A NEURAL NETWORK MODEL FOR THE TOMOGRAPHIC ANALYSIS OF IRRADIATED NUCLEAR FUEL RODS

摘要

A tomographic method based on a multilayer feedforward artificial neural network is proposed for the reconstruction of gamma-radioactive fission product distribution in irradiated nuclear fuel rods. The quality of the method is investigated as compared to a conventional technique on experimental results concerning a Canada deuterium uranium reactor (CANDU)-type fuel rod irradiated in a TRIGA reactor.