Sodium-Cooled Fast Reactor Severe Accident Modeling in Zero-Power Environment Based on the Representativity Theory

This project deals with first representativity studies of local reactivity effects in a fast lattice during the progression of severe core accident (SCA) in Sodium-cooled Fast Reactors (SFRs), with the aim of implementing new experimental programs in the French ZEPHYR ZPR project. The representativity study is based on sensitivity analysis of reactivity coefficients and local flux distributions against core degradation of the CFV type SFR inner core. The SCA sequence, e.g., voiding, fuel meltdown, molten pool formation, etc., might have pronounced influence on the neutronic characteristics of the fast reactor core, and can lead to prompt/superprompt recriticality. To predict the core behavior during such disruptions, it is necessary to develop accurate computational and experimental tools and methodologies. The assessment of reactivity behavior characterizing different stages of the representative disrupted core configurations is one of the challenges assigned to ZEPHYR. This requires a new methodology to be implemented in critical assemblies. The project is carried out in two steps. The first step treats the sub-assembly level of the CFV core, which undergoes three stages of degradation. The sub-assembly level evaluation is carried out in order to examine the translation of temperature effects in power systems into density effects in zero-power environment. By adopting an optimization method based on the particle swarm optimization (PSO) and the estimation of sensitivity coefficients made by Serpent Monte Carlo code, it is possible to identify highly representative configurations that could be loaded into the ZEPHYR reactor. However, the assembly calculations do not provide full information of the core behavior under the different SCA stages. Therefore, an extension of the previous assembly calculation is mandatory to provide a more realistic estimation of core behavior. Full core calculations of both the CFV type SFR and the ZEPHYR correspond to the second stage of the project, which requires particular modifications to the proposed methodology. As the core calculations require a longer computational time, a more robust optimization scheme is selected. The optimization based on the Nelder-Mead Simplex method was selected for this purpose. The algorithm, based on iterative Serpent MC calculations, estimates the representativity factor and locates the maximal value of the representativity factor in the search space. The presentation will provide an overview of the entire process.