terrapower / armicontrib-openmc Goto Github PK

ARMI has a sample thermal spectrum reactor (with a square lattice) here:

https://github.com/terrapower/armi/blob/main/armi/tests/tutorials/c5g7-blueprints.yaml

What do you think? Could we support that example with the current OpenMC benchmark?

We still need to allow for cartesian geometry type in inputWriters.py. Adding it shouldn't be too hard.

OpenMC tally results all have units of ____ per source particle. A bit of math is required to get results in usable units before sending them back to ARMI

Blocks with more component multiplicities than 1 and N are currently handled in a sub-optimal way by treating components as complex regions with many surfaces. It's better to define the block as a lattice with a small number of different universes. The necessary information is present in the blueprints input file as latticeIDs and lattice maps, but I so far haven't found this information inside the ARMI reactor.core object. I need to either find latticeIDs inside the core object or re-read the blueprints file when necessary.

So far we don't include lumped fission products in openmc materials.

Let's pull the Band-Aid off this. It's time.

The Python org has said that pyproject.toml will be the new way to package ARMI code, going forward.

So we should do that, as we did in ARMI.

Geometry transformations are currently handled inside OpenMCWriter. armi.GlobalFluxExecuter has _performGeometryTransformations() and _undoGeometryTransformations() that will be a better way to do this.

So far only hexagons, circles, and rectangles are supported. Adding other shapes may require making some reasonable assumptions about specific geometry.

It appears no one is using TOX locally, and TOX is only used on CI. This defeats the purpose of TOX entirely, so we might as well ditch it.

I was thinking that adding two things to the README would be really helpful:

1. A link to the OpenMC install guide.
   a. Did you use this? https://docs.openmc.org/en/stable/quickinstall.html
2. How to run your test case.
   a. I assume it's this pointing to this, but giving a command line makes it easier for people new to this project.

Forgive me if you think this is obvious.

OpenMC settings like mesh resolutions, number of particles, etc. are currently hard-coded. It would be good to have some reasonable default settings that are usually overwritten by optional settings in the ARMI settings input file.