Unstructured Grid Adaptation for Hypersonic Aerothermodynamics

The National Geospatial-Intelligence Agency (NGA) seeks innovative modeling and simulation concepts for estimating hypersonic flowfields and phenomenologies. In response to this need, Corvid Technologies (Corvid) proposes the investigation, development, and demonstration of a practical numerical framework for computational grid adaptation to improve modeling of hypersonic flowfields and thus greatly reduce the cost involved in computing accurate surface heating on hypersonic systems using Computational Fluid Dynamics (CFD). Ultimately, Corvid will develop a robust, flexible and production-ready grid adaptation capability that would significantly shorten the time required to predict accurate aerothermodynamic environments. To date, several adaptation strategies have been matured but no capability has been developed for use in a production CFD setting to resolve wall heat transfer on unstructured meshes. Furthermore, such a solution would provide a consistent process that could be leveraged across different CFD cases, CFD solvers, programs, and throughout industry. Since grid generation is typically one of the first steps required in all CFD simulations, this capability would be impactful for all types of CFD simulations including perfect gas cases, chemically reacting cases, multiphase flows, shock interaction problems, coupled fluid/thermal/structural simulations, turbulent flows, and others. \n\n Corvid’s approach will address one of the most time-consuming, but critical, components of hypersonic simulations: mesh generation. Corvid will leverage its in-house computational resources, software development capabilities, aerothermodynamic expertise, and grid-generation approaches to implement a robust, grid adaptation capability for mixed-element unstructured meshes. By simply providing an initial (unadapted) grid, flow conditions, vehicle attitude, and target error metrics, a final, converged CFD solution with accurately resolved wall heat transfer will be produced. During Phase I, Corvid will implement and demonstrate different mesh mechanic strategies on unstructured grids into a robust, comprehensive capability. The proposed capability will be able to adapt the mesh to bulk flow features (shocks, shear layers) and to boundary layer gradients in prismatic layers typical of unstructured grids. Corvid plans to implement this process around our in-house CFD solver (RavenCFD) to expedite the implementation, testing, and validation being proposed. This technology will be validated against aerothermal test data on select hypersonic cases in Phase I. Corvid estimates that an automated approach could improve overall time to solution by 50% with an accompanying increase in accuracy.