Failures Ignored In Series And Parallel System

The maintainability has been studied extensively in the theory of reliability. Maintenance is a usual way to improve system reliability. Reliability analysis of repairable system is an important branch of reliability theory. This thesis studies Markov repairable series system and parallel system consisting of two units under new assumptions with methods from the theory of probability and stochastic process. We establish two new models and then derive and prove some system reliability indexes.For Markov repairable systems, if repair times that are sufficiently short do not result in system failure, we can say that such a repair interval is omitted from the downtime record. We consider this situation in the new models in order to improve the old ones. We mainly suppose that the repair time less than a given critical value will be neglected. This supposition makes the new models more suitable to the reality.In the new models, the durations of working and repair of each unit no longer have negative-exponentials, which result to that the process is not Markov process. As new systems don't have the Markov properties, it's worth considering how to get reliability indexes of the new system. This thesis derives some indexes such as instantaneous availability, steady-state availability, reliability and so on. They are all important reliability indexes of repairable system.