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/)
-   -   [ICEM] Prism Mesh on a semisphere (http://www.cfd-online.com/Forums/ansys-meshing/118693-prism-mesh-semisphere.html)

Will Anderson June 2, 2013 13:27

Prism Mesh on a semisphere
 
2 Attachment(s)
Why does ICEM produce an erroneous Prismmesh on this simple semisphere?

Replay script and graphical result attached.

diamondx June 2, 2013 18:12

the problem resides in the geometry. make sure all the surfaces are checked for prism...
You can share your project so we can investigate with you...

Will Anderson June 2, 2013 18:15

Thank you diamondx.
In fact, i have found a different but related problem.
The symmetry surface and the ball must share an intersection curve.
They hadn't before.
Once the intersection curve between the symmetry plane and the ball was implemented (Geometry->Curve->Intersection of Surfaces), the mesh looks quite normal.

PSYMN June 2, 2013 21:11

5 Attachment(s)
Yes, but the fix is simpler than that...

The bad intersection caused the problem...
Attachment 22347
This is because your small circle was offset a bit, which crumbled the edges so prisms couldn't form.

Here you can see that the prisms are fine away from the bad edge...
Attachment 22348
See here in the script... Why did you set the center of the sphere 1 unit below the symmetry plane?
Attachment 22349
Once I fixed that first line (to go from 0 to 1 instead of -1 to 0), it worked no problem...
Attachment 22350

Will Anderson June 2, 2013 23:30

Thanks Simon. Why did I offset the sphere center by 1 mm? In order to be more sure that the surfaces intersected.

Thanks everyone.

PSYMN June 3, 2013 10:31

Quote:

Originally Posted by Will Anderson (Post 431536)
Thanks Simon. Why did I offset the sphere center by 1 mm? In order to be more sure that the surfaces intersected.

Thanks everyone.

OK, but it was probably unnecessary. ICEM CFD can handle some misalignment, but in this case, it was just too great compared to the mesh size.


All times are GMT -4. The time now is 14:09.