Polynomial Neural Network Approach for Prediction of Liquid Holdup in Horizontal Two-Phase Flow

Abstract

A new polynomial neural network (PNN) model for estimating liquid holdup in horizontal two-phase flow is proposed in this article. The PNN evolutionally synthesize network size, connectivity, processing element types, and coefficients for globally optimized structure through training. The framework is established using 330 data sets from different experimental conditions. This self-organizing approach automatically presents internal relationships among data in the polynomial forms, and enhances data approximation and explanation capabilities of resulting data-based learning models. The comparative studies with experimental correlations and artificial neural network applications reveal that the model exhibits significant improvement in the processing structure, and outperforms previous models in overall accuracy across liquid holdup ranges.