Installation of Code saturn with syrthes with the python scr

[Mathieu Genthon]

Hi everybody.

First of all, I'm glad to have found this forum,because it's very useful.

In fact, I have to study the air flow under photovoltaic panel(thermal transfer between air flow and panel). So I think I have to use syrthes coupled with saturn?
 
I have succesfully installed saturn 2.0.0rc1, by using the python script supplied in this forum (http://code-saturne.org/forum/viewtopic ... =940#p5239)

Unfortunately, syrthes has not been installed. As I'm  a beginner with Linux I'm wondering if it 's possible to install syrthes with the python script. maybe some lines to add in the setup files?

If it's not possible with the python script, how can I do it?

Thanks in advance

Mathieu  

[David Monfort]

Hello,

Indeed, you should use SYRTHES if you think "transversal" effects are important in the thermal conduction in your photovoltaic panel; but,  as a first approach you can also use the 1D-wall thermal model available in Code_Saturne (see the user subroutine uspt1d.f90)

Unfortunately, we do not provide the installation of SYRTHES through the Code_Saturne Python installer... so, you will have to do it manually.

First, download the SYRTHES archive on this page http://innovation.edf.com/recherche-et-communaute-scientifique/logiciels/syrthes-41220.html,

Second, read the README file at the top directory of the archive and follow the instructions to install the code, Third, install Code_Saturne with the Python script by providing the setup file with the path to your SYRTHES installation.

Feel free to come back on this forum if you have some trouble installing the codes.

David

[edit] DM: correct name for the "1d thermal wall module" subroutine

[Mathieu Genthon]

Hi,

Thanks for the advice. I've downloaded SYRTHES and installed it.

Now, when I try to install Saturn 2.0rc1 with the python script, it installs cgns, hdf5, etc... till NCS. After two hours the installation is still not over: (there is the screen of my terminal)

csf@calcul-csf:~/saturne/installer$ ./install_saturne.py

       Installation of Code_Saturne
        ____________________________

The process will take several minutes.
You can have a look at the log file meanwhile.

Check the setup file and some utilities presence.
Installation of CGNS
Installation of HDF5
Installation of MED
Installation of BFT
Installation of FVM
Installation of MEI
Installation of ECS
Installation of NCS

then I type enter and I got :

Error during generate documentation stage of NCS.
See install_saturne.log for more information.

Before using Code_Saturne, please update your path with:

  cspath=/home/csf/opt/cs-2.0-beta2/bin
  export PATH=$cspath:$PATH

The documentation should then be available through the commands:
  code_saturne info -g refcard
  code_saturne info -g user

Thank you for choosing Code_Saturne!

Do you have some suggestion? Is the code well installed?

[Claus Andersen]

Yes, the executables and libs of CS are installed and should work, the documentation is not installed. Take a quick look here http://code-saturne.org/forum/viewtopic.php?f=8&t=966 fix that issue.

[Mathieu Genthon]

Ok, thanks for the info.

Concerning the use of user subroutine ustpar1d.f90, I can't find it, but I found uspt1d.f90. Is it the same, which seems to be the 1D-wall thermal module.

I have some question concerning the use of this subroutine:

" if (iappel.eq.2) then
   if(iphas.eq.1) then
      do ii = 1, nfpt1d
        ifac = ifpt1d(ii)
        nppt1d(ii) = 8
        eppt1d(ii) = 0.01144d0
        rgpt1d(ii) = 1.d0
        tppt1d(ii) = 25.d0
      enddo
   endif
endif

if (iappel.eq.3) then
   if(iphas.eq.1) then
      do ii = 1, nfpt1d
         iclt1d(ii) = 1
! Physical parameters
         ifac = ifpt1d(ii)
         if (cdgfbo(2,ifac).le.0.025d0) then
           iclt1d(ii) = 3
           fept1d(ii) = -1.d4
         else
           iclt1d(ii) = 3
           fept1d(ii) =  1.d4
         endif
         xlmt1d(ii) = 31.5d0
         rcpt1d(ii) = 3.5d6
         dtpt1d(ii) = 0.3d0
      enddo
   endif
endif "

What is d0 and d4? How can we define it?

Are there others subroutines where I have to change value before launching the calculation in the GUI?

[Jean Deschodt]

Hi

d0 and d4 mean 10^0 and 10^4  there are not variables.

"d" is for double precision:

0.3D0 = 0.3
-1.d4 = -1.e4 = -10000.
1.d-4 = 1.e-4 = 0.0004

[David Monfort]

And the right name for the subroutine is actually uspt1d.f90... sorry for this mistake !

[Mathieu Genthon]

Hi again!

thanks for this answer Jean and David.

I've tried to use uspt1d.f90.  The trouble is that it did not work: i got this message in the listing file:

ALL PHASES  : 1D-WALL THERMAL MODULE NOT ACTIVATED                  NFPT1D =          0

I attach the subroutine to this messsage.

[David Monfort]

Hello Mathieu,

I've just read your subroutine and it seems correct AFAICT, except for two small "!" missing at the beginning (see the attached subroutine). Could you test with the routine I provide?

You can also try to print the value of nfpt1d at line 303, with something like: write(nfecra,*) nfpt1d

By the way, have you correctly put the Fortran file in the SRC directory so that it is compiled by the script? Indeed, you need to copy the user file from the REFERENCE/base directory to the SRC directory before modifying it, otherwise it won't be accounted for.

David

[Mathieu Genthon]

I've readded the two "!" that I had removed (line 71 and 81) . More over, the uspt1d routine was not in the right directory, so that it did not compile.

I've now put it in the SRC directory.

It stil does not work, but I've  got this message in the compile.log file:

/home/csf/tmp_Saturne/4p.4CMWTH.06030914/src_saturne/uspt1d.f90:288.17:

 CALL GETFBR(top,NLELT,LSTELT)
                1
Error: Symbol 'top' at (1) has no IMPLICIT type gfortran -I/usr/local/include -x f95-cpp-input -Wall -Wno-unused  -D_CS_FC_HAVE_FLUSH  -O -c /home/csf/tmp_Saturne/4p.4CMWTH.06030914/src_saturne/uspt1d.f90

I don't understand because I've defined 'top' as a group in SALOME.