Performance estimate for a Proton Exchange Membrane Fuel Cell: Sensitivity Analysis aimed to Optimization

A fuel cell based system performance is basically determined by the amount of current density the stack is able to produce, given a well-defined quantity of reactants flowing through it. Starting from a Proton Exchange Membrane fuel cell (PEMFC) distributed parameters model, considering all the aspects influencing the cell behavior, a Multidisciplinary Design Optimization (MDO) process based on a surrogate model is presented. A Monte Carlo Simulation approach is chosen to perform a sensitivity analysis to estimate the effects of key parameters on performances. This analysis allows the definition of a ranking Pareto plot to operate design variables reduction, decreasing the problem complexity and increasing the orthogonality of the input design matrix. The main purpose is to find a suitable and validated method able to reduce the time expense for a complete simulation, so to originate useful input to a multi-disciplinary design optimization.