Water Management in PEM Fuel Cells

Michigan State University, East Lansing MI

Investigators

Abstract

The goal of this project is the first mathematical analysis of the structure and scaling of the multiphase flow in the gas di_usion layer of PEM fuel cells. The weak flux regime identified from the scaling suggests a reduction of the coupled degenerate parabolic system of conservation laws for heat and mass transport with phase change to a sharp interface model, with a linear problem on either side of a slowly moving free interface. A rigorous analysis of this problem, based upon the renormalization group, will yield analytical expressions for vapor and liquid fluxes in terms of prescribed boundary conditions and material parameters. Based upon this reduction, a suite of fast, robust numerical schemes in 1+2 and 1+3D will be developed to investigate control strategies for water management. We estimate a speed up in computational time of 3-4 orders of magnitude over resolution of the full problem. A key application is the development of drying purge cycles, these minimize damage from ice formation after shut-down by dehydrating the gas di_usion layer and flow fields, while maintaining humidification levels in the membrane. Numerical simulation of drying purges requires fast and accurate capturing of the multiphase fronts, a feature unavailable in existing codes. In addition, a simplified model of the coupled parabolic-elliptic problem for water management in the ionomer membrane will be developed and analyzed. Recent experimental work has shown multiple steady states, hysteresis, and long-time oscillations in membrane hydration levels. At low inlet humidity levels associated with autohumidification the membrane hydration couples strongly to the electric potential through the membrane conductivity, leading to "dry-cell" bifurcations and slow moving fronts. Autohumidification criteria which avoid the dry-cell bifurcation will be developed, the inverse problem detecting the dry-cell bifurcation from external voltage measurements will be investigated. A model of the cathode catalyst layer including oxygen and hydronium transport, the overpotential, the electric potential, and build-up of liquid water layers will be developed. The connection between averaged reaction rates and catalyst layer micro-structure will be rigorously established through homogenization theory by solving key cell-problems which elucidate the role of triple junction points and catalyst layer geometry on over-all performance. A connection between micro-structure geometry and the experimental technique of AC impedance spectroscopy will be pursued. Proton electrolyte membrane (PEM) fuel cells are unique energy conversion devices with a broad range of applications. Powered by hydrogen gas or directly from methanol, they hold promise to revolutionize the automotive industry while at the same time yielding e_cient power generation for cell phones, laptop computers, camcorders, digital cameras, PDAs, and other applications in the 10-100 Watt range. However, optimization is required to achieve targets of cost reduction, unit power density, operational lifetime, and to improve performance for autohumidification regimes. Water management is key not only to e_cient cell operation but also to prolonging cell operational lifetime. The development of the proposed models and high-performance computational tools for water management in PEM fuel cells represent a significant breakthrough in the computational resolution of the delicate free-surface motion required to understand water management. These novel computational tools will improve resolution while reducing computation time from from weeks to minutes. This dramatic enhancement of computational performance will have substantial impact on the development of new materials and manufacturing processes for the fuel cell industry, including the PI's industrial partners, Ballard Power Systems andW.L. Gore. Moreover, the development of lightweight, durable, low thermal and acoustic signature power sources has been identified by the Intelligence Community as an important ingredient of homeland security.