ANALYSIS OF CRITICAL AND SUBCRITICAL NEUTRON NOISE EXPERIMENTS IN MINERVE USING ADVANCED NOISE TECHNIQUES

摘要

The Cohn-α and Feynman-Y methods are implemented to analyze critical and subcritical configurations of the MAESTRO core in the MINERVE zero power reactor in order to measure its integral kinetic parameters, i.e. effective delayed neutron fraction β_(eff) and the prompt neutron generation time A, in addition to the absolute reactivity ρ. Various methods are studied and implemented to obtain the variance-to-mean ratio curves and the cross-correlation power spectral densities (Cohn-α) of the detector's readings. Regarding the Cohn-α method, it is shown that the obtained kinetic parameters are highly sensitive to the numerical parameters used in the power spectra calculations. Despite their pronounce effect, very little considerations are usually given to their values, which are often determined rather arbitrarily. Generally, well-defined methodologies for tuning these parameters is overlooked. In this paper, a new methodology is proposed for analyzing the kinetic parameters' sensitivity to the power spectra calculations and for fine tuning of its numerical parameters in order to evaluate and reduce the associated uncertainty. Regarding the Feynman-Y method, the fitting models include a single-mode prompt reactivity model and a multi-mode delayed reactivity model. A direct derivation of the integral parameters (ρ and β_(eff)) is performed using a novel approach, in which the parameters are simultaneously considered as degrees of freedom within the curve fitting procedure. This novel approach enables a simultaneous direct estimation of β_(eff) and the absolute reactivity p, independent of prior reactivity calibrations, e.g. rod-drop or stable period analysis.