An experimentally derived calibration model for hydrogen-fueled solid oxide fuel cells at intermediate temperature

Abstract

The characteristic polarization curve (e.g. I–V curve) of a solid oxide fuel cell (SOFC) is a valuable outcome of experimental tests because it conveys crucial information about the SOFC's current-potential operational range and optimum performance point. Numerical analysis of the experimentally achieved I–V curve can yield information that is not readily available from it, such as parameters of operation, kinetics of electrochemical reactions, and mechanisms of mass and current transfer. Such details can be used to choose optimal operating conditions and improve the cell manufacturing process. Using experimental polarization curves of a commercial SOFC fed with hydrogen and air, this paper develops a calibration algorithm and extends the analysis to evaluate further operational parameters of the cell. The calibrated model was developed for a commercial SOFC operating at 700 °C and 800 °C. As target parameters, the calibrated model defined the anode and cathode exchange current densities as well as the effective diffusion coefficients of gas species within both electrodes and their respective pore radii. A comprehensive SOFC analysis model requires a detailed understanding of the morphology of the porous material inside the cell, and these parameters provided that insight. Using experimental and analytically optimized polarization curves, target parameters are defined with an accuracy of more than 98%. Based on the developed model, it was also possible to predict the maximum current density the cell could reach at temperatures of 800 °C and 700 °C, which was about 2 Acm−2 and 1.7 Acm−2, respectively. Compared with SOFC parameters reported in literature, the obtained values were fully consistent. In spite of this, the algorithm outcomes are considered “equivalent” parameters since they are solutions to a multivariable problem, and only their combination with other known parameters can result in fitting a benchmark curve. Additionally, their uniqueness criterion is not yet investigated as a means of describing actual cell parameters.