Retrofitting a two-phase flow pressure drop model for PEM fuel cell flow channel bends

During the operation of a proton exchange membrane (PEM) fuel cell, water and heat are produced as byproducts. The water produced in the cathode electrode can pass through the porous structure of the electrode and emerge from the surface of the gas diffusion layer, GDL. This forms liquid-gas two-phase flow in the cathode flow channels. The accumulation of liquid water in flow channels can be quantified by measuring the two-phase flow pressure drop. However, such quantifications require an accurate two-phase flow pressure drop model to correlate the water flow rate to the pressure drop. Literature on two-phase flow pressure drop in PEM fuel cell flow channels mainly focuses on the pressure drop in straight sections of the channel, but less attention has been paid to the pressure drop across the flow channel bend. However, many PEM fuel cell flow field designs have multiple bends across their channels. One common configuration is the 90° bend, which is used in serpentine flow fields. Therefore, the liquid-gas two-phase flow pressure drop across flow channel bends should be investigated properly. One of the most well-known models that has been developed to predict the two-phase flow pressure drop across 90° bends was proposed by Sookprasong et al. [1]. However, this model was developed for applications where the flow is dominated by the liquid phase in large pipes. As the two-phase flow in PEM fuel cell flow channels is dominated by the gas phase, the accuracy of this model for the proposed application should be carefully studied. Results obtained in this study indicated that this model cannot predict the two-phase flow pressure drop across PEM fuel cell flow channel bends. Instead, some modifications are needed to customize this model for this application. In this study the Sookprasong et al. model [1] was retrofitted for the application of PEM fuel cell flow channel bend. The modification included replacing the liquid-phase loss coefficient with the two-phase loss coefficient. When this adjustment was made the model’s prediction capability significantly improves. A mean absolute error in pressure drop prediction as low as 10.97% was obtained when the two-phase flow pressure drop in the straight section of the channel was used from the experimental data.