A methodology for probabilistic model-based prognosis

This paper deals with the prognosis of complex systems using stochastic model-based techniques. Prognosis consists in this case in computing the distribution of the Remaining Useful Life (RUL) of the system conditionally to available information. In so doing, three main challenges arise from the industrial context. First, the model should unify the two classical approaches to describing complex systems: the bottom-up and the top-down approaches. The former uses elementary interacting components whilst the latter models the system's physical behavior by means of a set of differential equations. Second, the prognosis must integrate online information to provide a specific result for each system depending on their life events. Online information can take different forms (e.g. inspections, component faults, non detection or false alarm, noisy signal) which must all be considered. Third, the prognosis must supply ready, meaningful numerical results, the error of which must also be under control. This paper proposes a method addressing those challenges. The method is illustrated with two different examples: a simplified spring-mass system and a pneumatic valve for aeronautical application.