Use ENDF/B-VII.1 MC2-3 nuclide IDs #1982
Conversation
Changes relative to VII.0 are: * V is no longer elemental (add V-50/51) * Zn is no longer elemental (add Zn-54/65/66/67/68/70) * Add Tm-168/169/170 * Add Ta-180 * Add W-180 * Add Tl-203/205 * Add Th-231 * Add Pa-229/230 * Add U-230/231 * Add Np-234 * Add Am-240 * Add Cm-240 * Add Bk-245/246/247/248 * Add Cf-246/248 * Add Es-251/252/254m So 32 new isotope evaluations and the elemental evaluation of V and Zn are gone in VII.1.
To support the VII.1 MC2-3 library, ARMI needs a lookup dictionary by library ID. Previously, there was just a single MC2-3 dictionary to do this, but now we need two - one for the VII.0 library and one for the VII.1 library. The preexisting `byMCC3Id` dictionary is still availabe and now returns VII.1 data. Also: * adds getter functions for the VII.0 and VII.1 IDs on a couple classes * modifies existing versionless getMcc3ID getters to wrap VII.1 getters * modifies readMCCNuclideData to parse the VII.1 nuclide labels
Previously, python could find the byMcc3Id in the global scope without needing to declare it in the readMccNuclideData function. After introducing the new byMcc3IdEndfvii.* dictionaries, it can't anymore. Declaring all the byMcc3Id* dictionaries as global to be safe.
These new tests are patterned off the old test for loading the MC2-3 data, they just check the dedicated VII.0 and VII.1 dictionaries and class functions, now.
This hopefully leads to fewer broken integration tests in Nala. If we point to ENDF/B-VII.1 isotopes, tests break since V is isotopic, rather than elemental.
This reverts commit 67be3f2. The sensible default for byMcc3Id is ENDF/B-VII.1 since that's what we're adopting for the project
This is a wrapper around the function that return the VII.1 MC2-3 ID
Previously ARMI assumed elemental data consistent to VII.0. Since we are transitioning to using VII.1 data by default, this logic needs to be updated to act on the VII.1 elemental list.
The diff vector was originally expected to be 477. With the change to ENDF/B-VII.1, elemental vanadium is gone, but two isotopes of vanadium are added. Each element/isotope has 3 pieces of data, so the new size of the vector is 477 - 3 * [2 - 1] = 480.
|
Aaron Reynolds seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account. You have signed the CLA already but the status is still pending? Let us recheck it. |
Co-authored-by: John Stilley <1831479+john-science@users.noreply.github.com>
Co-authored-by: John Stilley <1831479+john-science@users.noreply.github.com>
There was a problem hiding this comment.
Okay, @aaronjamesreynolds , I have determined that I like your PR. I just hate global nuclides.
I'm sorry it took me so long to get there.
Thanks for contributing!
| @@ -553,7 +553,7 @@ def eleExpandInfoBasedOnCodeENDF(cs): | |||
| ) | |||
|
|
|||
| elif cs[CONF_XS_KERNEL] in ["", "SERPENT", "MC2v3", "MC2v3-PARTISN"]: | |||
| elementalsToKeep.update(endf70Elementals) | |||
| elementalsToKeep.update(endf71Elementals) | |||
There was a problem hiding this comment.
I feel like SERPENT can be removed from this list since its not tested here. Additionally, doesn't this remove support of ENDF/B-VII.0 explicitly? It feels like there should be a setting for which library is being selected OR we should ensure that that changelog notes that this removes the backwards capability.
There was a problem hiding this comment.
@john-science, any concerns with removing SERPENT from this list? We could also just make another conditional that has SERPENT default to the endf70Elementals .
Having a setting for the library was my original plan, but global nuclides complicates this. John and I have are hoping to implement such a setting once global nuclides are addressed (#473).
There was a problem hiding this comment.
I understand the complications of the mixed library options based on that ticket and I agree that fixing the global nuclides issue would be a reasonable resolution.
@drewejohnson originally made this setting, so I think it should be appropriate to remove at this time since there is no use for it. I am OK with keeping it as an option though and adding endf70Elementals on that branch or removing SERPENT and updating the changelog.
There was a problem hiding this comment.
Alright Jake, I added a conditional so that VII.0 elementals are used with Serpent.
What is the change?
Currently, ARMI returns MC2-3 nuclide IDs consistent with the ENDF/B-VII.0 library. This PR makes it so that:
mcc-nuclides.yaml,getMcc3Id()or indexing thebyMcc3Iddictionary) is consistent with ENDF/B-VII.1, andgetMcc3IdEndfVII0()or indexing thebyMcc3IdEndfVII0dictionary.This paper provides a good overview of the differences between VII.0 and VII.1.
Tests have been updated or created to verify that
getMcc3Id(),getMcc3IdEndfVII0(), andgetMcc3IdEndfVII1()return the expected IDs. Additionally, two tests (test_compareDatabaseSim()andsetUp()[for the cross section manager]) needed to be updated since vanadium is elemental in VII.0 and isotopic in VII.1.Why is the change being made?
ENDF/B-VII.1 contains updated evaluations and entirely new evaluations compared to ENDF/B-VII.0. ANL has recently released VII.1-based MC2-3 libraries, so the most recent available data is the sensible default.
Checklist
docfolder.pyproject.toml.