Modeling and quantification of aging systems for maintenance optimization

This article deals with the maintenance optimization of a train air conditioning system. Indeed, SNCF (French Railway Company) which has in charge the maintenance of its rolling stock, is involved in research efforts in order to improve its techniques and efficiency in this field. In order to model this system, we use dynamic reliability method, the Piecewise Deterministic Markov Processes (PDMP). A deterministic method is used to calculate the reliability quantities : the finite volumes algorithm. The results found in this study are confidential, so we present results computed with fictive costs and laws. Thanks to this method, we have found a strategy which reduces the maintenance cost of 7% and the system failures number of 22%. Moreover, we observe that in this case, the finite volumes algorithm is faster than the Monte Carlo simulations.