A new hybrid deep signal processing approach for bearing fault diagnosis using vibration signals

Abstract

Signal processing is an important task for machine fault diagnosis. Over the recent years, many deep learning based signal processing methods have been developed for bearing fault diagnosis. However, these methods are facing some major problems when they are applied to machine fault diagnosis. In this paper, a new hybrid deep signal processing method for bearing fault diagnosis is presented. The presented method incorporates vibration analysis techniques into deep learning to form a deep learning structure embedded with time synchronous resampling mechanism. Data collected from real bearing test rig are used to validate and demonstrate the effectiveness of the presented method.