Mu-Analysis Based Controller Architecture Evaluation for the Fuel Processor in a Fuel Cell Powerplant

Fuel cell power plants with an integrated fuel processing system for hydrogen generation are an attractive option for stationary, distributed power generation. We consider a PEM fuel cell power plant with CPO based fuel processor for generating hydrogen from natural gas. During load transients, the temperatures at various points of the fuel processor need to controlled tightly so that the catalytic reactors function properly. The control of the CPO reactor temperature is particularly challenging due to the highly exothermic oxidation reactions, short residence time and the sensitivity to stoichiometry. We consider two architectures for the CPO exit gas temperature control: the first is a baseline controller designed using a decentralized approach and the second is an advanced controller, designed using multivariable control approach. The advanced controller gives better decoupling and disturbance rejection performance. We employ μ analysis to study the robust stability and robust performance of the two controllers. This analysis shows that the advanced controller is superior. Using the notion of skewed-μ, we establish that, for a prescribed level of performance degradation, the worst-case uncertainty that can be tolerated by the advanced controller is higher than the baseline controller.