A comparison of static and dynamic event tree analyses for SBO accident in VVER-1000/V446 NPP

Abstract

Event tree analysis is applied to quantify the core damage frequency (CDF) and assess the risk of nuclear power plants (NPPs) resulting from various postulated initiating events. To calculate this criterion, it is necessary to generate the probable scenarios according to the function of safety systems and the operator's actions. The classical event tree is currently used in PSA analysis. This method does not consider the accident's dynamics and scenarios. It considers only the availability/unavailability of the safety system functions and the operator's actions to calculate the frequency of each scenario. In contrast, the dynamic event tree method applies physical and probabilistic models to generate branches in the event tree, calculate the frequency of each scenario, determine the time profile of core damage, and time variation of physical parameters of the NPP for each scenario. This paper develops the dynamic event tree for the SBO accident at the VVER-1000/V446 NPP using the RELAP5 and RAVEN codes. The results are then compared with the outputs of the classical event tree. The results show that according to the assumptions, 3170 scenarios are evaluated in the dynamic event tree, while only 33 predetermined scenarios are examined in the conventional event tree. The calculated core damage frequencies are 3.61×10-6 (yr-1) and 1.97×10-6 (yr-1) for conventional and dynamic event trees, respectively.

Keywords

• static event tree (et)
• dynamic event tree (det)
• station black-out (sbo) accident
• vver-1000/v446 npp
• relap5/mod3.2
• risk analysis virtual environment (raven)