摘要
针对航空高可靠性产品现场故障数据样本量小和经常出现数据删失的问题,分析了平均秩次法与期望-极大值（EM）算法的有效性,给出了这两种方法的适用范围.应用蒙特卡罗仿真方法,在考虑维修的情况下,设计了一种评定可靠性评估方法有效性的仿真方法.该方法以威布尔分布为例,通过飞行年、飞机架数、日利用率的动态变化来模拟产生具有随机删失特点的航空装备现场故障数据,以此来动态驱动样本量和删失比的变化,并分别评估这两种可靠性评估方法的有效性.仿真结果表明,在样本量为10～30时的小样本随机删失数据参数估计中,EM算法应该被优先采用.
Application scope and validity of mean rank order method and expectation maximun （EM） algorithm are analyzed for the problem of high reliability aviation product of small sample sizes and random censoring.Considering the case of maintenance,a reliability assessment simulation method is designed based on the Monte Carlo method.Random censoring observations of aviation equipment are simulated through the dynamic changes of flight years,aircraft numbers,and daily utilization rate with the example of Weilbull distribution.Thus the validity of mean rank order method and EM algorithm are assessed through the changes of sample size and censoring rate.The simulation results show that EM algorithm should be preferred in parameter estimation of small sample and random censoring when the sample size is in the range of 10 to 30.
出处
《航空学报》
EI
CAS
CSCD
北大核心
2014年第5期1311-1318,共8页
Acta Aeronautica et Astronautica Sinica
基金
国家自然科学基金(51201182)
陕西省软科学技术基金(2011KRM122)~~
关键词
平均秩次法
EM算法
随机删失
威布尔分布
样本量
删失比
mean rank order method
EM algorithm
random censoring
Weibull distribution
sample size
censoring rate
