Application of convolution neural network to flow pattern identification of gas-liquid two-phase flow in small-size pipe

Abstract

Flow pattern is one of the most important parameters for gas-liquid two-phase flow. In this work, a new flow pattern identification method based on Convolution Neural Network (CNN) is presented. A 7-layer CNN structure is chosen, and the parameters of this network are determined by a training set. In order to verify the feasibility, experiments were carried out in horizontal pipe with the inner diameter of 4.0 mm. The results show that the presented method has better performance than traditional classification methods. The identification accuracies of four typical flow patterns are all above 92%. This work verifies the feasibility of applying CNN to flow pattern identification, and provides a good reference for parameter measurement of gas-liquid two-phase flow in small-size pipe.