Model Predictive Control of Reformate Composition for Use in Solid Oxide Fuel Cells

Fuel cells allow for increased efficiency in power production when compared to the thermodynamically limited efficiencies of heat engines. In the case of solid oxide fuel cells, they are also usable with the fuel infrastructure currently in place (natural gas). Although potentially transformative, solid oxide fuel cells are currently limited by engineering challenges related to operating temperature (> 600° C), durability, and load following ability. For example, the buildup of solid carbon in the stack, or coking, potentially limits one of the most desirable aspects of solid oxide fuel cells, which is their robustness to fuel type. As the working temperatures for SOFCs continue to decrease, in order to maintain fuel robustness, the need for control of the fuel reforming process increases. This work demonstrates the use of a model predictive control algorithm on a catalytic partial oxidation fuel reformer. The controller allows for load following demand changes from the fuel cell and meets those demand changes, while ensuring that the reformate composition is not prone to solid carbon formation. The controller meets current demand changes to within 0.1 amperes/second while maintaining compositional limits on the reformate flow and temperature limits on the stack and reformer.