Risk Analysis of Discrete Dynamic Event Tree Based on Dynamic Bayesian Network

With the development of science and technology, systems' complexity has increased sharply because of systems' characteristic of polymorphism and dependence. Although lots of classical methods have been proposed, they show great limitations in dynamic systems' probabilistic risk assessment. To solve this problem, a probabilistic risk assessment method that combines Discrete Dynamic Event Tree (DDET) and Dynamic Bayesian Network (DBN) is proposed in this paper. Compared with other methods, the new method has lots of advantages: (1) Based on DBN's mature mathematical foundation and software support, it provides a useful method for DDET's solution (2) Sensitivity analysis based on DBN helps analysts find high sensitivity events (key events) in system, which provides significant guidance for the development of preventive measures; (3) DDET focuses on the logical combination of events, while Dynamic Bayesian Network focuses on the correlation between different events. As a result, the combination of DDET and DBN helps analysts have a better understanding of risk events in complex systems.