Abstract
Abstract The paper presents reliability and risk analysis of the RBMK-1500 reactor accident localization system (ALS) (confinement), which prevents radioactive releases to the environment. Reliability of the system was estimated and compared by two methods: the conventional fault tree method and an innovative dynamic reliability model, based on stochastic differential equations. Frequency of radioactive release through ALS was also estimated. The results of the study indicate that conventional fault tree modeling techniques in this case apply high degree of conservatism in the system reliability estimates. One of the purposes of the ALS reliability study was to demonstrate advantages of the dynamic reliability analysis against the conventional fault/event tree methods. The Markovian framework to deal with dynamic aspects of system behavior is presented. Although not analyzed in detail, the framework is also capable of accounting for non-constant component failure rates. Computational methods are proposed to solve stochastic differential equations, including analytical solution, which is possible only for relatively small and simple systems. Other numerical methods, like Monte Carlo and numerical schemes of differential equations are analyzed and compared. The study is finalized with concluding remarks regarding both the studied system reliability and computational methods used.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call