Minimum Recommended Yplus High-Re Turbulence Models
 

Hi All -- relative beginner OpenFoam user.

I am attempting some numerical validation of a geometry with a tight gap/clearance. I was initially running the Low-Re LaunderSharma k-epsilon model with fully resolved boundary layers; maximum yplus of 1.

However, for unsteady and LES simulations, my required timestep for stability is incredibly small -- ~2 orders of magnitude smaller than what'd I'd like it.

So, I decided to abandon the idea of fully resolved viscous sublayers and use a wall function approach -- which I'm familiar with in the commercial world. However, I am struggling to get a mesh "coarse" enough for ensure my first cell lies in the logarithmic region.

If this were Fluent, I would switch on "scalable" or "enhanced" wall treatment to account for this. Knowing that this doesn't seem to be an option, is there is a lower recommended limit for Yplus using OpenFoam wall functions?

My average on my walls of interest is ~9, which my experience tells me is too low for standard wall function treatment, but I'm scared if I go any coarser I will lose ability to resolve flow through the gap properly.

Hi,

Although the normal recommended practice of wall y plus is somewhere between 30 and 500 (limits of log-law layer) for the wall function approach to work well, I remember having read somewhere that the conventional wall function approach may be used up to a y+ lower limit of 11 or so. This might result in reduced accuracy but how much it spoils the solution fidelity might be case-dependent.

If you already have an averge y+ of around 9, perhaps you could coarsen a them a bit more and try it out.

This depends on the wallfunctions, some are capable to work well in the viscous sublayer, low y+ region, but none work well in the region between 5<y+<20, where you are right in with y+=9

http://en.wikipedia.org/wiki/Law_of_the_wall