Hi Guys

I'm having a had time achieving a yplus value of <50 for my work. Hexa is used to generate the grid but it seems is not allowing me to contract the nodes towards the wall. It does allow me but the skewness has zero values and as such complain about volume disorientation when I did MESH EDITOR Diagnosis. The geometry is draf tube with 3 outlets. From CFX, I found out that the first node is 15mm from the wall but I needed <0.3mm to be able to get <50. Do anybody has an idea on this? I try Bigeo, Geomtry 1, etc. bunching law but none of them is able to solve the problem.

if you need more info let me know and I'll be much delighted to furnish you with it.

if you've idea to help me out, just dropped me few lines!!

Regards!!

If I need smaller mesh size at the wall I just use BiGeometric and specify required size(spacing) at the ends. Usual I get what I want, sometimes just need use of meshing options(projection limit) when I dealing with fine grid. (Of course sometimes geometry must not have gaps of element size...)

regards, deLuther

Hi DeLuther

Thanx!! But when I specified the required size, the skewness of some of the elements are less than 30 and even zero. Also the volume is less than zero. The problem is not the required size alone but getting the required size and at the same time fufilling the grid quality(skewness)is my problem. Did you check the skewness and got less than say 9 degrees or even zero and CFX5.7 accepted it without complaining about it? Please quick response on this will be much appreciated expecially about the skewness!!

thanx!!

When I adjust edge parameters I use "copy to parallel edges" option at begining ot tuning, by the end of tuning process, of course, this can destruct parameters of some edges. Usualy I get(in my cases) angles greater than 9 degrees(in HEXA). If you have volumes less than zero that cover large area, then you need you might need "meshing->invert block->inverted" operation. Sometimes block after some operations(like extruding) can have volumes less than zero. You can check it by using Display->Blocks->Worst option and look into messages window to check that block does not have negative determinant. (if so just invert blocks). Maybe this can help you... regards, deLuther

Hey deLuther,

Unfortunately for me, my "meshing->invert block->inverted" is not activated. I presumed because I selected "unstructured" mesh type under "file->msehing type". When I selected Unstructured or Structured meshing type, it was activated. I gave it a shot by ff what you said and which it did show the "worst" blocks but it didn't solved my problem. I still have some of skewness of my elements being less than 9degrees and even zero. But the strange thing is that when I specify my first spacing of Bigeo to say 15 or 20, all these crabs are removed but that gave me y+ value of thousands. Which means I mean the first spacing to be 0.2 or so to given a better y+ value. This is where the problem of skewness and negative volume(disorientation of element) crop up!!

what I don't understand is that when I was using CFX-Build, I could go as low as 0.02mm. Is it because the geometry was too simple since that one was just a conical diffuser but this is draft tube with 3 outlets? or because I used structured grid in CFX-Build and now I'm using unstructured grid? is there a way where I can change the "grid tolerance" in HEXA?

Thank a lot!! I appreciate your contributions!!

Regards

CFX-Build deal only with ustructured grid because it uses tetrahedral mesher. For further help, if need, you can send me geometry and topology files, it think these files not large.

regards, deLuther

Try splitting your blocks very near the wall and using edge parameters there to get a smooth boundary layer. Use Geometric 1 or 2 to get things to expand the right direction with an expansion factor of 1.2 to 1.4. Then use BiGeometric in the central block to match the edge lengths on either side.

Jeff

KKA,

I check your files. So I think some problems is in geometry. I also have similar problems with trimmed surfaces and intersection curves. So try to fix geometry. And also I noticed that there another problem with edges paremeters. It can be fixed by using copy to parallel edges option.

regards, deLuther

When grid accuracy counts, use Gridgen instead.

Thanks deluther! But I'm not clear with these sentences:

"And also I noticed that there another problem with edges paremeters. It can be fixed by using copy to parallel edges option"

please what those suppose to mean? Because I used the "edges parameters" by selecting the circular side edges and use "copy to llel edges" to affect all the parallel edges. so else are you talking about? I also noticed that the spacing and ration i specified at the circular is "not" exactly the same as the those in the outlet sections. say i specified spacing of 10 at the circular section, the outlet will be 9 or 9.5 despite I used the "copy to parallel option".

anyway, i'll give it a shot by trying to fix the geometry!!

Regards