Flow regime prediction using artificial neural networks for air-water flow through 1–5 mm tubes in horizontal plane

Abstract

Artificial Neural Network (ANN) based techniques for the classifications of flow regimes in air-water flow through 1–5 mm tubes are presented. 218 data points are based on the experimental investigation in 3 and 4 mm tubes and 2114 data points from various experimental results from the published literature for airwater two-phase flow in small diameter tubes have been used. Five different well known artificial neural network models have been used to predict the flow regime. The ANN model based on Radial Basis Function and Principal Component Analysis gives better predictability over the other networks used.