Drift-Velocity Closure Relationships for Slug

Two-Phase High-Viscosity Oil Flow in Pipes

B.C. Jeyachandra and B. Gokcal, University of Tulsa; A. Al-Sarkhi, King Fahd University of Petroleum & Minerals;
and C. Sarica and A.K. Sharma, University of Tulsa

Summary
The drift velocity of a gas bubble penetrating into a stagnant
liquid is investigated experimentally in this paper. It is part of
the translational slug velocity. The existing equations for the drift
velocity are either developed by using the results of Benjamin
(1968) analysis assuming inviscid fluid flow or correlated using
air/water data. Effects of surface tension and viscosity usually are
neglected. However, the drift velocity is expected to be affected by
high oil viscosity. In this study, the work of Gokcal et al. (2009)
has been extended for different pipe diameters and viscosity range.
The effects of high oil viscosity and pipe diameter on drift velocity for horizontal and upward-inclined pipes are investigated. The
experiments are performed on a flow loop with a test section with
50.8-, 76.2-, and 152.4-mm inside diameter (ID) for inclination
angles of 0 to 90. Water and viscous oil are used as test fluids.
New correlation for drift velocity in horizontal pipes of different
diameters and liquid viscosities is developed on the basis of experimental data. A new drift-velocity model/approach are proposed for
high oil viscosity, valid for inclined pipes inclined from horizontal
to vertical. The proposed comprehensive closure relationships are
expected to improve the performance of two-phase-flow models
for high-viscosity oils in the slug flow regime.
Introduction
The translational velocity (velocity of slug units), is one of the key
closure relationships in two-phase-flow modeling. It is described as
the summation of the maximum mixture velocity in the slug body
and the drift velocity. The drift velocity and translational velocity
are affected by high oil viscosity. High-viscosity oils are being
produced from many oil fields around the world. Oil-production
systems are currently flowing oils with viscosities as high as 10
Pas. Current multiphase-flow models and correlations are largely
based on experimental data with low-viscosity liquids. The laboratory liquids commonly used have viscosities less than 0.02 Pas.
Multiphase flows are expected to exhibit significantly different
behavior for higher-viscosity oils.
Gokcal et al. (2008) experimentally observed slug flow to be
the dominant flow pattern for high-viscosity oil and gas flows.
Accurate predictions of slug-flow characteristics are crucial in the
design of pipelines and process equipment. In order to improve
the accuracy of slug characteristics for high-viscosity oils, new
and improved models for flow characteristics such as drift velocity
and translational velocity are required.
Slug translational velocity is the sum of the bubble velocity in
stagnant liquid (i.e., the drift velocity vd) and the maximum velocity in the slug body. Research efforts have typically been focused
on the drift velocity in horizontal and upward-inclined pipes.
Nicklin et al. (1962) proposed an equation for translational
velocity as
vt = Cs vs + vd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
The parameter Cs is approximately the ratio of the maximum
to the mean velocity of a fully developed velocity profile. Cs is
Copyright 2012 Society of Petroleum Engineers
This paper (SPE 151616) was accepted for presentation at the SPE Annual Technical
Conference and Exhibition, Tuscany, Italy, 2022 September 2010, and revised for
publication. Original manuscript received 11 November 2010. Revised manuscript received
14 July 2011. Paper peer approved 21 July 2011.

June 2012 SPE Journal

approximately 1.2 for turbulent flow and 2.0 for laminar flow. vs
is the mixture velocity, which is the sum of the superficial liquid
and gas velocities. The drift velocity contributes to the translational
velocity of the slug unit for all pipe inclination. The value of the
drift velocity will be the translational velocity at zero mixture
velocity (the intersection point of the vertical axis on the vt-vs.-vs
curve). For horizontal pipes, the drift velocity is acting in the same
direction as the mixture velocity so it contributes to the magnitude
of the slug translational velocity.
Vertical Flow. Dumitrescu (1943) and Davies and Taylor (1950)
performed a potential ow analysis to nd the drift velocity for
vertical ow. Both derived the same dimensionless group (Froude
number) and found that Froude number has a constant value.
Davies and Taylor (1950) estimated the constant value as 0.328.
Dumitrescu (1943) made more-accurate calculations and theoretically determined this value as 0.351, which agreed well with the
air/water experimental data of Nicklin et al. (1962).
vd = 0.351 gD .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)

For vertical flow, Joseph (2003) proposed a model for the bubble-rise velocity in vertical flow, taking viscosity, surface-tension,
and shape of the bubble-nose effects into consideration. From the
experimental results, it is observed that the bubble nose is almost
spherical. When the bubble nose is spherical (axisymmetric cap),
the effect of the surface tension vanishes and the equation becomes
a function of only the fluid viscosity and the radius of the spherical-cap bubble, as shown in Eq. 3:
vd =

4
4
16 2
, . . . . . . . . . . . . . . . . . . . . . (3)
+
gr +
3 r
9
9 ( r ) 2

where r is the radius of cap and and are the density and viscosity of the liquid, respectively. It was shown that the experimental
data of Bhaga and Weber (1981) and the model predictions were
in good agreement.
Inclined Flow. For the inclined case, Zukoski (1966), Bendiksen
(1984), Weber et al. (1986), Hasan and Kabir (1986), and Carew
et al. (1995) experimentally studied drift velocity and found
that the drift velocity increases with inclination angle and then
decreases to its lowest value for vertical ow, reaching a maximum
value at an intermediate angle of inclination approximately 40
to 60 from the horizontal. This fact was explained qualitatively
by Bonnecaze et al. (1971). They discussed that the gravitational
potential rst increases and then decreases as the inclination angle
changes from the vertical to the horizontal position.
Weber et al. (1986) experimentally studied bubble-rise velocity
(in relatively small pipe diameters from 0.6 to 3.7 cm) for highviscosity Newtonian liquids. Froude number Fr was correlated as a
function of Eotvos number Eo, the Morton number (M=g4/ 3),
and the inclination angle .
Bendiksen (1984) performed an experimental study for velocities of single elongated bubbles in flowing liquids at different
inclination angles. The measured velocities were plotted against
the liquid velocity for each inclination angle. Then, drift velocities
were found by the extrapolation of the data to zero liquid velocity.
He correlated the drift velocity for inclined flow by using the drift
velocities for horizontal and vertical flow:
593

vd = vdh cos + vdv sin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (4)

Hasan and Kabir (1986) performed an experimental study in
the range of 90 > > 30 and proposed the relation
vd = vdv sin (1 + cos )1.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
Carew et al. (1995) studied the motion of long bubbles in
inclined pipes experimentally with viscous Newtonian and nonNewtonian liquids. They proposed an empirical correlation for
the drift velocity of an elongated bubble in inclined pipes. The
correlation depends on inclination angle and surface tension and
is valid at Re > 200 and Eo > 60.
Shosho and Ryan (2001) experimentally investigated the
effects of tube size and fluid type (including Newtonian and nonNewtonian fluids) on drift velocity for vertical and inclined tubes.
The drift velocity in terms of Froude number was correlated
with Eotvos and Morton numbers. Froude number increased and
decreased as the angle of inclination increased for both Newtonian
and non-Newtonian fluids. For non-Newtonian fluids with high
Morton number, Froude number was affected by both viscous
forces and tube size.
Alves et al. (1993) proposed a model for the drift velocity
including surface-tension effect in inclined flow using inviscid
flow theory. The model was compared against their experimental
data and Zukoski (1966) data. Gokcal (2008) used the Alves et al.
(1993) results and proved that the effect of surface tension on drift
velocity is negligible when the ID is 50.8 mm.
Van Hout et al. (2002) used an approach similar to that of this
study to identify inclination effects on drift velocity using optical
sensors. Two different pipe diameters (0.024 and 0.054 m) were
studied. The fluids were air and water. They found that the Bendiksen (1984) correlation predicted well for smaller diameter. But
as the pipe diameter increased, the correlation was valid only for
slightly inclined case. Substantial discrepancy was found when the
correlation was applied to higher inclination angles.
Horizontal Flow. Zukoski (1966) experimentally investigated the
effects of liquid viscosity, surface tension, and pipe inclination
on the motion of single elongated bubbles in stagnant liquid for
different pipe diameters. He also found that the effect of viscosity
is negligible on the drift velocity for Re = vd D/ > 200. Wallis
(1969) and Dukler and Hubbard (1975) claimed that there is no
drift velocity for horizontal ow because gravity cannot act in
the horizontal direction. However, Nicholson et al. (1978), Weber
(1981), and Bendiksen (1984) showed that drift velocity exists for
the horizontal case and the value of drift velocity can exceed the
vertical ow value. The drift velocity is a result of hydrostatic pressure difference between the top and bottom of the bubble nose.
Benjamin (1968) proposed the following relationship for the
drift velocity in horizontal pipes:
vd = 0.542 gD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (6)
Benjamin calculated the value of the drift-velocity coefficient
by using inviscid potential flow theory that inherently neglects
surface tension and viscosity. The drift velocity in horizontal slug
flow is the same as the velocity of the penetration of a bubble
when liquid is drained out of a horizontal pipe. Bendiksen (1984)
and Zukoski (1966) supported the study of Benjamin (1968)
experimentally.
Weber (1981) developed a correlation for drift velocity in
horizontal pipes on the basis of the experimental data of Zukoski
(1966) for liquids of low viscosities, as shown in Eq. 7:
vd / gD = 0.54 1.76 Eo

0.56

, . . . . . . . . . . . . . . . . . . . . . . . (7)

where Eotvos number is defined as Eo = D 2 g / .

A dimensional analysis for the drift velocity in horizontal pipes
was presented by Ben-Mansour et al. (2010). The step-by-step
594

method was used, and the following dimensionless numbers have

been found:
Fr =

D2g
vd
, N =
. . . . . . . . . . . . . (8)
3 2 1 2 , Eo =

D g
gD

The first dimensionless group is the Froude number Fr, the second
is the viscosity number N, and the third is Eotvos number Eo. It
was concluded that the drift velocity in a horizontal pipe can be
modeled using those three dimensionless groups.
From the literature review related to drift velocity for horizontal, inclined, and vertical pipes, it is apparent that detailed research
has been conducted on the effects of surface tension and pipe
diameter on drift velocity at different inclination angles. However,
for the effect of high viscosity on drift velocity, experimental and
theoretical studies are scarce and have been conducted for relatively small pipe diameter.
Shi et al. (2005a, b) recently conducted experimental and
modeling studies to determine the drift-flux model parameters of
water/gas, oil/water, and oil/water/gas flow in a 15-cm-diameter
pipe at different angles ranging from vertical to slightly downward. Experiments were performed with kerosene, tap water, and
nitrogen. The viscosity of the oil was 1.5 cp. A correlation for drift
velocity similar to that of Hasan and Kabir (1999) is used in their
drift-flux modeling of oil/water flow and similar to Wallis (1969)
for gas/liquid flows.
The approach in this study differs from that of the previous
studies in that single-viscosity oil was used to conduct the experiments at different viscosities by controlling the temperature of the
oil. The oil-viscosity range varied from 0.155 to 0.574 Pas. The
other salient feature is that the large diameters that are prevalent
in the field were selected. Experiments were conducted on 50.8-,
76.2-, and 152.4-mm-diameter acrylic pipe for inclination angles
from 0 to 90. The purpose of the present paper is to develop a unified drift-velocity closure relationship based on viscosity and pipe
diameter and compare it with experimental results for horizontal
and upward inclined pipes.
Experimental Setup and Procedure
The experimental facility consists of an oil-storage tank, a 20-hp
screw pump, a 3.05-m-long acrylic pipe, heating and cooling loops,
and transfer hoses and instrumentation. Details of the experimental setup are given in Gokcal et al. (2008) and shown in Fig. 1.
Experiments were conducted on 50.8-, 76.2-, and 152.4-mm-ID
pipes. The acrylic pipe is located close to the storage tank. The
inclination of the pipe can be varied using a pulley arrangement.
The pipe inclination can be changed from 0 to 90.
The heating and cooling loops are used to maintain the desired
temperature and thereby control the viscosity of the oil. The oil
pump supplies the pipe with oil. Then, the main inlet valve and
the auxiliary inlet valve are closed. The drainage valve is opened
to drain the residual oil captured and thereby create a gas pocket.
Next, the drainage valve is closed and the main inlet valve is
opened to release the gas bubble into the stagnant oil column. The
drift velocity is measured by two lasers (for 50.8- and 76.2-mmID pipe) or optical sensors (for 152.4-mm-ID pipe) separated by
a distance of 0.9144 m. The optical sensors work by the principle
that the light intensity changes when it reflects from/refracts
through the oil or the gas phase. This is stored as voltage readings
in a data-acquisition system with a frequency of 500 readings/sec.
The data are used to calculate the drift velocity by dividing the
distance between the two sensors by time difference between the
two voltage peaks.
The facility was modified for the horizontal case by replacing
the end plate of the pipe with a plug. This facilitated proper draining of oil as the gas bubble penetrated into the liquid.
Water and viscous oil were used as test fluids. The properties of
the oil are given in Table 1. The most important characteristic of
the oil is its large range of viscosity owing to strong temperature
dependence. The oil sample was tested before experiments, and
the surface tension remained constant at 29 to 30 dynes/cm for the
temperature change used in the present experiments. Oil-viscosity
June 2012 SPE Journal

Fig. 1Schematic of indoor high-viscosity test facility.

TABLE 1PROPERTIES OF OIL

Gravity at 15.6C
(API )

Density at 15.6C
3
(kg/m )

Flash Point
(C)

Pour Point
(C)

Viscosity at 40C
(Pas)

27.6

889

250

12.2

0.22

and density variation with temperature are shown in Figs. 2 and 3,

respectively.

conducted at different oil temperatures corresponding to a range

of viscosities from 0.154 to 0.574 Pas.

Results and Discussions

Initially, an experiment is conducted with water for horizontal pipe
to prove that the system is working properly. The results for water
are compared with the Benjamin (1968) model prediction. The predictions of drift velocity of the water from Benjamins model show
excellent agreement with the data. The rest of the experiments are

Drift Velocity in Horizontal Pipes. Fig. 4 shows the experimental

results for drift velocity vs. viscosity for horizontal pipes of different diameters. The rst point for all plots (at lowest viscosity =
0.001 Pas) is for the water case, and the other points are for oil.
Drift velocity decreases with increasing viscosity. As the pipe
diameter increases, the drift velocity increases. The decrease in

1
0.9
0.8

Viscosity (Pas)

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
10

20

30

40

50

60

70

Temperature (C)
Fig. 2Oil viscosity vs. temperature.
June 2012 SPE Journal

595

895

890

Density (kg/m3)

885

880

875

870

865

860

855
0

10

20

30

40

50

60

70

Temperature (C)
Fig. 3Oil density vs. temperature.

the drift velocity with viscosity is steeper in small pipes than in

large pipes. The plots tend to have an asymptotic level at very
high viscosity, which leads to a small variation in drift-velocity
variation with viscosity.
As shown by Ben-Mansour et al. (2010), the drift velocity in a
horizontal pipe can be correlated using Froude number, viscosity
number, and Eotvos number. Fig. 5 shows the correlation for drift
velocity in horizontal pipes at different viscosities, diameters, and
surface tensions. It is worth noting that data include oil at different
viscosities and water in different pipe diameters. The experimental
data for drift velocity in horizontal pipes of different diameters
with liquid of various viscosities are well correlated by Eq. 9:
Fr = 0.53 e

13.7 N 0.46 Eo0.1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . (9)

As mentioned in the Horizontal Flow subsection, the effect of

surface tension is negligible for a pipe diameter larger than 50 mm.
However, for horizontal flow, the effects of surface tension on drift
velocity were considered in the equation by using Eotvos number
(Eotvos number in Eq. 9 is defined as it was defined in Eq. 8).
Evaluation of Drift-Velocity Correlation in Horizontal Pipes.
Fig. 6 shows the comparison between the measured data and the
predicted result using Eq. 8. On the same plot, the prediction by
Weber (1981) for low viscosities is shown. It can be seen clearly that
most of the data collapse in a range of +8% and 13%. The Weber
(1981) correlation matches only for points at high values of Froude
number, which is expected because the correlation was developed on
the basis of the Zukoski (1966) data in a water/air system. The data
points where Fr is approximately 0.5 are for water experiments.

0.7
D=50.8 mm

Drift Velocity (m/s)

0.6

D=76.2 mm
D= 152.4 mm

0.5
0.4
0.3
0.2
0.1
0
0

0.1

0.2

0.3
0.4
Viscosity (Pas)

0.5

0.6

0.7

Fig. 4Drift-velocity variation with viscosity in horizontal pipes.

596

June 2012 SPE Journal

0.6
Data
Equ. (8)

Froude Number (Fr)

0.5
0.4
0.3
0.2
0.1
0
0.0001

0.001

0.01

0.1

N
0.46 / Eo 0.1
Fig. 5Horizontal Froude-number correlation vs. experimental data.

Drift Velocity in Vertical Pipes. Several correlations are available

in literature for rise velocity in stagnant uid (Davies and Taylor
1950; White and Beardmore 1962; Joseph 2003). Fig. 7 shows
the comparison of the model predictions of Joseph (2003) with
experimental data from Weber et al. (1986), Shosho and Ryan
(2001), and this study. The bubble radius and liquid viscosity must
be known to calculate the drift velocity from Eq. 3. It is experimentally observed that the radius of a bubble is approximately
0.55 to 0.6 times the radius of the pipe. This value is used for the
remaining calculations to compare model predictions with experimental results. Weber et al. (1986) performed their experiments
in 37.3-mm-ID pipe (relatively small) for viscosities between
0.051 and 0.183 Pas. The Shosho and Ryan (2001) experiments
were for same diameter for viscosities between 0.003 and 0.883
Pas. They are the only available data set for a higher-viscosity
range with comparable pipe diameter in the literature. The Joseph
(2003) model correlated the experimental data well, within 20%.
The present works experimental data agreed very well with this
correlation. A better match can be obtained if the curvature of the
bubble cap is included.
As mentioned in the Horizontal Flow subsection, the effect of
viscosity is negligible for pipe diameters larger than 50 mm. For
vertical flow, in the Joseph (2003) model, viscosity and surface

Froude Number -Predicted

0.6
0.5
0.4
0.3
+8%
0.2
-13%

Present

0.1

Weber(1981)

0
0

0.1

0.2

0.3

0.4

0.5

0.6

Froude Number -Measured

Fig. 6Comparison of correlation predictions with measured
Froude number for horizontal flow.
June 2012 SPE Journal

tension and shape of bubble-nose effects on drift velocity were

considered in his model in Eq. 3.
Drift Velocity in Upward-Inclined Pipes. Water/Air Case. The
drift velocity for water/air experiments in large diameter will be
a function of only pipe diameter and inclination angle (surfacetension effect is negligible). Fig. 8 shows the measured Froude
number in different pipe diameters for air/water system. Froudenumber plots vs. pipe inclination follow similar behavior for
all pipe diameters. The new closure relationship correlated with
Froude number at any inclination angle and pipe diameter is proposed in Eq. 10. Evaluation of Eq. 10 is shown in Fig. 9. All the
data are well predicted within 5%.
Fr = 0.248 2 + 0.299 + 0.497 . . . . . . . . . . . . . . . . . . . . . (10)
Oil/Air Case. Effect of pipe diameter and inclination angle at
certain oil viscosity is presented in Figs. 10 through 13. Effect of viscosity at certain pipe diameters is shown in Figs. 14 through 16.
From Figs. 10 through 16, it is observed that there is a clear
effect of diameter on drift velocity. For horizontal flow, when
diameter increases, the gravitational potential increases. This leads
to a stronger drive for the gas bubble to penetrate into the stagnant
liquid column, hence the higher drift velocity. As the inclination
gradually increases, so does the drift velocity, and it peaks at an
inclination of approximately 30 to 50. The gravitational potential
is at its maximum at this inclination angle. As the inclination is
increased further, the effect of drainage area comes into consideration. Even though the gravitational potential is high, the area
available for oil to drain is low. This, in turn, creates a resistance
for the air bubble to penetrate into the oil, effectively reducing
the drift velocity. With increasing pipe inclination, the extended
air bubble location, which is in contact with the upper part of the
pipe (at low inclination angle), moves toward the centerline of the
pipe. At a right angle, the extended air bubble location is in the
center of the pipe.
From Figs. 14 through 16, it is observed that as the viscosity increases, drift velocity decreases. As viscosity increases, the
resistance for the gas bubbles to intrude into the stagnant column increases. Thereby, the drift velocity is reduced. The same
parabolic trend of drift velocity increasing as the inclination angle
increases, reaching a maximum at 30 to 50 and then decreasing,
is observed.
Drift-Velocity Correlation in Upward-Inclined Pipes. Fig. 17
shows the comparison between the experimental data and the
597

Drift Velocity - Predicted (m/s)

0.5

Present Data
Weber et al.
Shosho and Ryan

0.4

+20%

0.3

-20%

0.2

0.1

0
0

0.1

0.2

0.3

0.4

0.5

Drift Velocity - Measured (m/s)

Fig. 7Comparison of Joseph (2003) model predictions with measured drift velocities for vertical case.
0.7

Froude Number - Predicted

0.7

Froude Number (Fr)

0.6
0.5
0.4
0.3

Present, 50.8-mm
Present, 76.2-mm

0.2

Present, 152.4-mm
Alves (1993)

0.1

Present
Zukoski (1966)
Alves (1993)

0.6
0.5
0.4

+5%

0.3

-5%
0.2
0.1

Zukoski (1966)

0
0

10

20

30

40

50

60

70

80

90

0.1

Inclination Angle ()

0.4

0.5

0.6

0.7

Fig. 9Comparison of measured Froude number with Eq. 9

predictions.

0.7

0.7

0.6

Drift Velocity (m/s)

0.6
Drift Velocity (m/s)

0.3

Froude Number - Measured

Fig. 8Measured Froude number vs. inclination angle for

water.

0.5
0.4
0.3
0.2

152.4 -mm
50.8-mm
10

20

30
40
50
60
Inclination Angle ()

70

80

0.4
0.3
0.2
152.4-mm
76.2-mm
50.8-mm

0
0

0.5

0.1

76.2-mm

0.1

90

Fig. 10Effect of pipe diameter on drift velocity for 0.574Pas oil.

598

0.2

10

20

30

40

50

60

70

80

90

Inclination Angle ()
Fig. 11Effect of pipe diameter on drift velocity for 0.378Pas oil.
June 2012 SPE Journal

0.7

Drift Velocity (m/s)

Drift velocity (m/s)

0.7
0.6
0.5
0.4
0.3
0.2

152.4-mm

0.5
0.4
0.3

50.8-mm

10

20

30

40

50

60

70

76.2-mm
50.8-mm

0
0

0
0

152.4-mm

0.2
0.1

76.2-mm

0.1

0.6

80

10

20

90

Inclination Angle ()

30

40

50

60

70

80

90

Inclination Angle ()
Fig. 13Effect of pipe diameter on drift velocity for 0.154Pas oil.

0.7

0.45

0.6

0.4

Drift Velocity (m/s)

Drift Velocity (m/s)

Fig. 12Effect of pipe diameter on drift velocity for 0.256Pas oil.

0.5
0.4
0.3

0.574

0.2

0.378

0.1

10

20

30

40

50

60

0.3
0.25
0.2

0.1

0.154

0.05

70

80

90

0.574

0.15

0.256

0
0

0.35

0.378
0.256
0.154

0
0

10

Inclination Angle ()
Fig. 14Viscosity effect on drift velocity for 152.4.-mm-diameter pipe (viscosities in Pas).

20

30 40 50 60 70
Inclination Angle ()

80

90

Fig. 15Viscosity effect on drift velocity for 76.2-mm-diameter

pipe.

0.45

Drift Velocity (m/s)

0.40

0.35

0.30

0.25

0.20

0.15

0.10
0

10

20

30

40

50

60

70

80

90

Inclination Angle ()
0.645 Pa.s

0.412 Pa.s

0.296 Pa.s

0.185 Pa.s

0.104 Pa.s

0.001 Pa.s

Fig. 16Viscosity effect on drift velocity for 50.8-mm-diameter pipe (Gokcal et al. 2008).
June 2012 SPE Journal

599

References

0.7

Froude Number - Measured

0.6

+12%
0.5
0.4

-12%

0.3
0.2
Oil Data

0.1

water Data

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Froude Number - Predicted

Fig. 17Comparison of the correlation predictions with measured Froude number for inclined flow.

predicted result using an approach similar to that of Bendiksen

(Eq. 4). The modified Bendiksen correlation shown in Eq. 11
for all pipe inclination simply uses the Froude number instead of
the drift velocity. The horizontal and vertical components of the
Froude number in the equation are obtained from the experimental
values for 0 and 90, respectively, which accounts for the effects
of viscosity.
Fr = Fr h cos + Fr v sin , . . . . . . . . . . . . . . . . . . . . . . . . . (11)
where Fr is the Froude number at angle of inclination, is the
angle of inclination, and h and v are for the horizontal and vertical
cases, respectively.
The predicted values in Fig. 17 are calculated by substituting
Eq. 3 in terms of Froude number for Fr v and Eq. 8 for Fr h.
It can be observed that most of the data points for high-viscosity
oil fall within the 12% limit. It can also be observed that all data
points for water lie very close to the 45 line. These observations
imply that there is good match between the predicted value and the
experimental data, provided that viscosity effects are considered.
Conclusions
Drift-velocity experiments were conducted on high-viscosity oil
for varying viscosities, pipe diameters, and inclination angles. It
was observed that viscosity has a profound effect on drift velocity.
Drift velocity increases with increasing pipe diameter and decreases
with increasing liquid viscosity. For horizontal flow, a dimensionless analysis was performed, and a new correlation for horizontal
drift velocity was developed and tested with the available data set.
For the inclined case, the effect of pipe inclination on drift velocity
was explained. As the pipe inclination increases, the drift velocity increases, and it peaks at an inclination near 30 to 50, then it
decreases again. For the inclined case of an air/water system, a new
correlation for different pipe diameter was developed. For the vertical
case of an oil/air system, the correlation developed by Joseph (2003)
was tested with the data set, and a very good match was obtained.
For inclined flow, it was observed that the correlation developed
by Bendiksen (1984) modified by using Froude number instead of
drift velocity worked well, provided that the viscosity effects are
considered for the horizontal and vertical components. Therefore,
the correlation developed for horizontal flow and the Joseph (2003)
correlation for vertical flow can be used in conjunction with the
modified Bendiksen equation to provide accurate values for drift
velocity in the horizontal and the upward-inclined case.
Acknowledgment
The authors would like to thank the member companies of the
TUFFP for their support.
600

Alves, I.N., Shoham, O., and Taitel, Y. 1993. Drift velocity of elongated
bubbles in inclined pipes. Chem. Eng. Sci. 48 (17): 30633070. http://
dx.doi.org/10.1016/0009-2509(93)80172-M.
Bendiksen, K.H. 1984. An experimental investigation of the motion of long
bubbles in inclined tubes. Int. J. Multiphase Flow 10 (4): 467483.
http://dx.doi.org/10.1016/0301-9322(84)90057-0.
Benjamin, T.B. 1968. Gravity currents and related phenomena. J. Fluid Mech.
31 (2): 209248. http://dx.doi.org/10.1017/S0022112068000133.
Ben-Mansour, A., Sharma, A.K., Jeyachandra, B., Gokcal, B., Al-Sarkhi,
A.S., and Sarica, C. 2010. Effect of pipe diameter and high oil viscosity
on drift velocity for horizontal pipes. Paper presented at the 7th
North American Conference on Multiphase Technology, Banff, Canada,
24 June.
Bhaga, D. and Weber, M.E. 1981. Bubbles in viscous liquids: shapes,
wakes, and velocities. J. Fluid Mech. 105: 6185. http://dx.doi.org/
10.1017/S002211208100311X.
Bonnecaze, R.H., Eriskine Jr., W., and Greskovich, E.J. 1971. Holdup and
pressure drop for two-phase slug flow in inclined pipelines. AIChE J.
17 (5): 11091113. http://dx.doi.org/10.1002/aic.690170516.
Carew, P.S., Thomas, N.H., and Johnson, A.B. 1995. A physically based
on correlation for the effects of power law rheology and inclination on
slug bubble rise velocity. Int. J. Multiphase Flow 21 (6): 10911106.
http://dx.doi.org/10.1016/0301-9322(95)00047-2.
Davies, R.M. and Taylor, G.I. 1950. The Mechanics of Large Bubbles
Rising through Extended Liquids and through Liquids in Tubes. Proc.
R. Soc. London, A 200 (1062): 375390. http://dx.doi.org/10.1098/
rspa.1950.0023.
Dukler, A.E. and Hubbard, M.G. 1975. A Model for Gas-Liquid Slug Flow
in Horizontal and Near-Horizontal Tubes. Ind. Eng. Chem. Fundam. 14
(4): 337347. http://dx.doi.org/10.1021/i160056a011.
Dumitrescu, D.T. 1943. Strmung an einer Luftblase im senkrechten Rohr.
ZAMMZeitschrift fr Angewandte Mathematik und Mechanik 23 (3):
139149. http://dx.doi.org/10.1002/zamm.19430230303.
Gokcal, B. 2008. An experimental and theoretical investigation of slug flow
for high oil viscosity in horizontal pipes. PhD dissertation, University
of Tulsa, Tulsa, Oklahoma.
Gokcal, B., Al-Sarkhi, A., and Sarica, C. 2009. Effects of High Oil Viscosity on Drift Velocity for Horizontal and Upward Inclined Pipes.
SPE Proj Fac & Const 4 (2): 3240. SPE-115342-PA. http://dx.doi.
org/10.2118/115342-PA.
Gokcal, B., Al-Sarkhi, A.S., and Sarica, C. 2008. Effects of High Oil
Viscosity on Drift Velocity for Horizontal Pipes. Paper presented at
the 6th North American Conference on Multiphase Technology, Banff,
Canada, 46 June.
Gokcal, B., Wang, Q., Zhang, H.-Q., and Sarica, C. 2008. Effects of
High Oil Viscosity on Oil/Gas Flow Behavior in Horizontal Pipes.
SPE Proj Fac & Const 3 (2): 111. SPE-102727-PA. http://dx.doi.
org/10.2118/102727-PA.
Hasan, A.R. and Kabir, C.S. 1986. Predicting Multiphase Flow Behavior
in a Deviated Well. SPE Prod Eng 3 (4): 474482. SPE-15449-PA.
http://dx.doi.org/10.2118/15449-PA.
Hasan, A.R. and Kabir, C.S. 1999. A Simplified Model for Oil/Water Flow
in Vertical and Deviated Wellbores. SPE Prod & Fac 14 (1): 5662.
SPE-54131-PA. http://dx.doi.org/10.2118/54131-PA.
Joseph, D.D. 2003. Rise velocity of a spherical cap bubble. J. Fluid Mech.
488: 213223. http://dx.doi.org/10.1017/S0022112003004968.
Nicholson, M.K., Aziz, K., and Gregory, G.A. 1978. Intermittent two phase
flow in horizontal pipes: Predictive models. The Canadian Journal
of Chemical Engineering 56 (6): 653663. http://dx.doi.org/10.1002/
cjce.5450560601.
Nicklin, D.J., Wilkes, J.O., and Davidson, J.F. 1962. Two-Phase Film Flow
in Vertical Tubes. Trans. Inst. Chem. Eng. 40: 6168.
Shi, H., Holmes, J.A., Diaz, L.R., Durlofsky, L.J., and Aziz, K. 2005b. DriftFlux Parameters for Three-Phase Steady-State Flow in Wellbores. SPE J.
10 (2): 130137. SPE-89836-PA. http://dx.doi.org/10.2118/89836-PA.
Shi, H., Holmes, J.A., Durlofsky, L.J., et al. 2005a. Drift-Flux Modeling of
Two-Phase Flow in Wellbores. SPE J. 10 (1): 2433. SPE-84228-PA.
http://dx.doi.org/10.2118/84228-PA.
Shosho, C.E. and Ryan, M.E. 2001. An experimental study of the motion
of long bubbles in inclined pipes. Chem. Eng. Sci. 56 (6): 21912204.
http://dx.doi.org/10.1016/S0009-2509(00)00504-2.
June 2012 SPE Journal

van Hout, R., Barnea, D., and Shemer, L. 2002. Translational velocities of
elongated bubbles in continuous slug flow. Int. J. Multiphase Flow 28
(8): 13331350. http://dx.doi.org/10.1016/s0301-9322(02)00027-7.
Wallis, G.B. 1969. One Dimensional Two-Phase Flow. New York:
McGraw-Hill.
Weber, M.E. 1981. Drift in intermittent two-phase flow in horizontal pipes.
Can. J. Chem. Eng. 59: 398399.
Weber, M.E., Alarie, A., and Ryan, M.E. 1986. Velocities of extended
bubbles in inclined tubes. Chem. Eng. Sci. 41 (9): 22352240. http://
dx.doi.org/10.1016/0009-2509(86)85073-4.
White, E.T. and Beardmore, R.H. 1962. The velocity of rise of single
cylindrical air bubbles through liquids contained in vertical tubes.
Chem. Eng. Sci. 17 (5): 351361. http://dx.doi.org/10.1016/00092509(62)80036-0.
Zukoski, E.E. 1966. Influence of viscosity, surface tension, and inclination
angle on motion of long bubbles in closed tubes. J. Fluid Mech. 25 (4):
821837. http://dx.doi.org/10.1017/S0022112066000442.
Benin Jeyachandra is a petroleum engineer with Schlumberger
Information Solutions in Houston, Texas, USA. He holds an MS
degree in petroleum engineering from The University of Tulsa.
Bahadir Gokcal is a senior flow assurance engineer with
ConocoPhillips in Houston, Texas, USA. His research interests are
multiphase flow in pipes, CFD modeling, and flow assurance.
Gokcal holds a BS degree in petroleum and natural gas engineering from Middle East Technical University in Turkey, and MS and PhD
degrees in petroleum engineering from The University of Tulsa.

June 2012 SPE Journal

Abdelsalam Al-Sarkhi is an associate professor at King Fahd

University of Petroleum and Minerals, Saudi Arabia. His research
interests are experimentation and modeling of multiphase
flow, thermodynamics, and heat transfer. He conducted several researches on the effect of drag-reducing polymers on
multiphase-flow behavior. He holds BS and MS degrees in
mechanical engineering from Jordan University of Science
and Technology and a PhD degree in mechanical engineering from Oklahoma State University, Stillwater, Oklahoma, USA.
Cem Sarica is a professor of petroleum engineering and the
director of two industry supported consortia at the University of
Tulsa (TU): Tulsa University Fluid Flow Projects (TUFFP) and Tulsa
University Paraffin Deposition Projects (TUPDP). He is also serving
as coprincipal investigator of Tulsa University High Viscosity Oil
Projects (TUHOP). He was as an associate professor of petroleum and natural gas engineering at The Pennsylvania State
University and an assistant professor of petroleum and natural
gas engineering at Istanbul Technical University (ITU) prior to
joining TU. His research interests are production engineering,
multiphase flow in pipes, flow assurance and horizontal wells.
He holds BS and MS degrees in petroleum engineering from ITU
and PhD degree in petroleum engineering from TU. He currently
serves as a member of SPE Projects, Facilities and Construction
Advisory Committee. He has previously served as a member
of SPE Production Operations and Books Committees and was
a member of SPE Journal Editorial Board between 1999 and
2007. He is the recipient of 2010 SPE International Production
and Operations Award. He has more than 100 publications
mostly in SPE Journals and Proceedings.

601