An expert system to assist with diagnosis and decision making in radiation protection

Most nuclear facilities now have sophisticated radiation monitoring systems providing information which can be used directly for radiation protection purposes and which is usually managed by a computer. However, a radiation protection expert is still needed at the facility to make a diagnosis of any incidents that occur on the basis of the signals given by the system. In the event of an accident, this expert, if physically present, has to act in an emergency situation, i.e. under conditions which are not ideal from the point of view of calmness and safety. As radiation protection experts at a research reactor, we thought that it would be useful to develop a computerized intelligent system which would use the data provided by the monitoring system to make an automatic diagnosis of the situation, without being affected by the state of emergency, and to assist officers in deciding on the actions to be taken to ensure the safety of personnel and the protection of the environment. The technical solution adopted is an expert system, an artificial intelligence solution based on a separation between the programming part, the control system simulating human reasoning, and the knowledge base containing the expert knowledge, largely in the form of ''if ..., then ...'' rules. The development of such a system has improved knowledge about radiation protection aspects of the facility and has permitted a sort of dynamic storage of this knowledge. The system will be used as a means of transmitting expert information to future radiation protection officers at the facility. A similar system is also planned for the entire Grenoble Nuclear Research Centre, to control incidents at the Centre and to assist those responsible in taking decisions to ensure the safety of personnel at the Centre and the protection of its environment. (author)