FENDL 2 vs 3 Comparisons

[eitan-weinstein]

Closes #81 .

This PR adds a CSV containing every isotope/isomer available in either FENDL2 or FENDL3 data, with the number of reactions per isotope listed for each row. Additionally, I included a histogram comparing the distributions of reactions per parent across the two datasets. The inclusion of this data will be informative to the users of ALARAJOYWrapper of the motivation of upgrading to FENDL3, given the much larger scope of the new data.

A summary of the differences between FENDL3 vs FENDL2 is listed below (updated following changes in #125 ):

• Number of parent isotopes/isomers included:
  • FENDL2: 1865
  • FENDL3: 2921
  • 1.566 times larger for FENDL3
• Total number of reactions included (excluding daughters with isomeric states > 9):
  • FENDL2: 15767
  • FENDL3: 79496
  • 5.042 times larger for FENDL3

[gonuke]

Thanks @eitan-weinstein - these kinds of visualizations will be helpful to think through what's changed, etc.

[gonuke]

I think we need to put a 0.0 for entries where there are no reactions to make sure we get the column ordering right. I'm a little skeptical at the number of entries that appear to have non-zero reactions in FENDL2 and no entry for FENDL3 here???

[eitan-weinstein]

Ok, I'll change it from NaN to 0.0.

The reason for a lot of the missing entries in FENDL3 is that the TENDL2017 neutron database for them have only the TENDL file, but not the additional required PENDF file, so GROUPR cannot convert them to groupwise data.

e.g. ¹H:

https://tendl.web.psi.ch/tendl_2017/neutron_html/H/NeutronH01.html

[eitan-weinstein]

As a follow up to @bohmt 's comparison of the single element response between ALARA with FENDL2 against FISPACT-II for the irradiation of elemental iron, I've used the comparison tool in progress from #144 to include a plot for ALARA with ALARAJOY-processed FENDL3 data.

As can be seen above, ALARA with ALARAJOY-processed FENDL3 has much greater agreement with FISPACT-II than ALARA with FENDL2 at 100 years, in part due to the lower tritium production that we've been seeing from ALARAJOY. If we are to be validating against FISPACT-II, perhaps this suggests that FENDL2's tritium production values are a meaningful overrepresentation, and the lessened values from ALARAJOY-FENDL3 is a valuable correction. Otherwise, in the shorter cooling times, there is a slightly higher total activity from ALARAJOY-FENDL3, which will be interesting to see if it appears across the board for other elements, or is unique to this case.

I'm going to start working on a tool to add to alarajoy_QA.py to generalize this kind of FENDL2 and FISPACT-II comparison for any given element, rather than just iron.