Machine Learning Based Anomaly Detection and Classification of Acoustic Emission Events for Wear Monitoring in Sliding Bearing Systems

Abstract The present study aims at monitoring and classifying the multi-variant wear behavior of sliding bearings. For this purpose, acoustic emission (AE) technique was applied to a test rig for sliding bearings. AE signals were evaluated with machine learning methods in order to detect anomalies from a hydrodynamic bearing operation. Furthermore, a deep learning approach based on convolutional neural networks was used for multi-class classification into three different wear failure modes, namely running-in, inadequate lubrication and particle-contaminated oil. A high accuracy and high sensitivity have been achieved in the detection and classification of three-body abrasion due to particle contamination. In the cases of running-in and inadequate lubrication, the incubation period during the onset of inadequate lubrication is sometimes mistaken for running-in and vice-versa, which reduces the overall accuracy of the classification.