In my experience a not too uncommon problem when using wall functions is that a cell with y+ ~30 is too large to resolve the flow domain geometry and you cannot afford to resolve the boundary layer (y+ ~1 in the first cell and a not too rapid cell growth in the wall normal direction).
I'm currently simulating a domain with a lot of jets with 2 mm diameter. I have just made a mesh with just below 10^6 cells. y+ in the wall cells in the region of interest is between 3 and 10. If i resolve this in a way that is suited for the two-layer model i will end up with something like 5*10^6 cells.
I know that Fluent modifies the wall functions if the first cell is located in the viscous sublayer. But does this really work or is it just a quick fix?
So dear CFD-community, what is your experience? Should I use too low y+ together with wall functions or should I use my badly resolved boundary layer together with the two-layer model? What will, generally of course, give the worst result?
Re: y+ Dilemma
The wall-function modification you mention is just a quick fix. It does not give good results. Fluent lacks a good model for this type of problems. I think that they are working on somthing better for version 6 though. If you use wall-functions try the non-eqilibrium wall-functions - they seem to behave a bit better in these cituations.
However, I would use the two-layer model and try to improve the resolution slightly - y+ around 3 should give you decent results. The Wolfstein two-layer model is a bit more tolerant with high y+ values than most other Low-Re models.
|All times are GMT -4. The time now is 15:25.|