Nonlinear Analysis, Control, and Modeling of the Two-Phase Flow Dynamics in Polymer Electrolyte Fuel Cells

• Main Author: Burkholder, Michael B.
• Format: Others
• Published: Research Showcase @ CMU 2015
• Online Access: http://repository.cmu.edu/dissertations/618; http://repository.cmu.edu/cgi/viewcontent.cgi?article=1656&context=dissertations

Polymer electrolyte fuel cells (PEFCs) generate clean, renewable power from hydrogen and oxygen with a byproduct of water. When the liquid water produced at the cathode is improperly removed, PEFCs experience unstable operation and power loss. Current techniques for cathode water removal result in excessive parasitic loads on the PEFC that contribute to high system costs. This thesis explores ways to improve the efficiency of PEFC cathode water management by developing a fundamental understanding of the dynamical stability of two-phase air and water ow in parallel cathode microchannels through experiments, analysis, and modeling. The dynamics of an experimental PEFC under varying degrees of cathode water removal were characterized by nonlinear invariants indicative of dynamical complexity and stability, the correlation dimension and Kolmogorov entropy. It was found that low-current operating conditionsm suffer from chaotic instability and performance loss due to cathode flooding. In order to detect and control dynamical instability in real time, a computationally-efficient reduced order Lyapunov exponent was formulated to indicate stability related to cathode water content. A stabilizing control algorithm was developed using feedback from real-time computation of the reduced order Lyapunov exponent of the PEFC voltage to trigger lowcost cathode pressure pulses. The control was demonstrated to stabilize flooding conditions with minimal parasitic expense for water removal. Two-phase air-water ow structures were visualized and pressure drops were measured in ex-situ microchannels under varying levels of transience. The pressure drops and their fluctuations were characterized with average values and fractal statistics, respectively, across air and water ow rates and ow regimes of relevance in PEFC operation. Dynamic pressure drop hysteresis was observed and measured, most likely for the first time. The statistical experimental results were used to develop a dynamical model of a PEFC cathode flow field with two-phase ow in parallel microchannels. The model included experimental values for two-phase pressure drops, a 1D + 1 PEFC model for water generation, and fractional Brownian motion for two-phase pressure uctuations. The model was used to understand ow maldistribution and Ledinegg instability in PEFC cathod flow fields, and to highlight methods for optimizing PEFC water management.