Water management in a polymer electrolyte fuel cell is a critical issue in ensuring high cell performance. The water production, due to both electro-osmotic drag and the chemical reaction, in the cathode side of the fuel cell leads to liquid water formation in the gas diffusion layer and the reactant flow channel. If this water is allowed to accumulate in the fuel cell, the transport of reactants to the membrane assembly will be inhibited and cell performance will suffer. In order to maximize the potential performance of a fuel cell, understanding of the liquid water dynamics is required, and a two-phase flow numerical model has been for this purpose. If an accurate numerical model can be created the development cycle for new flow channel designs can be accelerated.

The methodology adopted for the numerical simulation of dynamic two-phase flow is the volume of fluid (VOF) method. The major drawback of current VOF models is in the implementation of the three phase contact line. Current models use a constant static contact angle, which does not take into account real dynamics. This results in non-physical phenomena such as spherical and suspended droplets, instead of the experimentally observed attached semi-spherical droplets with the trailing edge of the droplet forming a tail.

To remedy this shortcoming, the implementation of a dynamic contact angle relation is required. The relations used in the current work follows the Hoffman formulation where the dynamic contact angle is obtained as θD. A function of the capillary number, based on the contact line velocity, and of the equilibrium contact angle. The function was implemented within the commercial CFD framework of Fluent using user defined functions.

The dynamic contact angle models were able to better predict the droplet dynamics, providing elongated droplet profiles. The dynamic contact angle model was also able to provide more realistic pressure profiles down the channel length. Parametric studies show the dramatic effects that air speed and static contact angle have upon the droplet dynamics. It was also observed that water injection velocity had a relatively small effect on the model. The dynamic contact angle model was found to be consistent with experimental work conducted in our laboratory in which the spinning motion of the fluid within the water droplet was observed [7].

The improved physical representation achieved with the new model results in more reliable simulations and provides a good foundation for the numerical modeling of fuel cell flow channels.