Fault Diagnosis and Optimization of Three-Dimensional Convolutional Neural Network Bearing Based on Differential Evolution Algorithm

Rolling bearings are a key component of rotating machinery, and their condition directly determines the operation of the entire rotating machinery. In the event of a malfunction, it will cause a major accident. After studying the common fault types of bearings and collecting relevant fault data, based on the analysis of the original fault data, a fault diagnosis model based on a three-dimensional convolutional neural network was proposed, and the parameters of the trained model were optimized by differential evolution. In the case of comparing the data input of one-dimensional and two-dimensional convolutional neural networks, the prediction accuracy of the 3D convolutional neural network is 99.16%, and the parameter optimization of the 3D convolutional neural network diagnosis model by differential evolution algorithm results in the optimized prediction accuracy of 99.51%, which is 0.35% higher than that of the fault diagnosis model before optimization.