Efficiency of Individual Targeting of a Three-Dimensional ITER Neutron Source in Monte Carlo Reactor Calculations

Abstract

The extension of the known source angular biasing technique by individual targeting for each of the source neutrons on a given arbitrary rectangular target on the reactor first wall, indicating the area of interest in the problem, is considered in search of the possibilities to reduce the time of Monte Carlo calculation for bulky ITER models. The technique is tested in the source unit of the 3D code BLANK implementing the direct Monte Carlo method. A simple cylindrical model of ITER with a rectangular generatrix and a thickness of the side wall of the shielding material (80% stainless steel + 20% H₂O) of 50 cm is calculated. The rectangular target matches the input hole of the equatorial port (2 × 2 m). The model calculations showed that the application of the technique did not result in a noticeable reduction of statistical uncertainty for the total neutron flux and its energy components in the first wall zone in the entire range of the targeting probability owing to the prevailing contributions of albedo neutrons. The error at the back surface of the model side wall was ~0.75 of the reference uncertainty (without targeting) in a targeting probability range of 0.2–0.4, which corresponds to a potential gain in computation time of up to ~2 times. These results yielded general empirical conclusions: the neutron flux estimated using this technique is unbiased, and a certain gain in the computer time, as compared to the calculation without targeting, is possible. The technique is expected to be more efficient in calculation of the neutron flux functionals in the depth of the shielding compositions, where the contribution of albedo neutrons from the first wall is insignificant, and the result is mainly determined by the intensity of the leading group of source neutrons. Since the considered technique is universal, it can be implemented as a useful option in any Monte Carlo code.