Uncertainty analyses of neutron noise simulations in a Zero-Power reactor

Artikel i vetenskaplig tidskrift, 2022

A comprehensive uncertainty analysis methodology has been established for the modeling of stationary neutron flux oscillations induced by fuel rods vibration in a zero-power reactor. The methodology includes uncertainty propagation and sensitivity analysis. The target event is based on an actual experimental campaign at the CROCUS zero-power reactor and corresponds to the simultaneous oscillation of 18 metallic uranium fuel rods in the periphery of the core. Both the uncertainty propagation and the sensitivity analysis commonly use a large part of the entire analysis process, from the selection of uncertain parameters to the actual code simulations. Applying a random sampling-based approach, the input parameters are sampled N times from their distribution information and used as inputs for N noise simulations using CORE SIM +. The quantity of interest (QoI) is the amplitude of the Auto-Power Spectral Density at various detector locations, which is normalized by the amplitude of the Cross-Power Spectral Density of the reference detector. Their uncertainties are determined following the 4th order Wilks’ formula for two-sided limits. Through the determination of correlations among QoI at the installed detector locations, it is demonstrated that the neutron noise near the area of oscillating fuel rods (noise source) have different behavior compared to the neutron noise further away from the noise source. The following sensitivity analyses are carried out using multiple correlation coefficients within grouped parameters. As expected from the QoI correlations, the QoIs at two different locations (near and far from the noise source) are influenced by different input parameters. Near the noise source, the QoI uncertainty is driven by the uncertainties in the position of the noise source, while the uncertainties in the nuclear data for U-235 and U-238 are the leading contributors further away from the source. This paper provides general information on how to perform the uncertainty analyses for neutron noise simulations, as well as quantitative estimates of the computational uncertainty required for the validation of the computer programs under development for the simulation of neutron noise.