小样本感度试验方法

提出了一种用于炸药及火工品感度试验的小样本方法。此方法较之传统的Ｂｒｕｃｅｔｏｎ升降法（估计方差一般需要３０ 个以上的试件，而且估计出的方差往往偏小），可以在少于３０ 个试件的小样本的条件下，对方差进行估计，而且估计精度高。建立了高置信度、高可靠度的Ｐ发火刺激量置信限公式，该公式可以综合利用当前试验数据和以往经验数据，从而仅需少量当前试验，便可达到Ｂｒｕｃｅｔｏｎ 升降法采用３０ 个以上试件方能达到的精度，大大节省了试验的费用及时间。大量的ＭｏｎｔｅＣａｒｌｏ 模拟试验及实例分析都进一步证明了上述结论。

A Method for Sensitivity Test with Small Sample

A small sample method for sensitivity test of explosive and initiating explosive devices is provided. This method gives a better estimate of variation with a specimen <30, while the traditional Bruceton method needs more than 30 specimens, moreover, the estimate is usually lower than that of the population variation. An estimate of p firing level of stimulus with high confidence and high reliability is given, which can make full use of both current data and past ones. So the information which could be used is much more than that of Bruceton method. As a result, fewer specimen are needed and the result is more precise than the Bruceton method which can only make use of the current test data. Finally, a Monte Carlo study and a contrast test example are given.