The Knowledge-Based Off-Site Emergency Response System for a Nuclear Power Plant

Abstract

A knowledge-based expert system for a nuclear power plant off-site emergency response system is described. The system incorporates the knowledge about the nuclear power plant behaviors, site environment and site geographic factors, etc. The system is developed using Chinshan nuclear power station of Taipower Company, Taiwan, ROC as a representative model. The objectives of developing this system are to provide an automated intelligent system with functions of accident simulation, prediction and with learning capabilities to supplement the actions of the emergency planners and accident managers in order to protect the plant personnel and the surrounding population, and prevent or mitigate property damages resulting from the plant accident. The system is capable of providing local and national authorities with rapid retrieval data about the site characteristics and accident progression. The system can also provide the framework for allocation of available resources and can handle the uncertainties in data and models. The prototype of the system is being built at INER using the off-the-shelf microcomputers. The Chinese user interface is also being developed. The system is also being implemented on an AI machine for a larger knowledge base and for the speed requirement.