CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   Prism Layers in the corner (http://www.cfd-online.com/Forums/star-ccm/78332-prism-layers-corner.html)

bengy July 19, 2010 04:27

Prism Layers in the corner
 
hello,
how to improve prism layers in the corner? I mean what I should do to create prism layers which are not convergent in the corner. I have very good mesh so informations and tips from help doesn't work (it works only for poor quality mesh).

Maddin July 21, 2010 02:55

Show a picture plz.

bengy July 21, 2010 03:18

I have some problem with this case
 
1 Attachment(s)
that's it.......

Maddin July 21, 2010 05:06

You can't do anything. Support also didn't know a work around...
I had the same problem...

bengy July 21, 2010 06:34

yeah that's true - support also didn't know. it's very strange so I decided to concentrate mesh closely to boundary (between prism layers and core mesh)

djack July 31, 2012 05:35

Prism Layer Crossover
 
Hi everyone,
I'm meshing a container using star ccm trimmer and prism layer model.
I'd like to know if it was possible to crossover two different prism layer in a corner to increase mesh definition in that region.
If possible that may be a solution for bengy
thanks,
Djack

siara817 July 31, 2012 05:59

I think if you reduce the prism layer stretching in the mesh/reference values/prism layer stretching, you can have finer meshing.

rwryne July 31, 2012 08:50

Quote:

Originally Posted by bengy (Post 268270)
that's it.......


Are you using 7.02 or 7.04? If so, try using the "advancing layer mesher". It is a new meshing model that will handle those corners better. read about it in the help file, they have screenshots of a Poly+Prism vs Advancing Layer for very similar geometry.

rwryne July 31, 2012 09:03

1 Attachment(s)
Here is a quick example I made comparing the two....I did exagerate the prism layer thickness to highlight the differences.

djack July 31, 2012 09:07

I'm using 6.04...
Anyway thanks for the answer.
djack

rwryne July 31, 2012 09:18

Quote:

Originally Posted by djack (Post 374581)
I'm using 6.04...
Anyway thanks for the answer.
djack


We don't get to say this often, but try coarsening your grid. It seems to happen as a function of OverallHeight / Surface Cell Size. Just keep increasing your surface cell size until the problem goes away, and then see if this is an acceptable size....if not, try decreasing your overall prism layer height.

eRzBeNgEl August 16, 2012 15:38

the new mesh type advanced layer meshing is the solution for this problem

djack August 17, 2012 03:38

I never heard about that mesh type. Is it available in Star ccm 6.04.
Djack

rwryne August 17, 2012 08:40

Quote:

Originally Posted by djack (Post 377455)
I never heard about that mesh type. Is it available in Star ccm 6.04.
Djack


It is the same mesh type I was referring to in my posts from 3 weeks ago, and it was introduced in 7.02 and improved in 7.04

djack December 10, 2012 16:26

I'm now using starccm 7.06 and as you said it works !
Thanks you all,
Djack

abdul099 December 14, 2012 21:34

The main question is: Why do you think, this mesh would be bad in the corner?

djack December 17, 2012 04:43

I see two arguments :
- Actually the velocity magnitude in the corner is smaller than far from the corner, then the solution may be to increase the prism layer height in the corner (law of the wall). That's definetly not the case for the old prism layer meshing system where the prism height was decreasing.
- I guess it's better to keep the same prism layer structure when geting closer to the corner.

renee November 7, 2013 09:12

I know this topic is a bit old but I am currently facing the same problem.
My Prism Layer Setup results in the same way bengy posted with his picture above. I did a short mesh study by variing different values. So far it looks like my prism layer thickness is to high.

Can somebody please explain me why the weird corner prism layer structure could cause a lower mesh quality? Sure, the y+ value should be constant but the y+ is only dependent on the first layer hight which not too much affected as you could see in the picture bengy posted.

Thanks a lot in advance.

Rene


All times are GMT -4. The time now is 05:31.