Bootstrap方法在小样本随机振动环境测量数据归纳中的应用研究

Application Research of Bootstrap Method in Small Sample Measured Data Reduction of Random Vibration Environment

针对传统正态容差上限方法在确定平台振动环境条件时,不适合处理小样本实测数据统计归纳的问题,提出了利用Bootstrap法对测量样本重采样,进而统计归纳振动环境条件的计算方法。首先利用Bootstrap法对来自正态分布的试验数据进行容差上限估计,通过与真值的对比分析,验证各种自助置信区间估计方法的适用性,而后提出使用纠偏百分位法和Bootstrap-t法对小样本振动测量数据进行归纳。最后,分别对某型飞机实测数据进行了统计归纳,并与传统容差上限估计方法进行了对比分析。

Traditional normal upper tolerance limit method is not suitable for statistical reduction of measured small sample data in determination of vibration environmental condition of platform.A calculation method to statistically reduce vibration environmental condition by using Bootstrap method resampling on measured specimen.Upper tolerance limit of test data with normal distribution was estimated using Bootstrap method.The applicability of confidence interval estimation methods was verified through comparison with true values.Bias-corrected bootstrap method and bootstrap-t method were put forward for statistical reduction of measured small sample data.The upper tolerance limits of vibration were estimated and compared by the above method and the traditional computation methods using flight measurements of some aircraft.