|
[Sponsors] |
May 26, 2011, 18:01 |
Geometry Tolerance and meshing
|
#1 |
New Member
Claudio
Join Date: Feb 2011
Posts: 20
Rep Power: 15 |
Hello there.
I worked for the past 2 months on the meshing of a very complex centrifugal compressor. After many days of frustration with Icem I managed to have a clean and smooth structured mesh of the geometry. However, whenever I try to refine it, I have negative determinants everywhere . I am trying to do a CFD simulation of a high pressure compressor with supercritical CO2. The Re is very high and the estimated cell size in order to have a reasonable y+ near the wall is between 1 and 0.1 microns . I tried to repair the geometry several times with icem, but even if I set the model tolerance and triangulation tolerance to 10^-07 I still can not fix it. Moreover, since Icem lacks of proper methods to define periodic planes, I had to manually crate all the periodic boundaries. Needless to say that all the geometric entities I created share the same tolerance of the geometry. Is there anybody who can help me? Does anybody have any experience with this kind of issue? thanks |
|
June 6, 2011, 08:34 |
|
#2 |
Member
jeevan kumar
Join Date: Mar 2009
Posts: 88
Rep Power: 17 |
If you can upload the tetin and the blocking file then i will have a look and can suggest you.
regards jeevan |
|
June 6, 2011, 10:33 |
|
#3 |
New Member
Join Date: Jun 2009
Posts: 11
Rep Power: 17 |
why don't you use turbogrid to mesh your compressor?
|
|
June 6, 2011, 11:03 |
|
#4 |
New Member
Claudio
Join Date: Feb 2011
Posts: 20
Rep Power: 15 |
Hello there!!!
Thank you very much for the reply. Unfortunately I am not allowed to disclose any of the geometry file as they belong to the industrial sponsor . Very sad, I know. Is there any way I can provide some information without giving up rights on the geometry file? In particular, I was trying to generate planes and lines in Icem. Even if I specify very small tolerance I can still see that lines and surfaces are not perfectly matching. Even if I keep tolerance to 10e-07 I can't refine the mesh to 10e-06 without having cells falling into the gaps. I actually thought about Turbogrid... but I am not sure how I can generate the initial file to be imported: The geometry I am working on is fairly complicated with seal passages and leakage flows. I have a STEP and parasolid file of the whole geometry. How do I generate the initial file to be imported in Turbogrid? Thanks a lot for your help!!! |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Importing CAD geometry and Editing edges to match | Ahmed | CFX | 15 | March 21, 2016 13:39 |
Problem Importing Geometry ProE to CFX | fatb0y | CFX | 3 | January 14, 2012 20:42 |
stitchMesh for uncongruent patches (stitch tolerance) | beugold | OpenFOAM | 0 | June 18, 2009 08:38 |
ICEM: Holes in Geometry | Daniel | CFX | 4 | February 22, 2009 10:40 |
GAMBIT - warn about geometric tolerance | CFDNewbie | FLUENT | 2 | July 7, 2007 10:48 |