Identification model of an accidental drop of a control rod in PWR reactors using thermocouple readings and radial basis function neural networks

Abstract

The accidental dropping of a control rod may cause the reactor to operate unsafely. In this type of event, there is a distortion in the distribution of power and temperature in the core may exceed operating limits reactor safe. This work aims to develop an alternative model capable of identifying, at any time of the cycle, the control rod that has accidentally dropped at the core of a PWR reactor, using the readings of the thermocouples in order to minimize possible losses.The model assumes that in a possible drop of a control rod, the largest temperature change occurs in the position where the control rod is inserted. Considering the fact that there are no temperature gauges in all control rod positions, the proposed model uses radial basis function (RBF) neural networks to make a reconstruction of temperatures in these positions from the measurements of the thermocouples at the time of the accidental drop.The study found that the predictions of the temperatures made by the RBF neural networks showed good results, which enables the identification of the control rod dropped accidentally in the core, by simple inference of the fuel assembly of lowest temperature among temperatures reconstructed.