基于地统计空间插值的三维辐射场重构研究  

Reconstruction of 3D Radiation Field Based onGeostatistical Spatial Interpolation

作　　者：龙泽宇 宋佩涛 冯青靓 李华 韦加富 刘瑞文 牛瑨博 Long Zeyu;Song Peitao;Feng Qingliang;Li Hua;Wei Jiafu;Liu Ruiwen;Niu Jinbo(Shanxi Key Laboratory for Radiation Safety and Protection,China Institute for Radiation Protection,Taiyuan 030006;Key Laboratory of Radiation Environment&Health of the Ministry of Ecology and Environment,China Institute for Radiation Protection,Taiyuan 030006;Research Center for Nuclear and Radiation Frontier Technology,China Institute for Radiation Protection,Taiyuan 030006)

机构地区：[1]中国辐射防护研究院辐射安全与防护山西省重点实验室,太原030006 [2]中国辐射防护研究院国家环境保护辐射环境与健康重点实验室,太原030006 [3]中国辐射防护研究院核与辐射前沿技术研究中心,太原030006

出　　处：《辐射防护通讯》2023年第5期18-24,共7页Radiation Protection Bulletin

基　　金：预先研究项目(32104040303);中核集团菁英项目(No.29)。

摘　　要：针对源项未知情况下的三维辐射场获取问题,根据辐射场具有空间自相关性的特点,利用地统计空间插值中结合对数变换的普通克里金方法来进行辐射场重构。通过两个实测的辐射场数据集对该方法进行了测试,结果表明,该方法在选取Exponential模型或Spherical模型来进行半方差拟合,且测量点为随机选取的情况下,其重构结果平均偏差的平均值为21.51%~74.01%,证明了该方法在实际应用中具有一定的有效性和可行性。另外,对测量点分布情况与辐射场重构准确性之间的关系进行了初步分析,确定了测量点的分布应尽可能均匀的选取方案。In this paper,according to the spatial auto-correlation characteristics of radiation field,the general Kriging method in geo-statistical spatial interpolation combined with logarithmic transformation is used to reconstruct the radiation field under the condition of unknown radioactive source.The method is tested through two measured radiation field data sets.The results show that when the Exponential model or Spherical model is selected for semi-variance fitting and the measuring points are selected randomly,the mean deviation of the of the reconstructed result ranges from 21.51%to 74.01%,which proves that the method is effective and feasible in practical application.In addition,the relationship between the distribution of measuring points and the accuracy of reconstructed radiation field is preliminarily analyzed and studied,and the selection scheme that the distribution of measuring points should be as uniform as possible is determined.

关 键 词：三维辐射场 地统计插值 克里金方法 

分 类 号：R144[医药卫生—公共卫生与预防医学]