Three-dimensional Multiphase Full-cell Modeling And Porous Media Flow Field Design Of Proton Exchange Membrane Fuel Cells

Proton exchange membrane(PEM)fuel cell is a high-efficient and environmental-friendly electrochemical conversion device that converts the chemical energy stored in reactants into electricity with only water as byproduct.During normal operation of PEM fuel cells,multiple heat and mass transfer processes and electrochemical reactions occur simultaneously,mainly including gas/liquid two-phase flow,species transport,water evaporation/condensation,membrane water absorption and release,ionic/electric conduction,and heat transfer,etc.Properly optimizing these transport processes to improve cell performance and durability,i.e.the water and thermal management,is of primary importance in promoting the large-scale commercial application of PEM fuel cell.In comparison with experimental characterization,developing multiphase multi-physics models has been a cost-effective and powerful tool in revealing the detailed transport mechanisms and is popular in fuel cell development.In this dissertation,three comprehensive three-dimensional(3D)multiphase non-isothermal full-cell models of PEM fuel cells were developed,which includes detailed channel two-phase flow,gas diffusion layer(GDL)anisotropy and complex mass transport in catalyst layer(CL).In these three full-cell models,the volume of fluid(VOF),Eulerian-Eulerian and two-fluid models are respectively used to solve the channel two-phase flow,in which the VOF and Eulerian-Eulerian models are able to take the surface tension and wall adhesion into account.But the liquid water amount in channel is underestimated due to the gas/liquid velocity ratio of 1,which is contrary to the reality.By treating fuel cell channel as porous media considering its small length dimension,the two-fluid model used in this dissertation is proposed based on the two-phase Darcy’s law,in which the gas/liquid velocity ratio is determined by the gas/liquid dynamic viscosity ratio and liquid saturation.A sub-model of liquid water coverage at the GDL surface is also developed and incorporated into the model to study the impact of various contact angles on cell performance.The liquid saturation profiles in the cathode channel at different relative humidities and stoichiometric ratios agree reasonably with literature results.Meanwhile,the incorporation of CL agglomerate model makes it more accurate in predicting the concentration loss and hence the performance of PEM fuel cell,especially at high current density regime.After the model validation regarding the polarization curves,ohmic loss and spatial current density distribution at various operation conditions,this model was used to investigate the effect of various flow field designs on transport process and cell performance,including conventional channel-rib configuration and porous media.In multi-channel flow field,the“dot matrix” in the manifold area is able to make the gas concentration distribution more uniform,but disadvantageous to liquid water discharge.In comparison,the performance of the multi-channel serpentine flow field was found to be a little higher,as a result of the enhanced convection effect from flow field to porous electrodes.Compared with the channel-rib configuration,the 3D fine mesh and foam material flow fields were found to increase fuel cell performance with more uniform current density distribution.This mainly benefits from that the porous structure largely increases the mass transfer volume.The incorporation of the full morphology of 3D fine mesh and metal foam material also enables accurate prediction of the ohmic loss at the GDL/flow field interface.And it was found that the decreased contact area may increase the ohmic loss due to the reduced current collection area,which may decrease cell performance at medium current density regime.A data-driven surrogate model based on support vector machine(SVM)was applied to optimize porous media flow field geometry.The complex structure was simplified as a 3D structured mesh consisting of a set of fibers with each perpendicular to the others at the joint.The results show that the data-driven surrogate model based on SVM predicts similar results as the 3D physical model.The genetic algorithm(GA)is further used for optimization,in which the data-driven surrogate model is selected as a fitness evaluation function.The optimal values obtained by the surrogate model were verified by the 3D physical model,indicating that the proposed data-driven surrogate model is effective in the design and optimization of porous media flow field.Finally,the 3D multiphase non-isothermal simulation is extended to a small 5-cell stack.And it was found that the temperature variation in stack is the main cause to the observed non-uniform distributions of the oxygen concentration,liquid water,current density and membrane hydration.In addition,the non-uniform oxygen concentration distribution in the stack is resulted from the water vapor dilution due to the increased water saturation pressure at high temperature.The poor and non-uniform cell performances in this 5-cell stack are mainly caused by the non-uniform and diluted oxygen concentration distribution and low membrane hydration.