Openfoam Wall Functions

Openfoam appears to have many different types of wall functions:

https://www.openfoam.com/documentation/guides/latest/api/group__grpWallFunctions.html

For the k-w SST model, which wall function actually used in Simscale? Do you have any documentation that maps which wall function is used foreach model in the Simscale framework?

The following documentation clams that [omegaWallFunctionFvPatchScalarField] is used in the k-w SST model but I wanted to get some confirmation on that.

Page 21:

As the theory mentioned before, the omega wall function provides the combination of viscous and
log equation. In the OpenFOAM omegaWallFunction is a special wall function which can switch
between viscous and logarithmic region according to the position of y> +. In the intersection of the
viscous sublayer and log-law region value is calculated through blending the viscous and log-law
sublayer value.

According to this k-w SST uses a blended wall function that is good for all y+ values. Is this accurate information for SimScale as well?

Thanks

Hello,

Briefly going over the questions:

  • I believe this is not explicitly in the docs, but it’s possible to check. By downloading a simulation result, the type of wall function is going to be in the result sets. For instance:

So yes, you are correct: omegaWallFunction is used for omega.

  • You are also correct about the blended wall function statement. This wall function contains a built-in mechanism to switch the wall treatment, based on the local y+.
  • Yes, there is a way to obtain the OpenFOAM inputs, by downloading the result set. Naturally some of the input may be custom/different from the OpenFOAM implementations.

Cheers

3 Likes

Thanks!

I’m only going to add something: The k-omega wall function is not appropriate for all y+ values. It is usually accepted that you should stay between 30 to 150. The Shear-Stress Transfer model (SST) blending the k-omega and k-epsilon models depending on the distance to the wall (among other factors). If you want to go through the theory, you can check: “Two-equation eddy-viscosity turbulence models for engineering applications”, by Menter (where it is discussed that the model is reliable as low as y+ = 3). However, be advised that this operation depends on how the wall function is programmed, so you will have to do your own tests. There is a juicy discussion here: y+ selection with epsilonWallFunction and omegaWallFunction -- CFD Online Discussion Forums and here: One more time: wall functions - SST -- CFD Online Discussion Forums. By the way, why would you like to go lower than y+ = 30 using wall functions?

2 Likes

By the way, why would you like to go lower than y+ = 30 using wall functions?

With the geometry I am working with, the velocity varies quite a bit within the volume so the y+ is changing quite a bit with location. Also, the inflation boundary seems to get deformed in a few locations so it is very difficult to keep y+ within 30 to 300 for 100% of the cells.

Good enough. I’d keep the correct y+ value around the most critical regions of the flow.