|
[Sponsors] |
April 21, 2009, 11:50 |
Fixing a poor quality mesh
|
#1 |
Member
Join Date: Mar 2009
Location: Texas
Posts: 33
Rep Power: 17 |
In a portion of my geometry, I have a volume that is not meshing very well. I am able to generate and "o-grid" (Blue mesh in the pictures) easily with the hex/map volume mesh. I want to use prismatic cells to fill the rest of the domain (the red mesh) using the cooper scheme. I have been somewhat successful getting this to work. Right now, the red mesh has a an equiskew size of 0.91. If I change the edge meshing straight piece of the source face even a little, this number goes up and sometimes I even get inverted cells. What is the best way to reduce my equiskew size for this type of mesh? Thanks in advance.
|
|
April 21, 2009, 14:58 |
|
#2 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
Why don't you try to fill it wit tet-hexcore?
Gambit will generate the outer layer with tetra and prism and the core with hexa. If you still want prism, then I would suggest splitting the volume along its axis (y). The problem here, is that Gambit has to project an half disk onto an half "rectangle". You can help gambit, if the projection is generated in more than one step |
|
July 15, 2010, 19:45 |
in order to have a better quality of your mesh
|
#3 |
Member
Domenico
Join Date: Sep 2009
Location: Cranfield
Posts: 48
Rep Power: 16 |
Hi man, I got a problem like yours, and then I started to think to divide my domain in several blocks. At the end of the day I achieved a good results by generating several blocks and the mesh is structured everywhere. First of all try to create the central part of your mesh at squared base instead of circular.
|
|
December 15, 2010, 12:48 |
|
#4 |
New Member
roger
Join Date: Oct 2010
Posts: 7
Rep Power: 15 |
What do you mean with inverted cells? I have a problem when doing the Grid in Fluent that says I have inverted elements and non-positive volumes. I'm trying to mesh an airfoil.
|
|
December 16, 2010, 10:44 |
|
#5 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
Go to examine Mesh and check where are those cells (choose Volume in quality type and set lower and upper limit for displaying only negative cells).
Most of time, you will understand the problem as you see where are bad cells
__________________
In memory of my friend Hervé: CFD engineer & freerider |
|
December 18, 2010, 13:44 |
|
#6 |
New Member
Sowmya K
Join Date: Nov 2009
Posts: 22
Rep Power: 16 |
Forrest, if this is some of a circular-to-rectangular transition duct, then it might be worth while for you to look at a Pointwise/Gridgen tutorial example, where creating a grid for the same geometry has been shown very clearly.
I struggled with very poor quality meshes for a transition duct as well (it is for my masters thesis). Why don't you use a 'Pave' hex unstructured mesh in the interior instead of a tet mesh? I did find it that it helps in reducing the skewness. |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
virtual geometry | richard | FLUENT | 4 | January 7, 2009 06:16 |
how to extend FSI 2D codes to 3D, need advises | abouziar | Main CFD Forum | 1 | May 30, 2008 05:08 |
basic of mesh refinement | arya | CFX | 4 | June 19, 2007 13:21 |
How to do Geometry from 3D mesh in ICEM | San24 | Main CFD Forum | 4 | August 8, 2005 03:49 |
General questions on grid-based computing | Adrin Gharakhani | Main CFD Forum | 21 | June 5, 2000 14:47 |