Open AccessFault Diagnosis of Permanent Magnet Synchronous Motor Inter Turn Short Circuit Based on Deep Reinforcement Learning
Author(s) -
Fangli Li,
Han Li,
Song Xu,
Yuanjiang Li
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2137/1/012004
Subject(s) - fault (geology) , computer science , convergence (economics) , reinforcement learning , adversarial system , control theory (sociology) , set (abstract data type) , artificial intelligence , sequence (biology) , electromagnetic coil , algorithm , engineering , electrical engineering , control (management) , seismology , geology , economics , programming language , economic growth , biology , genetics
With the increasing market share of Permanent Magnet Synchronous Motor(PMSM), the fault diagnosis and prediction technology for PMSM is becoming increasingly important. Firstly, in order to solve the problem of insufficient fault sample data consisting of negative sequence current, electromagnetic torque and other inter turn short circuit fault feature terms, the Conditional Generation Adversarial Network(CGAN) is used to expand the data set. Then, with sufficient data, Dueling_DQN algorithm of deep reinforcement learning is used to train and optimize the extended data set. Finally, the effectiveness of the algorithm in the field of PMSM fault diagnosis is verified by simulation training. The results show that the fault diagnosis accuracy of the algorithm can be reached 97.5%, while improved the convergence speed and saved the time cost of fault diagnosis.