摘要
反应堆生物屏蔽层的作用是将反应堆运行过程中产生的各类电离辐射降低到人员安全所允许的水平。对于反应堆生物屏蔽层的设计而言,通常需要在保证剂量安全的条件下找到满足重量最轻、结构稳定等设计目标的屏蔽方案。然而现有的两步式屏蔽设计方法需要花费大量时间成本进行屏蔽计算,极大地限制了反应堆生物屏蔽层的设计效率。本文提出基于神经网络算法的一步式设计方法——由辐射源参数与剂量率限值参数直接获得符合设计要求的屏蔽结构参数。通过本文研究,构建了以离散化辐射源参数和剂量限值为输入参量、屏蔽层重量为输出参量的神经网络模型,实现了反应堆生物屏蔽层的一步式设计,在保证屏蔽优化效果的同时实现了对反应堆生物屏蔽层设计效率的显著提升。
The main function of the reactor biological shielding layer is to reduce all kinds of ionizing radiation generated during the operation of the reactor to the allowable level of the personnel safety.For the design of reactor biological shielding layer,it is usually necessary to find a shielding scheme that satisfies the requirements of lightest weight and stable structure under the condition of ensuring the safety of dose.However,the existing two-step shielding design method requires a lot of time cost for shielding calculation,which greatly limits the design efficiency of reactor biological shielding layer.In this paper,a one-step design method based on neural network algorithm is proposed to obtain the shielding structure parameters that meet the design requirements directly from the radiation source parameters and dose rate limit parameters.Through the research in this paper,a neural network model with discrete radiation source parameters and dose limits as input parameters and shielding layer weight as output parameters is constructed,and the one-step design of the biological shielding layer of the reactor is realized.While ensuring the shielding optimization effect,the design efficiency of the reactor biological shielding layer is significantly improved.
作者
余鸿
吕焕文
应栋川
唐松乾
翟梓安
温兴坚
苗建新
Yu Hong;Lv Huanwen;Ying Dongchuan;Tang Songqian;Zhai Zian;Wen Xingjian;Miao Jianxin(Science and Technology on Reactor System Design Technology Laboratory,Nuclear Power Institute of China,Chengdu 610213,China)
出处
《电子技术应用》
2024年第S01期230-235,共6页
Application of Electronic Technique
关键词
生物屏蔽
一步式
神经网络
效率提升
reactor biological shielding
one-step design
neural network algorithm
efficiency promotion
作者简介
余鸿(1998-),硕士,主要研究方向:反应堆屏蔽设计工作。
