Modeling of single-phase flow through pipes and channels of arbitrary cross sections from laminar to turbulent regimes is fairly well understood and Computational Fluid Dynamics (CFD) is commonly employed to predict pressure drops and other flow behaviors. Several attempts have been made by many researchers to develop CFD based predictive tools to describe multi-phase flows with varying degree of simplification and a companion paper on the subject presents a two fluid model approach in the context of Fuel Cells.

Still a commonly employed engineering approach is to use Lockhart & Martinelli type of empirical correlations in two-phase fluid flow. Straightforward applicability of such empirical correlations for fuel cell application is much more doubtful as liquid water is added along the length of the channel due to electrochemical reaction and typically some estimate of liquid water present in the flow channels is required to correctly describe the pressure drop. In as much as this is a good engineering simplification the flow behavior in fuel cells is different from the classical two-phase flow where an incoming quality defines the second phase entering the domain, which is the starting point of such empirical correlations.

In this paper empirical correlations based on the experimental data for different mini-channels used in PEM fuel cell cathode flow fields are described. These correlations are used for describing the two-phase flow pressure drops in cathode flow fields of PEM fuel cells. These empirical correlations are also compared with Lockhart & Martinelli type models based on assumed liquid volume in channels. The developed correlations serve as a good water management design tool for the fuel cell designers.

Furthermore, in absence of such empirical correlations, it is not possible to account for significant pressure drops in the cathode channel where liquid water, that is the product of the electro-chemical reaction as well as due to other water transport mechanisms, is present in the flow channel and significantly influences the pressure drop. Typically in some applications, because of the liquid water, the pressure drops can be three times higher than that based on laminar flow consideration alone and this cannot be predicted by a single-phase CFD model in absence of a “laminar like” friction factor in the model. In addition to posing the water management problem for PEM fuel cell flow fields, this significantly impacts the concentration profile of reactants along the channel and the CFD based fuel cell model must be able to account for such pressure variations. In this paper the developed correlation is used in a CFD based fuel cell model and predicted pressure drops and performance of fuel cell are compared with experimental data.