Detailed measurementsof gas/liquid flow with lowliquid fractions in horizontaland near horizontal pipes

Abstract

This thesis presents detailed measurements of two phase gas/liquid flows in horizontal and near horizontal pipes. The measurement covers mixture velocities in the range from 5 m/s to 15 m/s, liquid fractions in the range from 0.0010 to 0.0100 and pipe inclinations in the range from -5° to +5° from horizontal.

Particle Image Velocimetry (PIV) is used to measure velocity profiles in the center plane of the pipe, and turbulence RMS profiles are found from the velocity measurements. PIV measurements are done in both the gas phase and in the liquid phase. It has been demonstrated that PIV is a powerful tool for measuring velocity fields of multiphase flow in experimental facilities. Rapid progress in digital camera technology is likely to make PIV an even better measurement technique in the future.

The phase distribution and interface position are measured by a traversable gamma densitometer. The gamma densitometer is traversed both horizontally and vertically. The interface position was identified in the center plane from the gamma measurements with vertical gamma beams. This information is very important when interpreting the PIV measurements, since it can be difficult to identify the interface position from the PIV measurements alone.

Flow patterns are determined by visual observations. Still pictures and high-speed videos are used for the visual observations. The flow patterns were divided into 7 groups: Stratified smooth, stratified wavy without entrainment, stratified wavy with gas bubbles in liquid phase, stratified wavy with liquid droplets in gas phase, stratified wavy with gas bubbles in liquid phase and liquid droplets in gas phase, transition flow pattern and slug flow.

The measurements are compared to simulations using two different commercial multiphase flow simulator tools. Two simulation series are carried out for both simulation tools. Simulation series A was modeled with a geometry that closely follows the physical setup. The results from this simulation series indicate that inlet effects in the simulations reduce the accuracy. In simulation series B, a 50 m long inlet section was added upstream the inlet in the physical setup to remove inlet effects from the simulation results. The results from this simulation series give a better agreement with the experimental results. This shows that the quality of the modeling work has a major impact on the accuracy of the simulations results.