Using it with OpenMC¶

OpenMC is not a dependency. It is imported inside the two interop functions and nowhere else, so installing this package pulls in nothing and every limit set works without it.

from radiological_material_clearance_finder.openmc_interop import from_openmc_material

material = from_openmc_material(openmc_material)

That reads the atom densities in atoms per barn-cm, which fixes the mass density as well, so the result supports both the Bq/g sets and the volumetric US sets with nothing else supplied. The OpenMC material's name and volume are carried across where it has them.

From a depletion¶

from_depletion_results turns a depletion straight into the series that time_to_clear expects:

import openmc.deplete
from radiological_material_clearance_finder import time_to_clear
from radiological_material_clearance_finder.openmc_interop import from_depletion_results

results = openmc.deplete.Results("depletion_results.h5")
series = from_depletion_results(results, material_id=1)

seconds = time_to_clear(series, "StrlSchV_metal_recycling")

The keys are seconds since the start of the results, and the values are Material objects. Nothing about the series is special, so a series assembled by hand from any other depletion code works identically.

Cooling time¶

The index falls roughly exponentially with cooling time, so time_to_clear interpolates logarithmically in the index between the two samples that bracket the threshold. Interpolating linearly instead places the crossing systematically late, by a factor that grows with how far apart the samples are.

This package computes no decay of its own. That is deliberate: the series comes from whatever produced it, so nothing here can disagree with the depletion the numbers came from.

An index that rises again

A daughter can grow in faster than its parent decays, so a material can meet the limits at one time and fail them later. Reporting only the first crossing would be actively misleading, so time_to_clear raises IngrowthError when the index climbs back above the threshold at a later sample. Pass allow_ingrowth=True to take the first crossing anyway.

What the agreement with OpenMC does and does not show¶

The US limit sets are checked against openmc.Material.waste_disposal_rating on every CI run. The agreement covers the index, the per-nuclide breakdown and the 10 CFR 61.55 waste classification, and holds to floating point round-off:

tests/test_agreement_openmc.py .......................  24 passed

Be clear about what that establishes. It shows that this package's path from an inventory to a specific or volumetric activity, and its sum of fractions, match an independent implementation exactly. It says nothing about whether the underlying limit tables are right, because both implementations read the same Fetter and NRC numbers. Table correctness is a separate question, addressed in where the data comes from, and it is checked against the regulations rather than against OpenMC.

Two of the three high severity defects found in this package during review were in the tables, not the arithmetic, and the arithmetic agreed with OpenMC exactly throughout. Machine agreement and correctness are not the same thing.