Numerical Study on Influence of Hydrofoil Clearance towards Total
Drag Reduction on Winged Air Induction Pipe for Air Lubrication
Yanuar
1
,Muhammad Akbar
1
, M. Alief
1
, Fatimatuzzahra
1
, and Made
1
1
Department of Mechanical Engineering, University of Indonesia, Depok, Indonesia
Keywords:
Drag Reduction, Air Lubrication, Hydrofoil, Multi-Phase Flow, Computational Fluid Dynamics.
Abstract:
A new device for air lubrication called Winged Air Induction Pipe (WAIP) is studied in the present work.
The device, which consists of angled hydrofoil uses the low-pressure region produced above the hydrofoil as
ship moves forward. The low pressure drives the atmospheric air into the water in certain velocities which the
pressure is negative compare to atmospheric pressure. A computational fluid dynamics approach is presented
to study the effect of hydrofoil clearance of Winged Air Induction Pipe in drag reduction experienced by the
plate which WAIP attached. The well-known volume of fluid’ model and k-ω SST (shear stress transport)
turbulence closure model have been used in the 2D numerical simulation in ANSYS Fluent. The numerical
simulation is carried out with different configuration of hydrofoil clearance and angle of attack. Effects of
these parameters on total drag force and drag reduction are reported. The reduction of drag force is found to
increase to about 10% compared to bare plate configuration.
1 INTRODUCTION
The methods of drag reduction using air lubrication
are becoming promising study due to the increase
of fuel efficiency produced as the result of reduced
drag(Cui et al., 2003). The principle of air lubrica-
tion method is to reduce the Reynold shear stress oc-
curs on the boundary layer of the flow (Yanuar et al.,
2012)(Toffoli et al., 2010). The magnitudes of the
Reynold shear stress can be moderately changed by
the dispersed phase for the dilute two-phase flow, but
the distribution pattern keeps unchanged (Muste et al.,
2009). Kodama et al. (2000) found promising result
using air lubrication in the form of microbubble for
drag reduction. It is well known that the presence
of the air in the turbulent boundary layer of the flow
leads to drag reduction for two reasons: first by low-
ering the average viscosity and density of the mixture
flow. The mixture of gas and liquid has lower density
and viscosity compare to the liquid itself; second, by
decreasing the magnitudes of the Reynold shear stress
through the interaction of the air and liquid.
Numerical study also can be performed to calcu-
late drag reduction produced by air lubrication. Nu-
merical study has been done as an alternative to ex-
perimental study as the numerical requires less time
and still gives accurate result by conducting valida-
tion towards the similar experimental result first. Var-
ious numerical study has been performed to calculate
the drag reduction using various air lubrication. Mo-
hanarangam et al. (2009) studied the phenomenon of
drag reduction by the drag reduction by the injection
of microbubble into a turbulent boundary layer us-
ing a Eulerian–Eulerian two-fluid model. Pang et al.
(2014) investigated microbubble drag reduction using
the Euler–Lagrange two-way coupling method in or-
der to understand the drag reduction mechanism by
microbubbles. Shereena et al. (2013)) and Evans et al.
(1992) conducted a numerical simulation using k-ω
SST to calculate the drag reduction produced by air
jet on an axisymmetric underwater vehicle.
The air lubrication requires an injection to dis-
perse air into the water. The injection requires en-
ergy due to the higher pressure in the water partic-
ularly in certain depth in the ship bottom hull. The
pressure from air compressor is required in order to
inject air into the water. However, the amount of en-
ergy required is large enough to cancel out a part of
the energy saved by the air lubrication. The injection
of the air into the water in certain depth requires vari-
ous source of energy: first the adiabatic compression,
the air generation in the water and mechanical losses
at the air compressor (Kumagai et al., 2015). As the
result, he net-power saving declines as little as 5%.
Kumagai et al. (2015) found a new device called
Winged Air Induction Pipe (WAIP). The WAIP con-
Yanuar, ., Akbar, M., Alief, M., , F. and , M.
Numerical Study on Influence of Hydrofoil Clearance towards Total Drag Reduction on Winged Air Induction Pipe for Air Lubrication.
DOI: 10.5220/0008544001630169
In Proceedings of the 3rd International Conference on Marine Technology (SENTA 2018), pages 163-169
ISBN: 978-989-758-436-7
Copyright
c
2020 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
163
sist of the air pipe and angled hydrofoil that has a
lower pressure in the upper surface due to the higher
velocity magnitudes. Previously, numerous studies on
the effect of the hydrofoil on air-water interface has
been performed. Duncan (1981) conducted and ex-
periment of the breaking waves produces by a towed
hydrofoil at constant depth and velocity. Kumagai
et al. (2011) and Muratoglu and Yuce (2015) found
that the hydrofoil also produces a negative pressure
that pull in the air above into the water as the hydro-
foil positioned near the water surface.
In the present work the WAIP from previous work
(Kumagai et al., 2015) is studied. The device pro-
duces natural air injection without using an air com-
pressor at critical velocity Uc that is defined as:
U
C
=
s
2gHα
C
P
α (L/h
b
)C
D
sinθ
(1)
where g is gravity acceleration, H is the depth of the
injection, α is the mean void fraction, C
P
is pressure
coefficient, and L, h
b
, C
D
, and θ are hydrofoil chord
length, the air-water mixture layer thickness, and hy-
drofoil angle of attack respectively. However, in their
study Kumagai et al. (2015) found that the hydrofoil
in some cases develop some problem regarding the
clearance to the bottom plate where the WAIP placed.
Additionally, it should be noticed that the present nu-
merical simulation is aimed to analyze the influence
of the hydrofoil clearance in Winged Air Induction
Pipe towards the amount of drag reduction produced
and the relationship between the angle of attack and
clearance of the hydrofoil in WAIP device.
2 NUMERICAL STRATEGIES
For simulating two phase flow, the Volume of Fluid
(VOF), as implemented in Fluent, is used. This can
be used to model the separation of air and water above
and below the ship respectively. The water is imple-
mented as the primary phase and air as the secondary
phase. The surface tension modeling also used in the
modeling to achieve representation of the air-water
contour. For a best viewing experience the used font
must be Times New Roman, on a Macintosh use the
font named times, except on special occasions, such
as program code (Section 2.3.7).
2.1 Governing Equations
Three dimensional, transient, viscous, incompress-
ible, and two-phase immiscible fluid flow is numer-
ically solved by discretizing RANS equations.
· U = 0 (2)
Figure 1: Schematic diagram from side-view of the compu-
tational setup
δρU
δt
+ (ρUU
T
) = p
+ (µU) + (ρτ) + S
(3)
where U = (u
x
+ u
y
+ u
z
) is the velocity vector, t
is time. is vector differential. p is relative pres-
sure. ρ and µ are fluid properties the density and dy-
namic viscosity, respectively τ is Reynold stress ten-
sor for turbulence flow. Closure of the turbulence
model for τ is k-ω Shear Stress Transport (SST). The
turbulence kinetic energy k and specific dissipation
rate ω are estimated from the boundary condition of
turbulence quantities turbulence intensity I and length
scale l. For simulating turbulent flow, the Shear-Stress
Transport (SST) k-ω is used to model the near wall
region of the flow. This is based on previous study
that found this model is well suited for simulating
two phase flows(Mohanarangam et al., 2009). The
k-ω SST model is an effective blend of robust and ac-
curate formulation of the k-ω in the near wall region
and k-ω model in the far field (Shereena et al., 2013).
The k-ω SST model gives more realistic result in pre-
diction of void fraction occurs in the near wall region
(Menter, 1994). Instead of using empirical wall func-
tion to correlate the near wall and far field region, k-ω
SST solved two turbulence scalar directly towards the
wall boundary (Mohanarangam et al., 2009). SST k-ω
can be described as: Kinematic eddy viscosity
ν
T
=
a
1
k
max(a
1
ωSF
2
)
(4)
Turbulent kinetic energy
δk
δt
+U
j
δk
δx
j
= P
k
β
kωt +
δ
δx
j
(ν + σ
k
ν
T
)
δk
δx
j
(5)
Dissipation rate
δω
δt
+U
j
δω
δx
j
= αS
2
βω
2
+
δ
δx
j
(ν + σ
ω
ν
T
)
δω
δx
j
+ 2(1 F
1
)σ
ω2
1/ω
δk
δx
i
δω
δx
i
(6)
As soon as the air introduced into the water, the
flow becomes two phase. The Volume of Fluid is
implemented in Fluent. This can be used to model
two phase flow and gives representation of the air wa-
ter interface. the continuity equation of the air-water
mixture can be defined as:
δρ
m
δt
+ ρ
m
v
m
= 0 (7)
SENTA 2018 - The 3rd International Conference on Marine Technology
164
where ρm id the density of the mixture, t is time and
v
m
is system average velocity. The formulation of
the density and system average velocity can be given
as in Eq.(8-9) where α
k
is the volume fraction from
the k phase, ρ
k
is the density of the k phase, and v
k
if the k phase average velocity. The momentum equa-
tion of the air-water mixture can be obtained by sum-
ming the individual momentum from each phase. The
equation given as in Eq.(10).
ρ
m
=
n
k=1
α
k
ρ
k
(8)
v
m
=
n
k=1
α
k
ρ
k
v
k
ρ
m
(9)
δ
δt
(ρ
m
v
m
) + · (ρ
m
v
m
·
v
m
) = p + · [µ
m
(
v
m
+
v
T
m
v
T
m
)] + ρ
m
g +
F + ·
n
k=1
α
k
ρ
k
v
dr,k
v
dr,k
(10)
where p is pressure, g is the gravity acceleration, F is
the body force intensity, and µ
m
is air-water mixture
dynamic viscosity that can be expressed as:
mu
m
=
n
k=1
α
k
µ
k
(11)
v
dr,k
is the drift velocity for secondary phase k, de-
fined as:
v
dr,k
=
v
k
v
m
(12)
where µ
k
is the dynamic viscosity of the k phase. The
relative velocity is defined as the velocity of a sec-
ondary phase p relative to the velocity of the primary
phase q.
v
pq
=
v
p
v
q
(13)
The mass fraction of any phase k given as:
c
k
=
α
k
ρ
k
ρ
m
(14)
Drift velocity and relative velocity v
p
q connected by:
v
dr,p
=
v
pq
n
k=1
c
k
v
qk
(15)
From the previous continuity equation for sec-
ondary phase p, the volume fraction of the secondary
phase p can be obtained as:
δ
δt
(α
p
ρ
p
) + · (α
p
ρ
p
v
m
) = · (α
p
ρ
p
v
dr,p
)
+
n
k=1
( ˙m
qp
˙m
pq
)
(16)
where m
qp
and m
pq
are the mass flow rates.
Figure 2: Part B: WAIP Set up
Figure 3: Computational Domain
2.2 Computational Domain
The analyze object is simplified by dividing the model
into four parts. As described in the Figure. 2, part A
(2,216 mm) is the bare plate, part B (80 mm) is the
WAIP, part C (2,600 mm) is bare flat plate that re-
ceived the effect of the WAIP, and part D (184 mm)
is the afterbody. The computational domain uses two
different in length of plane to separate the air and wa-
ter. As the segment EFHI is the water boundary and
segment FGIJ is the air boundary.
The air is treated with 0 velocity and the water is a
moving fluid with the velocity of U
c
= 5.6 m/s (Kuma-
gai et al., 2015). The details of the boundary condi-
tions are adopted from previous simulation (Shereena
et al., 2013) The boundary condition are: (a) segment
EF is a velocity inlet, i.e. where U is pre-described
in the –x direction; (b) segment DE is pressure out-
let. The rest of the computational domain edges are
treated as nonslip wall.
Table 1: Experimental Data with 15 mm of Clearance (Ku-
magai, et al., 2015)
Type θ
v (m/s) D
b
(N) D
b
(N) %R
Bare Plate Bare 5.6 181.49 - -
WAIP 12 5.6 167.75 -13.73 7.57
WAIP 16 5.6 166.77 -14.72 8.11
WAIP 20 5.6 162.85 -18.64 10.27
Numerical Study on Influence of Hydrofoil Clearance towards Total Drag Reduction on Winged Air Induction Pipe for Air Lubrication
165
Figure 4: Mesh Model (No. of Nodes = 201775; No. of
elements = 196401)
Figure 5: Enlarged View of Mesh for Part B (clearance =
20mm, angle of attack=20
)
2.3 Grid and Discretization
A sample of mesh model for the setup is shown in
Figure. 4. Quadrilateral is applied for the meshing
method as it gives more even separation for the air-
water interface of the domain. The value for the mesh
is done by trial and error to give the most appropriate
mesh model for the model. Because the mesh should
be evenly distributed around the model. Hence, edge
sizing and edge first layer thickness inflation also im-
plemented in the model to give narrower and more
even grid distribution towards the surface of the 2D-
model. The volume fraction of air is given as 1 for
the segment FGIJ, and 0 for the segment EFHI. In
this work, second order upwind scheme is used in all
calculation using a pressure-based computation. All
numerical simulation is done using transient solution.
The convergence criterion for numerical parameters
are all set to 10-3 for velocity, continuity, k, and ω.The
time step used in simulation is 0.001s with number of
time steps of 20 and 20 iterations.
The volume fraction of air is given as 1 for the seg-
ment FGIJ, and 0 for the segment EFHI. In this work,
second order upwind scheme is used in all calcula-
tion using a pressure-based computation. All numer-
ical simulation is done using transient solution. The
convergence criterion for numerical parameters are all
set to 10-3 for velocity, continuity, k, and ω The time
step used in simulation is 0.001s with number of time
Figure 6: Two Dimensional Representation of Air-water In-
terface on Computational Domain
Table 2: Numerical Data
Type θ
D
b
(N) D
b
(N) %R Error
Bare Plate Bare 173.48 - - 4.41%
WAIP 12 155.24 -18.24 10.51 7.46%
WAIP 16 156.63 -16.86 9.72 6.08%
WAIP 20 177.53 4.05 -2.33 9.02%
Error 6.74%
steps of 20 and 20 iterations. The numerical simula-
tion of the same geometry is adopted and similar sim-
ulations were performed for the validation purposes.
Since some of vital data such as plate draught, and the
length of afterbody were not reported in the paper, the
validation is somewhat approximate and shows error
of 6.74%. From the result obtained, it is concluded
that Volume of Fluid model and k-ω SST turbulence
model are the appropriate computational model for
this case of problem.
3 RESULTS AND DISCUSSIONS
The drag on the part C is showed in Table 3, 4,
and 5 for each configuration of clearance of 10 mm,
15 mm, 20 mm respectively, and angle of attack of the
hydrofoil of 12
, 16
, and 20
, where F
p
, F
v
, F
t
, C
p
are
pressure drag, viscous drag, total drag, and coefficient
of pressure respectively.
A visualization of total drag experienced by part C
is shown in Figure. 7 for each clearance of 10 mm, 15
mm, and 20 mm respectively. The simulation is con-
ducted without performing air injection from the hull
pipe in the WAIP device to proof the critical velocity
in (1) could produce the phenomenon of air entrain-
ment due to the negative pressure produced by the hy-
drofoil. The critical velocity of 5.6 m/s is used in the
simulation according to previous experiment (Kuma-
gai et al., 2010). The result of total drag from part
Table 3: Pressure and Viscous Drag of Part C (clearance =
10 mm; Re = 1.49 x 10
7
)
Clearance 10 mm
θ
F
v
F
t
F
0
v
F
0
t
12 97.03 97.03 158.41 158.41
16 97.64 97.64 159.42 159.42
20 96.80 96.80 158.04 158.04
SENTA 2018 - The 3rd International Conference on Marine Technology
166
C shows the distribution of the obtained data has no
particular tendency. On the clearance of 10 mm, the
smallest value of total drag is obtained in the angle
of attack of 20
which has a very small difference in
value compares to the 12
, and reach the maximum
value at 16
. Otherwise, on the clearance of 20 mm,
the minimum value is obtained at 16
.
Table 4: Pressure and Viscous Drag of Part C (clearance =
20 mm; Re = 1.49 x 10
7
Clearance 20 mm
θ
F
v
F
t
F
0
v
F
0
t
12 105.11 105.11 171.61 171.61
16 104.37 104.37 170.40 170.40
20 105.03 105.03 171.48 171.48
Table 5: Pressure and viscous drag of part C (clearance =
15 mm; Re = 1.49 x 10
7
)
Clearance 15 mm
θ
F
v
F
t
F
0
v
F
0
t
12 95.09 95.09 155.24 155.24
16 95.93 95.93 156.63 156.63
20 108.74 108.74 177.53 177.53
Table 6: Pressure and Viscous Drag of Part C (clearance =
10 mm; Re = 1.49 x 10
7
Clearance 10 mm
θ
F
p
F
v
F
t
C
p
F
0
v
F
0
t
12 49.13 3.41 52.54 80.22 5.57 85.79
16 74.85 3.18 78.04 122.21 5.19 127.41
20 91.41 2.60 94.01 149.24 4.25 153.49
However, on the clearance of 15 mm, the mag-
nitude of total drag increased drastically. This phe-
nomenon caused by the flow of the fluid that occurs
due to the presence of hydrofoil. The fluid flow oc-
curs on the downstream side of the hydrofoil is dif-
ficult to be mapped as a function. Hydrofoil always
has a tendency to experienced larger value of drag as
the angle of attack increased (Ockfen and Matveev,
2009). However, in this case study is performed to
analyze the effect produced by the hydrofoil towards
the total drag experienced on part C, the results could
not explain explicitly whether the data has a tendency
towards particular trends. A large number of varia-
tions of angle of attack is needed to really find the
exact plot of the phenomena. In the previous work
(Kumagai et al., 2011) also said that the fundamental
flow physics concerning this facility has not been clar-
ified yet because extremely complicated phenomena,
which are the free-surface effect of the hydrofoil.
However, from the result above it is can be ob-
tained that the 10 mm clearance gives the smallest
value of mean total drag. The zero value of the pres-
sure drag of the part C is due to the flow stream does
not cross the plate, instead the flow is parallel to the
plate. Thus, the plate only experienced viscous drag
due to the viscosity of the fluid. On the WAIP (part
B), the total drag is obtained by summing the pressure
drag and viscous drag experienced by the hydrofoil.
The drag component of each clearance of 10 mm, 15
mm, and 20 mm respectively is shown in Table 6, 7
and 8.
Table 7: Pressure and viscous drag of part C (clearance =
15 mm; Re = 1.49 x 10
7
Clearance 15 mm
θ
F
p
F
v
F
t
C
p
F
0
v
F
0
t
12 43.33 3.31 46.64 70.74 5.41 76.15
16 71.48 3.36 74.83 116.70 5.48 122.18
20 192.79 1.82 194.61 314.75 2.97 317.73
Table 8: Pressure and viscous drag of part C (clearance =
20 mm; Re = 1.49 x 10
7
Clearance 20 mm
θ
F
p
F
v
F
t
C
p
F
0
v
F
0
t
12 78.66 2.64 81.30 128.42 4.31 132.73
16 133.53 2.04 135.57 218.01 3.32 221.34
20 195.03 1.59 196.62 318.41 2.60 321.01
Table 9: Drag Force on Bottom Plate of The Model
Clear. θ
Do Db Dt Db
- bare 174.76 173.48 348.24 -
-
10mm
12 85.79 158.41 244.20 -15.07
16 127.41 159.42 286.82 -14.06
20 153.49 158.04 311.54 -15.44
15mm
12 76.15 155.24 231.39 -18.24
16 122.18 156.63 278.80 -16.86
20 465.42 177.53 642.95 4.05
20mm
12 132.73 171.61 304.34 -1.88
16 221.34 170.40 391.74 -3.08
20 321.01 171.48 492.49 -2.00
The hydrofoil experienced pressure drag due to its
surface intersecting the fluid’s streamline. Thus, the
molecule impacts the surface of the hydrofoil and re-
sulting non-zero value of pressure drag. The visual-
ization of the influence of clearance and angle of at-
tack on pressure coefficient of the hydrofoil is shown
in Figure 8. As can be seen on Figure 8, the pres-
sure coefficient of the hydrofoil has a tendency to in-
crease as the angle of attack increase. As theoretically
expected, the drag force increase as the angle of at-
tack is increased (Ockfen and Matveev, 2009). The
increased of angle of attack resulting in increase of
angle from intersection line between the hydrofoil’s
chord line and the fluid’s streamline. As a result, a
larger amount of energy to deflect the stream is re-
quired. The energy is obtained from kinetic energy
which in this case is the moving fluid. On the actual
Numerical Study on Influence of Hydrofoil Clearance towards Total Drag Reduction on Winged Air Induction Pipe for Air Lubrication
167
Figure 7: Total Drag Experienced by Part C
Figure 8: Pressure coefficient of the hydrofoil
condition, the model is moving through the water re-
sulting the kinetic energy is produced by the object
instead of the fluid. Thus, the energy loss is experi-
enced by the model. However, since the study is to
analyze the drag experienced by part C, further anal-
ysis of the hydrofoil’s drag is not performed.
A contour plot of dynamic pressure around part B
is shown in Figure. 9. As can be seen the pressure gets
Figure 9: Contour of dynamic pressure on part B
Figure 10: Drag reduction of part C
lower in the hull pipe above the hydrofoil. The phe-
nomenon is called the negative pressure, which the
pressure is lower compares to its surrounding. The
negative pressure is produced as the fluid moves to-
wards the model. The hydrofoil produces lower pres-
sure on its upper surface due to higher velocity caused
by longer curvature. experienced by part C for each
configuration of the model, where Do, is the hydrofoil
total drag. On the clearance of 10 mm 8-9% of drag
reduction is obtained. However, on the clearance of
15 mm, there is an added drag of 2% on angle of at-
tack of 20
. In this case the data is referred as outlier
or data that is outside the trend line (Mittnik et al.,
2001). Outlier is a common data in data distribution
that is not follow the normal distribution. This can be
caused by error from numerical computation that has
been performed where the mesh element relatively, is
not evenly distributed. As a result, the computational
result has a large value of error. The distribution of
mesh is depended on the element size of the mesh.
The element size has to be adjusted to fit the model
set up regarding the distance between two or more
surface/edge of the model. Generally, as the element
of the mesh get smaller the more accurate the result
gathered. However, this is could not be corresponded
that the finest possible gives the best result. Thus,
the grid independency test has to be performed in fu-
ture work. But as the error of the numerical data is
less than 10 percent the result obtained is somewhat
approximate regarding to the validation performed.
In some modelling, to reach solution’s grid indepen-
dency, numerical value in the computational domain
(the sizing parameter of mesh element) should be
much smaller than corresponding local value of the
model so that the numerical error could be minimized
(Wang and Zhai, 2012).
SENTA 2018 - The 3rd International Conference on Marine Technology
168
4 CONCLUSIONS
Computational Fluid Dynamics approach to estimate
the drag reduction by air lubrication using Winged Air
Induction Pipe (WAIP) is performed in the present
study and reasonably validated with experimental
works. By using nine configurations to achieve the
effect of hydrofoil clearance towards the drag reduc-
tion it is concluded that: the magnitude of drag reduc-
tion can be achieved when the contributing parameter
which are the angle of attack and hydrofoil clearance
chose at their optimum range. The optimum range
is achieved by modification of the parameter using
trial and error method. The modification of hydro-
foil clearance of the WAIP does not give a data trend
to a certain way. The application of WAIP gives result
of net drag reduction up to 10%. Figure 10 shows the
value of the drag reduction for each configuration of
the model. The clearance of the hydrofoil gives a sig-
nificant influence for the drag reduction. However, the
value of the drag reduction has no particular tendency
towards certain point. Therefore, the appropriate de-
sign is obtained by using trial and error method. This
is due to the unique flow characteristic produce by the
hydrofoil interacts with the plate in part C in different
ways depend on the clearance between hydrofoil
ACKNOWLEDGEMENTS
Authors are thanks to Department of Mechanical En-
gineering, Faculty of Engineering, Universitas In-
donesia for making facility available and also grant
PITTA No. 2561/UN2.R3.1/HKP05.00/2018
REFERENCES
Cui, Z., Fan, J. M., and Park, A. H. (2003). Drag coeffi-
cients for a settling sphere with microbubble drag reduc-
tion effects. Powder Technology, pages 132–134.
Duncan, J. H. (1981). An Experimental Investigation of
Breaking Waves Produced by a Towed Hydrofoil. Pro-
ceedings of the Royal Society of London. Series A, Math-
ematical and Physical Sciences, 377(1770):331–348.
Evans, G. M., Jameson, G. J., and Atkinson, B. W. (1992).
Prediction of the bubble size generated by a plunging liq-
uid jet bubble column. Chemical Engineering Science,
47:3265–3272.
Kodama, Y., Kakugawa, A., Takahashi, T., and Kawashima,
H. (2000). Experimental study on microbubbles and their
applicability to ships for skin friction reduction. In In-
ternational Journal of Heat and Fluid Flow, pages 582–
588.
Kumagai, I., Kushida, T., Oyabu, K., Tasaka, Y., and Murai,
Y. (2011). FLOW BEHAVIOR AROUND A HYDRO-
FOIL CLOSE TO A FREE SURFACE. Vis Mech Proc
Visualization of Mechanical Processes, 1(4).
Kumagai, I., Nakamura, N., Murai, Y., Tasaka, Y., Takeda,
Y., and Takahashi, Y. (2010). A New Power-saving De-
vice for Air Bubble Generation: Hydrofoil Air Pump for
Ship Drag Reduction. Proceedings of International Con-
ference on Ship Drag Reduction (SMOOTH-SHIPS).
Kumagai, I., Takahashi, Y., and Murai, Y. (2015). Power-
saving device for air bubble generation using a hydrofoil
to reduce ship drag: Theory, experiments, and applica-
tion to ships. Ocean Engineering, 95:183–195.
Menter, F. R. (1994). Two-equation eddy-viscosity turbu-
lence models for engineering applications. AIAA Jour-
nal, 32:1598–1605.
Mittnik, S., Rachev, S. T., and Samorodnitsky, G. (2001).
The distribution of test statistics for outlier detection
in heavy-tailed samples. Mathematical and Computer
Modelling, 34:1171–1183.
Mohanarangam, K., Cheung, S. C., Tu, J. Y., and Chen,
L. (2009). Numerical simulation of micro-bubble drag
reduction using population balance model. Ocean Engi-
neering, 36:863–872.
Muratoglu, A. and Yuce, M. I. (2015). Performance Anal-
ysis of Hydrokinetic Turbine Blade Sections. Technical
report.
Muste, M., Yu, K., Fujita, I., and Ettema, R. (2009). Two-
phase flow insights into open-channel flows with sus-
pended particles of different densities. Environmental
Fluid Mechanics, pages 161–186.
Ockfen, A. E. and Matveev, K. I. (2009). Aerodynamic
characteristics of NACA 4412 airfoil section with flap in
extreme ground effect. International Journal of Naval
Architecture and Ocean Engineering, 1:1–12.
Pang, M. J., Wei, J. J., and Yu, B. (2014). Numerical study
on modulation of microbubbles on turbulence frictional
drag in a horizontal channel. Ocean Engineering, 81:58–
68.
Shereena, S. G., Vengadesan, S., Idichandyxs, V. G., and
Bhattacharyya, S. K. (2013). CFD study of drag reduc-
tion in axisymmetric underwater vehicles using air jets.
Engineering Applications of Computational Fluid Me-
chanics, 7:193–209.
Toffoli, A., Babanin, A., Onorato, M., and Waseda, T.
(2010). Maximum steepness of oceanic waves: field
and laboratory experiments. Geophysics Research Let-
ters, 37.
Wang, H. and Zhai, Z. J. (2012). Analyzing grid indepen-
dency and numerical viscosity of computational fluid dy-
namics for indoor environment applications. Building
and Environment, 52:107–118.
Yanuar, Gunawan, Sunaryo, and Jamaluddin, A. (2012).
Micro-bubble drag reduction on a high speed vessel
model. Journal of Marine Science and Application,
11:301–304.
Numerical Study on Influence of Hydrofoil Clearance towards Total Drag Reduction on Winged Air Induction Pipe for Air Lubrication
169