Supersonic turbulent boundary layer: DNS and RANS

We assess the performance of a few turbulence models for Reynolds averaged Navier-Stokes (RANS) simulation of supersonic boundary layers, compared to the direct numerical simulations (DNS) of supersonic flat-plate turbulent boundary layers, carried out by Gao et al. [Chin. Phys. Lett. 22 (2005) 1709] and Huang et al. [Sci. Chin. 48 (2005) 614], as well as some available experimental data. The assessment is made for two test cases, with incoming Mach numbers and Reynolds numbers M = 2.25, Re = 365,000/in, and M = 4.5, Re = 1.7 × 10⁷/m, respectively. It is found that in the first case the prediction of RANS models agrees well with the DNS and the experimental data, while for the second case the agreement of the DNS models with experiment is less satisfactory. The compressibility effect on the RANS models is discussed.