No abstract

Measurement techniques are vital for the control and operation of multiphase oil-water flow in pipes. The development of such techniques depends on laboratory experiments involving flow visualization, liquid fraction ('hold-up'), phase slip and pressure drop measurements. They provide valuable information by revealing the physics, spatial and temporal structures of complex multiphase flow phenomena.This paper presents the hold-up measurement of oil-water flow in pipelines using gamma densitometry and electrical capacitance tomography (ECT) sensors. The experiments were carried out with different pipe inclinations from −5° to +6° for selected mixture velocities (0.2-1.5 m s −1 ), and at selected watercuts (0.05-0.95). Mineral oil (Exxsol D60) and water were used as test fluids. Nine flow patterns were identified including a new pattern called stratified wavy and mixed interface flow. As a third direct method, visual observations and high-speed videos were used for the flow regime and interface identification.ECT and gamma densitometry hold-up measurements show similar trends for changes in pipeline inclinations. Changing the pipe inclination affected the flow mostly at lower mixture velocities and caused a change of flow patterns, allowing the highest change of hold-up. ECT hold-up measurements overpredict the gamma densitometry measurements at higher input water cuts and underpredict at intermediate water cuts. Gamma hold-up results showed good agreement with the literature results, having a maximum deviation of 6%, while it was as high as 22% for ECT in comparison to gamma densitometry.Uncertainty analysis of the measurement techniques was carried out with single-phase oil flow. This shows that the measurement error associated with gamma densitometry is approximately 3.2%, which includes 1.3% statistical error and 2.9% error identified as electromagnetically induced noise in electronics. Thus, gamma densitometry can predict hold-up with a higher accuracy in comparison to ECT when applied to oil-water systems at minimized electromagnetic noise.

Flow patterns in oil-water carrying pipes vary due to the flow characteristics, fluid properties and pipe inclination. For inclined pipes, the gravity component along the pipe influences the flow patterns. Plug flow (PF) is one special flow pattern that occurs in slightly upward inclined pipelines. Mineral oil-Exxsol D60 (viscosity = 1.6 mPa.s, density =788 kg/m 3) and water (viscosity = 1mPa.s, density = 997 kg/m 3) were used as test-fluids. A test matrix was carried out to determine the possible flow patterns that occur at upward pipe inclinations +1° , +3° , +5° and +6° for low mixture velocities (0.2-0.5 m/s), and at water-cut 0.9. The plug flow regime was found only for +5° and +6° inclinations, while no plug flow was noticed at +1° and +3° inclinations. Plug flow was found only for lower flow velocity and higher water-cuts. Plug flow patterns were identified both through visual observation and by means of high-speed video imaging. Two new flow patterns 'oil droplet clusters in continuous oil and water (OC/O&W)' and 'distinct oil droplet clusters in water (D-OC/W)' were introduced, and they occur around the plug flow regime. High-speed images were post-processed for determination of the oil-water interface and subsequently used to calculate the water holdup. The time averaged water hold up decreased with increasing mixture velocity due to the decrease of oil-water slip as a result of increased degree of dispersion. The oil plugs entrained more droplets as mixture velocity was increased, leading to high-frequency fluctuations of volume fraction of the oil plugs. Holdup increased with increasing inclination due to the onset of plug flow, which leads to increased slip. The pressure drop over the test section was measured, and the frictional pressure drops were calculated using average water holdup values. The frictional pressure drop increased with increasing mixture velocity, due to increased mixing and subsequent increase of effective viscosity. The frictional pressure drop decreased with increased inclination due to the appearing of oil plugs and the drag reduction effect associated with the plug flow.

No abstract

Computational Fluid Dynamics (CFD) was used to analyze the flow behaviour of oil and water in the porous heavy oil reservoir with an Autonomous Inflow Control Valve (AICV) installed. 2D simulations in ANSYS FLUENT 13.0 for a 50m long heavy oil reservoir section along the well were carried out with four AICVs placed on the main production pipe. Two different cases were modelled by changing the volume fraction discritization scheme and the way of creating the pressure drop across the AICV. Both cases showed a successful imitation of AICV behaviour by automatically closing the valves completely for water and keeping them open for oil. The simulations performed under an implicit volume fraction scheme resulted in later breakthrough times and a diffused oil layer-like approach within the reservoir. The simulations performed under an explicit volume fraction scheme resulted in earlier breakthrough times and distinct flow behaviour of oil, which was identified as fingering phenomena. The oil production rates, breakthrough times, water migration within the reservoir and annulus were observed for both approaches. The accumulated oil production varies from 70 to 111.6m 3 due to different breakthrough times.