Demonstration of a random sampling approach to uncertainty propagation for generic pebble-bed fluoride-salt-cooled high temperature reactor (gFHR)

Abstract

Nuclear reactor safety analysis requires accurate propagation of uncertainty, a research topic that is rather underdeveloped for advanced reactors. We develop and demonstrate a framework for uncertainty analysis (UA) based on random sampling specific to a benchmark generic fluoride salt-cooled high temperature reactor (gFHR) core. This framework effectively encompasses uncertainties from nuclear data, manufacturing tolerances, and operational uncertainties such as thermophysical properties and material input uncertainties. These uncertainties are propagated to steady state figures of merit (k-effective/fuel temperature coefficient) and a single pebble depletion model.The gFHR core is modelled in SCALE 6.3b using core parameters from open-source data. Nuclear data input uncertainties are sampled from ENDF/B-VII.1 using the covariance-based XSUSA method via the Sampler sequence in SCALE. This article uses a standard statistical analysis based on normality and absolute margins to describe output probability distributions. The distribution of both steady state output quantities failed tests of normality; however, graphical inspection shows that the normal and uniform parameterizations sufficiently represent the output uncertainty. In practice, this may require more conservative engineering judgement. The mean and propagated uncertainty for k-effective and FTC were found to be 1.01271 ± 1.44% and −5.156 ± 8.22% (pcm/K) respectively. Distributions of depletion output quantities showed the trend of failing normality tests early and passing most often as burnup increased. Burnup (expressed in %FIMA) passed normality tests at end-of-life discharge (9 passes) with a mean and propagated uncertainty of 0.1877 ± 3.2% with 95% confidence that the true uncertainty falls between 3.012% and 3.312%. Absolute margins, or uniform parameterization, of the uncertainty distributions for steady-state and depletion quantities are also presented in the results section. Unique results from the nominal behavior of this depletion model are reflected in the depletion output distributions and explored further in a case study on the production chain of plutonium-239.