[OliverB]

Hello,

I was wondering if the use of nutUSpaldingWallFunction for k-e models in OF can be reasonably used instead of enhanced wall treatment of commercial codes ?
I couldn't find much information about it.

Thanks

[tas38]

The tentative answer is yes, this wall function attempts to provide a smooth transition between the laminar sublayer and log-layer. You should consult Spalding's original 1961 for a description.


However, this wall function is not the same blending/enhanced wall-function approach that is applied in commercial codes, say ANSYS fluent. Most notably, Spalding's universal function does not include adjustment's for adverse/favorable pressure gradients. If you take a look at the attached photo, I compare the Spalding wall-function with that of Fluent (less the pressure-dependent terms).


In practice, however, I have found it difficult to get grid-independent results using this nut..Spalding... wall function for y^+ > 1.0, the precise reason I chose to use it.

[cryabroad]

I did some similar tests a while ago and came to a conclusion that nutUSpalding is good for y+>5. For y+<5 I struggled to find a very good wall function though. I will add that my tests are not that rigorous, and my intention was to get a good mean velocity file at the end of a pipe (which feeds fluid into a much larger room).

Commerical codes tend to use a two-layer approach near the wall, and they use another nondimensional wall distance (y*) as the switch between these two layers. I think any wall function in FOAM that has this funtionality should be good. There's a very detailed document that explains wall functions in FOAM from Chalmers, just google it. I think it will help you a lot.

[OliverB]

Thank you both, your answers are very usefull.