基于置信限的武器性能试验方法

Test Methods of Weapon Performance Based on Confidence Limit

针对国军标中对不同质量水平装备的性能考核通常采用定数样本,用试验数据均值与指标对比判定是否合格,存在未给出试验结论的置信度以及不能体现装备性能优/劣引起的试验成本差异等问题。在分析传统假设检验方法、置信限评估方法和序贯检验方法的基础上,应用置信检验理论,提出了一种基于正态分布的装备性能置信检验序贯试验方法。该方法以“大概率”判定装备性能是否合格,避免了传统假设检验仅能给出“小概率”拒收或不拒收时却没有充分理由接受的尴尬,以及国军标定数样本一试到底的弊端,尤其对性能明显优于或低于指标,能提高试验效费比;并以试验仿真数据为例,诠释该方法的优势。

Aiming at the problem that the performance of different quality equipment are examined by using a fixed number of samples in the national military standard,and the problem whether the performance is qualified by comparing the test data mean value with the indicators,some problems still remain that the confidence degree is not given,and the test cost difference caused by good or bad performance is not reflected.The sequential test method based on the normal distribution of equipment performance confidence test was proposed by analyzing the traditional hypothesis testing method,confidence limit evaluation method as well as sequential testing method and by applying the confidence test theory.This method uses"high probability"to determine whether the equipment performance is quali-fied,which avoids the embarrassment that traditional hypothesis test can only give"low probability"as to whether to reject or not to reject,but has no sufficient reason to accept,as well as the disadvantage of one try to the end for the fixed sample test in the national military standard.The test cost-effectiveness ratio is improved for the equipment whose performance is significantly superior to or below the indicator.Finally,the advantages of this method are illustrated by experimental simulation data.