Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 9.3.1 - Interrogating the Effect of Bends on Two-Phase Gas-Liquid Flow using Advanced Instrumentation
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00222
 Material Information
Title: 9.3.1 - Interrogating the Effect of Bends on Two-Phase Gas-Liquid Flow using Advanced Instrumentation Experimental Methods for Multiphase Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Abdulkadir, M.
Zhao, D.
Sharaf, S.
Abdulkareem, L.
Lowndes, I.
Azzopardi, B.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: bend
ECT
WMS
void fraction
flow regime
 Notes
Abstract: When gas/ liquid mixtures flow around a bend they are subjected to forces additional to those encountered in a straight pipe. The behaviour of the flows at the inlet and outlet of the bend depends on the orientation of the pipes. Air/ silicone oil flows around a 90° bend have been investigated using advanced instrumentation: Electrical Capacitance Tomography (ECT), Hunt et al. (2004); Wire Mesh Sensor Tomography (WMS), Thiele et al. (2008); and high-speed video. The first two provide time and cross-sectionally resolved data on void fraction. The ECT probes were mounted 10 diameters upstream of the bend whilst the WMS was positioned either immediately upstream or immediately downstream of the bend. The downstream pipe was maintained horizontal whilst the upstream pipe was mounted either vertically or horizontally. The bend (R/D = 2.3) was made of transparent acrylic resin. From an analysis of the output from the tomography equipment, flow patterns were identified using both the reconstructed images as well as the characteristic signatures of Probability Density Function (PDF) plots of the time series of cross-sectionally averaged void fraction as suggested by Costigan and Whalley (1996). The superficial velocities of the air ranged from 0.047 to 4.727 m/ s and for the silicone oil 0.142 m/ s. Bubble/ spherical cap, slug, unstable slug and churn flows were observed before the bend for the vertical pipe and plug, slug, stratified flow when the pipe was horizontal. Bubble, stratified, slug and semi-annular flows are seen after the bend for the vertical 90 degree bend whilst for the horizontal 90 degree bend, the flow patterns remained the same as before the bend. These results are confirmed by the high-speed videos taken around the bend. Gardner and Neller (1969) proposed a modified Froude number (Fro =Um2/ Rgsino =1) criterion for the occurrence of stratification in a horizontal pipe downstream of a bend. However, the results obtained from the present study concluded that the downstream flow regime was stratified wavy flow.
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: VID00222
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 931-Abdulkadir-ICMF2010.pdf

Full Text

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


Interrogating the Effect of Bends on Two-Phase Gas-Liquid Flow using Advanced
Instrumentation


Mukhtar Abdulkadir, Donglin Zhao, Safa Sharaf, Lokman Abdulkareem, lan Lowndes and Barry
Azzopardi

University of Nottingham, Faculty of Engineering, Process and Environmental Engineering Research Division, Institute
University Park, Nottingham, NG7 2RD, United Kingdom
ian.lowndes @nottingham.ac.uk


Keywords: bend, ECT, WMS, void fraction, flow regime




Abstract

When gas/ liquid mixtures flow around a bend they are subjected to forces additional to those encountered in a straight pipe.
The behaviour of the flows at the inlet and outlet of the bend depends on the orientation of the pipes. Air/ silicone oil flows
around a 900 bend have been investigated using advanced instrumentation: Electrical Capacitance Tomography (ECT), Hunt
et al. (2i I'4); Wire Mesh Sensor Tomography (WMS), Thiele et al. (2008); and high-speed video. The first two provide time
and cross-sectionally resolved data on void fraction. The ECT probes were mounted 10 diameters upstream of the bend
whilst the WMS was positioned either immediately upstream or immediately downstream of the bend. The downstream
pipe was maintained horizontal whilst the upstream pipe was mounted either vertically or horizontally. The bend (R/D =
2.3) was made of transparent acrylic resin. From an analysis of the output from the tomography equipment, flow patterns
were identified using both the reconstructed images as well as the characteristic signatures of Probability Density Function
(PDF) plots of the time series of cross-sectionally averaged void fraction as suggested by Costigan and Whalley (1996).
The superficial velocities of the air ranged from 0.047 to 4.727 m/ s and for the silicone oil 0.142 m/ s. Bubble/ spherical
cap, slug, unstable slug and churn flows were observed before the bend for the vertical pipe and plug, slug, stratified flow
when the pipe was horizontal. Bubble, stratified, slug and semi-annular flows are seen after the bend for the vertical 90
degree bend whilst for the horizontal 90 degree bend, the flow patterns remained the same as before the bend. These results
are confirmed by the high-speed videos taken around the bend.

Gardner and Neller (1969) proposed a modified Froude number (Fre =Um2/ Rgsine =1) criterion for the occurrence of
stratification in a horizontal pipe downstream of a bend. However, the results obtained from the present study concluded
that the downstream flow regime was stratified wavy flow.


Introduction

Production and transportation engineers in the onshore and
offshore oil and gas industries have always been facing
technical and environmental challenges associated with
multiphase flows. For example, in an offshore
environment, it is economically preferable to transport
both gas and liquid through a single flow line and separate
them onshore. In this way, a significant cost can be saved
by eliminating the separate pipelines and phase separators
at the offshore platform of Floating Production and Storage
Operation. However, the instability problems caused by
the multiphase flow can ultimately damage the pipeline
system and this is unacceptable. The pipeline geometry
contains not just straight pipes but also fittings such as,
bends, valves, junctions and other fittings which make the
flow gas and liquid more complicated. These fittings may
lead to secondary flow, strongly fluctuating void fractions,
flow excursions, flow separation, pressure pulsations and
other unsteady flow phenomena. These phenomena can
cause problems such as bur-out, corrosion, and tube
failure, resulting in expensive outages, repairs, and early


replacement affecting plant reliability and safety. Among
these fittings, bends are often encountered in oil/gas
production system because of: as a result of terrain
undulation; flow line/riser combinations and at delivery
points to production facilities. The presence of a bend can
drastically change the flow patterns downstream of it.
The flow redistribution phenomenon in bends, however,
has received little attention. Literature on this subject is
very limited. Most of the investigations have been
restricted to single-phase flow (Eustice (1911); Dean
(1927; 1928); Jayanti (1990); Dewhurst et al. (1990) and
Spedding et al. (2k'1"4). Afew papers, Gardner and Neller
(1969), Carver (1984), Carver and Salcudean (1986), Ellul
and Issa (1987), Legius and Akker (1997), Azzi et al.
(2002), Azzi et al. (2005), Spedding et al. (2006), and
Benbella et al. (2008), address the issue of gas-liquid
systems but their experiments are confined to far smaller
pipes than industry size and to the working fluids with
physical properties very different from those dealt with by
industry.





Paper No


Single-phase flow in bends

Eustice (1911) published one of the earliest papers
reporting flow in bends. He employed dye visualization
studies with water. Laminar flow separation, reversal of
flow and greater turbulence were observed in a 900 bend
when R/d< 3. His experiments demonstrated the
existence of a transverse motion (secondary flow)
superimposed on the primary flow, represented in the form
of a pair of counter rotating longitudinal vortices.

Dean (1927; 1928) developed the first theoretical approach
on the motion of a fluid in a curved pipe. This has been
followed by several papers on this topic. Flow
measurements using a 3-Dimensional LDA system in a
square sectioned (0.1 x 0.1 m) 90 degree vertical bend was
reported by Dewhurst et al. (1990). Streamwise and
secondary velocities were obtained by these authors for
water flow upstream of the bend and at 800 into the bend.
Most of the studies of bends have been carried out for
single phase flow. Jayanti (1990) reviewed the research
into bends both under laminar and turbulent single phase
conditions.

Gas-liquid flow in bends

Two-phase flow patterns observed in bends are
qualitatively the same as those seen in straight pipes.
However, bends introduce a developing situation in the
flow pattern, whereby the relative positions and local
velocities of the two phases are redistributed.

According to Spedding and Benard (2006) designers
usually apply the general rule that a 900 elbow bend has a
pressure drop equivalent of 30-50 pipe diameters length of
straight pipe. However, a more exact method is desirable if
the estimation of pressure drop can have a critical impact
on operation or plant safety, such as on the downstream
side of a relief valve.

Gardner and Neller (1969) carried out visual and
experimental studies for bubble/ slug flow using a
transparent pipe of 76 mm diameter in a vertical 90 degree
bend of 305 and 610 mm radii of curvature, using air-water
at atmospheric pressure. The local air concentrations over
chosen cross-sections were measured. Their experimental
date were used to interpret the effect of the competition
between centrifugal and gravity forces on the flow
distribution in bends. They found out that gas can either
flow on the inside or outside of the bend depending on a
critical Froude number defined as
U2
Fr. =
Rg sin0 (1)

where Urn, the mixture velocity, m/ s; R, the radius of
curvature of the bend, m; 0, the angle of the bend. They
claimed that if Fr, is greater than unity, the air will hug
the inside of the bend, and if less than unity, the air will
hug the outside of the bend. In the case of Fr, =1, both
phases will be stratified. This conclusion, however, may
not be valid for working liquids with different viscosity,


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

density and surface tension.

Carver (1984) extended the work of Gardner and Neller
(1969) using a 2-Dimensional (D) numerical computation
method. They compared their results with those of Gardner
and Neller (1969). The agreement was not particularly
good. Carver and Salcudean (1986) recognized the
limitations of using the 2-D numerical approach to
simulate truly 3-D flows. They extended the 2-D model to
3-D and found that the 3-D model can give the similar
trend as observed by Gardner and Neller (1969). Ellul
and Issa (1987) developed an improved 3-D simulation
where substantially different solution algorithm was
adopted, a truncated gas momentum was added into their
momentum equations and both air-water and gas-oil
mixtures were considered. The simulation results showed
better agreement with the experimental data than those
obtained by the 2-D model in Carver (1984). No grid
sensitivity analysis was reported in the work of Ellul and
Issa (1987). The simulation result could not select the
optimum mesh size. No experimental data was available
at that time for them to validate their gas-oil simulation.

Legius and van der Akker (1997) carried out a numerical
and experimental study in a bend of 630mm radius of
curvature using air-water at atmospheric pressure. The
experimental facility consisted of a transparent acrylic
horizontal flowline (9 m long) connected to vertical riser
(4m in height) by the 630 mm radius bend. The diameter of
all pipes was 100 mm. Visual observation, 200 Hz digital
camera and an Auto-regressive modelling method were
used for flow regime identification. Slug and chur flow in
the riser and stratified, slug and a new regime called
"geometry enhanced slugging" in the flowline were
identified. The time dependent behaviour of two-phase
flow was modelled by an in-house code named Solution
Package for Hyperbolic Functions (SOPHY-2). Good
agreement between modelling and experiment results has
been found under almost all conditions except at higher
gas and lower liquid flow rates. However, the information
presented for the characteristics of slug flow is limited.
Important parameters like void fraction in liquid slug and
Taylor bubble were not presented. The dependence of the
Taylor bubbles and liquid slugs on the gas flow rates was
not examined. The sample frequency of 50-100Hz used by
these researchers seems too slow to get good spatial
resolution of signals.

It can be concluded that most of the work reported in the
literature on the multiphase flow in bends were carried out
in small diameter pipes with air-water as the model fluids.
Emphasis was on the determination of the pressure drop
and phase distribution inside the bends. The change of
flow structure before and after the bends was mainly
obtained by visualization and the underlining mechanism
for the change of flow patterns was not discussed. This
paper is aimed to provide a more complete understanding
on the flow phenomenon occurring in bends through the
comprehensive experimental investigation in both vertical
and horizontal pipes. The examined fluids are air and
silicone oil, through which the effect of liquid viscosity
and surface tension could be revealed. Advanced
instruments such as Electrical Capacitance Tomography





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


(ECT), Wire Mesh Sensors (WMS) and high speed video
camera have been used to measure void fractions before
and after the bend. The flow patterns were determined by
analyzing the Probability Density Function (PDF) of the
time series of void fractions. This analysis was validated
by the flow visualisation in the bend with the aid of a high
speed video camera.

Nomenclature


pipe diameter (m)
Dimensional
Froude number dimensionlesss)
gravitational constant (m/ s2)
radius of curvature of the bend (m)
velocity (m/ s)
cross-sectionally averaged void fraction


Subsripts
M mixture
SG superficial gas
SL superficial liquid
0 angle of bend

Experimental Facility

All experiments were carried out in an inclinable rig,
shown in Figure 1, which had been employed earlier for
annular flow studies by Azzopardi et al (1997), Geraci et al.
_(2'I I' Geraci et al. (2007b) and latterly for bubbly, slug
and chur flow studies by Hernandez Perez et al (2007),
and Abdulkadir et al. (2010). The experimental facility
consists of a main test section of the rig made from
transparent acrylic resin pipes of 67 mm inside diameter
and 6 m long to observe flow development over the test
section. For all experiments, testing pipe were maintained
at 6 m long from mixing section to make sure that they
were long enough for flow development. The test pipe is
set up so that it can be inclined from -5 to 900 being
mounted on a rigid steel frame which could be positioned
in intervals of 50, to enable the investigation of influence of
different inclinations on flow regimes. In this paper only
the experimental results obtained at the vertical and
horizontal positions are discussed.

Air was supplied from laboratory compressed air main and
measured by one of two variable area meters mounted in
parallel. In addition, the static pressure of air was
measured prior to entering the mixing section. Liquid was
taken from a liquid storage tank and pumped by a
centrifugal pump and metered by one of two variable area
meters mounted in parallel. Air was mixed with liquid at
the mixer before the mixture entered into the straight pipe,
passed through a 900 bend and finally goes to a separator
where air was released to the atmosphere from the top and
liquid, under the influence of gravity, returned to main
liquid storage. More details can be found in Abdulkadir
et al. (2010)


Air from compressed air main


Figure 1: Schematic diagram of inclinable rig. This rig
was converted to a vertical 900 bend and afterwards to
horizontal 900 bend.

A 900 bend with a radius of curvature of 0.1544 m was
mounted at the end of the straight pipe as shown in Figure
2. Downstream of the bend there is another straight pipe
from which the gas-liquid mixture enters a flexible pipe
that takes it to the phase separator. Two bend positions
were investigated: (1) vertical bend with upstream, vertical
riser and downstream horizontal flowline and (2)
horizontal bend with horizontal flow lines upstream and
downstream. The behaviour of the air-silicone oil
mixture was examined using the ECT and WMS.

A detailed description of the theory behind the ECT
technology is described by Hammer (1983), Huang (1995),
Zhu et al. (2003) and Azzopardi et al. (2008). The method
can image the dielectric components in the pipe flow
phases by measuring rapidly and continually the
capacitances of the passing flow across several pairs of
electrodes mounted uniformly around an imaging section.
Thus, the sequential variation of the spatial distribution of
the dielectric constants that represent the different flow
phases may be determined. In this study, a ring of
electrodes were placed around the circumference of the
riser at a given height above the injection portals at the
bottom of the 6 m riser section. This enabled the
measurement of the instantaneous distribution of the flow
phases over the cross-section of the pipe.

The WMS technology, described in detail by Da Silva et al.
(2010) can image the dielectric components in the pipe
flow phases by measuring rapidly and continually the
capacitances of the passing flow across several crossing
points in the mesh. It consists of two planes of 24
stainless steel wires with 0.12 mm diameter, 2.78 mm wire
separation within each plane and 2 mm axial plane
distance. This determined the spatio/ temporal resolution
of the sensor. Since the square sensor is installed in a
circular tube, only 440 of the total 576 wire crossing points
are within the radius of the tube (Abdulkadir et al. (2010)).
During the experiments, the horizontal transmitter lines are
pulsed one after another. By measuring the signal of all
crossing vertical receiver wires, the local capacitance
around the crossing points in the mesh is known. This
capacitance signal is a measure for the amount of silicone


Paper No






Paper No


oil, and thus indicates the local phase composition in the
grid cell.

The ECT was placed 4.49 m (67 diameters) away from the
mixing section while the WMS was placed at about 4.92 m
(73 diameters) away from the mixing section. The WMS
was afterwards located at a distance of about 0.021m (3
diameters) after the bend. The experiments were performed
at room temperatures (15-20 degree Celsius). The
properties of the two fluids used in the experiments are as
shown in Table 1. The influence of the bend orientation on
the flow behaviour was studied by changing the position of
the bend from vertical to horizontal.


Flanges


WMS1-



ECT: Plane 2

ECT: Plane T


Mixing
section


Probes


distance from mixing section


ECT plane 1
ECT plane 2
WMS1
Radius of Curvature


4.40 m
4.489 m
4.92 m
0.1544 m


Figure 2: Schematic diagram of vertical 90 degree bend
present study. The downstream pipe was maintained
horizontal whilst the upstream pipe was mounted vertically
or horizontally. This rig was instrumental to take physical
measurements that could be used to characterize the flow
regime existing before and after the bend when the flow
rates of both the oil and the air injection were varied

Table 1: Properties of the fluids
Fluid Density (kg/ Viscosity Surface
m3) (kg/ ms) tension (N/
m)
Air 1.18 0.000018
Silicone oil 900 0.00525 0.02

Results and Discussion

Several runs were carried out with air-silicone oil mixture
mainly for the purpose of defining and classifying all the
two-phase flow patterns attainable with the experimental
rig using the ECT and WMS. The liquid superficial
velocity investigated was 0.142 m/ s and gas superficial
velocities ranged from 0.047 to 4.727 m/ s. In each of


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

these runs, independent visual observations and flow
characteristics were recorded in detail for the vertical riser,
bends, and flowline (horizontal pipe). To test the accuracy
of the ECT and WMS, a comparison was made based on
bends, and flowline (horizontal pipe). To test the accuracy
of the ECT and WMS, a comparison was made based on
the Probability Density Function (PDF) of void fraction.
The results of the PDF of void fraction were further
reinforced by the cross-sectional slice view of void fraction
obtained from the WMS. Comparisons between the flow
regimes before and after the bends were made based on the
PDF of void fraction. All the flow regimes for the vertical
and horizontal 900 bends were then categorized based on
the basis of visual observations and supported by high
speed video image. It was observed that the transition from
one flow regime to another was gradual with respect to
fluid flow rates.

Comparison of PDFs of void fraction for the ECT and
WMS for the riser before the bend and reconstructed
images of the two-phase flow as depicted by the WMS


0 0.5
Void fraction
(a)
0.08


0.04 -
0.04


1 0 0.5
Void fraction
(b)


0.5
Void fraction
(c)


Figure 3: Comparison of PDFs of void fraction for the
ECT and WMS before the bend. In all cases the liquid
superficial velocity was 0.142 m/ s. Gas superficial
velocities: (a) 0.047m/ s (b) 0.544 m/ s (c) 4.727 m/ s.






Paper No


ta) (0) (c)
Figure 4: Reconstructed images of the two-phase flow
transition from spherical cap bubble to churn flow. In all
cases the liquid superficial velocity was 0.142 m/ s. Gas
superficial velocities: (a) 0.047m/ s (b) 0.544 m/ s (c)
4.727 m/ s.

Figures 3a to 3c presents the comparison between the
PDFs of void fraction for the ECT and WMS for same
flow conditions in a vertical riser. The plots show that at a
liquid superficial velocity of 0.047 m/ s and a gas
superficial velocity of 0.047 m/ s, both the ECT and WMS
defined the flow regime as spherical cap bubble.
However, the ECT provides a higher PDF while the WMS
predicts higher void fraction. Increasing the gas superficial


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

velocity to 0.544 m/ s whilst maintaining same liquid
velocity of 0.142 m/ s the characteristic signature of the
flow regime according to both the ECT and WMS is slug
flow. It can also be observed that the ECT still provide
higher PDF of void fraction than the WMS but with a
lesser range compared to at gas superficial velocity of
0.047 m/ s while the WMS provide a higher void fraction.
At a gas superficial velocity of 4.727 m/ s, the ECT and
WMS both defined the flow pattern as churn flow. It is
interesting to note that both the PDF of void fraction for
ECT and WMS show almost same peak. The degree of
agreement of the length of the PDF and void fraction
improves with an increase in gas superficial velocity. The
result therefore shows that both instruments predict the
same signature. Figures 8a to 8c show 2D slice view for
the void fractions observed for different gas superficial
velocities to support the result obtained in Figure 3a to 3c.
At a gas superficial velocity of 0.047 m/ s, there are still
bubbles of large size, but not as big as the pipe diameter.
When the gas velocity is increased to 0.544 m/ s,
coalescence starts leading to slug flow. At gas superficial
velocity of 4.727 m/ s, the slug flow transforms into churn
flow. This results obtained are in agreement with the
results obtained in Figures 3a to 3c.


Comparison of PDFs of void fraction before and after the bend using the WMS


Table 2: PDF of void fraction before and after the bend
Flow conditions PDF of void fraction before the bend PDF of void fraction after the bend
Liquid superficial o.3 0.3
velocity = 0.142 m/s u- u
Gas superficial a
velocity = 0.047 m/ s 0.2 0.2


0.1 ll 0.1


0 0
0 0.5 1 0 0.5 1
Void fraction Void fraction
Liquid superficial 0.06 0.06
velocity = 0.142 m/ s u_
Gas superficial
velocity = 0.544 m/ s 0.04 0.04


0.02 0.02


0
0 0 0.5 1
0 0.5 1 Void fraction
Void fraction





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


PDF is the rate of change of the probability that void
fraction values lie within a certain range (0 < e < 1)
versus void fraction. The PDF of time varying void
fractions has been used to classify the flow patterns in the
same manner as Costigan and Whalley (1996) and
Omebere-Iyari and Azzopardi (2007). A single peak at low
void fraction represents bubble flow and a double peak
feature with one at low void fraction whiles the other one
at high void fraction represents slug flow. A single peak at
low void fraction accompanied by a broadening tail
represents spherical cap bubble whiles a single peak at a
high void fraction with a broadening tail represents churn
flow.


The PDFs of the void fraction data obtained at a fixed
liquid superficial velocity 0.142 m/ s and variable gas
superficial velocities ranging from 0.047 to 4.727 m/ s are
shown in Table 2. The flow conditions are given in the first
column of the Table 2. In the second column, the PDF for
the riser before the vertical bend (the blue curve) are
compared with that for the flowline before the horizontal


bend (the red curve) at the same flow conditions. The third
column gives the same comparison for the scenario after
the bend.

At the low gas flow rate (gas superficial velocity = 0.047
m/ s), the PDF for the vertical riser presents a single peak
of the void fraction value at about 0.06 with a broadening
tail extending to a higher value of about 0.35. This defines
a "Spherical cap bubble flow" as in Costigan and Whalley
(1996). The flow pattern has been confirmed by the images
of high speed video camera as shown in Figure 5. The gas
bubbles indeed exhibit spherical cap shapes. When the
vertical riser was positioned to become horizontal
(flowline), the PDF of the void fraction at the same flow
rate, shows a dominant peak at 0.14 with a wide base
spanning from 0 to 0.36. This is the typical feature of plug
flow. The elongated gas bubbles are separated by sections
of continuous liquid moving downstream along the top part
of the pipe with almost zero void fractions in the liquid.
The variation of the void fraction reflects the different size
of the gas bubbles and the continuous liquid phase. After


Paper No






Paper No


the bend, the PDF for the horizontal pipe in the riser shows
a single peak, the signature of bubbly flow. The broaden
tail present in the PDF of the riser does not exist in the
PDF after the bend. The big cap bubbles break up in the
bend due to the balance of the centrifugal force and the
surface tension. The bubbles become more uniform. For
the horizontal setup, the PDF after the bend move to the
lower void fraction values with the dominant peak
frequency at the void fraction of 0.08. There is little change
in the size of the wide base compared with that before the
bend. With the same mechanism the elongated bubbles
break in to smaller bubbles when passing through the bend.
The flow patterns are still kept as plug flow though.

When the gas flow rate increases to 0.544 m/ s, the cap
bubbles coalescence into bullet-shaped Taylor bubbles and
the slug flow are formed. The PDF of the void fractions
in the riser gives two main peaks at the values of 0.2 and
0.5 respectively. These peaks are the signature of the
aerated liquid slugs and the Taylor bubbles with the
different sizes. Taylor bubbles have yet fully developed.
This claim is confirmed by the analysis of the video
images as shown in Figure 6. Similar to that for the riser,
the PDF for the horizontal flowline has one narrow peak at
the value of 0.02 and one wide peak with fluctuations at
around 0.7. With the increase of gas superficial velocity
from 0.047 to 0.544 m/ s, the elongated bubbles grow and
coalescence into bullet-shaped Taylor bubbles. The flow
pattern changes from the plug flow to the slug flow. After
the bend, the PDF of the void fraction in the flowline for
the riser has a "hill" shape. Stratified wavy flow was
observed. For the horizontal setup, compared with the PDF
before the bend, the lower void fraction peak moves to the
higher void fraction values and more peaks appear at the
void fractions 0.5 0.8. This can probably be attributed to
the collapse of the big Taylor bubbles while passing
through the bend. However, the flow pattern remained as
slug flow.


When the gas flow rate reaches 0.945 m/ s, again two
peaks appear on the PDF graph of void fractions for the
riser but the height of the lower void fraction peak
decreases more than 50% while the probability density of
the higher void fraction increases more than 40%
compared with those at 0.544 m/ s. The increase of gas
superficial velocity leads to the increase of Taylor bubbles
and the shrinkage of the liquid slugs. More and more
bubbles are entrained in the liquid slugs. This regime
according to Costigan and Whalley (1996) is defined as
unstable slug flow. For the case of the flowline
arrangement, the height of the lower void fraction peak
also decreases significantly. The PDF curve moves to the
higher void fraction values with the increase of the gas
superficial velocity. After the bend for the riser setup, with
the increase of USG from 0.544 to 0.945 m/ s, the stratified
wavy flow in the flowline becomes the developing slug
flow, which is featured by a small peak superimposed on a
big peak with a wide base. With the increase of USG from
0.544 m/ s the waves becomes stronger and as
consequence more bubbles are trapped inside. At a certain
point the Taylor bubbles are formed. For the horizontal
setup, no significant difference is present in the PDF
between before and after the bend except the significant


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

reduction in the height of the peak at the high void fraction
(-0.72).

At 2.363 m/ s, the PDF of void fraction for the riser has a
single peak at about 0.76 with broadened tails down to 0.2
and 0.9. This is the typical feature of chur flow. An
increase of the gas superficial velocity escalates the
instability of liquid slugs. When the speed of gas core
reaches to a certain point, the liquid slugs will be
penetrated and the integrity of the Taylor bubbles will be
destroyed. This leads to the transition to chum flow. For
the horizontal setup, released liquid from the collapsed
slugs accumulates on the bottom part of the pipe and
strong wave caused by the gas travel on the upper part of
the pipe. Stratified wavy flow is formed. After the bend,
the PDF of the void fractions for the riser significantly
shifts to the higher values compared with that before the
bend. The dominant void fraction from 0.75 increases to
0.83 and its count increases more than 40%. The chum
flow before the bend changes to stratified wavy after the
bend. This has been confirmed by visualization analysis.
The liquid film, however, is erratic and showed large
disturbances with void fractions down to about 0.7. For
the horizontal arrangement, no significant flow pattern
change was observed. It keeps the same stratified wavy
after the bend.


At the highest gas flow rate of 4.727 m/ s we examined, the
PDF of void fractions for the riser has very similar shape
to that at 2.363 m/ s but with much narrower tails. The
peak moves towards the higher void fraction. The flow
pattern is still the chur flow, but close to the transition to
the semi-annular flow. For the horizontal layout, with the
increase of gas flow rate from 2.363 to 4.727 m/ s, the
liquid film becomes more and more uniformly distributed
around pipe wall. This is reflected on the PDF curve where
the peak becomes narrower and exhibits less fluctuation.
The annular flow regime is approached when the gas
superficial velocity is big enough to distribute evenly the
liquid on the pipe wall. After the bend, semi-annular flow
is present in the flowline for the riser and stratified wavy
appears in the flowline for the horizontal arrangements.

Flow patterns using high speed video images
A. Flow regimes in vertical riser (vertical 90
degree bend)

Given below are descriptions of the various flow regimes
in the other that they might be expected to occur for
constant liquid flow rate and increasing air flow rate:

Spherical cap bubble: here as shown in Figure 5, there is
an increase in bubble coalescence and agglomeration. The
bubbles are now forming larger ones, but not big enough to
cover the pipe diameter. The velocity of a bubble may
differ substantially from that of the liquid phase.






Paper No


Figure 5: Video image of spherical cap bubble for a riser
at liquid superficial velocity = 0.142 m/ s and gas
superficial velocity = 0.047 m/ s.


Slug flow: the gas is observed to flow as large bullet
shaped bubbles separated by large silicone oil droplets with
some gas entrained in it. Figure 6 show that part of the
liquid from the Taylor bubble travel downwards as a film
on the pipe axis as thin lines.


Figure 6: Video image of slug flow for a riser at liquid
superficial velocity = 0.142 m/ s and gas superficial
velocity = 0.544 m/ s.

Unstable slug: the flow pattern shown in Figure 7 obtained
at higher gas flow rates represents the transition to chur
flow. An increase in gas coalescence in the liquid slug as a
consequence of an increase in gas flow rate brought about
oscillating of the liquid slug, thereby causing the liquid
slug to begin to collapse.


Figure 7: Video image of unstable slug flow at liquid
superficial velocity = 0.142 m/ s and gas superficial
velocity = 0.945 m/ s.


Figure 8: Video image of chur flow for a riser at liquid
superficial velocity = 0.142 m/ s and gas superficial
velocity = 2.363 m/ s.

B. Flow regimes in bend (vertical 90 degree bend)
The two dominant factors governing the flow structure of a
two-phase flow mixture in the 90 degree bends were the
flow regime as the mixture entered the bends and the
interaction of gravitational and centrifugal forces.

Spherical cap bubble: the flow regime of the mixture
entering the bend is spherical cap bubble with the size of
the bubbles almost occupying the entire cross-section of
the pipe. On entering the bend, the bubbles migrate to and
follow the inside of the bend while the liquid to the outside
of the bend. The liquid on striking the wall creates
secondary flow. When the secondary flow meets with the
primary (incoming) flow, bigger bubbles are created as a
consequence of high level of mixing. The created bubbles
as a result of low surface tension forces collapse almost
immediately. The gravity forces then takes the liquid to the
inside of the bend while the gas to the outside of the bend
as shown in Figure 9.


Figure 9: Video image of spherical cap bubble flow
passing through a vertical 90 degree bend at liquid
superficial velocity = 0.142 m/ s and gas superficial
velocity = 0.047 m/ s.

Slug and unstable flows: the centrifugal force takes the
Taylor bubble to the inside of the bend whilst the liquid to
the outside of the bend. The gravitational force then takes
the liquid to the inside of the bend while the Taylor


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

Chur flow: as the gas flow rate increases further as shown
in Figure 8, the unstable slug flow regime seizes to exist as
a result of breakdown of all of the liquid slugs. The
breakdown slugs are now distributed in the form of waves
on an annular film.





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


bubbles to the outside of it as shown in Figure 10a. The
interesting thing that happened here is that as the Taylor
bubble is moving up due to buoyancy and the liquid
downwards due to gravity forces, at equilibrium, the Taylor
bubble collapsed completely as depicted in Figure 10b.
The collapsed bubble created a dry patch occupying almost
100% of the bend. This region which is shown in Figures
10c and 10d is wetted periodically but can endure for a
reasonably long period of time. The bubble collapse could
be due to gravity forces overcoming the low surface
tension force of silicone oil. Liquid moving with high
momentum on the other hand, move in forward direction
while the ones with the lesser momentum moves
backward.


(a) (b)


Figure 10: Video image of stable slug flow passing
through a vertical 90 degree bend at liquid superficial
velocity = 0.142 m/ s and gas superficial velocity = 0.047
m/s.

Chum flow: the same mechanism could be taking place for
both the slug and chum flows. However, not all of the
bubbles collapsed, the bigger bubbles remained at the
centre of the bend while the smaller ones spread in all
directions as shown in Figure lla. Also, there was no dry
patch observed in this regime. Some of the bigger bubbles
are sent inside the bend by the centrifugal forces. As the
bubbles strike the bottom of the bend due centrifugal
forces, they collapse and drain downwards. This is
reflected in Figure 1 lb


Figure 11: Video image of chum flow passing through a
vertical 90 degree bend at liquid superficial velocity =
0.142 m/ s and gas superficial velocity = 0.047 m/ s.


C. Flow regimes in flowline (horizontal 90 degree
bend)
For flowline (horizontal) up flow, a different form of
classification had to be established; because gravity
introduces an asymmetry into the flow: the density
difference between both phases causes the liquid to travel
preferentially along the bottom of the pipe. These flow
regimes are described below also in order of increasing gas
flow rate.

Plug flow: gravity effect cause the gas plugs to move along
the top of the pipe. The level of liquid in the pipe is full to
almost the diameter of the pipe.

Slug flow: this regime is characterized by the intermittent
appearance of slugs of liquid which bridge the entire pipe
section and move at almost the gas velocity. Lots of
pressure fluctuations typify this flow regime observed at
much higher gas flow rates; where the gas pressure behind
the slugs is greater than in front of the slug. The length of
the Taylor bubble decreases with an increase in gas flow
rate whilst maintaining a constant liquid flow rate.

Stratified wavy flow: at higher gas velocities, the shearing
action of the gas at the interface generates large amplitude
waves on the liquid surface. Liquid is torn from the surface
of these waves giving rise to droplet entrainment in the gas
core. The deposition of these drops partially wet the top of
the pipe.

D. Flow regimes in bend (horizontal 90 degree
bend)
Plug flow: the gas plugs on entering the bend initially
migrate towards the inner radius of the bend under the
influence of centrifugal force, but subsequently it is forced
outward due to the increasing influence of gravity.

Slug flow: at higher gas flow rate, the centrifugal force
moves the liquid to the outside of the bend while the gas to
the inside of the bend. The gravity force then takes the
liquid to the inside of the bend and the gas to the outside of
the bend. The Taylor bubble then collapsed almost
immediately, leaving behind dry patches in the bend. This
same behaviour was observed for the slug flow in the
vertical 900 bend.


Paper No






Paper No


Stratified wavy flow: the level of liquid in the pipe has
dropped to less half the diameter of the pipe with an
increase in gas flow rate leading to collapse of the Taylor
bubble.

Competition between centrifugal and gravitational
forces
This section will aim to study the effect of bend on flow
separation. It will be based on a modified Froude number
proposed by Gardner and Neller (1969) that when Fr = 1,
there will be stratified flow after the bend. Also that when
Fr is greater than 1, air will move to the inside of the bend,
whilst when Fr is less than 1, air will move to the outside
of the bend.

Table 3: Comparison between present work and the works
of Gardner and Neller (1969) based on modified Froude
number.

Work Before the Fr, After the
bend bend
Gardner and Bubbly/ slug 1.00 Stratified
Neller flow flow
(1969)
Present work Bubbly flow 0.997 Stratified
wavy flow


It can be observed from Table 3 that the value of Fr from
the present study is 0.997 (= 1). The flow regime before
and after the bend are bubbly and stratified wavy flows
respectively whilst for Gardner and Neller, bubbly and
stratified flows are observed before and after the bend
respectively. This discrepancy may be due to the fact that
the viscosity of silicone oil is about 5 times that of water
and with a surface tension about 2/ 7 times that of water.


Conclusions

Interrogating the effect of bends on two-phase gas-liquid
flow has been successfully investigated using advanced
instrumentation, ECT and WMS. The WMS has 24 X 24
steel wires with 0.12 mm diameter and measures the
cross-sectional void fractions at 440 locations. The data
were taken at an acquisition frequency of 200 and 1000 Hz
for the ECT and WMS respectively over an interval of 60
seconds. The characteristic signatures of Probability
Density Function derived from the time series of
cross-sectionally averaged void fraction data were used to
identify the flow patterns. The results were validated by the
analysis of the recorded high speed video images. The
examined ranges of USG and USL are 0.047 to 4.727 m/ s
and 0.142 to 0.378 m/ s respectively. We have found that:

1) The ECT and WMS predicted same flow pattern
signatures.
2) With an increase of USG from 0.047 to 4.727 m/ s,
spherical cap bubble, slug, unstable slug and churn
flows were observed in the vertical riser while in the


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

horizontal flowline, plug, slug and stratified wavy
flows were developed. Certainly buoyancy force plays
an important role in the formation of the different flow
patterns.
3) Bends have significant effect on the gas-liquid flow. In
both the vertical and horizontal 90 degree bends
gravitational force tends to move the liquid to the
inside of the bend while the gas to the outside of the
bend. Some big spherical cap bubble and Taylor
bubbles break up in the bends due to the balance of the
centrifugal force and the surface tension. The bubbles
become more uniform. Dryness patch in the bend was
observed in the slug and unstable slug flows. As a
result, after the vertical bend the spherical cap bubble
flow became bubbly flow, and the chur flow turn to
stratified wavy and semi-annular flows. The horizontal
bend has less effect on the flow patterns compared
with the vertical bend.
4) Gardner and Neller (1969) proposed a modified
Froude number (Fre =Um2/ Rgsine =1) condition for
stratified flow after the bend. However, the result
obtained in the present study proved otherwise. The
result obtained was stratified wavy. This result is to be
expected in view of the fact that the viscosity of
silicone oil is about 5 times that of water and with a
surface tension about 2/ 7 times that of water.

Acknowledgements

M. Abdulkadir would like to express sincere appreciation
to the Nigerian government through the Petroleum
Technology Development Fund (PTDF) for providing the
funding for his doctoral studies.
Donglin Zhao and Safa Sharaf are funded by EPSRC under
grant EP/ F016050/ 1.
This work has been undertaken within the Joint Project on
Transient Multiphase Flows and Flow Assurance. The
Author(s) wish to acknowledge the contributions made to
this project by the UK Engineering and Physical Sciences
Research Council (EPSRC) and the following: GL
Industrial Services; BP Exploration; CD-adapco; Chevron;
ConocoPhillips; ENI; ExxonMobil; FEESA; IFP; Institutt
for Energiteknikk; PDVSA (INTEVEP); Petrobras;
PETRONAS; SPT; Shell; SINTEF; Statoil and TOTAL.
The Author(s) wish to express their sincere gratitude for
this support.

References

Abdulkadir, M., Hernandez-Perez, V, Sharaf, S., Lowndes,
I. S. & Azzopardi, B. J., Phase distributions of an
air-silicone oil mixture in a vertical riser, HEFAT 2010, 7t
International Conference on Heat Transfer, Fluid
Mechanics and Thermodynamics, Antalya, Turkey 2010.
Azzi, A. & Friedel, L. Two-Phase Upward Flow 90 Degree
Bend Pressure Loss Model, Forschung im Ingenieurwasen
Vol. 69, 120- 130 2005.
Azzi, A., Friedel, L., Kibboua, R. & Shannak, B.
Reproductive Accuracy of Two-Phase Flow Pressure
Loss Correlations for Vertical 90 Degree Bends, Forschung
im Ingenieurwasen Vol. 67, 109- 116 2002.






Paper No


Azzopardi, B. J. Drops in annular two-phase flow, Int. J.
Multiphase Flow, Vol. 23, S1-S53. 1997.
Azzopardi, B. J., Hernandez Perez, V, Kaji, R., Da Silva,
M. J., Beyer, M. & Hampel, U. Wire mesh sensor studies
in a vertical pipe, HEAT 2008, Fifth International
Conference on Multiphase Systems, Bialystok, Poland.
2008.
Azzopardi, B.J., Jackson, K., Robinson, J.P., Kaji, R.,
Byars, M. & Hunt, A. Fluctuations in dense phase
pneumatic conveying of pulverised coal measured using
electrical capacitance tomography. Chem. Eng. Sci., Vol.
63, 2548-2558 (2008).
Benbella, S., Al-Shannag, M. & Al-Anber, Z. A. Gas-liquid
pressure drop in vertically wavy 90 degree bend.
Experimental Thermal and Fluid Science 2008.
Carver, M.B., Numerical computation of phase separation
in two fluid flow, ASME Paper No. 82-FE-2, Vol. 106/ 153
1984.
Carver, M.B, & Salcudean, M., Three-dimensional
numerical modelling of phase distribution of two- fluid
flow in elbows and return bends, Numerical Heat Transfer,
Vol. 10, 229-251 (1986).
Costigan, G & Whalley, P.B. Slug flow regime
identification from dynamic void fraction measurements in
vertical air-water flows, Int. J. Multiphase Flow, Vol. 23,
263-282 (1996).
Da Silva, M.J., Thiele, S., Abdulkareem, L., Azzopardi, B.J.
& Hampel, U. High-resolution gas-oil two-phase flow
visualization with a capacitance wire-mesh sensor.. Flow
Measurement and Instrumentation, (2010Accepted


Dean, W.R. Note on the motion of a fluid in a curved pipe.
Phil. Mag., Vol. 4, 208-223(1927).
Dean, W.R. Stream- line motion of a fluid in a curved pipe.
Phil. Mag., Vol. 5, 673-695 (1928).
Dewhurst, S. J., Martin, S.R., Jayanti, S., & Costigan, G,
Flow measurements using 3-D LDA system in a square
section 90 degree bend, Report AEA-In Tech-0078 (1990).
Ellul, I.R., & Issa, R.I. Prediction of the flow of
interspersed gas and liquid phases through pipe bends.
Trans. Instn. Chem. Engrs, Vol. 65, 84-96 (1987).
Eustice, J. Flow of water in curved pipes. Proc. R. Soc.,
A84, 107-118 (1910).
Gardner, G.C. & Neller, PH. Phase distributions flow of
an air-water mixture round bends and past obstructions,
Proc. Inst. Mech. Engr., Vol. 184, 93 -101 (1969).
Geraci, G, Azzopardi, B.J. & Van Maanen, H.R.E.
Inclination effects on circumferential film distribution in
annular gas/ liquid flows. AIChE Journal, Vol. 53.
1144-1150 (21"1 I-Li
Geraci, G, Azzopardi, B.J. & Van Maanen, H.R.E. Effects
of inclination on circumferential film thickness variation in
annular gas/ liquid flows. Chem. Eng. Sci., Vol. 62,
3032-3042 (2007b).
Hammer, E.A. Three-component flow measurement in oil/


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

gas/ water mixtures using capacitance transducers, PhD
thesis, University of Manchester (1983).
Hernandez-Perez, V. Gas-liquid two-phase flow in inclined
pipes. PhD Thesis, University of Nottingham, United
Kingdom (2007).
Huang, S.M.. Impedance sensors-dielectric systems. In R.
A. Williams, and M. S. Beck (Eds.), Process Tomography,
Cornwall: Butterworth-Heinemann Ltd. (1995).
Jayanti, S. Contribution to the study of non-axisymmetric
flows. PhD Thesis, Imperial College London (1990).
Legius, H.J.W.M. & van den Akker, H.E.A. Numerical and
experimental analysis of translational gas-liquid pipe flow
through a vertical bend. Proceedings of the 8t
International Conference, BHR Group, Cannes, France
(1997).
Manera, A., Ozar, B., Paranjape, S., Ishii, M. & Prasser,
H.-M. Comparison between wire-mesh-sensors and
conducting needle-probes for measurements of two-phase
flow parameters. Nucl. Eng. Des. Vol. 239, 1718-1724
(2008).
Omebere-Iyari, N.K. & Azzopardi, B.J. A study of flow
patterns for gas/ liquid flow in small diameter tubes. Chem.
Eng. Res. Design, Vol. 85, 180-192 (2007).
Spedding, PL. & Benard, E. Gas-liquid two-phase flow
through a vertical 90 degree elbow bend, Exp. Thermal
Fluid Sci., Vol. 31, 761-769 (2006).
Spedding, PL. & Benard, E. & McNally, GM.. Fluid Flow
through 90 Degree Bends, Dev. Chem. Eng. Min. Process
Vol. 12, 107-128 (21.14).
Thiele, S., Da Silva, M.J., Hampel, U., Abdulkareem, L. &
Azzopardi, B.J. High-resolution oil-gas two-phase flow
measurement with a new capacitance wire-mesh
tomography. 5th International Symposium on Process
Tomography in Poland, Zakopane, 25-26 August (2008).
Zhu, K., Madhusudana Rao, S., Wang, C. & Sundaresan, S.
Electrical capacitance tomography measurements on
vertical and inclined pneumatic conveying of granular
solids. Chem. Eng. Sci. Vol. 58, 4225-4245 (2003).




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