Hybrid Reynolds-Averaged Navier–Stokes/Large-Eddy Simulation Models for Flow Around an Iced Wing

Abstract

This study evaluates the feasibility of applying a newly developed dynamic hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation modeling framework to predict the massively separated flow around a GLC-305 airfoil with a 22.5 min leading-edge glaze ice accretion. Three-dimensional numerical simulations were performed at , , and . Comparisons were made between experimental data and simulation results computed using two Reynolds-averaged Navier–Stokes models (Menter’s shear stress transport and Spalart–Allmaras) and two hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation models (delayed detached-eddy simulation and the dynamic hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation model). All models overpredicted the mean wall static pressures on the suction surface. Wall pressure predictions obtained using the Reynolds-averaged Navier–Stokes and dynamic hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation models exhibited qualitatively better agreement with experiments than did delayed detached-eddy simulation predictions, and the dynamic hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation model produced the best agreement for the mean streamwise velocity profiles. Turbulent intensity profiles showed substantial mesh sensitivity for the delayed detached-eddy simulation model simulations, whereas the dynamic hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation model results qualitatively compared well with the experiments and exhibited only a small degree of mesh sensitivity. Unique to the dynamic hybrid Reynolds-averaged Navier–Stokes/large-eddy simulation simulations was the accurate prediction of the flow reattachment location. None of the Reynolds-averaged Navier–Stokes models predicted flow reattachment, and the delayed detached-eddy simulation model predicted a substantially delayed flow reattachment.