Sensitivity of Fast-Reactor Economics to Uncertainties in Nuclear Data

The uncertainties in physics parameters and in fuel cost of fast power reactors due to current nuclear data uncertainties are considered for two sodium-cooled, oxide-fueled, 1000 MW(e) reactors. One reactor has a low core neutron leakage and a core composition that results in very low fuel cost and short doubling time for fissile material. The second reactor is forced to satisfy stringent safety criteria associated with the sodium voiding reactivity problem and, hence, has a high core neutron leakage and a large amount of moderating material (BeO) in its core composition, with resultant higher fuel cost. Ranges in uncertainties in fuel cost and doubling time for each recognized significant data uncertainty over a “correlated” energy interval are evaluated for each of the two reactors, using the highest and the lowest reasonable values of that nuclear data parameter, as well as the values recommended in the Brookhaven Evaluated Nuclear Data Files (ENDF/B). Combined uncertainties in data produce an uncertainty of ∼0.15 mill/kWh in fuel cost for the low leakage reactor and 0.25 mill/kWh for the reactor for which the design composition (BeO content) must be adjusted as the nuclear data are varied, in order to satisfy the stringent sodium voiding safety criteria. The current large uncertainty in the value of α for ²³⁹Pu below 15 keV is the largest single contributor to this overall fuel cost uncertainty, closely followed by the uncertainties in for ²³⁹Pu and σ_c(σ_c = σ_n,y) for ²³⁸U. On the basis of the calculated sensitivity of fuel cost to specific data uncertainties, a set of targets for nuclear data accuracy that would reduce the calculated fuel cost uncertainty to about ±0.03 mill/kWh is recommended.