Dynamic probabilistic risk assessment of seismically induced internal flooding considering aftershock effects using physics-based flood propagation and fragility models

This study develops a dynamic probabilistic risk assessment (PRA) method for combined earthquake and internal flooding events in nuclear power plants. The method explicitly integrates seismic fragility analysis with a flood propagation model, enabling evaluation of how earthquake-induced structural damage influences subsequent water intrusion, and vice versa, in a combined, interdependent manner. Seismic effects are evaluated using fragility functions across different ground-motion levels. These are implemented in a fault tree to estimate the failure probability of safety-related equipment. Internal flooding is evaluated by calculating the water level in each room and converting it into equipment failure probability using a water-level-dependent flood fragility model. The system behavior is treated as time-dependent. A state-transition-based dynamic PRA framework is used to update the probability of loss of function at each time step, consistent with a Markov-type state-transition formulation. Uncertainty in the leak rate and barrier damage is accounted for using a Monte Carlo simulation. By combining these models, the method can describe the progression of accident scenarios involving seismic damage and flooding. The results indicate that different scenarios can be governed by distinct mechanisms under different conditions. In some cases, seismic fragility is dominant, while in others, flooding becomes dominant. Mixed behavior is also observed. Therefore, this approach can be useful for evaluating complex scenarios that are difficult to treat with conventional PRA methods.