[Gmsh] For help on triangle vertices order

Umut Tabak u.tabak at tudelft.nl
Wed Aug 5 18:22:49 CEST 2009


zqyork wrote:
> Hi Tabak;
>   
Hi, still I am not clear on what your problem is, also CC your mails to 
the list as well, not only to me ;-)
>   Thank you so much.  My problem is when I try to load a .msh file. In 
>
> $Elements and $EndElements the order of vertices is prescribed for each
> triangle.
>   
This is the order of the '1st-order' elements

2
| \
|  \
|   \
0__\1

The triangle is a bit tilted but does no harm :) so you should see this 
reference ordering on the elements in the msh file...
> I am not very sure whether the order is of great importance. But during my
> test,
>
> the order should be important. That means 1,2,3 of the triangle vertices is
> different
>
> with 1, 3, 2 because of the orientation. 
If you do an FE computation, surely, the order does matter because it 
affects the shape function and related computations, inverse of Jacobian 
and so on.
> Now my mesh vertices order is
> arbitrary thus
>
> the mesh looks not right. 
>   
You still did not provide enough info. I did create a msh file for you 
and wanted to show ordering on that,
    ...
   25 2 3 0 5 0 10 2 11
   26 2 3 0 5 0 10 11 12
   27 2 3 0 5 0 10 12 9
   ..
These are elements 25, 26 see their vertices, [10 2 11] and [10 11 12] 
and look at the picture attached. They are right and ordered in the 
reference element ordering given in the reference manual, if you can not 
get these right in your program then there is an implementation problem, 
I would suggest the Matlab reader option as a quick check.

Hope this helps a bit.

Umut


  
-------------- next part --------------
A non-text attachment was scrubbed...
Name: toList.png
Type: image/png
Size: 13369 bytes
Desc: not available
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20090805/132897fc/attachment.png>