Speaker
Description
Evaluating nuclear data reduces the information of a set of experimental data and theoretical models to provide the best estimate of a physical quantity. In the last decade safety applications pushed the evaluation community toward better quantification of the uncertainties associated to evaluated nuclear data. The Bayesian inference appears as a standard and versatile mathematical framework to estimate such uncertainties. As a matter of fact, it is now a widespread tool to evaluate neutron cross sections, especially in the resolved resonance region [1]. On the other hand, the current JEFF evaluations related to the fission cross sections at higher energies as well as the neutron multiplicities and prompt neutron spectra do not often rely on a rigorous application of a Bayesian inference and if so only within some approximated treatments [1,2]. In a long term objective to better ground the determination of uncertainties in evaluations, we explore the feasibility of an exact Bayesian inference of the nuclear data associated to fission.
In this presentation, I will emphasize our preliminary attempts to perform an exact Bayesian inference to evaluate the total neutron induced cross section of an actinide in the continuum energy range as well as on the prompt neutron multiplicity and spectrum. I will emphasize our current results and conclude on the difficulties and challenges of such developments.
[1] C. DE SAINT JEAN, P. ARCHIER, E. PRIVAS, G. NOGUÈRE, B. HABERT, P. TAMAGNO, Nuclear Data Sheets 148, 383-419 (2018)
[2] D. ROCHMAN, E. BAUGE, A. VASILIEV, H. FERROUKHI, S. PELLONI, A.J. KONING, J.Ch. SUBLET, European Physical Journal Plus 133, 537 (2018)