Abstract
In order to evaluate nuclear power plant (NPP) safety, the nuclear industry has developed two safety analysis methods, deterministic safety analysis (DSA) and probabilistic risk assessment (PRA). However, carrying out DSA and PRA alone and relatively independently cannot accurately reflect the interactions between failures of equipment and the performance of system, thus safety and safety margin of NPP may be too conservative or under-conservative. In recent years, Riskinformed Safety Margin Characterization (RISMC)" technology was proposed by U.S. RISMC couples DSA and PRA to realistically analyze accident evolution paths, accident consequences, and probabilities under various accident scenarios. This paper aimed to solve the problem of huge calculation amount and computing resources in the RISMC analysis of NPPs. The dynamic event tree (DET) method was used and the concept of adaptive search was introduced, in order to balance the accuracy of limit surface positioning and total calculation amount. A typical accident case study of Small-break LOCA accident was carried out. Its results showed that using Adaptive DET algorithm in RISMC analysis could greatly improve the calculation efficiency of RISMC in accident scenarios and break through the limitations of traditional safety analysis of NPP.
