Original Manuscript

Journal of Vibration and Control

2023, Vol. 0(0) 1–10

Numerical investigation of structural © The Author(s) 2023

Article reuse guidelines:
sagepub.com/journals-permissions
models equipped with tuned liquid damper DOI: 10.1177/10775463231187693
journals.sagepub.com/home/jvc

Lidiya Bitew and Bedilu Habte

Abstract
Tuned liquid dampers (TLDs) are introduced in structural systems to suppress vibration of the structure due to dynamic actions.
Using a grid-based Finite Volume Method-Finite Element Method (FVM/FEM), efficiency of TLDs of cylindrical and rectangular
geometries was evaluated by performing a two-way Fluid-Structure Interaction (FSI) analysis. The nonlinear sloshing response of
fluid was simulated by employing a three-dimensional Navier–Stokes equation as the governing equation and occurrence of wave
breaking was accounted by using Volume of Fluid (VOF) method for free surface tracking. Utilizing the developed numerical
model, a series of three-dimensional transient analysis were conducted on a structure with and without TLDs aiming to
investigate effect of base excitation amplitude. The result show that maximum performance of TLDs at resonance is obtained
when the forcing ratio is limited to 0.1, and for this case, both cylindrical and rectangular TLDs displayed comparable efficiency.
Effect of excitation frequency fe was also investigated under a wide range of the forcing ratio showing that as fe increases, the
natural frequency of coupled TLD-structural system approaches the natural frequency of un-damped structural system.

Keywords
tuned liquid dampeds, fluid-structure interaction, volume of fluid, base excitation effect, finite-element/-volume,
rectangular/circular dampers

1. Introduction theories: potential flow theory, which assumes the flow to be

To reduce vibration-induced response of structures and incompressible, inviscid, and irrotational (Eswaran et al.,
make the structures serviceable, a number of passive vi- 2017; Faltinsen and Timokha, 2001; Fujino et al., 1992;
bration control mechanisms have been proposed by en- Novo, 2014; Tosaka and Sugino, 1991) and shallow-water
gineers (Soong and Dargush, 1997). Tuned liquid damper wave theory, which solves the nonlinear Navier–stokes
(TLD) is such a technology preferred for its economical and equation under the assumption of smaller wave height
operational advantages. TLD is liquid (usually water) filled compared to depth of liquid (Reed et al., 1998, 2019). When
tank that imparts indirect damping to the primary structure TLDs are subjected to larger amplitude base excitation,
through sloshing of liquid inside a container. During liquid inside tank will display larger deformation which is
sloshing, some portion of external force is dissipated characterized by free surface discontinuities and wave
through the action of viscous forces, such as wave breaking, breakings (Sun et al., 1992) and the models developed
free surface contamination, and liquid-boundary layer based on the above theories have limited applicability and
friction (Fujino et al., 1992), whereas some part of the load suffer in representing the actual sloshing response. Sun et al.
is counter balanced by the sloshing-induced base shear (1992) improved the later model by introducing two em-
force which is associated with the liquid impact on the tank pirical coefficients determined experimentally. Only wave
wall and is dependent on the developed wave height (Reed breakings due to steep waves were considered and
et al., 2019).
Container geometry affects sloshing response of liquid School of Civil and Environmental Engineering, Addis Ababa Institute of
inside container and its influence on the performance and Technology, Addis Ababa, Ethiopia
behavior of TLDs was investigated (Agresta et al., 2021; Received: 1 November 2022; revised: 17 April 2023; accepted: 26 June
Bhattacharjee et al., 2013; Sun et al., 1995). To assess effect 2023
of amplitude (A) and natural frequency fe, experimental,
Corresponding author:
analytical, and numerical approaches were used (Banerji Bedilu Habte, School of Civil & Environmental Engineering, Addis Ababa
et al., 2000; Fujino et al., 1988; Reed et al., 1998). Re- Institute of Technology, Addis Ababa, PO Box 385, Ethiopia.
searchers have developed models based on two main Email: [email protected]
2 Journal of Vibration and Control 0(0)

efficiency of TLDs determined using these models was was found to be sufficiently accurate (Ghaemmaghami et al.,
found to under-predict performances of TLDs subjected to 2012; Malekghasemi et al., 2015; Tuong and Huynh, 2020).
large amplitude harmonic motions (Banerji et al., 2000). The major aim of the present study is to investigate effect
TLDs are mostly assumed rigid and leave out influence of tank geometry, base excitation amplitude A, and exci-
of tank wall flexibility on the dynamic response. Some tation frequency fe on the efficiency of TLDs in reducing
recent studies indicate that flexibility of container affect the structural responses by developing a numerical model that
sloshing response as well as the developed hydrodynamic can (i) account all physical effects, (ii) accurately capture
pressure inside a tank (Gradinscak, 2009; Saghi et al., 2021; the three-dimensional nonlinear sloshing response, and (iii)
Seyoum, 2005). simulate two-way FSI. Unlike previous works, the present
Some researchers consider fluid-structure interaction study examines the structural frame and the TLD as an
(FSI) in order to ensure TLD’s efficiency and avoid tank integral unit offering a highlight on what effect the shape of
failure. A semi-experimental, real-time hybrid pseudo- the TLD container (rectangular vs cylindrical) would have
dynamic (PSD) study on effectiveness of TLD consider- on the response of a 3D framed structure.
ing FSI was conducted by Malekghasemi (2011).
Eswaran et al. (2017) developed a numerical model that
consider FSI phenomenon and studied floor wise performance 2. Model definition
of TLDs in mitigating vibration of multi-story structures. In this work, modal and time history analyses were con-
Although their model can achieve FSI, the transfer of data was ducted on a one-story structural model. The structural
limited to “liquid-to-structural domain,” not the other way, model is made up of steel with the properties given in
known as a one-way system coupling as shown in Figure 1(a). Table 1. Suitable dimensions for the models were obtained
FSI in TLD was considered when computing the through a series of modal analysis which correspond to
sloshing pressure acting on the tank wall. Marivani and a frequency mimicking a tall flexible building, with natural
Hamed (2014) conducted an FSI analysis and investigated frequency of fs ≈ 2 Hz. Through numerical experimentation,
effect of internal baffles on the inherent damping of TLDs. the dimensions of a one-story steel frame model given in
More recently, Finite Volume Method-Finite Element Table 2 were selected for this study.
Method (FVM/FEM) was employed to simulate the fluid The natural frequency of structural frame is computed to
and structural domain and study effective range of various be fs = 2.19, 2.19, and 3.65 Hz for x-axis, z-axis, and torsion,
parameters TLD-structure mass ratio, etc. The results were respectively.
verified by conducting experiments on a shaking table and Natural sloshing frequency of liquid inside the tanks was
tuned to the natural frequency of structure such that
a TLD-structure mass ratio of less or equal to 3% and TLD-
structure to TLD frequency ratio α is equal to unity (Fujino
structure frequency ratio (the ratio of the frequency of
et al., 1988, 1992; Reed et al., 1998). According to
excitation of the TLD and natural frequency of structure) Housner’s linear approximation method, equations (1) and
around 1.2 were determined to be optimum values (Ashasi- (2), the targeted natural sloshing frequency of liquid can be
Asokhabi, 2017; Malekghasemi et al., 2015). These studies attained by using different water depth ratios h/l; where h is
indicate that the FSI in the tank cannot be ignored, espe- liquid depth and l corresponds to dimension of TLD along
cially when designing relatively large TLDs. the direction of excitation (Housner, 1963). For this study,
The FVM/FEM is very suitable for simulating the liquid water depth ratio of h/l ≈ 0.2 is used.
sloshing of a TLD especially in the presence of breaking Studies on TLD performance recommended to use
wave, where the accuracy of a simplified model suffers. In a mass ratio μ, that is, ratio of liquid mass ml to structural
the case of low-amplitude ground motions, the FVM/FEM mass ms in a range less or equal to 3% (Ashasi-Asokhabi,

Figure 1. (a) One-way (b) Two-way FSI analysis approach (Tuong and Huynh, 2020).
Bitew and Habte 3

Table 1. Material property. where fl is natural sloshing frequency of liquid, h is liquid

depth, L is length of rectangular tank, D is diameter of
Description Symbol Value
cylindrical tank, and g is gravitational acceleration.
Steel modulus Es 210 GPa
Steel density ρs 7850 kg/m3
Steel poisson’s ratio υs 0.3 3. Time history analysis test cases
Acrylic density ρt 1000 kg/m3 External sinusoidal acceleration given by equation (3) was
Acrylic modulus Et 2.7 GPa applied to the structure system with and without TLD in x-
Acrylic poisson’s ratio υt 0.27 direction and the responses A and fe were examined. The
Water density ρ 1000 kg/m3 base excitation amplitude for the specified frequency range
is selected such that the forcing ratio (i.e., the ratio of vi-
bration amplitude of structure without TLD measured at the
Table 2. Dimensions of steel frame. installation point of damper to dimension of TLD along the
Description Symbol Value (mm) direction of excitation, uo/l) covers a wide range of am-
plitudes (Fujino et al., 1992; Reed et al., 1998; Sun et al.,
Slab width bs 650
1992).
Slab length ls 650 Accordingly, external sinusoidal acceleration loads with
Slab thickness ts 6.5 base excitation amplitude of A = 0.01 mm, 0.02 mm,
Frame height Hf 1040 0.05 mm, 0.1 mm, 0.25 mm, 0.5 mm, and 2.5 mm were
Column diameter dc 10 considered, where the structure without TLD at resonance
Column spacing Sc 520 vibrates at the steady-state amplitude uo of 1.02 mm,
2.04 mm, 5.11 mm, 10.21 mm, 25.53 mm, 51.18 mm, and
255.33 mm, respectively, resulting in a forcing ratio in the
2017; Malekghasemi, 2011). However, to achieve this
range of 0.008 < uo/l < 2.78.
optimal mass ratio for TLDs installed at tall structures,
considering the density of water and standard water depth X€ðtÞ ¼ we 2 Asinðwe tÞ (3)
ratio of 0.1 to 0.2, one would need larger space for TLD,
which is difficult to attain in view of space related issues.
we ¼ 2πfe (4)
Considering the above mentioned practical limitations,
a mass ratio around 1% is used for this study (i.e., μ = where X€ is external sinusoidal acceleration in x-direction,
1.12). we is circular excitation frequency, A base excitation am-
For the structural frame model, taking the above discussed plitude, t is time, and fe is external excitation frequency; this
values of water depth ratio and mass ratio, rectangular tank was done around the resonance, with frequency ratio β
with the size 92 mm × 160 mm × 120 mm (Length × Width × ranging from 0.8 to 1.2 with an increment of 0.1.
Height), and cylindrical tank of diameter 120 mm and height A total of 69 time history analysis were performed on the
170 mm were provided as TLD. To account for inertial effect models subjected to variable harmonic loads, and their
of container mass on the performance of TLD, mass of the efficiency is then evaluated by taking percentage reduction
tanks were made constant between the two geometries by of structure maximum top displacement and acceleration
using different tank height H. responses. Effect of excitation frequency on the TLD-
This study considers base excitation amplitude and structural system response is studied by plotting fre-
excitation frequency on efficiency of TLDs, keeping other quency response curve for the dimensionless frequency
parameters constant. Water depth of 18.4 mm and 24 mm ratio β and structural response in terms of displacement and
were provided for rectangular and cylindrical TLDs, re- acceleration.
spectively, with a natural sloshing frequency of 2.19 Hz
along x-direction.
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi 4. FVM/FEM analysis of TLD-structural
u
u
u3:16gtanh 3:16 Lh system
1 t
fl ¼ f or rectangular tank (1) For the simulation of TLD-structural system, ANSYS
2π L
software had been used, which contains mechanical and
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi fluent solvers that utilizes FEM and FVM in the workbench
u  
u environment. The grid-based FVM/FEM can be used to
u3:68gtanh 3:68 Dh
1 t track the sloshing liquid inside the tank and consider FSI. In
fl ¼ f or cylindrical tank (2) FVM/FEM, the solid domain (i.e., structural frame with
2π D
water tank) was meshed into elements which are
4 Journal of Vibration and Control 0(0)

represented by the governing equation of motion using (VOF) based on the concept of volume fraction was used.
FEM for their nodal displacement and acceleration results For every additional phase in the fluid domain, a variable
as shown in Figure 2. Whereas volume of fluid domain known as volume fraction F is introduced, representing the
inside TLD (including free space occupied by air) was volume occupied by each phase inside a computational cell
divided into a number of control volumes that have been between 0 and 1. For phases in each control volume, the
previously tested for grid independency using FVM method values of F are tracked throughout the domain by solving
as shown in Figure 2. continuity equation for volume fraction of phases, as
In order to account for nonlinear response of liquid motion, provided in equation (9).
a three-dimensional Navier–Stokes equation is employed as  
1 ∂ ∂ ∂ ∂
the governing fluid flow equation without any linearization ðFi ρi Þ þ ðFi ρi ui Þ þ ðFi ρi vi Þ þ ðFi ρi wi Þ ¼ 0
assumption. The general conservation of mass and momentum ρi ∂t ∂x ∂y ∂z
equations are expressed in equations (5) to (8). (9)

∂ρ ∂ðρuÞ ∂ðρvÞ ∂ðρwÞ X

n
þ þ þ ¼0 (5) Fi ¼ 1 (10)
∂t ∂x ∂y ∂z i¼1

where Fi is volume fraction at phase i, ρi is density at phase

∂ðρuÞ ∂ðρu2 Þ ∂ðρuvÞ ∂ðρuwÞ
þ þ þ i, ui,, vi and wi are velocity vector components at phase i in x,
∂t ∂x ∂y ∂z y, and z direction, respectively. The value of F is 0 for air, 1
 2  (6)
∂P ∂ u ∂2 u ∂2 u for water and 0 < F < 1 at the interface.
¼ þυ þ þ þ ρgx
∂x ∂x2 ∂y2 ∂z2 The velocity and density variables in equations (5) to (8)
are rewritten for each control volumes as given below and
∂ðρvÞ ∂ðρuvÞ ∂ðρv2 Þ ∂ðρvwÞ the model solves the modified governing equations of
þ þ þ multiphase fluid flow problems to give pressure and volume
∂t ∂x ∂y ∂z
 2 2  (7) fraction F results, which determine the wetted area for the
∂P ∂v ∂v ∂v
2
application of sloshing pressure generated in TLD at
¼ þυ þ þ þ ρgy
∂y ∂x2 ∂y2 ∂z2 a certain time instant.
X
n
∂ðρwÞ ∂ðρuwÞ ∂ðρvwÞ ∂ðρw2 Þ ρ¼ Fi ρi (11)
þ þ þ
∂t ∂x ∂y ∂z i¼1
 2  (8)
∂P ∂ w ∂ 2 w ∂2 w
¼ þυ þ 2 þ 2 þ ρgz X
n
∂z ∂x2 ∂y ∂z υ¼ Fi υi (12)
i¼1
where ρ is density of the fluid, P is fluid pressure, υ is fluid
viscosity, t is time, and u, v, and w are velocity vector where Fi, ρi , and υi , respectively, are volume fraction,
components in x, y, and z directions, respectively. density, and viscosity at phase i and n is number of phases
In order to account existence of multiple phases (i.e., present in the computational cell.
water and air) in the fluid domain and track the transient Once the governing equations (5) to (12) are integrated
water-air interface, a multiphase Volume of Fluid model over each control volumes, the solution variables are stored

Figure 2. Discretized structural and fluid mass for: (a) rectangular tank; (b) cylindrical tank.
Bitew and Habte 5

at the nodes, and Gauss’s divergence theorem is applied to analysis on a structural frame without TLD under a sinu-
transform these volume integrals in to surface integrals. soidal load given by equation (13). Top displacement of
In the TLD-Structure model, the interior walls of the structure without TLD obtained numerically was compared
water tank are all assigned as FSI faces. Fluid-structure with the experimental results as presented in Figure 5(a)
interaction is then achieved by coupling solution of the two which is in a good agreement.
problem domains (i.e., pressure and displacement) at the
interface through an exchange of solutions using a two-way X ðtÞ ¼ 0:5sinð2π2:07tÞ (13)
staggered coupling method and the iteration will continue where X is displacement (mm) and t is time.
up until solutions of the two domains have converged with The structure with TLD was modeled and analyzed in
in the coupling step. The transfer of solutions between the ANSYS workbench environment using the grid-based
two domains is controlled using ANSYS system coupling as FVM/FEM. The results obtained were compared with the
shown in Figure 3. experimental results presented in the work by Tuong and
Huynh (2020).
Figure 4(a) and Table 4 summarize numerical analysis
5. Model verification results with different element sizes and their percentage
Experimental test result presented by Tuong and Huynh difference from the experiment. The FVM/FEM method
(2020) are consulted, which was used to validate their with 2.5 mm element size of fluid domain produce results in
method of numerical analysis on the effectiveness of multi- good agreement with the experimental one as shown in
tuned liquid dampers (MTLD). The experiment was per- Figure 4(b), which demonstrate the validity of the numerical
formed on one-story steel moment resisting frame. To method used in this study for the transient analysis of
analyze the response of structure equipped with MTLD, structure with TLD.
both free vibration and forced vibration tests were con-
ducted on structural frame with and without MTLDs. Detail
characteristics of steel frame and MTLD used in the ex- 6. Results and discussion
periment are presented in Table 3. Effect of base excitation amplitude A, excitation frequency
A damping ratio of the structure, ζ, was computed from fe, and geometry on the behavior and efficiency of TLDs
the free vibration experimental result presented in (Tuong was investigated by performing time history analysis on
and Huynh, 2020). Using a logarithmic decrement method, structure with and without TLD subjected to small and large
damping ratio of structure ζ was determined to be 0.47% amplitude harmonic loads. To ensure numerical stability,
and the correctness of the value was later checked by time steps as low as 0.0001 were used and analysis is
comparing the experimental and numerical analysis results. continued until a steady-state condition is reached. Eval-
Figure 4(b) illustrates free vibration analysis results of steel uation of performance of TLDs was carried out by taking
frame without TLD, indicating a good agreement with the reduction of structural top displacement and acceleration
experimental result. upon the attachment of TLDs as criteria.
From modal analysis, the natural frequency of the
structural frame was determined to be fs = 2.10 Hz, which is
nearly equal to the value obtained from the shaking table
experiment (fs = 2.13 Hz). The validity of the numerical Table 3. Characteristics of steel frame and TLD
method is also checked by performing a time history
Description Symbol Value
Frame height Hf 1.10 m
Column spacing Sc 0.52 m
Slab thickness ts 5 mm
Column diameter dc 10 mm
Modulus of steel Es 210 GPa
Density of steel ρs 7850 kg/m3
Poisson’s ratio of steel υs 0.3
Width of TLD B 0.15 m
Length of TLD L 0.1 m
Density of acrylic ρt 1000 kg/m3
Modulus of acrylic Et 2.7 GPa
Poisson’s of acrylic υt 0.27
Density of water ρ 1000 kg/m3
Figure 3. Schematic view of two-way FSI analysis.
6 Journal of Vibration and Control 0(0)

Figure 4. Peak displacement without TLD: (a) under harmonic excitation; (b) free vibration.

Figure 5. Peak displacement with TLD (a) using FVM/FEM method (b) experimental and numerical analysis.

Table 4. Experimental and numerical analysis result. presented in Tables 5 and 6. As noted from Table 5, TLDs
performance depends on the base excitation amplitude and
FVM/FEM method
these dampers perform better when excited under smaller
Number of element
amplitudes compared to that of high amplitudes.
Structural response Experiment 115200 14400 1800 Another interesting phenomenon known as beating has
been observed from Figure 6 which is significant when A is
Top displacement (mm) 7.4 7.66 7.91 9.11
smaller. As A increases, the number of beating cycles re-
% Difference 3.48 6.96 23.11
duces and becomes insignificant when excited at A =
0.25 mm and dies out at A = 0.5 mm. This reduction of
beating phenomenon indicates the increasing of inherent
6.1. Observed effect of base excitation amplitude damping of TLD with the increase of excitation amplitude
A (Yalla and Kareem, 2001).
on TLD efficiency at resonance
A plot relating two dimensionless quantities, percentage
Displacement response histories for different cases are shown reduction of responses, and forcing ratio (uo/l), that is, the
in Figure 6. Considering peak structural displacement and ratio of amplitude at installation point of damper uo to size
acceleration responses for the different cases, percentage of TLD along the direction of motion l is illustrated in
reduction of structural responses were summarized as Figure 7.
Bitew and Habte 7

Figure 6. Time history of top displacement for amplitude A equals (a) 0.1 mm; (b) 0.5 mm.

Table 5. Maximum structural displacement (in mm) with and without TLD for β = 1
Structural displacement Percentage reduction
Base excitation amplitude A (mm) No TLD (uo) Cylindrical TLD Rectangular TLD Cylindrical TLD Rectangular TLD
0.01 1.02 0.10 0.10 89.86 90.69
0.02 2.04 0.20 0.18 90.43 91.12
0.05 5.11 0.51 0.48 90.01 90.51
0.1 10.21 0.97 0.92 90.51 91.02
0.25 25.53 3.48 3.20 86.38 87.45
0.5 51.18 11.79 8.11 76.96 84.15
2.5 255.33 116.99 101.23 54.18 60.35

Table 6. Maximum structural acceleration (mm/s2) with and without TLD for β = 1
Structural acceleration Percentage reduction
Base excitation amplitude A (mm) No TLD (uo) Cylindrical TLD Rectangular TLD Cylindrical TLD Rectangular TLD
0.01 194.05 19.94 18.04 89.72 90.70
0.02 388.37 37.55 34.84 90.33 91.03
0.05 970.92 97.32 93.01 89.98 90.42
0.1 1941.80 184.46 174.28 90.50 91.02
0.25 4854.60 638.93 601.48 86.84 87.61
0.5 9731.60 2254.50 1521.50 76.83 84.37
2.5 48546.00 21313.00 19159.00 56.10 60.53

As illustrated from Figure 7(a) and (b), efficiency of both surface with smaller wave heights as shown in Figure 8. For
cylindrical and rectangular TLDs depends on the forcing moderate excitation amplitudes (i.e., 0.022< uo/l < 0.1), the
ratio which in turn depends on A. Their efficiency is high liquid free surface developed smooth nonlinear waves with
when subjected to smaller excitation amplitudes. This ef- increased wave heights (Figure 9), which increases the
ficiency is maintained until forcing ratio reaches 0.1, at liquid-induced base shear force and make the efficiency of
which point a maximum efficiency was observed (90.43% TLDs unaffected with the increase of A. For cases with
and 91.12% for cylindrical and rectangular TLD, re- larger excitation amplitudes (uo/l > 0.1), clear bores and
spectively). However, as the forcing ratio increases beyond surface discontinuities were observed (Figure 10), and with
0.1, results showed high reduction in efficiency, since under further increase of A, wave breakings and turbulences at the
smaller excitation amplitudes (i.e., uo/l < 0.022), liquid wave crests were developed (Figure 11), which contributed
motion inside TLD displayed a continuous linear free to the inherent damping of the liquid (Fujino et al., 1992).
8 Journal of Vibration and Control 0(0)

Figure 7. Response reduction versus forcing ratio uo/l: (a) displacement; (b) acceleration.

Figure 8. Sloshing response inside cylindrical TLD for Figure 10. Sloshing response at 20 s for A = 0.25 mm for: (a)
A = 0.01 mm at: (a) 15 s; (b) 42.5 s. cylindrical; (b) rectangular TLD.

Figure 9. Sloshing response at 25 s for A = 0.05 mm for: (a) Figure 11. Sloshing at 14 s for A = 0.5 mm for: (a) cylindrical; (b)
cylindrical; (b) rectangular TLD. rectangular TLD

6.2. Comparison of rectangular and cylindrical then to 2.5 mm, damping efficiency of rectangular TLD was
higher compared to that of cylindrical TLD, as shown in
TLDs at resonance
Figure 6. Both cylindrical and rectangular TLDs have
As demonstrated in Figures 8 and 9, for A, from 0.01 mm to comparable efficiency in reducing structural responses
0.1 mm, cylindrical and rectangular TLDs displayed almost when amplitude is small (i.e., uo/l < 0.1). Whereas for
equal responses in terms of pattern and magnitude of re- structures under larger amplitude (i.e., uo/l > 0.1), efficiency
sponses. However, as A increases to 0.25 mm, 0.5 mm, and of rectangular TLD is higher than that of cylindrical TLD.
Bitew and Habte 9

Figure 12. Frequency response of top displacement (a) A = 0.1 (b) A = 0.5.

6.3. Observed effect of excitation frequency • TLDs perform well for lower excitation amplitudes
Effect of base excitation frequency fe on TLD performance showing up to 90% efficiency.
was studied through time history analysis on a structure • Increasing the base excitation amplitude so that the
with and without TLD subjected to harmonic load of forcing ratio is maintained within 0.1, the liquid-induced
constant amplitude and variable frequencies. Using the base shear force as well as inherent damping of TLD
maximum steady state response obtained from time history increase.
analysis, the corresponding frequency response curve for the • Both rectangular and cylindrical TLDs exhibit almost
dimensionless frequency ratio β and structural response is similar response reduction for lower forcing ratio. The
shown in Figure 12. As shown in the curves, maximum re- efficiency of TLDs reduce as the forcing ratio increases
duction of structural response was obtained when the structure beyond 0.1, with the rectangular TLD performing
is excited at a frequency equal to the natural frequency of the slightly better in this area.
un-damped structure (i.e., β = 1). Another thing evaluated from • For a better utilization of TLDs, the dimension along the
frequency response curve is the natural frequency of coupled direction of excitation should be proportioned such that
TLD-structural system. As shown in Figure 12, for smaller forcing ratio is below 0.1.
base, excitation amplitudes of A = 0.02 mm and 0.1 mm • Maximum reduction of structural response from TLDs is
maximum TLD-structural system response was recorded at obtained when the structure is excited at resonance.
β = 1.1, but as A increased to 0.5 mm and 2.5 mm, maximum • As base excitation amplitude increases, natural fre-
response is obtained at the β = 1. It can be concluded that for quency of the coupled TLD-structural system approaches
structures fitted with rectangular and cylindrical TLDs, the to the natural frequency of un-damped structure system.
natural frequency of coupled TLD-structural system ap-
proaches to the natural frequency of un-damped structure as This study can be extended further by considering the
the excitation amplitude increases. use of a nonlinear structural model as well as introducing
controllable screens and baffles for the development of
active TLDs, including investigation of multi degree of
7. Conclusion and recommendation system with varying TLDs settings.

Geometry of container and nature of external excitation force

are among the many factors that affect the behavior and ef- Acknowledgments
ficiency of TLDs in mitigating structural responses. A com- The authors would like to express their acknowledgement to the
bined FVM/FEM model was developed to account for School of Information Technology and Engineering (SiTE) at the
the nonlinear behavior of fluid sloshing and dynamic FSI. The Addis Ababa Institute of Technology (AAiT), and also to Dr
accuracy and robustness of the method was verified and the Abrham Gebre, Director of African Railway Center of Excellence
salient findings of the study are briefly summarized as follows: (ARCE) and all his colleagues for the facilitation and assistance
10 Journal of Vibration and Control 0(0)

they rendered in availing high performance computers of the center Gradinscak M (2009) Liquid Sloshing in Containers with Flexibility.
for this research work. Victoria University. Available at: https://vuir.vu.edu.au/15197/.
Housner GW (1963) The Dynamic Behaviour of Water Tanks.
Bulletin of the Seismological Society of America.
Declaration of conflict of interest Malekghasemi H (2011) Experimental and Analytical Inves-
The authors declared no potential conflicts of interest with respect tigations of Rectangular Tuned Liquid Dampers (TLDs).
to the research, authorship, and/or publication of this article. University of Toronto. Available at: https://tspace.library.
utoronto.ca/bitstream/1807/31329/6/Malekghasemi_Hadi_
201111_MASc_thesis.pdf.
Funding Malekghasemi H, Ashasi-Sorkhabi A, Ghaemmaghami AR, et al.
The authors received no financial support for the research, au- (2015) Experimental and numerical investigations of the dy-
thorship, and/or publication of this article. namic interaction of tuned liquid damper-structure systems.
Journal of Vibration and Control 21(14): 2707–2720.
ORCID iD Marivani M and Hamed MS (2014) Numerical study of slat screen
pattern effect on design parameters of tuned liquid dampers.
Bedilu Habte  https://orcid.org/0000-0002-8707-3514 Journal of Fluids Engineering 136(6).
Novo T, Varum H, Teixeira-Dias F, et al. (2014) Tuned liquid
References dampers simulation for earthquake response control of buildings.
Bulletin of Earthquake Engineering 12(2): 1007–1024.
Agresta A, Cavalagli N, Biscarini C, et al. (2021) Effect of bottom
geometry on the natural sloshing motion of water inside tanks: ReedYu DJ, YehGardarsson HS, Yeh H, et al. (1998) Investigation
an experimental analysis. Applied Sciences 11(2): 605. of tuned liquid dampers under large amplitude excitation.
Ashasi-Sorkhabi A, Malekghasemi H, Ghaemmaghami A, et al. Journal of Engineering Mechanics 124(4): 405–413.
(2017) Experimental investigations of tuned liquid damper- Reed D, Yeh H, Yu J, et al. (1998) “Tuned liquid dampers under
structure interactions in resonance considering multiple pa- large amplitude excitation,” Journal of Wind Engineering and
rameters. Journal of Sound and Vibration 388: 141–153. Industrial Aerodynamics, vol. 74–76, pp. 923–930, doi: 10
Banerji P, Murudi M, Shah AH, et al. (2000) Tuned liquid dampers .1016/S0167-6105(98)00084-1.
for controlling earthquake response of structures. Earthquake SaghiHirdaris RS, SaghiHirdaris H and Saghi H (2021) The in-
Engineering & Structural Dynamics 29(5): 587–602. fluence of flexible fluid structure interactions on sway induced
Bhattacharjee E, Halder L and Sharma RP (2013) An experimental tank sloshing dynamics. Engineering Analysis with Boundary
study on tuned liquid damper for mitigation of structural re- Elements 131: 206–217.
sponse. International Journal of Advanced Structural Engi- Seyoum S (2005) Dynamic Analysis of Liquid Containing Cy-
neering 5(1): 3–8. lindrical Tanks. Addis Ababa Institute of Technology. Avail-
Eswaran M, Athul S, Niraj P, et al. (2017) Tuned liquid dampers able at: http://etd.aau.edu.et/.
for multi-storey structure: numerical simulation using a parti- Soong TT and Dargush GF (1997) Passive Energy Dissipation
tioned FSI algorithm and experimental validation. S adhana Systems in Structural Engineering. Chichester: Wiley.
42(4): 449–465. Sun LM, Fujino Y, Pacheco BM, et al. (1992) Modelling of tuned
Faltinsen OM and Timokha AN (2001) An adaptive multimodal liquid damper (TLD). Journal of Wind Engineering and In-
approach to nonlinear sloshing in a rectangular tank. Journal of dustrial Aerodynamics 43(1–3): 1883–1894.
Fluid Mechanics 432: 167–200. SunFujino LMY, ChaiseriPacheco PBM, Chaiseri P, et al. (1995) The
Fujino Y, Pacheco BM, Chaiseri P, et al. (1988) “Parametric properties of tuned liquid dampers using a TMD analogy.
studies on tuned liquid damper (TLD) using circular containers Earthquake Engineering & Structural Dynamics 24(7): 967–976.
by free-oscillation experiments,” Doboku Gakkai Ronbunshu, Tosaka N and Sugino R (1991) Boundary element analysis of non-
vol. 1988, pp. 177–187. linear liquid motion in two-dimensional containers. Boundary
Fujino Y, Sun L, Pacheco BM, et al. (1992) Tuned liquid damper Integral Methods. Springer.
(TLD) for suppressing horizontal motion of structures. Journal Tuong BPD and Huynh PD (2020) Experimental test and nu-
of Engineering Mechanics 118(10): 2017–2030. merical analysis of a structure equipped with a multi-tuned
Ghaemmaghami AR, Kianoush MR and Mardukhi J (2012) liquid damper subjected to dynamic loading. International
Numerical Study on Annular Tuned Liquid Dampers for Journal of Structural Stability and Dynamics 20(07): 2050075.
Controlling the Response of Wind Towers Subjected to Seismic Yalla SK, Kareem A and Kareem A (2001) Beat phenomenon in
Loads. Computer-Aided Civil and Infrastructure Engineering. combined structure-liquid damper systems. Engineering
doi: 10.1111/j.1467-8667.2012.00772.x. Structures 23(6): 622–630.