发动机无失效数据可靠性评估研究

在定时截尾寿命试验特别是高可靠性、小子样试验中的无失效数据,引入了逆Weibull寿命分布函数,提出采用不完全Beta分布作为故障概率的先验分布,给出分布参数的最小二乘估计以及故障概率的Bayes估计和多层Bayes估计,从而得到可靠度的估计。最后针对实际问题进行了计算,结果表明Bayes估计稳定合理且容易计算,假定发动机寿命服从逆Weibull分布时的计算结果具有较高精度和较小的偏差。

Research on Reliability Assessment Zero-failure Data of Engine

In the time-censored life tests,zero-failure data is always obtained,especially in high reliability and small sampling tests.In the paper,Inverse Weibull distribution was introduced into the zero-failure data analysis.Concerning characters of zero-failure data,the incomplete Beta distribution was taken as an prior distribution of the failure probability,in order to estimate the failure probability using the Bayesian method.The least squares estimates of the distribution parameters were given,and hierarchical Bayesian estimation of failure probability of zero-failure data was presented,so that the reliability estimation of zero-failure data was obtained.Finally,calculations were performed concerning practical problems to illustrate the estimations developed in this paper.The result shows that the Bayesian estimations are stable,feasible and easy to operate,and the Inverse Weibull distribution is a reasonable alternative distribution to the life of engine,with less error.