Wall-Modeled Large-Eddy Simulation with Second-Order-Accurate Upwind Scheme

Abstract

In aircraft aerodynamic design, it is essential for certification by analysis to estimate aircraft aerodynamic characteristics with high fidelity across the full flight envelope using computational fluid dynamics. Improving accuracy of the simulation of unsteady flows such as transonic buffet is one of the problems. Wall-modeled large-eddy simulation (WMLES) is expected to improve the accuracy. In this study, three quantitative requirements for the matching points, where LES feeds the wall model with flow variables, are proposed to ensure the accuracy of the WMLES with a second-order-accurate upwind scheme that is often used for engineering codes. These requirements are related to 1) matching point height, 2) the number of cells below the matching point, and 3) grid resolution at the matching point. The proposed requirements are assessed on a flat-plate turbulent boundary layer and a shock/turbulent boundary-layer interaction flow. As a result, it is clarified that all of the requirements are necessary to ensure the accuracy in terms of mean velocity, Reynolds stresses, and skin friction.