A batch size selection strategy by checking stability of fission source distribution for Monte Carlo eigenvalue calculations

Abstract

ABSTRACTThe power iteration method is generally used for Monte Carlo eigenvalue calculations. In the process, the fission sources at each cycle are not independently sampled by the iteration scheme. This dependency of the sampled fission source can cause underestimation of response uncertainty. To solve these issues, previous studies have recommended the selection of large-size batches, which nonetheless, would require large calculation cost. This study proposes a method for selecting a reference criterion of batch size avoiding significant biases in the local responses, and mainly emphasizing the maintenance of the stability of fission source distribution. To determine the criterion for the diagnostics of stability, real-to-apparent standard deviations are analyzed with the stability of fission source distribution. The results show that the batch sizes determined by the proposed method allow the evaluation of local responses without large biases. The proposed method is expected to contribute to improving the reliability of the Monte Carlo criticality calculations.