SUBJECT: Ph.D. Dissertation Defense
BY: Kevin Davies
TIME: Thursday, August 1, 2013, 2:30 p.m.
PLACE: MARC Building, 114A
TITLE: Declarative Modeling of Modeling Proton Exchange Membrane Fuel Cells for System Design
COMMITTEE: Dr. Comas Haynes, Co-Chair (ME/GTRI)
Dr. Chris Paredis, Co-Chair (ME)
Dr. Tequila Harris (ME)
Dr. Sheldon Jeter (ME)
Dr. Tom Fuller (ChBE)
Dr. Robert Moore (Hawaii Natural Energy Institute)


In certain power applications, proton exchange membrane fuel cell (PEMFC) systems are preferable because they can convert fuel energy to work more efficiently than internal combustion engines and have energy-to-power ratios that can be easily adapted, unlike batteries. However, the cost and durability, and to a lesser extent, size and weight, of PEMFCs are generally not yet adequate for use beyond niche devices and special demonstrations. Much work is being done to investigate the modes of failure and degradation, develop new materials and structures, improve manufacturing processes, and design better systems. Mathematical models of PEMFCs are being used to help understand the relevant physical phenomena, study the effects of design choices, and perform model-based control. The breadth of these purposes has led to a multitude of specialized models. The vision for this research is to utilize recent advancements in equation-based, object-oriented modeling tools in order to design a reconfigurable PEMFC model that is effective for a variety of uses. It has first been necessary to determine a way to represent certain phenomena---electrical, chemical, fluid, and thermal---in a manner that respects the underlying physics, is well-organized for systematic implementation, and allows mathematical manipulation for efficient simulation. The method delineates the advective and diffusive contributions to transport and exchange processes while applying a novel form of upstream discretization. The equations have been implemented in the Modelica language as an open-source library for modeling fuel cells. The presentation will describe the design of the models and review the results under various operating conditions and model configurations.