This paper provides a discussion of several aspects of the construction of approaches that combine statistical (Reynolds-averaged Navier-Stokes, RANS) models with large eddy simulation (LES), with the objective of making LES an economically viable method for predicting complex, high Reynolds number turbulent flows. The first part provides a review of alternative approaches, highlighting their rationale and major elements. Next, two particular methods are introduced in greater detail: one based on coupling near-wall RANS models to the outer LES domain on a single contiguous mesh, and the other involving the application of the RANS and LES procedures on separate zones, the former confined to a thin near-wall layer. Examples for their performance are included for channel flow and, in the case of the zonal strategy, for three separated flows. Finally, a discussion of prospects is given, as viewed from the writer's perspective.
Large Eddy Simulation is now almost routinely used in Fluid Mechanics research to investigate fundamental aspects of turbulence mechanics, to help validate statistical closures and to obtain predictions for flows in which unsteady events associated with turbulence are of major interest or influence. Although LES continues to be an expensive approach at practically relevant Reynolds numbers, despite rapid advances in computer power, the expense is tolerable when the flow being simulated is remote from walls. However, flows which are substantially affected by near-wall shear and turbulence, pose serious resource challenges as a result of the need to increase the near-wall grid resolution in line with N = O Re 2 , to restrict the distance between the wall and the nodes closest to the wall to around y + = 2 and to maintain a cell-aspect ratio of order ∆y + /∆s + = O (2, 20), where s is any direction parallel to the wall. Thus, at high Reynolds numbers, the utility of LES in a practical context depends greatly on the availability of acceptably accurate near-wall approximations that allow the resolution requirements to be reduced to economically tenable levels.Over the past few years, a whole range of approaches to this problem have been proposed. These include log-law-based wall-functions, various zonal and seamless RANS-LES hybrid schemes, the DES method, and the two-layer approach. So far, no one particular method has been demonstrated to be definitely superior to others, and all involve restrictions and limitations which adversely affect the resulting solution in some circumstances. Even in a simple fully-developed channel flow at high Reynolds number, no method is able to give a solution that is without defect in the vicinity of the edge of the near-wall layer within which the approximate model is applied.In earlier work by Temmerman et al. (2004), a RANS-LES hybrid method has been investigated in which a conventional low-Re model is applied within a near-wall layer the thickness of which can be chosen freely, Fig. 1(a). Coupling to the LES domain is effected via compatibility constraints, including a dynamic process which adjusts the turbulent viscosity at the RANS side of the interface by reference to the Subgrid-scale viscosity in the LES layer.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.