open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I ‘m trying to use the vector PRIVE1 in my fortran file to save a new variable in the result file.
In my fortran file, I have copied the subroutines « NOM_VAR_TELEMAC2D » and « PRERES_TELEMAC2D ». In the "cas file" I have selected de variable N and set the variable "number of private arrays" to 1.
In « PRERES_TELEMAC2D » I use the BIEF library (CPSTVC) to call the vector PRIVE1 and I calculate my new variable (TAU).

I have written :
IF((LEO.AND.SORLEO(23)).OR.(IMP.AND.SORIMP(23))) THEN
CALL CPSTVC(ZF,PRIVE1)
DO I=1,NPOIN
CHEZY%R(I)=CHESTR%R(I)* H%R(I)**(1/6)
TAU%R(I)=(1000*9.81*T6%R(I)**2)/(CHEZY%R(I)**2)
PRIVE%ADR(1)%P%R(I)=TAU%R(I)
ENDDO
ENDIF

An error is detected during the compilation :
CALL CPSTVC(ZF,PRIVE1)
1
Error: Type mismatch in argument 'y' at (1); passed REAL(8) to TYPE(bief_obj)
ERREUR A LA COMPILATION

The special declaration in PRERES_TELEMAC2D  subroutine is :
TYPE(BIEF_OBJ), POINTER :: CHEZY, TAU

Thank for your help !!!

christophe

Can you share your PRINCI file with us ? Although you have put a lot of details in your message, I would like to have a look at the whole file.

Thanks.

Hi Christophe,

PRIVE1 is a pointer to a DOUBLE array, specifically
PRIVE%ADR(I)%P%R. This link is set in POINT_TELEMAC2D.

When you call CPSTVC, it should get two arguments that are both BIEF_OBJ objects. Because PRIVE1 is a double array, you got that error message during compilation.

If you replacewiththen you should hopefully have more success.

Hope that helps!
Chris

Hello,

That's it Chris !

I'll add that it is very curious that ZF is seen by the compiler as real*8, because it should there of type(bief_obj).

An other problem is that TAU and CHEZY are not really useful, and should be allocated by a call to BIEF_ALLVEC before using their %R part.

I would just suggest that TAU and CHEZY are declared as double precision, they need not be an array because they are just intermediate variables aren't they ?

With best regards,

Jean-Michel Hervouet

Hi,

I just come back from holidays.

Sebastien, here is it the fortran file.

fortran_T2D_v2.f


Chris with your solution, the compilation is Ok, but these is an error in the first iteration.

The message is :
=======================================================================
ITERATION 0 TEMPS : 0.0000 S
PRERES : NOMBRE DE COURANT MAXIMUM : 0.000000
0.08user 5.54system 0:09.23elapsed 61%CPU (0avgtext+0avgdata 148672maxresident)k
696inputs+0outputs (3major+9676minor)pagefaults 0swaps
## Erreur : Fin anormale : time ./out2276_gFortran.exe :2816
## Error : System command failed for /home/sysadm/sharepoint/Test1_T_E/cas.txt2276_tmp/telemac2d.bat :6400
________________________________________________________
Execution terminee: telemac2d.bat
________________________________________________________
No compilation/linking/file errors detected.
Execution errors detected.
Please see messages in stdout above or study stderr output.

Working directory: /home/sysadm/sharepoint/Test1_T_E/cas.txt2276_tmp
can be manually deleted with: /home/sysadm/sharepoint/Test1_T_E/delete_cas.txt2276

Returning exit status 1

===================================================================

Thank for your help.

with the file...





Thank you !

Hi
In your code, you use T6.
I'm not sure but if in your result file you don't ask for the Scalar Velocity (M in your steering file) I think T6 was not compute. That could generate a crash when you try to use it.
Try to replace T6 by the Velocity.

Hope this helps

Nice to see you back !

Just a thought -- did you add the keyword NUMBER OF PRIVATE ARRAY = 1 in your CAS file ?

Sébastien.

yes I did !!

I agree with Jean-Michel - the crash looks to be coming from trying to use the unallocated double arrays TAU%R and CHEZY%R. I also don't see why you need these to be arrays for your current calculation - a simple double precision variable would suffice.

If for some reason you do want TAU and CHEZY to be arrays, then inserting the linesafter your variable declarations in PRERES_TELEMAC2D should reference TAU and CHEZY to temporary BIEF vectors (with dimension NPOIN). Your existing code should then work without further modification.

Good luck,
Chris