble learning.

• Diverse DTNs are built by different kernel MMDs to learn transferable features.

• A comprehensive metric is designed to guide PSO to assign voting weights for DTNs.

摘要

•OEDTN integrates parameter transfer learning, domain adaptation, ensemble learning.•Diverse DTNs are built by different kernel MMDs to learn transferable features.•A comprehensive metric is designed to guide PSO to assign voting weights for DTNs.

论文关键词：Rolling bearing,Fault diagnosis,Optimal ensemble deep transfer network,Domain adaptation,Kernel maximum mean discrepancy

论文评审过程：Received 15 April 2020, Revised 15 December 2020, Accepted 16 December 2020, Available online 17 December 2020, Version of Record 24 December 2020.

论文官网地址：https://doi.org/10.1016/j.knosys.2020.106695