Isı Bilimi ve Tekniği Dergisi, 35, 1, 1-9, 2015

J. of Thermal Science and Technology

©2015 TIBTD Printed in Turkey
ISSN 1300-3615

HYDRODYNAMIC CHARACTERISTICS OF TWO-PHASE FLOW THROUGH

HORIZONTAL PIPE HAVING SMOOTH EXPANSION
Emrah DENİZ* and Nurdil ESKİN**
*Istanbul Technical University, Faculty of Mechanical Engineering
34437 Beyoglu, Istanbul, Turkey, [email protected]
**Istanbul Technical University, Faculty of Mechanical Engineering
34437 Beyoglu, Istanbul, Turkey, [email protected]

(Geliş Tarihi: 23.09.2013, Kabul Tarihi: 18.02.2014)

Abstract: In this study, characteristics of adiabatic two-phase flow through horizontal pipe having smooth expansion are
investigated numerically and experimentally. Effects of volumetric void fraction and internal diameter of the pipe on
hydrodynamic characteristics of the two-phase flow are examined using average diameter of the bubbles injected. Air and
water are chosen as the fluid couple. Flow rate for water is kept constant at 3 lt/s while those for air are taken as 50 and 61
lt/min. Thus, volumetric void fraction of 21.74 and 25.31 % are obtained for the two-phase flow considered, respectively.
Eulerian-Eulerian Model and Reynolds Stress Model (RSM) are employed for modeling of two-phase flow and turbulence.
Numerical results are then compared to the experimental data that was previously obtained by dual optical probe that
measures local parameters of the two-phase flow (Deniz, 2009). The comparison shows that using the assumption of the
average bubble diameter for numerical modeling gives reasonable results for developed and stratified flow.
Keywords: Two-phase flow, Numerical simulation, Dual optical probe, Smooth expansion, Void fraction, Pressure drop

TEDRİCİ GENİŞLEMEYE SAHİP YATAY BORU BOYUNCA İKİ FAZLI AKIŞIN

HİDRODİNAMİK KARAKTERİSTİĞİ
Özet: Bu çalışmada, tedrici genişlemeye sahip yatay kanaldan geçen adyabatik iki fazlı akışın karakteristikleri sayısal ve
deneysel olarak incelenmiştir. Hacimsel boşluk oranı ve boru iç çapının iki fazlı akış karakteristiğine etkisi, enjekte edilen
kabarcıkların ortalama çapı kullanılarak irdelenmiştir. Akış çifti olarak hava ve su seçilmiştir. Hacimsel debi değerleri, su
için 3 lt/s; hava için 50 ve 61 lt/dak olarak alınmıştır. Böylece incelenen iki fazlı akıştaki hacimsel boşluk oranı değerleri
sırasıyla % 21.74 ve % 25.31 olarak elde edilmiştir. İki fazlı akışın ve türbülansın modellenmesinde Euler-Euler Modeli ve
Reynolds Gerilim Modeli kullanılmıştır. Sayısal sonuçlar, iki fazlı akışın yerel parametrelerini ölçen ikili optik prop
kullanarak, daha önceden elde edilen deneysel sonuçlarla karşılaştırılmıştır (Deniz, 2009). Karılaştırma sonucunda, sayısal
modellemede kullanılan ortalama kabarcık çapı yaklaşımının gelişmiş ve katmanlaşmış akış için kabul edilebilir sonuçlar
verdiği bulunmuştur.
Anahtar Kelimeler: İki fazlı akış, Sayısal modelleme, İkili optik prop, Tedrici genişleme, Boşluk oranı, Basınç düşüşü

NOMENCLATURE Greek Symbols

A Area [m2] α Void fraction [Ag/Atotal]

D Diameter [m] β Volumetric void fraction [ Qgas / Qtotal ]
f Friction factor
ε Turbulence dissipation [m2/s3]
s Mass flux [kg/m2s]
ε Surface roughness [m]
g Gravitational acceleration [m/s2, cm/s2]
k Turbulence kinetic energy [m2/s2] µ Dynamic viscosity [Pa.s]
L Length [m] υ,v Kinematic viscosity [cm3/s, m2/s]
ρ Density [kg/m3]
N Number of the holes in the injector
Φ Two phase flow factor
P Pressure [Pa]
X Lockhart-Martinelli factor
S Slip ratio [UG/UL]
u,U Velocity [m/s]
V Volume [cm3] Subscripts
x Quality[mpas/mtotal]
l Liquid phase
y Position on the y-axis [m]
g Gas phase
Volumetric flow rate [m3/s]
INTRODUCTION with the data that was previously obtained by using dual
optical probe (Deniz, 2009). Construction of the
Gas-liquid two-phase flow takes the leading role in numerical domain and the analysis are performed in
applications including evaporation, condensation and flow GAMBIT (v. 2.3.16) and ANSYS FLUENT (v. 12),
through cross-sectional area change (i.e. nozzle, diffuser) respectively. For simulations, the average diameter of the
and sudden pressure drop (i.e. flashing). These systems bubbles injected to the flow is calculated via a correlation
normally have complex geometries composed by proposed by Kunii (1991). Referring to the comparison
singularities like expansion, contraction, bends, orifices, with the experimental results, it can be concluded that the
etc. Thus, two-phase flow characteristics, such as presented model based on average bubble diameter can
distribution of local void fraction and phase velocities, describe the local values and the flow pattern reasonably.
pressure drop and flow pattern of aforementioned
singularities should be identified for design of such EXPERIMENTAL SETUP
systems. However, there is limited research documented
in this subject. Theoretical analyses of air-water flow The schematic representation of the experimental
through straight horizontal and/or vertical pipe and facility is given in Figure 1. Water and air were chosen
comparison of results with experiments were performed as the fluid couple. Water at atmospheric conditions was
by Ekambara et al. (2008), Ghorai et al. (2006), Morel et taken from the tank by a pump and sent to the pipe
al. (2010) and Chahed et al. (2003). The effects of phase attached to an electro-magnetic type flow meter in order
velocities, volume fraction of gas and interfacial to measure its volumetric flow rate, .
roughness on two-phase flow characteristics were Meanwhile air at the atmospheric conditions was
investigated in these studies. Winterton et al. (2001), compressed by the compressor and delivered to an
Ahmed et al. (2008), Aloui et al. (1999) and Balakhrisna immersed injector where the two-phase flow was
et al. (2010), researched two-phase flow through straight obtained. It had 28 holes with 1 mm diameter positioned
pipes with or without sudden expansion, while Fossa and on four rods inserted through the cross-section of the
Guglielmini (1998) and Bertola (2004) worked on pies upstream pipe inlet. Air-water mixture then flowed
with sudden contraction, experimentally. They used through the test section where the measurements were
instruments such as electrical impedance probe (Fossa performed by optical probe. Accuracy of the probe is
and Guglielmini 1998), hot-film anemometer (Ahmed et given as ± 7 % by the manufacturer company, RBI
al. 2008), fiber optical probe (Bertola 2004), and CCD (France). Measurement principles of the optical probe
camera (Winterton et al. 2001) for measurements. can be found in François et al (2003). All pipes in the
Winterton et al. (2001), Aloui et al. (1999) and Fossa and facility are made of transparent acrylic in order to
Guglielmini (1998) performed local measurements observe the flow. Picture of the injector and schematic
whereas Balakhrisna et al. (2010) measured average. of the test section are shown in Figure 2.
Liquid couples that were studied were air-oil (Ahmed et
al. 2008), water-air (Ekambara et al. 2008, Ghorai et al.
2006, Morel et al. 2010, Chahed et al. 2003, Winterton et
al. 2001, Aloui et al. 1999, Fossa and Guglielmini 1998,
Bertola 2004) and oil-water (Balakhrisna et al. 2010).

As summarized above, it can be seen in literature that the

numerical and/or experimental studies investigating the
gas-liquid flow characteristics generally contain flow
through straight pipes or pipes with abrupt changes.
There are only few studies considering the progressive
cross-sectional changes. Additionally, except the ones
listed above, it can be stated that the majority of the
experimental studies published in the literature consists of
averaged values of two-phase flow parameters. One also
can note that many theoretical studies are validated by Figure 1. Schematic of the experimental facility
referring to the averaged experimental values although
they are constructed on the separated flow model that The test section consisted of three pipes called upstream
provides the local values. Therefore, the theoretical pipe, singularity pipe and downstream pipe with
investigation of two-phase flow through singular prescribed length and diameters. Inner diameter of the
geometries and its validation with local measurements is upstream channel (i.e. 32 mm) enlarged to that of the
a gap in literature that needs to be addressed. downstream channel (i.e. 40 mm) smoothly, with an
expansion angle of 9°. Area ratio for the singularity was
In this study, characteristics of adiabatic air-water two- calculated to be 0.64. In the measurements, flow rate of
phase flow through horizontal pipe with progressive water was taken as 3 l/s, while that of air were 50 and
expansion are investigated. Results of the theoretical 61 l/min. Thus, volumetric void fractions of 21.74 and
analysis based on Eulerian-Eulerian model are compared 25.31 % were achieved for the flow, respectively.

2
 ul  ul  ul (1)

ug  ug  ug (2)

l  (1   )ul A  0 (3)

 g   ug A  0 (4)

l ul (ul (1   ))  P(1   )  l g sin  (1   )  F1 (5)

 g ug (ug )  P( )   g g sin  ( )  F2 (6)

For modeling of turbulence, Reynolds Stress Model

(RSM) is utilized in the simulations. The exact
equations used in the model are given below.

Duiu j   u u  
Dt

X k 
p

 
 uiu j uk   jk ui   ik u j   i j
X k


 
 u j ui  ui u j
  uiu j  u j uk   2 (7)
 X k X k  X k X k
Figure 2. Test section and the injector used in the
experimental study p  u  u j 
  i  
As shown in Figure 2, there are four locations where   X j X i 

experimental and numerical results are compared, two at
  
the inlet and outlet of the singularity pipe, and the other
Dk  p dk ui
two at 10  D1 (i.e. 320 mm) before and after the former   uk  k        uiuk  (8)
Dt X k     dX  X k
ones. For clarity, these locations of these positions will  k 
be defined in terms of γ, throughout the text, γ being the
distance from the injector divided by the total length of D   2 uk p  
the test section. Thus, from first location to the forth, γ    uk   
Dt X k   X X Xk 
takes the values of 0.4, 0.58, 0.82 and 1, respectively.  j j 
The geometrical details of the test section and the ui  uk uk u uj  u  2ui
operational conditions are given in Table 1. 2   i   2 ul i (9)

X j  X i X j X k X k  X j X k X j

Table 1. Operating conditions of the experimental study 2
u  ui u j   2 ui 
L1 = 1000 mm 2 i  2 
L2 = 320 mm X j X k X k  X k X j 
 
Geormetry L3 = 410 mm
D1 = 32 mm Numerical Analysis
D2 = 40 mm
Gas phase/Liquid Air/water at atmospheric The numerical analysis of the flow under consideration is
phase conditions performed based on the parameters and the geometrical
Qair Qwater 50 and 61 [l/min] / 3 [l/s] restrictions mentioned in the experimental setup section
/
 of this paper. Since characteristics of the bubbles in two-
21.74 % and 25.31 %
phase flow are determined by the design parameters of
the injector utilized in the experimental study (Figure 2.),
THEORETICAL MODEL the numerical analysis in the present study is performed
in two steps including the simulation of the flow through
Governing Equations the injector, and then through the pipes.
In this study, the simulation of air-water flow through Modeling of the flow through the injector
the test section is performed by two-fluid Eulerian-
Eulerian model at steady-state condition. Considering The numerical geometry constructed for the injector is
no mass transfer between the phases and no heat given in Figure 3. Here, the holes of the injector are
transfer from the surroundings to the flow or vice versa, represented by the means of rods of 32 mm length that
the governing equations of the model are given below. are inserted to a pipe, which was all created within 3D
domain. In order to ease the step of creating the structural

3
mesh, the rods are constructed with square cross-sections
having the same hydraulic diameter with that of the
circular holes, i.e. 1 mm. According to the simulation, air
flows through these rods while water enters the pipe from
the remaining cross-section at the inlet, and the two-phase
flow of interest is obtained just after the outlet of the rods
where the phases infiltrate each other. At the end of the
simulation for each air flow rate, the phase velocity, and
the void fraction profiles obtained at the downstream of
the rods are extracted from the simulation, and are
introduced as the inlet conditions for the analysis of the
flow through the pipes. Figure 4. y-directional node densities for the tested grids
having 83,817 (left) and 185,513 (right) nodes

Figure 3. Numerical geometry of the injector (left) and mesh

structure at water-inlet

370,000 nodes are considered adequate for the injector

simulation according to the grid independency
examination performed by testing several domains with
Figure 5. Grid independency test based on void fraction
different node densities. The mesh structure at water-inlet distribution
is also shown in Figure 3. In ANSYS FLUENT (v.12),
the boundary condition “velocity inlet” is taken as the In order to determine the turbulence model to be used,
inlet condition for water and air, while the boundary flow through the upstream pipe is modeled via Standard
conditions “interior” and “outflow” are employed as the k-ε and Reynolds Stress Model (RSM), and is compared
outlet condition for the rods and the pipe, respectively. It with the related experimental results illustrated in Figure
was only needed to model the half of the injector by using 6. According to the comparison, RSM turbulence model
the symmetry condition. Eulerian multiphase model is is chosen for the numerical simulation of the flow
used for the analysis. Reynolds Stress Model (RSM) is considered. As in the injector simulation, Eulerian
employed to model the turbulent flow. Standard Wall multiphase model is used for the analysis. Standard Wall
Functions are utilized for near wall treatment. Phase Functions are utilized for near wall treatment. Phase
Coupled SIMPLE scheme for pressure-velocity coupling, Coupled SIMPLE scheme for pressure-velocity coupling,
Green-Gauss Cell Based option for gradients, and Second Green-Gauss Cell Based option for gradients, and Second
Order Upwind for spatial discretization are chosen. Order Upwind for spatial discretization are chosen.
Modeling of the flow through the pipes

The second step of the analysis is modeling of the two-

phase flow through the test section having three regions,
i.e. upstream, through and downstream of the singularity.
The numerical domain of the system is constructed in 3D,
and is symmetrical with respect to y-axis. 83,817 nodes
are determined to be adequate for upstream and
downstream pipes, while twice the nodes are used for the
singularity section, according to the grid independency
study. In each region, the axial length in z-direction is
divided into 400 pieces. For the upstream pipe, the node
densities in y-direction for each grid of the independency
study, and the corresponding void fraction distributions
are given in Figure 4 and Figure 5.
Figure 6. Comparison of turbulence models with the
experimental data

4
Modeling of the bubbles location (i.e. γ = 0.82) since the velocity of the phases
decreases with an increase in the cross-section of the pipe.
Due to momentum of the flow and forces between the
phases, the injected bubbles break up or coalesce,  = 0.58
causing a variation in bubble diameter along the flow. In   21.74 %   25.31 %
this study, it is assumed that bubbles are spherical and
that they do not interact with each other. By using the 1
0.9
former assumption, an average diameter value for the
0.8
bubbles at the injector outlet is extracted from the 0.7
volume estimated via the correlation written in literature 0.6
(Kunii 1991) and is given below. y/D 0.5
0.4

 / N 
6/5 0.3

Vbubble  1.138 (10) 0.2

g 3/ 5 0.1
0
0 10 20 30 40 50 60 70 80 90 100
RESULTS AND DISCUSSIONS
 %

Experimental Results Figure 8. Effect of volumetric void fraction at   0.58

The experimental results for the radial distribution of the

The flow with lower volumetric void fraction decelerates
local void fraction along the upstream and downstream of
more than that with the higher one. This effect becomes
the smooth expansion are plotted below. In the following
significant at top of the channel, where most of the
figures, y / D stands for the non-dimensional position and bubbles gather. Thus, local void fraction values for
is defined as the position on the y-axis divided by the   21.74 % are greater than those for   25.31 % a
channel diameter. Thus, y / D varies from zero to the can be observed in Figure 9. The discrepancy decreases
unity representing the bottom and top of the channel, as the flow develops through the downstream pipe as
respectively. shown in Figure 10.

The experimental values obtained at upstream of the

smooth expansion are given in Figure 7 and Figure 8.
From the figures, it can be seen that the local void
fraction increases with increasing y / D value due to the
density difference between air and water, and the effect of
gravitational force. As the flow moves along the upstream
pipe (towards γ = 0.58), the effect of volumetric void
fraction becomes obvious at the top of the channel due to
the development of stratification. As expected, local void
fraction rises with increasing volumetric void fraction.

 = 0.4

  21.74 %   25.31 %

1
0.9 Figure 9. Effect of volumetric void fraction at   0.82
0.8
0.7
0.6
y/D 0.5
0.4
0.3
0.2
0.1
0
0 10 20 30 40 50 60 70 80 90 100
 %

Figure 7. Effect of volumetric void fraction at   0.4

The radial distribution of local void fraction at

downstream of the smooth expansion is illustrated in
Figure 9 and Figure 10. Once compared with the values
obtained upstream of the expansion, it can be stated that
Figure 10. Effect of volumetric void fraction at   1
the local void fraction values increase after the singularity

5
Numerical Results the singularity outlet (   0.7 ). Then a new
stratification begins to develop in the downstream pipe.
The numerical values of two-phase flow parameters
along the radial direction of the pipes and their
comparison with the experimental results are given in
the following sections.

Radial Distribution of Void Fraction

The numerical and experimental local void fraction

distributions for   25.31 % at   0.4 and   0.58
are shown in Figure 11 and Figure 12, respectively.

Figure 12. Local void fraction distribution at   0.58

Figure 11. Local void fraction distribution at   0.4

From the figures, it can be seen that the numerical and

experimental values are comparable and are getting closer
to each other as the flow develops. In a developing
bubbly flow, the diameter of the bubbles are affected by
the interactions (i.e. break-up, coalescence) and the forces
(lift and drag) more than it does in a developed flow, and Figure 13. Numerical local void fraction distribution along the
therefore more variation would be observed. Also, using singularity pipe for   21.74 %
the optical probe for experiments is an intrusive method
that causes an error of ± 7 % in the measurements due to The numerical results for the downstream pipe are
its blockage effect, contributing to the difference the compared with the experimental results and the
numerical values. Referring to the explanations written comparison is shown in Figure 14 and Figure 15.
above and to the comparison between the values plotted
in Figure 11 and Figure 12, it can be concluded that the
model based on the constant bubble diameter assumption
simulates the developed and the stratified two-phase flow
more successfully than that for the developing flow, for
the cases investigated.

The numerical local void fraction distribution through

the singularity pipe is plotted for   21.74 % in
Figure 13. Stratified water-air flow enters the singularity
pipe at   0.58 and begins to decelerate at singularity-
inlet (   0.69 ) due to the smoothly enlarging cross-
section. Since deceleration of water is greater than that
of air due to the higher density of water, local void
fraction values decrease, the stratification developed
along the upstream pipe collapses, and the local void
fraction distribution takes the form of a curve. The Figure 14. Local void fraction distribution at   0.82
curve gets narrower as the decelerating continues along

6
From Figure 14, it can be observed that the numerical already performed in the present study. The effects of
simulation and the experimental measurements give the singularity section and the volumetric void fraction
similar profiles for the local void fraction distribution at on the local slip ratio distribution are given in Figure 17.
downstream of the singularity. As the stratification
develops, numerical and experimental values get closer to
each other as seen in Figure 15. Furthermore, the
stratification in the lower half of the downstream channel
develops earlier in the numerical simulation as in the
upstream pipe because of the constant bubble diameter
assumption.

Figure 17. Distribution of numerical local slip ratio

before and after the singularity

Two-Phase Pressure Drop

Figure 15. Local void fraction distribution at   1 The numerical values for the pressure drop along the
upstream and the singularity pipes are compared with
Slip Ratio the results computed by the correlations existing in the
literature, in this section. In two-phase flow, pressure
Slip ratio is an important parameter in two-phase flow drop has three components, as written below.
since it determines the type of the model (i.e.
homogeneous or separated) that the flow characteristics Ptotal  Pstatic  Pmomentum  Pfriction (11)
can be analyzed with. Slip ratio variation with respect to
 for   21.74 % and   25.31 % is given in For the upstream pipe, it can be stated that the friction
Figure 16. pressure drop ( Pfriction ) is the dominant parameter since
the pipe is horizontal and the mass flow rate of the flow
through the pipe is constant. For the singularity pipe, in
addition to the frictional pressure drop, momentum
pressure drop is also important due to the deceleration
caused by the enlarging singularity. The correlations
used to evaluate the friction and the momentum pressure
drops are listed in below equations.

Pfriction  2L PL (12)

PL  4 f L  L / di  GL2 1/ 2L  (13)

f L  0.079 / Re0.25
L where Re  GL di / L (14)

2L  1  C X  1 X for ReL  4000

2
(15)
Figure 16. Variation of numerical slip ratio with respect to 
0.5 0.1
 1  x   G   G 
0.9

For both cases shown in the figure, the slip ratio takes X       (16)
the maximum value at the injector outlet due to the high  x   L   L 
velocity of the air during injection. As the flow
develops, S varies between 1.24 and 0.58. Thus, it can d  1  G 2 1 dA
Pmomentum  G 2   (17)
be concluded that the separated flow model must be dz  m  m A dz
used in order to analyze the flow characteristics, as is

7
Eq. 12 to Eq. 16 are proposed by Lockhart and acceptable for the case investigated. The numerical
Martinelli (1949). In Equ. 15, C depends on the flow results for the pressure drop through the singularity are
regimes of the phases and takes the values given in close to the estimations by Lockhart and Martinelli
Table 2. (1949) and Kourakos et al. (2009) within a range of
9-19 % and 14-37.5 %, respectively. It must be noted
Table 2. Values of C that the greater discrepancy stands for a flow with
Liquid Gas C higher volumetric void fraction for all cases mentioned.

Turbulent Turbulent 20
Laminar Turbulent 12
Turbulent Laminar 10
Laminar Laminar 5

The pressure drop through the smooth enlargement is

estimated with the correlation proposed by Kourakos et
al (2009).

Psin gularity  0.061  0.8917  10717  ReL0.8283

1  0.378
G12 2   
 1    1  x  L  1 (18)
2L   G  
Figure 19. Pressure drop for   25.31 %
The numerical and calculated values of the pressure
drop in the flow are given for   21.74 % and The effect of volumetric void fraction on two-phase
pressure drop is illustrated with the numerical values
  25.31 % in Figure 18 and Figure 19, respectively. plotted in Figure 20. As expected, the pressure drop
increases by increasing volumetric void fraction.

Figure 18. Pressure drop for   21.74 %

Figure 20. Pressure drop with respect to void fraction
Once compared with the results computed by Lockhart CONCLUSIONS
and Martinelli (1949), it can be stated that the numerical
model overestimates the pressure drop along the In this study, the characteristics of air-water flow
upstream pipe within a range between 21-25 %. The through the horizontal pipe with smooth expansion are
model utilized here is derived for fully developed, investigated numerically and experimentally. According
stratified flow, which does not take the entrance effects to the experimental results, it is concluded that the local
and developing flow conditions into consideration.
void fraction increases with increasing y / D value due
However, in the numerical study, air-water flow is
developing and stratifying along the upstream channel to the density difference between air and water, and the
beyond outlet of the injector. Besides, in the numerical effect of gravitational force. The local void fraction
study, the surface roughness for the pipes is not taken increases with increasing volumetric void fraction and
into consideration considering that the pipes used in the the effect of volumetric void fraction is more clearly
experimental study are made of acrylic and “roughness / observed at the top of the channel due to the
diameter” (  / D ) ratio is too small. Therefore, the error development of stratification. It is found that the
within a range in 21-25 % can be considered as numerical and experimental values are comparable and

8
are getting closer to each other as the flow develops.
The discrepancy mostly occurs at the bottom half of the Chahed J., Roig V., Masbernat L., 2003, Eulerian-
channel since the numerical stratification develops Eulerian Two-Fluid Model for Turbulent Gas-Liquid
earlier than the experimental stratification. Void fraction Bubble Flows, Int. J. Multiphase Flow, 29, 23-49.
distribution through the singularity pipe takes the shape
of a curve due to deceleration of the water is greater Deniz E.,2009, Experimental Study of Bubbly Flow in
than deceleration of the air. The slip ratio between the Singular Geometries (Research Project Report 2009-
phases varies between 1.24 and 0.58 in the flow though 09), VKI Library, Belgium.
the pipes. Thus, the separated flow model must be Ekambara K., Sanders R. S., Nandakumar K. and
employed in order to analyze the characteristics of the Masliyah J. H., 2008, CFD Simulation of Bubbly Two-
flow, as is performed in the present study. The Phase Flow in Horizontal Pipes, Chem. Eng. J., 144,
numerical model overestimates the pressure drop in the 277-288.
flow compared to the results calculated by the Fossa M. and Guglielmini G., 1998, Dynamic Void
correlations provided in literature (Lockhart and Fraction Measurements in Horizontal Ducts with
Martinelli 1949, Kourakos et al. 2009). Sudden Area Contraction, Int. J. Heat Mass Transfer,
41, 3807-3815.
ACKNOWLEDGEMENT
François F., Garnier J. and Cubizolles G., 2003, A New
We would like to express our appreciations to von Data Acquisition System for Binary Random Signal
Karman Institute for Fluid Dynamics, and Prof. J. M. Application in Multiphase Flow Measurements, Meas.
Buchlin, Assoc. Prof. Patrick Rambaud, and Dr. Vasilis Sci and Tech., 14, 929-942.
Kourakos for their support and cooperation during the
experimental study. Ghorai S. and Nigam K. D. P., 2006, CFD Modeling of
Flow Profiles and Interfacial Phenomena in Two-Phase
REFERENCES Flow in Pipes, Chem. Eng. Process, 45, 55-65.

Ahmed W. H., Ching C. Y. and Shoukri M., 2008. Kourakos V. G., Rambaud P., Chabane S. and Buchlin
Development of Two-Phase Flow Downstream of a J. M., 2009, Modeling of Pressure Drop in Two-Phase
Horizontal Sudden Expansion, Int. J. Heat Fluid Flow. Flow within Expansion Geometries, AIP Conf. Proc.,
29, 194-206. 1207, 802-80.

Aloui F., Doubliez L., Legrand J. and Souhar M., 1999, Kunii D., 1991, Fluidization Engineering, Butterworth-
Bubbly Low in an Axisymmetric Sudden Expansion: Heinemann, Boston.
Pressure Drop, Void Fraction, Wall Shear Stress,
Bubble Velocities and Sizes, Exp. Therm Fluid Sci., 19, Lockhart R. W. and Martinelli R. C., 1949, Chem. Eng.
118-130. Prog., 45, 39-48.

Balakhrisna T., Ghosh S., Das G. and Das P. K., 2010, Morel C., Ruyer P., Seiler N. and Laviéville J. M.,
Oil-Water Flows through Sudden Contraction and 2010, Comparison of Several Models for Multi-Size
Expansion in a Horizontal Pipe – Phase Distribution and Bubbly Flows on an Adiabatic Experiment, Int. J.
Pressure Drop, Int. J. Multiphase Flow, 36, 13-24. Multiphase Flow, 36, 25-39.

Bertola V.,2004, The Structure of Gas-Liquid Flow in a Winterton R. H. S. and Munaweera J. S., 2001, Bubble
Horizontal Pipe with Abrupt Area Contraction, Exp. Size in Two-Phase Gas-Liquid Bubbly Flow in Ducts,
Therm Fluid Sci., 28, 505-512. Chem. Eng. Process, 40, 437-447.