基于顺序统计量的加速度计恒加寿命试验优化设计

Optimization design method of accelerometer’s constant-life test based on sequential statistics

针对模型参数先验信息未知下的加速度计恒加寿命试验优化设计方法缺乏的问题,提出了一种加速度计恒加寿命试验的优化设计方法。首先,分析了加速度计的加速寿命预实验结果,并初步制定加速度计的加速寿命试验方案;而后,基于顺序统计量和整体线性无偏估计理论,以每个应力水平下的样本分配为设计变量,同时以加速度计在正常应力水平下中位寿命估计值的方差最小和模型参数的协方差阵行列式最小为优化目标,建立加速度计定数截尾恒加寿命试验优化设计的数学模型。模型计算结果表明,两种优化目标的求解结果一致,确定加速度计的试验方案为(8|4,2,2),即8个样本在应力水平S1、S2、S3下分配的样本量分别为4、2、2,验证了所提出方法的可靠性。

An optimization design method for accelerometer’s constant life test is proposed to overcome the problem of lacking the optimization design method when without prior information of model parameters.Firstly,the accelerometer’s accelerated life pretest results are analyzed,and the accelerometer’s accelerated life test scheme is preliminarily established.Then,based on the sequential statistics and the overall linear unbiased estimation theory,and taking the sample distribution at each stress level as design variables,a mathematical model for optimization design of accelerometer’s constant-life test with fixed truncation is established based on the optimization objective of minimizing both the variance of the accelerometer’s estimated median life and the covariance matrix’s determinant of the model parameters under normal stress.The model calculation results show that the results of the two optimization targets are consistent,and the experimental scheme for the accelerometer is determined as(8|4,2,2),that is,the allocation of 8 samples under the stress level S1,S2,S3 are 4,2,2,respectively.The reliability of the method is verified.