Application of hybrid RANS/LES methods for the simulation of shock-induced turbulent boundary layer separation

The results of shock-induced turbulent boundary layer separation simulations using scale-resolving methods with wall modelling (WMLES) are presented. A test case is chosen for which experimental data are available as well as reference large eddy simulation results. The IDDES method based on the SST turbulence model in the WMLES mode is compared with a similar method in which the explicit subgrid scale model is turned off. Within the framework of the second approach, two schemes of the WENO family are considered. In each case, two grids of different density are employed. Separation size, pressure and friction coefficient distributions along the wall, and one-point pressure probability density functions in two reference points are compared. An unexpected result is demonstrated: in all the scale-resolving simulations, the separation region length turned out to be underestimated by 15–28%, and no significant reduction in the error with mesh refinement was observed. Possible reasons for this phenomenon are discussed.