Monitoring of damage development of a ball bearing in service with self-organization mapping method for AE waveforms.

Abstract

The purpose of this study is to evaluate damages in a ball bearing with an acoustic emission technique. Many techniques for estimating bearing damages have been proposed ,and utilize vibration or acoustice emitting from a bearing in service. Acoustic emission technique methods is expected to detect the damage in early stage because the technique can detect high freqrency small vibration. However AE method can not identify the damage modes such as wear, flaking, and fatigue. In this studuy, identification of the damage mode on bearings in service with the AE waveform classfication was curried out with a self organization mapping (SOM) method which is one of neural networks. SOM migiht be classified the AE signals into three groups corresponding to each damage modes on the raceway obsereved with an optical microscope.