code_saturne User's Forum

Hi
 
I recently installed Code Saturne 2.0-rc1 and i started doing the tutorial "CFDSTUDY" (cf "Annonces" section) but i didn't linked Salome-5.1.3 and Saturne 2.0-rc1 using the script and archive proposed
1) the Code runs fine when the fluid properties are set to "constant" in the Saturne GUI.
2) when i changed the "constant" properties to "user law" and enters the tutorial's value, the simulation diverges after a while
 
Is expected because i didn't liked Salome and Saturne the way it was advised in CFDSTUDY , or is there something else?
Which files would you need to clarify this?

The "Ok" case pictures.
 
Best regards

Hello,
 
Could you post your mesh (MED file) and XML file here so that we can
see if we reproduce the problem (without needing to go through the
tutorial) on a standalone version ? Being linked with SALOME or not should not make any difference regarding the calculation (the GUI has additional features under SALOME, but the calculation is the same).
 
Best regards,

Hi
 
Here is the command file

Here is the file

My mesh looks a bit different from the one in the Tutorial but i used similar option (or so i think)

I tried with a finer mesh and the same issue arises.[with one thing bogging my mind: one area of the mesh was not refined, so i will try to construct a better mesh and launch again]

Regards

Oops :o ! There is mistake in the formula of the viscosity in the tutorial embedded in SALOME. In fact one must read:
mu = 0.0015452 - 3.2212e-5 * TempC + 2.45422e-7 * TempC^2;
Thanks a lot for the feedback !

Thanks
It does work fine.
The thing is, i wasn't very carefull also, not checking the fluid properties: i was just looking for a test case.
Would Code Saturne provide a list of test cases like Code Aster does? For one install, the user can check if the code works well by trying 1case, 29 cases or the daunting 1890 cases or so.
 
Thanks again for your help.

Providing end-users with a list of tests cases if something we have in mind, but we need some more time to make our mind on how to do it...

Hi,
 
A quick question, for mu we have only one pre-defined variable TempC, does that mean we cannot use others, such as rho ?
 
because if I have a rho expression and a nu expression, I would expect mu = rho * nu.
 
I tried to use rho in the expression editor but I got an error.
 
Many thanks :)
 
Wayne
http://code-saturne.blogspot.com/

Hello,

You cannot indeed use others, such as rho. Alexandre could probably give you more details on why this choice was made, but one reason is that the order in which physical properties are computed is unspecified, so using one for another would be dangeous (maybe building a dependency graph for expressions could be imagined ?).

With user subroutines, this is not an issue, as it is the programmer's responsibility to define things in the right order.

With MEI, the only workaround available today would thus be to expand rho's and nu's definitions in mu.

Regards,

  Yvan