A Comparison of CREATE™ Kestrel and CFD++ for Resolving Hypersonic Flow

The Army Corp’s Engineering Research and Development Center has partnered with Lockheed Martin Aeronautics Company and the Air Force Research Lab Multidisciplinary Science & Technology Center (AFRL-MSTC) as part of the EXPanded MDO for Effectiveness-based DesIgn TEchnologies (EXPEDITE) program to simulate hypersonic flight in a computational environment that supports multi-physics coupling. More specifically, a hypersonic multi-physics workflow was sought that coupled high fidelity fluid, thermal, and structural analyses with an emphasis on using Department of Defense High Performance Computing Resources and government owned software tools. The goal of this paper is to detail the comparison of the CREATE™-AV Kestrel fluid solver, kCFD, against the baseline fluid solver CFD++ by Metacomp Technologies. CFD++ was chosen as the baseline aerodynamic tool due to the EXPEDITE team’s familiarity and experience with the software. This comparison was necessary for the purpose of ultimately incorporating Kestrel in the overall multi-physics workflow. A Lockheed Martin test vehicle similar to a sphere-cone geometry with aerodynamic control surfaces was used for the comparison. Identical CAD models and meshes were used in both solvers to remove the effects of geometric modeling and/or meshing from the comparisons. The comparison therefore focused on identifying physical modeling and/or numerical algorithm differences that could potentially affect the nature of the final solution in the hypersonic flight regime. Several aerodynamic and aerothermodynamic quantities were compared from a notional hypersonic trajectory at various vehicle and control surface configurations. These quantities included integrated lift, drag, and moment coefficients, heat transfer, and node-by-node boundary data comparisons for pressure and temperature. The results show that, despite the numerous differences in numerical algorithm implementations, integrated values for forces and moments, as well as boundary data for pressure and temperature, remain closely aligned across a wide range of Mach number, vehicle AoA, and control surface deflection settings. However, the results show a much wider dispersion for heat flux values between the two solvers. Further analyses show that the aerothermodynamic quantities are quite sensitive to modeling coefficients and implementation details intrinsic to the various turbulence models available in the solvers and great care is necessary when heat transfer effects are sought. Finally, it was concluded that the Kestrel software was a suitable tool for modeling hypersonic flow of interest to the EXPEDITE program.