Reliability Modeling And Simulation Of Multi - State Repairable System Based On Markov Process

The research of the reliability is significant to ensure the safety of production, promote the structure optimization and the improvement of performance, enhance the system of reliability, safety and Maintenancability. Initially, there are only "duty" and "fail" state in the concept of scholars. Such concept is one-sided, because the results of the analysis in the concept would lead to a gap with actual situation. In other word, the analysis in the concept would not enough accurate. With the complicating of mechanism and the appear of various property, the various states and the various level of property for system or component are analysed. Statistical analysis and preventive maintenance(PM) is made to the fault of multi-state system for increasing the level of reliability and production efficiency, reducing the failure rate and the cost of running. This thesis take engine as object and focuses on the modeling risk of a multi-state repairable system. The main work of this paper is as follows:(1) To present an opinion about the using of a Markov process and Markov property with multiple states. The fuel supply system is analysed directly by the improved Markov model. The sudden transfer, which is based on the analysis of previous data of the fuel supply system, will be directly considered in the improved Markov model, which will eliminate the "0"state. Matlab will be used to get the reliability parameters in the process for simplifying counting process, finally, preventive maintenance decision will be made for the system.(2) Preventive maintenance decision will be made for the fixed transition probabilities of multi-state system. The Initial Model is builded on Markov and check the correctness of the Initial Model. The reliability parameter of each state is got from the Initial Markov Model.The Model A is got by way of inserting the CostMatrix into the Initial Model. Its Markov property is be checked. The model is runned for obtaining cost by RENO. Taking the optimal average total cost as object, the feasibility of preventive maintenance(PM) is decided.(3) Preventive maintenance decision will be made for the non-fixed transition probabilities of multi-state system.The Model B, a periodically decreasing reliability model, and Model C, a continuously decreasing reliability model, are builded on the basis of the non-fixed transition probabilities of multi-state system. The two models are all not follow Markov process and their Markov property is analysed. The CostMatrix is included by models. Taking the optimal average total cost as object, the decision about preventive maintenance(PM) and more detail analysis is made. Finally, the opinions of preventive maintenance is presented for the non-fixed transition probabilities of multi-state system.