Lattice Boltzmann Simulation of Multiphase Transport in Nanostructured PEM Fuel Cells

As the fossil fuel crisis becomes more critical, it is imperative to develop renewable sources of power generation. Polymer electrolyte membrane (PEM) fuel cells are considered a viable option. However, the cost of the platinum catalyst has hindered their commercialization. PEM fuel cells with platinum loading of >0.4 mg cm2 are common. Efforts towards further reducing this loading are currently underway utilizing nanostructured electrodes. A consequence of increased platinum utilization per unit area and thinner nanostructured electrodes is flooding, which is detrimental to fuel cell performance. Flooding causes a two-fold impact on cell performance: a drop in cell voltage and a rise in parasitic pumping power to overcome the increased pressure drop, which together result in a significant reduction in system efficiency. Proper water management is therefore crucial for optimum performance of the fuel cell and also for enhancing membrane durability.;The goal of this thesis is to simulate the multiphase fluid transport in the nanostructured PEMFC of H2O in air with realistic density ratios. In order to pursue this goal, the ability of the pseudopotential based multiphase lattice Boltzmann method to realistically model the coexistence of the gas and liquid phases of H2O at low temperatures is explored. This method is expanded to include a gas mixture of O2 and N 2 into the multiphase H2O systems. Beginning with the examination of the phase transition region described by the current implementation of the multiphase pseudopotential lattice Boltzmann model. Following this, a modified form of the pressure term with the use of a scalar multiplier kappa for the Peng-Robinson equation of state is thoroughly investigated. This method proves to be very effective at enabling numerically stable simulations at low temperatures with large density ratios. It is found that for decreasing values of kappa, this model leads to an increase in multiphase interface thickness and a reduction in maximum spurious velocities. It is shown that although low temperatures and large density ratios are attainable using these modifications, the lowest temperature results should be discarded due to the non-physical density variations along the phase interface. Building on this insight the simulation of a liquid droplet of H2O at 73 degrees C surrounded by humidified air (a mixture of H2O, N2, and O 2) with realistic density ratios is demonstrated.