SBIR-STTR Award

This proposal offers to provide NASA with an automatic mesh generator for the simulation of aerodynamic flows using Reynolds-Averages Navier-Stokes (RANS) models. The tools will be capable of generating high-quality, highly-stretched (anisotropic) grids in boundary layer regions and transition smoothly to inviscid flow regions even in an adaptive context. The objective of the work is to offer a unified view for generating quality and robust RANS meshes coupled naturally with anisotropic mesh adaptation. Our innovation is to view the anisotropic mesh generation within the Riemannian metric framework which thus far has been used exclusively in anisotropic mesh adaptation. Using the metric-based framework allows much easier handling of the large mesh size ratios involved in the computation, whereas traditional methods use the Euclidean framework to compute distance and volume. This innovative view to generate these meshes makes the entire procedure more generic and much more robust. The emphasis is being put on deriving a completely automatic process to generate quality and robust anisotropic meshes. Our existing and proven software package will be modified to include these innovative methods. A NASA test case will be computed for validation of the methods. The software will be delivered in Phase II.

The innovation of our Phase II STTR program is to develop and provide to NASA automatic mesh generation software for the simulation of fluid flows using Reynolds-Averaged Navier-Stokes codes. As a result of the successful Phase I work, these new tools are now capable of generating high-quality, highly-stretched (anisotropic) meshes in boundary layer regions and transition smoothly to inviscid flow regions, even in an adaptive context. The significance is that our method has the ability to generate a boundary layer mesh while keeping intact the previous adaptation procedures from non viscous simulations. This leads to a natural coupling between boundary layer mesh generation and anisotropic mesh adaptation. All of the Phase I objectives were met and all tasks were completed successfully. The Phase II project will include improvements in surface remeshing, coding for optimal speed and increased robustness of the solvers, adding a mesh optimization module, providing a link to general CAD packages, include unsteady coupling where the boundary layer mesh refinement evolves in time, conduct further validation and verification on NASA models by running flow cases with our solver, documenting the project, and delivering the new meshing software to NASA.