An Echo State Network for fuel cell lifetime prediction under a dynamic micro-cogeneration load profile

Improving Proton Exchange Membrane Fuel Cell durability is a key that paves the way to its large scale industrial deployment. During last five years, prognostics discipline emerged as an interesting field for state of health prediction and lifetime estimation. The information provided by prognostic module crucial optimizing control strategy extend fuel cell lifetime. In this paper, approach based on Echo State Network under variable load developed. novelty paper perform profile without prior knowledge latter. Two solutions are developed in work. first one consists evaluating remaining useful repeated cycle. second using Markov chains generate estimations future profile, allowing thus overcome need real knowledge. Both proposed give accurate results proton exchange membrane lifetime, with low uncertainties.