Wall-modeled large eddy simulation in complex geometries with application to high-lift devices

In large eddy simulations (LES) of wall-bounded flows, the need to resolve energycarrying eddies in the vicinity of the wall leads to a prohibitive cost which prevents its use at very high Reynolds numbers. To overcome this limitation, many attempts have been made in the last two decades to model the unresolved scales near the wall. These works mainly focused on two different approaches: hybrid RANS/LES computations and direct modeling of the wall-stress. In hybrid RANS/LES, the unsteady Navier-Stokes equations are solved on a single grid with the eddy viscosity computed from a RANS model near the wall and a LES model away from the wall. In the wall-stress modeling approach, the LES equations are formally defined everywhere in the domain, with RANS equations solved on an embedded (overlapping) grid near the wall. The coupling between RANS and LES is weaker and acts in a fashion similar to the widely used wall-functions in RANS: the RANS solver takes information from the computed LES flow field, and returns back the result in the form of wall fluxes, i.e., the shear stress and heat transfer at the wall. The present contribution falls into the second category and is aimed at evaluating the benefits of this approach in situations involving complex geometries and underlying complexity in the flow field, such as transition and/or separation. Wall-modeling in the context of LES dates back four decades, but has primarily been applied to problems with simple geometry, e.g., the turbulent flow in a channel. Balaras et al. (1996) extended the modeling to account for some non-equilibrium effects and applied this to duct flow. Wang & Moin (2002) added a dynamic procedure to the non-equilibrium model and applied it to the flow over a trailing edge. More recently, a non-equilibrium model has been developed and tested in a channel flow with periodic hills (Duprat et al. 2011). To the authors’ knowledge, no attempts have been made so far to use a wall-model approach in a configuration where the geometric complexity requires the use of an unstructured grid. Here we aim at evaluating the concept for external aerodynamics at high Reynolds number and identify the remaining challenges.