Adaptive and Parallel Computational Techniques in Materials Science.
Abstract
We developed mathematical models for problems involving the fabrication of ceramic- matrix composites, the coating of ceramic fibers, and the oxidation of ceramic-matrix composites. To solve these models, we developed an adaptive and parallel computational framework for the finite element and finite volume solution of partial differential equations. The parallel adaptive framework handles mesh generation, refinement, and coarsening; dynamic load balancing; distributed database management; interprocessor communications; data migration; time integration; and algebraic solutions. The models and solutions revealed improved processing techniques for a reactive vapor infiltration application and potentially superior strategies for coating fibers. Oxidation studies are investigating the 'self healing' of cracks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 09, 1996
- Accession Number
- ADA312778
Entities
People
- J. E. Flaherty
Organizations
- Rensselaer Polytechnic Institute