Simulation-based probabilistic risk assessment for spent fuel pool

This report presents an approach for simulation-based probabilistic risk assessment (PRA) of a spent fuel pool, and analyses a loss of offsite power scenario using the approach. In the simulation-based event tree, accident timings, such as failure times of components and durations of manual actions, are simulated to analyse time-dependencies. The time windows for probabilistic analysis, namely mission times for safety functions and available times for manual actions, are calculated based on spent fuel pool conditions affected by the timings of previous events. The model combines deterministic and probabilistic analysis; the spent fuel pool conditions are calculated by a simplified, but sufficiently realistic deterministic model.

In this report, the simulation-based model is used to quantify minimal cut sets of a static PRA model more realistically. The results of dynamic and static analyses of a loss of offsite power scenario are compared. The dynamic analysis decreases the frequencies of minimal cut sets significantly. The decrease is particularly related to more realistic definition of mission times and crediting the operation of the cooling/make-up systems before they fail, which gives more time for the subsequent manual actions. The results also indicate that crediting repairs can greatly decrease the frequencies, though those can also be modelled in static manner to some extent.

There are some challenges related to application of the approach for full-scope spent fuel pool PRA. The simulation-based event tree becomes easily very complex when there are many failure combinations to analyse, and there is no good tool support to integrate the minimal cut sets of static PRA and the simulation results. One possibility would be to develop simulation-based event trees as independent PRA model so that there would be no need for static PRA model. However, the minimal cut set information would be lost, and the identification of all relevant failure combinations could be a challenge. Another potential solution would be to develop a simulation module for automatic quantification of minimal cut sets. This would provide more flexibility than a simulation-based event tree and would give wider possibilities to perform advanced minimal cut set quantifications.