RESEARCH ON ARTIFICIAL INTELLIGENT NUCLEAR POWER PLANT EMERGENCY OPERATING GUIDE

抄録

Artificial intelligent (AI), as a mature high technology, had little used in nuclear power plant (NPP) field that it has a wide application scope to improve the management of NPPs. This paper analyzes the method and feasibility of using AI on field of NPP emergency operating guide, i.e. making the emergency operating strategy after accident occurred by AI system. This paper adopts artificial neural network (ANN) to calculate the emergency operating strategy, in which the post-accident status parameters vector is the input of the model and the emergency operating strategy is the output. In order to satisfy the requirement of ANN, emergency operating strategy is modeling based on the opinion that the fundamental strategy of an accident is to control the parameters which might effect on the safety functions. Controlling of the safety related parameters can also be deeply cataloged into control methods and system/component control, and forming an emergency operating strategy tree in the final. This paper also argues aspects of input data format, judgment of safety shutdown, ANN application arithmetic optimization, etc. A simple ANN model has been constructed in this paper and performing a practice calculation. The result verified that using AI on making decision during emergency condition is feasible. It is indicated that, currently AI system can provide auxiliary information for post-accident occasion, and this paper also illustrates the further research aspects for engineering practice on the next phase.