CFD Online Discussion Forums - About the yPluw and wallFunctions in the RAS calculation

Hi, foamers here!

I have a question about the calculation using RAS model.
briefly the question is, how to set the wall boundary conditions when the yPlus have a wide range (like 0.1-100)?

As we know, for the ke model, the yPlus must located in the log region which means yPlus>30, and then wall functions should be used.
However, if we can not make sure that the yPlus of the mesh all >30, what should we do?

For example, in my calculation of a T type pipe flow, the Re is larger than 50000 so I think I should use a RAS model.
Because a stagnation region appears at one end of the pipe ( this end is closed), the velocity in that region is very small and the yPlus also becomes small too, about 0.001 order.
Other places where the flow velocity is large, the yPlus ranges from 25 to 100.

At first, I did not notice the yPlus problem and set all the walls like:
epsilon, epsilonWallFunction
k, kqRWallFunction
nut, nutkWallFunction
and used chtMultiRegionSimpleFoam (OF-v1712) with RNG ke model , the result converged, the residual of p_rgh is smaller than 1e-6.

However, after check the yPlus I found some place did not meet the criteria >30 as I said above.
I read the source code of the wallFunctions and found that some functions can hold both low- and high-Re boundaries.

so I tried again with:
epsilon, epsilonWallFunction with lowReCorrection on
k, kLowReWallFunction
nut, nutkWallFunction
But this time the result does not converge, the residual of p_rgh oscillates around 0.001 order.

( linearUpwind is used for div schemes of all variables, nNonOrthogonalCorrectors in SIMPLE is 1, relaxationFactors for p_rgh is 0.3 and for U is 0.7)

so, I hope someone can give me some hints about this problem.
Thank you!