Simulation approach and code functionality assessment using deterministic and Monte Carlo codes system for U-235 core-follow depletion calculation

Abstract

Improving the performance of reactor simulation codes and adequately predicting the fuel composition are crucial to increase the accuracy of simulations and the quality of the analysis. McMaster Nuclear Reactor (MNR) follows a strict daily operational and shutdown schedule, except for Sundays. This results in a buildup of neutron poison during the reactor shutdown and hence the Control Rods (CRs) are extracted in the time of the reactor startup. Therefore, tracking MNR core composition using faithful Control Rods (CRs) positions data to estimate the axial fuel inventory profile is not yet established. In this study, nodal diffusion (OSCAR-4) and Monte Carlo (Serpent-2) simulations codes were employed. This study investigates: (1) code-to-code differences in the multiplication factor (keff) and in the axial fuel inventory considering the same burnup step sizes and total energy release with all CRs out; (2) for long cycles of core calculation, Monte Carlo stochastic code is computationally expensive, therefore an adequate temporal discretization steps are studied and identified. The keff values for the fresh reactor when all rods are out is: 1.1339 and 1.1283, and the results at the end of the burnup calculation is: 1.0101 and 1.0049, for Serpent-2 and OSCAR-4, respectively. In addition, the findings showed that averaging the CRs travel distance, for calculation the fuel inventory, can provide results similar to those with most tracked CRs motion.