CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   meshing an annulus (http://www.cfd-online.com/Forums/fluent/28361-meshing-annulus.html)

atholl March 27, 2001 05:08

meshing an annulus
 
Can anyone suggest any strategies for meshing a concentric annulus in GAMBIT? I've been playing about with edge meshing the boundary layers up until now and was wondering if there are any better methods.

Anthony Wachs March 27, 2001 07:51

Re: meshing an annulus
 
I am currently working on annular geometries including concentric ones. My strategy is to cut the annular in two equal volumes (two half annulus), following the axial direction and to mesh similarly those two parts. With this solution, you are able to prescribe the mesh refinement in the radial direction, as also in the axial direction. Finally, when I have meshed five of the six faces of each volume (the remaining unmeshed face is one of the two annular sections of the pipe), I use the Cooper scheme to project the mesh of the annular section in the whole volume.

If you'd like to see an example, do not hesitate to contact me on my personal email

I hope that will help you Sony

Ashutosh March 27, 2001 10:37

Re: meshing an annulus
 
The meshing strategy will depend on the type of flow problem you are solving. If you are expecting BL on the annulus walls then you may have to refine the mesh near the wall by applying Boundary layer. Then you can mesh using Cooper meshing scheme taking the two concenteric faces( radial faces) as source faces. This way you don't need to divide the volume into two as done by Anthony. Are you solving 3D or 2D problem? With Cooper you can get the good quality mesh.

Bye!

Ashutosh

atholl March 27, 2001 11:50

Re: meshing an annulus
 
Thanks for the quick replies! I'm modelling in 3-D and have tried applying a boundary layer at the walls, but cannot seem to get good aspect ratio on the cells without having an excessively high interval count on the edges, which as a result takes a very long time to cooper the volume. Is there any way around this?


All times are GMT -4. The time now is 15:50.