基于k-means聚类的核电荷数甄别方法研究  

Identification Method for Nuclear Charge Based on k-means Clustering

作　　者：杨凤虎 黄智晟 师全林[1] 李志明[1] 徐江[1] 李雪松[1] 解峰[1] 张小林[1] 梁建峰[1] 余功硕[1] 关清帝 王春杰 余青江 康泰 张云鹤 姜文刚[1] YANG Fenghu;HUANG Zhisheng;SHI Quanlin;LI Zhiming;XU Jiang;LI Xuesong;XIE Feng;ZHANG Xiaolin;LIANG Jianfeng;YU Gongshuo;GUAN Qingdi;WANG Chunjie;YU Qingjiang;KANG Tai;ZHANG Yunhe;JIANG Wengang(Northwest Institute of Nuclear Technology,Xi’an 710024,China;Department of Engineering Physics,Tsinghua University,Beijing 100084,China)

机构地区：[1]西北核技术研究所,西安710024 [2]清华大学工程物理系,北京100084

出　　处：《现代应用物理》2024年第6期50-57,共8页Modern Applied Physics

摘　　要：核工业和裂变理论研究的快速发展,对准确、可靠的裂变碎片核电荷数分布数据的需求不断提高。然而,至今仍没有可靠的方法测定裂变碎片核电荷数分布。为了克服这一难题,提出了利用电离室脉冲波形进行聚类,实现粒子核电荷数甄别的方法。用已知核素(127I,128Xe,129Xe,134Xe和133Cs)不同入射能量粒子在电离室中输出的波形作为数据集,采用过阈定时方法排除粒子入射能量对波形的干扰,然后通过k-means方法进行聚类,成功实现了83%的甄别正确率,优化处理后正确率达到99%,该方法有助于裂变碎片核电荷数分布数据的测量。With the rapid development of nuclear industry and fission theory,the demand for accurate and reliable nuclear charge distribution data of fission fragments is increasing.However,there is still no reliable method to determine the nuclear charge distribution of fission fragments.To overcome this problem,a new method to identify the nuclear charge of particles is proposed in this paper by clustering the pulse waveforms of ionization chamber.The waveforms of the incident particles(127 I,128 Xe,129 Xe,134 Xe and 133 Cs)in the ionization chamber is chosen as the data set,the proper data pre-process and k-means clustering are used to realize 83%discrimination rate,and 99%after optimization.This method is helpful for measuring the nuclear charge distribution of fission fragments.

关 键 词：K-MEANS聚类 波形分析 核电荷数甄别 低气压屏栅电离室 

分 类 号：O571[理学—粒子物理与原子核物理] TL816[理学—物理]