Verification of OpenMC TBR evaluation using MCNP as reference

[yogesh]

Hi all,

I am a PhD student at University of Tennessee, Knoxville. Right now I am trying to actively learn OpenMC, as an exercise I am conducting a verification study. I am evaluating TBR for multiple solid breeder concepts (8 different breeders and 3 multiplier - Be, Be12Ti, Be12V) and comparing them to MCNP results published in this paper .

In a short summary, the tokamak design in reduced to infinite shells. Source(14 MeV, 100cm radius) is surrounded by homogenized blanket. It starts off as just breeder+multiplier, and then other materials are added within the blanket region with a relative volume fraction while maintaining 200cm width throughout.

note: Geometry and Cross-sections were kept consistent throughout.

Through the simulations, I have received a pretty decent agreement with the MCNP data, but I noticed some huge difference in handful of cases (9/24 cases). I would be open to meeting on zoom or sharing more comparisons. I had some questions, I am posting some of them.

1. In all 8 breeders with Be, the OpenMC evaluation is significantly higher than MCNP. (I have rechecked the code multiple time and as other multipliers show reasonable agreement, I am inclined to trust the code for now). In the comparisons below, there are no materials added in the breeding zone. I was wondering what would be best way to move forward and what other things can be checked, I think I am stuck here.

2. I also wanted to visualize where the neutrons are being absorbed in the geometry, if I could be directed to the relevant documentation, that would be really helpful

3. Can a simulation be run with two different cross sections, if yes, how? if not, is there a way around it?

I apologize for the long post. Let me know if you would be interested in looking at more comparisons to get a better picture of the questions.

Thank you.

Yogesh

[Shimwell]

Hi Yogesh

Here are some random thoughts on your three points.

1. I guess it is good to check the material isotopes fractions, material densities, neutron energy distribution, geometry sizes are all the same between the two codes. Also check you have the same nuclear data for both simulations. Check the density has been corrected when mixing multiplier and breeder. Check the li6 enrichment is the same

2. To visualize absorption I would recommend a RegularMesh tally and setting the tally.score to ['absorption']. These are easy to plot with matplotlib.imshow or there is a nifty package for plotting RegularMesh tallies

3. yes you can run different cross sections. download or process the h5 cross section files and then set the openmc.Config['cross_sections'] to the path of the cross_sections.xml
   relevant link to docs 3. Data Configuration — OpenMC Documentation
   link to nuclear data download Official Data Libraries | OpenMC
   link to nuclear data processing (official, and my own fork)

[paulromano]

@yogesh Two thoughts/questions come to mind:

• How are you evaluating the TBR in OpenMC?
• Are you using any S(α,β) table on the material with Be? Since it is a low Z material, you will likely need to account for molecular effects on thermal scattering with a S(α,β) table.