Characterization of low and high frequency phenomena in a PEM fuel cell using singularity analysis of stack voltage

Abstract

The aim of this work is to propose a new development of the fuel cell characterization tool based on Voltage Singularity Spectrum (VSS), a pattern that can advantageously be estimated from the ``free'' evolution of the stack voltage. A Polymer Electrolyte Membrane Fuel Cell (PEMFC) designed to operate in micro Combined Heat and Power units is investigated to create an experimental database made of normal and abnormal operating conditions. Some VSS signatures based on Wavelet Leader Multifractal Analysis (WLMA) are computed on low- and high-frequency filtered stack voltage signals reflecting different physical phenomena dynamics. The resulting low- and high-frequency VSS computed for six different operating conditions are then compared to the Electrochemical Impedance Spectra (EIS) recorded in the same conditions. In this way, we intend to establish some links between the VSS patterns and the EIS based on well-known physical analyses. We also show that the dominant singularity strengths of the low (and possibly also high) frequency VSS can be considered as relevant clustering features to identify various PEMFC operating states (linked with poorer/better air diffusion, with lower/higher MEA hydration), as it is usually done with the use of the characteristic resistances intercepted on the EIS graph x-axis.