Integrated PRA methodology to advance fire risk modeling for nuclear power plants

This paper presents the current progress in the development of an Integrated Probabilistic Risk Assessment (I-PRA) methodology, an integration of classical PRA with simulation techniques, for the analysis of fire risk at nuclear power plants (NPPs). I-PRA replaces fire-related logic in existing Fire PRA models with a Fire Simulation Module (FSM) that explicitly simulates the underlying physical phenomena and their dependencies and dynamic interactions. FSM includes three sub-modules: fire initiation, fire progression, and post-fire failures. The fire progression sub-module uses the Fire Dynamic Simulator (FDS) code, developed by the National Institute of Standards and Technology, and uncertainties in input parameters are propagated using Latin Hypercube Sampling. Large-scale, high-performance computers are utilized to overcome the excessive computational demand of FDS. The methodology is applied for one common NPP scenario (motor control center panel fire in a switchgear room). Global sensitivity analysis is conducted to rank the criticality of input parameters.