An Application of the Dynamic Neural Network to Accident Diagnosis Systems for Nuclear Power Plants

摘要

In this work, we developed a dynamic neural network for an accident diagnosis system. The dynamic neural network is efficient to handle continuous varying situations such as accidents in NPP. And a prototype of accident diagnosis system was developed for validation. But in this prototype, just only alarms are considered as input patterns. In order to make more reliable and efficient accident diagnosis systems, much more variables such as values and trend of instruments in an MCR should be considered.