Experimental Study of Horizontal Gas-liquid Two-phase Flow in Two Medium-diameter Pipes and Prediction of Pressure Drop through BP Neural Networks

Abstract

Horizontal gas-liquid two-phase flow is of crucial importance in oil-gas storage and transportation. In view of previous researches with the lack of horizontal gas-liquid two-phase flow in medium-diameter pipes (60-75 ㎜), experiments were conducted in DN 60 and DN75 at medium-high liquid velocity (50-250 m3/d) and high gas-liquid ratio (20-500 m3/m3). Comparison between flow patterns maps in DN60 and DN75 and the Taitel-Dukler model showed that only flow patterns map in DN60 agreed well. On the basis of experimental data analysis, variation laws of liquid holdup and pressure drop were determined. Besides, this paper developed a method to predict liquid holdup and pressure drop through BP neural networks. Results proved the capability of BP neural networks. The further validation can be made in practical applications.