ADVANCED PROBABILISTIC SIMULATION OF NPP FIRES

The consequences of fires at NPPs can be estimated using the numerical simulation. When combined with Monte Carlo technique, the probability of target events can be calculated. The computational cost related to the use of most advanced fire models has prevented their application in this probabilistic manner. In this work, a new technique, based on an intuitive approach, is proposed. The technique allows the use of two different models in the same Monte Carlo simulation, and is therefore called Two-Model Monte Carlo (TMMC). In TMMC, simulations are first performed using both models in relatively small number of points of the random space, and the scaling functions are formed. Full Monte Carlo is then performed using the faster but less accurate model, and the results are corrected using the scaling functions. The performance of the numerical technique is studied in a simple room fire scenario, where two-zone and computational fluid dynamics (CFD) models are used. TMMC can reduce the computational cost with a factor of 30, compared to a full Monte Carlo using CFD. As a real NPP fire scenario, the distribution of component failure time is calculated in a relay room fire. In this application, CFD is used in both models but the difference in accuracy and computational cost is achieved by the refinement of the computational mesh.