CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (https://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [GAMBIT] most efficient mesh for a combustor with holes (https://www.cfd-online.com/Forums/ansys-meshing/82086-most-efficient-mesh-combustor-holes.html)

perdita November 16, 2010 10:07

most efficient mesh for a combustor with holes
 
1 Attachment(s)
Hi everyone,

So, I'm trying to model a combustor with dilution holes (see attached picture). I'm continuing the work someone else began, and that geometry did not have the holes in it, and therefore had been meshed using hex/cooper method (thats what the mesh on the two front volumes is.)

However once I add the holes to the model, I keep getting errors if I try to use hex or hex/wedge and cooper to mesh it. [Face [x] of volume [x] is not appropriate for use as a face to project along. Either it is not submapped, or the choice of source faces is incorrect]. (I have tried pre-meshing the faces, nothing seems to work. It also then gives the same error for one of the faces at the end of the volume, which I find a bit strange.)

So my question is - what would be the most efficient meshing scheme to use for the volume, given that the available computational power is not very high. At the moment, I have only been able to get a mesh using Tet/Hybrid and TGrid as my options. And as I want to have the mesh quality reasonably fine, Im worried that this will end up taking too long when I eventually run the simulation in FLUENT. (It takes quite some time just making the mesh itself.)

Does anyone have any advice for me?

Thanks!

-mAx- November 17, 2010 01:04

Your geometry is decomposed so that you will try to get an hexa-mesh based on 2d-quad-map. That's why you have a square section on the middle.
You can handle your geometry with pave instead of quad (your mesh will be still hexa, but cooper)
Keep all splits along x direction (3 splits), and merge all subvolumes in each x-split.
You have now 3 cylinders which can be meshed with cooper (just mesh one souirce (cap) with quad-pave).
YOur problem with holes can be fixed by splitting twice the left volume along hole's axis.
Check tutorial 7 in Gambit Modeling Flow In a Tank (there is one split similar as yours)

perdita November 17, 2010 04:48

Quote:

Originally Posted by -mAx- (Post 283717)
Your geometry is decomposed so that you will try to get an hexa-mesh based on 2d-quad-map. That's why you have a square section on the middle.
You can handle your geometry with pave instead of quad (your mesh will be still hexa, but cooper)
Keep all splits along x direction (3 splits), and merge all subvolumes in each x-split.
You have now 3 cylinders which can be meshed with cooper (just mesh one souirce (cap) with quad-pave).
YOur problem with holes can be fixed by splitting twice the left volume along hole's axis.
Check tutorial 7 in Gambit Modeling Flow In a Tank (there is one split similar as yours)

i'll give this a shot, thanks! and thanks for the link of the tutorial - it may come in handy


All times are GMT -4. The time now is 11:08.