You are on page 1of 7

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

Numerical Study of Fuel Tank Sloshing Phenomenon in


Automobile Vehicles Using CFD Code
R. Thundil Karuppa Raj
School of Mechanical
and Building Sciences,
VIT University, Vellore,
Tamil Nadu, India.
632014
thundil.rajagopal@vit.ac.in

T. Bageerathan
School of Mechanical
and Building Sciences,
VIT University, Vellore,
Tamil Nadu, India.
632014
bageerathan.t@gmail.com

Abstract The basic phenomenon in an automobile is


acceleration and deceleration. The fuel in the tank flows
irregularly with a splashing sound. This phenomenon is called as
sloshing. Due to sloshing the vehicle may lose its stability.
Sloshing noise is also a factor in this regard. This bothers the
passengers as the fluid in the tank is gasoline or diesel. This work
mainly concentrates on suppressing the sloshing noise in the fuel
tank through baffles. The tank is made up of High Density
Polyethylene (HDPE) in which the full height baffles cannot be
constructed. This work mainly focuses on the simulation of slosh
experiments to analyze the baffle design. Solid works 2011 is
used to model the fuel tank, discretized using Ansys ICEM CFD
12.Ansys CFX 12 is used to solving the problem. The turbulent
kinetic energy, velocity of water and force are calculated and the
fuel tanks are analyzed. The baffles reduced the sloshing inside
the tank to a great extent.

Keywords Fuel tank, Baffle, Automobile, Computational


Fluid Dynamics, Sloshing.
I. INTRODUCTION
The phenomenon of acceleration and braking generally
occurs in the automobiles. So, the vehicle is subjected to
inertia effects. Sloshing occurs due to this and unpleasant
noise is also generated. Gasoline or diesel is the fluid in the
fuel tank which bothers the occupants. Some measures has to
be taken to reduce the turbulent kinetic energy and other
relative parameters. The flow of the fluid has to be controlled
in order to have them reduced. In early days steel fuel tanks
were used where full height baffles are used to reduce
sloshing. In recent days High Density Polyethylene (HDPE)
tanks are used as they are light in weight and durable. In
HDPE tanks, full height baffles are difficult to be constructed
as they are manufactured by blow molding process. So the
baffles are designed such that they can be constructed in the
HDPE tanks and as well as they reduce sloshing. The works
mainly focuses on the design of baffle in the fuel tank and

N. Kulasekharan
Saveetha Engineering College,
Saveetha Nagar, Thandalam,
Chennai, Tamilnadu, India.
602105
sekarannk@gmail.com

analyze it. We considered a fuel tank model and computer


simulations will be used to validate the design.
II. LITERATURE REVIEW
M. Eswaran et al., [1] analyzed sloshing waves in baffled
and un-baffled tanks. They carried numerical simulations
based on the volume of fluid (VOF) technique with arbitraryLagrangian-Eulerian formulation. The response of the coupled
system is obtained using the ADINA software. They also
compared numerical data with experimental data which were
found to be agreeable. Hakan Akyldz et al., [2] investigated
liquid sloshing in a cylindrical tank with various fill levels and
ring baffles under roll motions. The main work of them is to
verify the relative effectiveness of baffle arrangements. They
have done experimental setup to study non-linear behavior
and the damping characteristics of the liquid sloshing. They
found that the ring baffles are found to be effective in
reducing the sloshing loads. Hakan Akyildz et al., [3] carried
out numerical and experimental work on pressure variations
and three-dimensional effects on liquid sloshing loads in a
partially filled rectangular tank. VOF technique is used to
study characteristics. Computational data were compared with
the experimental data to access the accuracy. They also
studied configured many baffled and no-baffle tanks are
studied.
P.K.Panigrahy et al., [4] carried out series of experiments
in a developed liquid slosh setup and estimated the pressure
development on tank walls and free surface displacement of
water from the mean static level. They also conducted the
experiments with and without baffles and the parametric
changes were observed. Mahesh Balthy et al., [5] carried out a
study to predict the dynamic behaviour of the fuel inside fuel
tank during the transient driving conditions using commercial
CFD code. They carried out an experimental study in actual
vehicle. The work is simulated and analyzed in CFD using the
inputs from actual testing and flow pressure distribution. Hoi

ISBN 978-1-4799-3158-3
419

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

Sum IU et al., [6] estimated kinetic energy and average


turbulent kinetic energy from the experimental and simulated
works. The fuel tank with different baffle design is analyzed
with different water levels. Noise values were also recorded
during experimental work. These values were used to compare
the experimental and simulated work.
Korang Modaressi-Tehrani al., [7] formulated and
analyzed three-dimensional nonlinear model of a partially
filled circular tank with and without baffles and derived the
pressure distribution over wetted tank surface. They analyzed
with 40% and 60% fill volumes and different types of the
baffles using FLUENT software. The results were found
satisfactory. Masashi Kamei et al., [8] conformed the
correlation of the slosh noise performance and factors related
to mechanism of slosh noise. They established a method to
predict the slosh noise in the early design phase. Kouji
Kamiya et al., [9] combined experiments with simulation to
establish a sloshing analysis technique. The experimental and
simulation works were validated. The fuel tank can be
simulated using coupled analysis between fluid and structure.
Won-Joo Roh, et al., [10] analyzed the flow of the fuel in the
tanks by Marker and Cell. They compared and verified by
experiments with transparent fuel tanks. The impact pressures
on the tank wall were analyzed to find noise source of the
wall. They also conducted a parametric study to decrease the
impact pressure by changing its dimension and shape.
Manuel J. Fabela-Gallegos [11] studied the dynamic
behavior of the heavy vehicle moving on roads. Baffles were
placed in the fuel tank to suppress sloshing. An experimental
study is also conducted on the fill level and number of baffles
in the fuel tank. Ben-Ren Tang [12] studied the phenomenon
of internal fuel and pressure distribution of the fuel using the
Arbitrary Lagrangian Eulerian methodology and improved the
fuel tank design. Three dimensional deformation modes of
fuel and tank were utilized to analyze the simulation results.
Nabih E. Bedewi and Tarek Omar [13] used Finite element
model utilizing the smoothed particle hydrodynamics (SPH)
algorithm in LS-DYNA to simulate the behavior of
automotive fuel tank in drop tests. The SPH algorithm is
applied to model a variety of fuel tanks as well as test
specimens. Fan Li et al., [14] used the combination of CFD
(Computational Fluid Dynamic), FE (Finite Element) and
Acoustic simulation methods, to evaluate the radiated fuel
tank slosh noise performance using CAE methods.

without hampering the slosh noise phenomenon.


Pierre De
Man and Jules-Joseph Van Schaftingen, [18] presented an
experimental validation of the source path- receiver approach
to slosh noise from a fuel tank on a commercially available
passenger vehicle. They correlated between predicted and real
noise, confirming the validity of the approach. Eric Frank et
al., [19] identified the current capabilities and discussed
optimal parameters of testing component level fuel slosh
noise, and explored the merits of various NVH analysis
methods that can be used to quantify slosh noise. Veera
Venkata Sunil Vytla and Yuya Ando [20] compared the force
acting on the fuel tank and found that they have same trend as
the mechanism that triggers the sloshing. They also compared
the deformation of the node located on one of the two tank
chambers and found that the deformation of a node on the
tank did not change much for the given braking scenario with
the changing water levels. G. R. Yan et al., [21] used transient
slosh approach to examine the effectiveness of the baffle
designs. They developed a computational fluid dynamic fluid
slosh model using volume of fluid technique coupled with a
Navier-Stokes solver. They analyzed fluid slosh responses in
terms of slosh frequency, resulting forces and moments under
different fill volumes and liquid cargos. They found that
partial baffle designs was effective.
III. CFD SIMULATION
A. Fuel tank geometry
The fuel tank dimensions are from Hoi Sum IU et al., [6]. The
dimensions are as shown in Figure 1.

Stephen Sibal et al., [15] presented both LS-Dyna's


Arbitrary Lagrange-Euler (ALE) and Abaqus' Coupled
Eulerian- Lagrange (CEL) methods for predicting the
structural performance of a fuel tank system and demonstrated
that a fuel tank systems and their components can be
numerically evaluated before the products release. Jong-Suh
Park et al., [16] verified the reliability of the FSI method and
suggested a new CAE analysis processes to predict fuel
sloshing noise. Mahesh Balthy et al., [17] evaluated the slosh
performance of plastic tanks, carried out a study to predict the
dynamic behaviour of the fuel inside a fuel tank during
transient driving conditions. The developed a new model

ISBN 978-1-4799-3158-3
420

Fig. 1 Fuel tank dimensions [6]

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

Fig. 2 Dimension of the baffle (Dimensions in mm)


Fig. 4 Initialization of model w/o baffle in CFX-Pre

B. Modelling and meshing


The tank is modelled in Solidworks 2011. The tank
model is discretized using Ansys ICEM CFD. The meshed
tank is shown in Figure 3. The tank model is initialized in
Ansys CFX Pre as shown in Figure 4 and 5. Ansys CFX is
used for simulation purposes. Water is used as the working
fluid. The entire model is discretized using hexahedral mesh
elements which are accurate and involve less computation
effort. Hexahedral mesh near the wall surface allows
capturing the boundary layer gradient accurately. Fine control
on the hexahedral mesh near the wall surface allows capturing
the boundary layer gradient accurately.

Fig. 5 Initialization of model with baffle in CFX-Pre

IV. GRID INDEPENDENCE STUDY


The grid independent study is carried out starting with
100000 nodes till 500000 nodes. The results is independent to
the number of grids after 400000 nodes. So 496000 nodes are
used to capture the boundary layer.
V. GOVERNING EQUATIONS
Fig. 3 Meshed Fuel Tank

The governing equations (equ 1, 2, 3 &4) are used to


simulate the 3D flow inside the fuel tank. Turbulence is taken
care by shear stress transport (SST) model.

ISBN 978-1-4799-3158-3
421

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

Conservation of mass( V ) 0 1

VII. CORRELATION ANALYSIS

yx

p xx

zx 2
x
x
y
z
p xy yy zy
y momentum

g 3
y
x
y
z
p xz yx zz
z momentum

4
z
x
y
z
x momentum

The Hoi Sum IU et al., [6] has not mentioned the number
of grids and turbulence model they used. So the correlation
analysis is carried out between the simulation results of Hoi
Sum IU et al., [6] and this work. The results are taken at the
time steps of 2 sec, 2.5 sec and 3 seconds. The correlation
factor is 0.97. This makes it visible that there is a high
correlation between these simulation values.

VI. BOUNDARY CONDITIONS


The fluid domains are defined in Ansys CFX Preprocessor. Here Shear stress Transport model is used as the
turbulence model. The initial conditions are as in Table I.

TABLE II
CORRELATION ANALYSIS RESULT OF TURBULENT KINETIC ENERGIES WITH
6.5 CM WATER LEVEL

TABLE I
INITIAL CONDITIONS AND VALUES

Initial conditions

Density of the fluid

Time
(Sec)

Corresponding values

2.0
2.5
3.0

0.001 kg/cm3, density of water

Number of fluid
Present

1, air is considered to be the void


region.

Gravity

981 cm/s2 downward along the zaxis

Time step size

0.0005 sec.

Duration of the
simulation

4 seconds

Fluid Temperature

293.0 K

Compressibility of
the fluid

Incompressible, Water is
considered to be an in
compressible fluid

Void region pressure

1013 kg/cm-s2, 1 atmosphere


pressure

Initial pressure field

Hydrostatic pressure in the zdirection

Displacement

-15.5[cm]*sin((pi*t)/(1[s])) cm

Turbulence model

Shear stress turbulence model


(SST)

Height of water

6.5 cm

Present
work
(cm2/s2)
895.261
762.382
646.666

VIII.

Hoi Sum
IU et al.,
(cm2/s2)
39.6
13.1
3.4

Correlation
Factor
0.975469

RESULTS AND DISCUSSION

Comparison between the No-baffle design and baffled


design are carried out using simulation results. The simulation
results of No-Baffle configuration is validated with the results
from the experiments conducted by Hoi Sum IU et al., [6] by
correlation analysis. The correlation factor is 0.975469 which
conforms the validity of the present model to the model of Hoi
Sum IU et al., [6]. The average values of turbulent kinetic
energy, velocity of water and force of the baffled model are
reduced when compared with the No-baffle configuration. The
figure 5, 6, 7 and 8 shows that the pressure is very high when
they are at the ends. The readings are taken from 1.25 to 1.55
seconds because the fuel moves from one end to other end. In
order to reduce the turbulent kinetic energy, baffles are used.
These baffles also reduces the force of water which is about to
hit the wall. The baffles are placed near the deep and shallow
end of the tank. So it decreases the average values of turbulent
kinetic energy, velocity and force at the time when the fuel
travels from one end and hits the other end of the tank. The
table III and table IV shows the values of the baffled model
and No-model configuration. The effect of baffle is
satisfactory when the fuel height is lesser than the baffles. The
efficiency of the baffle is null when the fluid level is higher
than the baffles height. Bottom mounted baffles reduce the
sloshing and other parameters to a great extent. Top mounted
baffles are not considered in this study as they are not
efficient. The air bubbles behind them try to escape and
generate sloshing noise [6]. Side mounted baffles are used in
this work. These side mounted baffles reduce the velocity.
The values of the baffled model is lesser when compared to
the no-baffle model.
The model is simulated in a computer with CORE 2DUO
2.2GHz processor 4GB RAM. The system took
approximately 120 hours to simulate the No-baffle model and

ISBN 978-1-4799-3158-3
422

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

280 hours approximately to simulate the baffled model. The


baffles height is 5cm. The width of the baffle doesnt
influence the sloshing [6]. Bottom mounted baffles separates
the fuel into sections and they are not continuous. The
turbulence created is only within the separated fuel section. So
the flow from one end to other end is supressed. Baffles have
openings in their portions that wont allow the fluid to stay
behind them which will be useful in reading the present fuel
level.

TABLE IV
PARAMETERS AND VALUES OF THE BAFFLED CONFIGURATION

Time
(sec)

Average
Turbulent
kinetic
energy
(cm/s2)

Average
Velocity
of water
(m/s)

Average
force (N)

0.5

75.431

0.576815

0.00269705

1.0

164.592

0.656029

0.00232842

1.5

291.159

0.609976

0.00192577

2.0

825.445

1.070730

0.00330000

2.5

597.987

0.516010

0.00134000

3.0

444.881

0.679411

0.00282506

Fig. 6 Time (sec) vs. Displacement (cm) graph

TABLE III
PARAMETERS AND VALUES OF THE NO-BAFFLE CONFIGURATION
Fig. 7 Turbulent kinetic energy vs. time

Time
(sec)

Average
Turbulent
kinetic
energy
(cm/s2)

Average
Velocity of
water
(m/s)

Average
force (N)

0.5

4939.00

1.4324

0.0029957

1.0

1617.15

0.8806

0.0025992

1.5

1000.68

0.5467

0.0034853

2.0
2.5
3.0

895.26
762.38
646.66

0.9009
0.6348
0.7602

0.0033333
0.0072200
0.0062500

Fig. 8 Average force vs. Time

ISBN 978-1-4799-3158-3
423

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

NOMENCLATURE
V
Velocity vector [-]
x, y, z position co-ordinates, [-]
g
gravitational force [ms-2]
p
pressure [pa]
Greek symbols

density, [kgm-3]

shear stress, [Nm-2]


REFERENCES
[1]

M. Eswaran, U.K. Saha , D. Maity, Effect of baffles on a partially


filled cubic tank: Numerical simulation and experimental validation,
Elsevier : Computers and Structures 87,pp.198205, December 2008.

[2]

Hakan Akyldz, N.Erdem Unal, Hafzullah Aksoy, An experimental


investigation of the effects of the ring baffles on liquid sloshing in a
rigid cylindrical tank, Elsevier: Ocean engineering 59, pp. 190-197,
January 2013.

[3]

Hakan Akyildz and N. Erdem Unal, Sloshing in a three-dimensional


rectangular tank: Numerical simulation and experimental validation,
Elsevier: Ocean engineering 33, pp. 2135-2149, March 2006.

[4]

P.K.Panigrahy, U.K.Saha, D.Maity,Experimental studies on sloshing


behavior due to horizontal movement of liquids in baffled tanks,
Elsevier: Ocean engineering 36, pp. 213-222, November 2008.

[5]

Mahesh Balthy, Chetan Prakash Jain, Mohan Makana and Raghavendra


Katti, Development of a Quantification Methodology for Slosh Noise
Associated with Dynamic Fuel Flow Behavior in Passenger Car
Tank, SAE TECHNICAL PAPER SERIES, 2011-28-0089, 2011.

[6]

Hoi Sum IU, W. L. Cleghorn and J. K. Mills, Design and Analysis of


Fuel Tank Baffles to reduce the Noise Generated From Fuel Sloshing,
SAE TECHNICAL PAPER SERIES, 2004-01-0403, 2004.

[7]

Korang Modaressi-Tehrani, S. Rakheja and I. Stiharu , Role of


Transverse Baffle Designs on Transient Three-Dimensional Liquid
Slosh in a Partly-Filled Circular Tank, SAE TECHNICAL PAPER
SERIES, 2005-01-3594, 2005.

[8]

Masashi Kamei, Junichi Hanai, Wataru Fukasawa and Takaomi


Makino, Establishment of a Method for Predicting and Confirming
Fuel Tank Sloshing Noise, SAE TECHNICAL PAPER SERIES,
2007-01-1538, 2007.

[9]

Kouji Kamiya and Yoshihisa Yamaguchi, Simulation Studies of


Sloshing in a Fuel Tank, SAE TECHNICAL PAPER SERIES,
2002-01-0574, 2002.

[10]

Won-Joo Roh, Sok-Hyun Cho and Jae In Park, Noise Reduction by


Decreasing Impact Pressure, SAE TECHNICAL PAPER SERIES,
2005-01-1913, 2005.

[11]

Manuel J. Fabela-Gallegos, Andres Ramirez-Valencia, Carlos A.


Favela-Gallegos, David Vazquez-Vega and Miguel Martinez-Madrid,
Experimental Assessment of Baffles and Their Effect on the
Longitudinal Sloshing Force in a Scaled Elliptical Tank, SAE
TECHNICAL PAPER SERIES, 2005-01-3577, 2005.

[12]

Ben-Ren Tang, Sarba Guha, Tau Tyan, Jiamaw Doong, Len Shaner,
Dilip Bhalsod, Simulation of Sloshing and Ballooning in Fuel Tanks
for High Speed Impacts, SAE TECHNICAL PAPER SERIES, 200601-0314, 2006.

[13]

Nabih E. Bedewi and Tarek Omar, Modeling of Automotive Fuel


Tanks Using Smoothed Particle Hydrodynamics, SAE
TECHNICAL PAPER SERIES, 2007-01-0682, 2007.

Fig. 9 Average velocity of water vs. Time

The average turbulent energy shown in Figure 7


shows that the value at 2 seconds is approximately similar.
This is because at that particular time the water flows from the
shallow end to the deep end. The baffle is placed at the deep
end region. So the flow is not restricted when it flows from
the shallow end region. So the baffle should be placed at the
two ends so that the flow is restricted when it flows from
either ends. The average velocity of water of the baffled
model at 2 seconds is more because the flow from the shallow
end to deep end, makes the fluid to flow above the height of
the baffle. The height of the baffle is 5 cm and the water
height is 6.5 cm. So the baffle height should be more than the
height of the water level to suppress sloshing.
The work arrives to the following conclusions:
1. The turbulent kinetic energy is high in the no-baffled
tank model.
2. The average turbulent kinetic energy is reduced up to
80.20% in the baffled model.
3. The average force is reduced up to 33.51% in the
baffled model.
4. The average velocity of water is reduced up to
55.23% in the baffled model.
5. The baffle height should be more than the height of
the water level to suppress sloshing.
So the implementation of the baffles reduces the
turbulent kinetic energy as similar to the case of full height
baffles in the steel fuel tank. Also the velocity and average
force of water is reduced to a great extent. So baffles of
height more than the height of water should be used to get
better efficiency.
ACKNOWLEDGEMENT
We wish to acknowledge the expertise of Mr. Balaji
Ranganathan from FL Smidth, CFD Division, and Chennai for
his valuable suggestion in carrying out this work.

ISBN 978-1-4799-3158-3
424

International Conference on Recent Advances in Mechanical Engineering and Interdisciplinary Developments [ICRAMID - 2014]

[14]

Fan Li,Stephen D. Sibal, Ian Francis McGann and Raphael Hallez,


Radiated Fuel Tank Slosh Noise Simulation, SAE TECHNICAL
PAPER SERIES, 2011-01-0495, 2011.

[15]

Peiran Ding, Fan Li and Stephen Sibal, An Integrated Fuel Tank


System Simulation, SAE TECHNICAL PAPER SERIES, 2011-010792, 2011.

[16]

Jong-Suh Park, Seung-Chan Choi and Seok-Gil Hong, The Prediction


of Fuel Sloshing Noise Based on Fluid-Structure Interaction Analysis,
SAE TECHNICAL PAPER SERIES, 2011-01-1695, 2011.

[17] Mahesh Balthy, Chetan Prakash Jain, Mohan Makana and Raghavendra
Katti , Development of a Quantification Methodology for Slosh Noise
Associated with Dynamic Fuel Flow Behavior in Passenger Car
Tank, SAE TECHNICAL PAPER SERIES, 2011-28-0089, 2011.
[18]

Pierre De Man and Jules-Joseph Van Schaftingen, Prediction of


Vehicle Fuel Tank Slosh Noise from Component-Level Test Data,
SAE TECHNICAL PAPER SERIES, 2012-01-0215, 2012.

[19]

Eric Frank, Chris Moon Jason Rae and Michael Popovich,


Optimization of Test Parameters and Analysis Methods for Fuel
Tank Slosh Noise, SAE TECHNICAL PAPER SERIES, 2013-011961, 2013.

[20]

[21]

Veera Venkata Sunil Vytla and Yuya Ando, Fluid Structure


Interaction Simulation of Fuel Tank Sloshing, SAE TECHNICAL
PAPER SERIES, 2013-01-0641, 2013.
G. R. Yan, S. Rakheja and K. Siddiqui, Baffle Design Analysis for a
Road Tanker: Transient Fluid Slosh Approach, SAE TECHNICAL
PAPER SERIES, 2008-01-2670.

ISBN 978-1-4799-3158-3
425