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] Uncovered edges on Fluent mesh (http://www.cfd-online.com/Forums/ansys-meshing/117424-uncovered-edges-fluent-mesh.html)

Fontan May 8, 2013 08:16

Uncovered edges on Fluent mesh
 
2 Attachment(s)
Good morning,

I'm trying to create a 2D mesh for a scramjet intake to be used in Fluent. In order to simulate the viscous flow, we need to condense the mesh along the walls, so I had to divide the geometry in multiple blocks in order to condense where needed, as seen in the figures, along with the pre-mesh with focus on where I need it condensed.

The issue is, whenever I try to export it as a .msh file for Fluent, it tells me I have uncovered edges. I have tried rebuilding the geometry edges in ICEM from the points coordinates, but to no success.

Maybe it has to do with the interior edges I have created to separate the blocks? I'm out of ideas as to what to do.

Thanks in advance for any help.

Far May 8, 2013 09:30

It is the association problem. Check all your boundary edges are green.

Fontan May 8, 2013 09:33

Far, thanks for the reply.

I've checked and all the edges are green. I've been careful furing association in order to avoid any issues specially at the wedges.

Far May 8, 2013 09:34

Attach files... let me check them

Fontan May 8, 2013 09:51

Files are too big for the forum, I've uploaded them HERE.

I can also send them directly to your email if you prefer. If there are any files missing, please let me know.

Thanks again for the help.

FJSJ May 8, 2013 11:50

Hi Fontan,

How did you do the geometry? I think the problem is the geometry. Iīve tried to do the mesh too and I got the same problem. Then Iīve tried "Build Diagnostic Topology" because Iīve seen some weird connections and if you click in "Delete unattached curves and points" all the curves dissapears. So.. how did you do it?

Fontan May 8, 2013 12:04

Thanks for the reply, Javi.

I did it using AutoCAD 2013 and exported as an .iges file.

I suspected the geometry to be the issue and tried deleting the edges and reconstructing it entirely in ICEM using the intersecting points, but it didn't work as well.

Far May 8, 2013 12:23

1 Attachment(s)
Ok. Problem was solved with merge vertices operation. You have disconnected blocks and issue resolved with merging vertices.

Another issue was association of blocks edges with internal edges, this should be avoided. Only associate perimeter edges where boundary is defined.

Also there were no points, so also made them and vertices were associated with respective point, again be careful in over-associating vertices.

Files are attached and you can specify spacing as deemed necessary. :)

FJSJ May 8, 2013 12:51

Hi Far,

I donīt understand you when said "you have disconnected blocks". Is it possible? And... when you solve this issue.. you use "merge vertices"... but.. what are those vertices?

About points. I created the points and then associated them with vertices. I think is better in that way. But, are they necessary in this case?

Far May 8, 2013 14:09

please refer to following pics ...

http://imageshack.us/a/img703/7604/22419471.png
http://imageshack.us/a/img577/6518/36538004.png
http://imageshack.us/a/img843/4904/45615419.png

FJSJ May 8, 2013 15:33

Thank for the explanation Far.When I saw the blocking strategy it seemed a bit weird. In fact, I merged the blocks (exactly where you write "no need" in purple). And.. whatīs about points? Are they necessary?

Far May 8, 2013 15:37

Actually it was by default in my mind that outermost vertices should be associated to points as explained in ICEM tutorials (static mixer and sphere cube).

Quote:

blocking strategy it seemed a bit weird
Overall blocking strategy is perfect. But when he made split horizontally (making three blocks), he deleted blocks which are not being used but forgot to merge vertices in the interior of intake.

Better approach would have been to collapse blocks in that region, thus avoiding the need of merging vertices. Just think if there were 100's of parallel blocks with same problem, it would have been cumbersome to manually merge each vertex pair.

Fontan May 8, 2013 15:45

Far, thank you very much. I'm heading home right now, but I'll check it out as soon as I get there.

Thank you as well for the explanation regarding the merging of the vertices, I'll be sure to keep that in mind in the future.

FJSJ May 8, 2013 15:49

Thank you again for you advise!


All times are GMT -4. The time now is 11:47.