code_saturne User's Forum

Hello everyone,

I would like to know if anyone had this kind of error, I had it after 5hours and 46 minutes of calculations:

../../../../src/code_saturne-5.0.7/src/alge/cs_sles_it.c:4932: Fatal error.

Jacobi: error (divergence) solving for k

I have attached the listing file below

Thanks a lot for any answer,

Hello,

You case is probably diverging, but it seems you reduced the logging (listing) output frequency, so we do not have details about intermediate time steps.

Could you re-run your computation with listing output frequency set to 1 ?

Regards,

Yvan

Hello,

Thanks for your answer, and as requested here is the listing with an output every 1 step in the attached file, (and the error file).

Hello,

It still seems the output frequency is different than 1 (I see NTLIST = 90) in the listing.

Check again under "calculation control/Output control/Log frequency" in the GUI (and check you do not have a setting which overwrites this in user source files if present).

Could you also attach your XML file ?

Regards,

Yvan

Hello,

This calculation is for the same game, and for 10 iterations, I simplified, more my geometry, but still get the same ERROR, I have already generated with the same characteristics of mesh for other geometry and it works well, so I don't get this "Jacoby error",

I have attached files requested int he last reply,

Thanks for any answer, best regards

Houda

Hello,

Your CFL number is already very high at the second iteration, so you probably have a much too large time step. There are also warnings about gradient non-convergence (a mesh issue) so using a least-squares gradient reconstruction (instead of iterative reconstruction) might help here.

Regards,

Yvan