Large-Eddy Simulation of Separated Vortical Flows

Traditional numerical computations of turbulent engineering or geophysical flows are based on the Reynolds-averaged Navier-Stokes
(RANS) equations. They need to be closed by statistical turbulence models which often have difficulties handling complex flow situations.
More recently, the Large-Eddy Simulation (LES) approach has been developed in which the large scales of turbulence are resolved numerically in space and time while their interaction with the non-resolved subgrid scales is modeled. Computationally, LES are typically much more expensive than RANS but still considerably less expensive (often by two orders of magnitude) than fully resolved Direct Numerical Simulations (DNS) of the same flow, without an essential loss of accuracy. LES are therefore expected to play a major role for the future prediction and analysis of certain turbulent flows in which a representation of unsteady turbulent fluctuations is important. Examples include large-scale flow separation in aerodynamics, fluid-structure interaction, turbulent flow control, aeroacoustics or turbulent combustion. Multiphase flows are another important area of LES applications.

The lecture will present some LES results obtained using our recently developed approximate deconvolution (ADM and ADM-RT) subgrid models. A range of fundamental problems have been simulated successfully with flow features such as wall turbulence, laminar-turbulent transition, massive flow separation, swirl, compressibility and shocks, as well as aeroacoustics. The subgrid models have also been implemented into the semi-industrial code NSMB for compressible flows which can handle geometrically complex flow configurations. It was applied to investigate jets in crossflow and film cooling which is crucial for gas turbines.
Another problem investigated relates to the transport and settling of suspended particles in model estuaries. As details of the settling mechanisms are still not fully understood, there is a great demand for accurate simulations of such flows.