Estimation of Axial DNBR Distribution at the Hot Pin Position of a Reactor Core Using Fuzzy Neural Networks

Abstract

The pressurized water reactor (PWR) generally operates in the forced convection or nucleate boiling regime. However, if the fuel rod is operating at a high power density, the nucleate boiling that is characterized by extremely high heat transfer rates becomes film boiling with severely reduced heat transfer capability, which is called Departure from Nucleate Boiling (DNB). In this work, the axial DNB Ratio (DNBR) distribution at the hot pin position is predicted by the fuzzy neural networks using the measured signals of the reactor coolant system. The fuzzy neural network is a fuzzy inference system equipped with a training algorithm. The fuzzy neural network is trained by a hybrid method combined with a back-propagation algorithm and a least-squares algorithm. The proposed method is applied to the first cycle of the Yonggwang 3 nuclear power plant. The relative 2-sigma error averaged for 13 axial locations of the hot rod is 1.97%. The fuzzy neural networks estimate DNBRs more accurately at central parts that have relatively lower DNBR values which are more important in safety aspects. From these simulation results, it is known that this algorithm can provide reliable protection and monitoring information for the nuclear power plant operation and diagnosis by accurately predicting the DNBR each time step.