随机加权法在可靠性参数评估中的应用研究

介绍了随机加权法的基本思想,并将其应用于正态分布小样本的参数置信限估计。采用MonteCarlo仿真方法,验证了估计准确度与样本量之间的关系,指出了随机加权法的适用范围。结果表明,对于样本均值的置信下限估计,在样本量大于一定值时,随机加权法的估计准确度能够达到置信度的要求,而对于样本方差的置信上限估计,准确度较低,因此随机加权法不适用于方差的置信上限估计。算例证明,随机加权法能够在满足置信度要求的基础上,得到更加合理的均值的置信下限估计,与传统方法相比,优势明显。

Research of Reliability Assessment Based on Random Weighting Method

The basic idea of random weighting method is introduced,and the method is used to evaluate the reliability parameters of small sample in normal distribution.By Monte Carlo simulation,the connection between accuracy and sample size is validated,and the applicable range is given.Results include that,for the confidence lower limit of sample mean,when the sample size is large enough,the accuracy can reach the request of the confidence.But for the confidence upper limit of sample variance,the accuracy is too low to accept,thus the method is inapplicability here.An example shows that,the random weighting method can get a more logical confidence lower limit of sample mean,and the advantage compared to the classical method is evidence.