A case comparison of a proportional hazards model and a stochastic filter for condition-based maintenance applications using oil-based condition monitoring information

Abstract

The ability to predict the expected time remaining before a component fails is crucial when scheduling maintenance activities and component replacements. The current paper presents a comparison of the proportional hazards model and a probabilistic filtering approach when applied to the estimation of a components residual life using stochastically related oil-based wear information. The condition information is collected at irregular monitoring points from aircraft engines and consists of the concentrations of various contaminating metallic particles in an oil sample. Issues regarding the use of multiple information parameters are also addressed.