A Model for Extraction of Spatially Resolved Data from Impedance Spectrum of a PEM Fuel Cell

Tatyana Reshetenko,Andrei Kulikovsky

doi:10.1149/2.0511805jes

Abstract

We report a novel approach to processing of impedance spectra of a PEM fuel cell. We split the cell into N virtual segments and let each segment to have its own set of transport and kinetic parameters. The impedance of a single segment is calculated using our recent physics–based impedance model; the segments are “linked” by equation for the oxygen mass balance in the cathode channel transporting the local phase and amplitude information from one segment to another. Thanks to this transport, the total cell impedance contains information on the local transport and kinetic properties of the cell. We show that fitting the model cell impedance to the experimental spectra yields the parameters of individual segments, i.e., the shape of the cell physical parameters along the cathode channel.

Highlights

Expected wide penetration of polymer electrolyte membrane fuel cell (PEMFC)-based power sources on a market puts forward a problem of fast and reliable characterization of cells and stacks
Fuel cell flooding is a complex phenomenon, which includes accumulation of water produced in the oxygen reduction reaction (ORR) in the cathode catalyst layer (CCL) and the gas–diffusion layer (GDL), formation of small liquid water droplets in the cathode channel, merging of these droplets to form a liquid water slug, and transport of the slugs toward the outlet by the pressure gradient
Eq 9 has been fitted to the experimental spectra using the trust region reflective (TRF) algorithm for nonlinear least–squares fitting procedure from the Python library SciPy

Summary

Introduction

Expected wide penetration of polymer electrolyte membrane fuel cell (PEMFC)-based power sources on a market puts forward a problem of fast and reliable characterization of cells and stacks. According to a quasi–2D approach, the segments are connected by the flow in the cathode channel; the respective equation for the oxygen mass transport along the channel takes into account spatial variation of the cell local parameters. Cell segmentation.—The model Equations 1–6 contain five fitting parameters: b, Cdl , σ0, Dox and Db. The system of Equations 1–6 has been nondimensionalized, linearized and Fourier–transformed to yield a system of linear equations for the perturbation amplitudes of the overpotential and oxygen concentration in the ω–space.[24,25]

Results

Conclusion