Convergence assessment of the MCNP5 model for NUR research reactor

Abstract

In recent years, the Monte Carlo (MC) methods have become an extensive and reference tool for criticality calculations. Despite the high precision that could be reached in MC modeling, the criticality calculations are still faced with the problem of fission source convergence leading, thereby, in some cases to false convergences and, in particular, to an erroneous effective multiplication factor. This is a challenging issue from a criticality safety point of view. Thus, to prevent from these specific issues and to obtain correct results, a convergence assessment has been carried out, for the first time, in this work, to the MC model of the NUR research reactor initial core. To this end, both effective multiplication factor (keff) and Shannon entropy of the fission source (Hsrc) were evaluated for various M, the number of neutrons per cycle, and for different locations of the initial fission source in the reactor core. From the obtained results, it reveals that: a number of 10000 neutrons/cycles (or more) are needed to reduce bias on keff and at least 30 inactive cycles are required to allow keff convergence. Whereas, the choice of a point source in each fuel element or in each fuel plate are the most preferred assumptions. As for the number of active cycles, it can be fixed according to the reasonable uncertainty that a user may accept for the effective multiplication factor, keff. ©