Model-based prognosis for rolling element bearings

Rolling element bearing is one of the most critical components that determine the machinery health and its remaining lifetime. Rolling contact wear is responsible for the damages initiating on and beneath the contacting surfaces of rolling element bearings. Wear prognosis aims to predict the defect and its evolution progress before it occurs. It addresses the use of automated methods to predict the degradation of physical system performance and the remaining lifetime. Actually, the fundamental need for predictive intelligence tools is to monitor the degradation rather than just detecting the defects, otherwise it will be hard to optimise the asset utilisation in cost effective manner. The current prognosis models are based on predetermined probabilistic damage functions or constant damage factors. Therefore, the purpose of this paper is to develop a prognosis model which is able to predict the evolution of rolling contact wear, using systems dynamics approach. Instead of pre-determined damage functions, the dynamic development of wear process is proposed. The dynamic development of wear considers multiple wear mechanisms and their interactions with respect to surface topological and tribological changes. Moreover, it considers the stress concentration mechanisms and their propagation processes. Therefore, the paper utilizes a five-stage model to simplify the dynamic development of wear progress over the lifetime. The five stages are running-in, steady-state, defect initiation (dentations, pits, and inclusions), defect propagation (extended pits, propagated cracks), damage growth (spalls). The paper is relevant in enhancing the effectiveness of prognostics procedures of rolling element bearing wear.