Radiative transfers P1-model

[Jonathan Gerardin]

Hello,
 
I try the radiative transfers, it works well for the DOM, but I can't make the calculation with the P1-model.
 
I've got this error :
 
@                                                            @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @                                                            @ @@ ATTENTION : RAYONNEMENT : ARRET A L'ENTREE DES DONNEES @    =========                                               @    NOMBRE DE DIRECTIONS NON ADMISSIBLE                     @                                                            @  Le nombre de directions doit etre 32 ou 128 (NDIREC)      @    Il vaut ici NDIREC =         -1 @                                                            @  Le calcul ne sera pas execute.                            @                                                            @  Verifier usray1.                                          @                                                            @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @  
 
In the P1-model, we don't need  ndirec (which is the number of direction of the DOM). What can I do to avoid this error?
 
Thank you.
 
Jonathan GERARDIN

[Alexandre Douce]

Do you use the GUI or the usray1.f90 for your data input ?

[Jonathan Gerardin]

I use the GUI

[Alexandre Douce]

Could you attach the xml file and the mesh ?

[Jonathan Gerardin]

k

[Alexandre Douce]

There is a little bug on the verification of the user data input: the number of the directions for the angular discretization is only usefull for the DO model, but the code check this value also for the P-1 model.  To avoid this trouble please copy this file rayopt.f90 in your SRC directory of all your cases.

Could you please test this patch and tell me if is it OK ?

Thanks Jonathan for your feedback.
 
PS: warning the forum change the name of the attached file, don't forget to rename the file to rayopt.f90

[Jonathan Gerardin]

Thank you Alexandre, but I've got a new error now :
 

At line 422 of file /home/jonathan/Telechargements/install-200-beta2/ncs-2.0.0-beta2/src/base/codits.f90
Fortran runtime error: Array reference out of bounds for array 'nomvar', lower bound of dimension 1 exceeded (0 < 1)
Error running the calculation.
 
Check Kernel log (listing) and error* files for details

 
The listing seems to be incomplete (I join it)

[Alexandre Douce]

I will test your case and try to reproduce  the trouble. Do you have sources in SRC, have you begin your development?
Your files of parameters test1 and test2,  are they different?

[Alexandre Douce]

Jonathan, I ran your data input test1 with your mesh with Code_Saturne 2.0-rc1. It does not crash! Therefore I believe that you have some sources in SRC. If I am right, could you please open a new thread in the development forum with your sources attached? If not, you should use the 2.0-rc1 instead of the 2.0-beta2.

[Jonathan Gerardin]

I haven't got source in SRC, I wanted to try the P1 before starting the development. I can't download the 2.0-rc1. (on my laptop, it's maybe the 2.0-rc1 version of Saturne??)