Fonctionne avec Indico
v3.2.8
10h: Mohammad Rababah
Computation of Sensitivity in Monte-Carlo codes: Investigation of Convergence for Reactor Cores with High Dominance Ratio
Estimation of sensitivities plays a crucial role in reactor core design and safety assessments, with a comprehensive understanding of associated uncertainties being essential for ensuring result reliability and establishing design margins for improved operational safety. The utilization of Monte Carlo-based codes for sensitivity computation has expanded notably with the integration of perturbation theory and the advancements in computation power. Despite the method's growing application, research on the statistical convergence of sensitivities remains relatively limited. In this presentation, we highlight the challenges associated with sensitivity computation, such as the statistical convergence of sensitivity in large nuclear reactor cores, and we present our approach to tackle the issue. We provide different methods employed to compute sensitivities and we present our findings from simulating various reactor configurations, highlighting the factors impacting the statistical convergence of sensitivities, and proposing a method to accelerate convergence and lower computation cost.