Diagnosis and prognosis of bearing failure in rotating machinery using acoustic emission and artificial neural network

Abstract

Bearing failure is well-known as a common problem in industries. Therefore, timely diagnosis and prognosis (DAP) of bearing fault is very crucial in order to prevent sudden damages. This paper proposes the practical method of bearing fault DAP using acoustic emission (AE) technique assisted with artificial neural network (ANN). The bearings failure data is measured based on the AE in terms of decibel (dB) and Distress levels, which are used as input for ANN of a bearing fault DAP. For this purpose, an experimental rig is setup to collect data from target bearing by using Machine Health Checker (MHC) Memo assisted with MHC Analysis software. In this work, Elman network with training algorithm, Levenberg-Marquardt Back-propagation is used for ANN DAP. The obtained results indicates that the proposed methods are suitable to inform the remaining useful life time of a faulty bearing. © 2010 The Institute of Electrical Engineers of Japan.