Journal of Petroleum Science and Engineering 127 (2015) 409420

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering

journal homepage: www.elsevier.com/locate/petrol

Experimental investigation of oilwater two-phase ow in horizontal

pipes: Pressure losses, liquid holdup and ow patterns
Ahmad Shamsul Izwan Ismail a, Issham Ismail a,n, Mansoor Zoveidavianpoor b,
Rahmat Mohsin a, Ali Piroozian a, Mohd Shahir Misnan a, Mior Zaiga Sariman c
a

Malaysia Petroleum Resources Corporation (UTM-MPRC) Institute for Oil and Gas, Universiti Teknologi Malaysia, Johor 81310, Malaysia
Department of Petroleum Engineering, Faculty of Petroleum and Renewable Energy Engineering, Universiti Teknologi Malaysia, Johor 81310, Malaysia
c
Department of Process Engineering, Development Division, PETRONAS Carigali Sdn. Bhd., Kuala Lumpur 50088, Malaysia
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 8 April 2014
Accepted 31 January 2015
Available online 7 February 2015

Flow experiments have been conducted for oilwater two-phase ow in a horizontal 5.08 cm ID ow
loop at a length to diameter ratio of 1311. The uids were light Malaysian waxy crude oil from the
offshore Terengganu (o 818 kg/m3 , mo 1.75 mPa s and wax content 16.15 wt%) and synthetic
formation water. The water-cut was varied between 10 to 90% at nine mixture ow rates of 2.0 to
16.2 cm3/s. Measuring the changes in pressure drop and liquid holdup at different ow rates of oilwater
two-phase ow, a new ow pattern was identied. Strong dependence of the oilwater slippage on the
minimum ow rate was observed. The highest pressure drop of 11.58 kPa was obtained at maximum
ow rate of 16.21 cm3/s and oil fraction of 0.9; while the lowest pressure drop of 1.31 kPa was recorded
at the lowest ow rate of 2.03 cm3/s and water fraction of 0.9. The experimental results could be used as
a platform to understand better a more complex case of gas/oil/water concurrent ow in a pipeline.
& 2015 Elsevier B.V. All rights reserved.

Keywords:
Waxy crude
Oil/water ow
Pressure drop
Holdup
Flow pattern

1. Introduction
The need for reliable experimental studies on many engineering
applications of ow assurance has been the driving force behind
extensive research efforts in the area of multiphase ow. Liquid
liquid ow could be dened as the simultaneous ow of two
immiscible liquids in a pipe. Previously, multiphase ow research
works were mainly focused on gasliquid ow; among the earliest
studies in the gasliquid eld were Beggs and Brill (1973), Wicks and
Dukler (1960), Hagedorn and Brown (1964), Gregory and Aziz (1975)
and Cornish (1976). Nevertheless, the industry attention has shifted
towards the understanding of the simultaneous ow of gasoil
water mixtures (Trallero et al., 1997). Despite the extensive studies on
gasliquid two phase ow, liquidliquid ow has received inadequate research attention (Atmaca et al., 2009). In the oil and gas
industry, simultaneous transport of water and oil in pipelines occurs
frequently. For oil elds operating at high water-cuts and low
wellhead pressures, the effect of the water phase with respect to
pressure drop is of particular importance. Lack of knowledge of the
ow patterns, pressure drop and in-situ distributions of the liquids
could be hampered the safe and economic transport of these uids.

Corresponding author. Tel.: 60 7 553 5506; fax: 60 7 5581463.

E-mail address: issham@petroleum.utm.my (I. Ismail).

http://dx.doi.org/10.1016/j.petrol.2015.01.038
0920-4105/& 2015 Elsevier B.V. All rights reserved.

The gained knowledge via experimental analysis can contribute to

accurate modelling and prediction of oilwater ow in pipes.
Due to the dwindling of conventional light crude oil or easy oil
reserves and the existence of lots of mature oilelds around the
globe, especially in the Malaysian oilelds, the phenomenon of
concurrent ow of oil and water in pipelines has been the main
subject of research studies in petroleum production and enhanced
oil recovery with water injection. Furthermore, there are many
cases where high water cut is present but the wells are still
considered economically viable to operate. Understanding the
behaviour of oilwater ow in pipelines, such as ow pattern,
pressure drop, and liquid holdup is crucial for many engineering
applications such as design and monitoring of the separation
process, interpretation of production logs, and operation of ow
lines and wells (Atmaca et al., 2009).
Some of the oilelds around the world are producing waxy
crude oil. This phenomenon is due to the presence of parafn (C18
C36) and/or naphthenic (C30C60) hydrocarbons in the crude oil
(Mansoori, 1993). When a crude oil contains waxes, the properties
of the oil, especially the viscosity, will greatly change. There were
numerous two phase ow experimental studies on the signicance
of viscosity, such as Russell et al. (1959), Arirachakaran et al. (1989),
Oglesby (1979), Trallero (1995), Alkaya (2000) and Mckibben et al.
(2000). All of these prominent researchers have found that the
viscosity was greatly affecting the ow pattern, pressure drop, and
liquid holdup.

410

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

Aside, dealing with an oilwater mixture in a pipeline leads to

unique and complex problems in the oil and gas industry due to its
complicated rheological behaviour, and vast difference in pressure
gradient encountered for different ow patterns (Arirachakaran
et al., 1989). Although two phase ow of oil and water is normally
occurred in pipes during production or transportation of petroleum
uids, its hydrodynamics behaviour under a wide range of ow
conditions and inclination angles still creates a relevant unresolved
issue for the oil industry (Flores et al., 1999). Actually, the main
reported laboratory works on liquidliquid two-phase ow were
accomplished using gas oil, mineral oil, or rened oil, and limited
experimental studies were performed on waxy crude oil. Therefore,
an unexplored territory arises in terms of the ow behaviour in a
pipeline when a waxy crude oil is introduced in a two-phase ow
system. Accordingly, this crude that contains waxes would affect the
ow behaviour due to its viscosity changes, complex interfacial
chemistry, and natural emulsion effect.
Flow pattern is a particular type of geometric distribution of the
components in a pipe and many of the names given to these ow
patterns are now quite standard (Brennen, 2005). As emphasized by
Trallero et al. (1997), the subject of oilwater ow cant be addressed
in a unied way. Thats because of the diversity of oil properties (e.g.,
viscosity, density, rheological behaviour, etc.), which makes their
investigation not only too broad and contentious but also important
and worthwhile.
Generally, we need an accurate prediction of waxy crude oil
multiphase ow behaviour to produce and transport the waxy crude
oil safely and economically. Waxy crude oils have complex ow
properties; although considerable research has gone into the solution
of specic industrial pipelining problems, a study devoted to the
understanding of the behaviour of this material has not been
appeared in the literature. Thus, an experimental investigation has
been conducted to study the ow behaviour (i.e., ow pattern,
pressure drop, and water holdup) of Malaysian waxy crude oilwater
ow in horizontal pipes. This study addresses the determination of
oilwater ow pattern for Newtonian and low-viscosity Malaysian
waxy crude oil above wax appearance temperature (WAT).

2. Literature review
Crude oil from reservoirs is pushed to the surface by the high
underground pressure (natural drive) and is owed through their
respective wellheads and pipelines for further processing. Normally,
crude oil pipelines contain a fraction of water due to water
encroachment from an aquifer and among others; and water
percentage tends to increase in pipes over time. The situation is
worsened when the wells are still operating even though the
production stream is producing at high water-cut (Ngan, 2010). In
liquidliquid ow studies, the necessity to understand the nature
and ow behaviour of this type of multiphase ow is crucial due to
the existence of different mechanisms governing them and various
ow patterns conguration. Russell and Charles (1959), Russell et al.
(1959) and Charles et al. (1961) were among the earliest researchers
who conducted studies on liquidliquid ows. Most of their results
became a reference for the subsequence studies and also provided a

Oil

Water
Fig. 1. Examples of oilwater ow in a horizontal pipe.

basic knowledge in understanding better the behaviour of a liquid

liquid ow. This scenario has attracted numerous extensive research
works on this area after a decade, such as Guzhov and Medredev
(1971), Guzhov et al. (1973) and Hughmark (1971). Brauner (2002)
found that a liquidliquid system is characterized by a low density
difference between phases and this nding was supported by
Atmaca et al. (2009). They explained that the oilwater system
usually has similar densities, a large difference in viscosities, and
more complex interfacial chemistry compared to gasliquid systems (Fig. 1). However, a small density difference in terms of oil
properties (e.g., API 45 to API 10) implies tremendous differences in
composition, viscosity, etc.
Nadler and Mewes (1997) explained that the ow behaviour of oil
and water in pipes is heavily relied on the droplet distribution of the
dispersed phase and volume fraction of the phases. This dependency is
due to the effect of nite density difference between the oil and water
phases that is contrary to gasliquid ow system that possesses a
great density difference. A simultaneous ow of oil and water will
create an oilwater emulsion since they mix together when owing in
pipes. This phenomenon completely changes the physical properties of
the liquids. An emulsion which is formed in a dispersed system
consists of two immiscible liquids. An unstable emulsion formed
during a dispersed ow could be separated into its original phases
when it was left in stationary at a reasonable amount of time
(Arirachakaran et al., 1989). Besides, these emulsions may appear to
be a non-Newtonian or Newtonian rheological behaviour (Brauner,
2002). The differences in characteristics are triggered mainly by the
small buoyancy effect, lower free energy at interface, smaller dispersed
phase droplet size, and high momentum transfer capacity in liquid
liquid ows (Vielma et al., 2007).
An accurate prediction of oilwater ow behaviours, such as
pressure drop, ow pattern, and water holdup are imperative in
many advanced engineering applications (Brauner, 2002); such as
designing and monitoring downhole metering, water-lubricated
pipelines, production optimization, articial lift design and modelling, optimum string selection, and production-logging interpretation
(Flores et al., 1999). Russell and Charles (1959) extensively studied
the ow behaviour of oilwater system by considering the ow
pattern, pressure drop, and liquid holdups. Russell et al. (1959) have
successfully observed the ow characteristics of oilwater in a
horizontal condition using a 2.05 cm ID pipe. They found three types
of ow pattern, namely bubble, stratied, and mixed ows. They also
observed that water holdup was greatly inuenced by liquid input
ratio and viscosity. In principle, co-current ows of liquidliquid
mixtures in pipes are stable by considering the ow parameters (i.e.,
supercial velocity of each phase, the mixture owrate, the pipe
diameter, the surface tension, the nite density difference, pipe
wettability, and the ratio of viscosity of uids, as well as the shear
stress between the liquid phases). In spite of the parameters
mentioned above, pipe plane inclinations also affect the ow pattern.
This ow pattern includes horizontal and vertical ow conditions
which has signicant differences in terms of ow pattern identications (Oddie et al., 2003). Charles et al. (1961) conducted a study on
oilwater ow in horizontal pipelines, and they encountered four
types of ow pattern namely; water droplets in oil, concentric water
with oil owing in the core, oil slugs in water and oil bubble in water.
They explained that the viscosity and low density difference between
oil and water were affecting the ow patterns signicantly. Generally,
in an experimental study, there are many possible ow patterns that
can be observed in horizontal conditions apart from those mentioned
by Charles et al. (1961) as found by other researchers like Brauner
(2002) and Trallero et al. (1997).
Researchers like Vuong et al. (2009), Vielma et al. (2007) and
Trallero et al. (1997) found that the pressure drop was strongly
depended on the ow patterns and ow rates. On the other hand,
Atmaca et al. (2009) and Sridhar et al. (2011) stated that pressure

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

gradient was mainly inuenced by the oil viscosity as well as the

effect of ow patterns transition. Nonetheless, Vielma et al. (2007) in
their experimental study of liquidliquid ow behaviour horizontally
believed that the dominant effect for pressure losses was the ow
regime which associated with the mixing uid distribution in the
cross sectional area of the pipe. They found that in a turbulent ow,
the mixing uids shifted from separated phases to dispersed ow,
which has resulted in a greater pressure drop. Vielma et al. (2007)
also stated that there was no signicant effect of pressure drop
during high water supercial velocity at any oil supercial velocity
because of the droplets that present at all owing conditions.
Rodriguez and Oliemans (2005), Poesio et al. (2008), Grassi et al.
(2008), Angeli and Hewit (1998) and Stapelberg and Mewes (1994)
have studied thoroughly pressure drop across horizontal, vertical,
and inclined pipes under various operating conditions using oil
samples of low to high viscosity. The change of pressure gradient
was greatly related to phase inversion or ow pattern transitions
(Arirachakaran et al., 1989). This phenomenon was supported by
many researchers in their studies on oilwater ow in pipes, such as
Trallero et al. (1997), Flores et al. (1999), Poesio et al. (2008), Vuong
et al. (2009), Wang and Gong (2010) and Ngan (2010). Trallero et al.
(1997) highlighted that an insignicant pressure drop was obtained
after the transition period from a stratied region to a dispersed
region. This was attributable to the loss of oil continuity where the
water has fully wetted the pipe wall. Consequently, the drag of the
ow, which is caused by shear stress, is reduced and eventually
reduced the pressure loss. Generally, when high viscous oil ows
together with water, the mixture of high oil viscosity tends to form a
thin oil lm on the pipe wall, thus it increases the shear stress and
substantially elevates the pressure drop. Other elements that contribute to pressure drop are the pipe wettability and roughness. A
stainless steel pipe can be water wetted, while owing oilwater
together in pipes will cause lower pressure drop. Soleimani et al.
(2000) in their study on spatial distribution of oil and water in a
horizontal pipe ow have explained succinctly the pipe wettabilitys
effect on pressure drop.
Water holdup (Hw) is known as in-situ volume fraction of water
over a total mixing liquids in a specic length of a test section and it
is measured using quick closing valves. A value for liquid holdup or
specically water holdup cannot be obtained analytically; it has to be
determined using an empirical technique. This is a function of
variables such as liquid properties, ow pattern, pipe diameter, and
pipe inclination. Via an experimental study, Vielma et al. (2007)
revealed that the Hw was found to have increased with the increase
of the supercial water velocity, vsw, and decrease of supercial oil
velocity, vso. However, Soleimani et al. (2000) stated that in a twophase ow system, the arithmetic mean velocity of a mixture for
each phase does not fully represent the total supercial velocity.
Thus, the in-situ fractions of oil and water were differing from the
corresponding input fractions. They found that the holdup was
altered from the input volume fraction at a minimum ow rate
and increased slip ratio. This trend, according to Charles et al. (1961)
and Martinez et al. (1988), was due to the liquid that was in contact
with the pipe. It was likely to accumulate in the tubes at a slower
velocity. This also occurred in high viscous oilwater ow in a
horizontal system (Zhang et al., 2010). They discovered that the oil
ows at the top of the pipe wall at lower velocity values. The total
liquid holdup can be calculated using equation of, HL Hw Ho.
There are many researchers such as Trallero et al. (1997), Zhang
and Sarica (2006), McKibben et al. (2000), Rodriguez and Oliemans
(2005) and Keskin et al. (2007) who have conducted experimental
and modelling studies focused on low-medium oil viscosity-water
ow systems of the multiphase ow. However, the study of the high
viscous oilwater ow was limited in numbers. Even though other
researchers such as Vuong et al. (2009) and Sridhar et al. (2011)
came out with a comprehensive and practical classication of ow

411

patterns based on high viscous liquidliquid relationship, they did

not encompass the effect of using actual waxy crude oil in their
respective research studies. In spite of being a permanent interest
for the petroleum industry, the issue of waxy crude oilwater ow
system in pipes has barely been addressed in the technical literatures. This issue is worsened when coupled with the oilwater
mixture ow behaviour, which presents a complex problem for
pipeline transportation and production of waxy crude oil. The
reason for this complexity is due to the inuence of crude oil
compositions that makes a vast difference in the pressure gradient
encountered for different ow patterns (Wang and Gong, 2010).
In early experimental research, Oglesby (1979) reported there were
14 ow patterns, while others described only three to four (Russell
et al., 1959; Malinowsky, 1975). Since the 1990s, with the advanced
instruments and techniques, different ow pattern parameters have
been measured more accurately, and ow patterns of oilwater ow
have been analyzed objectively (Trallero, 1995; Trallero et al., 1997;
Nadler and Mewes, 1997; Angeli and Hewit, 1998; Shi et al., 1999; Shi,
2001). Wang and Gong (2010) have thoroughly studied the differences
and similarities of ow patterns and transition characters between
crude oilwater and high viscosity oilwater owed in horizontal
pipes. They found that the mineral oilwater ow has been widely
used in the indoor experiments to study the two phase ow
characteristics. Even though its density and viscosity were approximately similar to the actual crude oil, they observed that the ow
pattern and transition characters between mineral oilwater ow and
actual crude oilwater ow were behaving differently. The similarities
in terms of ow pattern between both pairs could be classied as
water-dominated, oil-dominated, intermittent, and stratied regions.
However, the differences were largely contributed by the transition
character between crude oilwater ow and mineral oilwater ow.
There was a delay in transition and switch of transition region in crude
oilwater ow, both of these phenomena were greatly inuenced by
the presence of surfactants or some molecules in the crude oil, such as
asphaltenes, resin, etc. (Gafanova and Yarranton, 2001).
The design of subsea pipelines carrying waxy crude oil, especially in Offshore Malaysia, is largely controlled by ow patterns and
pressure drop requirements. The correct measurement of waxy
crude uid behaviour in pipelines is accordingly important for
proper design of pumping capacity and pipeline dimensions. In this
paper, the behaviour of waxy offshore Terengganu crude oil was
studied above its WAT in terms of pressure drop, water holdup and
ow patterns for nine ow rates in a 5.08 cm ID horizontal ow
loop at a length to diameter ratio of 1311.

3. Experiments
3.1. Experimental set-up
The experimental set-up where all the tests and measurements
conducted for the Malaysian waxy crude oilwater ow in pipes
was located at the Malaysia Petroleum Resources Corporation
Institute for Oil and Gas (UTM-MPRC for Oil and Gas), Universiti
Teknologi Malaysia (UTM). Fig. 2 shows the complete experimental set-up.
The experimental investigations began with characterization of the
test uids as shown in Table 1. The test uids used throughout the
experimental works were waxy crude oil from offshore Terengganu of
Malaysia and synthetic formation water. The crude oil used for the
research works was considered as mild waxy crude oil since it has low
WAT and pour point of 26 1C and 18 1C, respectively. The pour point
was measured by ASTM D97/93 standard test method. WAT and wax
content (%) were determined using the Differential Scanning Calorimetry (DSC) (Visintin et al., 2005). The DSC measurements of original
wax were performed with Perkin- Elmer Pyris 7 DSC thermal analyser

412

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

Fig. 2. Schematic diagram of the oilwater test facility.

3.2. Flow pattern observation

Table 1
General properties of the test uids.
Properties

Crude oil

Water

API gravity (1API)

Density (kg/m3)
Dynamic viscosity @ 30 1C (mPa s)
Flash point (1C)
Pour point (1C)
WAT (1C)
Wax content (wt%)
Asphaltenes (wt%)

41.4
818
1.75
o19
18
26
16.15
0.06

10
1000
1.0

system, which was calibrated with an indium standard before use. The
crude components were provided from Gas ChromatographyMass
Spectrometry (GCMS) analysis (Chen et al., 2006).
The facility comprised oil and water tanks of capacity of 209 m3
each. The ow of oil and water from their respective tanks to the
test section were driven by two three-phase induction motors of
1.49 kW that could generate 2900 rpm with maximum owrate of
16.2 cm3/s. Meanwhile, a positive displacement mixture pump
was installed in the closed-loop system to circulate the mixture
uids in the pipeline. The test loop, as depicted in Fig. 3, comprised
two 33.5 m long straight stainless steel pipes, connected by a 1.5 m
long U-shape stainless steel pipe to reduce disturbances on the
ow pattern due to the sharp turn.
The pipe has a 5.08 cm internal diameter (ID) and has been
instrumented to permit continuous monitoring of temperature
and pressure drop. Thermocouples were installed along the test
section to continuously monitor the thermal condition inside the
pipeline. The outer pipeline up to the settlement tank of test
section was wrapped with bre cloth to mitigate heat loss. The
closed-loop was also connected to the Y mixing section at the
entrance, as depicted in Fig. 4.

In the experiments of oil/water ow patterns, investigators have

almost exclusively used visual observation and photography related
techniques (Russell et al., 1959; Charles and Lilleleht, 1966;
Arirachakaran et al., 1989). The most common way to identify the
different ow patterns is to observe the ow in a transparent
channel or through a transparent window on the pipe wall. As an
extension to visual observation, photographic or video techniques
have also been widely used. For very rapid phenomena, high-speed
photography or video is necessary. In this experiment, ow patterns
were identied by observing the ow behaviour through the
observation window (i.e., the transparent acrylic pipe) using The
Sony cyber-shot DSC-H9 high denition video camera. The camera
has the capability of producing 100 shots at 2.2 frames per seconds
at 1920  1080 pixel. The camera was used together with the video
at HDTV (1080i) video output to capture the ow of the waxy crude
oil and water through the 3.4 m section of observation window.
Since waxy crude oil has a tendency of opaqueness, a uorescein
powder was introduced into the water phase to enhance the
observation of ow behaviour between phases. The mixture of
uorescein powder with water produced a luminous colour (i.e.,
for free water identication) and it was transmitted with ultraviolet
(UV) light. In this experiment, 0.1 g of uorescein powder per
1500 ml of water was used and observations were made by UV light.

3.3. Pressure drop measurement

Pressure drop was measured differentially using two pressure
transducers of type Wika Model A-10. The model A-10 includes the
following features; output signal 420 mA, power supply1430 V;
pressure ranges 0103.4 kPa and accuracy r 71% of span. Two
pressure transmitters were located over a length of 3 m after the
observation window and 12 m from the pump. The National

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

413

Fig. 3. Schematic diagram of multiphase ow loop test section.

Fig. 4. Y mixing section

Instrument data acquisition system equipped with Labview Software

was used to record the pressure drops.
3.4. Water holdup measurement
Holdups were measured using a 5000 ml graduated beaker.
The test uids owing in the test section were trapped in the
removable spool using two quick closing valves (also called
solenoid valves). The quick closing valves were specially designed
to close and open instantly using an electronic switch. The trapped
liquids were drained into a container and were left to settle for a
day prior to taking the measurements.
3.5. Experimental procedures
A total of 72 runs were accomplished in the experimental
works of waxy crude oilwater ow using a closed-loop pipeline
at horizontal condition. The oil percentages were varied from 10 to

90%, while ow rates were ranged from 2.03 to 16.21 m3/s. The
experimental work was carried out as follows:
(1) Oil and water were stored in their 209 m3 respective storage
tanks. Water mixed with uorescein powder was pumped into
the system until it completely lled up the closed-loop section.
(2) Once the closed-loop section was fully occupied by water, oil
was then injected gradually into the system according to the
stipulated percentage (i.e., 10 to 90% of oil) and the displaced
water was then collected at the discharge line of the test
section. The amount of oil injected must be equivalent with
the amount of water collected at the discharge point.
(3) Both the oil and water were kept on circulated in the closedloop system at the predetermined ow rate (i.e., 2.03 m3/s) for
15 min until a stabilized ow was achieved.
(4) The ow patterns were recorded via the observation window
using a high denition video camera and simultaneously the
pressure drops were recorded using pressure transducers
during the 15 min circulations. Lastly, the mixing pump was

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

switched off after the solenoid valves were closed to trap the
mixture uids in the removing spool for water holdup
measurement.
(5) Steps (1) to (4) were repeated for different ow rates (i.e., 2.03
to 16.21 m3/s) and oil water ratios (i.e., 10 to 90%).

crude oil was found to obey the Newtonian uid characteristic.

From the graph, an empirical equation was established to determine the viscosity of the specic crude oil at 30 1C from the shear
stress and shear rate relationship as follows:

0:0172_

The viscosity of oil (mo can be calculated as follows:

3.6. Crude oil characteristic

Contextually crude oil is known as a cocktail since various
substances are existed in the liquid. The behaviour of the crude oil
itself is considered relatively complex due to the presence of
parafn or asphaltenes and other elements in the oil which may
greatly inuence its characteristic. Table 2 shows the composition
of the waxy crude oil used in this research work.
According to their research work on oilwater ow in horizontal conditions, Wang and Gong (2010) discussed the differences between mineral oilwater ow and crude oilwater ow in
a two phase ow system. They found that there were great
alterations in terms of two phase ow characteristics when using
mineral oil in comparison to crude oil due to the crude oil
compositions. Since natural surfactants, such as asphaltenes and
resin, are present in crude oil, the oil and water phases have a
strong propensity to form water-in-oil emulsion and this kind of
emulsion is usually found to be quite stable, such that it contributes to a complex oilwater two phase ow in a pipeline
system. Nevertheless, Wang and Gong (2010) did not detail out the
crude oils composition which depending on its composition,
might pose a substantial effect on the crudes behaviour. The
details of crude oil composition are imperative since they inuence the oils characteristics. According to Table 2, there are few
complex elements contained in the crude oil which can be found
at number of components of 18, 24, 26, 27, and 28.
The waxy crude oil used in this research work was found to
behave as a Newtonian uid at 30 1C as shown in viscosity graph in
Fig. 5. From the graph of shear stress versus shear rate curve, it
shows that the crude oil exhibits a linear curve. This showed the

mo

where and _ are stand for shear stress and shear rate,
respectively.
It should be noted that this study did not investigate the effect
of each chemical component of the crude oil towards their
behaviour. We conned ourselves to the investigation of the ow
pattern, pressure drop, and water holdup of the waxy crude oil
water ow in a horizontal pipe.

4.0
3.5
Shear stress (D/cm 2)

414

3.0
2.5
2.0
1.5
1.0
0.5
0.0
0

20

40

60

80

100 120 140 160 180 200 220

Shear rate (1/sec)

Fig. 5. Dynamic viscosity of the crude oil at 30 1C.

Table 2
Crude oil composition detected using GCMS equipment.
Number of components

Components

Carbon, C

Percentage (%) by weight

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

Nonane
Decane
Undecane
Hexacosane
Pentacosane
Tetradecane
Pentadecane
Hexadecane
Heptadecane
Tetracosane
Octacosane
Nonadecane
9,10-Dihydrophenanthracene
Icosane
Heneicosane
Benzeneacetic acid
Triacontane
Allylpentaspiro[3.0.3.0.3.0.3.0.3.1] henicosan-21
Tricosane
Tetracosane
Docosane
Pentacosane
20 ,40 -Dimethyloxanilic acid
2-Methyl-3-phenyl-1H-indole
Dodecahydropyrido[1,2-b]isoquinolin-6-one
1,3-Dimethyl-4-azaphenanthrene
2-(Acetoxymethyl)-3-(methoxycarbonyl)biphenylene
3,3-Diisopropoxy-1,1,1,5,5,5-hexamethyltrisiloxane

C9H20
C10H22
CH3(CH2)9CH3
CH3(CH2)24CH3
CH3(CH2)23CH3
CH3(CH2)12CH3
CH3(CH2)13CH3
CH3(CH2)14CH3
CH3(CH2)15CH3
CH3(CH2)22CH3
CH3(CH2)26CH3
CH3(CH2)17CH3
C14H12
C20H42
CH3(CH2)19CH3
C8H8O2
CH3(CH2)28CH3
C21H30O
CH3(CH2)21CH3
H(CH2)24H
CH3(CH2)20CH3
CH3(CH2)23CH3
C8H16O2
C15H13N
C13H21NO
C15H13N
C17H14O4
C12H32O4Si3

2.40
3.94
1.64
3.31
3.09
3.68
4.20
5.07
4.38
4.71
3.89
4.33
4.00
4.56
4.60
3.96
4.08
3.38
4.64
3.94
2.72
4.28
3.32
3.19
5.97
0.55
0.95
1.23

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

4. Results and discussion

4.1. Flow pattern analysis
Flow pattern identication comprised the static image capturing
and video recording throughout the respective mixture velocities
and water cuts. Since the crude oil was black once dispersed, it
affected visual discrimination. Thus, to enhance the visual observation of free water and images, a predetermined quantity of
uorescein powder (i.e., 0.1 g per 1500 ml of water) was added to

the water. After the water was mixed with uorescein powder, it
became luminous once introduced with the UV light. This could
distinguish the ow patterns better when oil and water owed
together in the pipeline.
Various types of ow pattern have been observed in the case of
oilwater ow in horizontal pipes in the literature (Arirachakaran et
al., 1989; Trallero et al., 1997; Atmaca et al., 2009; Vuong et al., 2009).
The basic ow patterns observed were stratied ow, large slug
elongated or spherical, dispersed ow, annular ow and combination
of these four basic ow patterns (Brauner, 2002). Many of the

Table 3
Photographic examples of observed ow pattern.
Types of ow pattern
Stratied wavy ow (STW)

Stratied wavy with semi dispersed ow at interface and oil lm (STSD&O)

Semi dispersed ow with semi emulsion at interface and thin oil lm

(SDSE&TO)

Dispersion of water in oil and emulsion (DWE)

Dispersion of oil in water with water continuous (DO)

415

Examples of observed ow pattern

416

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

prominent researchers like Trallero et al. (1997), Flores et al. (1999),

Lovick and Angeli (2004), Atmaca et al. (2009) and Angeli and Hewit
(1998, 1999) have discussed the ow patterns by referring to mineral
oil or rened oil as the test uids in an oilwater system. However,
there was limited discussion about the waxy crude oilwater two
phase ow patterns in horizontal pipes. It should be noted that
different types of uid behave differently in a liquidliquid system
and consequently, this issue should be given a serious consideration. In
this study, when a waxy crude oil was introduced into the oilwater
two phase ow system, the basic ow patterns have been expectedly
observed. This research has revealed a new ow pattern during the
continuous ow; that may be due to the complexity of the Malaysian
waxy crude oil. This phenomenon was known as oil lm and emulsion
and such a ow pattern was never discussed in open literature. Even
though Vuong et al. (2009) and Sridhar et al. (2011) have discussed
about the presence of oil lm but, the oil lm discovered by them was
just a very thin layer without the presence of emulsion. Different oil
properties, water ratios, ow rates and the complexity of waxy
Malaysian crude may be involved in the presence of the new ow
pattern. The oil lm and emulsion was found to have formed during
the dispersed ow of oil in water.
This research has identied ve types of ow pattern formed
during the waxy crude oilwater ow in a horizontal 5.08 cm ID
pipe at different oilwater ratios and mixture ow rates. The ve
ow patterns were: (1) stratied wavy ow (STW), (2) stratied
wavy with semi dispersed ow at interface and oil lm (STSD&O),
(3) a newly found semi dispersed ow with semi emulsion at
interface and thin oil lm (SDSE&TO), (4) dispersion of water in oil
and emulsion (DWE) and (5) dispersion of oil in water with water
continuous (DO). Table 3 shows the ow patterns observed during
the waxy crude oilwater two phase ow in a horizontal pipe at
ambient condition (30 1C). It was found that the waxy crude oil
used in the experimental work has the tendency to stick on the
wall of pipe due to its viscosity and pipe wettability effect. This
effect is due to the oil-wet nature of the pipe (i.e., oil is adsorbed to
the pipes inner surface) and might not translated to the wax
formation in the ow experiment. As shown in Table 1, the WAT is
26 1C and the experiment was conducted at ambient condition (i.
e., 30 1C). That is to say the washing pipe with some additives
should convert the pipes inner surface from oil-wet into waterwet state. As the oil fraction and ow rate increased, the oil lm
was found to have encapsulated the water phase by forming a
circular layer around the pipe with an emulsied layer at the
interface of oil and water, as depicted in SDSE&TO.
The STW commonly occurs when two layers of oil and water
are completely separated with the oil ows at the top and water
ows at the bottom of the pipe (Sridhar et al., 2011). This type of
ow pattern was clearly seen in this study during the low and
intermediate mixture velocities ranging from 2.03 to 8.10 m3/s for
Cw 40.5. While waviness was observed at the interface of crude oil
and water layers due to their density and viscosity difference at
moderate velocities. Trallero (1995) explained that the oilwater
ow was fully segregated at low velocities due to the effect of
gravity. This behaviour could be clearly seen when the lighter
liquid occupied the upper part of the test section while the heavier
density liquid is in the lower part of the test section.
The STSD&O was found to have occurred as the ow rate
increased the amplitude of the wave at the interface; thus creating
a semi dispersed ow (i.e., oil droplets in water phase) at the
interface while water and oil phase were owing in the pipeline.
Elseth (2001) also found that as the velocity was increased from
slow to moderate, there was a large number of droplets emerged
and the interface became even wavier. As sufcient energy was
supplied, the droplets broke away from the waves. The droplets
became bigger and eventually formed a semi dispersed ow at the
interface. Considering the crude oil consisted of natural emulsiers

as discussed by Wang and Gong (2010), therefore the tendency for

emulsion (water in oil) to develop was higher. It could as well
affect the energy breakage of the wave and droplets at the
interface between oil and water.
The oil lm started to appear as the water cut and the mixture
ow rate decreased. The oil lm was formed on the inner surface of
the pipeline due to its viscosity and pipe wettability. The acrylic
pipe is an oil wet material thus it inuences the oil to adhere to the
surface of the pipe. This behaviour could be observed clearly at
mixture velocities ranging from 2.01 to 14.18 m3/s for all water cuts.
This phenomenon was more favourable in the water dominated
region. The Malaysian waxy crude oil behaves like a gluey liquid due
to its viscosity, wax contents and complex natural chemical compositions. Therefore, oil tends to stick on the surface of the pipe in a
continuous oil ow and eventually creating an oil lm. This phenomenon happened while the water cut decreased and simultaneously the ow rate increased. The static images of ow rate at
8.0 m3/s, as water cut was increased from 50% to 70% show the
presence of oil lm on the inner surface of the acrylic pipe.
Above moderate ow rates (i.e., 10.0 m3/s), a SDSE&TO ow was
encountered because of sufcient energy which has been distributed to the oil and water phases. This phenomenon has created
partial droplets at the interface and is distributed across the pipe.
The mixing rate between oil and water layers was found to have
increased as the ow rates were increased, which results in creating
a stabilized semi dispersed ow. This phenomenon encouraged the
formation of oilwater emulsion. Semi dispersed ow with partial
emulsion occurred at water cuts ranging from 10% to 80% with ow
rates from 6.80 to 14.18 m3/s. The SDSE&TO was considered a newly
found ow pattern for oilwater ow in this experiment because
none of the ow researchers have ever explained or mentioned
such a ow pattern in their published works. This pattern was
largely contributed by the heavier parafn components, asphaltenes, resin, and natural emulsiers in the Malaysian waxy crude
(see Table 2).
As the mixture velocities further increased up to vm 412.15, a
full dispersed ow was found to have formed due to the breakup of
semi dispersed ow and increased mixing rate between phases.
This phenomenon produced either oil in water dispersion (DO) or
water in oil dispersion (DWE) depending on the dominant phase.
Since the ow rate was relatively high, all the droplets were
distributed across the pipe and this has created a full dispersed
ow. Such a situation denitely induced the formation of oilwater
emulsion. Generally, waxy crude oil contains a large amount of
higher molecular weight hydrocarbons such as parafn and asphaltenes. Therefore, the physical interaction between the crude oil and
water above WAT tends to develop an emulsion of water in oil
(DWE). The physical interaction refers to contact between two
immiscible liquids that mix spontaneously during a liquidliquid
ow in which one is present as small droplets or microscopic size
globules that have distributed in the other phase. Since the utilized
waxy crude oil contains some polar molecules (see Table 2) and
shows Newtonian characteristics at above WAT thus the interaction
becomes apparent and eventually emulsion has been created. There
was no chemical reaction supposed to take place because dead
crude and the experimental works were conducted at ambient
condition. But the waxy crude oil starts to crystallize below WAT
and this phenomenon will become more complex and hard to form
emulsions; this was beyond the scope of the study.
At high water cuts, it was expected to have a full dispersed ow of
oil in water with water continuous (DO) since the continuous phase
was dominating the ow at this condition. This behaviour was
observed at higher water fraction (i.e., 90%) and higher ow rate (i.
e., 16.21 m3/s). Trallero (1995), Nadler and Mewes (1997) and
Arirachakaran et al. (1989) also encountered such a phenomenon
in their research work.

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

4.1.1. Flow pattern map

A ow pattern map, as shown in Fig. 7, has been successfully
generated using an actual waxy crude oil in this study. Other
researchers who had experienced the same procedure were Wang
and Gong (2010) in their comparative studies. However, their
experiment was limited to non-waxy crude oilwater ow and
mineral oilwater ow in horizontal conditions. They decided to
use the mineral oilwater ow as the point of comparison with crude
oilwater ow behaviour because mineral oil was widely used in the
area of liquidliquid studies. The use of crude oil is of utmost
importance because it represents the actual behaviour of the oil
during oil production and transportation (Vielma et al., 2007).
The boundary layers generated on the ow pattern map were
to show the phase dominated region. In comparison of the
generated map shown in Fig. 8 with the boundary layer produced
by Wang and Gong (2010), it was found that different uids used
for the two phase ow system gave different ow patterns at
different conditions. The discrepancy between the generated ow
pattern map and Wang and Gong ow pattern map proved that
different uids behaved differently. The boundaries were mainly
affected by viscosity of the crude oil as well as the existence of
surface active (Yarranton et al., 2000) such as asphaltenes, resin,
etc. which could decrease signicantly the surface energy between
oil and water phases. This phenomenon would enhance the oil
domination region and phase inversion from oil in water to water
in oil emulsion.
The crude oil also has a different hydrodynamic ow when it
ows together with water in a horizontal pipe. It is obviously
shown in Fig. 8 that those parameters have greatly inuenced the
formation of boundaries for water dominated and oil dominated
regions. The experimental results demonstrated that the oil
dominated region began at moderate ow rates while intermittent
was found to have formed from as low as 2.0 m3/s ow rate for low
water cuts. On the other hand, Wang and Gong (2010) found that
the dispersed ow of water in oil started as early as 2.0 m3/s ow
rate and 50% water cuts. The reasons why actual crude experiments were limited are due to its hazardous nature, presence of
emulsion and difculties in ow pattern identication because of
its opaqueness. The challenges faced by the oil and gas industry in
producing and transporting waxy crude oil is another complemented factor that deters researchers from using waxy crude oil in
their studies.

4.2. Water holdup analysis

The in-situ volume fraction of water over total mixing liquids in
a specic length of a test section is commonly known as water
holdup, Hw. The Hw obtained from the research work is shown in
Fig. 9. Generally, water holdup cannot be calculated analytically,
but can be determined empirically because it is a function of
variables such as liquid properties, ow pattern, pipe diameter,
and pipe inclination. Two quick closing valves were installed at
each end of the transparent section to trap the water holdup. For
accurate measurement of liquid holdup, type of quick closing valve
used has to be chosen carefully. This is the aspect that many
researchers overlooked or unnoticed during the liquid holdup
determination. A solenoid valve was found to be the best option.
Varseveld and Bone (1997) highlighted that a solenoid valve has
less than 1% error of control accuracy and is highly recommended
for a quick closing valve system. It is fast and directly actuated as
the switches turned on and therefore could avoid human error as
compared to the use of a typical ball valve which might experience
a slip during opening and closing of both valves at two separate
points during an oilwater ow in a pipe.

18.0
16.0
14.0

Flow rate (m3/s)

The two phase ow experimental work conducted on the

Malaysian waxy crude oilwater in a horizontal pipeline has
revealed a new ow pattern called SDSE&TO. The ow pattern
was found at water fraction, C w o 0.8, and ow rates ranging from
6.08 to 14.18 m3/s. The cross sections of all the ow patterns are
depicted in Fig. 6. For SDSE&TO ow, there was a third phase
present; an emulsion between oil and water phases. The emulsion
phase grew thicker as the oil fraction and mixture velocities
increased. In the stage of intermittent ow, water dispersed in
oil as discrete droplets and was distributed over the entire pipes
cross section. The shape of the dispersed water droplets was from
small spherical to irregularly globules. Nevertheless, the size and
distribution depended on the ow rate.

417

12.0
STW

10.0

STSD&O
SDSE&TO

8.0

DWE

6.0

DO

4.0
2.0
0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Water fraction
Fig. 7. Flow pattern map generated for the waxy crude oilwater system at 30 1C.

Fig. 8. The sketch map of ow regions.

Fig. 6. Cross-sectional view of ow patterns chronological behaviour: (a) STW (b) STSD&O (c) SDSE&TO (d) DWE (e) DO.

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

The water holdup, as shown in Fig. 9, was found to have

increased with water fraction. It was also found that no signicant
effect of viscosity on water holdup can be inferred from the
experimental results; these ndings are in good agreement with
those found by Vuong et al. (2009), Atmaca et al. (2009) and Sridhar
et al. (2011). For the stratied ow, the slippage effect was found to
be signicant since oil was lighter than water in terms of density
thus it moved faster than water in a horizontal condition. However
in a dispersed ow, the slippage was insignicant. Soleimani et al.
(2000) found that the holdup was inuenced by the input volume
fraction at the minimum ow rate as well as the increase in slip
ratio. This trend, according to Charles et al. (1961) and Martinez
et al. (1988), was due to the contact between the liquid and the
pipe, and the liquid was likely to be accumulating in the pipe at a
slower velocity. When oil and water travelled concurrently in a
pipe, oil was found to travel faster than water causing a slippage
between the phases. Because of the slippage, the in-situ water
volume fraction at any particular section in the pipe could not be
computed directly from the input conditions.
The water holdup reduced at a lower ow rate due to the drag
caused by the pipes wall. Generally, the drag force slows down the
ow of the water with respect to oil and water ows as a thin layer
at the bottom of the pipe at a very slow supercial velocity (Vielma
et al., 2007). The water holdup was found to have increased with
the ow rate and water cut because the water layer became thicker
and owed faster due to its low viscosity. As for transition from
water dominated to oil dominated regions, the holdup experienced
a gradual change due to the presence of emulsion at moderate to
higher ow rate. Whereby at the stratied ow, the water holdup
was found to have increased but upon transition to oil dominated
regions, the water holdup reduced considerably. Table 4 tabulates
the statistics of Hw values for each ow rate from 10% to 90% water
holdup.
4.3. Pressure drop analysis
Flow patterns obtained from this research were found to be
strongly depended on ow rates and ow pattern, as supported by
Vuong et al. (2009), Vielma et al. (2007) and Trallero (1995). As
depicted in Fig. 10, the pressure drop increased as oil fraction and
ow rates were increased. Pressure drop started to increase
gradually at a ow transition from stratied ow to dispersed
ow (intermittent region) specically from STW to STSD&O with
ow rate 46.08 m3/s and water fraction, C w o 0.8. As approaching
to the dispersed ow in the oil dominated region known as semi
dispersed ow with SDSE&TO, there was a marginal increase in
pressure drop because the waxy crude oil at this condition has
formed a thin layer on the inner surface of the transparent pipe
1
0.9

Water holdup, Hw

0.8

10% oil
20% oil

0.7

30% oil
40% oil

0.6

50% oil

0.5

60% oil

0.4

70% Oil

0.3

90% oil

80% oil

0.2
0.1
0
0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

Flow rate (m3/s)

Fig. 9. Water holdup versus ow rate at various oil fractions.

Table 4
Statistics of Hw values for different ow rate and oilwater fractions.
Flow rate (m3/s)

Min

Max

Average

Std. dev

2.03
4.05
6.08
8.10
10.13
12.15
14.18
16.21

0.22
0.26
0.24
0.19
0.17
0.16
0.15
0.14

0.76
0.85
0.89
0.91
0.91
0.92
0.92
0.95

0.49
0.55
0.58
0.57
0.54
0.51
0.48
0.46

0.18
0.18
0.20
0.23
0.24
0.25
0.25
0.25

12.00
10.00

Pressure drop (kPa)

418

90% water
80% water

8.00

70% water
60% water
50% water

6.00

40% water
30% water

4.00

20% water
10% water

2.00
0.00
0.0

5.0

10.0

15.0

20.0

Flow rate (m3/s)

Fig. 10. Pressure drop versus ow rate at various water fractions.

and thus increasing the drag force as well as the viscosity due to
the presence of non-stabilized emulsion. The existence of nonstabilized emulsion of water in oil has altered the hydrodynamic
behaviour of the mixture uids that has eventually increased the
pressure drop.
As the water fraction reduced from 80% to 10% and ow rate
increased to 12.15 m3/s, there was an increase in pressure drop. This
increase was due to the dispersion of water in oil and loss of water
continuity, during the formation of dispersion of water in oil and oil
continuous with emulsion (DWE). At water fraction 0.9 and ow
rate 0.8 m3/s, the pressure drop decreased marginally because the
water phase owed and wetted continuously the pipes wall while
dominating over a lighter phase (i.e., oil). Water was found to have
fully wetted the transparent pipes wall, thus this phenomenon has
reduced the drag force and subsequently mitigated the pressure
drop compared to high fraction of oil ow (C o 40.2). A similar case
was observed in the water dominated region. Trallero et al. (1997)
explained that at water dominated region (i.e., water in continuous
ow), the drag force of the ow caused by the shear stress between
the water phase and pipes wall was insignicant, and eventually
reduced the pressure drop. It was also found that the ow rates
with a thin oil lm would cause a higher pressure drop due to the
existence of shear stress near the pipes wall which was contributed
by parafn and other natural complex compounds of the waxy
crude oil. The water holdup at this type of ow pattern (with the
presence of wax) was found to have reduced because a thin layer of
wax has formed on the pipes wall and subsequently reduced the
effective ow area of the water.
Apart from ow rate and ow pattern which were found to be
the main factors affecting the pressure drop, Vielma et al. (2007)
highlighted that ow regime was another factor which inuenced
the pressure drop. They found that in the turbulent ow, the mixing
uids shifted from separated phases to dispersed ow that has
resulted in a greater pressure drop. The research nding was found
to be in good agreement with Vielma et al. (2007). It increased as

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

Table 5
pressure drop statistics for different ow rate and oilwater fractions.
Flow rate (m3/s)

Min

Max

Average

Std. dev

2.03
4.05
6.08
8.10
10.13
12.15
14.18
16.21

0.44
0.64
0.76
0.87
1.01
1.17
1.33
1.56

1.36
1.59
1.86
2.02
2.16
2.39
2.87
3.86

0.85
1.07
1.25
1.42
1.56
1.73
2.05
2.71

0.27
0.29
0.34
0.38
0.39
0.40
0.50
0.71

the ow changed from stratied to dispersed ow. Meanwhile, the

pressure drop changes were found to be closely related to phase
inversion or ow pattern transitions (Arirachakaran et al., 1989;
Trallero et al., 1997; Flores et al., 1999; Poesio et al., 2008; Vuong
et al., 2009). Wang and Gong (2010) and Ngan (2010) also have
encountered the same phenomenon. Trallero et al. (1997) who
highlighted that insignicant pressure drop was experimented
during the transition from stratied to a dispersed region. This
was due to the loss of oil continuity where the water was found to
have fully wetted the pipe wall and thus reducing the drag of the
ow that caused by the shear stress. On the other hand, when waxy
crude oil (emulsion) owed together with water, the crude oil
formed a thin layer of oil lm on the pipes wall, thus increasing the
shear stress and substantially caused an increase in pressure drop.
In the semi dispersed ow with semi emulsion at interface and thin
oil lm (SDSE&TO), when the oil became more dominant, it
encapsulated the water phase and subsequently formed a thin layer
of oil lm on the pipes wall. This phenomenon would increase the
pressure drop due to the shear stress between the oil and pipes
wall. Strazza et al., 2011, Brauner (1991), Bannwart (1998), Vuong
et al. (2009) and Sridhar et al. (2011) also have explained thoroughly this scenario. Table 5 tabulates the statistics (maximum,
minimum, average and standard deviation) of pressure drop values
for each ow rate (2.03 to 16.21 m3/s) with different oilwater
concentration (i.e., 10% to 90%).

5. Conclusions
The reported experiment is the rst stage study to evaluate the
behaviour of crude oil in the subsea ow lines (horizontal pipes)
carrying waxy uid from the offshore Terengganu to onshore
facilities above WAT. A set of experiments with light Malaysian
waxy crude and synthetic formation water have been conducted
and the following conclusions can be drawn:
(1) A new ow pattern map (i.e., SDSE&TO) was established for
the Malaysian waxy crude oilwater ow in a horizontal pipe
at ambient condition (30 1C). In general, ve ow patterns
were observed during the experimental works, namely STW,
STSD&O, SDSE&TO, DWE and DO.
(2) The water holdup was found to be highest (95.1%) at dispersion of oil in water with water continuous (DO) and lowest
(1.90%) at water dispersed in oil and oil continuous with
emulsion (DWE). A higher pressure drop was obtained at the
highest ow rates (i.e., 16.21 m3/s) and oil fraction, C o 0.9
where P 11.58 kPa, while the lowest pressure drop was
recorded at the lowest ow rates (i.e., 2.03 m3/s) and at
highest water fraction, C w 0.9 where P 1.31 kPa.
As the present study represents a rst attempt to study liquid
liquid ow pattern in a light Malaysian waxy crude, while the
available data is limited, wide experimental data for various two-

419

uid pairs is still required in order to further validate and extend

the proposed models. The experimental results could be used to
develop a theoretical model as there was limited published works
on waxy crude oilwater ow in open literature. On top of that, the
ow behaviour of a liquidliquid system which uses waxy crude is
still obscure and incomprehensible due to the substantial effect of
emulsion and wax deposition. Despite that, prediction of the ow
behaviour, such as ow pattern, pressure drop, and water holdup, is
crucial in economic consideration because an accurate prediction
can lead to production optimization for transport of these uids. In
the current experimental study, no signicant effect of viscosity on
water holdup was inferred. In order to obtain an in-situ phase
distribution and consequently provide more information on the
phase holdups, instrumentation is deantly required to obtain an
in-situ phase distribution of waxy crude oilwater ow system. In
the future, we expect even more relevant studies, such as investigation of waxy crude oil behaviour at or below WAT, to evaluate the
gel up condition and restart pressures needed after long-term ow
line shutdown. Accordingly, accurate models will be possible to be
developed based on the actual crude oil representation to address
current and future challenges, and also to support the understanding of a more complex case of gas/oil/water system.

Acknowledgments
The authors would like to express their sincere gratitude to The
Ministry of Higher Education (MoHE) of Malaysia and Universiti
Teknologi Malaysia (UTM) for funding this research project via the
Fundamental Research Grant Scheme (FRGS) Vote 4F136 and
Research University Grant (RUG) Vote 01H68, respectively.

References
Alkaya, B., 2000. Oilwater Flow Patterns and Pressure Gradient in Slightly Inclined
Pipes (Master Thesis). University of Tulsa, Oklahoma.
Angeli, P., Hewit, G.F., 1998. Pressure gradient in horizontal liquidliquid ows. Int.
J. Multiphase Flow 24, 11831203.
Angeli, P., Hewit, G.F., 1999. Flow structure in horizontal oilwater ow. Int.
J. Multiphase Flow 26, 11171140.
Arirachakaran, S., Oglesby, K.D., Malinowsky, M.S., Shoham, O., Brill, J.P., 1989. An
analysis of oil/water ow phenomena in horizontal pipes. In: SPE-18836
Presented at the SPE Production Operation Symposium, Oklahoma, U.S.A.
Atmaca, S., Sarica, C., Zhang, H.Q., Al-Sarkhi, A.S., 2009. Characterization of oil/water
ows in inclined pipes. In: SPE-115485-PA-P Presented at the SPE Annual
Technical Conference and Exhibition, Denver, U.S.A.
Bannwart, A.C., 1998. Wavespeed and volumetric fraction in core annular ow. Int.
J. Multiphase Flow 24, 961974.
Beggs, H.D., Brill, J.P., 1973. A study of two-phase ow in inclined pipes. SPE 4007.
J. Pet. Technol. 255, 607617.
Brauner, N., 1991. Two phase liquidliquid annular ow. Int. J. Multiphase Flow 17,
5976.
Brauner, N., 2002. Modeling and Control of Two Phase Phenomena: LiquidLiquid
Two Phase Flow System. CISM Center. School of Engineering, Tel-Aviv University, Udine, Italy http://www.eng.tau.ac.il/  brauner/publication.html.
Brennen, C.E., 2005. Fundamentals of Multiphase Flow. Cambridge University Press
p. 2005.
Charles, M.E., Govier, G.W., Hodgson, G.W., 1961. The horizontal pipeline ow of
equal density oilwater mixture. Can. J. Chem. Eng. 39, 2729.
Charles, M.E., Lilleleht, L.V., 1966. Correlation of pressure gradients for the stratied
laminar-turbulent pipeline ow of two immiscible liquids. Can. J. Chem. Eng.
44, 4749.
Chen, S., Oye, G., Sjoblom, J., 2006. Characterization and rheological properties of
waxy oils. Annu. Trans. Nordic Rheol. Soc., 14.
Cornish, R.E., 1976. The vertical multiphase ow of oil and gas at high rates. SPE
5791. J. Pet. Technol. 28, 825826.
Elseth, G., 2001. An Experimental Study of OilWater Flow in Horizontal Pipes (Ph.
D. Thesis). Norwegian University of Science and Technology.
Flores, J.G., Chen, X.T., Sarica, C., Brill, J.P., 1999. Characterization of oil/water ow
pattern in vertical and deviated wells. SPE-56108. J. SPE Prod. Facil. 14 (2),
102109.
Gafanova, O.V., Yarranton, H.W., 2001. The stabilization of water-in-hydrocarbon
emulsions by asphaltenes and resins. J. Colloid Interface Sci. 241, 469478.

420

A.S. Izwan Ismail et al. / Journal of Petroleum Science and Engineering 127 (2015) 409420

Grassi, B., Strazza, D., Poesio, P., 2008. Experimental validation of theoretical models
in two-phase high-viscosity ratio liquidliquid ows in horizontal and slightly
inclined pipes. Int. J. Multiphase Flow 34, 950965.
Gregory, G.A., Aziz, K., 1975. Design of pipelines for multiphase (gas-condensate)
ow. JCPT 750302 J. Can. Pet. Technol. 14, 2833.
Guzhov, A.I., Grishin, A.D., Medredev, V.F., Medreva, O.P., 1973. Emulsion formation
during the ow of two immiscible liquids. Int. J. Multiphase Flow 8, 5861.
Guzhov, A.I., Medredev, O.P., 1971. Pressure losses in ow of two mutually
immiscible liquids. Int. Chem. Eng. 11, 104106.
Hagedorn, A.R., Brown, K.E., 1964. Experimental study of pressure gradients
occurring during continuous two-phase ow in small-diameter vertical conduits. SPE 940 J. Pet. Technol. 17, 475484.
Hughmark, G.A., 1971. Drop breakup in turbulent pipe ow. J. AIChE 17, 10001009.
Keskin, C., Zhang, H.Q., Sarica, C., 2007. Identication and classication of new three
gas/oil/water ows patterns. In: SPE 110221 Presented at the SPE Annual
Technical Conference, California, U.S.A.
Lovick, J., Angeli, P., 2004. Experimental studies on the dual continuous ow
pattern in oilwater ows. Int. J. Multiphase Flow 30, 139157.
Malinowsky, M.S., 1975. An experimental study on oil-water and air-oil-water
owing mixtures in horizontal pipes. (Master Thesis). University of Tulsa,
Oklahoma, U.S.A.
Mansoori, G.A., 1993. Parafn/wax and Waxy Crude Oil: Principles of
Nanotechnology-Molecular Based Study of Condensed Matter in Small Systems.
World Scientic, pp. 14.
Martinez, A.E., Intevep, S.A., Arirachakaran, S., Shoham, O., Brill, J.P., 1988. Prediction
of dispersion viscosity of oil/water mixture ow in horizontal pipes. In: SPE
18221 Presented at the SPE Annual Technical Conference, Houston, U.S.A.
McKibben, M.J., Gillies, R.G., Shook, C.A., 2000. A laboratory investigation of
horizontal well heavy oilwater ows. Can. J. Chem. Eng. 78, 743751.
Nadler, M., Mewes, D., 1997. Flow induced emulsication in the ow of two
immiscible liquids in horizontal pipes. Int. J. Multiphase Flow 23, 5568.
Ngan, K.H., 2010. Phase Inversion in Dispersed LiquidLiquid Pipe Flow (Ph.D.
Thesis). University College London, London.
Oddie, G., Shi, H., Durlofsky, L.J., Aziz, K., Pfeffer, B., Holmes, J.A., 2003. Experimental
study of two and three phase ows in large diameter inclined pipes. Int.
J. Multiphase Flow 29, 527558.
Oglesby, K.D., 1979. An Experimental Study on the Effect of Oil Viscosity, Mixture
Velocity and Water Fraction on Horizontal OilWater Flow (Master Thesis).
University of Tulsa, Oklahoma.
Poesio, P., Strazza, D., Sotgia, G., 2008. Very-viscous-oil/water/air ow through
horizontal pipes: pressure drop measurements and prediction. J. Chem. Eng.
Sci. 64, 11361142.
Rodriguez, O.M.H., Oliemans, R.V.A., 2005. Experimental study on oilwater ow in
horizontal and slightly inclined pipes. Int. J. Multiphase Flow 29, 527558.
Russell, T.W.F., Charles, M.E., 1959. The effect of the less viscous liquid in the
laminar ow of two immiscible liquids. Can. J. Chem. Eng. 37, 1824.
Russell, T.W.F., Hodgson, G.W., Govier, G.W., 1959. Horizontal pipeline ow of
mixture of oil and water. Can. J. Chem. Eng. 37, 917.

Shi, H., Cai, J., Jepson, W.P., 1999, Oil-water distributions in large diameter
horizontal pipelines.Presented in 4thMultiphase Flow and Heat TransferInternational Symposium, Xian, China. pp. 73-80.
Shi, H., 2001. A study of oil-water ows in large diameter horizontal pipeline. (Ph.D.
Dissertation). University of Ohio.
Soleimani, A., Lawrence, C.J., Hewitt, G.F., 2000. Spatial distribution of oil and water
in horizontal pipe ow. In: SPE 66906, SPE Annual Technical Conference and
Exhibition, Houston, U.S.A.
Sridhar, S., Zhang, H.Q., Sarica, C., Pareyra, E., 2011. Experiments and model
assessment on high-viscosity oil/water inclined pipe ows. In: SPE 146448,
SPE Annual Technical Conference and Exhibition, Colorado, U.S.A.
Stapelberg, H.H., Mewes, D., 1994. The ow regimes transitions, pressure loss and
slug frequency of liquidliquidgas slug ow in horizontal pipes. Int. J. Multiphase Flow 20, 285303.
Strazza, D, Grassi, B., Demori, M., Ferrari, V., Poesio, P., 2011. Core-annular ow in
horizontal and slightly inclined pipes: existence, pressure drops, and hold-up.
J. Chem. Eng. Sci. 66 (2853-10).
Trallero, J.L., 1995. Oilwater Flow Patterns in Horizontal Pipes (Ph.D. Thesis).
University of Tulsa. Oklahoma.
Trallero, J.L., Sarica, C., Brill, J.P., 1997. A study of oil/water ow patterns in
horizontal pipes. SPE 36609J. SPE Prod. Facil., 165172.
Varseveld, R.B., Bone, G.M., 1997. Accurate position control of a pneumatic actuator
using on/off solenoid valves. J. IEEE/ASME Trans. Mechatron. 2, 195204.
Vielma, M., Atmaca, S., Sarica, C., Zhang, H.Q., 2007. Characterization of oil/water
ows in horizontal pipes. In: SPE 109591, SPE Annual Technical Conference and
Exhibition, Anaheim, U.S.A.
Visintin, R.F., Lapasin, R., Vignati, E., DAntona, P., Lockhart, T.P., 2005. Rheological
behavior and structural interpretation of waxy crude oil gels. Langmuir 21,
62406249.
Vuong, D.H., Zhang, H.Q., Li, M., 2009. Experimental study on high viscosity oil/
water ow in horizontal and vertical pipes. In: SPE-124542-MS-P, SPE Annual
Technical Conference and Exhibition held in New Orleans, LA, U.S.A.
Wang, W., Gong, J., 2010. Flow regimes and transition characters of the high
viscosity oilwater two phase ow. In: SPE-130850-MS-P, CPS/SPE International Oil & Gas Conference and Exhibition held in Beijing, China.
Wicks, M., Dukler, A.E., 1960. Entrainment and pressure drop in concurrent gas
liquid ow: I. Air-water in horizontal ow. J. AIChE 6, 463468.
Yarranton, H.W., Hussien, H., Masliyah, J.H., 2000. Water-in-hydrocarbon emulsion
stabilized by asphaltenes at low concentration. J.Colloid & Interface Science
228, 5263.
Zhang, H.Q., Sarica, C., 2006. Unied modeling of gas/oil/water pipe ow-basic
approaches and preliminary validation. In: SPE 95749-PA-P, SPE Annual
Technical Conference and Exhibition held in Dallas, U.S.A.
Zhang, H.Q., Vuong, D.H., Sarica, C., 2010. Modeling high viscosity oil/water ows in
horizontal and vertical pipes. In: SPE-135099-MS-P, SPE Annual Technical
Conference and Exhibition held in Florence, Italy.