Fuel Cell Manifolds

In a fuel cell system that needs to use two or more stacks to achieve the desired power output, it is convenient to be able to connect the stacks' fluid ports (air, hydrogen, and coolant) in parallel: this allows many balance of plant components to be shared between the stacks, often resulting in weight, space, and cost savings. This was done on the Phase 2 fuel cell bus, and to minimize space needed, the connections were made using a set of manifolds designed at UD. In order for the system to function well, air and hydrogen must be evenly distributed between the stacks, pressure drop in the manifolds must be low, and liquid water must be effectively removed from the hydrogen. To accomplish this in a readily manufacturable design, a multi-layered structure was chosen, as seen below. CFD analysis was carried out with ANSYS FLUENT 12 to optimize the design.

Materials and sealing are interesting and important topics for these manifolds. Simply making the manifolds from metal would not work: the coolant ports are separated from the stacks' bipolar plates by approximately 10mm of coolant, and the bipolar plates are 50 to 200 volts more negative than chassis ground, so a significant current would flow, corroding the manifolds and further increasing coolant conductivity by the addition of metal ions.

Making the manifolds from an engineering polymer to avoid this problem would cause difficulties in maintaining sufficient pressure to keep the manifolds sealed. This occurs since the layers must be held together with metal threaded rods or bolts (polymer fasteners suffer from too much stress relaxation to maintain sealing over time). Since the coefficients of thermal expansion of engineering polymers are much larger than any commonly used metal, simply heating the manifolds from room temperature to the stack's operating temperature would excessively compress the gaskets (if used) or plastically deform the manifolds themselves. Either case causes potential loss of sealing force upon cooling, as was observed in the manifolds supplied with the vehicle by EBus.

To get around this problem, a hybrid solution was devised: the first layer of the manifold handles coolant, and is made of Delrin, with the other layers made of anodized aluminum (one of the few engineering materials that can resist the corrosive effects of wet hydrogen). The Delrin layer is bolted to the stack, and the aluminum layers are connected to the Delrin layer using stainless steel threaded rods and T-section nuts that compress only a small part of the Delrin layer's thickness.

A cross-section view of a tie rod and its installation.

The 0.25 inch of Delrin contributes only a very small amount of thermal expansion, so the elasticity of the stainless rods and the gaskets is sufficient to take up the thermal strain without losing sealing. An even better match could theoretically be achieved using aluminum rods, but it was found that aluminum threads did not allow enough sealing force.

To protect the aluminum layers from corrosion, a graphite electrode was installed in the coolant layer, connected to the vehicle's 12 volt hybrid power circuit through a resistor. The electrode thus becomes a powered anode, so that the grounded aluminum components are cathodically protected.

An exploded view of the inlet manifold for the Phase 2 bus.

The manifolds have been installed on the Phase 2 fuel cell bus and operated successfully, with no leaks and significantly reduced pressure drop over the original manifolds.