Fault Diagnosis Method of Permanent Magnet Synchronous Motor Based on CNN-LSTM-Attention

In this paper, the fault diagnosis method of permanent magnet synchronous motor (PMSM) is studied, and a fault diagnosis model based on CNN-LSTM-Attention is proposed for common faults such as inter-turn short circuit, permanent magnet demagnetization and rotor eccentricity. The model combines the feature extraction ability of Convolutional Neural Network (CNN), the temporal data processing ability of Long Short-Term Memory Network (LSTM) and the feature weighting ability of attention mechanism, which can effectively extract and utilize useful information from fault signals. Experimental results show that the diagnostic accuracy of the model on the test set is not less than 97%, which has the advantages of high accuracy, wide generalization and strong adaptability compared with traditional fault diagnosis methods. The research in this paper is of great significance to ensure the reliable operation of the motor and improve the stability and safety of the whole system.