Fault Feature Extraction Method of a Permanent Magnet Synchronous Motor Based on VAE-WGAN

Liu Zhan,Feng Qian,Xiaowei Xu,Xue Qiao,Qiong Luo

doi:10.3390/pr10020200

Abstract

This paper focuses on the difficulties that appear when the number of fault samples collected by a permanent magnet synchronous motor is too low and seriously unbalanced compared with the normal data. In order to effectively extract the fault characteristics of the motor and provide the basis for the subsequent fault mechanism and diagnosis method research, a permanent magnet synchronous motor fault feature extraction method based on variational auto-encoder (VAE) and improved generative adversarial network (GAN) is proposed in this paper. The VAE is used to extract fault features, combined with the GAN to extended data samples, and the two-dimensional features are extracted by means of mean and variance for visual analysis to measure the classification effect of the model on the features. Experimental results show that the method has good classification and generation capabilities to effectively extract the fault features of the motor and its accuracy is as high as 98.26%.

Highlights

Permanent magnet synchronous motors (PMSMs), which have the advantages of high efficiency, small size, large power density, and wide speed range, have a wide range of applications in production and life
In view of the characteristics of non-stationary, non-linear, multi-source heterogeneity, low value density, and imbalanced fault data collected by online monitoring equipment of permanent magnet synchronous motors, which makes the fault mechanism analysis difficult, this paper proposed a fault feature extraction method based on variational auto-encoder (VAE)-WGAN for a permanent magnet synchronous motor
VAE-WGAN was selected as the fault feature extraction model and its network parameters were set

Summary

Introduction

Permanent magnet synchronous motors (PMSMs), which have the advantages of high efficiency, small size, large power density, and wide speed range, have a wide range of applications in production and life. In the process of long-term operation of the motor some faults may occur, such as electrical faults, single faults of demagnetization faults, mechanical faults, and coupling faults in which multiple faults affect each other [1]. These faults make the motor damaged during use [2], resulting in economic losses and even casualties. It is difficult to simulate all fault types and fault degrees, resulting in incomplete samples; second, the motor may be permanently damaged, resulting in high test costs, and may even affect the safety of laboratory personnel and sites. In view of the difficulty that the number of collected fault samples is too low and seriously imbalanced compared with the normal data, it is necessary to find a reliable sample data expansion method to avoid under-fitting or over-fitting

Objectives

Methods

Findings

Conclusion