Numerical methods for seismic fragility analysis of structures and components in nuclear industry - Application to a reactor coolant system

In the nuclear industry, the probabilistic risk assessment approach has become the most commonly used methodology for evaluating the seismic risk of nuclear plants. In this framework, fragility curves express the conditional probability of failure of a structure or component for a given seismic input motion parameter. In engineering practice, fragility curves are determined in a simplified way using safety factors with respect to design earthquake. For critical components, however, it can be important to have a more accurate evaluation of seismic fragility based on numerical simulation. In this paper, we will present a complete probabilistic study of a nuclear power plant component based on Monte Carlo simulation and evaluate fragility curves. Moreover, we discuss different methods available for statistical estimation of fragility curves and comment on the choice of the ground motion parameter. Indeed, choosing an appropriate ground motion parameter can reduce uncertainty.