Fuel Cells fault diagnosis under dynamic load profile using Reservoir Computing

Fuel cell (FC) is considered as one of the most interesting solutions to overcome future energy crisis announced by the International Energy Agency. However, various bottlenecks, whether technological or societal, slow the industrial interest for this technology and therefore the mass production of fuel cells. One of these bottlenecks is related to the limited lifetime of FC system. To counter it, the implementation of fault diagnosis and identification methods is interesting. This paper presents an original and experimentally compatible diagnosis approach, named Reservoir Computing. This paradigm, coming from the Artificial Intelligence domain, is an evolution of traditional Artificial Neural Network, with an untrained reservoir of neurons (the Read-Out layer is trained only) instead of the succession of different all-trained layers. Targeted fault types are stoichiometry value faults, pressure drop, temperature drop and failure on the cooling circuit. Experimental results show the simplicity and efficiency of RC method to detect these faults under a dynamic load profile.