Choquet integral parameters inference for health indicators fusion within multi-levels industrial systems: application to components in series
Résumé
Data fusion is a cornerstone step for monitoring component health state within a prognostics and health management (PHM) vision. It enables gathering disparate health information in order to build a global view for the situation evolving over time. Nevertheless for complex systems, efficient health monitoring should be addressed not only at component level but also at higher system abstraction levels since the decision making procedure rely on the global industrial system state. Herein, a novel approach based on the computation of a health index at each system layer function by fusing values within its related indicator vector is proposed. The fusion operator used for this approach is the Choquet integral that requires identifying capacities (Choquet integral parameters) for each function within the system decomposition. In order to take into account relationships between functions in accordance with the considered system layer, a bottom-up approach allowing inferring capacities of each system function from those computed at component level is introduced. This work is the follow-up of a previous one dealing with the case of system elements arranged in parallel. The feasibility and interest of this approach are shown on an application structured with two system layers.