The objective of this research is to develop hierarchical models for composites molding manufacturing processes that will explicitly consider the heterogeneous nature of the constituent materials. Starting from detailed descriptions of composite media, permeability models will be developed, which will be used to fully define the effective permeability field at length scales typical of computer aided design. The approach is to use high-performance parallel super-computing to solve the equations of fluid flow in microstructures of realistic complexity. This involves solutions of very large systems of equations and will be achieved using novel algorithms for fast, parallel computation. Upscaling will be achieved using analytical and numerical Monte-Carlo stochastic techniques.
This research will enable the design of novel and defect-free composites molding manufacturing processes that, instead of being limited by the heterogeneous nature of the material, take full advantage of it to create unique products with tailored property profiles. Ability to accomplish this will provide the US industry with a currently unavailable and significant competitive advantage. Technology transfer will be enhanced by collaboration between the PIs and DOE National labs as well as by the co-funding of the project by NSF as DOE's FreedomCAR program. In addition, the parallel supercomputing aspects of this work fit well with the need to support cutting edge research on high-performance computing in order to maintain and expand the Nation's lead in this important area. Educational impact is assured by ongoing participation at National and Regional undergraduate research programs, including programs targeting minorities.
