航空发动机MTBF的Bayes评估

Bayesian assessment on MTBF of aero-engine

针对航空发动机的试验样本量和故障数据少,采用传统的数学平均值法对其平均故障间隔时间(MTBF)评估不能反映其真实可靠性水平的问题,基于Bayes理论,把历史试验数据视为先验信息,采用矩等效方法确定先验分布,然后通过Bayes理论综合现场试验数据,建立了一种基于Bayes理论的航空发动机MTBF评估方法。该方法可以扩大MTBF评估所需的信息量。采用所提出的Bayes方法对某航空发动机MTBF进行评估,得到其MTBF评估值为302.68h,比采用数学平均值法约提高了18.7%,评估结果更符合实际。表明该方法可应用于航空发动机MTBF的评估。

It is difficult to evaluate the mean time between failures （MTBF） of aero-engine with traditional mathematical mean time method because of its small samples.To solve this problem,the historical test information was regarded as prior information,and prior distribution was determined by the moment equivalent method,then field test data were combined with prior distribution through Bayesian theorem,lastly a Bayesian method to evaluate MTBF of aero-engine was presented.The information used in the new method to evaluate MTBF was expanded.An example showed the MTBF evaluation result with the new method was 302.68 h,18.7 % higher than that with the traditional mathematical mean time method.The engineering experts believed that the new result is closer to the real value than that by the traditional method.It is feasible to use the new method to evaluate MTBF of aero-engine.