最大值指标截尾正态分布精度换算方法

Precision conversion methodology with truncated normal distribution theory assumption oriented to maximum-error specification

提出了面向最大值指标的截尾正态分布精度换算方法,为最大值指标与常用精度指标间的精度换算以及真值测量系统精度指标的确定提供了参考依据。该方法假设系统输出序列中各观测点的合格概率服从对数截尾正态分布;根据给定最大值指标的置信水平及序列样本量,证明并推导了截尾正态分布之截尾上限、截尾下限、均值及标准偏差的计算公式,导出了最大值精度指标与1σ等常用精度指标间的换算关系;结合精密仪器有关理论给出了最大值指标下真值测量系统精度指标的确定方法。实例应用的实验结果表明,该方法是可行的。

A precision conversion methodology with truncated normal distribution theory assumption oriented to maximum-error specification was brought forward, and it could be taken as a reference frame for the precision conversion between maximum-error specification and other precision measurement specifications, so that the precision class of according true value measurement systems could be determined in advance. The method assumes that the conformity probability of the observation sequence is subjected to logarithmic truncated normal distribution ; based on the aimed confidence level for maximum-error specification and the given sample size of target sequence, the calculation formulation of upper truncated limit, lower truncated limit, mean and standard deviation of the truncated normal distribution were proved and derived, thus the precision conversion relationships between maximum-error specification and other precision measurement specifications, such as 1 tr, were turned out; through referring to the corresponding theories on precision instrument fields, the determination methodology for precision class of true value measurement systems under maximum-error specification was given. The application on related example eases proved the feasibility of the proposed method.