Search

This talk is about an ongoing programme of research on detailed performance calculations for two-phase liquid-gas flow in the gas channel and porous transport layer, as found, for example, on the cathode side of a polymer electrolyte fuel cell. The porous geometry is typically obtained by digital reconstruction from nano-computer tomography images. The domain may then tessellated with a computational mesh, whereupon the equations of mass and momentum are solved ,e.g., by means of a volume-of-fluid method. Liquid water is produced at the same time as gaseous oxygen is consumed by electrochemical reduction at the electrode surface, which is to be considered a boundary condition in the present problem. The problem was originally formulated in ref. [1]. Liquid-gas counter flow in the porous transport layer results in liquid drops being entrained in co-flow in the gas channels and convected downstream by the gas. The flow in the channels and adjacent parts of the porous transport layer is transient-periodic, but with some significant randomness, associated with the interactions between the different fluid streams percolating into the channel from the pores. In this presentation, the complex micro-scale flow field is described in some detail. Consideration is also given as to the mechanisms for construction of macro-homogeneous properties such as absolute and relative permeability, capillary pressure vs. saturation for porous media, as well as macroscopic drag laws for two-phase channel flows based on calculations on a micro-scale. These are required for macro-scale homogeneous models, as typically employed at a cell-level. A discussion is also given of the impact of salient physical properties such as surface tension, and how these may be manipulated to improve future performance for electrochemical conversion devices such as fuel cells and electrolysers.

Polymer electrolyte electrochemical cells (PECs) offer exciting possibilities as energy converters in the form of both fuel cells and electrolyzers. Computational fluid dynamics can be used to analyse the performance of virtual prototypes at the micro-scale, as well as the cell and stack levels. This presentation details a comprehensive program of development of open source tools to describe and predict PEC operation. The software employed is OpenFOAM, which mimics the governing partial differential equations that describe the physics of PEC behaviour. Models range from detailed two-phase flow in porous transport layers based on digital reconstructions and the volume of fluid method, through single and multi-fluid cell models employing Eulerian-Eulerian solvers, to stack-level models whereby volume averaging also known as local homogenization is employed for both single-phase and two-phase formulations. The development of a multi-phase stack model represents an original and novel scheme not previously published. The results of mathematical calculations are compared with analytical solutions for one-dimensional idealizations, previous three-dimensional (3-D) numerical work by other researchers, and detailed experimental data gathered in-house. Comparisons are made in terms of local current density and species mass/molar fractions, as well as current density vs. voltage, i.e., polarization characteristics. Different levels of model complexity are also considered, for instance comparisons of a 2-D electrochemical formulation based on a Kirchhoff-Ohm equation (ideal potential less losses/over-potentials) vs. a fully 3-D simulation of both ionic and protonic potential fields. The presentation concludes with an assessment of the needs/requirements of the next generation of PEC analysis tools.