Running gamma transport simulation

Separate section for discussion on gamma transport
User avatar
Jaakko Leppänen
Site Admin
Posts: 2387
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Running gamma transport simulation

Post by Jaakko Leppänen » Sun Sep 22, 2013 9:54 pm

Seems a bit strange that the buffer runs out. It means that one photon history creates 10 secondaries that are put in que to wait for the main history to complete, and the buffer containing new particle structures is underrun. The only process that currently creates secondary photons is pair production, and it should be physically impossible for the annihilation gammas to induce new pair production reactions. The buffer limit can be adjusted by changing the last value of TestParam() on line 6704 of readinput.c:

Code: Select all

	      if (k < np)
		WDB[DATA_PART_GBUF_FACTOR] =
		  TestParam(pname, fname, line, params[k++], PTYPE_REAL, 
			    1.0, 10.0);
Do you see the source region in your geometry plots?
- Jaakko

User avatar
markdehart
Posts: 40
Joined: Fri May 17, 2013 6:10 pm
Security question 1: No
Security question 2: 92
Location: Idaho National Laboratory

Re: Running gamma transport simulation

Post by markdehart » Tue Sep 24, 2013 5:34 am

This last patch got me off and running. I set gbuf to 50 (sometime I'll experiment to see how low I can take it and get it to work). But now I have a case that runs to completion, and I'll now start looking at detector scoring.

I also got the tracks option to work. Have to say that is pretty darned cool. I ran a coarse plot case initially, and it looked like my source was being correctly placed. I just started a finer case with more tracks to see what kind of "art" I can make! :ugeek:

User avatar
markdehart
Posts: 40
Joined: Fri May 17, 2013 6:10 pm
Security question 1: No
Security question 2: 92
Location: Idaho National Laboratory

Re: Running gamma transport simulation

Post by markdehart » Mon Sep 30, 2013 6:04 pm

It appears that the gbuf buffer issue is related to the number of processors in an OpenMP calculation. I came in this morning and tried to run a case that ran last week, and kept finding that a value of 50 was too low (although the same value worked last week). I had made a minor change in the input but when I set it back the way it was it was still failing with gbuf too small errors. Then I decided to step down from 20 processors to 12 and the case ran (so I assume I was running with just 12 processors last week). I did some experimenting and found that on one processor (all that is really needed for a gamma calculation anyway) I can make gbuf as small as 3 and the case will run just fine.

User avatar
Jaakko Leppänen
Site Admin
Posts: 2387
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Running gamma transport simulation

Post by Jaakko Leppänen » Mon Sep 30, 2013 8:24 pm

Thanks! I'll try to fix the problem for the next update.
- Jaakko

samuelshaner
Posts: 1
Joined: Wed Aug 06, 2014 6:27 am
Security question 1: No
Security question 2: 92

Re: Running gamma transport simulation

Post by samuelshaner » Wed Aug 06, 2014 6:55 am

I'm working with Serpent 2 to perform assembly burn up and subsequent gamma transport calculations on the spent fuel assembly for the OECD/NEA depletion calculation benchmark (provided with serpent as PWR_BURNUP example). The burn up calculation is rather straightforward, but I am having some issues with setting up the gamma transport simulations:

- I am trying to use the MCNP cross section data since it includes processed gamma libraries (mcplib63 and mcplib84, I believe). Is there an "xsdirconvert.pl" script to properly convert the MCNP xsdir file to an xsdata file that includes photon transport data? If not, do you have any suggestions on generating an xsdata file that includes photon transport data using MCNP cross section data? A previous post on this thread mentioned this issue so I'm curious if it was resolved with an updated xsdirconvert perl script.
- I'm trying to model a gamma source from all isotopes in spent fuel (mainly Ba-137m, Cs-134, and Eu-154) and get the photon flux and energy spectra at various locations surrounding the assembly. Will I need to explicitly set a gamma source for each cell containing decaying isotopes or will the calculation take these into account?
- Are there any example input files you could provide for gamma transport calculations?

Thanks!
Sam

User avatar
Jaakko Leppänen
Site Admin
Posts: 2387
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Running gamma transport simulation

Post by Jaakko Leppänen » Tue Aug 12, 2014 10:43 am

I checked the xsdirconvert.pl script available at the Serpent website:

http://montecarlo.vtt.fi/download/xsdirconvert.pl

and it seems to be able to set the correct data type for photon libraries. Recently there hasn't been much work done for gamma transport mode in the official beta version of Serpent 2, mainly because the work on the physics routines is still under way. Once that is completed (should be sometime later this year), we'll fix the other issues as well. So please note that even if you are able to run the simulation, it is still missing some important physics (e.g. production of secondary bremsstrahlung), so the results cannot be expected to yield very good results.

At the moment the source spectrum needs to be defined using the "src" card, and the most flexible option is to read the source points from a separate file. We'll start working on a radioactive material source option after the transport part has been completed and tested. The idea is that Serpent could generate the source spectrum based on material composition and nuclide-wise gamma spectra from ENDF decay files.
- Jaakko

User avatar
Jaakko Leppänen
Site Admin
Posts: 2387
Joined: Thu Mar 18, 2010 10:43 pm
Security question 2: 0
Location: Espoo, Finland
Contact:

Re: Running gamma transport simulation

Post by Jaakko Leppänen » Tue Aug 12, 2014 1:12 pm

I uploaded a simple photon transport example at: http://virtual.vtt.fi/virtual/montecarl ... les/GAMMA/
- Jaakko

Post Reply