An automated approach for bearing damage detection

Abstract

Bearings are an essential component in all types of rotary equipment. Erosion in the bearing is unavoidable due to radial and axial forces permanently acting on the bearing during the course of rotation. In order to avoid catastrophic breakdown of the equipment, it is a fundamental requirement to monitor the bearings. The scope of the existing bearing fault diagnosis techniques in the literature is limited to only pre-known bearing and machinery. On the contrary, this research develops a generalized protocol for detecting ‘inner’ and ‘outer’ race bearing faults for any unknown rolling element bearing. This automated bearing failure detection model tunes itself adaptively to any type of rotary equipment and the bearing. Automated bearing failure detection is based upon using wavelet transform to scan the spectral contents and applies envelop detection. The raised asynchronous energy in the envelop spectrum is a potential indication for the bearing faults.