Comparison of two modelling tools for the evaluation of tritium and hydrogen transfers in nuclear reactors or complex systems

Abstract In the framework of the TRANSAT project, started in 2017 (TRANSversal Actions for Tritium supported within the H2020 Euratom program), the ability to measure and to assess the tritium inventory and migration within different kinds of reactors or processes is one major challenge to control the potential releases and personal dosimetry in nominal operating conditions. A benchmarking activity between two calculation tools developed either for the fusion machines (EcosimPro developed between CIEMAT and EAI) or for the fission reactors (KUTIM developed at CEA), was initiated to improve the level of confidence in the tritium and hydrogen balances estimated by such codes in complex systems. An application to a conceptual fission reactor based on the technology of Sodium Fast Reactors was carried out on the base of a shared data set used for the design characteristics of circuits and components as well as for physico-chemical properties of liquid sodium, of wall materials and of operating conditions. The modeling of specific components or physical equilibriums involved in sodium cooled reactors was adapted and implemented in EcosimPro libraries. In particular the major recovery of tritium and hydrogen impurities in purification systems (cold traps) is evaluated as well as the dissolution equilibrium of both hydrogen isotopes at the primary sodium surface between the liquid metal and the gas plenum. Calculations results with both codes are presented and compared in terms of tritium activities and hydrogen concentrations in each circuit of the reactor, but also in terms of transfer fluxes and releases. In addition, the distribution of the various transfer contributions (between primary, secondary and tertiary circuits) is analyzed. Comparisons between both codes results associated with sensitivity studies on the influence of main parameters (such as tritium and hydrogen source terms, temperature profiles) are also used to propose perspectives of modeling improvements.