Automated Preparation of Geometry Models for Computational Applications

In this proposed effort we seek to develop and integrate CAD repair algorithms with our existing highly capable geometry manipulation and grid generation software with the goal of greatly decreasing the time spent generating grids for complex geometry configurations as part of the numerical simulation problem setup. Highly detailed CAD models are critical for design and manufacture of a wide variety of geometries, from antennas to vehicles, bridges and buildings. However, in many cases, small features in the CAD model are not needed for the numerical simulations carried out using those models, such as those involving structural mechanics, fluid dynamics, or electromagnetics. De-featuring, or simplifying CAD models and performing associated clean-up of the geometry entities remains a large bottleneck in the simulation process. In addition, deficiencies exist in exchanging data among CAD tools. Computational engineers may also spend a large amount of time setting up and performing surface and volume grid generation to obtain grids that form the input to computational solvers. This proposal outlines methods for greatly reducing the entire CAD to mesh process using new and existing technology developed at HyPerComp, or open source solutions, if available.