To find effective and practical methods to distinguish gas-liquid two-phase flow patterns, new flow pattern maps are established using the differential pressure through a classical Venturi tube. The differential pressure signal was first decomposed adaptively into a series of intrinsic mode functions (IMFs) by the ensemble empirical mode decomposition. Hilbert marginal spectra of the IMFs showed that the flow patterns are related to the amplitude of the pressure fluctuation. The cross-correlation method was employed to sift the characteristic IMF, and then the energy ratio of the characteristic IMF to the raw signal was proposed to construct flow pattern maps with the volumetric void fraction and with the two-phase Reynolds number, respectively. The identification rates of these two maps are verified to be 91.18% and 92.65%. This approach provides a cost-effective solution to the difficult problem of identifying gas-liquid flow patterns in the industrial field.

Chemical, petroleum and nuclear systems are only a few of the industrial processes that utilize gas-liquid flow in annular closed channels. However, concentric horizontal annuli flow patterns have received little attention. The ability to precisely characterize two-phase flow patterns using computational techniques is crucial for the production, transportation, and optimization of designs. This current research aims to establish the accuracy of the computational fluid dynamics (CFD) model in predicting the gas-liquid flow pattern in the concentric annulus pipe and validating the flow pattern of liquid holdup with experimental results from the literature. The simulations were done on a test section of a 12.8 m length pipe with a hydraulic diameter of 0.0168 m using air and water as the working fluids. The volume of fluid (VOF) model in Ansys Fluent based on the Eulerian- Eulerian approach in conjunction with the realizable k-ε turbulence model was used to model the gas-liquid flow pattern, i.e. dispersed bubble, elongated bubble, and slug in a horizontal annulus. A comparison of the model with the experimental high-speed video images shows a reasonable agreement for the flow pattern and liquid holdup data.