摘要
核电厂周边大气放射性核素监测数据变异性强、偏度峰度大、基本不符合正态分布,直接使用各类统计方法会产生较大的误差,有必要对核素样本数据的正态性进行分析。并对非正态分布核素数据序列采用简单变换、Box-Cox变换和Johnson变换等方法逐步实现正态化。经分析,沉降物核素正态性优于气溶胶核素,14 C和总放数据序列正态性优于γ放射性核素。Johnson变换表现出了更强的正态变换能力,对研究核电厂周边大气核素监测数据变异性有一定优势,是数据变异性强的γ放射性核素数据正态变换的理想工具。
The monitoring data of atmospheric radionuclides around nuclear power plants have strong variability,large skewness and kurtosis,and basically do not conform to normal distribution.Direct use of various statistical methods will produce large errors.It is necessary to analyze the normality of nuclide sample data.The non-normal distribution nuclide data sequence is gradually normalized by simple transformation,Box-Cox transformation and Johnson transformation.According to the analysis,the normality of sedimentation nuclides is better than that of aerosol nuclides,and the normality of 14 C and total emission data sequences is better than that ofγradionuclides.Johnson transform shows stronger normal transformation ability,which has certain advantages in studying the variability of atmospheric nuclide monitoring data around nuclear power plants.It is an ideal tool for normal transformation ofγradionuclide data with strong data variability.
作者
王春梅
温莉琴
郑晓荧
林明贵
Wang Chunmei;Wen Liqin;Zheng Xiaoying;Lin Minggui(Fujian Radiation Environment Supervision Station,Fuzhou 350013,China)
出处
《环境科学与管理》
CAS
2024年第11期116-121,共6页
Environmental Science and Management
基金
福建省新型智库项目(项目编号:23MZKB26)
福建省环保科技计划项目(项目编号:2022R017)
2024-2025年度国家环境保护辐射环境监测重点实验室开放基金项目(环辐监[2024]19号)。
