A global approach for a consistent identification of static and dynamic phenomena in a PEM Fuel Cell

Abstract In this paper, we propose a parameterization process for static and dynamics models dedicated to the analysis of two typical characterizations of a Proton Exchange Membrane Fuel Cell (PEMFC): the polarization curve (V–Icurve) and a set of Electrochemical Impedance Spectroscopies (EIS) carried out for several current densities of this curve. The specificity of the proposed approach is to consider conjointly both characterizations during all the proposed analysis process. This global strategy ensures the separation of the different fuel cell phenomena (activation, diffusion and ohmic) in the static and dynamic domains by imposing, with different ways, an expected consistency between both characterizations. Starting from the measured polarization curve, a first parametric identification is carried out leading to a static model of the Fuel Cell (FC). The EIS data are, at this stage, used to obtain the ohmic resistance of the static model. This is the first consistency forced between both characterizations. By fixing the transfer coefficient to 0.5 (platinum catalysts), it is possible to separate unequivocally the different fuel cell phenomena (activation, diffusion and ohmic) and to identify the different parameters of the laws which describe them in steady state for a given Membrane Electrode Assembly (MEA). In the objective of describing the dynamic of these three phenomena, an identification process for generic models involving RC cells in series and without a priori (number of RC cells not presupposed) is secondly applied to the set of measured EIS. Thanks to time-constant spectra (not classically used in fuel cell world) handled in this approach, the time-constants related to the three phenomena are extracted. The first specificity is to force the consistency between V–I curve and EIS by calculating the activation and diffusion resistances thanks to the derivation of the physical laws used for the static model and parameterized in the first step of our approach. The second specificity consists in forcing the equality between the V–I slope and the “zero-frequency resistor” of the EIS for a given characterized density current. This results in the apparition of a residual part for the impedance involving different suffered phenomena (potentially platinum oxidation, channel pressure oscillations…). The characteristics of this residue are analyzed. To calibrate the proposed method and to demonstrate its sensitivity to changes that may occur in the FC components, experiments concerning a single cell with different sets of components (different membrane thicknesses and different platinum loadings in the Active Layer (AL)) were achieved and analyzed by applying this method. Cyclic voltammetries were carried out in addition of V–I curves and EIS to check the relevancy of the parameters identified through our approach.