Application of advanced Rossi-alpha technique to reactivity measurements at Kyoto University Critical Assembly

Abstract

This study presents the first application of the advanced Rossi-alpha method (theoretically introduced by Kong et al., 2014) on the reactivity measurements in a research reactor: detector count signals at the Kyoto University Critical Assembly (KUCA) facility. The detector signals in the KUCA A-type core are analyzed by three subcriticality measurement methods: (1) Feynman-alpha (F-α) method, (2) Rossi-alpha (R-α) method, and (3) advanced Rossi-alpha (advanced R-α) method. Four cases are analyzed for two different subcritical states of the core and two different neutron source locations. Two different negative reactivity ρ values are obtained by the measurements of control rod worth and regarded as the reference reactivity values, comparing the results by the four methods.The F-α shows reactivity errors ranging between 7.1 and 7.3% due to its use of variance-to-mean ratios of detector count signals, which are not very sensitive to neutron background noise. However, the fitting uncertainties associated to the F-α results are large, ranging between 5.4 and 12.8% at one standard deviation. The R-α shows small fitting uncertainties ranging between 2.8 and 3.8%, although reactivity errors are in the range of 3.5–26.5% due to the neutron background noise. Finally, the advanced R-α that explicitly models the neutron background noise contrary to the previous methods shows the reactivity errors in the range of 1.0–11.8%, and provides the lowest uncertainties of the measured ρ in the range of 0.4–0.9%. In conclusion, among the four methods applied to the reactivity measurements at KUCA, the advanced R-α reveals the best accuracy with the lowest uncertainties.