PhD Thesis - Diagnosis of PEMFC by Elechrochemical Impedance Spectroscopy

The proposed PhD thesis is a contribution to the diagnosis and the prognostic tools.

Description of activities:
The work carried out in the former project Health-Code has led to the development of an algorithm for PEMFC which is able to:
– process the EIS data (imaginary and real parts of the impedances and corresponding frequencies),
– extract features which are relevant indicators of the state of health of the fuel cell,
– classify the state of health of the fuel cell in a feature space, previously built and trained
– identify a fault and its intensity

Concerning the prognostic algorithm :
– state of the art on literature basis,
– data mining on long duration tests of PEMFC,
– testbench of different approaches based on artificial intelligence
– development of the prognostic algorithm,
– developing a Human-Machine Interface (HMI),
– implementing the algorithm on an embedded device (hardware to be defined),
– testing the algorithm in real time and evaluating its impact on the lifetime of the PEMC system

For both algorithms, the link with physical phenomena in fuel cell system must be established, which needs a deep knowledge of operation, faults and degradation of such systems.