Mesh Generation Using a Correspondence Distance Field
A Thesis Presented for the Master of Science in Computational Engineering Degree, The University of Tennessee at Chattanooga
Nicholas Szapiro, August 2012
The central tool of this work is a correspondence distance field to discrete surface points embedded within a quadtree data structure. The theory, development, and implementation of the distance field tool are described, and two main applications to two-dimensional mesh generation are presented with extension to three-dimensional capabilities in mind.
First is a method for surface-oriented mesh generation from a sufficiently dense set of discrete surface points without connectivity information. Contour levels of distance from the body are specified and correspondences oriented normally to the contours are created. Regions of merging fronts inside and between objects are detected in the correspondence distance field and incorporated automatically. Second, the boundaries in a Voronoi diagram between specified coordinates are detected adaptively and used to make a Delaunay tessellation. Tessellation of regions with holes is performed using ghost nodes.
Images of meshes for each method are given for a sample set of test cases. Possible extensions, future work, and CFD applications are also discussed.
Click here to access a full copy of Nick's thesis.