Diagnosis of Inter-turn Short Circuit of Permanent Magnet Synchronous Motor Based on Deep learning and Small Fault Samples

摘要

An efficient and accurate method based on a conditional generative adversarial net (CGAN) and an optimized sparse auto encoder (OSAE) is proposed to detect the inter-turn short circuit (ITSC) problem for permanent magnet synchronous motors (PMSMs). In order to achieve an accurate detection of the ITSC, the CGAN is adopted to augment the few fault samples, and a noise injection strategy is applied to enhance the generalization ability of the network in the framework of the OSAE. Specifically, we made a combination of two types of signals to create a training set that is augmented by the CGAN, and the parameters of the OSAE are determined by the training process of networks. The experimental results indicate that the proposed method for the fault diagnosis of this fault achieves high accuracy 98.9%.(c) 2021 Elsevier B.V. All rights reserved.