High cost is currently a barrier to PEM fuel cell commercialization. Therefore, changes must be made to existing designs in terms of architecture, materials or manufacturing methods. Traditionally, PEM fuel cells are water-cooled. However, ultimately all of the fuel cell's waste heat needs to be rejected to the surrounding air. This thesis focuses on the feasibility of alternative direct air-cooled designs. Thus, combining the previously separate fuel cell and radiator devices. This leads to a larger stack design, but offers the benefits of increased system simplicity, and potentially improved reliability and lower cost. Air-cooled systems eliminate an intermediate cooling medium, have fewer components, and with effective integration, can use materials well-suited for heat rejection. Both ambient air and tandem, or dual, air systems were examined. Performance goals for the new Radiator Stack Architectures (RSA) developed by Next­-Generation Fuel Cells for Transportation (NGFT) were based on the best existing PEM fuel cell technology, New Electric Car II (NECARII), jointly developed by Daimler Benz and Ballard Power Systems in 1995.

System modeling efforts examined the functional performance of six air-cooled architectures. Modeling and experiments concluded that low-grade heat produced by the fuel cell stacks described, could be rejected, with acceptable parasitic losses using a variable speed fan. The TERS97 design, a focus of much work, was found to have volumetric and gravimetric power densities 37 percent and 19 percent lower, respectively, than NECARII.

A method of rapidly prototyping gas delivery plates was also explored using CAD/CAM and screen printing or CNC machining as target manufacturing methods. The feasibility of the rapid prototyping method was proven through two programs, one which examined general flow field design, and another which generated various computer models for Ballard Power Systems ' MK-5 stack design.