基于贝叶斯神经网络方法对^(13)C(α,n)^(16)O反应数据分析  

Data analysis for ^(13)C(α,n)^(16)O reaction based on Bayesian neural network method

作　　者：唐诺程 孙小军[1,2] TANG Nuo-Cheng;Sun Xiao-Jun(College of Physics and Technology,Guangxi Normal University,Guilin 541004,China;Guangxi Key Laboratory of Nuclear Physics and Nuclear Technology,Guangxi Normal University,Guilin 541004,China)

机构地区：[1]广西师范大学物理科学与技术学院,桂林541004 [2]广西师范大学广西核物理与核技术重点实验室,桂林541004

出　　处：《中国科学：物理学、力学、天文学》2024年第11期160-168,共9页Scientia Sinica Physica,Mechanica & Astronomica

基　　金：国家自然科学基金(编号:12065003);广西重点研发计划项目(编号:2023AB07029);桂林科技研发项目(编号:20210104-2);中央引导地方科技发展专项(编号:Guike ZY22096024)资助。

摘　　要：在核天体领域,^(13)C(α,n)^(16)O反应是超铁元素合成过程中非常关键的中子源反应之一,该反应截面会对s-过程和i-过程核素丰度产生十分重要的影响.迄今为止,国际上有多个团队测量该反应的数据,但在质心系能量E_(c.m.)=0.27 MeV以下的数据分歧最高达130%,导致外推存在极大的不确定性.本文基于贝叶斯神经网络方法,通过甄别数据,分别选取两类数据进行训练,系统地对^(13)C(α,n)^(16)O反应在质心系能量E_(c.m.)=0.1-6 MeV区域内的测量数据(共13个团队的测量数据总计2210实验点)进行比对分析.研究结果表明:(1)两种数据处理方式均能很好地重现测量数据,特别是共振峰信息;(2)我国锦屏深地核天体测量平台(JUNA)测量的数据具有更可靠的外推置信度.In the field of nuclear astrophysics,the ^(13)C(α,n)^(16)O reaction is one of the key neutron source reactions for synthesizing super-iron elements.Furthermore,the cross-section of this reaction has a significant impact on the abundance of the sand i-processes.So far,more than a dozen international groups have measured the cross-section data of this reaction.However,the largest difference of the experimental data below E_(c.m.)=0.27 MeV reaches 130%,leading to significant uncertainty of the theoretical prediction in the lower incident energy region.This paper systematically compares and analyzes the measurement data(2210 experimental points contributed by 13 groups)for ^(13)C(α,n)^(16)O reaction in the center-of-mass system energy range E_(c.m.)=0.1–6 MeV using two different data classification methods based on the Bayesian neural network approach.The results indicate that(1)both classification methods can accurately reproduce the measured data,particularly the details of resonance peaks,and(2)the data measured by JUNA in China exhibit a higher degree of reliability and confidence in extrapolation.

关 键 词：^(13)C(α n)^(16)O反应 贝叶斯神经网络方法 反应截面 S因子 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] P14[自动化与计算机技术—控制科学与工程]