Rapid Development of Advanced High-Speed Aerosciences Simulation Capability

The current state of the art for production computational fluid dynamic (CFD) simulations in both the Entry, Descent and Landing (EDL) communities and modeling of hypersonic vehicles is the solution of steady-state problems on fixed computation grids. However, the majority of relevant challenge problems are unsteady. Accurate simulations of such unsteady phenomena currently require computational grids that are overly refined, i.e., in large portions of the domain or even globally. An effective method to alleviate over-refinement, and associated cost, is the application of automatic mesh adaptation. Mesh adaptation applies refinement in the vicinity of critical flow features while coarsens at other areas and can significantly reduce the time-to-solution in terms of both overall computation time and hands-on gridding. Corvid, RPI, and Simmetrix will continue development of an in-memory, parallel, mesh adaptation framework for unsteady problems that will enable rapid turnaround of large sets of CFD runs in a significantly shorter time frame than current workflows.