A predictive-based maintenance approach for rolling stocks vehicles

In this paper, a model-based maintenance approach is developed for rolling stocks vehicles operating along railway networks. By considering the high management costs in the modern and complex railways fleets as a primary requirement, the key goal of the proposed approach consists in efficiently integrating maintenance actions with the capability to satisfactorily keep railway services. Here, this is achieved by means of multi-layer approach that combine into a single framework the following ingredients: interpolation procedures, machine learning algorithms and prediction arguments that take advantage of an accurate model description of the rolling stock dynamics. Experiments on a PV7 EVO - Matisa, owned by the Italian Railways Network, have been conducted with the aim to show the effectiveness of the proposed maintenance architecture.