A new bearing fault diagnosis approach combining sensitive statistical features with improved multiscale permutation entropy method
作者：
Highlights：
• A new idea is provided to identify sensitive statistical features.
• Application of XGBoost method is introduced in condition monitoring of bearings.
• The proposed method demonstrates anti-noise ability through simulation analysis.
• Proposed method excel over IMPE, AAPE and WPE methods from experimental analysis.
• New statistical features can be introduced as the methodology allows flexibility.
摘要
•A new idea is provided to identify sensitive statistical features.•Application of XGBoost method is introduced in condition monitoring of bearings.•The proposed method demonstrates anti-noise ability through simulation analysis.•Proposed method excel over IMPE, AAPE and WPE methods from experimental analysis.•New statistical features can be introduced as the methodology allows flexibility.
论文关键词：Ball bearing fault diagnosis,Improved multiscale permutation entropy,Multiscale statistical parameters,Complementary ensemble empirical mode decomposition,Extreme gradient boosting
论文评审过程：Received 17 July 2020, Revised 13 February 2021, Accepted 16 February 2021, Available online 18 February 2021, Version of Record 24 February 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.106883
