中文摘要：

针对绝大多数机械系统部件的破坏形式为疲劳破坏,采用由P-S-N曲线转化而来的通用疲劳寿命分布更适于表示系统中部件的寿命分布。研究的对象为并联系统,部件寿命不服从指数分布,部件的维修时间服从指数分布,并且为每一个部件提供一个维修工,修复后可以恢复如新。利用Markov更新过程以及数值积分求得单部件的瞬态可用度计算公式,再利用单调关联系统的特点求得系统可用度与部件可用度之间关系,得出并联系统的瞬态可用度计算公式。以两个部件组成的并联系统为例,部件寿命分布服从某种材料的通用疲劳寿命分布,并利用Monte Carlo方法对该系统进行了验证。与模拟方法相比,利用Markov更新过程可以在更短的时间求得更精确的瞬态可用度计算结果。因此为部件寿命分布形式多样的复杂系统求解瞬态可用度提供理论基础。

英文摘要：

Failure mode of most mechanical systems is fatigue failure, so the fatigue life distribution function obtained by using the P-S-N curve is suitable to describe the work life distributions of units in a mechanical system. In this paper, the parallel system is studied. The work life doesn＇t obey exponential distribution. Every unit has a maintainer. The repairing time obeys the exponential distribution and one unit after repair is ＇ as good as new＇. According to the Markov renewal process, the calculation formula of transient availability for single unit can be figured out, then by the use of the features of coherent system and numerical integration, the calculation formula of transient availability for parallel system can be figured out. Taking a parallel system as an example, the work life distributions of units follow fatigue life distributions of some certain engineering materials, and transient availability of the system is verified by Monte Carlo method. Comparing with the Monte Carlo method, the Markov renewal processes takes less time to get more precise result of transient availability. So it can provide theoretical basis for the availability calculation for complex system, in which units work life obey various distributions.