Training strategies of Multilayer Perceptron network for study in three-phase flow


César Marques Salgado
Roos Sophia de Freitas Dam
William Luna Salgado


This report presents a methodology based on nuclear techniques combined with artificial neural network (ANN) that have been used in order to predict fluid volume fractions (FVFs) using data obtained from gamma-ray radiation detectors. This research proposes investigations and comparisons of training strategies based on dual-modality principles using two NaI(Tl) detectors.


Application of Nuclear Techniques in Industry