Gas-water two-phase flow pattern characterization with Multivariate Multiscale Entropy

Gas-liquid two-phase flow is a commonly encountered multiphase flow phenomena in process industries. The flow behavior of each typical flow pattern is urged to know in order to further understand its inherent flow mechanism. This paper aims to characterize the flow pattern of gas-water two-phase flow within a horizontal pipe by the Multivariate Multiscale Entropy (MMSE) analysis with the measured data of a dual-plane Electrical Resistance Tomography (ERT) system. The measured data from each plane of ERT is jointly processed as a bivariate time series set, and the Multivariate Multiscale Sample Entropy (MSampEn) is then applied to characterize the inner flow behavior of the two-phase flow system. The results demonstrate that the MSampEn is capable of providing insight characterization of the decomposition of the measurement signal of each flow pattern, and that the flow patterns and also their complexity can be easily identified at different scales.