[Gmsh] Transfinite line algorithm

David Colignon David.Colignon at ulg.ac.be
Thu Jul 30 11:12:45 CEST 2009



Martin Vymazal wrote:
> Hello Dave,
> 
>  thank you very much for your reply. I had an idea how 'progression' 
> works, the problem is that this is not what I see in my mesh when I 
> visualize the nodes on the wall of the airfoil. 

can you send us an example (.geo plus picture or an indication on where to zoom) ?


As I wrote in my
> previous email, making progression slightly bigger than 1.0 (say, 1.3 or 
> bigger), gmsh gives me a completely incorrect mesh.


the spacing between the 1D meshing nodes obeying to a geometric law, it is very sensible to the progression parameter, 
and 1.3 is a rather high value


Sometimes it even
> tries to mesh the inside of the airfoil.
>  I also found the question/reply about 'bump' in gmsh mailing list 
> archives to which you pointed me, but it does not help me to understand 
> whether I'm doing something wrong or if gmsh simply cannot mesh the 
> airfoil the way I want. My only question is simply: can I control the 
> number and distribution of the mesh points on the airfoil in the geo 
> file I attached yesterday? If so, how?

With the transfinite command, you can precisely control the number (and spacing) of the _1D_ meshing nodes on the 
boundary  (splines) of your airfoil.

But the problem is that you mix the 1D transfinite mesh with (various) Fields for the 2D meshing, and if there is a huge 
discrepancy between the spacing of the 1D nodes and the background 2D field of characteristic lengths, the 2D meshing 
algorithm sometimes do not succeed to match the conflicting constraints, and you sometimes get these strange results.

So Gmsh is perhaps able to mesh the airfoil the way you want, but it is a little bit "tricky" and need a lot of trials 
and errors...

Good luck !

Dave




>  Once again, thanks for any help.
> 
>  Best,
> 
>    Martin Vymazal
> 
> 
> 
> Cituji David Colignon <David.Colignon at ulg.ac.be>:
> 
>>
>> Martin Vymazal wrote:
>>> Dear gmsh users,
>>>
>>> I'm trying to generate a second order (P2) unstructured mesh around a 
>>> symmetric airfoil. I managed to create a mesh which is fine so far, 
>>> but I would like to control precisely the number of points on the 
>>> airfoil. To make things a bit more complicated, I'd also like to have 
>>> more points close to the leading and trailing edges of the airfoil. I 
>>> thought I could use the transfinite algorithm with 'progression' or 
>>> 'bump' options, but the progression gives me a bad mesh for a 
>>> progression ratio which is not close to 1.0 (I attached my *.geo 
>>> files, progression 1.3 at the end of the naca0012 geo file already 
>>> gives strange results.
>>
>> Hi Martin,
>>
>> I think displaying the nodes of the 1D mesh on the lines with 
>> "shift-alt-p" or Tools->Options->Mesh->Visibility->Nodes could help 
>> you understanding what happens, the way "bump" and "progression" works 
>> in Gmsh and what you call "strange" results (I see what you mean).
>>
>> http://geuz.org/gmsh/doc/texinfo/gmsh-full.html#SEC45
>>
>> The optional argument `Using Progression expression' instructs the 
>> transfinite algorithm to distribute the nodes following a geometric 
>> progression (Progression 2 meaning for example that each line element 
>> in the series will be twice as long as the preceding one). The 
>> optional argument `Using Bump expression' instructs the transfinite 
>> algorithm to distribute the nodes with a refinement at both ends of 
>> the line.
>>
>> and if you search for "bump" in the mailing-list archives, you get this:
>>
>> http://www.geuz.org/pipermail/gmsh/2008/003526.html
>>
>>
>> btw , point 5000 is exactly the same as point 381. It is perhaps not a 
>> problem but you should avoid it.
>>
>> Cheers,
>>
>> Dave
>>
>> -- 
>> David Colignon, Ph.D.
>> Collaborateur Logistique du F.R.S.-FNRS
>> CÉCI - Consortium des Équipements de Calcul Intensif
>> ACE - Applied & Computational Electromagnetics
>> Institut Montefiore B28
>> Université de Liège
>> 4000 Liège - BELGIQUE
>> Tél: +32 (0)4 366 37 32
>> Fax: +32 (0)4 366 29 10
>> WWW:    http://hpc.montefiore.ulg.ac.be/
>> Agenda: 
>> http://www.google.com/calendar/embed?src=david.colignon%40gmail.com
>>
>>
>>
>> It seems
>>> that gmsh calculates the lengths of the lines on the wall of the 
>>> airfoil incorrectly (not precisely enough?). I hoped I could somehow 
>>> set the precision with which the spline representing the wall of the 
>>> airfoil is meshed and thus fix the problem, but found nothing 
>>> relevant in the documentation.
>>> On the other hand, the 'bump' option seems to have no effect whatever 
>>> value I use (I tried values between 0.5 and 3.5, but I must admit I 
>>> don't know what this number actually means).
>>>
>>> I'd appreciate any help/hints or at least an info whether this can be 
>>> done in gmsh or not.
>>>
>>> Thank you very much,
>>>
>>>        Martin Vymazal
>>>
>>> ----------------------------------------------------------------
>>> This message was sent using IMP, the Internet Messaging Program.
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> gmsh mailing list
>>> gmsh at geuz.org
>>> http://www.geuz.org/mailman/listinfo/gmsh
>>
>>
> 
> 
> 
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.