(528f) Multi-Scale Modeling and Control of a PEM Fuel Cell

At the heart of a Polymer Electrolyte Membrane Fuel Cell (PEMFC) is a Nafion membrane with a functionality that greatly depends on water content. When membrane hydration is low, fuel cell performance degrades due to low ionic conductivity. When membrane hydration exceeds a certain limit, the membrane saturates and flooding occurs. During flooding, oxygen is prevented from reaching the reaction sites and again performance drops. We have found that load changes greatly influence this hydration profile. While advanced flow channel design can improve the situation for a given load condition, a high level controller will be needed to manage load requests while simultaneously maintaining a specific hydration profile.

During simulation of the PEMFC, we have observed two distinct time scales within the dynamics of the fuel cell process. The fast time scale consists of electrochemistry and the dynamics of the gases within the flow channels. The slower time scale contains the solid material temperature dynamics as well as the dynamics of hydration within membrane. The method of singular perturbations allows one to approximate the fast time scale aspects with algebraic equations and then fold these relations into the dynamics of the slower portion of the process. The resulting low order process model has comparable accuracy, and will form the basis of a computationally tractable nonlinear model predictive controller for PEMFC applications.