Prediction of the thermal dynamic parameters fluctuation of coolant system of the IBR-2M reactor using neural networks

This paper presents an artificial neural network method for long-term prediction of the thermal dynamic parameters of primary coolant circuit of the IBR-2M reactor. The main goal is to predict the temperature and liquid sodium flow rate through the core and thermal power. It is shown that the prediction can reduce three times the effects of slow reactivity fluctuations in power and decrease the requirements for the automatic power stabilization system. Nonlinear autoregressive neural network (NAR) with local feedback connection has been considered. The results of prediction error ~ 5% coincide with the experimental ones.