DEVELOPMENT OF SEVERE ACCIDENT MANAGEMENT SUPPORTING SYSTEMS USING QUANTIFIED CONTAINMENT EVENT TREES

Abstract

The Probabilistic Risk Analysis (PRA) for a nuclear power plant can provide so much valuable information including plant-specific vulnerabilities against severe accident, which is quite useful in developing severe accident management (SAM) strategies. In this study, the information regarding Containment Event Trees obtained by performing PRA is introduced to develop the suitable event-oriented SAM strategies. For these, it is crucial to identify the exact state of severe accident progression which is called Plant Damage State as an entrance condition. However, there are several areas where insufficient knowledge exists at the moment and thus it can be impossible to identify the event correctly. In this study, symptom-oriented strategies are also provided as an alternative methodology by using safety objective tree method to circumvent such drawback. Both event-oriented and symptom-oriented approaches are synthesized into a computerized supporting system for SAM, and linked with a severe accident simulator for on-line verification. A sample sequence is selected to show the simulated results performed by the system. Simulation results show that this kind of methodology might be quite useful in developing and implementing SAM strategies.