DESIGN AND MODELING OF A LARGE PROTON EXCHANGE MEMBRANE FUEL CELL WITH HIGH HYDROGEN UTILIZATION FOR AUTOMOTIVE APPLICATIONS
OMID BABAIE RIZVANDI
Mechatronics Engineering, MSc. Thesis, 2016
Prof. Dr. Serhat Yeşilyurt (Thesis Advisor), Assoc. Prof. Dr. Selmiye Alkan Gürsel
, Asst. Prof. Dr. Mustafa Fazıl SERİNCAN
Date & Time: August 4th, 2016 – 3:00 PM
Keywords : Proton exchange membrane fuel cell, Flow field design, Species transport, Dead-ended operation, Ultra-low stoichiometric flow condition.
Performance of proton exchange membrane fuel cell (PEMFC) depends on several factors, such as flow fields design, cooling technique, species transport, and water management. In order to enhance the performance of a high power (automotive) PEMFC, three-dimensional model of the anode flow field with ultra-low stoichiometric flow condition and without the effect of species transport, two-dimensional model of the anode flow field with species transport, and three-dimensional serpentine flow fields for the cathode and cooling domains are studied and optimized. In the anode models, widths of the channels and ribs and configurations of their headers are investigated to obtain a uniform flow and hydrogen concentration distribution through the channels. For the anode flow field, two approaches lead to different optimum designs, however, we prefer the one from the two-dimensional model with the mass transport. In the final design of the anode flow field, the hydrogen-depletion region ratio is less than 0.2%. In the cathode model, an unstructured search is used to obtain a design that has a pressure drop within 30% of the output power. In the cooling model, dimensions of the channels and ribs, and pressure difference between the inlet and outlet manifolds are investigated to find a uniform temperature distribution through the cooling plate with index of uniform temperature (IUT) less than 3 °C. Finally, a one-dimensional model of species and liquid water transport and distribution through the anode and cathode channels and their gas diffusion layers (GDLs) is studied. Results of this model agree reasonably with experimental data.