模糊复合抽检贮存数据的Bayes融合评估方法

目的针对包含模糊样本的复合抽检型产品开展贮存寿命评估。方法针对批次产品中的出厂失效数据和贮存失效数据,开展批次数据的相容性检验。通过出厂失效样本数随机化处理,量化贮存过程中模糊样本的不确定性。将出厂试验数据作为先验信息,贮存过程中的出厂试验数据作为观测信息,基于Bayes融合方法,更新出厂失效概率。通过更新后的出厂失效概率,确定模糊样本的组成,筛选出贮存失效概率样本。针对筛选后的样本,基于样本量加权最小二乘方法,开展贮存寿命评估。结果将所提方法应用于某弹箭产品案例,有效评估了批次出厂失效概率及其估计方差,并给出了可靠寿命评估结果。结论所提Bayes评估方法融合了出厂抽检数据和贮存抽检数据,有效解决了含模糊样本的失效概率估计问题,提高了估计的精确性,基于样本量权重的加权最小二乘法,考虑了样本的可信程度,提升了方法的科学性。

Bayes Fusion Evaluation Method for Storage Data of Composite Inspection with Fuzzy Samples

The work aims to evaluate the storage life of composite inspection products with fuzzy samples. The compatibility test was carried out on the ex-factory failure data and storage failure data of batch products. The uncertainty of fuzzy samples in the storage process was quantified by randomizing the number of ex-factory failure samples. With the failure data in the ex-factory inspection as prior information and the ex-factory failure data in the storage inspection as observation information, the ex-factory failure probability was updated based on Bayes fusion method. The composition of fuzzy samples was determined by the updated ex-factory failure probability, and the storage failure samples were screened out. The storage life of the products was evaluated based on the screened storage failure probability by the weighted least square method. The proposed method was applied to a missile product case. The ex-factory failure probability and its estimation variance were accurately estimated and the reliable life evaluation results were given. The proposed Bayes fusion method combines the ex-factory inspection data and the storage inspection data and effectively solves the failure probability estimation problem containing fuzzy samples and improves the estimation accuracy. The weighted least square method based on sample size considers the reliability of samples, which is more scientific.