Inverter and Sensor Fault Diagnosis of PMSM Drive System based on Improved WOA-LSTM

Aiming at the problems of low fault diagnosis accuracy and poor model generalization of permanent magnet synchronous motor (PMSM) drive system, a fault diagnosis model method combining Improved Whale Optimization Algorithm (IWOA) and Long Short Term Memory neural network (LSTM) was proposed. Firstly, the multi-dimensional features of motor signals are extracted by multi-scale feature fusion method, and the high-dimensional feature sample set is obtained. Secondly, the proposed Improved WOA is used to optimize the LSTM network model hyperparameters. Then, the improved WOA-LSTM model is used to realize the fault diagnosis of PMSM drive system inverter and sensor. Finally, the effectiveness of the proposed method is verified by comparing it with typical fault diagnosis methods.