Over the last decade, probabilistic risk assessment (PRA) for nuclear power plants shifted its focus toward simulation-based methods of evaluating safety and risk aspects. Classical reliability modeling based on static Boolean structures does not capture the dynamic nature of plants experiencing accident sequences. Dynamic PRA extends conventional reliability modeling, adding a higher degree of variability to events occuring during accident sequences. This paper demonstrates the capability of a new dynamic event tree (DET) analysis tool applicable to severe accident sequences. A cyberattack was assumed to penetrate the instrumentation system of a boiling water reactor during its hot shutdown phase, altering the thermal–hydraulic variables being processed. The analysis focused on scenarios that lead to the core being uncovered, in which the peak cladding temperature is reached due to the disabling of the cooling systems. A limiting surface is defined, associating the cyberattack timing to the recovery time limit that would compromise the core integrity.
Read full abstract