PITA Fiscal Year 2009 Projects - Environmental Technologies

A Large-Scale Integration of Physical Sub Models for Energy Processes: Focus on Hydrogen Polymer Electrolyte Fuel Cells

Principal Investigators: Lorenz T. Biegler, Myung S. Jhon

All energy systems comprise phenomena at different time and length scales, and often can be described through a hierarchy of scale-specific models. Such physics-based models use specific numerical techniques to solve the governing equations, covering atomistic, molecular, mesoscopic, continuum, device, and plant levels. The objective of this work is to achieve fulfillment of stringent performance requirements to maintain leadership in energy science and global markets. For these, we need to develop holistic multi-scale models, which ensure that knowledge generated at one scale is transmitted to the other, producing synergistic knowledge-bases. While there is tremendous interest in coupling models of different scales, especially with nano-science and technology, the vision of the proposed work goes far beyond what is currently considered in multi-scale modeling, by focusing on novel future energy applications.

To integrate multi-scale models for system design, we select hydrogen energy and polymer electrolyte fuel cell (PEFC) devices as benchmark case studies; these highly sought applications serve as excellent cases of multi-scale hierarchical phenomena [1]. Our current modeling capabilities have been limited to developing continuum level models for individual PEFC devices connected to large-scale state-of-the-art optimization algorithms [2, 3]. Through these efforts we examined water management issues and resolved inverse parameter estimation problems to obtain phenomenological models for catalyst layer (CL) [4]. We will first perform molecular dynamics (MD) simulations on the polymer electrolyte membrane (PEM), and develop lattice Boltzmann methods (LBMs) at mesoscopic level, superior to continuum computational fluid dynamics (CFD) in the multi-scale integration of porous media flow in gas diffusion layers (GDLs). Later we will examine the atomistic to supra-molecular levels via quantum mechanics and ab-initio MD descriptions for CL and PEM, and efficient multi-scale approaches using reduced order models (ROMs) that link sub-models in a parallel computing framework. This opens new paradigms in modeling and design of novel energy systems.