CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [GAMBIT] Cannot mesh with cooper scheme (http://www.cfd-online.com/Forums/ansys-meshing/116928-cannot-mesh-cooper-scheme.html)

ghost82 April 28, 2013 14:29

Cannot mesh with cooper scheme
 
4 Attachment(s)
Hi!
I have to mesh the domain you can see in the attached pictures, with full hexa.
It id a "double helical coil" immersed in an aluminum block".
The domain is composed by 4 volumes.
I successfully meshed faces and I wanted to mesh the each volume with the cooper scheme, but gambit returns:

"Volume cannot be meshed because two sets of source faces which form natural starts for the projections could not be found."

I don't understand, in my opinion the 2 sources are of (bottom and top)..
What's wrong?

Thank you,

Daniele

-mAx- April 29, 2013 01:37

try to split the volume along (yz)
Instead of 180 rotation volume, you will have 2x 90 (it may help).

ghost82 April 29, 2013 03:30

1 Attachment(s)
Quote:

Originally Posted by -mAx- (Post 423743)
try to split the volume along (yz)
Instead of 180 rotation volume, you will have 2x 90 (it may help).

Hi Max, thanks for your reply,
I managed to split each of the 4 volumes into different volumes by projecting the spiral edges on the top/bottom surfaces and mesh them with hex submap.
I think of meshing the central zone of the domain with map, but I have too coarse regions in that zone, so the only solution I find is to split each of these zones into "3" volumes, assign fluid names and refine them directly into fluent: do you think is it acceptable (see sketch in the picture)?.

Thank you,

Daniele

-mAx- April 29, 2013 03:55

-Are you now able to mesh the volumes, which gave you problem?
-can you refine non-tetrahedral zones in fluent?

ghost82 April 29, 2013 03:59

Quote:

Originally Posted by -mAx- (Post 423768)
-Are you now able to mesh the volumes, which gave you problem?

Yes, now problem is solved; all volumes are meshed with hexa; my doubts are about the central zone(s).

Quote:

-can you refine non-tetrahedral zones in fluent?
I think so, I never tried, but in the past I was able to refine 2d quadrilateral cells, I have to try..

Thanks,

Daniele

ghost82 May 1, 2013 09:23

4 Attachment(s)
Dear all, dear Max,
my idea does not work very well...
For completeness I can say that fluent let you refine hexa mesh (see attach image): the cube is splitted into 8 regions and refined different times; when refining 2 times fluent automatically adds refining to adjacent zones, in other word you can have adjacent cells (edges) smaller/greater than 50%/200%.

As I wrote I was successfully in meshing "peripheral volumes" with hexa, but my problem remains at the center of the spiral; any comments on splitting the central volume is welcome.

Daniele

-mAx- May 2, 2013 02:05

Hello Daniele,
Once you are sure that your "already meshed volumes" are hexa-meshable, I would delete those mesh, and I would start meshing the center of the spiral.
If you don't have any meshed volume, then you are "free" with hexa-start-meshing.
Once the center of spiral is meshed, you won't have any problem to mesh the rest (but I think you already know that).
Regarding decomposition, maybe you could post a picture with only the center of the spiral (shaded mode on), to see more details.

ghost82 May 2, 2013 02:48

Quote:

Originally Posted by -mAx- (Post 424552)
Hello Daniele,
Once you are sure that your "already meshed volumes" are hexa-meshable, I would delete those mesh, and I would start meshing the center of the spiral.
If you don't have any meshed volume, then you are "free" with hexa-start-meshing.
Once the center of spiral is meshed, you won't have any problem to mesh the rest (but I think you already know that).
Regarding decomposition, maybe you could post a picture with only the center of the spiral (shaded mode on), to see more details.

Thank you Max,
I'm attaching directly the dbs file, with part of geometry (center + adjacent volume): volumes to be problematic for me are volume 49 and 63.

Here the dbs:
http://www45.zippyshare.com/v/47192229/file.html

Thank you,

Daniele

-mAx- May 2, 2013 09:37

1 Attachment(s)
Hello Daniele,
I don't have time enough to really test and mesh your geometry, but did you try such a decomposition?
The "Y" could be more reproduced, if your mesh get coarser.
For volumes 49 & 63, you will have to split them with plane parallel at (xz) with y-offset 10.334201 (you already have some points at this level)
Attachment 21398

ghost82 May 2, 2013 09:57

Quote:

Originally Posted by -mAx- (Post 424676)
Hello Daniele,
I don't have time enough to really test and mesh your geometry, but did you try such a decomposition?
The "Y" could be more reproduced, if your mesh get coarser.
For volumes 49 & 63, you will have to split them with plane parallel at (xz) with y-offset 10.334201 (you already have some points at this level)
Attachment 21398

Thank you Max for your input, I'm working on it.


All times are GMT -4. The time now is 09:45.