Automation of Analysis Model Creation

The present text proposes the development of a computational infrastructure to enable rapid creation of high-quality conforming finite-element computational models---compatible with all existing finite-element simulation tools. The specific proofs-of-concept to be pursued in the Phase I will revolve around generation of tetrahedral meshes, but extensions to other types of finite-element types (e.g. hexahedral meshes) will be developed fully as part of the Phase II work. The proposed infrastructure will be built upon a novel surface-representation methodology based on Fourier analysis and a certain continuation method for the resolution of the Gibbs phenomenon. The resulting software will be able to produce both 1) High-quality, large-patch parametrizations of general surfaces, and, from these, 2) Conforming surface and volume finite element representations for complex three-dimensional structures. We expect that, in view of the ability of the continuation method to parametrize large portions of a given surface starting a CAD representation - while, at the same time, repairing, abstracting and completing surface features, the resulting infrastructure will allow for creation of finite-element models of high quality in human-operator times of the order of a small fraction of those required by procedures in use at present.BENEFITS: We expect that, in view of the ability of the proposed software to develop three dimensional meshes starting from a point cloud arising, e.g, from a CAD representations while, at the same time, repairing, abstracting and completing surface features, the resulting infrastructure will allow for creation of finite-element models of high quality, addressing issues concerning element quality, surface repair, and coarsening/refinement around boundaries and geometric singularities, in human-operator times of the order of a small fraction of those required by procedures in use at present. There is a well recognized need for a computational capability of the type proposed in this text within the multi-billion dollar Computer Aided Design and graphics world. One of the goals of the proposed work is to produce software that interfaces transparently with the substantial CAD infrastructure already in place within the Navy, DoD and commercial concerns, to facilitate design and virtual testing of products, vehicles, weapon systems, etc.