Can I expect accurate turbulence modelling if my y+ is in the buffer layer?

Using SimScale
1

If you’re y±value happens to land in the buffer layer (1<Y+<30), we dynamically blend between wall functions and fully resolving the turbulent boundary layer. This delivers acceptable results, however we still generally recommend that your y±value lands within the log layer of the wall (Y+>30) or the viscous sublayer (Y+<1). If you are confident that your Y+ will be less than 1 and want to force a full resolution solve we also allow for this. Still our default behavior with K-Omega SST is to do dynamically blend wall functions with fully resolving the turbulent boundary layer.

You can find more information in our documentation:
How to Control, Export, and Analyze Y+ During Meshing, Simulation and Post-Processing?
What is y+?

1 Like
2

Hi @ataleb , really nice post. I want to add that this advice is specific to the default turbulence model for most solvers, K-omega SST. There are different requirements for different turbulence models, for different solvers, and for different applications. So always best check documentation if you are in doubt or contact the forums for advice when operating outside the default settings.

Best,
Darren

1 Like