中文摘要：

可修串联系统是一类经典的可靠性模型,在实际工程中较为常见.为满足工程实际需求,所研究的系统由2个部件和1个维修工组成.假设工作时间、维修时间和部件更换时间均服从指数分布,在维修之后系统不能修复如新.系统采用部件故障N次之后将被更新的维修策略.在这些假设下利用几何过程和马尔科夫过程对系统进行建模,再利用半马尔科夫过程和数值积分方法给出系统瞬态可用度和（0,t]时间内故障次数的计算公式.最后给出算例并应用Monte Carlo方法对所得公式进行验证.为进一步研究复杂机械系统的可靠性提供了理论基础.

英文摘要：

A series repairable system is one of the classical reliability models and is usually used in practice. To meet the need of practical engineering application, a repairable system consisting of two components and a single repairman was studied. We assumed that the working time distribu- tion, the repair time distribution and the replacing repair time distribution of the system were all exponential. After repairing, the system was not ＂as good as new＂. A repair replacement policy N under which the system was replaced when the number of failures of the system reaches N was studied. Under these assumptions, by using the geometrical process and Markov process tech- nique, we derived explicit expression for system availability and failure times in （0, t] by semi- Markov process and numerical integration. Finally, a numerical example for policy N is given and verified by Monte Carlo. This research provides evidence for further studies of complex mechani- cal system reliability.