Difference between revisions of "Sig Turbulence / Flow over Periodic Hills"

Revision as of 09:50, 29 May 2009

Olivier Brugiere, Universite Joseph Fourier, Grenoble, France

1 Motivation

The flow over bodies with massive separation constitutes an important area of applications for LES. In many geometrie we can find this kind of flow like the 2D backward-facing step, the asymmetric diffuser or the periodic hill. We know experimental's data which was made by Almeida et al. [1] but in Mellen and al. [2] et Temmerman et al. [3] we can find that they have an influence of the side wall so the 1995 ERCOFTAC/IAHR workshop have cast a doubt on the true periodicity of the experimental configuration. But we have some data from LES with a good mesh like [[1]]. Flow over 2D periodic is in fact an experimental configuration with 9 hills. For the computational, we represent this configuration by a channel periodic with two half's hills like on the figure 1.

The probleme of this flow is the separtion bubble which is due to a adverse pressure gradient that's why this is a test case for many subgrid scale models (SGS) and for near-wall low.

2 Testcase description

2.1 Flow configuration

2.1.1 Geometrical Parameters

Hill's heigh : $h = 28 mm$
Streamwise distance : $L_{x} = 9,0 h$
Normal wall heigh : $L_{y} = 3,035 h$
Spanwise distance : $L_{y} = 4,5 h$

2.1.2 Boundary condition

Streamwise condition : periodicity
Spanwise condition : periodicity
Normal to streamwise : two walls

2.2 Simulation details

Figure 2 : Hill mesh

We have make the geometry on Gambit Fig 2 and the mesh is composed by $Nx \times Ny \times Nz = 118 \times 33 \times96$ .

The aim of my study is testing a posteriori near-wall low. To compare we are running three cases :

With near-wall low
With the Spalding Low [4]
Wiht the Mahart et al. low [5]

3 Numerical results

4 References

[1] Almeida, Durao and Heitor, 1993, Experimental thermal and fluid science, Vol. 7, pp. 87-101.

[2] Mellen, Frohlich, Rodi, 2000, Large Eddy Simulation of the flow over periodic hills, 16th IMACS World Congress , Lausanne.

[3] Temmerman, Leschziner, 2001, Large Eddy Simulation of separated flow in a streamwise periodic channel construction, Int. Symp. on Turbulence and Shear Flow Phenomena, Stockholm, June 27-29.

[4] Spalding, 1961, A single formula for the law of the wall, Jl. Appl. Mech., vol 28, pp. 455-457

[5] Manhart Peller and Brun, 2008, Near-wall scaling for turbulent boundary layers with adverse pressure gradient, Theor. Comput. Fluid Dyn., vol 22 , pp. 243-260.

Back to Sig Turbulence

@@ Line 3: / Line 3: @@
 == Motivation ==
-The flow over bodies with massive separation constitutes an important area of applications for LES. In many geometrie we can find this kind of flow like the 2D backward-facing step, the asymmetric diffuser or the periodic hill.
+The flow over bodies with massive separation constitutes an important area of applications for LES. In many geometrie we can find this kind of flow like the 2D backward-facing step, the asymmetric diffuser or the periodic hill. We know experimental's data which was made by Almeida et al. [1] but in Mellen and al. [2] et  Temmerman et al. [3] we can find that they have an influence of the side wall so the 1995 ERCOFTAC/IAHR workshop have cast a doubt on the true periodicity of the experimental configuration. But we have some data from LES with a good mesh like [[http://cfd.mace.manchester.ac.uk/cgi-bin/cfddb/prpage.cgi?81&LES&database/cases/case81/Case_data&database/cases/case81&cas81_head.html&cas81_desc.html&cas81_meth.html&cas81_data.html&cas81_refs.html&cas81_rsol.html&1&1&1&1&1&unknown]]. Flow over 2D periodic is in fact an experimental configuration with 9 hills. For the computational, we represent this configuration by a channel periodic with two half's hills like on the figure 1.
-* [[Sig Turbulence / 2D backward-facing step |2D backward-facing step]] : it's a quick test case because of it's a 2D model but the separation bubble is determined by the spet geometry.
-* [[Sig Turbulence / Asymmetric diffuser |Asymmetric diffuser]] : we know experimental's data which was made by Buice and Eaton [[http://www.grc.nasa.gov/WWW/wind/valid/buice/buice02/buice02.html]]. Unfortunatly, the Reynolds number of the fow was necessarily slow (<math> Re_{\tau} = 180 </math>).
-[[Image:hilloutlined.jpg|thumb|Figure 1 : periodic hill]]
-* [[Sig Turbulence / Flow over Periodic Hills |Periodic Hills]] : we know experimental's data which was made by Almeida et al. [1] but in Mellen and al. [2] et  Temmerman et al. [3] we can find that they have an influence of the side wall so the 1995 ERCOFTAC/IAHR workshop have cast a doubt on the true periodicity of the experimental configuration. But we have some data from LES with a good mesh like [[http://cfd.mace.manchester.ac.uk/cgi-bin/cfddb/prpage.cgi?81&LES&database/cases/case81/Case_data&database/cases/case81&cas81_head.html&cas81_desc.html&cas81_meth.html&cas81_data.html&cas81_refs.html&cas81_rsol.html&1&1&1&1&1&unknown]]. Flow over 2D periodic is in fact an experimental configuration with 9 hills. For the computational, we represent this configuration by a channel periodic with two half's hills like on the figure 1.