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] Problems with mesh for annular-type geometry (http://www.cfd-online.com/Forums/ansys-meshing/98790-problems-mesh-annular-type-geometry.html)

Kristin_N March 19, 2012 11:17

Problems with mesh for annular-type geometry
 
2 Attachment(s)
Hello,

I'm new to ICEM CFD and I have encountered some problems I hope someone can help me with.

I'm trying to mesh the cylinder/annulus shown in the attached picture. It has an annular region at one end, but the inner cylinder doesn't extend to the opposite end of the outer cylinder. Also, the inner cylinder has a "cap" region with zero-wall thickness (I'm hoping it will be sufficient to prescribe a wall boundary condition for the simulations, as applying a wall thickness resulted in a skewed mesh with cells extending into the no-mesh zone).

My approach has been:
  • Create an initial block around FLUID and then split it at the end of the inner cylinder; associate edges to curves; make the end vertices periodic.
  • Create the first O-Grid for the inner cylinder so that I can then assign that part to SOLID and avoid meshing it.
  • Create a second O-Grid for the large cylinder and associate edges.
  • I then modify the edge parameters to refine the mesh along the walls.
When I check the mesh quality however, the cells at the interface between the two sets of blocks are not good. I tried to merge the side blocks to improve this (I don't want the inner part of the small cylinder to be meshed, so didn't merge the central blocks). Merging the side blocks results in another problem however: the node count on the merged edges no longer matches the node count on the inner edges of the blocks.

Am I going about blocking this geometry all the wrong way? Is there a way to improve the cell quality at the interface of the split?

I hope this post makes sense and that the attached images help.

Thank you in advance for your help,

Kristin

Attachment 11950

Attachment 11951

Kristin_N March 19, 2012 11:20

2 Attachment(s)
A couple more images

Attachment 11952

Attachment 11953

Kristin_N March 19, 2012 11:21

2 Attachment(s)
Last two, I hope they clarify what I tried to describe in my post

Attachment 11954

Attachment 11955

Far March 19, 2012 11:31

whats the problem exactly, the procedure you have adopted seems to be ok . Attach blk and tin

Kristin_N March 19, 2012 12:21

1 Attachment(s)
Thank you for your quick reply Far.

I have attached the blk and tin files to an email sent to you on turboenginner@gmail.com.

The problem is that, due to the node count not matching in the axial direction following the merge of the blocks, a section view of the mesh looks like the image attached.


I look forward to helpful advice,

Kristin

Attachment 11968

Far March 20, 2012 15:28

Done. I shall upload video on youtube.

Some working on geometry was necessary as it helps to control the blocking.

Attaching tin and blk files.

http://img138.imageshack.us/img138/4558/34177561.png

http://img688.imageshack.us/img688/9061/49428069.png

http://img252.imageshack.us/img252/1117/59824151.png

http://img215.imageshack.us/img215/6531/91600559.png

http://img638.imageshack.us/img638/5479/31807988.png

Far March 20, 2012 15:40

Similar problem is also discussed here.
http://www.cfd-online.com/Forums/ans...rt-fluent.html

Far March 21, 2012 00:29

Quote:

The problem is that, due to the node count not matching in the axial direction following the merge of the blocks, a section view of the mesh looks like the image attached.
This problem can be solved with linked bunching, if every thing else is ok.

Far March 27, 2012 00:48

Detailed procedure is given here.

http://www.cfd-online.com/Forums/ans...nner-wall.html

Kristin_N March 28, 2012 17:50

Far,

Thank you very much for all your help, my mesh finally works!

Kindest regards,

Kristin


All times are GMT -4. The time now is 13:22.