Influence of nuclear data uncertainties on reactor core calculations

Abstract

Abstract To investigate the influence of nuclear data uncertainties on reactor core calculations systematically, the sampling based uncertainty and sensitivity software package SUSA developed at GRS was extended for the use with nuclear covariance data. Varied nuclear data are generated randomly corresponding to the uncertainty information from the covariance matrices. After performing a large number of calculations with these data, the results are statistically evaluated; this can be done not only for integral, but also for local output quantities like the assembly power distribution. The method is applied to multi-group Monte Carlo calculations stationary states of the PWR MOX/UO2 core transient benchmark, and to corresponding nodal diffusion calculations. Unexpectedly large uncertainties result for the radial power distribution. The uncertainties in the nodal results agree very well with those in the Monte Carlo reference results; thus, it is possible to apply the random sampling method to determine the influence of nuclear data uncertainties on transient core calculations.