Deep Convolution Neural Network Motor Fault Identification Based on Generative Adversarial Network under Unbalanced Sample

Abstract

Based on the problem that the traditional motor fault diagnosis method relies on the signal processing power and the model generalization ability is poor, this paper proposes a fault diagnosis method based on generative adversarial network under unbalanced data sets. It builds a small sample training set to train generative adversarial network, and adds the generated sample to the original small sample training set. A deep convolutional neural network (DCNN) model that is suitable for motor fault diagnosis is proposed, and the fault characteristics are learned from the original data layer by layer, so as to realize the identification of different faults. After a lot of experimental analysis, the method is better than the existing depth model in terms of detection rate and error rate.