Interturn Short Circuit Fault Diagnosis of Brushless DC Motor Based on Image Feature Extraction and Transfer Learning

In this paper, a method for interturn short circuit fault (ISCF) diagnosis of brushless DC motor (BLDCM) based on image feature extraction and transfer learning is proposed. First of all, the three-phase current signals of the motor stator windings are collected synchronously to convert the one-dimensional current signals into two-dimensional image signals. Then the convolutional neural network based on transfer learning is used to diagnose the ISCF. This method reduces the cost of stator winding fault detection in permanent magnet motor system and has potential application value in precise location and diagnosis of stator winding fault.