Small sample reactivity worths calculation exact perturbation theory and Monte Carlo transport

International audienceThe goal of this work is to calculate with a decent precision and accuracy the effect induced by the oscillation-or the substitution-of a pure sample of material at the center of a critical nuclear reactor core. The difference between the calculated and measured reactivity worth might hopefully be used to feedback nuclear data. Frequently, Monte Carlo transport codes fail in that task due to their inefficiency to reach convergence on such small reactivity differences (e.g. measured effects are sometimes lower than 1 pcm). Various deterministic and stochastic approaches have been used in the past, but they failed to provide unquestionable results for all isotopes. In this paper, we present a new methodology that allows exact calculation (i.e. not limited to first order) of these effects using Monte Carlo Neutron Transport codes. This method has several advantages: (1) It does not require significant code developments and it can be applied with almost any standard Monte Carlo codes. (2) It does not relied on any assumptions other than standard assumptions made by these codes: thus, it can be used to calculate any kinds of perturbation (e.g. concentration, geometry, medium, cross-sections) with a variable degree of effectiveness; still, it converges to the true value obtained from the difference of two independent runs. (3) In contrary of direct Monte Carlo perturbation technics, the more perturbations are small and locally contained, the more it performs. Validation cases confirm the numerical good behavior of the method named Black Body Exact Perturbation