lowering y+ on surfaces...
 

Hi I just started using CFX5.5 (CFX4 before) but I have a problem. With CFX4 I could build a block next to a wall and make a fine mesh and in that way keep low y+ values (around 11) but I don't know which parameter I should "play" with in CFX5. I've tried the mesh control and inflated boundary (were I adjusted the geometric progression on the relevant boundaries) but still my y+ values are too high. Does anyone have a good idea of what I can play around with in the Mesh menu? Every advice is welcome!!!

Dimitris

Re: lowering y+ on surfaces...
 

Dimitros

What is high?

CFX 5.5 provides two values: 1)SOLVER YPLUS and 2) YPLUS. Forget SOLVER YPLUS. Just look at YPLUS which should be between 11 and 100, provided you use a turbulence model with waal functions. If you use a sublayer model, it should be below 2.

Improving your YPLUS can be established using an inflation layer with 10 elements with a progression of 1.3. You need at least 10 elements in your boundary layer. If YPLUS is too high, just increase the progression factor to 1.4 or 1.5. With a simple Excel-sheet you should be able to calculate your smallest element and estimate the expected YPLUS.

Astrid

Re: lowering y+ on surfaces...
 

Thank you very much Astrid....it works great!!!

Re: lowering y+ on surfaces...
 

Also note that the fixed y-plus (scalable) wall functions in CFX-5.5 are the default when running k-epsilon. You can't get y-plus below 11 with these because the code is assuming that the first control volume at the wall is at the boundary of the viscous sub-layer, not inside it.

You might try running SST instead of k-epsilon which automatically integrates to the wall where it can (i.e., you have enough grid to resolve the boundary layer), and automatically uses scalable where you don't have enough grid.

Either way though, Astrid is right. You should use inflation and try and get some control volumes into your boundary layer if resolving the boundary layer is important for your application.

Neale