Scouting the feasibility of Monte Carlo reactor dynamics simulations

Usual space dependent reactor dynamics simulations still rely on heavily simplified transport modelling, often restricted to two energy groups prepared with considerably debatable presumptions. The radical increase in computing power availability of recent times allows for more detailed modelling, more energy groups, better physics treatment or for choosing Monte Carlo (MC) methods. Developing MC methods for reactor dynamics are not supported up to now due to the lack of confidence of feasibility for realistic cases fearing unpractical computing times, but the capability of easy parallelization of the MC computing flow may offer hope of real applications. Our investigations here treat the options for performing and optimizing dynamic MC calculations including thermal feedback with the aim of determining the necessary computer resources for realistic simulations. We focus here on methodological questions as previous papers on the subject did not elaborate on such details.