61 A Study on an Accident Diagnosis Methodology Using Influence Diagrams (PSAM-0261)

In nuclear power plants, operators are allowed to follow Emergency Operating Procedures ( EOPs) when reactors are tripped because of accidents. However, it's very difficult to diagnose accidents and select appropriate procedures to mitigate current accidents in the short time given. Even if accidents are diagnosed quickly, there is the possibility of misdiagnosis. To cope with this problem, methodology using Influence Diagrams has been developed and applied to represent the dependency behaviors and uncertain behaviors of complex systems. An example to diagnose the accidents such as SLOCA and SGTR with similar symptoms has also been introduced. From the constructed model, operators were able to diagnose accidents at any stages. Also, the integrated design of Thermal Hydraulics Online Monitoring Advisory System (THOMAS) is introduced in this paper. THOMAS, for improved Korea Standard Nuclear Power Plant, is an advisory monitoring system that improves performance by using virtual reality and databases. It is believed that the new design will improve the operability and maintainability as well as simplify the design process of the monitoring and diagnosis system of the nuclear power plant. THOMAS also offers improved information about accidents with given symptoms. The model is expected to help operators diagnose problems more correctly and rapidly while reducing human error.