FLOW RATE CALCULATION ON A TWO-PHASE STRATIFIED FLOW REGIME USING RADIOACTIVE PARTICLE TRACKING AND ARTIFICIAL NEURAL NETWORK

Resumo

This report presents a simulated methodology for predicting the volume fraction of two-phase flows and calculating the superficial velocity based on the radioactive particle tracking technique. To measure the volume fraction, a sealed radioactive particle inside the pipeline is used. The multiphase flow test section consists of oil and salt water on a stratified flow regime, a polyvinyl chloride pipe, five NaI (Tl) detectors, and a Cs-137 radioactive particle emitting gamma rays of 662 keV. The prediction of the volume fraction was performed using a multi-layer perceptron neural network with a backpropagation algorithm. The novelty of this article is the combination of radioactive particle tracking and artificial neural networks to predict the volume fraction of multiphase flows. Furthermore, the Cross Correlation function was used to determine the time delay to obtain the superficial velocity in three different volume fractions, which allowed the calculation of the flow rate for each phase.