分层性物质的组合取样误差与份样数目之间的关系及其Monte Carlo模拟

本文从理论上研究了分层性物质的组合取样误差与份样数目之间的定量关系，表明份样数目与组合取样相对标准偏差的平方之积为一常数，称为组合取样常数，它与Ｉｎｇａｍ ｅｌｌｓ 和Ｓｗ ｉｔｚｅｒ 提出的取样常数是相似的，表示６８％ 置信度下组合取样的相对标准偏差达到１％ 时的份样数目。应用Ｍｏｎｔｅ Ｃａｒｌｏ 模拟技术对该理论进行了验证，结果令人满意。

Investigation on the Relationship between Composite Sampling Error and the Number of Increments for Chemical Analysis of Stratified Materials with Monte Carlo Simulation Technique

The theoretical study was carried out for the relationship between composite sampling error and the number of the increments for chemical analysis of stratified materials. It reveals that the product of the number of increments and the square of the relative standard sampling deviation is a constant, which is named as the composite sampling constant. Monte Carlo simulation technique was used to testify the above theory. The stratified populations studied in this paper were Gaussian , uniform random and multi nomial (taking silicon carbide material as an example) distributions. The results were found to be similar for the different populations.