File Format Definitions
-
I'm starting to develop a new FE-Solver and would like to use Salome for Pre-/PostProcessing. So I need the definition (Reference) for MED and/or UNV files. Where can I get them?-
Hi!
Who can help to get med-2.2.2 library sources? I found a lot of version above 2.2.2 in french only..
-
You can find a documentation about med file at http://www.code-aster.org/outils/med/
But this documentation is in french only...
-
Thanks a lot, but as I don't understand french i can't use that documentation. Is there any german or english documentation of the file structure? What is the unv-format, is it I-DEAS-Universal?
-
-
Hi everybody
I am starting writing C-code in order to read and write MED files. In fact, I am interesting in reading "Champs" entities. According to Y. Fricaud, it makes sense to use med-2.2.2 version directly.
For that purpose, I mimic the examples given in test10.c (in med-2.2.2/share/doc/tests) or in test27.c.
I am surprised to see that the files (*.MED) directly created by this code is not readable by SALOME (that I launch using the command SALOME/KERNEL-2.2.8/bin/salome/runSalome).
For example, if I consider the test27 which contains examples of Cartesian or Standard Grids, when I try to read them using the MED panel of SALOME (using MED Import facility), I get the following error:
The file is incorrect, some information will be missing
In the Object Browser, the three components are creates (names: grille_cartesienne, grille_standard and maillage_vide) but they seem empty (this is what Show Advanced Info procedure tells me). So It does not seem for me to be good thing to copy this example.
If I try the Post-Pro panel of SALOME and use "Import rom File" option, I get the error:
There is an error in imported file
and this time, the Object Broser remains empty.
Any idea what is going wrong there.
Thanks in advance for your help.
Didier
-
Hello,
I had the same problem. I never succeed in importing a MED file containing some 'grille' af any type. So I used 'non_strucuture' mesh to represent a grid and it works fine.
Christophe
RENARD Didier wrote:
Hi everybody
I am starting writing C-code in order to read and write MED files. In fact, I am interesting in reading "Champs" entities. According to Y. Fricaud, it makes sense to use med-2.2.2 version directly.
For that purpose, I mimic the examples given in test10.c (in med-2.2.2/share/doc/tests) or in test27.c.
I am surprised to see that the files (*.MED) directly created by this code is not readable by SALOME (that I launch using the command SALOME/KERNEL-2.2.8/bin/salome/runSalome).
For example, if I consider the test27 which contains examples of Cartesian or Standard Grids, when I try to read them using the MED panel of SALOME (using MED Import facility), I get the following error:
The file is incorrect, some information will be missing
In the Object Browser, the three components are creates (names: grille_cartesienne, grille_standard and maillage_vide) but they seem empty (this is what Show Advanced Info procedure tells me). So It does not seem for me to be good thing to copy this example.
If I try the Post-Pro panel of SALOME and use "Import rom File" option, I get the error:
There is an error in imported file
and this time, the Object Broser remains empty.
Any idea what is going wrong there.
Thanks in advance for your help.
Didier
-
suteau christohe wrote:
Hi Christophe
Hello,
I had the same problem. I never succeed in importing a MED file containing some 'grille' af any type. So I used 'non_strucuture' mesh to represent a grid and it works fine.
Christophe
Thank you for your answer. But I am puzzled.
In fact, I am more interested in setting variables in a 'Champ' that must be realted (if I understood correctly) to a maillage. Then I found an interesting and easy way to proceed by creating this "regular grid" and letting MED do the work to consider it as a 'maillage' (The Cartesian or standard grid seemed to me by far to be the easiest).
Now, according to your advise, I would have to change it into a Non-Structured 'Maillage' where I need to specify each single grid node and probably their mutual connectivity relationship. If I am correct, this will be, by far, more demanding.
Am I correct ?
-
Hello Users and Developers,
I do also have a question concerning the unv/med file formats. Regarding the
exported unv/med-files from a "meshed" geometry, I miss some information which would
be quite a pleasure to have. These are, in especial, information regarding the boundary
entities of a grid.
For instance, nodes lying on the boundary of the geometry could be separated from inner nodes
(if inner nodes exist). it would be also great to have the faces lying on a surface seperated and enumerated. Then,
one could provide information which surfaces the boundary node is related to.
There is probably the possibility to export such information yet... Does anybody have any hints for me?
I tried to study the med format doucmentation, but my knowledge of french is rather rudimental =)
Thanks in advance, best regards
Dan-
Hi,
If anyone is interested, you can download MED format manual in English and a short tutorial on how to use MED from my home page. Here is a link to the archive that contains it all:
http://www-tm.wbmt.tudelft.nl/~andrico/tutorial.tgz
Regards,
Andriy
P.S. I received this tutorial thanks to Nicolas Crouzet (thanx Nicolas)
-
Dear Andriy,
let me say "thank you so much" to you for announcing this link to the english pdf-tutorials -- which are really
a great help =)
I'm sure they will be very useful for other salome-users as well.
Best regards,
Dan
-
Dear Andriy
I had the same issue. Trying to develop my own solver based on Salome .med mesh data.
I'd almost lost hope when I saw your post. Thank you Thank you Thank you-
Hi again
I am very new to Salome. I am developing my own FEM engine and intend to use .med format
for importing mesh data from Salome.
I just wonder how to associated attributes like material properties to the model and beyond that
how does this information pass to the output (.med files) so that It can be read into c/c++ for example.
cheers
-
DAVOOD ANSARI OGHOL BEIG wrote:
Hi again
I am very new to Salome. I am developing my own FEM engine and intend to use .med format
for importing mesh data from Salome.
I just wonder how to associated attributes like material properties to the model and beyond that
how does this information pass to the output (.med files) so that It can be read into c/c++ for example.
cheers
Hi Davood,
In order to prescribe material properties to the elements I use groups. In that manual you can find how to handle groups. Hence, in Salome you define different groups and then in your FE program you can determine which group each element belongs to and
prescribe corresponding properties.
Regards,
Andriy
-
Hi
Thank you for the reply.
This is indeed a good idea and mostly working.
I am not sure, yet, when you have a material property or physical load-field as a continuously changing field
over the domain then you'd find it rather straight forward to have tools like the "DATA module" used.
The only problem is that this DATA module is very weakly supported. I have had no luck getting it compiled.
You can find info about it by searching "DATA module" in this forum. Probably you may know more than I do.
Kindly let me know
Davood-
Hi Davood,
If you search the forum you'll find a post where it says that OpenCascade does not support DATA module. It was developed as an academic project that had finished some time ago. Instead they have a development module XDATA (Python based), that allows you to do comparable (or, may be more powerful stuff). But I've never used it.
Although those two modules (DATA, XDATA) provide an elegant way to work with boundary conditions and physical properties, essentially they probably only work on MED object and work with groups. Hence, if you do the same with groups you can achive the same, but the actual mat. properties and B.C's have to be read from somewhere else.
Regards,
Andriy-
Hi andreykiv
I should really thank you for this clarification.
I was completely lost in trying to get the old DATA module working.
I burt in the "DATA-module" fever for a week.
Based on what you say, I would rather use the group-base method.
Many people may get similarly confused. I have searched the entire web about the DATA module (including the forum)
and did not see any thing saying that DATA is no longer being supported. If you are right (I believe you are) then
the page http://www.salome-platform.org/home/presentation/data/ provided a very misleading point of view
and should somehow be updated.
Deeply appreciated.
-
Hi andreykiv
I should really thank you for this clarification.
I was completely lost in trying to get the old DATA module working.
I burnt in the "DATA-module" fever for a week.
Based on what you say, I would rather use the group-base method.
Many people may get similarly confused. I have searched the entire web about the DATA module (including the forum)
and did not see any thing saying that DATA is no longer being supported. If you are right (I believe you are) then
the page http://www.salome-platform.org/home/presentation/data/ provided a very misleading point of view
and should somehow be updated.
Deeply appreciated.
-
Hi alll
I seem not to be able to use the group based method quite well.
Does anyone have some examples ready there?
Let's say we have the boundary of a 2D object and we want to make it into groups so that the
boundary can be listed in the mesh as line segments (assuming triangular elements).
How should we specify the groups in Salome so that they can be retrieved in MED (c/c++).
Many thanks
Davood
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Hello,
Versions before 2.2.2 were in French also. In addition, med version 3.0 still has a French documentation, but the function names are now in English, so med 3 seems really preferable to med 2.3.
Though it is not as practical as having an up to date documentation, looking at the source code of the MED_SRC file drivers in SALOME may be a good start.
Regards,
-
