Flow patterns identification of horizontal oil-water two-phase flow based on multivariate variational mode decomposition-Hilbert Huang Transform

Aiming to the problem of low identification in dealing with multivariate time series of oil-water two-phase flow by traditional time-frequency analysis technology, we propose a time-frequency analysis method of oil-water two-phase flow based on multivariate variational mode decomposition (MVMD) - Hilbert Huang transform (HHT). Decomposing the multi-channel time series that containing hydrodynamic information by multivariate variational mode, and further determine the inherent mode functions (IMFs) through spectrum analysis. Then, we perform HHT on IMFs and select the time-frequency characteristics of the signals to distinguish the flow patterns. The experimental results show that by selecting the Hilbert spectrum and marginal spectrum as the time-frequency characteristics, it can recognize the oil-water two-phase flow patterns under different flow velocities and map the nonlinear variation of fluid effectively, which has certain guiding significance for multivariate time series analysis of relative complex systems.