Least-Squares Parameter Estimation for Catalyst Layer Agglomerate Models

A framework is presented to estimate the micro-structural parameters of cathode fuel cell electrodes by means of a nonlinear least-squares method. This work represents the first attempt in the literature to characterize the structure of the catalyst layer by numerical parameter estimation using a two-dimensional membrane electrode assembly model with an ionomer-filled agglomerate catalyst layer approximation. The framework is developed by coupling a two-dimensional model to an optimization based least-squares algorithm in DAKOTA. The algorithm, NL2SOL, minimizes the sum-of-squares of the residuals for any number of data points and parameters. Employing the proposed methodology allows for accurate characterization of the electrode structure and quantification the quality of the curve fit. Extension of this methodology allows for parameter estimation as novel materials are incorporated into fuel cell construction. Results indicate that curves can be fit using micro-structural and electrochemical parameters consistent with values published in the literature. However, the quality of the fit deteriorates for large data sets over the entire range of operating conditions.