An End-to-end Fault Diagnosis Method of Rolling Bearing Based on Convolution Neural Network

Rolling bearing is an important part of industrial machinery, and it’s vital to diagnose what the bearing’s defect is when it fails. With the help of deep learning, the deep neural network can be used to diagnose kinds of faults with automatic diagnosis. However, the neural network can’t identify the faults by using raw vibration data. This paper proposes a novel fault diagnosis method of rolling bearing that using the frequency data as the input of the improved convolution neural network model, where the frequency data is transformed from fast Fourier transform. No manual experience is required in this method. Experiment verification is tested on the two datasets, CRWU dataset and DLUT dataset. And the experiment results show the superiority of the suggested technique which realizes the end-to-end diagnostic method on rolling bearing.