Abstract
Abstract
Aiming at the difficult problem of inter-turn short-circuit fault diagnosis of permanent magnet synchronous motors without stopping the machine, this paper takes timing anomaly detection as the starting point, and proposes an online fault detection method based on transfer learning deep self-coding network. First, a migration deep self-coding network model is constructed, and a combined sample of negative sequence current and electromagnetic torque is used to extract common features in different domains. Then, the online detection model of abnormal timing is established. Finally, the permutation entropy value of the normal state of the motor is used to construct an alarm threshold to improve the matching speed of abnormal sequences in the online data. Experimental results show that the method in this paperhas better detection practicality.
Reference12 articles.
1. Research status and challenges of deep learning in the field of fault diagnosis [J];Hao;Control and Decision,2017
2. Application of wavelet noise reduction and Hilbert transform in motor bearing fault diagnosis [J];Feng;Journal of Motor and Control,2017
3. Time-frequency atoms-driven support vector machine method for bearings incipient fault diagnosis;Liu;Mechanical Systems and Signal Processing,2016
4. Extraction of early fault features of rolling bearings using optimal wavelet scale cyclic spectrum [J];Rui;Journal of Mechanical Engineering,2018
5. Weak fault diagnosis of rotating machinery based on feature reduction with Supervised Orthogonal Local Fisher Discriminant Analysis;Li;Neurocomputing,2015
Cited by
2 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献