External source mode: support of MCNPX and GEANT4

New ideas for code development

External source mode: support of MCNPX and GEANT4

Postby Andrei Fokau » Wed Aug 18, 2010 3:47 pm

Citing Jaakko from another topic
Jaakko Leppänen wrote:The most resent version of the code has an external source calculation mode that is still under development. One of the features that is planned but not yet implemented is the capability to read and write source distributions.

I would like to ask for a possibility to load source distribution created by MCNPX (by WSSA card). In this case Serpent can be used for modeling Accelerator-driven systems, the area where MCNPX is the only code applied.

I have never used GEANT4, but I believe that support of its source distributions as well could be another great feature of Serpent breaking monopoly of MCNP/X in this field.
KTH Reactor Physics (Stockholm, Sweden) neutron.kth.se
Andrei Fokau
 
Posts: 77
Joined: Thu Mar 25, 2010 12:25 am
Location: KTH, Stockholm, Sweden

Re: External source mode: support of MCNPX and GEANT4

Postby Jaakko Leppänen » Wed Aug 18, 2010 5:20 pm

Yes, I have certainly considered this possibility. Using the same format would also allow source files generated by Serpent to be used with MCNP.

The problem with ADS applications is that I have very limited experience in neutron physics above 20 MeV where most ACE format cross section libraries end. I have no idea what happens beyond that. Serpent should be able to handle almost every ENDF reaction law in the data format, but the physics is so different at higher energies that there are probably a lot of errors yet to be discovered.
- Jaakko
User avatar
Jaakko Leppänen
Site Admin
 
Posts: 1962
Joined: Thu Mar 18, 2010 10:43 pm
Location: Espoo, Finland

Re: External source mode: support of MCNPX and GEANT4

Postby Andrei Fokau » Wed Aug 18, 2010 6:01 pm

Jaakko Leppänen wrote:The problem with ADS applications is that I have very limited experience in neutron physics above 20 MeV where most ACE format cross section libraries end. I have no idea what happens beyond that. Serpent should be able to handle almost every ENDF reaction law in the data format, but the physics is so different at higher energies that there are probably a lot of errors yet to be discovered.

Some data libraries have evaluations up to 150 MeV and there are very intensive research activities to make it as a standard. However, I cannot say for sure that the physics below 150 MeV corresponds to the one below 20 MeV. As a temporary solution, one would reject all neutrons from the source file with energy above 20 MeV*, and also include an additional option for testing to simulate all neutron energies for which the chosen data library has necessary cross-sections.

*) This is how it was done earlier, when source from LAHET was used in MCNP4C. We also used to calculate source with MCNPX and then use it in burn-up calculations with MCB, loosing high-energy neutrons (normally 0-2% of the neutrons leaking from the spallation target boundary).
KTH Reactor Physics (Stockholm, Sweden) neutron.kth.se
Andrei Fokau
 
Posts: 77
Joined: Thu Mar 25, 2010 12:25 am
Location: KTH, Stockholm, Sweden

Re: External source mode: support of MCNPX and GEANT4

Postby Jaakko Leppänen » Wed Aug 18, 2010 6:36 pm

Andrei Fokau wrote:Some data libraries have evaluations up to 150 MeV and there are very intensive research activities to make it as a standard. However, I cannot say for sure that the physics below 150 MeV corresponds to the one below 20 MeV. As a temporary solution, one would reject all neutrons from the source file with energy above 20 MeV*, and also include an additional option for testing to simulate all neutron energies for which the chosen data library has necessary cross-sections.

*) This is how it was done earlier, when source from LAHET was used in MCNP4C. We also used to calculate source with MCNPX and then use it in burn-up calculations with MCB, loosing high-energy neutrons (normally 0-2% of the neutrons leaking from the spallation target boundary).


Just out of curiosity, have you tried running Serpent calculations with your 150 MeV data libraries?
- Jaakko
User avatar
Jaakko Leppänen
Site Admin
 
Posts: 1962
Joined: Thu Mar 18, 2010 10:43 pm
Location: Espoo, Finland

Re: External source mode: support of MCNPX and GEANT4

Postby Andrei Fokau » Wed Aug 18, 2010 7:08 pm

Jaakko Leppänen wrote:Just out of curiosity, have you tried running Serpent calculations with your 150 MeV data libraries?

According to the information about JEFF 3.1.1 on NEA website, some of evaluations have upper incident neutron energy of 200 MeV (for example, Fe-56, Pb-204). For Serpent calculations, we always use JEFF 3.1 data library prepared by you for earlier version of Serpent (v1.1.0 if I remember correctly), and always in the fission source mode, so none of us has tried the higher energies.

Jaakko, we are expecting your newest set of data (for v1.1.7) with JEFF3.1.1 to arrive soon from NEA data bank. Does it have the energies up to 200 MeV (for some nuclides)? As far as I understand, Serpent doesn't have the energy cut-off at 20 MeV, so in principle it can be done.

Time is very hard to cheat with, but probably someone could make a benchmark with Serpent and MCNPX below 150 MeV using JEFF3.1.1 in pure lead, or in mixture of all nuclides with cross-sections up to 200 MeV.
KTH Reactor Physics (Stockholm, Sweden) neutron.kth.se
Andrei Fokau
 
Posts: 77
Joined: Thu Mar 25, 2010 12:25 am
Location: KTH, Stockholm, Sweden

Re: External source mode: support of MCNPX and GEANT4

Postby Jaakko Leppänen » Wed Aug 18, 2010 8:31 pm

Andrei Fokau wrote:Jaakko, we are expecting your newest set of data (for v1.1.7) with JEFF3.1.1 to arrive soon from NEA data bank. Does it have the energies up to 200 MeV (for some nuclides)? As far as I understand, Serpent doesn't have the energy cut-off at 20 MeV, so in principle it can be done.


Looking at the ACE files, it seems that at least some of the nuclides in the Serpent JEFF-3.1.1 library go up to 200 MeV. Serpent doesn't have a cut-off at 20 MeV, but there may be some "sanity checks" that need to be revised before the code agrees to run the simulation.

Andrei Fokau wrote:Time is very hard to cheat with, but probably someone could make a benchmark with Serpent and MCNPX below 150 MeV using JEFF3.1.1 in pure lead, or in mixture of all nuclides with cross-sections up to 200 MeV.


I also think it would be a good idea to run some tests at higher energies. Any problems that are discoverd are probably there at lower energies as well, the physics is just so different that different reaction laws dominate and problems with negligible contribution are extremely hard to catch.
- Jaakko
User avatar
Jaakko Leppänen
Site Admin
 
Posts: 1962
Joined: Thu Mar 18, 2010 10:43 pm
Location: Espoo, Finland


Return to Development

Who is online

Users browsing this forum: No registered users and 2 guests