Development of a RANS-Based Wall-Modeled LES Approach for Hypersonic Flows

Accurate prediction of turbulent flows is essential for the design of the next generation of hypersonic vehicles. There is a growing consensus within the aerospace community that in the next few years there will be a substantial increase in the use of hybrid RANS-LES and wall-modeled LES (WMLES) methods given that RANS models, when used on their own, are unable to accurately capture unsteady separated flows. While these methods have mainly been developed, and tested for incompressible or weakly compressible flows they lack maturity for hypersonic flow applications. A wellintegrated wall-modeling approach is an essential ingredient in order to attain computationally efficient solution approaches for highly complex high-speed WMLES flow simulations. In this research, we will be testing and further developing an advanced ODEbased wall model (including energy equations, pressure gradient and convective terms) for WMLES of hypersonic flows around complex geometries integrated inside body-fitted and Cartesian grid based Navier-Stokes solvers. The proposed method provides the RANS-based near-wall modeling for an adaptive WMLES approach, as well as for pure RANS solvers. These wall models will be tested in our body-fitted CFD solver, as well as they will be integrated in a Cartesian grid based solver with Adaptive Mesh Refinement (AMR), which provides a fully-automated mesh generation process independent of the complexity of the geometry. By employing a multi-resolution approach and hybrid programming paradigms this approach can meet the high-performance computing capabilities expected from modern simulation codes. Within the scope of the proposed research we will provide insight into the current state-of-the-art of wall models for turbulent hypersonic flows and thereupon provide significant improvements to these models. For validating these models, we will use available experimental and simulation data as well as our own Direct Numerical Simulations (DNS). Long term, this research has the potential of being truly transformative because the proposed method will significantly accelerate CFD work flow and the degree of automation provided will extend the ability to conduct reliable, highfidelity, turbulent flow simulations of hypersonic vehicles