Hey, I'm somewhat new to the whole CFD thing and though im learning quickly i've got to a part that has stumped me! see i have this 2d model where there is a small hill that wil produce a seperation in flow. I've opted to use a k-e viscous model as thats what dassault aviation used for a similar model. And according to the userguide an RNG would give better results! anyway.. i've been told that i want a y+ value of between 30 and 300 so that it follows the law of the wall. The problem i find is that when you get to the parts where it seperates the y+ values just dip to values around 0 -5 depending on the grid. The only grid that i've managed to get higher that 30 values for y+ is one that is rather large relative to the model. hence im loosing accuracy due to the large boundary layer mesh! people have said i want a fine B.L mesh though all that i've tried give me really small y+. Its realy stumped me! any suggestions? would the low y+ values at the speration points really effect the result so much?

well if you have yplus near to 3 (manual say 1 would be better), what you can do is use enhanced wall treatment for it. where it needs the mesh in the yplus you are talking about.

1) Use enhanced wall treatment as Zxaar mentioned. For this will need a finer mesh.

2) Adapt the grid that you are already using for Y+ so that you Y+ lies between say 30-60. Then use non-equilibrium wall funtions.