Reliability quantitative analysis of protection system of nuclear power plant based on BDMP

Digital instrument and control system is the central nervous system of nuclear power plant, and shutdown protection system is one of the most critical subsystems, and its reliability evaluation is one of the key tasks to improve the safety level of nuclear power plant operation. Testability and maintainability are one of the basic requirements of protection system design. Failure of a single component will not lead to loss of protection function in the actual operation process, and timely repair of the fault components can improve the reliability of the system. Therefore, the repair of failure components should be considered in reliability evaluation. The Boolean Logic Driven Markov Process (BDMP) has the ability to describe complex correlations and dynamic event timing. The model is similar to the fault tree structure and is easy to set up. Based on the limitations of the existing BDMP modeling software and the lack of validation of the BDMP quantitative analysis software YAMS, In this paper, a method of transforming Boolean Logic Driven Markov Process (BDMP) into Markov model is proposed, and the probabilistic model detector PRISM is used for quantitative analysis of Exported model. The BDMP model of 2oo3 and 2oo4 shutdown protection system of nuclear power plant is established, and the corresponding equivalent Markov model is derived by the proposed method. The quantitative analysis of the model is realized by using PRISM, and the failure probability of the system is calculated. The results of PRISM and traditional BDMP quantitative analysis software YAMS are compared. The comparison shows that the cumulative failure probability obtained by YAMS fluctuates greatly with time, and the cumulative failure probability curve obtained by PRISM is roughly in the middle of the YAMS curve envelope. It can be considered that the PRISM calculation method is more robust and the result is more reliable. (authors)