CFD Online Discussion Forums
Page 2 of 2 1 2
Show 40 post(s) from this thread on one page

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Wall treatment with OpenFOAM (https://www.cfd-online.com/Forums/openfoam-solving/102246-wall-treatment-openfoam.html)

KateEisenhower November 20, 2015 03:13
Hello foamers,

As I understood from the source code, the omegaWallFunction is continuous and so can be used for a low-Re simulation as well. My boundary field in the omega dictionary looks as follows:

Code:
no_slip_wall
{
    type            omegaWallFunction;
    value            uniform ???;
    Cmu              0.09;
    kappa            0.41;
    E                9.8;
    beta1            0.075;
}
I am not sure what I should set as value. Is it reasonable to take the value I have calculated for my velocity inlet?

Best regards,

Kate

RodriguezFatz November 20, 2015 03:16
As far as I know, "value" is only used for the very first iteration of the very first time step. So unless the solution does diverge, I think you can set any value you want. Since this is a wall, I would go for a very high value.

akidess November 20, 2015 03:43
A possible guideline is
\omega_w = 6 \nu / (\beta * y^2)

I'm not sure if the standard kOmega model in OpenFoam is suitable for lowRe setups. Anyway, as Rodriguez said, it's just an initial guess, but why not make an educated one ;)

See also: http://www.cfd-online.com/Wiki/Near-...k-omega_models

RodriguezFatz November 20, 2015 03:45
Anton, as you can see from the formula you posted, omega goes to infinity at the wall. So it will never be a good guess to set omega to zero, not for high-Re and not for low-Re. It's always pretty high, such as 1e8.

akidess November 20, 2015 03:54
Right, I mixed something up there. Fixed now, thanks!

KateEisenhower September 5, 2016 05:41
Quote:
Originally Posted by RodriguezFatz (Post 574143)
As far as I know, "value" is only used for the very first iteration of the very first time step. So unless the solution does diverge, I think you can set any value you want. Since this is a wall, I would go for a very high value.

Hello all,

can anyone confirm this? Is this also the case for other turbulent paramters at a wall?

Best regards,

Kate

Giuseppe92 March 2, 2017 07:38
Hello,
I found this post very intersting and I would like to clarify something about the k-omega SST model in OpenFoam.
I have understood that if I want to use wall function I need a y+ between 30 and 100 for a good result. In order to do this I can write like this in 0/k file for a boundary:

liner
{
type kqRWallFunction;
value uniform 4e-07;
}

If I want to solve the boundary layer I need a y+<=1 first of all and not to call the wall function.
My question is: How can I manage k-omega SST without wall function for example on the above boundary for k?

Thank you in advance.

Best Regards,
Giuseppe

crizpi21 July 21, 2018 21:00
Quote:
Originally Posted by Giuseppe92 (Post 639171)
My question is: How can I manage k-omega SST without wall function for example on the above boundary for k?
Hi Giuseppe,

Did you find any answer to your question?
I am also interested in how to model the boundary layer using k-omega SST
(I am running a simulation of the flow past a bluff body), either using wall functions or directly solving the boundary layer.
Maybe it would also be helpful if you could share your case.

Thanks in advance

Cristina

guanjiang.chen May 28, 2021 11:38
Quote:
Originally Posted by Giuseppe92 (Post 639171)
Hello,
I found this post very intersting and I would like to clarify something about the k-omega SST model in OpenFoam.
I have understood that if I want to use wall function I need a y+ between 30 and 100 for a good result. In order to do this I can write like this in 0/k file for a boundary:

liner
{
type kqRWallFunction;
value uniform 4e-07;
}

If I want to solve the boundary layer I need a y+<=1 first of all and not to call the wall function.
My question is: How can I manage k-omega SST without wall function for example on the above boundary for k?

Thank you in advance.

Best Regards,
Giuseppe
I think the type should be fixedValue or calculated.


All times are GMT -4. The time now is 07:42.
Page 2 of 2 1 2
Show 40 post(s) from this thread on one page