DYNAMIC APPROACH ON MULTI-UNIT PROBABILISTIC RISK ASSESSMENT USING CONTINUOUS MARKOV AND MONTE CARLO METHOD

摘要

A risk assessment of multi-unit by a typical event tree (ET) method is insufficient because the state of the plant will vary by a plant status of the adjacent unit. In this study, a new approach of scenario quantification method of multi-unit Nuclear Power Plants (NPPs) using Markov process and Monte Carlo method is proposed to evaluate interactive time-transient accident scenario progression which considered the effect of an adjacent unit. The impact (positive and negative effect) of the adjacent unit (multi-unit) on a single unit is modeled to change failure probability of each heading in the ET. Markov process is adopted to estimate transient plant status using a fault tree (FT) analysis. Markov process decides the plant status at the present time step only based on the status of the previous time step. Also, Monte Carlo method is used to decide the current plant status by comparing a random variable with the transient plant state. Rather than giving branch probabilities in the ET, the current failure probabilities of each heading of the ET are used to decide accident progression. By using this methodology, the influence of multi-unit and dynamic and interactive accident scenario progression can be evaluated. The new methodology applied for a scenario quantification of the primary containment vessel (PCV) failure event. It is showed that the new methodology is an effective approach to evaluate the interaction among multi-units and the particular time sequence of each unit.