Error not displayed and not found in log file either

MFiX Bug report
#1

Hello

I keep receiving this error and I can’t find any additional information to help recognize the issue:

*** The MPI_Comm_f2c() function was called after MPI_FINALIZE was invoked.

*** This is disallowed by the MPI standard.

*** Your MPI job will now abort.

[mariam-Precision-Tower-7910:457583] Local abort after MPI_FINALIZE started completed successfully, but am not able to aggregate error messages, and not able to guarantee that all other processes were killed!

mpirun detected that one or more processes exited with non-zero status, thus causing

the job to be terminated. The first process to do so was:

Process name: [[34104,1],54]

Exit code: 1

Here is the archive:
pinewood_gasification_2022-04-06T144552.231706.zip (149.1 KB)

I only need to know the root cause of this error.

Kindly help,
Thanks

#3

Hi Mariam. I ran this case on 8 cores, after about 14 hours we got a divergence.

    DT < DT_MIN.  Recovery not possible!

Prior to this there were a large number of:

    The sum of mass fractions is not equal to one.

This is a pattern we’ve been seeing quite a bit of lately here on the forum, where the mass fractions get out of line and then the whole simulation diverges. I’m not 100% if the mass fractions are causing the divergence, but I’m starting to suspect that it is related. We are working on this problem and will follow up.

Thanks for the report, and for your patience

– Charles

1 Like
#4 
% grep ':-(' */LOG

PINEWOOD_GASIFICATION.LOG: t=     1.0428 Dt= 0.2912E-03 NIT=  1MbErr%=  2.078    : Energy Equation diverged :-(
PINEWOOD_GASIFICATION.LOG: t=     1.7417 Dt= 0.1254E-03 NIT=  2MbErr%= -11.02    : Negative gas density detected :-(
1 Like
#5

Dear Charles

Thank you very much for your reply. I will try to investigate what causes the energy equation to diverge, however is there anything I can do with relaxation factor (for example) to solve this issue?

Thanks again,
Mariam

#6

Mariam - I just found these “sad face” messages in the log. I hadn’t noticed them previously. Not all of them indicate a fatal error - in some cases, MFiX is able to recover by reducing dt and the “sad face” is followed by a “happy face”:

t=     1.0428 Dt= 0.2912E-03 NIT=  1MbErr%=  2.078    : Energy Equation diverged :-(
   Recovered: Dt= 0.26211E-03 :-)

I think what we need to find is the final :-( message, one that is not followed by Recovered:

#7

It failed after reaching DTMIN at about time=3.1 seconds for me. Before failing, the residual for k and then the residual for gas temperature failed to converge. Your residuals are fairly small (1E-6) so you could restart with larger residuals to see if you can get past the non-convergence and then lower the residuals again if you want to. You can also temporarily switch to first order discretization or reduce the relaxation factor. My suggestion is to backup your files and do a restart with larger k-epsilon and energy equation residuals.

1 Like