Large-Eddy Simulation of Flow over a Wall-Mounted Hump with Separation and Reattachment

Abstract

Wall-resolved large-eddy simulation of a model turbulent flow involving favorable and adverse pressure gradients, imposed by surface curvature of a wall-mounted hump, is performed for a spanwise-periodic computational domain. The flow acceleration over the front portion of the hump is strong enough to exceed the relaminarization criterion of Narasimha and Sreenivasan (“Relaminarization in Highly Accelerated Turbulent Boundary Layers,” Journal of Fluid Mechanics, Vol. 61, No. 3, 1973, pp. 417–447) but only over a relatively short streamwise extent. Although the flow does not relaminarize, turbulent skin-friction variation exhibits a plateau, also observed in the experiment of Greenblatt et al. (“A Separation Control CFD Validation Test Case, Part 1: Baseline and Steady Suction,” AIAA Journal, Vol. 44, No. 12, 2006, pp. 2820–2830), before it again continues to rise. The subsequent adverse pressure gradient is strong enough to cause flow separation. The location and extent of flow separation, including skin-friction distributions, compare reasonably well with experimental data. Computed velocity and Reynolds stress profiles are also compared with the experimental results. A systematic study of the effect of computational domain span, subgrid-scale model, tunnel backpressure, incoming boundary layer, grid refinement, Mach number, and top tunnel wall contour (which models the blockage effect of the experimental setup) is carried out, and sensitivity of the results to these parameters is discussed.