A Fusion Prognostics Framework Based on FMMEA and Bayesian Theory

Abstract

Prognostics approaches can assess system reliability in its actual life-cycle conditions, provide advance warning of failure, and reduce system maintenance cost. In all prognostics approaches, identification of appropriate monitored parameters, which can be employed to predict imminent failure, is critical. However traditional approaches, data-driven prognostics and model-based prognostics have their limitations. This paper proposes a fusion prognostics framework, which identifies the appropriate parameters monitored by failure modes, mechanisms and effects analysis (FMMEA), and predicts system remaining useful life by data-driven prognostics based on Bayesian theory. The fusion prognostics framework leverage the strength from both approaches to provide better predictions of remain useful life.