Particle Filter-Based Method for Prognostics with Application to Auxiliary Power Unit

Particle filter (PF)-based method has been widely used for machinery condition-based maintenance (CBM), in particular, for prognostics. It is employed to update the nonlinear prediction model for forecasting system states. In this work, we applied PF techniques to Auxiliary Power Unit (APU) prognostics for estimating remaining useful cycle to effectively perform APU health management. After introducing the PF-based prognostic method and algorithms, the paper presents the implementation for APU Starter prognostics along with the experimental results. The results demonstrated that the developed PF-based method is useful for estimating remaining useful cycle for a given failure of a component or a subsystem.