ABSTRACT
The aim of this work is to propose a modeling framework to call into question the maintenance regulations of a train fleet and improve an enforced maintenance policy without disrupting its main structure. The considered fleet is constituted of identical trains. A fixed periodic preventive maintenance planning is imposed by the manufacturers for safety and operational quality requirements. Each overhaul includes a set of pre-defined tasks such as inspection, minimal repair activities and preventive replacement of some components. During inspection if the degradation level of a component exceeds a primary replacement threshold it is precautionary replaced. In case of failure of at least one component when the train is in operation, it is immediately replaced. In this paper we propose the introduction of opportunistic maintenance activities to improve the periodic maintenance strategy while operating within the constraints of imposed planning. Opportunistic maintenance can include early revisions or precipitate preventive replacements of some component (with respect to scheduled preventive replacement or their primary replacement threshold). Dynamic maintenance task can be applied additionally during corrective activities or pre-scheduled revisions. Dependence matrices are used in order to consider stochastic dependencies between the system’s components. Economic dependence is basically derived from a treelike set-up model. A specific replacement threshold dedicated to opportunistic preventive replacement as well as a flexibility degree of revision dates are introduced. Multiple constraint are also considered for optimization: availability, work hours and number of simultaneously revised systems. A comparison between the periodic maintenance strategy with and without opportunistic tasks is presented.
