CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   LES wall function (https://www.cfd-online.com/Forums/openfoam-solving/63122-les-wall-function.html)

ayoros March 30, 2009 05:07
LES wall function
 
Hi all, and good Monday morning for everyone (at least for those close to my meridian time) !

About the nuSgsWallFunction, I still have some doubts.

Do the yPlus values have to be higher than say 11 for the wall function to be effective ? Or can LES wall function cope with yPlus values from say 1 to 30 ?

Also, if I don't use wall function, am I right applying 'zeroGradient' for nuSgs boundary condition at walls ?

Thanks in advance,
Fabien

cedric_duprat March 31, 2009 04:36
Hi Fabien,

In OpenFOAM, the nuSgsWallFunction correspond the Spalding's law.
It's a fit of the laminar and logarithmic regions for an equilibrium flow. So, if your boundary layer is in equilibrium, you can use the wall model for a large value of y+.
But because it is never the case, I would advise to resolve as accurately as possible your boundary layer by putting your first of the wall grid point in the viscous sub-layer (Y+<7)

the point is that between 7 and 30, it's the buffer layer and .... people don't really know what happend there that's why people usually avoid these values.

I hope this will help you

Cedric

PS: more information can be found in the turbulenceSIG wiki http://openfoamwiki.net/index.php/Si...Implementation

ayoros March 31, 2009 07:55
Hi Cedric,

Thanks for the information.

I looked after the Spalding's law and now I get it.

Regards,
Fabien

maysmech January 25, 2011 15:05
Quote:
the point is that between 7 and 30, it's the buffer layer
Hi,
Which B.C should be used for nu and K for yPlus 20?
Thanks in advance

openfoammaofnepo April 4, 2014 18:41
Quote:
Originally Posted by cedric_duprat (Post 211368)
Hi Fabien,

In OpenFOAM, the nuSgsWallFunction correspond the Spalding's law.
It's a fit of the laminar and logarithmic regions for an equilibrium flow. So, if your boundary layer is in equilibrium, you can use the wall model for a large value of y+.
But because it is never the case, I would advise to resolve as accurately as possible your boundary layer by putting your first of the wall grid point in the viscous sub-layer (Y+<7)

the point is that between 7 and 30, it's the buffer layer and .... people don't really know what happend there that's why people usually avoid these values.

I hope this will help you

Cedric

PS: more information can be found in the turbulenceSIG wiki http://openfoamwiki.net/index.php/Si...Implementation
Hi All,

I have a question about the wall functions of LES in Openfoam. Theoretically, when wall function is used, the velocity at the wall is not zero and instead the velocity is friction velocity which is solved from the log-law or linear law. Then the wall shear stress is obtained tau_w=rho*u_tau**2. Then this wall shear stress can be used for the discretization of the momentum equations. However, in openfoam's specifications, when wall function is applied, the velocity at the wall is still (0 0 0).

Why can we still use (0 0 0) for the wall velocity here?

Thank you in advance.
OFFO

openfoammaofnepo April 11, 2014 06:50
Hi cedric_duprat,

About the LES wall function in openfoam, there is a question I need to ask you:

in nuSgsWallFunction, the velocity at the first cell near the wall is instantaneous, not mean. Is there any possible to implement the time averaged velocity for that quantity?

OFFO

Quote:
Originally Posted by cedric_duprat (Post 211368)
Hi Fabien,

In OpenFOAM, the nuSgsWallFunction correspond the Spalding's law.
It's a fit of the laminar and logarithmic regions for an equilibrium flow. So, if your boundary layer is in equilibrium, you can use the wall model for a large value of y+.
But because it is never the case, I would advise to resolve as accurately as possible your boundary layer by putting your first of the wall grid point in the viscous sub-layer (Y+<7)

the point is that between 7 and 30, it's the buffer layer and .... people don't really know what happend there that's why people usually avoid these values.

I hope this will help you

Cedric

PS: more information can be found in the turbulenceSIG wiki http://openfoamwiki.net/index.php/Si...Implementation

ashvinc9 September 4, 2014 15:44
Wall-function in LES
 
Hi OFFO,

I doubt it’s not possible to implement the time-averaged velocity for the wall-function in LES, because LES resolves the instantaneous flow not the mean (or time-averaged) flow. Opposite to RANS. But in any case, one use the wall-function approach based on the instantaneous flow in LES.

People often say it’s not good to use the wall-function in LES, but what about the high Re flows, such as the atmospheric flow having the Reynolds number~10^(8-9). It is simply too much computationally demanding even using the modern super-computers.

Most importantly, when you use the wall-function try to put your first node at, say y^+>50 from the wall. Moreover, the nuSgsWallFunction is available for LES in OF. But be carefully this wall function is poor in predicting the flow separation.

Hope this helps!

BR, Ashvin

openfoammaofnepo October 1, 2015 16:51
Dear Ashvin,

Thank you. In "muSgsUSpaldingWallFunction", the mu_w (which is for the boundary wall surface) is updated as follows:

Code:
            muSgsw[facei] =
                max
                (
                    rhow[facei]*sqr(utau)/magFaceGradU[facei] - muw[facei],
                    0.0
                );
My understanding is: at the walls mu_w should be zero since there will be no any turbulence there. The mu that should be calculated from the wall function should be for the near wall cells. Could you please give me some comments for what I say above. Thanks!

Quote:
Originally Posted by ashvinc9 (Post 509116)
Hi OFFO,

I doubt it’s not possible to implement the time-averaged velocity for the wall-function in LES, because LES resolves the instantaneous flow not the mean (or time-averaged) flow. Opposite to RANS. But in any case, one use the wall-function approach based on the instantaneous flow in LES.

People often say it’s not good to use the wall-function in LES, but what about the high Re flows, such as the atmospheric flow having the Reynolds number~10^(8-9). It is simply too much computationally demanding even using the modern super-computers.

Most importantly, when you use the wall-function try to put your first node at, say y^+>50 from the wall. Moreover, the nuSgsWallFunction is available for LES in OF. But be carefully this wall function is poor in predicting the flow separation.

Hope this helps!

BR, Ashvin

ashvinc9 October 7, 2015 10:56
Hi,

In current framework of OpenFOAM, all the wall-functions are implemented using subgrid-scale (or eddy) viscosity concept. We therefore write like (e.g. in your case):

Code:
wall
    {
        type            muSgsUSpaldingWallFunction;
        Cmu              0.09;
        kappa            0.41;
        E                    9.8;
        value          uniform 0;
    }
Therefore, muSgs=0 on the wall boundary means "NO wall-function at all". In fact, the mu that is calculated on the wall using

Code:
muSgsw [facei] =           
    max ( rhow[facei]*sqr(utau)/magFaceGradU[facei] - muw[facei],                    0.0 );
is for the near-wall cell centers. I agree that there is no turbulence on the wall-surface and this can be defined by applying the velocity U=0 on the wall. To understand the OF wall-function theory (for LES and rough wall-fucntion) in bit detail see e.g. page no. 31 of this thesis: http://www.doria.fi/handle/10024/103100

openfoammaofnepo April 23, 2016 08:10
Thank you for your reply and the link for the thesis. The rough wall function is investigated. Do you have any comment about the exisiting wall function in OpenFOAM for smooth walls? I mean if you also tested the spalding's law in LES of near wall turbulence?

cheer, OFFO

Quote:
Originally Posted by ashvinc9 (Post 566998)
Hi,

In current framework of OpenFOAM, all the wall-functions are implemented using subgrid-scale (or eddy) viscosity concept. We therefore write like (e.g. in your case):

Code:
wall
    {
        type            muSgsUSpaldingWallFunction;
        Cmu              0.09;
        kappa            0.41;
        E                    9.8;
        value          uniform 0;
    }
Therefore, muSgs=0 on the wall boundary means "NO wall-function at all". In fact, the mu that is calculated on the wall using

Code:
muSgsw [facei] =           
    max ( rhow[facei]*sqr(utau)/magFaceGradU[facei] - muw[facei],                    0.0 );
is for the near-wall cell centers. I agree that there is no turbulence on the wall-surface and this can be defined by applying the velocity U=0 on the wall. To understand the OF wall-function theory (for LES and rough wall-fucntion) in bit detail see e.g. page no. 31 of this thesis: http://www.doria.fi/handle/10024/103100

ashvinc9 June 15, 2016 03:30
Hi, Sorry for my late reply. Yes, I have made a comparison between the two (rough and smooth) wall functions. In the above mentioned thesis, see Figure 2.3 for smooth and 2.4 for rough wall functions results on simple open-channel flow. In more detail, Figures 3.8, 3.10 and 3.11. However, it must be clear that because of the lack of roughness, the smooth wall-function (i.e. nuSgsUSpaldingWallFunction) should not perform as good as the rough one. For this reason the agreement with wind-tunnel measurements using the smooth wall-function is poor.

eb19 February 9, 2017 06:34
Hi All,

Do we still need to use nuSgSpaldingWall if we use dynamic Smagorinsky in OpenFOAM? I use homogenousDynamic smagorinsky model. Let's say the boundary layer in my model is not resolved.

Please help. Thanks

Eb19


All times are GMT -4. The time now is 00:12.