Dynamic Risk Assessment Framework to Support Decision-Making for Nuclear Safety

Probabilistic safety assessment (PSA) is widely used in the risk assessment field of nuclear power plants (NPPs) using static event trees (ETs) and fault trees. Traditional PSA evaluates NPP safety by analysing representative scenarios with conservative assumptions in complex cases. Over the past several decades, research has been performed to evaluate realistic risk assessments considering dynamic characteristics. This is expected to help understand the differences between traditional PSA and dynamic PSA. However, to date, there have been no dynamic PSA applications that have influenced regulatory decision-making, partly due to the complexity of modelling and simulating dynamic accident scenarios. Moreover, dynamic PSA requires time-dependent reliability functions not only for system components but also for human actions to evaluate dynamic risk. For these reasons, this work proposes a novel dynamic PSA methodology, a dynamic risk assessment through automatic accident sequence generation using optimized simulations of NPPs (DRAGON). DRAGON consists of three main components: (i) an optimized simulation algorithm to reduce the number of required simulations, (ii) an automatic accident sequence generation method to present the results in an interpretable manner while controlling the scenario coverage and accident sequence complexity using alpha shape algorithm, and (iii) a time-dependent human reliability evaluation framework to assess dynamic risk quantitatively. The integration of three elements enables decision-makers to determine the desired level of scenario coverage, the complexity of the generated accident sequences, and evaluate the resulting dynamic risk, thereby supporting adaptive decision-making for nuclear safety. DRAGON was demonstrated with a case study for station blackout with dynamic scenarios including the timing of portable facility installation. The results showed that DRAGON can effectively support decision-making to select the accident sequences for dynamic scenarios.

Keywords: Dynamic Probabilistic Safety Assessment, Decision-Making, Station Blackout, Dynamic Scenario, Automatic Accident Sequence Generation, Alpha Shape method.