Atomic and Nuclear Data for Codes
Fusion codes are only as accurate as the atomic, molecular, and nuclear data they consume: cross-sections, rates, and decay data underpin every result.
Data behind the physics
Beneath every simulation lies a library of measured and evaluated data: the probabilities that particles react, the rates of atomic processes, and the properties of nuclei. Neutronics needs cross-sections; edge codes need ionization and recombination rates; activation needs decay data. The code's accuracy is bounded by the quality of these inputs.
Kinds of data
- Nuclear reaction cross-sections versus energy, for transport and breeding
- Atomic rate coefficients for ionization, recombination, and excitation
- Molecular data for dissociation chains in the cold divertor
- Radioactive decay data: half-lives, branching, and emitted radiation
Evaluation
Raw measurements are combined with nuclear-model calculations into evaluated data libraries that provide a recommended, internally consistent value with uncertainties. Different evaluations can disagree, so which library a calculation uses is a real choice that affects the result, and it must be recorded as part of provenance.
Processing
Evaluated data are processed into the formats codes consume: continuous-energy files for Monte-Carlo, multigroup sets for deterministic transport, and temperature-broadened cross-sections for the operating conditions. Errors in processing can corrupt an otherwise correct code, so processing is itself verified.
Uncertainty propagation
Data carry uncertainties that propagate into predictions. For quantities sensitive to a particular cross-section, such as tritium breeding to the lithium reactions, the data uncertainty can dominate the total. Careful studies propagate these uncertainties rather than treating the data as exact.
Because a breeding or shielding number depends on the data library, Kronos records the library and version alongside every neutronics result so the number can be interpreted and reproduced.