| Among all applications of hydrogen energy,new energy vehicles are the key application of hydrogen energy in the transportation field.As the key technical equipment of new energy vehicles,fuel cells,especially proton exchange membrane fuel cells(PEMFC)are more potential and have more considerable commercial value.However,there are still obstacles needed to be solved in engineering practice.In this thesis,the porous layer modifications and 3D flow channel structure optimizations based on finite element software and genetic algorithm were investigated in order to improve cell performance.Firstly,two porous layer optimization methods of groove modification and porosity distribution were compared.It was found that the porosity optimization effect is more obvious than the groove modification.Therefore,porosity distribution along single direction and along two directions(along flow direction and perpendicular in flow direction)were designed and simulated.Among the simulation results,an optimal case was found with the porosity distributed by ε=0.5,0.7,0.9,k=0,-4000/3,-2000/3.Compared with the constant porosity,the current density uniformity of the interface between the cathode catalytic layer(CCL)and the cathode diffusion layer(CGDL)was more homogenous and the nonuniformity of the current density was decreased by 18.3%,the saturation of liquid water was decreased by 4.9%,and the current density was increased from 1.61952(constant porosity)to 1.69765 A m-2 with an increase of 4.8%.Then,in the flow field optimization work,three different 3D(three dimension)flow channel structures were compared.It was found that compared with the straight channel and the corrugated plate channel,although the wave channel structure was simple,the cycle and intensity of the speed enhancement along the Y direction(perpendicular to the flow)could be well controlled by the change of the cross section through the adjustment of the structural parameters in order to balance the influence on the transport of reactants and the transport rate of liquid water to the flow channel.Thus,the potential of wave channel on optimization of fuel cell efficiency is relatively high.Finally,based on the wave 3D flow field,a comparative analysis of the flow channel structure parameters was performed and the structure parameters for optimization were determined.By training the neural network as the proxy model of PEMFC,using genetic algorithm,the inclination angle θ,the radius r of the rib fillet and the porosity gradient of CGDL along the Y(perpendicular to the flow)direction were selected as optimization parameters.During the optimization,the porosity gradient was optimized collaboratively with structure parameters to obtain higher fuel cell efficiency while reducing the negative effects such as uneven distribution of reactant concentration and higher pressure drop caused by the design of the flow channel structure.It was found that the optimal model obtained by the genetic algorithm had lower pump power loss and better liquid water removal capacity.As for the optimal case obtained by the genetic algorithm,compared with the constant porosity,the evaluation metric,cell efficiency was increased by 16.8%,the power loss was decreased by 17.3%,and an increase of 8.41% in current density.This showed that not only the inhomogenous distribution of reaction caused by the 3D flow channel design could be solved by the application of the porosity distribution,but also more liquid water in CGDL was removed,the negative impact on the current density caused by the reduce of effective electric conductivity was compensated.In a word,it was beneficial to improve the comprehensive performance of PEMFC. |
PDF Full Text Request