A fusion prognostics strategy for fuel cells operating under dynamic conditions

Transportation-oriented proton exchange membrane fuel cells (PEMFC) are attracting much attention, but the strong dynamic operating conditions in transportation applications limit durability improvement and wide commercialization of (FC). Prognostics dedicated to predicting FC remaining useful life (RUL) can facilitate early provision control/maintenance programs improve reduce costs. However, credible degradation indexes for prognostics difficult access or observe directly from under conditions. Moreover, long-term prediction performance state-of-the-art models is often not satisfactory. This paper proposes a fusion strategy address these challenges. Specifically, system dynamics identified by using an electrochemical mechanism model extracted based on parameters. Subsequently, reduced-dimensional symbolic representation long short-term memory network developed evolution degradation. The proposed approach validated accelerated stress test data vehicle-oriented PEMFC. results show that be used identify Based model, accurate RUL further achieved over indexes.