How to handle near wall flow behavior for turbulent flow?

I was looking through the fpga and limerock cases. I noticed that for both cases use the zero equation formulation to handle the turbulence modelling, the zero equation model incorporates wall effects by damping the mixing length close to the wall.

For the other turbulence model defined, the k - epsilon model I don’t see the the equivalent damping terms used to handle near wall behavior.

My question is if its possible to handle no slip wall cases such as fpga with alternative turbulence models such as k - epsilon? If so, what’s the recommended way of doing so.

The k-epsilon equations that are currently in the source code are under development. Hence they do not have any special treatment to handle the near wall behavior as you pointed out. Among the two ways of handling them (low Re models and wall functions), we recently have seen better success using the wall functions approach. Resolving the near wall variation using a Low-Re model might require very high point cloud densities and hence wall functions seem to be an attractive alternative. We are working more towards it and hopefully we can have a few examples showing its use in upcoming releases.

2 Likes

Thanks for the response. I have also been working on a low Re formulation of the K-e model as it seemed like it was more straightforward to implement within the SimNet framework despite the mesh/sampling penalty, its interesting to hear that the wall function approach has been more effective.

I have 2 further questions regarding your wall treatment function:

  • Are you able to share your wall function in it current state? I have a few 2D and 3D cases I would like to try, I’d be happy to share them with you.

  • Additionally, are there any resources that detail the release schedule? This would help decide whether its worth waiting for your release or implementing our own.