|
[Sponsors] |
November 22, 2013, 11:39 |
Mesh Generating
|
#1 |
Member
Join Date: Nov 2013
Posts: 43
Rep Power: 12 |
Hello,
while meshing Quad/Pave Gambit gives out an error saying that "exact evaluation failed, try using facet evaluation" the mesh will be created but also the geometry seems to change, which is odd. As you can see in the picture the green line (straight) is the original geometry and after applying the mesh, the yellow line (curve) is created which leads to elements with really high skewness. Is there a default setting or something like that which could stop Gambit from auto-changing the mesh like this? Are there any angle etc limitations for quad-pave-meshing of the face/edges which I might have violated? Thx and regards, Persil |
|
November 25, 2013, 01:50 |
|
#2 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
try to modify the topology overthere. For example split the surface where it degenerates.
It may help
__________________
In memory of my friend Hervé: CFD engineer & freerider |
|
November 25, 2013, 04:34 |
|
#3 |
Member
Join Date: Nov 2013
Posts: 43
Rep Power: 12 |
okay, I did that and now I can mesh the 2 created surfaces separately but it is not a satisfying result. but thx for the reply
|
|
November 26, 2013, 09:12 |
new question
|
#4 |
Member
Join Date: Nov 2013
Posts: 43
Rep Power: 12 |
hey,
when meshlinking the hub and tip of a turbine blade (faces) I get elements with really hight skewness when using hex/cooper as 3D mesh. That is caused by the twist in the blade itself as you can see in the picture. Does anybody know a way how to make these 2 face meshes look more alike? I tried meshing both separetly but this causes cooper to disfunction... |
|
November 26, 2013, 09:34 |
|
#5 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
Do they belong to the same volume? I mean, is the top (poor quality) surfaces the projection of the bottom one?
If yes try to split the volume and in this area. Maybe you should decompose leading edge area like C-grid, or something like that
__________________
In memory of my friend Hervé: CFD engineer & freerider |
|
November 26, 2013, 09:56 |
|
#6 |
Member
Join Date: Nov 2013
Posts: 43
Rep Power: 12 |
yes, the top face is the projection of the one below. I meshed the bottom 1 and mesh linked it to the other.
Do you mean splitting the volume in half so I would get another face in between the to faces in the picture? This way I would split the whole model in half which sounds a bit odd. Or where does "splitting in that area" refer to? I am pretty new to Gambit so I never did C-Grid before. I imagine it as half circles around the leading/trailing edge. But when reaching the kind of rectangular border there will occur problems, I guess... |
|
November 26, 2013, 10:35 |
|
#7 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
C-grid is just a "way" to decompose your mesh in proximity of the leading edge
here an example for O-grid: http://www.cfd-online.com/Forums/ans...ured-mesh.html If you want you can provide your dbs file, I will check it tomorrow at the office (I am not an aero expert)
__________________
In memory of my friend Hervé: CFD engineer & freerider |
|
November 27, 2013, 09:16 |
|
#8 |
Member
Join Date: Nov 2013
Posts: 43
Rep Power: 12 |
Sorry, I cannot send you the file. But thx for the offer.
I will give it a new try with a better designed igs-file. Is it possible to use 1 journal for different igs-files or do I have to start all over? Because the blade will look exactly the same. i Sot should only be a matter of the amount of vertices, edges and faces right? |
|
November 27, 2013, 09:32 |
|
#9 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
theorically yes.
But as you will import a geometry, it is possible that geometry index won't be regular (understand if you import a circle, then gambit will call it edge_5 or whatever. And on next import it will call it edge_10 --> and your journal file will fail) But you can try
__________________
In memory of my friend Hervé: CFD engineer & freerider |
|
November 28, 2013, 03:27 |
|
#10 |
Member
Join Date: Nov 2013
Posts: 43
Rep Power: 12 |
yes, that is exactly what I consider as the problem. but at least for 1 file the indices of edges and so on should always be the same, right?
it only differs if you compare different files of the same geometry with each other |
|
November 28, 2013, 03:44 |
|
#11 |
Super Moderator
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,297
Rep Power: 41 |
I am not familiar with journal, but I think it is more appropriated with geometries which are built in Gambit (and you decide which name you give to topology).
With import you have the problem mentionned above. Regarding your question, just try to generate a simple journal with one split with volume, on with surface and one with edge. Then try to call this journal with another (different) import file. And I would also add, that igs import aren't always with good quality, and you need to clean the geomety with "clean Up" tool, whichs may generates some virtual entities with another random index-assignment
__________________
In memory of my friend Hervé: CFD engineer & freerider |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Moving mesh | Niklas Wikstrom (Wikstrom) | OpenFOAM Running, Solving & CFD | 122 | June 15, 2014 06:20 |
[Gmsh] 2D Mesh Generation Tutorial for GMSH | aeroslacker | OpenFOAM Meshing & Mesh Conversion | 12 | January 19, 2012 03:52 |
Doxygen documentation | Tanay | OpenFOAM Installation | 9 | September 23, 2011 11:40 |
ParaView and Qt 4.3.5 on Mac OS X 10.6 | Adrian | OpenFOAM | 3 | August 8, 2010 03:16 |
Icemcfd 11: Loss of mesh from surface mesh option? | Joe | CFX | 2 | March 26, 2007 18:10 |