A prediction method for high speed turbulent separated flows with experimental verification

Abstract

It is well known that the level of current engineering computational aerodynamic analysis for turbulent hypersonic separated flows is not satisfactory. Computations using standard two-equation turbulence models do not give sufficiently accurate predictions. Existing compressibility corrections do not provide general improvements. In fact, they seem to render predictions worse, at least for Mach numbers up to five. Possible improvements using Explicit Algebraic Turbulence Models (EARSM) have been investigated. This class of models allow for a more general coupling between the mean field and the turbulent Reynolds stresses, compared to the standard two-equation models based on the Boussinesq hypothesis. A new class of EARSMs is proposed and compared to results of existing standard models. It is demonstrated that the new proposed EARSM gives improved prediction over standard two-equation models and previously published EARSMs for validation cases relevant for high speed separated flows. A new complete EARSM turbulence model based on the k-G) twoequation model is recommended as giving the overall best prediction of the validation test cases. As part of the project a new series of detailed validation experiments, with turbulent boundary layer-shock wave interaction on a flat plate at Mach 5, with different degree of flow separation, have been performed.