哈尔滨工程大学学报2025，Vol.46Issue(4)：667-674,8.DOI:10.11990/jheu.202302011
蒙特卡罗中子物理计算代理模型研究
Surrogate modeling of Monte Carlo neutronics calculation
摘要
Abstract
This work presents a stochastic Kriging surrogate modeling method for Monte Carlo(MC)neutronics cal-culation to address the stochastic effects on the outputs of MC neutronics calculation.The objective is to examine uncertainty quantification for the effective multiplication factor kinf of a spherical fuel element in an infinite medium.Based on the finite calculation results from the neutronics MC codes OpenMC,a stochastic Kriging surrogate model was established to predict the effective multiplication factor in the infinite medium,replacing the need for OpenMC.This model was then used to quantify the uncertainty of the single pebble multiplication factor calculation.The re-sults show that,compared to OpenMC results,the multiplication factor predicted by the stochastic Kriging model has a deviation within±85×10-5,indicating higher prediction precision.The empirical probability distribution of the effective multiplication factor in the infinite medium is highly consistent with the results from OpenMC,and the statistics of the multiplication factor also demonstrate consistent estimation results.The presented model in this study can effectively address the random effect in the output response of MC neutron physics calculations,and the established surrogate model can be applied to the uncertainty analysis in MC neutronics calculations to reduce the calculation cost.关键词
蒙特卡罗中子计算/随机Kriging方法/代理建模/抽样统计不确定性分析/无穷介质下的有效增殖因子Key words
Monte Carlo neutronics calculation/stochastic Kriging method/surrogate model/stochastic sampling-based uncertainty analysis/effective multiplication factor in an infinite medium分类
能源科技引用本文复制引用
王毅箴,王迎飞,崔梦蕾,郭炯,张汉,李富..蒙特卡罗中子物理计算代理模型研究[J].哈尔滨工程大学学报,2025,46(4):667-674,8.基金项目
国家自然科学基金项目(12405203). （12405203）
