基于成败试验的导弹贮存可靠性Bayes分析模型

Bayesian model for missile storage reliability analysis based on success-failure test

针对传统导弹贮存可靠性分析方法的缺点,提出了一种Bayes分析模型.这种模型以狄氏分布为先验分布,综合利用了历史信息和专家信息,结合导弹检测时间段的成败试验数据,给出了各检测时间段可靠性的联合后验分布,然后利用Gibbs抽样算法进行后验推断,得到了各阶段可靠性的Bayes估计和区间估计.此外,基于当前试验数据,利用这种模型以预测未来导弹贮存可靠性.算例表明这种Bayes模型参数含义清楚,利于工程应用.

In view of the disadvantages of the classical analysis method for missile storage reliability analysis, a Bayesian model was proposed in this paper. The Dirichlet distribution was taken as prior distribution in the model. The historical information and expert information were synthetically used. Combined with the missile success-failure test data for every stage, the joint posterior distribution of reliability was generated. Then, the Gibbs sampling algorithm was used to compute the posterior inference. The Bayesian estimators and Bayesian lower bound were gained for the reliability of every stage. Furthermore, based on the test data, the model can be used to predict the missile storage reliability. The example shows that the parameters of the Bayesian model have apparent meaning and are convenient to use for engineering applications.