摘要
以萨瓦纳船用核动力堆为原型,等比构建了中子-γ混合辐射场多目标优化模型,使用非支配排序遗传算法(NSGA-Ⅱ)与神经网络相结合的屏蔽智能优化方法,将屏蔽层总重量和屏蔽后的剂量率作为优化目标,进行多目标寻优,得到了pareto最优解;选取其中1组最优解分别利用蒙特卡罗方法计算和神经网络预测进行可行性对比验证,在神经网络预测误差允许的范围内,得到的剂量率均满足寻优时设置的约束限值。研究结果表明,该屏蔽智能优化方法对反应堆中子-γ混合射线的屏蔽参数优化是可行的,相比于传统的纯蒙特卡罗方法而言,能在计算准确的前提下极大减少计算时间。
Based on the Savannah marine nuclear power reactor,the multi-objective optimization models of neutron-γmixed radiation were constructed which take the weight of the shielding layers and the dose rate after shielding as the optimization objectives.And the self-developed intelligent shielding optimization method that combines the non-dominated sorting genetic algorithm(NSGA-Ⅱ)and neural network was used for the multi-objective optimization models.Thereafter,pareto-optimal solutions were obtained,and a set of the optimal solutions were chosen to calculate with Monte Carlo method and neural network respectively for feasibility verification.The obtained dose rates all meet the limits within the allowable error of neural network prediction.These results show that the intelligent shielding optimization method is feasible for shielding parameters optimization of the reactor neutron-γmixed radiation,and it can reduce the calculation time compared with the traditional pure Monte Carlo method without reducing calculation precision.
作者
毛婕
宋英明
张泽寰
杨力
韩嵩
赵均
MAO Jie;SONG Yingming;ZHANG Zehuan;YANG Li;HAN Song;ZHAO Jun(School of Nuclear Science and Technology University of South China,Hengyang 421001 China;China Nuclear Power Technology Research Institute,Shenzhen 518026,China)
出处
《原子能科学技术》
EI
CAS
CSCD
北大核心
2021年第5期892-900,共9页
Atomic Energy Science and Technology
关键词
屏蔽设计
中子-γ混合辐射场
非支配排序遗传算法
智能优化
多目标优化
shielding design
neutron-γmixed radiation field
non-dominated sorting genetic algorithm
intelligent optimization
multi-objective optimization
