[yambanshee]

Good day all,


I am modelling an axial flow fan, and am experimenting with different turbulence models. High on my interest is that of the k-omega sst model.


I have searched around, but mostly have found old and often conflicting views on the correct way to implement the k-omega sst model for a low-Re type simulation (solving through boundary layer).


k is simple enough to set to a low value, but omega I've heard recomendations to use omegaWallFunction or setting it to a large, constant value ( omega = (2*nu)/(C_w2*y_1^2) ). Which is the more correct implimentation for solving through the boundary layer (y+<1). And if the former, what value is used in the value field?


As for nut, I've seen recommendation to use nutSpalartAllmarasWallFunction, nutWallFunction, calculated or 1e-12. And if using any of the wall funcs or calculated, what would be the value once again?


Thanks!

[tjuAeron]

Hi Zander Meiring,
I'm using the same turbulent model at the moment, applying low-Re function.

I found a discussion from
https://www.researchgate.net/post/Sh..._wall_function

It seems that omegaWallFunction is suitable for both low and high Re.The condition based the blending of viscous and log layer.

For nut, in fact, there is a nutLowReWallFuntion.

tmp<scalarField> nutLowReWallFunctionFvPatchScalarField::calcNut() const
{
return tmp<scalarField>(new scalarField(patch().size(), 0.0));
}

The code shows that nut is set to zero