COMPARISON OF THE POINT-KINETICS AND QUASI-STATIC METHODS WITH SIMMER-Ⅲ- APPLICATION TO A ULOF CALCULATION

摘要

The neutron kinetics methods used for the calculation of severe accidents of sodium-cooled fast reactors usually differ depending on the phase of the accident. During the primary phase, when the core degradation is limited to the sub-assembly, the point-kinetics model is mostly used. During the secondary phase, as there can be movement of materials in the core, the more precise quasi-static method is generally used. Compared, to quasi-static model, the point-kinetics approach has the important advantage of having a low computational cost. However, it relies on the strong assumption that the flux shape does not change during the transient. This creates discrepancies in the reactivity calculation, and thus in the evolution of the core power, but up to now, point-kinetics has often been used, despite the lack of quantification of these approximations. The goal of this paper is to give an estimation of the error performed when using point-kinetics in spite of the quasi-static method. The SIMMER-Ⅲ code, which features quasi-static kinetics, is used to make this evaluation. To make a correct comparison between both kinetic approaches, an adaptation of a quasi-static model for simulating a point-kinetics calculation scheme has been done with the SIMMER code. This enables to compare the two kinetics approaches without introducing any additional bias in the thermohydraulics or structural mechanics models during the transient. This comparison was performed in the framework of the ASTRID development on an Unprotected Loss Of Flow accident (ULOF) of a sodium-cooled fast reactor. As expected, the results show a good agreement at the beginning of the transient, up to the sodium boiling. Afterwards, an important change in the flux shape occurs, which cannot be taken into account by point-kinetics and leads to rather important discrepancies. Consequently, the results show that point-kinetics is irrelevant once the sodium boiling has begun. Nowadays, as the computational capacities have increased, quasi-static kinetics must be privileged.