Kunal Jain

Lattice-Boltzmann Simulations of Multiphase Flows in PEM Fuel Cell GDLs and Micro-channels

INTRODUCTION AND SIGNIFICANCE Effective water management is essential for improved power density and freeze-thaw durability in automotive applications of Proton Exchange Membrane Fuel Cells (PEMFCs). This requires efficient removal of excess water from the cathode catalyst by capillary action through the porous gas diffusion layer (GDL). Understanding of the two-phase transport of liquid water and gaseous reactants within these porous materials is limited, due primarily to the challenges of in-situ diagnostics for such thin, optically opaque materials. Transport is typically analyzed by fitting Darcys Law type expressions for permeability in conjunction with semi-empirical capillary pressure relations. After liquid water emerges from the GDL, it must be removed from the cell through micro-channels that are being continually reduced in characteristic size to meet power density requirements. As the water emerges from the GDL into these channels, it may evolve as droplets, form films, or produce liquid slugs. Correlations based on two-phase flow regime maps are typically used to estimate the ‘wet’ pressure drops in the micro-channels, and the effects of liquid water on both pressure drops and mass transport of reactants in the micro-channels are typically neglected in fluid dynamics models of the cell performance.

A Multiphase, Two-Fluid Model for Water Transport in a PEM Fuel Cell

INTRODUCTION AND SIGNIFICANCE Water management is one of the main challenges in PEM Fuel Cells. While water is essential for membrane electrical conductivity, excess liquid water leads to flooding of catalyst layers and consequent reduced cell power. Despite the fact that accurate prediction of twophase transport is critical for optimizing water management, understanding of the two-phase transport in fuel cells is relatively poor. Wang et al. (1,2) have studied the two-phase transport in the channel and diffusion layer separately using a multiphase mixture model. The model fails to accurately predict saturation values for high humidity inlet streams. Nguyen et al. (3) developed a twodimensional, two-phase, isothermal, isobaric, steady state model of the catalyst and gas diffusion layers. The model neglects any liquid in the channel. Djilali et al. (4) developed a three-dimensional, two-phase, multicomponent model. The model is an improvement over previous work, but neglects drag between the liquid and the gas phases in the channel. To enable model-based design and optimization of PEM fuel cells, particularly for automotive applications, models must address a broad range of conditions including operation with significant liquid water in the channels as droplets or films.

Power map modeling in integrated circuits and power devices

In the following paper, a flip-chip package structure is modeled using electronics cooling solver ESI Presto. Two numerical approaches are investigated to model power sources referred to as “Power Map”. First approach is to model sources as rectangular patches on a surface. This approach requires geometry re-meshing. Second approach is to model sources as points which eliminates the need to re-mesh. Hence, large number of sources can be modeled without increasing the grid count or simulation execution time. Results from these two approaches are compared and they match very well for the current mesh. A parameter sensitivity analysis is performed by varying power map parameters such as power amplitude, Gaussian pulse width and location on the die. The relation between input power and average temperature rise is linear in these simulations as expected. The point source method is used to demonstrate a case with very large number (10,000) of sources.