Domain Decomposition By the Advancing-Partition Method for Parallel Unstructured Grid Generation

A computer program product for domain decomposition for generating unstructured grids, the computer program product comprising at least one computer-readable storage medium having computer-readable program code stored therein, the computer-readable program code comprising:

• computer-usable program code for (a) generating a surface mesh of a spatial domain for which a grid is to be generated;
• computer-usable program code for (b) determining a location of a partition plane dividing the domain into two sections;
• computer-usable program code for (c) identifying triangular faces on the surface mesh that intersect the partition plane;
• computer-usable program code for (d) generating a partition grid of tetrahedral cells using a marching process in which a front comprises only faces of new cells which intersect the partition plane, the partition grid being generated until no active faces remain on the front, the partition grid dividing the domain into two sub-domains;
• computer-usable program code for (e) collecting triangular faces on each side of the partition plane into two separate subsets;
• computer-usable program code for (f) renumbering each subset of triangular faces locally and creating a local/global mapping for each sub-domain;
• computer-usable program code for (g) generating a volume grid for each sub-domain;
• and
• computer-usable program code for (h) merging the partition grid and volume grids using the local-global mapping.

A system for domain decomposition for generating unstructured grids, the system comprising:

• a processor configured for (a) generating a surface mesh of a spatial domain for which a grid is to be generated;
• the processor further configured for (b) determining a location of a partition plane dividing the domain into two sections;
• the processor further configured for (c) identifying triangular faces on the surface mesh that intersect the partition plane;
• the processor further configured for (d) generating a partition grid of tetrahedral cells using a marching process in which a front comprises only faces of new cells which intersect the partition plane, the partition grid being generated until no active faces remain on the front, the partition grid dividing the domain into two sub-domains;
• the processor further configured for (e) collecting triangular faces on each side of the partition plane into two separate subsets;
• the processor further configured for (f) renumbering each subset of triangular faces locally and creating a local/global mapping for each sub-domain;
• the processor further configured for (g) generating a volume grid for each sub-domain;
• and
• the processor further configured for (h) merging the partition grid and volume grids using the local-global mapping.