Abstract

The discharge of two-phase flow from a co-currently flowing gas-liquid region through single or multiple branches is an important process in many industrial applications including oil-gas production and nuclear power plants. Accurate physical descriptions of the flow phenomena involved, along with the quality and mass flow rate of the discharging streams, is necessary to adequately predict the different phenomena associated with the process.

A test facility was developed, consisting of a horizontal pipe with an inlet diameter of 50.8 mm and three 6.35 mm diameter branches located at a distance of 1.8 m from the pipe inlet. The branches were machined perpendicularly into the test section wall, and oriented at 0, 45, and 90 degrees down from horizontal. Air and water, operating at 206 kPa, were used to provide a two-phase flow regime. Both fluids flowed co-currently within the inlet, and mainly in the stratified regime, but transitions to wavy and slug regimes were observed.

Extensive experimental data are reported for the three branch orientations. The relation between the air-water interface height, the inlet superficial gas and liquid velocities, and the branch two-phase quality and mass flow rate are presented for each branch orientation. The critical inlet conditions leading to beginning of two-phase flow in the branch, the onsets of gas and liquid entrainment, respectively, were reported in both single and dual branch cases. Effects of inlet measurement location, the secondary branch Froude number, and branch fluid phase on the critical conditions were investigated. A novel map relating the dual discharge branch Froude numbers, the inlet superficial liquid velocity, and the related dual branch phenomena was developed. The map presented the three observed modes of gas entrainment during dual discharge.

A two-fluid separated theoretical model was developed in order to predict the critical height at the onset of gas entrainment in a bottom branch. Potential flow theory leads to the branch being simulated by a point-sink, while the flowing liquid upstream of the branch was simulated by a uniform constant crossflow velocity. Two analytical criteria were used to predict the dip position (height and offset distance) relative to the branch. Inaccuracies with experiments lead to the inclusion of empirical terms to satisfy the local crossflow velocities within the inlet. A digital imaging technique was also developed in order to record local interface profiles at the onset of gas entrainment, and was used to satisfy the relationship between the dip height and offset distance. The semi-empirical approach provided a significant improvement over the purely analytical model, and demonstrated that the critical height to be predicted within a reasonable error.