[lac]

Hello!

This is my first thread on this great forum (although I'm reading the posts since a while), so welcome everyone.

My problem/question is (as the title) related to the wall functions for epsilon in OpenFOAM. I'm simulating internal flow in a duct. The flow should not be very complex, but the geometry, which I meshed with snappyHexMesh, is. Therefore there are some regions of the geometry where y+ is very low, but at other locations it is higher than 30.
Therefore, although using the k-epsilon model, I'm trying to used low Re wall functions for epsilon, which, according to my understanding is default in OFv5. The problem is, that at some locations I have experienced (by monitoring the solution with probes) frequent switching between the 'laminar' and the 'turbulent' values. As it seems to be like, this switching is not continuous between the two values and I suspect this to be the reason the solution stuck at a residual level. On the other hand with OFv4, it goes further down. I have attached four pictures. The first two show the results from OFv4 and the third and fourth the results from OFv5.

I hope that someone has a solution for this problem, or the code can be modified in such way to avoid this in the future (if it is possible), but unfortunately I lack the skills to do that.

Thank you for your suggestions and answers.


PS. The valleys in the residuals and the discontinuities are the result of my experimentation with different schemes.