[Gmsh] unv format difference.
Christophe Geuzaine
cgeuzaine at ulg.ac.be
Fri Nov 14 17:31:25 CET 2008
Roland Tollenaar wrote:
> Hi Christophe,
>
> I've listed on the mailing list but while this thread is up still this
> offline reply still.
>
>> Can you check in the UNV file and make sure that you have sections
>> with type "2477" (with label "PERMANENT GROUP")? If you do, maybe we
>> chose a bad group type in the UNV spec? Which one should we use?
> 2467 by the look of things.
>
> This is a fragment out of a unv file from gmesh:
>
> (2412).............
> .........
> 2953 2432 2959 3896
> 26405 111 1 0 7 4
> 2970 2953 2959 3896
> -1
> -1
> 2477
> 1 0 0 0 0 0 0
> 255
> PERMANENT GROUP1
> 7 211 0 0 7 67 0
> 0
> 7 66 0 0 7 221 0
> 0
> etc
>
>
> Whereas a unv that is saved with groups from Salome (and is convertible
> with unv2abaqus) looks something like this:
>
> (2412).............
> .........
> 2953 2432 2959 3896
> 26405 111 1 0 7 4
> 2970 2953 2959 3896
> -1
> -1
> 2467
> 0 0 0 0 0 0 0
> 7
> support1
> 7 211 0 0 7 67 0
> 0
> 7 66 0 0 7 221 0
> 0
> etc
>
>
> I.e. as a layman I would say there are two differences. The 2477 is 2467
> and the separation line between the groups (the line directly following
> the type number) starts with zero intead of 1. But after that counts up
> in the same manner.
>
>> If I remember correctly we actually change "e" or "E" to "D" on
>> purpose for the UNV format, as requested by the UNV spec...
> Ouch, this would be a pity. Anyhow Salome gives out "e". Maybe Salome is
> taking liberties with the .unv format in which case you must check the
> correct type number. Maybe Gmsh uses the correct number. This would be a
> bit of a disaster, the "e" or "d" can be sorted out automatically with a
> script, but replacing the 2477 with 2467 with sed is rather dangerous.
I found this, which seems to corroborate your findings: Salome seems to
be using a deprecated tag (2467):
http://www.salome-platform.org/forum/?groupid=10&forumid=10&thread=1205
>
> Please let me know whether I must hack about or whether there might be
> changes coming to Gmsh in this regard?
>
> Thanks
>
> Regards,
>
> Roland.
>
>
>>
>>
>>
>>>
>>> Kind regards,
>>>
>>> Roland Tollenaar.
>>>
>>>
>>>
>>>
>>>>
>>>>
>>>>
>>>>> Best regards,
>>>>>
>>>>> Roland
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>>> I would be highly obliged if you could briefly point out what
>>>>>>> approach I have to take here.
>>>>>>>
>>>>>>> Kind regards,
>>>>>>>
>>>>>>> Roland Tollenaar
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>
>
--
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science
http://www.montefiore.ulg.ac.be/~geuzaine