测量不确定度有多种评定模型,采用不同的评定模型可能得到不同的结果。为了保证求得的不确定度数值有足够的精度,必须对不同模型计算出的测量不确定度进行有效的验证。文中首次研究了测量不确定度评定模型的验证,该方法首先通过埃奇沃思级数展开来表示出测量数据的分布函数,然后由蒙特卡罗模拟法产生大量符合此分布函数的测量数据的模拟值,计算出该模拟值的标准差作为不确定度评定的验证值,从而实现对各种不确定度评定模型的验证。最后通过多个宴例分析说明了该方法的右效件.
There are many models used in measurement uncertainty evaluation. The verification of uncertainty evaluation is necessary in order to ensure the accuracy of the measuring results. A novel verification method of the measurement uncertainty evaluation is described. The distribution function of the measuring data is expressed by Edgeworth expansion firstly. A large number of simulative measuring data according with this kind of distribution is then generated by Monte Carlo simulation. The standard deviation of these simulative values is calculated, and which is considered as the valid value of the uncertainty evaluation. Therefore the verification of the uncertainty evaluation model is realized. Finally, the rationality and validity of the applied methods is compared and confirmed by various measurement examples.