竞争性故障模型可靠性评估的非参数估计方法

Nonparametric estimation method of reliability evaluation in competitive fault model

针对航空发动机现场数据的可靠性评估面临着小子样与随机删失问题,对可靠性评估中的非参数估计方法进行了对比分析.应用蒙特卡罗方法,在竞争性故障模型下,设计了一种对比可靠性评估方法优劣性的仿真方法.该方法以三重威布尔分布为例,通过模拟产生具有随机删失特点的航空发动机现场故障数据,分别对比4种可靠性评估非参数估计方法的优劣性.最后通过一个工程应用案例验证了模型的有效性.仿真结果表明:在三重威布尔竞争性故障模型下,非参数估计方法的优劣顺序依次是Kaplan-Meier估计、平均秩次法、Herd-Johnson估计、Nelson-Aalen估计.

The nonparametric estimation method of reliability evaluation was compared and analyzed for the problem of small sample and randomly censoring in the field data of aer- o-engine. Monte Carlo method was applied and a simulation method was designed to compare the reliability evaluation method within competitive fault model. Through the example of trebling Weilbull distribution, random censoring observations of aero-engine were simulated through this method by comparing four nonparametric estimation methods of reliability eval- uation. An example of engineering application was used to illustrate the validity of the meth- od. Results show that the excellence order of nonparametric estimation method is Kaplan- Meier estimation, mean rank order method, Herd-Johnson estimation and Nelson-Aalen esti- mation under the trebling Weilbull competitive fault model.