测量不确定度最大残差系数的一种新算法  

A Novel Method for Calculating the Coefficient of Maximum Residual in Measurement Uncertainty

作　　者：王中宇[1] 张海滨[1] 刘智敏[2] 

机构地区：[1]北京航空航天大学仪器科学与光电工程学院,北京100083 [2]中国计量科学研究院,北京100013

出　　处：《计量学报》2006年第3期201-205,共5页Acta Metrologica Sinica

基　　金：国家自然科学基金(50375011)

摘　　要：在测量不确定度的评定中,用测量数据的最大残差乘以适当的系数获得标准差的方法称为最大残差法。通过对最大残差的分布函数及其数字特征的分析研究,建立了最大残差的概率模型,并且利用蒙特卡罗模拟及Matlab软件,提出了计算测量不确定度最大残差系数的一种新方法,求出了当测量次数小于50时,最大残差所对应的分布函数、均值、标准差和自由度的数值,并给出了最大残差法的系数表。最后通过测量实例验证了理论分析的正确性。用文中提出的方法可简单、迅速、可靠地计算出所需要的标准差数值。The method of maximum residual error for measurement uncertainty can be used to calculate the standard deviation in a simple, rapid and reliable way. The general method of maximum residual error and its distribution are analyzed . The probability model of the maximum residual error is also built. A novel method of calculating the coefficient of maximum residual error is introduced by using the Monte Carlo simulation and Matlab software. Calculations of the distribution function and the numerical characteristics of maximum residual are also developed. The coefficients of maximum residual error are calculated when the measuring number is less than 50. Finally some practical examples are given and the correctness of the applied methods is validated by the measurement experiments.

关 键 词：计量学 测量不确定度 标准差 最大残差 蒙特卡罗模拟 

分 类 号：TB9[一般工业技术—计量学]