y+ values
 

Hi, Can someone confirm my understanding of scaleable wall functions and the importance of y+ values. Is it true that there is no recommended minimum y+ value? So y+ values of around 1 to 3 are ok? If I want to resolve the flow profile in the boundary layer, then I should just make sure there are sufficient nodes, right?

Thanks in advance!

Mark

Re: y+ values
 

Hi, Mark

To my understanding, yes, you are right. Scalable wall function can overcome the problems of inconsistencies with fine mesh (the first node is down to the sub-layer). If you are trying to resolve the boundary including the sub-layer, sub-layer models with y+ values lower than 2 are recommended.

Regards, Forrest

Re: y+ values
 

Thanks!

Re: y+ values
 

This is not the whole story.

k-e with scalable wall functions was a first step towards "y+ independent" meshes. In case of a y+ value below 11 anywhere in the model, y+ is set to 11 locally. So the node is shifted virtually out of the viscous sublayer. So wall functions are applied everywhere. This is a pragmatic approach but it leads to results better than ordinary k-e with meshes where y+ locally lies outside a range between 20 and 200, the well known y+ criterion.

The next step was the introduction of the SST model in which the wall treatment is managed automatically. Where y+ locally is below 11 the SST model switches from a wall function formulation to a low-Re formulation where the flow field is integrated down to the wall (no wall functions are used)

If you have a mesh where y+ is around 1 allover (which is better regarding accuracy), you also can use SST because the model then behaves as a full low-Re model

Re: y+ values
 

Hi Helge,

I just look at the solver theory. It used Y* instead of Y+. ie. the alternative velocity scale is used instead of friction velocity in Y+.

I wonder if Y+ & Y* are the same. If I used friction velocity in the Y+ formula, can I still follow the criteria stated in the manual (i.e. 20<y*<100 if wall function approach is used)?

Thanks for advice.