High-Pass Filtered Eddy-Viscosity Models for Large-Eddy Simulations of Compressible Wall-Bounded Flows

2005 ◽  
Vol 127 (4) ◽  
pp. 666-673 ◽  
Author(s):  
Steffen Stolz

In this contribution we consider large-eddy simulation (LES) using the high-pass filtered (HPF) Smagorinsky model of a spatially developing supersonic turbulent boundary layer at a Mach number of 2.5 and momentum-thickness Reynolds numbers at inflow of ∼4500. The HPF eddy-viscosity models employ high-pass filtered quantities instead of the full velocity field for the computation of the subgrid-scale (SGS) model terms. This approach has been proposed independently by Vreman (Vreman, A. W., 2003, Phys. Fluids, 15, pp. L61–L64) and Stolz et al. (Stolz, S., Schlatter, P., Meyer, D., and Kleiser, L., 2003, in Direct and Large Eddy Simulation V, Kluwer, Dordrecht, pp. 81–88). Different from classical eddy-viscosity models, such as the Smagorinsky model (Smagorinsky, J., 1963, Mon. Weath. Rev, 93, pp. 99–164) or the structure-function model (Métais, O. and Lesieur, M., 1992, J. Fluid Mech., 239, pp. 157–194) which are among the most often employed SGS models for LES, the HPF eddy-viscosity models do need neither van Driest wall damping functions for a correct prediction of the viscous sublayer of wall-bounded turbulent flows nor a dynamic determination of the coefficient. Furthermore, the HPF eddy-viscosity models are formulated locally and three-dimensionally in space. For compressible flows the model is supplemented by a HPF eddy-diffusivity ansatz for the SGS heat flux in the energy equation. Turbulent inflow conditions are generated by a rescaling and recycling technique in which the mean and fluctuating part of the turbulent boundary layer at some distance downstream of inflow is rescaled and reintroduced at the inflow position (Stolz, S. and Adams, N. A., 2003, Phys. Fluids, 15, pp. 2389–2412).

Author(s):  
Steffen Stolz

Eddy-viscosity models such as the Smagorinsky model [1] are the most often employed subgrid-scale (SGS) models for large-eddy simulations (LES). However, for a correct prediction of the viscous sublayer of wall-bounded turbulent flows van-Driest wall damping functions or a dynamic determination of the constant [2] have to be employed. Alternatively, high-pass filtered (HPF) quantities can be used instead of the full velocity field for the computation of the subgrid-scale model terms. This approach has been independently proposed by Vreman [3] and Stolz et al. [4]. In this contribution we consider LES of a spatially developing supersonic turbulent boundary layer at a Mach number of 2.5 and momentum-thickness Reynolds numbers at inflow of approximately 4500, using the HPF Smagorinsky model. The model is supplemented by a HPF eddy-diffusivity ansatz for the SGS heat flux in the energy equation. Turbulent inflow conditions are generated by a rescaling and recycling technique proposed by [5] where the mean and fluctuating part of the turbulent boundary layer at some distance downstream of inflow is rescaled and reintroduced at inflow.


PAMM ◽  
2008 ◽  
Vol 8 (1) ◽  
pp. 10099-10102
Author(s):  
Nikolaus Peller ◽  
Michael Manhart

2011 ◽  
Vol 133 (2) ◽  
Author(s):  
Hassan Raiesi ◽  
Ugo Piomelli ◽  
Andrew Pollard

The performance of some commonly used eddy-viscosity turbulence models has been evaluated using direct numerical simulation (DNS) and large-eddy simulation (LES) data. Two configurations have been tested, a two-dimensional boundary layer undergoing pressure-driven separation, and a square duct. The DNS and LES were used to assess the k−ε, ζ−f, k−ω, and Spalart–Allmaras models. For the two-dimensional separated boundary layer, anisotropic effects are not significant and the eddy-viscosity assumption works well. However, the near-wall treatment used in k−ε models was found to have a critical effect on the predictive accuracy of the model (and, in particular, of separation and reattachment points). None of the wall treatments tested resulted in accurate prediction of the flow field. Better results were obtained with models that do not require special treatment in the inner layer (ζ−f, k−ω, and Spalart–Allmaras models). For the square duct calculation, only a nonlinear constitutive relation was found to be able to capture the secondary flow, giving results in agreement with the data. Linear models had significant error.


Sign in / Sign up

Export Citation Format

Share Document