An oil-water flow pattern classification and characterization for wellbores is proposed based on the integrated analysis of experimental data, including frictional pressure drop, holdup, and spatial phase distribution, acquired in a transparent test section (2-in. i.d., 51-ft long) using a refined mineral oil and water ?o/?w=0.85, ?o/?w=20.0, and ?o-w =33.5 dyn/cm at 90°F). The tests covered inclination angles of 90°, 75°, 60°, and 45° from horizontal.

The oil-water flow patterns have been classified into two major categories given by the status of the continuous phase, including water-dominated flow patterns and oil-dominated flow patterns. It was found that most water-dominated flow patterns show significant slippage but relatively low frictional pressure gradients. In contrast, all the oil-dominated flow patterns exhibit negligible slippage but significantly larger frictional pressure gradients. Six flow patterns have been characterized in upward vertical flow; three were water dominated and three were oil dominated. In upward inclined flow there were four water-dominated flow patterns, two oil-dominated flow patterns and a transitional flow pattern. Flow-pattern maps for each of the tested inclination angles are presented. A mechanistic model to predict flow-pattern transitions in vertical wells is proposed. The transitions to the very-fine-dispersed flow patterns were evaluated by combining the concepts of turbulent kinetic energy with the surface free energy of the droplets, while the transitions to the churn flow pattern and the phase inversion were predicted based on the concept of agglomeration. The model compares favorably with the measured data.


Two-phase flow of oil and water is commonly encountered in wellbores; however, its hydrodynamic behavior under a wide range of flow conditions and inclination angles constitutes a relevant unresolved issue for the oil industry.

Multiphase flows are characterized by the existence of diverse flow configurations or flow patterns, which can usually be identified by a typical geometrical arrangement of the phases in the pipe. Inherent to each flow pattern are characteristic spatial distributions of the interface, flow mechanisms, and distinctive values for design parameters such as pressure gradient, holdup, and heat-transfer coefficient. There is clear evidence that accurate knowledge of oil-water flow patterns, their ranges of existence as a function of flow rates and pipe inclination angles, and values for their associated hydrodynamic parameters are crucial in a number of production-engineering applications. These include production optimization, optimum string selection, production-logging interpretation, downhole metering, and artificial lift design and modeling. Additionally, the understanding of oil-water flow in wellbores is fundamental in determining the volumes of free water in contact with the pipe that could cause scaling and corrosion of the pipe after prolonged exposure. This paper addresses the fundamental problem of identifying and characterizing oil-water flow patterns and predicting flow-pattern transitions for conditions pertinent to oil-water producing wells.

Literature Review

Despite being a subject of permanent interest for the petroleum industry, the issue of oil-water flow patterns in wellbores has barely been addressed in the technical literature. A limited number1–3 of experimental studies have been found that provide a description of the flow patterns for low-medium viscosity oil and water. Govier, Sullivan, and Wood1 and Zavareh, Hill, and Podio2 conducted investigations in vertical flow, while Scott3 and Zavareh, Hill, and Podio covered inclinations near the horizontal and near the vertical, respectively. A larger number of related studies concentrated on measuring the holdup, mainly for oil-dispersed systems, focusing on finding expressions for the slip velocity that could be used in the interpretation of production logs. More recently, Hasan and Kabir4 and Tabeling et al.5 provided the first insights into the modeling of oil-water flow in wellbores.

The vertical oil-water data available in the literature1,2,6 suggest that the flow patterns can be grouped into two major categories, including a water-dominated region and an oil-dominated region. Also, Govier, Sullivan, and Wood reported a highly turbulent region between the oil- and water-dominated regions, that was designated as slug flow, in analogy to gas-liquid flow. However, the flow-pattern classification and designation in analogy to gas-liquid flow does not provide a clear and distinct description applicable to oil-water flow. For instance, the so-called slug flow denotes a disorganized distribution of different size droplets but without the presence of a clearly defined bullet-shaped droplet that characterizes slugs and Taylor bubbles in gas-liquid slug flow. Moreover, the fundamental slugging mechanisms of pickup and shedding are not likely present in oil-water flow.

The investigation of Zavareh, Hill, and Podio supports the existence of water- and oil-dominated regions; however, it did not distinguish any flow pattern at the transition between the two flow regions in vertical flow. Since their study was conducted in a large-diameter pipe, Zavareh, Hill, and Podio gave only limited information of the flow patterns occurring at high flow rates. Nevertheless, they identified a finely dispersed flow pattern called dispersed-bubble flow, occurring at high water flow rates, that was not reported by Govier, Sullivan, and Wood.

The two studies2,3 conducted in inclined pipes were limited to flow patterns in the water-dominated region. There are no published flow pattern data in the oil-dominated region for any inclination other than vertical and horizontal, and no information on the oil-water flow patterns existing for inclinations ranging from 30° to 75°. There are very limited data relating flow patterns and their associated hydrodynamic parameters, including holdup and pressure drop, for most operational oilfield conditions.

This content is only available via PDF.
You can access this article if you purchase or spend a download.