The work will focus on the 3D implementation of the Phase 1 CHARM mesher, with solution-adaptive iteration for CFD and non-CFD applications. The proposed 3D method will incorporate and extend a previously developed method of generating field-guided hexahedral elements from a metric tensor field. While the fundamental technical approach a combination of metric tensor conditioning, metric-tracing mesher, and cell-packing mesher remains the same, there are many technical challenges specific to the 3D domain, including the following: - Investigation into conditioning of volume metric tensor fields - Investigation into the topology (structure) of volume metric tensor fields - Developing algorithms for the generation, repair, and adjustment of streamsurface arrangements - Developing algorithms to convert streamsurface arrangements to hex-dominant meshes - Developing algorithms to combine streamsurface- and packing-based meshes - Investigation into designing these algorithms for mesh adaptation rather than adaptive remeshing - Investigation of time and storage efficiency of these algorithms in a large-scale parallelism context In addition to the above, the goal is to generalize the solution in order to support its packaging and commercialization for a number of problem sets and target applications. This includes generalization of the solver-adaptive framework, creation of APIs to programatically expose core functions, and provide UI access to appropriately control and configure the application.
More »