Motor fault diagnosis and analysis method based on transfer learning and multi-source information fusion
The invention provides a motor fault diagnosis analysis method based on transfer learning and multi-source information fusion. The method comprises the following steps: designing two variable working condition motor fault diagnosis models based on parameter transfer and feature transfer; establishin...
Gespeichert in:
Hauptverfasser: | , , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention provides a motor fault diagnosis analysis method based on transfer learning and multi-source information fusion. The method comprises the following steps: designing two variable working condition motor fault diagnosis models based on parameter transfer and feature transfer; establishing a motor fault diagnosis model based on multi-source information fusion, and performing feature extraction and diagnosis identification on the current signal and the vibration signal by using a deep residual network; and a multi-source information decision-making layer fusion method is introduced. According to the method, the JMMD is introduced to be combined with the maximum average deviation to realize model parameter adjustment and feature space distribution adaptation, the improved D-S evidence theory is utilized to fuse the identification results and output the final classification result, and the reliability of the diagnosis result can be effectively improved.
本发明提出了基于迁移学习和多源信息融合的电机故障诊断分析方法,包括以下步骤:设计基于参数迁移与基 |
---|