Nuclear data sensitivity for reactor physics parameters in a lead-cooled reactor. (October 2018)

• Record Type: Journal Article
• Authors: Trottier, Alexandre; Adams, Frederick P.; Levinsky, Alex; Roubtsov, Danila
• Is Part Of: Annals of nuclear energy. Volume 120(2018)
• Page Start: 333
• Page End: 347
• Publication Date: 2018-10
• Subjects: Lead-cooled reactor -- Nuclear data -- Uncertainty analysis -- Sensitivity analysis -- Monte Carlo; Nuclear energy -- Periodicals; Nuclear engineering -- Periodicals; 621.4805
• DOI: 10.1016/j.anucene.2018.05.047
• Languages: English
• ISSNs: 0306-4549

Abstract

Highlights: Lead-cooled reactor parameters evaluated using SCALE, MCNP6 and SERPENT. Nuclear data sensitivity coefficients from SCALE and MCNP6 are compared. U-235 radiative capture uncertainty in the 2.5 keV to 1 MeV range dominates. A comparison against a new Stochastic sampling framework is presented. Abstract: Sensitivity of fast reactor physics coefficients to nuclear data differs from thermal reactors. A conceptual lead-cooled reactor has been modelled and assessed for nuclear data sensitivity. A TSUNAMI sensitivity analysis was performed to estimate uncertainty and sensitivity of the neutron multiplication factor to the nuclear data of all nuclides and reactions relevant to the reactor concept. The results were compared against MCNP6, which also employs an adjoint-based method for sensitivity calculations. Sensitivity of the neutron multiplication factor was found to agree closely between MCNP6 and TSUNAMI, for lead cross sections and other important nuclear data. The TSUNAMI assessment of nuclear data impact was also examined through a brief survey of nuclear data evaluations. The evaluations were found to have differences of interpretation that impact the TSUNAMI calculations. The consequent uncertainties in fuel-temperature and coolant-voiding coefficients were also examined in TSUNAMI. Based on the TSUNAMI ranking of important data, a subset of nuclides and reactions were selected for stochastic sampling of cross-section data using NJOY. The sampled cross-section data were then used in SERPENT simulations to assess uncertainties in calculated parameters, which were consistent with the TSUNAMI results.