Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

Published 05/18/2017
Copyright © 2017 SAE International
doi:10.4271/2017-01-9678
saepcmech.saejournals.org

An Experimental Study of Sloshing Noise in a Partially Filled Rectangular Tank

G Agawane, Varun Jadon, Venkatesham Balide, and R Banerjee
Indian Institute of Technology Hyderabad

ABSTRACT
Liquid sloshing noise from an automotive fuel tank is becoming increasingly important during frequent accelerating/decelerating
driving conditions. It is becoming more apparent due to significant decrease in other noise sources in a vehicle, particularly in hybrid
vehicles. As a step toward understanding the dynamics of liquid sloshing and noise generation mechanism, an experimental study was
performed in a partially filled rectangular tank. A systematic study was performed to understand the effects of critical parameters like
fill level and acceleration/deceleration magnitude. Response parameters like dynamic pressure, dynamic force, dynamic acceleration
and sound pressure levels along with high speed video images were recorded. The proposed experimental setup was able to
demonstrate major events leading to sloshing noise generation. These events in the sloshing mechanism have been analysed from the
dynamic sensor data and correlated with high speed video images.

CITATION: Agawane, G., Jadon, V., Balide, V., and Banerjee, R., "An Experimental Study of Sloshing Noise in a Partially Filled
Rectangular Tank," SAE Int. J. Passeng. Cars - Mech. Syst. 10(2):2017, doi:10.4271/2017-01-9678.

INTRODUCTION baffles on sloshing dynamics has been studied both using xperimental
and numerical techniques by several groups. Akyildiz and "U" ¨nal
Liquid sloshing can be defined as the movement of liquid free-surface
[16] studied sloshing in a rectangular tank with varying exciting
inside a partially filled container due to any external disturbance [1].
parameters to analyse the effects of baffles and fill levels on impact
Sloshing has been an important area of research in various
pressure and fluid flow inside tanks. Liu and Lin [4] and Cho and Lee
engineering fields like transportation of liquid by all major modes of
[17] numerically investigated the effect of baffle height on sloshing
transportation [2, 3, 4, 5, 6, 7], safety of structures against seismic
dynamics. Rebouillat and Liksonov [18] reviewed different numerical
loading [8, 9], vehicle stability for aerospace application [10, 11], etc.
techniques that are being used to study sloshing dynamics and the
More recently, slosh noise generation under sudden acceleration/
associated fluid-structure interaction.
deceleration condition is getting increased research attention for
premium and hybrid vehicles [12, 13]. In automobiles, sloshing noise
As mentioned before, effect of liquid sloshing in an automotive fuel
in a fuel tank can be radiated as structure borne noise and air borne
tank is gaining importance as it contributes to overall vehicle noise
noise. Structure borne noise is generated due to wall vibration
level. Noise generation and radiation within a vehicle due to liquid
(Fluid-structure interaction) and Air borne noise is radiated due to
sloshing inside a tank is quite complex. Researchers till recently have
fluid-fluid interaction inside the tank. The generated vibrations
typically relied on experimental measurements of such noise
propagate through mounting attachments to the passenger cabin. As a
characterization [19, 20]. Wachowski et al [13] concluded on the
result of advancements in automotive technology, noise radiation
basis of experiments that sloshing noise can be classified into three
from engine, driveline, tires, brakes, and wind sources has
categories: Splash, Hit and Clonk noise and the three were
considerably reduced. Hence sloshing noise has become as one of the
differentiated on the basis of their dominant frequency range. Kamei
major noise source in a vehicle for particular operating conditions.
et al [21] experimentally determined a correlation that relates fuel
tank, body parts and tank mounting structure to slosh noise. The tank
A number of experimental studies has been performed to investigate
mounting structure, which is related to both the fuel tank and the
sloshing phenomenon in tanks. Ibrahim et al [14] have reviewed
body parts has the largest contribution to sloshing noise.
several engineering applications where liquid sloshing dynamics is an
important engineering consideration. Theories based on non-linear
However, with the advent of powerful computational tools,
dynamics is being extensively used to understand the complex
researchers are now trying to develop computational based
behaviour during sloshing. However, a majority of studies are related
methodologies for noise estimation which can then be used in early
to work towards understanding dynamic loading and stability of
design phase of a fuel tank. Wiesche [12] performed experiments and
structures. Since this review, several other studies have been reported
numerical simulations of sloshing in automotive fuel tanks. He
in open literature on sloshing dynamics. Lugni et al [15] performed
reported a correlation between slosh noise and dynamic pressure
experiments to study the behaviour of flip-through events generated
fluctuations. Additionally, he used two-phase Computational Fluid
upon a vertical rigid wall when subjected by impact waves. Effect of
Dynamics (CFD) to track liquid interface within realistic tank
391
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

392 Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017)

geometries to determine these pressure fluctuations. Park et al [22] would travel. Special care was taken to minimize lateral movement of
have studied a vibration of automotive floor panels due to sloshing the vehicle. The track was maintained at a horizontal position with
loads using Fluid - Structure Interaction (FSI) analysis. Vaishnav et al respect to the ground with the help of a spirit level. A transparent
[23] used Eulerian based Computational Fluid Dynamics (CFD) and rectangular tank made of Acrylic with the dimensions of 0.238 (L) ×
Arbitrary Lagrangian-Eulerian (ALE) based vibrational analysis to 0.220 (W) × 0.238 (H) m3 was fabricated to perform the experiments
perform FSI studies of fuel tank sloshing. Similarly, Marriott et al in this study. The tank wall was 6 mm thick and it was rigidly placed
[24] have used FSI based computational methods to study fuel tank on the platform to avoid any relative motion. The tank and carriage
sloshing behaviour. together is referred as vehicle in rest of this paper.

As can be seen from the above literature survey, most of the research A three-axis linear inertia acceleration sensor (3g-ADXL335) and a
on automotive slosh noise generation is restricted to test conditions, line triggering sensor were mounted on the wooden platform. The
which are very empirical in nature and therefore, do not clearly line sensor was used to trigger the dynamic sensors, which include
explain the mechanism of slosh noise generation and its propagation. the dynamic pressure, force, acceleration and microphones.
There is a need to develop a multi-physics model to address the Additionally, a high speed camera, which was used to record the
above lacuna. Recently, Jadon et al [25] performed an integrated liquid motion inside the tank was also triggered using this line sensor.
study of liquid sloshing behaviour, its impact on structural vibration
and noise radiation on the rectangular tank that was subjected to a The vehicle was accelerated when a dead weight falls under gravity.
sudden deceleration. They performed both experimental and The vehicle was tethered to this dead weight by a string and pulley
computational study on this geometry. The current work is an mechanism. A band brake was used to apply the brake when the
extension of that study, and an in-depth analysis has been performed vehicle approached the end of its traverse. The liquid inside the tank
to understand various factors affecting noise generation and its sloshes due to the sudden application of this brake. Different sensors
radiation from a partially filled rectangular tank. Experiments have are mounted on the vehicle to record the effect of this sloshing
been performed under controlled laboratory conditions. An impulse activity. Sensor specifications are given in Table 1. Data from
based dynamic loading experimental rig was designed for this study. dynamic pressure sensors was acquired using HBM DAQ Quantum
Parametric studies were conducted to understand the effect of fill MX410. Data from all other sensors were acquired using a NI
level and vehicle deceleration on slosh noise generation. As this is the cDAQ-9178 data acquisition system. The line sensor was activated
first step in a comprehensive study to understand the mechanism of by a change in the colour of a strip mounted along the vehicle track,
slosh noise generation and its propagation, the authors have currently which in turn triggers other sensors. Phantom V12.1 high speed
restricted themselves in reporting their experimental work. Numerical camera was used to capture the liquid sloshing behaviour, which was
study is expected to be reported in near future. also triggered by the line sensor. Various camera settings were used to
acquire the video. The present work reports the images that were
obtained with a camera frame rate of 1000 fps. This frame recording
EXPERIMENTAL SETUP setting was chosen based on test duration and memory required for
online recording. Noise radiated due to liquid sloshing from the tank
Test Rig was monitored using four microphones that were placed in front, left,
right and top direction of the tank at a distance of 1 m from the final
resting position of the tank wall.

Table 1. Sensors Specifications

Figure 1. Schematic Diagram of Experimental Setup

Figure 1 shows the schematic diagram of the sloshing noise test rig,
which simulates the sloshing phenomenon under controlled braking
load to measure dynamic force and dynamic acceleration on the tank
wall. It also measures sloshing noise radiated from the tank. The
system consists of four major subsystems: (i) carriage and track, (ii)
rectangular tank with provision for sensor mounting, (iii) loading The acceleration and deceleration of vehicle due to the loading
mechanism and (iv) braking mechanism. The carriage consists of a mechanism and the band brake arrangement as described above is
wooden platform mounted on four polymer wheels designed for low estimated from Eq. (1a) and 1(b) respectively
rolling noise in order to minimize background noise. An aluminium
track of approximately 1.5 m was prepared on which the vehicle
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017) 393

RESULTS AND DISCUSSION

The current test setup was used to perform several parametric studies
(1a)
to understand the effect of tank fill level and deceleration magnitude
on slosh noise generation. The effects of these parameters on fluid
and
structure interaction of the tank was recorded using the dynamic
sensors; liquid slosh visualization using the high speed video camera;
and radiated sloshing noise using the microphones. A minimum of
three trials on at least two different days were performed to establish
(1b)
the repeatability of the current test methodology.
where, m1 is the mass of the tank with water, m2 is the mass of the
dead weight, a is the acceleration of the tank, d is the deceleration of
Background Noise Measurement
the tank, t1 is the time required to reach the required acceleration, t2 is
Before measuring the liquid sloshing noise, a separate test was
required time for deceleration, and μ is coefficient of friction.
performed to estimate the background noise due to vehicle motion on
the track. This background noise incorporates tyre noise, noise due to
Data Acquisition loading and braking mechanisms. Two different tests were performed
Figure 2 shows the mounting arrangement of the various sensors on to estimate the background noise: (a) tank filled with water and (b)
the vehicle. The dynamic force and dynamic acceleration sensors tank filled with sand of equivalent mass. Both these tests were
were attached on the external wall surface of the tank. The dynamic performed under the same loading conditions.
pressure sensors were mounted such that they were in direct contact
with the liquid inside the tank. The sensor mounting were so designed
to avoid liquid leakage from the tank.

Figure 3. Estimation of Vehicle Background Noise

Figure 3 shows the recorded sound pressure levels as recorded by the

front microphone. As can be seen from this figure, multiple peaks
were observed for the case where the tank is filled with water.
Figure 2. Tank dimensions, (b) sensors location for 20% fill, (c) sensors for However, only one peak is observed when the tank is filled with sand.
60% fill, (d) actual tank with sensor configuration In both cases, the first peak happens at the same time and occurs
before or at the time of braking. Hence, it can be concluded that the
Three dynamic pressure sensors were mounted at a height of 10% first peak is due to the background noise. Subsequent events
below the liquid free surface. These sensors were mounted on the corresponds to liquid sloshing and the effect of background noise is
front, back and right walls of the tank as shown in Figure 2. Another insignificant during these instances. High speed camera recordings
sensor was placed at 0.1 times of the tank height and was mounted on were not made during noise measurements due to significant
the tank front wall. The acceleration and forces were placed on the background noise contribution from its cooling fan.
front and back walls of the tank and are adjacent to the dynamic
pressure sensor that was placed 10% below the liquid free surface.
The high speed camera was placed normal to the track. The camera Experimental Results
field of view could capture the complete tank at the time of braking. Table 2. Set of Experiments shows the set of experiments that was
performed to understand the effect of fill level and deceleration
magnitude on slosh noise radiation. As mentioned previously, one
dynamic pressure sensor was placed at 0.1 times of the tank height in
all the cases and all remaining dynamic sensors were placed at 10%
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

394 Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017)

below the liquid free surface. Major events that were recorded in
inertial acceleration sensor and dynamic pressure sensors were
analysed using high speed camera images of that instant.

Table 2. Set of Experiments

Base Case
Figure 4 shows results from the base case, which is defined as fill
level of 40% and deceleration of 0.25 g. In this case the sensors were
placed at 30% of tank height. Figure 4 (a) shows the inertial
acceleration data in three directions. It was observed that the vertical
component of inertial acceleration sensor is within the range 9.8 ±0.2
m/s2 (1g) and therefore captures the acceleration due to gravity with
sufficient accuracy. Lateral acceleration is significantly lower
compared to the axial acceleration and therefore it can be assumed
that vehicle lateral movement is negligible.

Figure 5. High Speed Video Images at Different Time Instances from Base Case
Figure 4. Dynamic Measurements from Base Case, (a) Inertial acceleration,
(b) Dynamic force, (c) Dynamic acceleration, (d) Dynamic Pressure
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017) 395

duration. However, the dynamic signals were captured for longer

duration of time. The previous experiment was repeated for the same
fill level of 40% and deceleration of 0.25g. However, in this case the
dynamic sensors were placed at 10% of the tank height. The
corresponding dynamic sensor plots are shown in Figure 6. As can be
seen by comparing the dynamic plots from Figure 4 it is clearly
observed that the dynamic response is lower when the sensors are
mounted away from the liquid free surface. This is because most of
the dynamic events are restricted to a region close to the liquid free
surface. The dynamically active region varies significantly with fill
level and deceleration. More rigorous studies are required to identify
the dynamically active region.

a.

Figure 7. Spectrogram Analysis of SPL data from Base Case, (a) Inertial
Acceleration, (b) SPL Waveform, (c) SPL Spectrogram, (d) Impact Zone
Spectrogram

Spectrogram analysis has attracted attention for its ability to analyse

rapidly changing transient signals and it provides localized temporal
b. and frequency information in the signal. Figure 7 shows the
Spectrogram analysis for 40% fill level and 0.25g deceleration. Sound
Figure 6. Dynamic Measurements from 40% fill, sensor at 10% and 0.25g
Pressure Level (SPL) variation with respect to time acquired from the
deceleration, (a) Dynamic force, (b) Dynamic Pressure
top microphone is considered for Spectrogram analysis as shown in
Based on the high speed video images, it was observed that the liquid Figure 7(b). Based on the sound wave form, the total measurement
inside the tank impacts the front wall immediately after braking and duration is divided into three zones: (a) pre-braking zone, (b) impact
the corresponding image is shown in Figure 5. This impact event was zone and (c) linear zone. Impact zone starts at the instance of start of
recorded by all dynamic sensors. Dynamic sensors includes dynamic braking which is identified based on the sudden decrease in the
pressure, dynamic force, dynamic acceleration sensors and inertial acceleration data. Pre-braking zone is identified as the zone
microphones. Subsequently, the liquid moves towards the rear wall before this instance. Linear zone starts when the SPL values reduces
and it attains its maximum height along the rear wall at t=0.74 sec by approximately 10 dB with respect to the peak SPL value in the
after sensor triggering. Part of the liquid hits the tank roof and a Impact zone. It can be observed from the figures that the impact zone
corresponding peak in the recorded sound pressure level of top is confined to 2-3 sloshing periods. Spectrogram analysis is suitable
microphone is observed. Figure 5 shows the liquid position at this to classify different sloshing noise i.e. hitting noise, clonk noise and
moment of time. The fluid then again returns back towards the front splash noise [13]. It is observed that hitting noise is low frequency
wall and it becomes progressively more bubbly as observed in Figure noise and is significant till 800 Hz. First peak in impact zone as
5 at t=1.113 sec. These major events were well captured by the shown Figure 7(b) is associated with hit event on the back wall at 1.2
dynamic sensors. After two set of events, the sloshing phenomenon sec for 40% fill level and 0.25g deceleration. Subsequent hit event on
inside the tank transitions towards linear sloshing regime and is well the front wall occurs at 1.5 sec. As can be seen from the same figure,
supported by high speed camera images. Therefore, active duration of higher noise levels are confined to the active slosh duration and in the
the dynamic activities is typically restricted to two or three sloshing remaining time only background noise is recorded. Hence, acoustic
periods after application of brake. This is defined as the active slosh analysis should be performed for the active slosh duration. The SPL
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

396 Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017)

spectrogram shown in the Figure 7(c) is also divided into the three Transition Zone
zones as described above. Zone 3 does not show any major noise This type of nonlinearity arises due to oscillations of large amplitude
generation activity. Though zone 1 shows significant amount of noise in which the free liquid surface experiences non-planar motion. In
generation activity, however as mentioned previously, these noise dynamic pressure sensor data this region can be identified as region
generation is due to vehicle background noise. Hence, zone 2 has where the pressure value decays gradually from peak towards the
been analyzed thoroughly and therefore an exploded view is given in linear regime. The high speed video images show a clear liquid free
Figure 7(d). surface as seen in Figure 8. Similarly, the same figure shows the three
flow regimes as identified from a typical dynamic pressure data. The
Parametric Studies following parameter effects are discussed below:
An extensive parametric studies was performed for various fill levels,
deceleration magnitudes and sensor locations as given in Table 2. Set Effects of Fill Level & Deceleration Magnitude
of Experiments. Similar to the base case, measurements were recorded Figure 9 and Figure 10 shows the effect of fill level on peak values of
for dynamic sensors along with high speed video. Two different dynamic pressure and force at different deceleration level. The
sloshing regimes were observed during this parametric study: relative magnitude of the peak values at the front and rear walls show
a strong dependency on the fill level. The magnitude of peak dynamic
1. Linear flow Regime
pressure and force at the front and rear walls is almost same for 20%
2. Non-linear Flow Regime
fill level at 0.2g deceleration. However, when the fill level is 40%, the
front wall value is higher than the rear wall value. In contrast, the rear
Linear Flow Regime
wall value is higher than the front wall value in case of 60% fill level.
This regime is associated with small oscillations of fluid free surface In case of 0.25g deceleration, the relative magnitude trends are
in which the surface remains planar without rotation. This regime is similar to the previous case, but the difference between the front and
recorded in the form of smooth waves as seen in the high speed video rear wall values are subdued. Dynamic force plot also show the same
image in Figure 8. behaviour for different fill levels. However, for 0.3 g, the relative
trends are significantly different than the previous two cases. Based
Non-Linear Flow Regime
on this observation, fill level and deceleration magnitude have
This flow regime can be further classified into impact zone also non-linear relationship with dynamic pressure and force.
termed as strongly non-linear zone and transition zone or weakly
non-linear zone. This non-linear behaviour between deceleration magnitude and
dynamic parameters is due to the non-linearity of the fluid motion. As
can be seen from the high speed video images of Figure 11, at 20%
fill level, wave front impact takes place on both the walls. However,
in case of 40% fill level, liquid surges smoothly over the back wall
but it makes an impact on the front wall. The available ullage space
over the liquid surface is limited in case of 60% fill level. As can be
seen from the figure, the amount of liquid interacting with back wall
is higher than the front wall. Therefore, the impact force on the back
wall is higher. Similarly, as can be observed from the Figure 9, the
dynamic pressure on the rear wall is higher than the front wall for
20% and 60% fill level and deceleration of 0.25g and 0.3g. However,
the dynamic pressure at front wall dominates over the dynamic
pressure at back wall for 40% fill level. It must be noted that 80% fill
level was only tested for 0.25g deceleration. In this case, it was noted
that the dynamic activities are less compared to 60%. This is because,
the rigid mass, which is defined as the volume of fluid whose centre
of gravity does not change significantly relative to the vehicle centre
of gravity, is higher in case of 80% fill level and therefore
correspondingly has lower dynamic mass and this results in lower
Figure 8. Definition of Three Flow Regimes in Sloshing
dynamic activity. Hence, dynamic peak values of 80% is lower than
60% fill level. Hence it can be concluded from the above observations
Impact Zone
that the fluid interaction with tank walls may be smooth or impact
This non linearity is mainly due to rapid velocity changes associated dominated depending on the vehicle deceleration and fill level in a
with hydrodynamic pressure impacts of the liquid motion close to the non-linear fashion.
free surface. This zone is characterized by highly bubbly liquid
free-surface and rotational flow. Illustrative example of this flow Effect of Sensor Location
regime is given in Figure 8. This zone dominates for first few cycles Table 2. Set of Experiments summarizes the location of the sensors
of sloshing phenomenon and the number of cycles depend upon fill for the different experiments performed in the current study. This
level and deceleration value. arrangement was done to determine the effective liquid height for
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017) 397

dynamic events. Figure 12 shows the dynamic pressure data from

these two sets of sensors for different fill levels at 0.25 g deceleration.
The following observations can be made from this study:
• Other than the 20% fill case, the dynamic pressure observations
were significantly muted for the first set of sensors.
• Significant dynamic events can be observed for the second set of
sensors as they are closer to the free surface.

Figure 10. Effect of Fill Level on Dynamic Force for Various Deceleration(a)
0.2g (b) 0.25g (c) 0.3g

Figure 9. Effect of Fill Level on Dynamic Pressure for Various Deceleration

(a) 0.2g (b) 0.25g (c) 0.3g
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

398 Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017)

Sloshing Time Period

Time period of sloshing is different for impact and linear regime due
to the difference in the fluid flow characteristics in these two regimes.
Theoretical sloshing time period can be determined using the
following equation [1]:

(2)

where,

km = (2m - 1)π/L for asymmetric modes

km = 2mπ/L for symmetric modes

fs = sloshing natural frequency (Hz),

L = maximum dimension of base of the tank (m),

h =height of fluid filled in the tank (m),

g=acceleration due to gravity (m/s2)

Figure 13 show the comparison of experimentally determined time

period in the impact and the linear flow regimes along with
theoretical values as calculated from Eq. (2) for different fill levels at
a deceleration of 0.25g.The time period in impact regime is higher
Figure 11. High Speed Video Images for different Fill Level than the linear regime for all fill levels. However, the time period in
linear regime is close to theoretical time period. Hence, Eq. (2) is
valid only for linear sloshing regime.

Figure 12. Effect of Sensor Mounting Location on Dynamic Pressure

Fluid mass within the tank can be divided into two zone: (a) rigid mass
and (b) dynamic mass as given in ref [1]. The rigid mass is expected to Figure 13. Comparison of Experimental and Theoretical Sloshing Time Period
have very low relative velocity with respect to the tank and because of
this the dynamic activities observed by the first set of sensors are Spectrogram Analysis
muted. On the contrary, dynamic mass will have a high relative Figure 14 shows the SPL waveform and Spectrogram in the impact
velocity with respect to the tank and therefore the dynamic pressure zone for different fill levels at 0.25 g deceleration. It can be observed
sensor shows elevated dynamic activities. However, further that the SPL values in the pre-braking zone is associated with the
investigations are required to quantify these two regions within a tank. background noise due to the vehicle movement. The values of
background noise increases with increase in fill level due to increased
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017) 399

inertial effects. The peak SPL values in the impact zone decreases understanding sloshing noise generation mechanisms. All dynamic
with increase in fill level. At 20% fill level, two distinct peaks in the sensors synchronously showed large peaks during major events in
SPL data is observed. However, the values of the second peak relative sloshing. High speed image analysis with integrated sensor data
of the first one decreases with increase in fill level and at 60% and shows that major events in sloshing noise is restricted to active
80% fill level it becomes marginal. These peaks are associated with duration time period and effective liquid height.
the fluid impact on the tank walls. As can be observed from Figure
14, the decrease in SPL values corresponds to a similar decrease in The sloshing phenomenon inside the tank was classified as non-linear
the dynamic force values at 0.25g. The major noise generation due to and linear flow regime. Dynamic activities in the non-linear regime
impact is restricted to low frequency region. As can be observed from are the source of sloshing noise generation. Sloshing noise is a
the figure, higher noise level is confined between 200-600 Hz. These non-linear function of fill level and acceleration/deceleration.
observations are consistent with observations made by [13] and they Spectrogram analysis of microphone data shows that hit noise is
termed it as hit noise. dominant in a low-frequency region. Sloshing period for non-linear
and linear regime was measured from dynamic response data and
compared with analytical results. There is a good match between
analytical and experimental results in the linear regime. However,
further studies will be required to develop the validated sloshing
noise prediction model.

REFERENCES
1. Ibrahim, R.A., Liquid Sloshing Dynamics. 2005, Cambridge: Cambridge
University Press.
2. Delorme, L., Colagrossi A., Souto-Iglesias A., Zamora-Rodriguez R.,
and Botia-Vera E., A set of canonical problems in sloshing, Part I:
Pressure field in forced roll-comparison between experimental results
and SPH. Ocean Engineering, 2009. 36(2): p. 168-178.
3. Lee, D., Kim M., Kwon S., Kim J., and Lee Y., A parametric sensitivity
study on LNG tank sloshing loads by numerical simulations. Ocean
Engineering, 2007. 34(1): p. 3-9.
4. Liu, D. and Lin P., Three-dimensional liquid sloshing in a tank with
baffles. Ocean engineering, 2009. 36(2): p. 202-212.
5. Popov, G., Sankar S., Sankar T., and Vatistas G., Liquid sloshing in
rectangular road containers. Computers & fluids, 1992. 21(4): p. 551-569.
6. Wasfy, T.M., Wasfy H.M., and Peters J.M.. Coupled Multibody Dynamics
and Smoothed Particle Hydrodynamics for Predicting Liquid Sloshing for
Tanker Trucks. in ASME 2014 International Design Engineering Technical
Conferences and Computers and Information in Engineering Conference.
2014. American Society of Mechanical Engineers.
7. Wei, W., Junfeng L., and Tianshu W., Modal analysis of liquid sloshing
with different contact line boundary conditions using FEM. Journal of
Sound and Vibration, 2008. 317(3): p. 739-759.
8. Karamanos, S.A., Patkas L.A., and Platyrrachos M.A., Sloshing effects
on the seismic design of horizontal-cylindrical and spherical industrial
vessels. Journal of pressure vessel technology, 2006. 128(3): p. 328-340.
9. Shekari, M., Khaji N., and Ahmadi M., A coupled BE-FE study for
evaluation of seismically isolated cylindrical liquid storage tanks
considering fluid-structure interaction. Journal of Fluids and Structures,
2009. 25(3): p. 567-585.
10. Cheng, X.-d., Hu M.-z., and Wen J.-h., Dynamic characteristics of liquid
sloshing in a transversely vibrating spherical tank with a spacer under
low gravity. Journal of Hydrodynamics, Ser. B, 2008. 20(6): p. 762-769.
11. Veldman, A.E., Gerrits J., Luppes R., Helder J.A., and Vreeburg J., The
numerical simulation of liquid sloshing on board spacecraft. Journal of
Computational Physics, 2007. 224(1): p. 82-99.
12. aus der Wiesche, S., Noise due to sloshing within automotive fuel tanks.
Forschung im Ingenieurwesen, 2005. 70(1): p. 13-24.
Figure 14. SPL Waveform and Impact Zone Spectrogram for Different 13. Wachowski, C., Biermann J., and Schala R., Approaches to Analyse and
Predict Slosh Noise of Vehicle Fuel Tanks, in 24th International Conference
of Noise and Vibration Engineering (ISMA2010). 2010: Belgium.
CONCLUSIONS 14. Ibrahim, R.A., Pilipchuk V., and Ikeda T., Recent advances in liquid
sloshing dynamics. Applied Mechanics Reviews, 2001. 54(2): p. 133-199.
An experimental setup was developed to study liquid sloshing noise 15. Lugni, C., Brocchini M., and Faltinsen O.M., Wave impact loads: The role
generation and its propagation from a rectangular tank. Proper of the flip-through. Physics of Fluids (1994-present), 2006. 18(12): p. -.
instrumentation was done to record dynamic pressure, acceleration, 16. Akyildiz, H. and Ünal E., Experimental investigation of pressure
distribution on a rectangular tank due to the liquid sloshing. Ocean
force and sound pressure level along with high speed images. The Engineering, 2005. 32(11): p. 1503-1516.
current experimental setup has the capability to simulate hit and 17. Cho, J. and Lee H., Numerical study on liquid sloshing in baffled tank
splash noise for various fill levels and acceleration/deceleration by nonlinear finite element method. Computer methods in applied
conditions. Extensive parametric studies were conducted for mechanics and engineering, 2004. 193(23): p. 2581-2598.
 Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018

400 Agawane et al / SAE Int. J. Passeng. Cars - Mech. Syst. / Volume 10, Issue 2 (July 2017)

18. Rebouillat, S. and Liksonov D., Fluid-structure interaction in partially

filled liquid containers: a comparative review of numerical approaches.
CONTACT INFORMATION
Computers & Fluids, 2010. 39(5): p. 739-746. Contact:
19. De Man, P. and Van Schaftingen, J., "Prediction of Vehicle Fuel Tank
Slosh Noise from Component-Level Test Data," SAE Technical Paper Raja Banerjee
2012-01-0215, 2012, doi:10.4271/2012-01-0215. Department of Mechanical & Aerospace Engineering, Indian Institute of
20. Frank, E., Moon, C., Rae, J., and Popovich, M., "Optimization of Technology Hyderabad, Kandi, Sangareddy, Telangana, India 502285
Test Parameters and Analysis Methods for Fuel Tank Slosh Noise,"
SAE Int. J. Passeng. Cars - Mech. Syst. 6(2):1306-1312, 2013, [email protected]
doi:10.4271/2013-01-1961. Ph: +91-40-2301 6015
21. Kamei, M., Hanai, J., Fukasawa, W., and Makino, T., "Establishment
of a Method for Predicting and Confirming Fuel Tank Sloshing Noise,"
SAE Technical Paper 2007-01-1538, 2007, doi:10.4271/2007-01-1538. ACKNOWLEDGMENTS
22. Park, J., Choi, S., and Hong, S., "The Prediction of Fuel Sloshing Noise
Based on Fluid-Structure Interaction Analysis," SAE Int. J. Passeng. The authors would like to acknowledge the financial support of
Cars - Mech. Syst. 4(2):1304-1310, 2011, doi:10.4271/2011-01-1695. Mercedes Benz India Pvt. Ltd. for this work.
23. Vaishnav, D., Dong, M., Shah, M., Gomez, F. et al., "Investigation and
Development of Fuel Slosh CAE Methodologies," SAE Int. J. Passeng.
Cars - Mech. Syst. 7(1):278-288, 2014, doi:10.4271/2014-01-1632.
24. Marriott, D., Ohtomo, T., and Wako, T., "Complete Multi-Discipline
Simulation for Sloshing Noise," SAE Technical Paper 2015-01-0672,
2015, doi:10.4271/2015-01-0672.
25. Jadon, V., Agawane, G., Baghel, A., Balide, V. et al., "An Experimental
and Multiphysics Based Numerical Study to Predict Automotive Fuel
Tank Sloshing Noise," SAE Technical Paper 2014-01-0888, 2014,
doi:10.4271/2014-01-0888.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or
otherwise, without the prior written permission of SAE International.

Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the paper.