Maintenance grouping models for multicomponent systems

International audienceThis chapter provides some contributions for maintenance optimization of multicomponent system in the arising context of on-line monitoring and dynamic decision-making under specific constraints. The key point is, on one hand, to develop maintenance models taking into account economic dependence and some maintenance constraints. On the other hand, it is important to provide calculation and optimization tools in order find an optimal grouped maintenance planning and to update a it in presence of a “dynamic context” which may randomly occur on-line. The modeling framework that is chosen is mainly based on reliability theory and stochastic modeling. In that way, the chapter presents firstly several models for maintenance grouping based on the rolling horizon. These models are then developed taking in consideration of maintenance constraints such as limited maintenance teams and/or restricted maintenance duration which may significantly impact the maintenance planning. The proposed models can help not only to establish an optimal grouped maintenance planning but also to efficiently update it in dynamic context (occurrence of maintenance opportunities, changes of repairmen, availability level, etc.). To reduce the computing time, optimal algorithms such as MULTIFIT, generic algorithms are implemented. Several numerical examples are also introduced to show the use and the performance of the proposed models for maintenance grouping of multicomponent systems. The application area is very large and can encompass systems in the process, production, transportation industries as well as in critical infrastructures