Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: P3.60 - Influence of the pipe diameter in dispersed oil-water flows
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00544
 Material Information
Title: P3.60 - Influence of the pipe diameter in dispersed oil-water flows Non-Newtonian Multiphase Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Plasencia, J.
Nydal, O.J.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: oil-water flow
phase inversion
dispersed flow
multiphase flow
 Notes
Abstract: Oil-water flows are commonly encountered in oil production wells and in transport pipes. During dispersed flows the apparent viscosity of the oil-water mixture increases towards the point of phase inversion, where the continuous phase switches from one to the other. Large scale experiments with real crudes to characterize the flow around the phase inversion point are costly and not practical. Hence, it is of interest to investigate the possibility of characterizing the oil-water flow in small scale setups. The present work reports the investigation of the influence of pipe diameter on dispersed oil-water flows in horizontal pipes. Pressure gradients are measured for oil-water flows in acrylic pipes with diameters of 16, 32 and 60 mm. Experimental results are presented for the oil Marcol 52 (μ = 11,0 mPa.s - ρ = 835 kg/m3) and tap water varying the oil/water fractions at fixed mixture velocities. Impedance probes are used to identify the continuous phase and the occurrence of phase inversion. The results show a significant increment of pressure gradient at the phase inversion point that occurs at higher water cuts for the smaller diameter pipes. The influence of pipe diameter is observed when two similar Reynolds number situations are compared. The results show that the friction factors and the mixture’s apparent viscosity are considerably higher in the smaller pipe.
General Note: The International Conference on Multiphase Flow (ICMF) first was held in Tsukuba, Japan in 1991 and the second ICMF took place in Kyoto, Japan in 1995. During this conference, it was decided to establish an International Governing Board which oversees the major aspects of the conference and makes decisions about future conference locations. Due to the great importance of the field, it was furthermore decided to hold the conference every three years successively in Asia including Australia, Europe including Africa, Russia and the Near East and America. Hence, ICMF 1998 was held in Lyon, France, ICMF 2001 in New Orleans, USA, ICMF 2004 in Yokohama, Japan, and ICMF 2007 in Leipzig, Germany. ICMF-2010 is devoted to all aspects of Multiphase Flow. Researchers from all over the world gathered in order to introduce their recent advances in the field and thereby promote the exchange of new ideas, results and techniques. The conference is a key event in Multiphase Flow and supports the advancement of science in this very important field. The major research topics relevant for the conference are as follows: Bio-Fluid Dynamics; Boiling; Bubbly Flows; Cavitation; Colloidal and Suspension Dynamics; Collision, Agglomeration and Breakup; Computational Techniques for Multiphase Flows; Droplet Flows; Environmental and Geophysical Flows; Experimental Methods for Multiphase Flows; Fluidized and Circulating Fluidized Beds; Fluid Structure Interactions; Granular Media; Industrial Applications; Instabilities; Interfacial Flows; Micro and Nano-Scale Multiphase Flows; Microgravity in Two-Phase Flow; Multiphase Flows with Heat and Mass Transfer; Non-Newtonian Multiphase Flows; Particle-Laden Flows; Particle, Bubble and Drop Dynamics; Reactive Multiphase Flows
 Record Information
Bibliographic ID: UF00102023
Volume ID: VID00544
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: P360-Plasencia-ICMF2010.pdf

Full Text

Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Influence of the pipe diameter in dispersed oil-water flows


J. Plasencia and O.J. Nydal

Norwegian University of Science and Technology, Department of Energy and Process Engineering
Trondheim, NO 7491, Norway
jose.l.plasencia @ntnu.no and ole.j.nydal @ntnu.no


Keywords: oil-water flow, phase inversion, dispersed flow, multiphase flow




Abstract

Oil-water flows are commonly encountered in oil production wells and in transport pipes. During dispersed flows the apparent
viscosity of the oil-water mixture increases towards the point of phase inversion, where the continuous phase switches from
one to the other. Large scale experiments with real crudes to characterize the flow around the phase inversion point are costly
and not practical. Hence, it is of interest to investigate the possibility of characterizing the oil-water flow in small scale setups.
The present work reports the investigation of the influence of pipe diameter on dispersed oil-water flows in horizontal pipes.
Pressure gradients are measured for oil-water flows in acrylic pipes with diameters of 16, 32 and 60 mm. Experimental results
are presented for the oil Marcol 52 ([t = 11,0 mPa.s p = 835 kg/m3) and tap water varying the oil/water fractions at fixed
mixture velocities. Impedance probes are used to identify the continuous phase and the occurrence of phase inversion. The
results show a significant increment of pressure gradient at the phase inversion point that occurs at higher water cuts for the
smaller diameter pipes. The influence of pipe diameter is observed when two similar Reynolds number situations are
compared. The results show that the friction factors and the mixture's apparent viscosity are considerably higher in the smaller
pipe.


Introduction


Oil-water flows are commonly encountered in oil
production systems. The amount of water being produced
together with the oil normally increases during the life time
of the well. In some cases it is economically possible to
continue the well production up to around 90% of water in
volume (Xu, 2007).
In oil-water flows at low water cuts, dispersed droplets of
water flow in the oil continuous phase. As the water cut
increases the phase inversion point is reached, the dominant
phase switches, resulting in droplets of oil flowing in the
newly formed water continuous phase.
The phase inversion process is detailed by Arirachakaran et
al. (1989). The occurrence of this phenomenon depends on
different factors. The oil and water physical properties such
as density, viscosity and interfacial tension appear to be the
most important. The wettability effect of the two phases on
the pipe wall material are also pointed as relevant, having an
influence on the water cut where phase inversion take place
(Angeli and Hewitt, 1998; Ioannou et al., 2005).
Chemical composition also plays a very important role in
the stabilization of the oil-water mixtures. The presence of
surface active agents (surfactants), which tend to
accumulate at the oil/water interface, lead to a reduction of
the interfacial tension promoting dispersion and
emulsification of the droplets. Fine solids can also act as
mechanical stabilizers (Kokal, 2005).


The occurrence of the phase inversion phenomenon is
accompanied by a sudden increment of the pressure gradient,
and as a consequence, the production efficiency declines.
Hence, special considerations must be taken in the pipelines
production system design to cope with the occurrence of
phase inversion.
In order to accurately predict and prepare the production
system to handle the phase inversion phenomenon, an
appropriate characterization of the crude must be done. The
first fluid samples from a well are in only small amounts. It
is then necessary to find a practical way to characterize the
fluid in small scale experiments.
An interesting characterization method was developed by
Johnsen & Ronningsen (2003). They measured the apparent
viscosity of different types of water in crude oil emulsions
under pressurized and temperature controlled flow
conditions. A flow simulator shaped as a hollow wheel was
used. The emulsion viscosity was obtained by a calibration
of the torque acting on the rotating wheel. This small scale
method proved to be useful.
A limited amount of work have been reported on oil-water
flows in pipe diameters smaller than 25 mm. Mandal et al.
(2007) compared the flow patterns using transparent
polymethyl-methacrylate (PMMA) pipes with diameters of
12 mm and 25 mm for a mixture of water and kerosene. The
objective was to study the influence of pipe diameter on the
oil-water flow pattern in tubes of diameter less than 25 mm.






Paper No


Slug flow was observed for the small diameter due to the
increased effect of the surface tension and equilibrium
contact angle in the narrow pipe.
Wegmann and Von Rohr (2006) investigated the flow
patterns that develop inside glass pipes of 5.6 mm and 7.0
mm. It was observed that transition from stratified flow to
other flow structures occurred for lower velocities in the
smallest pipe. It was proposed that this is due to the
increasing impact of the surface tension forces compared to
the buoyancy forces with decreasing diameter. The smaller
the pipe diameter is, the easier interfacial tension forces can
overcome gravitational forces.
Angeli and Hewitt (1998) measured pressure gradient in
oil-water flows using acrylic and steel pipes with internal
diameters of 24 mm with particular interest on the effect of
the tube material. It was concluded that the tube wall
material can strongly influence the pressure gradient during
two-phase liquid-liquid flows. A drag reduction effect was
also reported.
Ioannou et al. (2005) performed oil-water flow experiments
in acrylic and steel pipes with diameters of 32 and 60 mm. It
was concluded that the initial continuous phase has an
influence on phase inversion, which is not an instantaneous
phenomenon but lasts for some time. Also, the pressure
gradient peak observed during phase inversion was sharper
and larger as the mixture velocity increases.
The experiments reported here are the initial results of an
ongoing project aimed at investigating small scale flow
characterization methods for dispersed oil-water flows.
Pressure gradient measurements are obtained for oil-water
flows using acrylic pipes with internal diameters of 16, 32
and 60 mm and a model oil. Further work will include
experiments with real crudes, and with model fluids which
are modified to resemble real fluids.


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Experimental Facility

The experiments were performed using the experimental
facility at the Multiphase Flow Laboratory, Department of
Energy and Process Engineering, NTNU, presented in
Figure 1. The flow loop consists of a set of pumps which
supply single phase oil and water from the top and the
bottom of the main separator tank respectively. The fluids
are pumped through independent lines to a "Y" junction that
is used as a mixer. At the single phase lines before the mixer,
flow meters and control valves assist the metering and flow
rate control of each phase. Two different sets of flow meters
and control valves were used for each phase depending on
the flow rate range needed in the experiment. Additionally,
pump frequency converters were also used to adjust the
superficial velocities of each phase.
The "Y" junction is used to gently mix the fluids and serves
as the entrance to the main test section. Three different test
sections in the horizontal position were used. Two 16 m
long acrylic pipes with internal diameters of 32 and 60 mm,
and a 4 m long pipe with an internal diameter of 16 mm
which was connected in series at the last meters of the 60
mm pipe when used. The acrylic pipes are mounted on a
support beam and are connected to a small initial separator
at the exit. This first separator, vented to the atmosphere,
helps to absorb any pressure fluctuations during the
experiments and also separates air in the case of liquid-gas
experiments. Finally, the oil-water mixtures are transported
back to the main 3 m3 tank where are separated by gravity
with the help of coalescing plates. These coalescing plates
reduce any remaining vortices and turbulence inside the
tank improving separation and preparing both phases to be
pumped again into the test section.


Nomenclature


e roughness
D diameter
ID internal diameter
IP impedance probes
L pipe length
P pressure
Re Reynolds number
U velocity
WC water cut


Greek letters


f Darcy's friction factor
p density
A gradient
t dynamic viscosity


Subscripts
app apparent
m mixture


Water pump


[ Control Valve r1 Diff. pressure transd.
SFlow meter Co Conductance probe


Figure 1: Multiphase flow facility at NTNU.


Pressure gradient experiments were performed for the oil
Marcol 52 and dyed water. Physical properties of Marcol 52
and dyed water are available in Table 1. Pressure gradient
was measured over 1.22, 0.68 and 0.50 m in the 60, 32 and
16 mm pipe ID respectively. Pressure taps are located at 11
m from the entrance of the 32 and 60 mm pipes and at 3 m
downwards for the 16 mm pipe (at 344D, 183D and 188D
respectively).






Paper No


Table 1: Fluid properties


Properties (i 200C Oil Marcol 52
Density 835 kg/m3
Viscosity 11 mPa.s

Properties (i 200C Dyed Water
Density 998 kg/m3
Viscosity 1.1 mPa.s

Oil-water Interfacial tension @ 200C 35.4 mN/m
Marcol 52 properties as specified by the manufacturer. Water
properties (Chupin, 2003). Oil-water interfacial tension was
measured with the Du Nouy Ring method for oil samples taken
from the separator.

Pressure gradients were measured using a differential
pressure transducer with 0 5.50 kPa range (+0.07 % FS) at
the specified locations. The pressure transducer was
connected to the pressure taps by transparent hoses filled
with air. Entrainment of liquid into the hoses due to pressure
variations was eliminated by injecting some pressurized air
prior to each pressure gradient measurement.
The experiments were conducted varying the water cut from
0 to 100%, when possible, starting from water as the
continuous phase, for different fixed mixture velocities. A
typical experiment was initiated by flowing single phase
water at the desired mixture velocity. Then, the water cut
was gradually decreased keeping the mixture velocity
constant by increasing the oil superficial velocity. In a very
few cases at the higher mixture velocities it was not possible
to reach the single phase oil condition due to a capacity
limitation of the oil pump.
The acrylic pipes were used to allow visual observation of
the flow. Figure 2 shows the typical oil-in-water and
water-in-oil dispersions obtained during the experiments for
the 60 mm pipe. In this picture water appears in green due
to the dye used to help in the identification of each phase. In
Figure 2 (a) it is possible to see oil droplets flowing with a
higher concentration in the upper half of the pipe and (b)
water droplets flowing in the oil continuous phase with a
higher concentration in the lower half of the pipe due to the
gravitational effects. The mixture became more
homogeneous as the mixture velocity increases.












(a) (b)

Figure 2: Typical oil-water dispersions observed during the
flow experiments at low mixture velocities (approx. 1 m/s).
Water appears in green. (a) oil-in-water (b) water-in-oil
(Pipe ID = 60 mm).

The oil-water dispersions obtained with Marcol 52 and
water were not stable. The resultant oil-water mixture


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

normally separates quite fast, in the order of seconds.
Nevertheless, after running the experiments for some time it
was possible to observe small amounts of the opposite phase
in oil or water.


Phase inversion identification

During the experiments at low mixture velocities it was
visually possible to identify the actual continuous phase and
the occurrence of phase inversion. For the higher mixture
velocities this was not clear, and the use of impedance
probes was necessary.
The impedance probes (IP) consist of two ring electrode
pairs with the same internal diameter as the acrylic pipes.
The conductivity between the flush mounted ring probes
can be measured and calibrated to the cross-sectional
holdup for separated and for homogeneous bubble flows
(Andreussi et al., 1988). As the oil phase is non-conducting
and the water is conducting, the ring probes can be used to
detect the point of phase inversion.
Impedance probes were placed in two different positions of
the test section. The first one at 7 m from the pipe entrance
and the second one immediately before the differential
pressure transducer, as shown in Figure 1.
Normalized impedance signals are shown in Figure 3 for a
mixture velocity of 3.5 m/s in the 60 mm pipe. This curve
was obtained by averaging the impedance signals at each
water cut and normalizing them in such a way to get the
highest signal value equal to one and the lowest equal to
zero. An interpretation of the impedance signals reveals that
phase inversion begins at a point where a change in the
slope of the normalized signal appears (40% WC). In Figure
3 the standard deviations of the pressure gradient
measurements are also presented as a function of water cut.
The higher deviations found at 40% water cut are an
indication of the pressure oscillations that are characteristic
of the initiation of the phase inversion process which is
confirmed by the impedance probe signals.


10 020
0.9 018
B +Stand Dev.- 016
0.7 0.14
06 012 ,c
05 010 .I
04 008
03 006 4
02 004
01 002
00 000
0 20 40 60 80 100
WC(%)

Figure 3: Impedance probe normalized signal and
pressure gradient standard deviation as a function of water
cut for a mixture velocity of 3.5 m/s in the 60 mm pipe.


The results obtained from the impedance probes located at
the two different points along the pipe reveal that phase
inversion occurred simultaneously.






Paper No


Results and Discussion

Pressure gradient

Pressure gradient results are presented as a function of water
cut for different constant mixture velocities. As reported by
Ioannou et all (2005), depending on the initial continuous
phase the water cuts where phase inversion takes place may
vary. In the present work results starting from water as the
dominant phase are presented.
Figures 4, 5 and 6 present the pressure gradient results for
the pipe diameters of 16, 32 and 60 mm respectively, for a
range of mixture velocities. The pressure gradient values
correspond to the average over a period of 20 seconds. The
water cut is the ratio of the water superficial velocity to the
total superficial velocity. The oil pump capacity limitation
was at a high mixture velocities and low water cuts for the
16 mm pipe.

60 -
| 1 5 m/s -- 2.0 m/s 2. 5 m/s
5.0

40
3.0
S2.0

1.0


0 20 40 60 80 100
WC(%)


Figure 4: Pressure gradient as a function of water cut for
different mixture velocities in the 16 mm pipe.


" 0 2.5 m/s
9.0m/s


6.0 Z ---- ------------




0.0 -------------------------------------------

0 20 40 60 80 100
WC(%)


Figure 5: Pressure gradient as a function of water cut for
different mixture velocities in the 32 mm pipe.

2.5
-2.0m/s
20 2.5m/5

15 --9-3.5 m/
03. m





10
0 20 40 60 80 100
WC(%)


Figure 6: Pressure gradient as a function of water cut for
different mixture velocities in the 60 mm pipe.


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

As expected the results show that the pressure gradient
increases with the increment of mixture velocity for all
water cuts. Also, higher pressure gradients result as the pipe
diameter is reduced for the same mixture velocity. It is
possible to observe a sudden increment in pressure gradient
at around 40% of water cut in the 60 mm pipe and at 50% in
the 32 mm pipe. This characteristic in the curves is clearly
more evident as the mixture velocity increases.
Two pressure peaks are seen for the 16 mm pipe. The first
one occurs at around 45% water cut and is more significant
as the mixture velocity increases. The second one is larger
and occurs at 30% of water cut.
Drag reduction was also detected in the results at higher
water cuts. As the oil fraction increases it is possible to see a
gradual reduction of pressure gradient until a sudden
increment in pressure gradient occurs. This effect if exists is
not so evident in the oil dominated region. Drag reduction
was also reported by Angeli and Hewitt (1998).



Phase inversion

Figure 7 shows the results for two different pipe diameters
(32 and 60 mm) for the mixture velocity of 3.5 m/s.
Pressure gradient and the normalized impedance probes
signal are plotted as a function of water cut. Following the
previous interpretation of the IP normalized signal it is
possible to say that the inversion process in the 32 mm pipe
starts between 40 and 50% WC accompanied by an
increment in the pressure gradient. Similar behaviour is
observed in the 60 mm pipe where inversion starts between
35 and 40% WC. This is an indication that phase inversion
occurs at higher water fractions in smaller pipes.

80 1
7 32mm -*- m 09
7,0




3.0 0.4








probe signal in the 32 and 60 mm pipes at 3.5 m/s.


Impedance probes were not available for the 16 mm pipe.
00.3
















However, pressure oscillations, quantified by the standard
a40----- 0y----5 s















deviation of the measurements, were used to identify the
0.1
00 0
0 20 40 60 80 100
WC(%)









occurrence of phase inversion. Figure 8 shows the standard
deviations of the pressure gradient and nomeasurementslized impedan the 16
mm pipe previously presented in Figure 4. It can be
Impedance probes were not available for the 16 mm pipe.
However, pressure oscillations, quantified by the standard
deviation of the measurements, were used to identify the
occurrence of phase inversion. Figure 8 shows the standard
deviations of the pressure gradient measurements in the 16
mm pipe previously presented in Figure 4. It can be
observed that the maximum deviations, more evident for the
higher velocities, occur at around 45% of water cut
matching with the pressure peaks seen in the pressure
gradient curves of Figure 4. The same type of data can be
observed in Figure 9, this time for the 32 mm pipe. The
pressure fluctuations show a maximum at the phase
inversion point predicted by the impedance probes (Figure
7). The pressure oscillations may suggests that the phase






Paper No


inversion in the smaller pipe diameter occurred at around
45% of water cut and the pressure increment observed at
30% might be due to other reasons.
Results presented in Figure 9 also suggest that the initiation
of phase inversion is accompanied by pressure gradient
instabilities that are stronger as the mixture velocity
increases. The standard deviation of the measurements in
the 60 mm pipe presented the same type of behaviour. This
can be observed for the case of the mixture velocity of 3.5
m/s in Figure 3.

0.30
0.25

0 20


A er
0.15
M 0.10

0.050
0.00 -----------------I ----------
0 20 40 60 80 100
WC(%)

Figure 8: Standard deviation of the differential pressure
measurements as a function of water cut for different
mixture velocities in the 16 mm pipe.

0.90
0.80 -0-40m/s
0o70 -35 m/s
060 -- 3 0 m/s
0.50 2.5m/s
0.40
030
0 20
o.10
0.00
0 20 40 60 80 100
WC(%)

Figure 9: Standard deviation of the differential pressure
measurements as a function of water cut for different
mixture velocities in the 32 mm pipe.



Flow condition

The pressure gradient increment observed at 30% WC in the
16 mm pipe (Figure 4) might be associated with the change
of flow condition from turbulent to transitional or laminar
flow due to the simultaneous variation of mixture density
and viscosity with water cut. A simple calculation can be
carried out for single phase oil flowing in the 16 mm pipe
for the three different velocities.

U= 1.5 m/s, Re =1822
Um= 2.0 m/s, Re =2429
Un= 2.5m/s, Re=3036

Where Um is the mixture velocity and Re is the Reynolds
number. From the calculations it can be concluded that for
single phase oil the flow condition is laminar at 1.5 m/s and
falls in the transition region at 2.0 m/s. In addition, for low
water cuts the flow condition might be laminar or


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

transitional for the mixture velocity of 1,5 m/s and will fall
inside the transition region for the case of 2,0 m/s. Thus, the
pressure increase seen at water cuts less that 30% for 1.5
m/s might be due to the flow undergoing transition from
turbulent (Re 60000, 29% WC) to laminar (Re=1822, 0%
WC). This was not the case of the bigger pipe diameters (32
and 60 mm) where all the experiments were carried out in
turbulent conditions.



Friction Factor

Figures 10, 11 and 12 show the friction factor results for the
three pipe diameters. Friction factor coefficients were
calculated from pressure gradient data using Equation 1.


2.AP.D
f=


In this equation f is the Darcy's friction factor, AP is the
pressure gradient, D is the internal diameter of the pipe, p,
is the density of the mixture, U, is the mixture velocity and
L the pipe length.
It can be observed from the figures that the friction factor
values close to the single phase points are consistent. On the
other hand, these values are quite dispersed around the
region where phase inversion takes place. In the 16 mm
pipe (Figure 10) the friction factors are only slightly higher
at the water cut determined earlier as the phase inversion
point (around 45%). At 20% WC the friction factors
increases abruptly to around 0.04 and a subsequent fall to
around 0.03 for single phase oil (mixture velocity equal to
1.5 m/s).
Looking at Moody's chart (1944), friction factors greater
than 0.04 fall into the transition flow region for Reynolds
numbers between around 2300 and 4000 and low pipe wall
roughness. This might be an indication that the friction
factor values at 20% WC may correspond to a transition
flow region. On the other hand, Reynolds numbers below
2300 which is the case of the single phase oil flowing at 1.5
m/s in the 16 mm pipe (Re = 1822), falls into the laminar
region and the friction factor results to be around 0.03 being
in agreement with the experimental results. These two facts
can indicate that the pressure gradient increment found at
30% water cut (Figure 4) and the decrement for single phase
oil occurred due to a switch in flow condition from
turbulent to transitional and laminar.

0.060
t1.5 m/s
00 00 2.m/s

0.040 A25m/

0 0030
S0020 ** P .U AA

0010
0000
0 20 40 60 80 100
WC(%)

Figure 10: Friction factor as a function of water cut for
different mixture velocities in the 16 mm pipe.






Paper No


0.050
0.045
0-040
0.035
S0.030
c0.025
t 0.020
0.015
0.010
0-005
0.000


0 20 40 60 80 100
WC(%)


Figure 11: Friction factor as a function of water cut for
different mixture velocities in the 32 mm pipe.

0.040
0.035
S2.0m/s
0.03 2.
0m030 . A.- 3 0 m/s
S0.025 -- .35m/5
0.020
o.o15 2


0.005
0.000
S20 40 60 BO 100
WC(%)


Figure 12: Friction factor as a function of water cut for
different mixture velocities in the 60 mm pipe.

With the data acquired it is possible to compare two friction
factor curves, obtained from two different pipe diameters
and mixture velocities, where the Reynolds numbers are
similar for the two pipe diameters both at 0% and at 100%
WC. This is the case of the mixture velocity of 4.0 m/s in
the 32 mm and 2.0 m/s in the 60 mm diameter pipes. In
these two situations the single phase Reynolds numbers
result to be 9716 & 116131 for 32 mm and 9109 & 108873
for 60 mm pipes at 0 and 100% water cut respectively (all in
turbulent conditions). Starting with the assumption that the
wall roughness of the acrylic pipe can be neglected (case of
smooth pipes, which is not far from the reality for acrylic
pipes) and assuming that the flow is fully dispersed so that
the homogeneous model is a good approximation, then the
friction factors should only be a function of the Reynolds
number.

0-040 -
/ ID 60mm- 2 0m/s
o.oo ----ID-32mm -40 -

I 0025
S0.020


Co o
G.005 --
0o-0
coMoo,,,


1 25 /s
S30m/s
* = ._= "= 0. A43 m,/s
-w'A35 *-



A.


Colebrook equation,

1 e/D+
I-- -2.log 3.+

f 12 3.7


2.51 2

Re. f2


For turbulent flow in smooth pipes,

1- 2log(Re. f2)- 0.8


Apparent viscosity,

Pm .Um .D
S Pm Re
Papp Re


Where e is the pipe wall roughness and pum is the apparent
viscosity.

20 -
S8 -ID 60mm -20 m/s
1 -W-0ID 32mm 40 ms
E14
12
S10 -

a 1



0 20 40 60 80 100
WC(%)


Figure 14: Apparent viscosity as a function of water cut for
a mixture velocity of 2.0 and 4.0 m/s in pipe IDs of 60 and
32 mm respectively (similar Re at 0 and 100% WC).

Following the friction factor behaviour observed in Figure
13 the higher discrepancies in the viscosity are found at the
phase inversion region and are evidently higher in the


0 20 40 6o so 100 smaller pipe. Since the fluid properties are the same in both
WC(%) cases, the only variable that might be influencing the
apparent viscosity seems to be the droplet size which is a
Figure 13: Friction factor as a function of water cut for a result of the shear stresses induced by the flow. Possible
mixture velocity of 2.0 and 4.0 m/s in pipe IDs of 60 and 32 separation of the phases into a more stratified flow
mm respectively (similar Re at 0 and 100% WC). configuration also gives additional uncertainties in the
discussion.


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Figure 13 presents the friction factors comparison for the
two cases considered as a function of water cut. It is
observed that the friction factor values are the same in both
single phase situations but deviate in the region where phase
inversion take place.



Apparent viscosity of the dispersion

Mixture's viscosity or apparent viscosity has been
calculated for the friction factors presented in Figure 13.
The results were calculated using Equations 2 and 4 which
take into account the roughness of the pipe wall.
Nonetheless, the results were similar to those predicted by
Equation 3, used for smooth pipes. It confirms that the
smooth pipe assumption adopted for the analysis in the last
section for the acrylic pipes was not far from the results
obtained considering its roughness.






Paper No


Figure 15 presents the apparent viscosity results calculated
for the 32 mm pipe from the pressure gradient data shown in
Figure 5. Similar to that in Figure 14, the results show a
discrepancy that is higher around the phase inversion region.
As shown in Figure 16, the experiments presented here for
the 32 mm pipe were performed in turbulent flow conditions.
In this figure the vertical axis (Re) is presented in
logarithmic scale to improve the presentation.


-$-2.5 m/s
16 ---Ia3.0 -m/s
S14 3.5 m/s
12 4 .0 m/s





0
0 20 40 60 80 100
WC(%|

Figure 15: Apparent viscosity as a function of water cut for
different mixture velocities in the 32 mm pipe ID.

1,000.000


100000oo

-2 5m/s
10,000
-3 0 nVs
3 5 m/s
S--4.0 m/s
1.000 ,-
0 20 40 60 8O 100
WC(%)

Figure 16: Reynolds number as a function of water cut for
different mixture velocities in the 32 mm pipe ID.


Conclusions

Pressure gradient measurements were conducted in pipes
with 16, 32 and 60 mm of internal diameter. The results
show that phase inversion occurs at higher water cuts in the
smaller diameter pipes. The phase inversion phenomenon is
accompanied by pressure gradient fluctuations that are
stronger as the mixture velocity increases.
A comparison between two conditions where similar
Reynolds numbers are found for the two different pipe
diameters at 0 and 100% WC respectively shows that the
resultant friction factors and apparent viscosities are
considerably higher in the smaller pipe in particular in the
region where phase inversion takes place. This might be a
result of the shear stresses induced by the flow giving
different droplet size distribution leading to different
apparent viscosity and possibly to different degrees of
separation.
For the smallest pipe the pressure gradient curves showed
two peaks with water cut. It was concluded that in addition
to the phase inversion phenomenon, a turbulent-laminar
transition may influence pressure gradient. Hence, initial
turbulent flow conditions may become laminar due to the


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

density and viscosity variations as a function of water cut.
The discussion in this paper is based on an assumption of
dispersed flow regimes. However, as the flow rates are
reduced a gradual transition to separated flow will occur
with Marcol-water flows.
This work is part of a study on phase inversion. Marcol 52
was chosen as a model fluid which will be modified to
allow for laminar flow studies with stable oil-water
dispersions.

Acknowledgements

The authors acknowledge the Flow Assurance Innovation
Centre (FACE) for the financial support and the permission
to publish this work.

References

Andreussi, P., Di Donfrancesco, A. & Messia, M. An
impedance method for the measurement of liquid hold-up in
two-phase flow. International Journal of Multiphase Flow
Vol. 14, No. 6, pp. 777-785 (1988).

Angeli, P & Hewitt, GF. Pressure gradient in horizontal
liquid-liquid Flows. International Journal of Multiphase
Flow 24, 1183-1203 (1998).

Arirachakaran, S., Oglesby, K.D. & Brill, J.P An analysis of
oil/water flow phenomena in horizontal pipes. SPE 18836
(1989).

Chupin, G. An experimental Investigation of multiphase
gas-liquid pipe flow at low liquid loading. PhD Thesis.
Norwegian University of Science and Technology. Faculty
of Engineering Science and Technology (2003).

Ioannou, K., Nydal, O. & Angeli, P Phase inversion in
dispersed liquid-liquid flows. Experimental Thermal and
Fluid Science 29, 331-339 (2005).

Johnsen, E.E. & Ronningsen, H.P Viscosity of "live"
water-in-crude-oil emulsions: experimental work and
validation of correlations. Journal of Petroleum Science and
Engineering 38, 23-36 (2003).

Kokal, S. Crude-oil emulsions: A state-of-the-art review.
SPE 77497-PA. Production and Facilities (2005).

Mandal, T K. Chakrabarti, D. P & Das G. Oil water flow
through different diameter pipes: similarities and differences.
Trans IChemE, Part A, Chemical Engineering Research and
Design, 85(A8): 1123-1128 (2007).

Moody, L.F. Friction factors for pipe flow. Semi-Annual
Meeting, Pittsburgh, Pa., June 19-22, ASME (1944).

Wegmann, A. & Von Rohr, PR. Two phase liquid-liquid
flows in pipes of small diameters. International Journal of
Multiphase Flow 32, 1017-1028 (2006).

Xu, X. Study on oil-water two-phase flow in horizontal
pipelines. Journal of Petroleum Science and Engineering 59,
43-58 (2007).




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs