复杂系统平均剩余寿命综合评估方法

Comprehensive Evaluation of Mean Residual Life of Complex System

针对工程中广泛关注的平均剩余寿命问题,首先利用信仰推断给出单个设备可靠度的置信分布,包括指数分布、威布尔分布和对数正态分布;然后基于组成系统的设备试验信息,结合系统的组成结构,给出系统可靠度的置信分布,接着根据可靠度和平均剩余寿命之间的关系,建立复杂系统平均剩余寿命综合评估模型并给出具体的评估算法。对于在设备试验中经常出现的右删失等不完全数据,给出了将不完全数据虚拟成等效的完全数据的转化方法。最后,通过模拟研究和实例验证,结果表明本文的方法较为精确,能够满足工程的实际需求。

The mean residual life of complex system is an important reliability index which is focused on very much in engineering, but it is rarely studied so far. The confidence distribution of single equipment’s reliability, such as exponential distribution, Weibull distribution and log-normal distribution, is derived by fiducial distribution. Based on test information of components composed the complex system, the confidence distribution of the system reliability is obtained via the structure of the system. The comprehensive evaluation model of the mean residual life of complex system is established from the relationship between reliability and mean residual life, and the evaluation algorithm is given in detail. For incomplete data like censored data which is often occurred during component test, one pseudo-complete sample technique is given. In the end, the good performances of the methods discussed in the paper are shown by simulation and a real example.