Experimental Thermal and Fluid Science 83 (2017) 69–77

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science

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

Experimental measurements of bubble convection models in two-phase

stratified liquids
Hassan Abdulmouti
Department of Mechanical Engineering Division, Sharjah Men’s College, Higher Colleges of Technology, P.O. Box 7946, Sharjah, United Arab Emirates

a r t i c l e i n f o a b s t r a c t

Article history: The surface flow generated by a bubble plume is a technique proposed to collect the surface-floating sub-
Received 29 July 2015 stances. This has a great importance when handling the oil layer formed during large oil spill accidents due
Received in revised form 9 December 2016 to the need to protect naval systems, rivers, and lakes. The motivation of this research is to broaden the
Accepted 11 December 2016
understanding of oil flow in the stratified layer of oil on the free surface. Laboratory experiments have been
Available online 16 December 2016
carried out in order to investigate the multi-dimensional motion of water and oil due to bubbles. The flow
structure of bubble-induced convection in a stratified liquid is investigated by using particle imaging
Keywords:
velocimetry (PIV) measurements and pathline measurements. It is confirmed by this paper that the flow
Multiphase flow
Bubble plume
structure is strongly modulated by the gas flow rate and bubble size. The velocity of the surface flow
Surface flow induced by the bubble plume in the vicinity of the oil-water interface is about 50% larger and stronger than
Flow visualization that inside the oil layer. Moreover, the surface flow is particularly rapidly generated in the vicinity of the
Stratified liquid oil-water interface. The highest kinetic energy (which is 0.012 (m/s)2) is generated at a far distance inside
Particle imaging velocimetry (PIV) the bubble plume and in the vicinity of the oil-water interface. This observation confirms the idea that the
Thermal stratified liquid bubble plume can indeed generate a strong and wide surface flow over the bubble generation system. The
oil layer is easily broken by bubbles, especially with high void fraction and small bubble size and high gas
flow rates. The experimental results were compared with the numerical results of our earlier paper. It is
confirmed by this paper that the experimental results resemble the numerical results. As a reference, in
a thermally stratified liquid, the liquid flow pattern of bubble-induced convection has also been measured
by using thermo-sensitive liquid crystal tracer particles in which a color-image processing is combined to
measure the temperature field. These results were compared to the immiscible stratified cases.
Ó 2016 Elsevier Inc. All rights reserved.

1. Introduction metal refining process, the reduction of surfactants in chemical

Bubble dynamics is of great importance in optimizing the engi-
neering design and operating parameters of various adsorptive
bubble system operations, such as dissolved air flotation and dis-
persed air flotation, for the separation of solids from a liquid. Bub-
ble rise in liquids under different conditions is one of the oldest of
scientific investigations [40]. It is of interest as the characteristics
of a rising bubble (e.g. size, shape, velocity and trajectory) give
insight into the dynamics of a system [12]. Rise velocity is one of
the most important characteristics of a rising bubble. Numerous
studies have been performed on the motion of single bubbles
[26,4]. Flows induced by a bubble plume are observed in various
engineering disciplines, e.g. in industrial, material, chemical,
mechanical, and environmental applications. That would be situa-
tions such as chemical plants, nuclear power plants, applications in
naval engineering, for the accumulation of the surface slag in the
E-mail addresses: habdulmouti@hct.ac.ae, hassanabujihad@hotmail.com
reactive processes, waste treatment, gas mixing and resolution,
heat and mass transfer, aeronautical and astronautical systems,
biochemical reactors as well as distillation plants and so on
[20,27,9]. The main features of these kinds of flows are the follow-
ing: (1) a large scale circulation of the liquid phase can be gener-
ated in circulation systems like lakes, agitation tanks, etc. [31].
(2) Strong rising flows can be induced by the pumping effect, as
in air-lifting pumps [33]. (3) High speed surface flows are devel-
oped at the free surface, by which the density and the transporta-
tion of the surface floating substances can be controlled [2,14]. (4)
A high turbulence energy can be produced in the two-phase region
due to the strong local interaction between individual bubbles and
surrounding liquid flow [19,15]. On the other hand, the improve-
ment and development in the performance of oil barriers (oil
fences) is required, especially for high values of current velocity,
wave height and wind velocity in order to protect the environment
and the sea life from oil pollution [11,6,29,13]. Hence, a bubble
plume is considered to possibly be an effective way to control

http://dx.doi.org/10.1016/j.expthermflusci.2016.12.010
0894-1777/Ó 2016 Elsevier Inc. All rights reserved.
70 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77
Nomenclature
A area of calculation in the injector region, (injector sur-
face of the bubble generator)
D equivalent (mean) bubble diameter (mm)
H thickness of the total height or the height of the strati-
fied liquid in the tank (two layers of water and oil) (mm)
Hwater = h1 thickness (initial height) of the first layer (water)
(mm)
Hoil = h2 thickness (initial height) of the second layer (oil) (mm)
c ratio of the thickness of the second layer (oil) to that of
the total height of the immiscible stratified liquid
Qg gas volume flow rate (m3/s)
Vb rising bubble velocity (m/s)
uP particle velocity (m/s)
uL liquid velocity (m/s)
uG gas velocity (m/s)
x horizontal coordinate
the density and transportation of surface-floating substances, and
it is expected to be an effective tool to support the function of an
oil fence, since it can generate a strong and wide surface flow over
the bubble generation system and it can damp the wave motion.
Many researchers have carried out extensive modeling experi-
ments by focusing on the flow field using air bubbles as gas injec-
tion through a bottom nozzle is the most popular and has wide
applications. Since bubble plumes have been used with varying
degrees of success more information on these subjects should be
accumulated [38,39,1,24,16,17].
There are many applications for bilayer convection; the most
important one is the oil fence application. Two-layer systems have
been studied theoretically and experimentally by many research-
ers. Bilayer convection is also an interesting transport process to
be studied for both theoretical and application related reasons.
Theoretically, a bilayer convection is full of nonlinear dynamics
with a wide range of parameters to be investigated. Johnson and
Narayanan [8] studied the effects of boundaries on bilayer convec-
tion in a cylinder with idealized boundary conditions. Using as
examples two bilayer systems, a plot of the Rayleigh number ver-
sus aspect ratio (radius-height) was calculated. For certain sys-
tems, they found that the type of convection coupling, either
thermal or viscous, will change as the width of the container
changes, even as the height is fixed. Additionally, the interfacial
structure was calculated to help identify the driving force for the
convection. Sudden changes in the convection coupling and the
dominant driving force of convection were found as the container
radius increased. Interfacial tension gradients respond to the tem-
perature field and cause convection of the ‘Marangoni’ type and
this in turn is affected by buoyancy driven convection or Rayleigh
convection. Degen et al. [7] experimentally studied two-layers
heated from below. Johnson et al. [22] presented a brief review
of convective phenomena associated with material processing,
and they explained several instability phenomena that can occur
in a bilayer of two fluids heated from either the top or the bottom
and the effect of laterally and vertically confined geometries. Busse
and Sommermann [5], Fujimura and Renardy [10], Johnson and
Narayanan [23], Renardy and Stoltz [36], Ozawa et al. [35] dis-
cussed Marangoni effects and studied purely buoyancy driven con-
vection for specific two-layer systems. They also investigated the
velocity vectors of a water-silicone oil system in thermal convec-
tion. The presence of the interface and the coupling between the
fluids has received attention by both experimentalists and theo-
reticians. Beyond that, the interaction between the bubble gener-
ated liquid convection and a stratified layer is known to be
y vertical coordinate
a void fraction
qL1 density of the first layer (water) (kg/m3)
qL2 density of the second layer (silicone oil) (kg/m3)
qP density of particle (kg/m3)
d particle diameter
toil kinematic viscosity of the oil (m2/s)
twater kinematic viscosity of water (m2/s)
Ke Kinetic energy calculated per unit mass (m/s)2
b added mass coefficient for traceability of particles
against the flow in the oil layer
FDL drag force from the liquid
FDG drag force from the gas
FAL added inertial force from the liquid
FAG added inertial force from the gas
important for generating large circulation in lakes, rivers, seas,
dams and many other processes. It is very important to know the
interaction between the surface flow generated by bubbles and
an oil layer for application to an actual oil fence. Usually the oil
covers the water surface for a long time. However, in the past there
were very few reports concerning this interaction mechanism and
the bubble convection pattern in immiscible two-phase stratified
liquids as an application to an actual oil fence. The motivations
then, of the present work is the demand to broaden the knowledge
of the separation mechanism of oil due to bubbling and the flow
pattern inside the stratified two-phase liquids. In regards to appli-
cations, our aim is to focus on the actual oil transportation effect
considering an oil layer on a free surface. Laboratory experiments
have been carried out in order to investigate the multi-
dimensional motion of water and oil due to bubbles. The present
paper consists of two kinds of fundamental research. The first type
is the experimental measurement of the oil transportation phe-
nomenon due to the bubble plume (the motion of oil-water inter-
face). The flow structure of the bubble-induced convection in a
stratified liquid is clarified by using PIV measurements and path-
line measurements. The second one is the experimental study of
bubble-induced convection in a thermal stratified liquid layer
using thermo-sensitive liquid crystal tracer particles, a procedure
which has been performed in order to compare the convection pat-
tern with the immiscible stratified cases. It is confirmed by this
paper that the experimental results resemble the numerical results
that were obtained by our earlier paper [31,32]. The results show
that the flow structure depends strongly on the gas flow rate and
bubble size. The velocity of the surface flow induced by the bubble
plume in the vicinity of the oil–water interface is about 50% larger
and stronger than that inside the oil layer. Moreover, the surface
flow is particularly rapidly generated in the vicinity of the oil–wa-
ter interface. The highest kinetic energy (which is 0.012 (m/s)2) is
generated at a far distance inside the bubble plume and in the
vicinity of the oil–water interface. This observation confirms the
idea that the bubble plume can indeed generate a strong and wide
surface flow over the bubble generation system. The bubbles break
the oil layer easily, especially with high void fraction and small
bubble size and high gas flow rates.
2. Experimental apparatus and conditions
An experimental apparatus for investigating the bubble convec-
tion pattern in immiscible two phase stratified liquid is con-
 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77 71

structed as shown in Fig. 1. The tank is 500 mm in length, 750 mm Table 1

in height, and 24 mm wide. Experimental conditions of the two layers.

The stratified liquid in the tank, which consists of water and sil- Parameter Value
icone oil, stands H = 500 mm high. The experimental conditions of Density of water qL1 = 1000 kg/m3
the two layers are summarized in Table 1. The density of the oil is Kinematic viscosity of water twater = 106 m2/s
about 10% lighter than that of the water, and the kinematic viscos- Initial water thickness h1 = 350–500 mm
ity of the oil is ten times that of water. These conditions have been Density of silicone oil qL2 = 900–935 kg/m3
Kinematic viscosity of oil toil = 105–5
104 m2/s
selected to be similar to heavy oil in the sea. The bubble generator Initial oil thickness h2 = 0–150 mm
consists of 110 needles (each needle being 0.15 mm in diameter) Temperature of environment 22–24 °C
installed at the bottom of the tank. The injector surface of the bub- Maximum gas flow rate 20.0
106 m3/s
ble generator is the area A = (55
22) mm2. The gas flow rate is Density of air 1.25 kg/m3
precisely controlled by a pressure regulator and a flowmeter. A
lighting setup with a black background and metal halide lamps is
used for taking pathline images and PIV measurements. The visu- 3. Flow visualization experiment
alized flows are recorded by a digital video camera (Sony, DCR-
VX1000) that captures 30 fps. The digital images are preprocessed In order to clarify the flow pattern of the internal liquid flow
through the NIH image software version 1.60 (produced by the (the flow field around the bubble plume) in immiscible two-
National Institutes of Health of the United States of America). phase stratified liquids, the spherical tracer particles (made of a
The preprocessing entails sharpening, binarizing and smoothing high-porous polymer with diameters of 200–600 lm, and a density
of the images. In this paper, the horizontal direction is considered of 1010 kg/m3, which is almost the same as that of water) are
to be the x axis, and the y axis is the vertical direction centered in seeded in the entire tank as tracer particles for both the pathlines
the bubble plume. The point of origin of the x, y coordinate system and the PIV measurements then the flow is visualized. The trace-
is located at the center of the bubble generator, as shown in Fig. 1. ability of these particles against the flow in the oil layer is esti-
The bubble injector conditions are shown in Table 2; the values in mated theoretically by using the equation of translational motion
this table are calculated by using the time average of 120 consec- of spherical particles (Eq. (1)). Eq. (1) consists of the inertial force
utive frames in the image processing (4 s). This reflects the accu- of the particle at the left hand side, the drag force from the liquid,
racy of the experimental results. The averaged bubble diameter the drag force from the gas, the added inertial force from the liquid.
and the standard deviation have been calculated by measuring While the added inertial force from the gas at the right hand side is
more than 1200 bubbles in the local VTR images in the bubble neglected because of the low mass density of the gas, only the
plume using image processing. These images are taken by record- added inertial force from the liquid FAL is acting at each instant
ing local pictures of the injector region of the bubble generator. The of time since the particle is immersed in the liquid (surrounded
bubble diameter is defined by the equivalent bubble diameter by liquid). On the other hand, the value of the velocity of the par-
using ellipsoidal approximations for the bubble shapes. The equiv- ticles to that of the oil liquid is calculated to be more than 97%.
alent bubble diameter is estimated by the vertical and the horizon- Using these estimations, it has been confirmed that the particles
tal lengths of each bubble, which are obtained by using the NIH sufficiently follow the flow in the oil layer when the frequency of
software after binarizing the images. The measurement uncer- the velocity fluctuation is lower than 20 Hz.
tainty for the bubble diameter is estimated to be around  3
(0.01 mm) according to the pixel resolution. The void fraction (a) 4 d Du
p qP P ¼ F DL þ F DG þ F AL ð1Þ
is calculated by using the equation (a = Qg/A
Vb), 3 2 Dt
[2,13,14,15,32,21,34,28], where A is the area of calculation in the The recorded images are ported to a computer. The flows are
injector region, (injector surface of the bubble generator), A = measured in this section by using the PIV technique and the path-
(55
22) mm, Vb is the rising bubble velocity. The measurement lines measurements for 3 cases of (c = h2/H), where c is the ratio of
uncertainty for the void fraction is estimated to be about 2%. the height of the second layer (oil) to that of the total liquid height.

Metal halide light

h1
Oil
h H
Air 2

Water

Bubble injector
Flowmeter
Pressure controller

Fig. 1. Schematic diagram of experimental apparatus.

72 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77

Table 2
The bubble injector conditions.

Case No. Gas flow rate Qg (m3/s) Mean bubble diameter D (mm) Standard deviation of D (mm) Void fraction a
6
Case 1 0.28
10 0.20 0.008 0.0020
Case 2 1.39
106 0.30 0.012 0.0035
Case 3 2.78
106 0.45 0.020 0.0050
Case 4 4.17
106 0.60 0.028 0.0070
Case 5 5.56
106 0.75 0.037 0.0090
Case 6 6.94
106 0.80 0.045 0.0105
Case 7 8.33
106 1.00 0.051 0.0115
Case 8 9.72
106 1.15 0.058 0.0130
Case 9 11.11
106 1.35 0.065 0.0145
Case 10 12.50
106 1.55 0.072 0.0160
Case 11 13.89
106 1.75 0.080 0.0180
Case 12 15.28
106 1.90 0.089 0.0190
Case 13 16.67
106 2.00 0.098 0.0200

Fig. 2 shows sample of the recorded images of the flow field for the
cases: (a) c = 0 and for case-4 of the experimental conditions, (b)
c = 0.1 and for case-3 of the experimental conditions and (c)
c = 0.3 and for case-1 of the experimental conditions respectively.
In these images the bubble plume is located in the middle of the
image, while the particles are distributed around the bubble
plume. Our observation showed that the particles and the bubbles
have an almost two-dimensional motion in the x-y plane. The
time-averaged flow is almost two-dimensional especially because
the tank width is small and there are no perpendicular components
to the front and back walls in the time averaged flow field. The flow
is strong in the bubble plume regions, which are far from the wall
side. Hence, a two-dimensional measurement using PIV and path-
lines is helpful for grasping the time-averaged internal flow struc-
ture. This kind of flow is 2 dimensions as reported by many
researchers [2,13,14,15,34,28,31,32,33,21,9,27,30,12,19,38,39,18],
among others.
The images are ported to a computer and the pathlines are cal-
culated for 60 consecutive frames (2 s) averaging the movement of
the particles after preprocessing the digital images through NIH
software as shown in Fig. 3. Moreover, the time-averaged velocity
vector maps of these cases are obtained by using the BDCC (Bright-
ness Distribution Cross-Correlation) method [3,25], (which repre-
sents the two-phase flow pattern) as shown in Fig. 4. The grid
averaging method is used in order to get grid-rearranged vector
maps. Although the PIV velocity vectors obtained may be of the
gas–liquid interfaces rather than the liquid, it is thought reason-
able to assume that in the plume the liquid velocity is the same
as the bubble velocity. These results are obtained by using more
than 250,000 velocity vectors for each case, which are captured
during 120 frames (4 s). This reflects the accuracy of the experi-
mental results since the number of frames and the time allocated
for either the calculation and the measurement is rather long when
compared to those by others [21,34,28,31,32,33,9,27,30,12,19,38,
39,18], among others.
According to these figures the detailed flow mechanism inside
the two layers can be explained as follows. The resulting flow is
steady and symmetric relative to the bubble plume center because
a small gas flow rate is given (Qg = 4.17
106 m3/s); the main
upward liquid flow in the water layer is driven along the bubble
plume by the rising bubbles. The flow evolves as follows: The
momentum of the upward flow attains a maximum near the oil–
water interface where the velocity of bubble reaches 0.9 m/s. Just
under the oil–water interface the upward flow changes its orienta-
tion rapidly (with a velocity about 1 m/s) into a horizontal flow.
Then, in the water layer a pair of liquid circulations is generated
next to the bubble plume. After time passes by, the pair of circula-
tions induces a whole scale circulation in the entire water layer. At
the same time, some bubbles accumulate on the oil–water inter-
face, and a part of the bubble plume penetrates and passes through
the oil stratum causing a secondary flow inside the oil layer by the
buoyancy of the penetrating bubbles. The velocity of the bubbles

Fig. 2. Sample of recorded images.

 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77 73

Fig. 3. The pathlines of the flow pattern.

Fig. 4. The time averaged velocity vector map of the flow pattern.

inside the oil layer is (10–20%) less than that inside the water layer
due to the high oil viscosity. Thus, the velocity of the bubbles
decreases quickly (to about 0.2 m/s) when the bubbles enter the
oil layer. Therefore, the void fraction inside the oil layer increases.
The flow pattern inside the oil layer is still very weak and basically
it is only rising bubbles and particles for the case (b). When the
ratio reaches c P 1/3 case (c), a different flow pattern results
inside the second layer, while the first layer preserves almost the
same flow pattern. When the ratio reaches c P 1/2, the flow pat-
tern becomes complicated because double convection occurs near
the interface of the two layers. Also, inside the bubble plume the
bubble penetrates the oil layer faster with a velocity of 0.3–
0.5 m/s. However, the detailed mechanism of generating the sec-
ondary flow inside the oil layer (which is similar to that induced
in the water layer) for the case (c) can be explained as follows:
The oil flow is driven by the bubble plume. The momentum of this
flow becomes maximum near the free surface (but is still about
50% smaller than that induced in the vicinity of the oil–water inter-
face). Just under the free surface the upward flow inside the oil
layer changes its orientation into a horizontal flow. Then, a pair
of liquid circulations is generated in the oil layer besides the bub-
ble plume. Around this pair of circulations, the accumulated bub-
bles on the oil–water interface slowly rise through the oil layer
to the free surface and float. However, the pair of circulations
inside the oil layer is smaller than that induced inside the water
layer due to the effect of the oil viscosity. Hence, the surface flow
induced by the bubbles inside the oil layer at the free surface is
weaker and smaller in scale than that induced inside the water
layer in the vicinity of the oil–water interface.
In order to clarify the detailed structure of the flow in the two
layers, the two-dimensional distribution of the kinetic energy is
calculated from the measured averaged velocity vector map as
shown in Fig. 5. The kinetic energy is calculated per unit mass
using the following equation:
1 2
Ke ¼ ðu þ v 2 Þ ðm=sÞ2 ð2Þ
2
Fig. 5 shows that the highest kinetic energy (which is 0.012 (m/
s)2) is generated at a long distance in the center of the bubble
plume and in the vicinity of the oil–water interface. These results
correspond with the numerical results that were obtained by our
earlier papers [31,32]. Where in the most typical four structures
of the simulated results, which were obtained by the numerical
simulation, are summarized as follows: Each figure is the instanta-
neous flow at 120 s after starting bubbles injection. The figure is an
example for the case where the bubbles separate and break the oil
layer and reach to the free surface. In the second figure, the bubbles
only pass the oil layer while the water does not reach the surface.
74 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77
Fig. 5. The kinetic energy distribution (m/s)2.
In Third figure, the bubbles cannot pass the oil layer. Some of the
bubbles start to accumulate in the vicinity of oil–water interface
and the other part of the bubbles is transported with the descent
of the water. The forth figure is of the same condition as the third
one, but it is the flow at 60 s after. At this time, the bubbles, which
were accumulated near the oil–water interface, start to float slowly
through the oil. From these simulation results, it is confirmed that
the oil layer can be separated by a bubble plume when the bubble
plume has a high void fraction and small bubbles. Those numerical
results agree with our experimental results.
The convection due to the bubble plume inside the oil-layer is
different from thermal convection. In the case of thermal convec-
tion (Lower Dragging Mode, Viscous Coupling, Upper Dragging
Mode and Convection induced by a bubble plume in a thermal
stratified liquid) [32,7,22,23,36], oil has an opposite circulation to
that of water. Therefore, the oil–water interface is quite stable,
because the flow in the vicinity of oil–water interface has the same
direction; hence, there is no strong shear stress acting on the inter-
face. In clear contrast, the flow for the case of bubble-convection
close to the interface is counter-current. Thus, the flow at the inter-
face is outwards in the water, and inwards in the oil, and hence
opposing for bubble convections, as shown in Fig. 4. Therefore,
the interface becomes unstable and experiences large displace-
ments, especially when the gas flow rate increases, since the direc-
tion of the two flows inside the oil and the water are opposite to
each other. Another reason is the strong local momentum interac-
tion between the bubbles and the liquid. This phenomenon
enhances the mixing of the two liquid layers due to the bubble
plume.
The maximum speed of the surface flow induced by a bubble
plume is governed by the flow structure in the initial region where
the rising flow changes into a surface flow. The experimental
observation clarifies the following: (1) Due to the penetration of
the bubbles through the interface, the oil–water interface heaves
up in the center part when a small gas flow rate is given with
parameters Qg 6 12.50
106 m3/s, D 6 1.55 mm and a 6 0.016.
In this case the bubbles enter and penetrate the oil stratum and
reach the upper free surface, while the water does not. Moreover,
some bubbles accumulate in the vicinity of the oil–water interface
and then start to float slowly through the oil around the center
part. (2) The oil layer is separated by the rising water in the center
part and descends downwards when a large gas flow rate is pro-
vided with parameters Qg 6 15.28
106 m3/s, D 6 1.90 mm and
a 6 0.019. (3) The oil layer is completely separated and broken
by the strong convection when larger gas flow rates are given with
parameters Qg = 16.67
106 m3/s, D = 2.0 mm and a = 0.020. The
power efficiency of breaking the oil layer is the necessary power to
separate the oil layer and is calculated by dividing the potential
energy of the interface by the buoyancy energy of injected the bub-
bles. This efficiency is about 60–75%. We can therefore say that the
bubbles can easily break and separate the oil layer. This confirms
that the bubble plume can separate the oil stratum when the bub-
ble plume has a high void fraction and a high gas flow rate.
To review and summarize, the results presented in the above
mentioned figures confirm the following points: (1) First, there
are two circulating flows of liquid near the bubble plume for both
the water and the oil layers. The circulation flows inside the water
layer are larger than that inside the oil layer. (2) The velocity of liq-
uid is high (about 0.8 m/s) inside the bubble plume in the water
layer and near the oil–water interface, while low in other regions
(about 0.1 m/s). (3) It is recognized that as the gas flow rate
increases, the magnitude of velocity increases and the effective
area of the bubble plume (the width of the surface flow) expands
in horizontal direction. Here we can recognize two effective areas:
the first one is located at the oil–water interface and the second is
located on the free surface. Hence, the surface flow of the oil–water
interface which is induced by the bubble plume is stronger and lar-
ger than that induced in the free surface. (4) The highest kinetic
energy (which is 0.012 (m/s)2) is generated at a long distance
inside the bubble plume and in the vicinity of the oil–water inter-
face. This observation confirms the idea that the bubble plume can
indeed generate a strong and wide surface flow over the bubble
generation system. These results agree with the numerical results
that were obtained by our earlier papers [31,32]. Where those sim-
ulated results showed the flow structure at 120 s in the case that
the void fraction is fixed to 0.01 and the bubble size is changed
from 0.2 to 2.0 mm. The curves are the streamlines of liquid phase.
The bubble plume is straightly formed and the liquid flow is
strongly induced in case of large bubbles, while in case of small
bubbles the liquid flow is only induced in the water since the bub-
ble plume becomes unstable. This instability is due to the local
strong momentum interaction between the bubbles and the liquid.
As further time passes, the bubbles distribute widely in both the oil
and the water regions. In the oil layer, the bubbles rise slowly with
a configuration of small mushroom clouds. The next figure of the
numerical results picked up the two patterns of the flows, which
were drawn by the velocity vectors, and the vorticity of the liquid
phase. The first part of this figure stands the case that the bubbles
penetrate the oil layer where the liquid flow changes from stable
straight jet flow under the interface to unstable flow involving vor-
tices over the interface. The vorticity distribution indicates the
presence of strong rotation inside the oil layer or near the oil–wa-
 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77 75

300 K 303 K

Blue High

Green
Middle
Yellow

Red Low

(a) Raw image recorded in DVC (b) Measure temperature

Fig. 6. Test of color-to-temperature transformation.

ter interface. While the second part of the figure stands the case
where the bubbles accumulate under the oil–water interface, here
there is no flow inside the oil layer, but a large scale circulation
occurs in the water region. Also, there is a strong vorticity genera-
tion at the oil–water interface, which enhances the flow fluctua-
tion. The aforementioned points make it clear that the entire
flow structure depends on bubble penetration in the oil.
4. Bubbly flow in thermal stratified liquid
Convection induced by a bubble plume in a thermal stratified
liquid is measured by using liquid crystal tracer particles in order
to find out the difference between immiscible and miscible fluids.
The test tank size is 200 mm in length, 100 mm in height and
12 mm in width. Silicone oil of 10 cSt in kinematic viscosity, and
960 kg/m3 in density is used as liquid media. The liquid crystal tra-
cer particle is cholesteric type, which illuminates vividly in the Sil-
icone oil at the temperature range between 300.4 K and 303.4 K,
made by Japan Capsular Products (Product No. RM2830). Fig. 6
shows the raw image and the temperature distribution without
bubble injection. The bottom and the top plane temperatures are
300.4 K and 303.4 K respectively, which are controlled by two con-
stant temperature bathes (Taitec, Ltd, type TR2114). The second
part of the figure shows the color distribution related to the tem-
perature. The red color indicates high temperature 303.4 K and
the green refers to middle while the blue indicates low tempera-

300 K 303 K

(a) T=0.5 s (d) T=20.0 s

(b) T=5.0 s (e) T=30.0 s

(c) T=10.0 s (f) T=60.0 s

Fig. 7. Temperature transportation due to chain bubbles.
76 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77
Surface Flow
(a) Water only (b) Oil-Water Phases
Fig. 8. Flow pattern in three environment cases.
ture 300.4 K. The color distribution was produced based on an
algorithm which is estimated from dimensionless intensities of
the three color elements, i.e., red, green, and blue. The algorithm
was used to perform the color-to-temperature transformation by
16 bit type color image processing.
Fig. 7 shows the time history of the measured temperature dis-
tribution. In the initial stage, cold liquid rises with the bubbles and
reaches the free surface. Then, a pair of liquid circulations convec-
tion is generated next to the bubble plume around the chain-
bubbles. After 60 s passes, the temperature distribution becomes
uniform only in the circulation area. If there is no temperature gra-
dient, the surface flow induced by the bubbles occupies a wide
area. The reason why it is not kept in this experiment is that the
high temperature liquid near the free surface does not descend
due to the buoyancy.
Fig. 8 illustrates the flow patterns induced by a bubble plume in
three kinds of environments. When there is neither oil layer nor
thermal stratification, the bubble plume generates high speed sur-
face flow in the vicinity of the water surface (Fig. 8(a)). When the
water surface is covered with an oil layer, a secondary flow is
induced inside the oil layer by the buoyancy of the penetrated bub-
bles. In this case, the oil–water interface becomes unstable since
strong shear stress acts near the interface because the direction
of the two flows inside the oil and the water are opposite to each
other. Therefore, the bubbles can easily break the oil layer. When
the liquid is single phase with a temperature gradient, the convec-
tion induced by the bubble plume is limited in the circulation area
even after passing a long time. In the circulation area the temper-
ature becomes uniform due to the mixing effect, however, there is
no mixing with the external area. Therefore, it can be said that the
presence of immiscible interface enhances the mixing of the liquid
due to the bubble plume.
5. Concluding remarks
Flow visualization of the bubble plume in two immiscible strat-
ified fluids is carried out in order to improve the applicability of the
bubble plume as an oil fence. The circulation flow driven by bubble
in a thermal stratification is elucidated by using PIV measure-
ments. The covering effect of the oil layer on the free surface and
the convection due to the bubble plume are investigated by using
image processing, PIV measurements and pathlines measurements.
Shear Stress Shear Stress
(c) Thermal Gradient
The experimental results and the numerical results that were
obtained by our earlier paper [31,32] show good analogy and the
main results can be summarized as follows:
1. The flow structure is strongly modulated by the gas flow rate
and bubble size.
2. The velocity of the surface flow induced by the bubble plume in
the vicinity of the oil–water interface is twice larger and stron-
ger than that inside the oil layer. Moreover, the surface flow is
particularly rapidly generated in the vicinity of the oil–water
interface.
3. The bubble plume can indeed generate a strong and wide sur-
face flow over the bubble generation system because the high-
est kinetic energy (which is 0.012 (m/s)2) is generated at a far
distance inside the bubble plume and in the vicinity of the
oil–water interface.
4. The oil layer is easily broken by bubbles, especially with high
void fraction and small bubble size and high gas flow rates.
References
[1] H.K. Abdel-Aal, G.B. Stiles, C.D. Holland, Formation of interfacial area at high
rates gas flow through submerged orifices, AICHE J. 12 (1966) 174–180.
[2] H. Abdulmouti, Y. Murai, Y. Ohno, F. Yamamoto, Measurement of bubble plume
generated surface flow using PIV, J. Visual. Soc. Jpn. 21 (2) (2000) 31–37.
[3] R.J. Adrian, Particle-imaging techniques for experimental fluid mechanics,
Annu. Rev. Fluid Mech. 23 (1991) 261–304.
[4] Amir Arash Rafiei Mehrabadi, Effects of frother type on single bubble rise
velocity. A thesis submitted to McGill University in partial fulfillment of the
requirements for the degree of Master of Engineering McGill University.
Montreal, Canada, June 2009.
[5] F.H. Busse, G.V. Sommermann, Double-layer convection: a brief review and
some recent experimental results, in: Y.Y. Renardy, A.V. Coward, D.
Papageorgiou, S.M. Sun (Eds.), Advances in Multi-Fluid Flows, SIAM,
Philadelphia, 1996.
[6] A.K. Chesters, M. Van Doorn, L.H.J. Goossens, A general model of unconfined
bubble plumes from an extended source, Int. J. Multiphase Flow 6 (1980) 499–
521.
[7] M.M. Degen, P.W. Colovas, C.D. Andereck, Time-dependent patterns in the two-
layer rayleigh-benard system, Phys. Rev. E 57 (1998) 6647–6659.
[8] D. Johnson, R. Narayanan, Geometric effects on convective coupling and
interfacial structures in bilayer convection, Phys. Rev. E 56, 5462–1 56 (5)
(1997), http://dx.doi.org/10.1103/PhysRevE.56.5462.
[9] T.K. Fannelip, S. Hirschberg, J. Kuffer, Surface current and recirculating cells
genelated by bubble curtains and jets, J. Fluid Mech. 229 (1991) 629–657.
[10] K. Fujimura, Y.Y. Renardy, The 2:1 Steady/Hopf mode interaction in the two-
layer Benard problem, Physica D 85 (1995) 25–65.
 H. Abdulmouti / Experimental Thermal and Fluid Science 83 (2017) 69–77
[11] R.W. Gross, J.M. Kuhlman, Three-component velocity measurements in a
turbulent recirculating bubble-driven liquid flow, Int. J. Multiphase Flow 18
(1992) 413–421.
[12] S. Hara, M. Ikai, S. Namie, Fundamental study on an air bubble type of oil
boom, Trans. Ship-Making Soc. Kansai-Jpn. 194 (1984).
[13] Abdulmouti Hassan, Visualization and Image Measurement of Flow Structures
Induced by a Bubbly Plume Ph. D. thesis, Fukui University, 2003.
[14] Abdulmouti Hassan, Surface flow generation mechanism induced by bubble
plume, Yanbu J. Eng. Sci. (YJES) (2011), Second issue. PS-M02-28 (50–67).
[15] Abdulmouti Hassan, Measurement of Flow Structures Induced by a Bubbly
Plume Using Visualization, PIV and Image Measurement, Scholar’s Press, 2013,
ISBN-13: 978-3-639-51490-2. 7, June.
[16] Abdulmouti Hassan, Bubbly two-phase flow: Part I-characteristics, structures,
behaviors and flow patterns, Am. J. Fluid Dynam. 4 (4) (2014) 194–240, http://
dx.doi.org/10.5923/j.ajfd.20140404.03, Scientific and Academic Publishing.
[17] Abdulmouti Hassan, Bubbly two-phase flow: Part II-characteristics and
parameters, Am. J. Fluid Dynam. 4 (4) (2014) 115–180, http://dx.doi.org/
10.5923/j.ajfd.20140404.01, Scientific and Academic Publishing.
[18] A. Hassan Yassin, T.K. Blanchat, C.H. Seeley Jr., R.E. Canaan, Simultaneous
velocity measurement of both components of a two-phase flow using particles
image velocimetry, Int. J. Multiphase Flow. 18 (3) (1992) 371–395.
[19] N.A. Hussain, B.S. Narang, Simplified analysis of air-bubble plumes in
moderately stratified environments, ASME. J. Heat Transfer 106 (1984) 543–
551.
[20] N.A. Hussain, R. Siegel, Liquid jet pumped by rising gas bubbles, J. Fluids Eng.
12 (1976) 49–61, March.
[21] K. Isao, S. Akimi, D.C. Besnard, Prediction of turbulence suppression and
turbulence modeling in bubbly two-phase flow, Nuclear Engineering and
Design, vol. 141, North-Holland, 1993, pp. 145–158.
[22] D. Johnson, R. Narayanan, P.C. Dauby, Geometric effects on convective
coupling and interfacial structures in bilayer convection, Phys. Rev. E 56
(1997) 5462–5472.
[23] D. Johnson, R. Narayanan, Marangoni convection in multiple bounded fluid
layers and its application to materials processing, Phil. Trans. R. Soc. Lond. A
356 (1998) 885–898.
[24] W.T. Jones, Air barriers as oil-spill containment devices, J. Soc. Pet. Eng. (1972)
126–142.
77
[25] I. Kimura, H. Kimura, T. Takamori, Image processing of flow around a circular
cylinder by using correlation techniques, in: Proc. 4th Int. Int. Symp. Flow
Visualization, Hemisphere, 1986, pp. 221–226.
[26] Lawrence K. Wang, Nazih K. Shammas, William A. Selke, Donald B. Aulenbach,
Gas dissolution release and bubble formation in flotation systems, in: L.K.
Wang et al. (Eds.), Handbook of Environmental Engineering, Flotation
Technology, vol. 12, Springer Science + Business Media LLC, 2010, http://dx.
doi.org/10.1007/978-1-60327-133-2_2.
[27] A.M. Leitch, W.D. Baines, Liquid volume flux in a weak bubble plume, J. Fluid
Mech. 205 (1989) 77–98.
[28] Y. Matsumoto, Y. Murai, Numerical simulation of bubble plume in a tank with
free surface, Trans. Jpn. Soc. Mech. Eng. 61B (588) (1995) 54–61.
[29] T.J. McDougall, Bubble plumes in stratified environments, J. Fluid Mech. 85
(1978) 655–672.
[30] J.H. Milgram, Mean flow in round bubble plumes, J. Fluid Mech. 133 (1983)
345–376.
[31] Y. Murai, H. Abdulmouti, Y. Ohno, J. Ohta, F. Yamamoto, Two-phase flow
induced by a bubble plume in the vicinity to a free surface, in: ASME, FEDSM-
11275, 2000, pp. 1–6.
[32] Y. Murai, Y. Ohno, S. Bae Dae, H. Abdulmouti, F. Yamamoto, Bubble-generated
convection in immiscible two-phase stratified liquids, in: Proceeding of ASME
FEDSM. May 29- June 1–18180. New Orleans U.S.A, 2001.
[33] Y. Murai, Y. Matsumoto, Numerical analysis of detailed flow structures of a
bubble plume, JSME Int. J. Ser. B 41 (3) (1998) 568–575.
[34] Y. Murai, Y. Matsumoto, Eulerian analysis of bubbly two-phase flows using CIP
scheme, Comput. Fluid Dynam. J. 8 (1) (1999).
[35] M. Ozawa, U. Muller, I. Kimura, T. Takamori, Flow and temperature
measurement of natural convection in a Hele-Shaw cell using a thermo-
sensitive liquid-crystal tracers, Exp. Fluids 12 (1992) 213–222.
[36] Y.Y. Renardy, C.G. Stoltz, Time-dependent pattern formation for convection in
two layers of immiscible liquids, Int. J. Multiphase Flow 26 (2000) 1875–1889.
[38] T.Y. Sun, G.M. Faeth, Structure of turbulent bubbly jets-I. Methods and
centerline properties, Int. J. Multiphase Flow 12 (1986) 99–114.
[39] T.Y. Sun, G.M. Faeth, Structure of turbulent bubbly jets-II. Phase property
profiles, Int. J. Multiphase Flow 12 (1986) 115–126.
[40] G.I. Taylor, The action of a surface current used as a breakwater, Proc. Roy. Soc.
A231 (1955) 466–478.