Sensitivity analysis and uncertainty quantification in predictive modeling of proton-exchange membrane electrolytic cells

Insights into degradation mechanisms, lifetime prolongation, and efficiency boost-up of proton-exchange membrane (PEM) electrolytic cells (ECs) is crucial for a quick technology ramp up. Besides experiments, macro-scale modeling can contribute to the better understanding of complex mechanisms in PEM electrolytic cells (PEMECs). Modeling of PEMECs is subjective to uncertainties in operating conditions, material and design properties. Uncertainty quantification (UQ) and sensitivity analysis (SA) can highlight how uncertainties propagate from the input parameters to the output of PEMEC models. However, UQ has not found its way into PEMEC modeling yet. Thus, we apply a quasi-Monte Carlo (QMC) and polynomial chaos expansions (PCEs) framework to quantify uncertainty in a highly nonlinear PEMEC model. Specifically, we implement a fast 1D python solver, which solves a system of six nonlinear coupled ordinary differential equations (ODEs) in PEMECs. The solver robustness is an asset for the conducted UQ and global SA study since the computational model needs to be evaluated several times. The UQ and SA analysis underlines how stochastic simulations improve the prediction accuracy of PEMEC models. It also enlightens the function of PEMECs. For example, the predicted cell voltage is very sensitive to the thickness of PEM, especially when the current density increases.