The Optimal Source Bias Method based on RMC code

Abstract

When criticality calculation is performed with Reactor Monte Carlo (RMC) code for reactor analysis, the fission source iterative method is selected. There are two strategies to sample the fission neutrons in each fissile region: the conventional Proportional Stratified Sampling Method that focuses on the real physical process, and the author-developed Optimal Stratified Sampling Method that focuses on the optimal mathematics treatment. The Optimal Stratified Sampling Method have the effect of global variance reduction and could be easily implemented with the use of optimal source bias coefficients. However, the calculation of optimal source bias coefficients needs the values of variances in advance, which is almost impossible. To meet the requirement, a Batch Approximate Method is then put forward. The Batch Approximate Method based on the Batch algorithm and Shannon Entropy Diagnosis, eliminates the correlation among cycles during the criticality calculation. Combining the Optimal Stratified Sampling Method and the Batch Approximate Method, the Optimal Source Bias Method is established and developed in RMC code. The Optimal Source Bias Method was tested with a standard Pressurized Water Reactor assembly. The effect of global variance reduction was found, as well as the flattening of variance distribution. So, the new method is valuable for the high fidelity computation of the RMC code.