IMPR_TABLE and IMPR_RESU with DEFI_FICHIER failing to provide saved files

Mishael

I am working on an extension of verification study FORMA06 (v3.02.112) in Code Aster reference. I am performing growth of the multiple cracks in a 2D plate under tension. I am trying to extract my results, and I am facing some difficulties.

I am using IMPR_TABLE together with DEFI_FICHIER command to save Stress intensity factor (SIF) values into tables. Unfortunately, it is not saving tables into external files. I can see that the file is created in message file.
I am also facing issues in saving the MED files using the IMPR_RESU and DEFI_FICHIER command for the visualization mesh after the crack growth. Unfortunately, during my simulation some warnings are issued and ended without generating the MED files requested.
I also tried to run a verification case SSLV319D directly from the library, in which the same commands are used to save MED file. The verification case also failed to save the MED file to external file. How can I resolve this issue? I also referred the Beginning with Code_Aster book.

Please see the comm and message files for the cases mentioned. I am also uploading the initial MED file (for my case study) if someone wants to test it. For SSLV319D, I used the same MED file associated with it from the library.
I am running my simulations on Ubuntu (installed in WSL2) with the latest version of SALOME MECA and Code Aster.

Mishael

I found a statement in the explanation of DEFI_FICHIER command description. I think this is cruicial in defining the output files. The description is as follows: "The user must declare the physical names of the files and the associated logical units. This declaration is made in the astk interface prior to launching the execution of the job. The files must be added to the study profile by selecting the libr type and associating them with the chosen logical unit number".

Once I added the names of all files together with logical units under 'Data Files' for my case study (I think this can also be added via astk interface), the script executed successfully and saved all my expected results. I am attaching herewith the mesh file, command file, and export file for someone who wants to test it. I didn't have to specify the DEFI_FICHIER command in my command file.

Now, my doubt is why we should use the "DEFI_FICHIER" command if we need to externally define the required files and their logical units.

hberro

Mishael

code_aster commands always take a logical unit for i/o files handling (fortran)... For example LIRE_MAILLAGE UNITE=20

DEFI_FICHIER allows to "dynamically" associate files or paths on your disk to logical units that can be used in the code_aster commands... Doing so makes it possible to use a minimal export file where only the .comm file is referenced (UNIT 1), or even to start a run_aster python session and then interactively work your way around your calculation command by command and whenever a file is needed as input or output, to create it through DEFI_FICHIER.

I hope this makes it clearer for your use case.

Mishael

hberro

I understood the concept with DEFI_FICHIER now. Thank you for the clarification.

But, I am still confused about the application of it via comm file. For example, if I am not defining the files beforehand either in astk or export file, then I always see a message in my 'message' file as "Deleting 'file_name.med': No such file or directory". Then the IMPR_RESU file failed to save the data to the file created through the DEFI_FICHIER command. My problem is that I can't define the name of the file beforehand (inside export file) as the file name is created inside a 'for' loop and the upper limit of the 'for' loop is calculated based on the input variables which are varying. I also tested on one of the validation cases SSLV319d by simply running the simulation without altering any code. Similar message is printed inside 'message' file.