Code_saturne don't calculate in LES case !!!

[rezki]

Hello,

I am using 4.0.0-rc2 version, I try to run simulation using "LES votrex method" and "Wale" model.
For the moment, I use simple venturi type geometry, with square inlet section which has an edge of 5 mm.

When I run the simulation, the code remains blocked in "Computing geometric quantities" step (see the attached listing file) without mentioning any error.
I use 6 processors and apply 4 time steps.
usvort.f90 is attached for more details.

Does anyone happen to have any idea about what the issue might be ?
Thank you in advance.
RC

[Yvan Fournier]

Hello,

Please upgrade to 4.0 (current version is 4.0.1) in case the issue is caused by a bug which is already fixed.

Did you try running with a different number of processors ?

Also, did you first try running without or with minimal user subroutines (i.e. without the vortex method) ?

To understand the issue, it is important to know where it come from. If you interrupt the run, you may get a backtrace which can be useful, to determine where the code is hanging.

Regards,

Yvan

[rezki]

Thank you for your help.
I already tried with RANS models and it works fine. I will follow your suggestions (upgrade to 4.0 or 4.0.1 and increase the number of processors) and check if that resolves the issue.

Is it possible with code_saturne to run LES simulation without specifying the inlet turbulence boundary conditions (i.e without "usvort.90" for Vortex method or "cs_user_les_inflow.f90" for Synthetic Eddy Method) ?!

Regards.
RC

[Yvan Fournier]

Hello,

If you want turbulence to set up quickly enough, you need some some sort of specific (non-smooth) inlet method...

Did you try the synthetic eddy method ? It is probably more often used than the vortex method.

You can also try using the mapped inlet method (new in 4.0) to "recycle" the flow and help turbulence develop in a short distance (see user examples or Doxygen).

Regards,

Yvan