Inverse problem method applied to flow field optimization of proton exchange membrane fuel cells

An optimization approach,combining a simplified conjugate-gradient method(inverse problem solver) and a three-dimensional,two-phase and non-isothermal fuel cell model(direct problem solver),was developed to determine the key geometric parameters of a proton exchange membrane fuel cell.In this approach,with channel height as the searching variable(optimized object) and the reciprocal of cell output power density as the objective function,the optimum channel height(optimized design variable) was derived from searching the minimum of the objective function.The results show that for the optimized serpentine design,except the outlet channel being diverging,the other channels should be tapered.The cell performance is,meanwhile,improved by 11.9% compared to the convectional serpentine flow field under the same operating conditions.A detailed investigation of local transport characteristics reveals that the tapered channel design enhances sub-rib convection,leading to more oxygen transport over the cell and more effective liquid water removal out of the cell;however,the diverging outlet channel can provide relatively proper sub-rib convection to prevent reactants from "short-circuit",which means that reactants directly flow out of the cell and thus results in reactant waste.