Multilayer limited penetrable visibility graph for characterizing the gas-liquid flow behavior

Abstract

The research of the dynamical flow behavior of gas-liquid two-phase flow remains a challenge of great importance. We first carry out the vertical upward gas-liquid two-phase flow experiment in a 50 mm inner-diameter pipe and capture flow information via four-sector distributed conductance sensor. Then, we develop multilayer limited penetrable visibility graph (MLPVG) to analyze multichannel measurements, aiming to characterize the complicated flow behavior. We calculate two projection networks of MLPVG from two different perspectives (i.e. the overlapping information and the whole information). The ratio between the global clustering coefficients of these two projection networks is calculated as a quantitative indicator to characterize the multilayer network. We find that the proposed multilayer network measure can effectively characterize the evolution of gas-liquid flow, from randomly dispersed bubbles to periodically arisen gas slugs. The results suggest that our method can effectively fuse multichannel measurements and reveal the evolution of the complicated flow behavior.