Multichannel 1D-CNN and Hilbert Transform Based Bearing Fault Classification

Bearings rank among the most vital components of rotary machines, approximately 30% of total defects leading to machine breakdown are attributed to their failure, making it one of the most prevalent issues in machinery equipment that results in severe damage. Therefore, bearing faults detection and diagnosis has been a heated research topic. Recently, the utilization of deep learning algorithms has become widespread in the field of machine health monitoring. Convolutional neural network (CNN) has been extensively studied and applied recently in this paper a one-dimensional CNN (1D-CNN) based Hilbert transform is proposed in order to explore the effect using of the envelope spectrum with the Multichannel input CNN. For the purpose of evaluating the performance of the proposed approach using the accuracy and loss function rate are analyzed, the suggested approach has demonstrated a notable level of classification accuracy.