Investigation of Flow Dependent Blending Functions in Hybrid LES/RANS Simulations

Abstract

A blending function that depends on the ratio of the von Karman length scale and the Taylor microscale in conjunction with k‐ turbulence model is used to study two Mach 3 shock wave/boundary layer interactions on 20 degree compression ramp and 25 degree compression-expansion ramp. The results are compared with two other blending functions: one in which the von Karman length scale is replaced by the distance to the nearest wall while the other employs a grid dependent expression proportional to the ratio of the RANS eddy viscosity to that of the LES. The results were further compared with a calculation where the underlying RANS model is that of Menter’s shear stress (SST) model with a blending function that depends on the distance to the wall. All inflow boundaries employ the rescaling-reintroducing procedure developed by Xiao et al. for hybrid LES/RANS approaches. In general, the blending function based on the von Karman length scale gives the best results when compared with available measurements. The skin friction measurements showed the highest sensitivity to the various blending functions and the two RANS models employed.