[alexeym]

Dear all,

During preparation of the OpenFOAM validation cases based upon http://turbmodels.larc.nasa.gov, I came across http://turbmodels.larc.nasa.gov/openfoam_issues.html (http://turbmodels.larc.nasa.gov/ChangesToOpenFOAM.pdf).

To be short, OpenFOAM's implementations of SA, SST, and Wilcox models are not quite correct (or "...did not match the equations given on this website").

And my question is: should it be reported as a bug? Was not able to find any discussion on this topic. Or maybe bug report should just contain patch with code corrections?

[wyldckat]

Hi Alexey,

I know that sometime ago there was a report about a 6.0 or 60 factor... it's this one: http://www.openfoam.org/mantisbt/view.php?id=1245

And I know there was a fix done a few weeks... er, months ago? Related to f23 or f31... ah, here it is: https://github.com/OpenFOAM/OpenFOAM...7bf8ad06b280ba
But this is odd, because I remember another similar change done more recently... it might have been this: https://github.com/OpenFOAM/OpenFOAM...23b343f30465d4

Either way, there are two details that come to mind:

1. It's necessary to diagnose which fixes have already been implemented.
2. It's necessary to validate what these changes affect, if anything at all.

Reporting this as a bug is probably a good idea, at the very least for keeping it on record.
The problem might then be if and how these issues can be fixed or new code can/should be added with these specific implementations.

The other detail that I'm find very strange is that the PDF file does not specify which OpenFOAM version it was tested with...

Best regards,
Bruno

[alexeym]

Hi Bruno,

Thank for the reply. Guess, I should start with verification/validation cases and through comparison of the sources code (let's say 2.3.x branch) and proposed corrections. And after file a bug report.

Quick look at Spalart-Allmaras implementation revealed that the first correction in PDF is already in the code

tmp<volScalarField> SpalartAllmaras::fv2
(
    const volScalarField& chi,
    const volScalarField& fv1
) const
{
    if (ashfordCorrection_)
    {
        return 1.0/pow3(scalar(1) + chi/Cv2_);
    }
    else
    {
        return 1.0 - chi/(1.0 + chi*fv1);
    }
}

But default value of ashfordCorrection_ is true, so, I think, correction proposed in PDF is usually ignored.

[sfernaferna]

Hi,

Does anyone know if by OpenFOAM 7 these issues were fixed? Thanks!

[dweaver123]

I'm not sure about Spalart but for k-omega SST it seems they have not been changed and I'm not sure why. I have changed them to match the original papers [1] along with the updates[2]. and can send you the files if you'd like.

As for the k-omega Wilcox model: It seems the 1988 model is implemented instead of the most recent model which has some developments that make it less sensitive to freestream conditions. I personally haven't had success with even the newest k-omega Wilcox model [3] because of the sensitivity to free stream conditions[maybe check: 4].

All this being said, if you're not limited on the size of your mesh or restrictions because of multi-phase flows then I'd suggest you use a fine grid with LES .


[1]Menter, F. R., 1994, “Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications,” AIAA Journal, 32(8), pp. 1598–1605.
[2]Menter, F. R., Kuntz, M., and Langtry, R., “Ten Years of Industrial Experience with the SST Turbulence Model,” Heat and Mass Transfer, p. 8.
[3]Wilcox, D. C., 2008, “Formulation of the K-w Turbulence Model Revisited,” AIAA Journal, 46(11), pp. 2823–2838.
[4]Menter, F. R., 1992, “Influence of Freestream Values on K-Omega Turbulence Model Predictions,” AIAA Journal, 30(6), pp. 1657–1659.