A fault diagnosis method based on one-dimensional data enhancement and convolutional neural network

Abstract

When the mechanical fault diagnosis data has the problems of complex fault types and insufficient training samples, it is of great significance to maintain the efficiency and accuracy of fault diagnosis to ensure the safe operation of equipment. In this study, a fault diagnosis method based on one-dimensional data enhancement and convolutional neural network (CNN) is proposed. Firstly, a stacked autoencoder (SAE) improved by a back propagation neural network (BPNN) with a softmax classifier is used to enhance the one-dimensional vibration signal. Then the enhanced data set is directly converted into two-dimensional images to train the CNN. Compared with the existing models in the simulation experiments, it is proved that the model can improve the training accuracy to 98.89% and the test accuracy to 97.25% under the premise of ensuring the diagnosis efficiency when there are many fault types and few training data in the experimental data set.