Artificial Intelligence Approach for Fuel Cell Model Parameter Calibration

Electrochemical model of a fuel cell involves several parameters which influence its polarization curve. For a numerical fuel cell model to match experimental polarization curve, it is critical to find the right values of these parameters. It is hard to find the values of all the parameters experimentally, and hence parameter calibration is required. A fully automated workflow for calibration of fuel cell model parameters in a three-dimensional Computational Fluid Dynamic (CFD) simulation is created. The CFD model captures detailed electrochemistry and water phase change. The CFD polarization curve is generated by sequentially running a series of simulations starting from low current densities to high current densities. Experimental polarization curve is used as the validation target. An objective function is defined as the L² norm of the difference between the experimental and the CFD generated polarization curve measured at various current densities. For calibration, eight fuel cell parameters are chosen as input parameters namely, the cathode and the anode exchange current densities, protonic conduction coefficient, cathode and anode gas diffusion layer (GDL) porosity, exponent of pore blockage, effective active area exponent and liquid removal coefficient. An Adaptive Metamodel of Prognosis (AMOP) is used to drive the optimization objective function to its global minimum. The automated workflow runs multiple simulations with different parameter values in parallel on high performance computing cluster, thus accelerating the calibration process. The polarization curve generated by the calibrated model matches accurately with the experimental data with an error L² norm of less than 3%. The automated workflow eliminates the need for manual trail-and-error processes which can be uncertain and laborious.