radiological material clearance finder

Give this package a nuclide inventory and it tells you whether the material meets the clearance, exemption or waste disposal limits of the UK, German, US, EU and IAEA regulations, by which route, and what is driving the answer.

It exists because the arithmetic is trivial and the tables are not. An activation calculation hands you a few hundred nuclides; the regulation hands you a table with its own units, its own catch-all rules, its own list of daughters a parent already accounts for, and its own idea of what is in scope at all. This package implements the arithmetic once and treats all 22 tables (4706 limit values) as data, generated from the official sources.

Install

Not on PyPI yet, so install from a checkout:

pip install radiological-material-clearance-finder

There is nothing else to install: in a fresh virtual environment, pip freeze afterwards lists exactly one package, this one.

The shortest useful example

A gram-scale piece of activated steel, as atom counts. Stable Fe-56 is in the inventory because the mass of everything you pass is the Bq/g denominator.

from radiological_material_clearance_finder import Material, clearance_index

steel = Material({"Fe56": 8.4e22, "Co60": 2.1e9, "Cs137": 3.1e8})
result = clearance_index(steel, "UK_EPR16_out_of_scope")

print(result.index)
print(result.clearable)
print(result)

11.244691988476553
False
UK_EPR16_out_of_scope: index 11.24 of 1  [NOT CLEARABLE]
  nuclide        activity        limit    share
  Co60              1.122          0.1   99.7%
  Cs137           0.02901            1    0.3%

Eleven times over the limit, and Co-60 is 99.7% of the reason. To ask which of the shipped routes this material would already pass:

from radiological_material_clearance_finder import clearable_routes

print(clearable_routes(steel))

['UK_EPR16_norm', 'UK_IRR17_natural', 'StrlSchV_exemption_activity', 'UK_EPR16_exempt_material', 'UK_IRR17_registration', 'StrlSchV_incineration_100', 'StrlSchV_landfill_100', 'StrlSchV_landfill_1000', 'StrlSchV_incineration_1000']

Two of those are not the good news they look like. See what it could not account for below.

The sum of fractions

Every regulation here uses the same calculation: divide each radionuclide's activity by its tabulated limit and add up the ratios. Below one, the material meets the limits. Germany calls it the Summenformel, the UK the summation rule, the NRC the sum of fractions rule, and Fetter calls the result a waste disposal rating. They are one calculation over different tables, so it is implemented once.

Specific activity in Bq/g is scale invariant, which is why the example above needed no density and no volume: atom counts, atom densities in atoms/barn-cm and mass fractions all give the same index, to floating point round-off. The volumetric sets (the US ones, in Ci/m3) do need a density, and atom densities supply one on their own.

What is different about this one

Zero dependencies

dependencies = [], and importing the package pulls in nothing outside the Python standard library. OpenMC is not required and is imported only inside the two functions of openmc_interop, if you call them. That is deliberate: a clearance check should be runnable anywhere, including somewhere that cannot install a neutronics stack.

The tables are data, with provenance

Adding a jurisdiction means adding a JSON file and a build script, never a branch in the calculation. Every set records where its numbers came from and when they were fetched, and you can ask it at runtime:

from radiological_material_clearance_finder import get_limit_set

limits = get_limit_set("StrlSchV_metal_recycling")
print(limits.label)
print(limits.source)
print(limits.url)
print(limits.retrieved)
print(limits.limits["Co60"], limits.units)

Metal scrap for recycling
Strahlenschutzverordnung (StrlSchV) 2018, Anlage 4 Tabelle 1, Spalte 14
https://www.gesetze-im-internet.de/strlschv_2018/anlage_4.html
2026-09-10
0.6 Bq/g

Nothing reaches the network at run time. The scripts in tools/ regenerate the tables from the official sources on demand, and each asserts that every limit is one significant figure times a power of ten, which is the shape these regulations use without exception, so a misparsed exponent fails the build instead of shipping quietly.

The result reports what it could not account for

A bare number cannot tell you that it left something out. ClearanceResult can, and this is the part worth knowing before you trust an index. Recall UK_EPR16_norm from the passing list above:

norm = clearance_index(steel, "UK_EPR16_norm")
print(norm)
print(norm.uncovered_fraction)

UK_EPR16_norm: index 0 of 1  [CLEARABLE]
  2 nuclide(s) with no limit, 100.00% of total activity
1.0

An index of zero, formally clearable, and every becquerel in the material outside the sum. That table is the NORM one and covers only natural-series radionuclides, so neither Co-60 nor Cs-137 appears in it, and this set has no catch-all limit, so the two of them contribute nothing at all. uncovered_fraction of 1.0 is how you find that out. A result carries five such fields:

Field	What it records
uncovered	activity with no limit at all, absent from the index
defaulted	nuclides that fell back on the set's catch-all limit
excluded	daughters a parent's limit already accounts for in full
credited	activity a parent could only partly account for, the rest still in the sum
out_of_scope	the regulation excludes this material outright

The last three exist because the regulations publish their own parent-to-daughter equilibrium tables, and those tables genuinely disagree: EPR 2016 pairs Zr-95 with Nb-95m where IRR 2017 pairs it with Nb-95. Each limit set therefore carries its own, so no decay chain is assumed on your behalf.

Include the stable isotopes

The Bq/g denominator is the mass of everything you pass. Drop the stable Fe-56 from the example above and the same two radionuclides give an index of 3.14e+14 instead of 11.24, because the inventory now weighs 2.8e-13 g rather than 7.8 g. An inventory filtered to its radioactive nuclides is the single easiest way to get a wrong answer here. Material warns when nothing you passed is stable.

Verification, honestly

The US sets agree with openmc.Material.waste_disposal_rating to floating point round-off, all 81 Fetter lower bounds match the 1990 paper, the IAEA PDF extraction and the EU XHTML parse agree on all 257 values they have in common by independent routes, and four regulatory tables re-derived from the live sources match the committed data. The suite currently stands at 277 tests, all passing.

All of that is machine-to-machine agreement. No human has checked any of these tables against the published regulations, and this package is not a substitute for reading them or for your own regulatory advice.

Where to go next

• API reference, the package overview and every public symbol.
• Material for the five ways to build an inventory: atom counts, atom densities, specific activities, masses and mass fractions.
• Limit sets for the 22 shipped sets, their provenance, and registering your own.
• Cooling for time_to_clear, which interpolates a crossing time from a series of cooled materials and refuses to report one that ingrowth later reverses.
• Classification for NRC waste classes and UK waste categories.
• OpenMC interop for starting from an openmc.Material or a depletion results file.