Degradation assessment of bearing fault using SOM network

Rolling element bearings are used in a wide variety of rotating machineries, such as pumps and compressors. Bearing fault or even failure is one of the foremost causes of breakdown in such rotating machines, resulting in costly systems downtime. However the bearing fault conditions, e.g., different types or positions of bearing fault and especially different degradation levels have different impact on the machine's operation, causing different level of safety issues. This paper presents a methodology for degradation assessment of bearing fault based on self-organization mapping (SOM) network. The specific faults are considered as cracks in outer race, inner race and ball with different fault diameters. Wavelet analysis is firstly used to extract features especially weak fault characteristics. Using SOM network, different fault mappings are then presented considering various failure modes, degradation severity levels and the possible degradation trajectory, which helps to differentiate each fault condition. It is proved to be a powerful and comprehensive tool for machine's elaborate condition monitoring management that combines in a unified representation both current and predictive perspectives of the fault degradation behavior.