| Author | Hugues L. de Cougny |
|---|---|
| Title | Parallel Unstructured Distributed Three-Dimensional Mesh Generation |
| Year | 1998 |
| School | Civil Engineering |
| Institution | RPI |
| Abstract | Parallel unstructured three-dimensional mesh generation is a challenging problem for many reasons, the most obvious coming from the complexity of "partitioning" geometric models such that meshing is load balanced at all times. The approach taken here is to separate surface and volume meshing and present two different methodologies to perform these tasks in parallel. For surface meshing, which is performed on a model face by model face basis in the sequential case, partitioning considers the distribution of individual model faces to processors with possible subdivision if adequate load balance cannot be achieved. The technique used in sequential surface meshing is based upon vertex insertion using the Delaunay criterion in parameter space. In volume meshing, an octree built considering the surface mesh and meshing size attributes is partitioned in parallel using the Recursive Inertial Bisection methodology and "portions of space" are handed out to processors for meshing. Volume meshing operators on two techniques: (i) octant template meshing for interior octants and (ii) face removals (advancing front) to fill the space in between the surface mesh and the "templated" octants. Due to the mesh being distributed, domains corresponding to the interfaces of the initial partitioning are left unmeshed. In order to complete volume meshing, the repartitioning of these domains or interfaces is necessary. This is referred to as hierarchical repartitioning since face, edge, and vertex interfaces are repartitioned in order (followed each time by a face removal phase). The obtained mesh is then repartitioned using the Recursive Inertial Bisection methodology, prior to be optimized with respect to element quality. Parallel mesh generation involves many different processes. Focus has been given to the detailed presentation of scalable procedures to support each described process. |