Sensitivity analysis of proton exchange membrane fuel cell performance to operating parameters and its applicability assessment under different conditions

Highlights: • The parameter combinations were optimized using Box-Behnken Design (BBD). • The effects of operational parameters on the PEMFC are analyzed. • The sensitivity of each parameter is calculated and its applicability is verified. The performance of proton exchange membrane fuel cells (PEMFCs) varies greatly under different operating conditions. Reasonable combination of operating parameters can extend the life of fuel cells and promote the output efficiency. In this paper, a three-dimensional multi-component transport CFD model coupled with electrochemical calculation is developed for cell simulation, and then the sensitivity of PEMFCs performance to each parameter are analyzed. Firstly, to avoid a large number of parameter combinations and reduce computations, Box-Behnken design (BBD) is adopted for the arrangement of numerical cases. Then based on numerical simulation, current density and oxygen molarity distribution on the cathode film surface of each case are obtained. The fitted equations of current density and gas concentration to operating parameters are deduced through multiple quadratic regression equations, and then the optimal combination of parameters for current density is obtained on the basis of Response surface methodology (RSM). Subsequently, the sensitivities of each parameter under optimal condition to current density and internal gas molar concentration are calculated, respectively. The results show that the change in operating pressure has the most significant impact on both current density and gas distribution for PEMFCs and the next are the changes in temperature and relative humidity on both sides. Besides, the change of porosity will remarkably affect the uniformity of gas distribution. To further assess the applicability of PEMFCs sensitivity under different conditions, the sensitivities of each parameter at optimal condition are compared with that originates from wide range of conditions. It is found that there is not much difference between the sensitivity under the optimal condition and sensitivity deduced from the initial condition where each parameter has a wide range of scope, which confirms the generality of the sensitivity obtained under the specified conditions.