Uncertainty Calculations

[Hafssa ZIANI]

 

    Hello,

          I was running CE-tsunami-3d-k6  for a reactor core. But SAMS failed to execute and the program gives me just the sensitivities coefficients in the file (.sdf).

          So, how can I calculate the uncertainty in keff with SAMS as a stand-alone program, using the sensitivity profile that already I had?

          " If it is possible ".

  Best Regards,

[William BJ J. Marshall]

Hafssa,

What error message(s) did you get from SAMS?  If SAMS failed it is almost certain that there are problems with the SDF and you need to sort that out before you try to use the sensitivity data.

B.J. Marshall

Senior R&D Staff

Reactor and Nuclear Systems Division, ORNL

[Hafssa ZIANI]

 

      Hello,

         I have modeled a reactor core which consists of a lattice of 101 fuel elements. I am trying to perform some sensitivity calculations using the CE TSUNAMI-3D          CLUTCH approach, with parallel execution in a cluster environment.
        In general,  when I adopted the average material it works perfectly. However, the problem appears when I used the detailed material to specify individual unit for         each lattice element. 

        Even  SAMS failed to execute. I didn't find any problems with the SDF file.

        I got this error (*.out file):

 **********************************************************************************************************************************

           Congratulations!  You have successfully traversed the perilous path through KENO-VI in 23359.79297 minutes

 **********************************************************************************************************************************
SAMS failed to execute. error code 256
The following data cards precede an = card
      '*********************************END Fil input***************************************

    SCALE is finished on Mon Mar  2 22:39:21 2020.
-------------------------- Summary --------------------------
tsunami-3d-k6 finished. used 1.66697e+06 seconds.
SCALE driver required a maximum of 93.86 MiB of RAM.
------------------------ End Summary ------------------------

        And I received the following message (*.msg file):

sh: copy_graphic: command not found
sh: copy_graphic: command not found
[node-1-13.local:64528] PMIX ERROR: BAD-PARAM in file src/dstore/pmix_esh.c at line 491
[node-1-4.local:188929] PMIX ERROR: BAD-PARAM in file src/dstore/pmix_esh.c at line 491
sh: i_sams: No such file or directory

Thanks for your help
Best regards

[William BJ J. Marshall]

I'm still concerned that your SDF has errors in it.  The SDF is ASCII-formatted, so you should be able to search (or grep for that matter) for clearly bad data like entries that are printed as "nan".  I'm not 100% sure what capitalization you might get for your "not a number" entries, so a case-insensitive search is best.  Other bad data are likely difficult to detect without a detailed review.  It is possible that your SDF is fine, but that a problem in the SAMS block input with TSUNAMI-3D caused the problem.

The total sensitivities are also reported in the SDF, so you can perform direct perturbation calculations to confirm that the sensitivities are correct for the major fission, absorbing, moderating, and reflecting species in the problem.

The best path forward to generate data-induced uncertainties, once you have confirmed that you have accurate sensitivities, is to run TSUNAMI-IP.  The input is fairly simple: uncert_long in the parameter block, and the path to your SDF in both the experiment and application blocks.  As I've said, you really don't want to jump to this step until you've confirmed that the sensitivity data you generated is not only well formed but also correct.

B.J.

[Hafssa ZIANI]

    

   Hello,

  

   Thank you very well. It works  perfectly.

   I already confirmed that the sensitivities coefficients are correct for the major fission, absorbing and moderating between the average and detailed material.

   Also, the uncertainty calculations for the both are approximately equal.

   Thank you again for your help.