No abstract

No abstract

A method to judge whether the source iteration in Monte Carlo calculations is converged or not by applying a ''sandwich method'' is described in this report. A procedure is also introduced to determine the number of skip generations in the source iteration satisfying the convergence criterion. Moreover, the effectiveness of this method is demonstrated by calculating using the fresh fuel vault model in the OECD source convergence benchmark problems.

No abstract

The effect of Pu-rich agglomerates in U-Pu mixed oxide (MOX) fuel on a lattice calculation has been demonstrated. The Pu-rich agglomerate parameters are defined based on the measurement data of MIMAS-MOX and the focus is on the highly enriched MOX fuel in accordance with increased burnup resulting in a higher volume fraction of the Pu-rich agglomerates. The lattice calculations with a heterogeneous fuel model and a homogeneous fuel model are performed simulating the PWR 17 Â 17 fuel assembly. The heterogeneous model individually treats the Pu-rich agglomerate and U-Pu matrix, whereas the homogeneous model homogenizes the compositions within the fuel pellet. A continuous-energy Monte Carlo burnup code, MVP-BURN, is used for burnup calculations up to 70 GWd/t. A statistical geometry model is applied in modeling a large number of Pu-rich agglomerates assuming that they are distributed randomly within the MOX fuel pellet. The calculated nuclear characteristics include k-inf, Pu isotopic compositions, power density and burnup of the Pu-rich agglomerates, as well as the pellet-averaged Pu compositions as a function of burnup. It is shown that the effect of Pu-rich agglomerates on the lattice calculation is negligibly small.