Deep transfer network for rotating machine fault analysis

作者：

Highlights：

• Both the first and higher-order moments contribute in distribution alignments.

• Soft labels can also align conditional distributions effectively.

• Joint distribution alignments work better than marginal distribution alignments.

摘要

•Both the first and higher-order moments contribute in distribution alignments.•Soft labels can also align conditional distributions effectively.•Joint distribution alignments work better than marginal distribution alignments.

论文关键词：Intelligent fault diagnosis,Rotating machine,Deep transfer network,Auto-balanced high-order KL divergence,Smooth conditional distribution alignment,Weighted joint domain adaptation

论文评审过程：Received 17 January 2019, Revised 25 July 2019, Accepted 31 July 2019, Available online 1 August 2019, Version of Record 8 August 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.106993