Ocean Engineering 167 (2018) 55–64

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Tuned liquid dampers with porous media T

∗
Wen-Huai Tsao, Wei-Shien Hwang
Department of Engineering Science and Ocean Engineering, National Taiwan University, Taiwan

A R T I C LE I N FO A B S T R A C T

Keywords: A tuned liquid damper (TLD) filled with porous media is developed to promote an additional damping force for
Porous media the conventional system. According to Darcy's law, the permeability of porous media and liquid viscosity form a
Tuned liquid damper damping force linearly proportional to fluid velocity. In this paper, an equivalent mechanical system considering
Equivalent mechanical system the damping due to porous media in rectangular TLD is established. In contrast to the other operating TLDs with
Potential flows
damping devices, the natural frequency and damping ratio of TLD in the present paper can be obtained without
using empirical formulas. The dynamic characteristics and performances are also discussed.

1. Introduction
In the past decades, the vibrational control technologies for struc-
tures have been deeply studied in civil engineering (Housner et al.,
1997). The tuned mass damper (TMD) is one of the most popular and
effective dynamic vibration absorber (DVA). It is easy to build, so it has
been widely adopted. However, TMD suffers some frequency-related
limitations. For instance, it could be inappropriate for aging structures,
because its mistuning could cause the reduction of performance. Be-
sides, TMD has limited effort during earthquakes (Housner et al., 1997).
Therefore, some modified TMD, such as active and semi-active TMD,
had been developed (Chang and Soong, 1980; Hrovat et al., 1983). A
tuned liquid damper (TLD), which is a tank filled with liquid (usually
water) in a structure, is known as a much more economical DVA. In the
early 1950s, it was originally designed for stabilizing vehicles and sa-
tellites in space (Graham and Rodriguez, 1952; Abramson, 1966).
Afterwards, TLD was applied to structures from 1980s (Bauer, 1984;
Modi and Welt, 1988) and soon became very attractive since it needed
relatively low manufacture costs and almost no maintenance required.
Nowadays, several high-rise buildings have adopted TLD systems
(Wakahara et al., 1992; Tamura et al., 1995). The applications of TLD
are also extended to different kinds of structures including bridges
(Chen et al., 2008) and offshore structures (Jin et al., 2007; Ha and
Cheong, 2016).
To understand the interaction between TLD and structures, the
dynamic response of TLD was studied in advance. Graham and
Rodriguez (1952) developed an equivalent mechanical system for liquid
sloshing based on the linear wave theory. Then Housner (1957) revised
the equivalent masses and their locations. Afterwards, this simple
analytic model has been widely adopted in practice (Veletsos, 1984;
Peek and Jennings, 1988; Dodge, 2000; Ibrahim et al., 2001; Ibrahim,
2005). Until recently, Li and Wang (2012) and Li et al. (2012) derived a
numerical fitting method and exact solutions for the equivalent linear
model of sloshing in a rectangular tank. On the other hand, the non-
linear wave problems became more and more important due to the
large liquid motion (Chwang and Wang, 1984). Many numerical
schemes have been developed to solve these sloshing problems such as
finite difference method (FDM) (Liu and Lin, 2008), finite element
method (FEM) (Okamoto and Kawahara, 1992), boundary element
method (BEM) (Amano et al., 1993), method of fundamental solution
(MFS) (Wu et al., 2016), and regularized boundary integral method
(RBIM) (Chen et al., 2017). Overall, both analytic and numerical
methods can offer the satisfactory predictions for the linear or nonlinear
sloshing problems.
The traditional fresh-water TLD is well known to be inferior to TMD
due to its insufficient damping. As a matter of fact, researchers have
focused on increasing the damping effect of TLD to achieve an optimal
control by additional devices, such as rough boundary (Fujino et al.,
1988), wall baffles (Chang et al., 1998; Ju et al., 2004; Zahrai et al.,
2012), vertical screens or baffles (Warnitchai and Pinkaew, 1998; Tait
et al., 2005; Tait, 2008; Tait and Deng, 2010; Faltinsen and Timokha,
2011; Maravani and Hamed, 2011; Crowley and Porter, 2012; Nayak
and Biswal, 2015; Kumar and Sinhamahapatra, 2016; Chen et al.,
2018), horizontal screens (Jin et al., 2014), bars (You et al., 2007), or
floating roof (Ruiz et al., 2016). In recent years, many researchers
prefer slotted screens owing to their convenience in installation. The
damping coefficient of a screen basically depends on its solidity ratio,
location, number, and tank dimensions as well as the response of liquid.
However, those parameters related to damping effect need to be de-
termined by experiments; therefore it somehow increases the difficulty

∗
Corresponding author.
E-mail address: wshwang@ntu.edu.tw (W.-S. Hwang).

https://doi.org/10.1016/j.oceaneng.2018.08.034
Received 1 February 2018; Received in revised form 15 August 2018; Accepted 19 August 2018
0029-8018/ © 2018 Elsevier Ltd. All rights reserved.
W.-H. Tsao, W.-S. Hwang Ocean Engineering 167 (2018) 55–64

during the design stage. In addition, the change of natural frequency consisting of fluid only, the porosity is defined as the fraction of the
induced by additional damping device is always ignored. If the damping total volume of the medium that is occupied by void space, and the
ratio has to be further increased by installing more screens or adjusting permeability is defined as a quantity measuring the influence of a
the screen solidity ratio, the natural frequency may be somewhat mis- substance on the fluid flux in the region it occupies. The second term on
tuned. the right-hand side of Eq. (2) resulting from Darcy's law implies the
Instead of screens, the porous media provides another choice to linear damping force proportional to the velocity of fluid relative to the
increase the damping of TLD. This idea originated from the application tank. Assuming the flow field is irrotational, the velocity potentials φ
of breakwater, which is developed to mitigate the response of ocean and ψ can be defined as:
wave. Huang applied Biot's theory of poroelasticity (Biot, 1962) to ⇀
U = ∇φ (3)
consider both inertial force and viscous damping for the problems of
breakwaters and porous wave-makers (Huang, 1991; Huang and Chao, ⇀
v = ∇ψ (4)
1992; Huang et al., 1993; Hsu et al., 2005). According to Darcy's law
⇀ ⇀
(Nield and Bejan, 2013), the resistant force resulted from the viscosity where U = γ u is the average of the fluid velocity over the total volume
and permeability is equivalent to a damping force linearly proportional of the medium. Due to the continuity of fluid and rigid-body excitation
to the fluid velocity. Abbaspour and Hassanabad (2010) verified this of tank, both velocity potentials satisfy the Laplace equation such as:
force decreased the dynamic reaction of liquid in storage tank very ∇2 φ = 0 (5)
rapidly. Therefore, it is more convenient to adjust the specific porosity
and permeability to achieve the required damping. ∇2 ψ =0 (6)
The analytic method for porous TLD system is developed in this
paper. The governing equations and boundary conditions are in-
2.2. Boundary conditions
troduced in Section 2; the equivalent mechanical model which takes the
damping effect into account is established in Section 3; several ex-
The kinematic and dynamic boundary conditions of the free surface
amples are verified and discussed in Section 4; and some conclusions
can be described as (Nield and Bejan, 2013):
are briefly drawn in Section 5.
⇀
DR
=⇀
u on the free surface
2. Sloshing in porous media Dt (7)

∂φ 1
2.1. Governing equations of liquid within porous tanks + ∇φ⋅∇φ + γα (φ − ψ) + g (z − h) = 0 on the free surface
∂t 2γ (8)
⇀ μ
The free surface motion in an isotropic porous media inside a rec- where R is the location of particle on the free surface, and α = κρ . Since
tangular tank subjected to a horizontal excitation along the x-direction the given excitation is horizontal and unidirectional, the impermeable
can be simplified as a two-dimensional sloshing problem. Assume this condition can be expressed as:
tank has length L and water depth h as shown in Fig. 1. The skeleton of
∂φ
porous media is rigid and fixed with the tank and water inside is in- (x , 0, t ) = 0 on the bottom
∂z (9)
compressible; the continuity and momentum equations of water within
the porous tank are (Nield and Bejan, 2013): ∂φ L
⎛± , z, t ⎞ = ⇀
v on the side walls
∂x ⎝ 2 ⎠ (10)
∇⋅⇀
u =0 (1)

∂⇀u γμ ⇀ ⇀ 3. Analytic solutions

ρ⎡ + (⇀
u ⋅∇) ⇀⎤
u = −∇P − ( u − v ) + ρ⇀
g
⎢
⎣ ∂t ⎥
⎦ κ (2)
3.1. Homogeneous solution of linear waves in rectangular porous TLD
in which ⇀u is the intrinsic fluid velocity, ⇀
v the velocity of tank, γ the
porosity, μ the dynamic viscosity, κ the permeability, P the pressure, ρ The linearized kinematic boundary condition on the free surface in
the fluid density, and g the gravitational acceleration. The intrinsic porous media can be simply obtained by substituting Eq. (3) into Eq. (7)
fluid velocity is defined as the fluid velocity over a volume element as (Huang et al., 1993):
∂η 1 ∂φ
=
∂t γ ∂z (11)
where η is the vertical deviation from the mean free surface. Combining
Eq. (11) and the linearized dynamic boundary condition, i.e. the line-
arized form of Eq. (8) (Beck, 1972), the velocity potential at z = h sa-
tisfies:
∂ 2φ ∂φ ∂ψ ⎞ g ∂φ
+ γα ⎛ − + =0
∂t 2 ⎝ ∂t ∂t ⎠ γ ∂z (12)
If only the relative motion between fluid and tank is considered, the
velocity potential of tank can be given as a constant (such as a sta-
tionary tank), therefore ∂ψ/ ∂t = 0 . Since the tank is only subjected to
lateral force, the even-mode fluid motions will vanish due to the anti-
symmetrical fluid motion. Hence the general solutions of fluid motions
from the Laplace equation are of the form:
φn (x , z , t ) = Dn sin an x cosh an z e sn t , n = 0,1,2... (13)
Fig. 1. Coordinate system and definition of length and filling depth of a rec- where Dn is a constant determined from the initial condition and
tangular tank. an = (2n + 1) π / L . Therefore, φn has to satisfy the constraint of Eq. (12)

56
W.-H. Tsao, W.-S. Hwang Ocean Engineering 167 (2018) 55–64

as: n 3
⎛ 2 + g an tanhan h ⎞⎟ D1n + γαωD2n = 4(−1) ω
⎜−ω
2
⎝ γ ⎠ an Lcoshan h (25)
g
Dn ⎡sn2 cosh an h + sn γα cosh an h + an sinh an h ⎤ sin an x e sn t = 0
⎢
⎣ γ ⎥
⎦ g
− γαωD1n + ⎜⎛−ω2 + an tanhan h ⎟⎞ D2n = 0
(14) ⎝ γ ⎠ (26)
Since all the values of x and t have to satisfy Eq. (14), the quantity g
Let ωn = a tanhan h
γ n
be the undamped natural frequencies,
inside the bracket must be zero. Hence,
pn = ωn2 − ω2 , and q = γαω, then the coefficients D1n and D2n can be
−γα γ 2α 2 g solved as:
sn = ± − an tanhan h
2 4 γ (15) pn ω3 4(−1)n
D1n =
In general, the damping device of a vibration absorber will be an pn2 + q2 an2 Lcoshan h (27)
underdamped system. Therefore, sn should be a complex number, α
should be small enough, and the permeability should satisfy: qω3 4(−1)n
D2n =
pn2 + q2 an2 Lcoshan h (28)
μγ γL π
κ> coth h
2ρ gπ L (16) Under the assumption of linear wave theory, the horizontal re-
sultant force F per unit width applied to the tank due to the side-wall
Then Eq. (15) is rewritten as: pressure can be obtained as:
h
−γα g γ 2α 2
sn =
2
±i
γ
an tanhan h −
4 (17)
F = γρ ∫ ⎡ ∂∂φt ⎛⎝ L2 , z, t ⎞⎠ − ∂φ −L
⎛
∂t ⎝ 2
, z , t ⎞ ⎤ dz = −γAρLhω2 sin ωt
⎠⎦
0 ⎣
∞
The nth mode of velocity potentials can be expressed as: 8 tanh an h
+ γAρ ∑ (−pn ω4 sin ωt + qω4 cos ωt )
−γαt n=0 an3 L pn2 + q2 (29)
φn (x , z , t ) = Dn sin an x cosh an z e 2 eiωdn t (18)

in which the frequency of the n th mode is defined as: 3.3. Equivalent mechanical system for rectangular porous TLD
g γ 2α 2
ωdn = an tanhan h − In this section, an equivalent mechanical system is constructed to
γ 4 (19) replace the porous TLD. In contrast to the traditional undamped system,
Eq. (18) also shows the decay of envelope for every mode is in the the damping resulted from porous media is discussed. This system as
same exponential rate. shown in Fig. 2 consists of a fixed mass mf which is attached to the tank
and infinite sets of moving masses mn , equivalent damping coefficients
3.2. Particular solution of linear waves in rectangular porous TLD cn and equivalent stiffness kn . This model is dynamically equivalent to
the sloshing motion in the porous tank; therefore it has to exert the
As a harmonic excitation is applied to the tank in the x-direction, its same force on the tank. When the tank is subjected to a surge motion
horizontal displacement relative to the ground x t can be expressed as: (horizontal excitation), the infinite sets of motion equations on this
system are given by:
x t (t ) = A sin ωt (20)
mn x¨n + cn x˙ n + kn x n = −mn x¨t (30)
where A is the amplitude of displacement. The velocity potential of the
where x n is the horizontal displacement of the moving masses relative
rigid skeleton of porous media is:
to the tank. From the viewpoint of structural vibration control, only the
ψ = Aωx cos ωt (21) steady-state response is considered when constructing the frequency
response curve. Therefore the particular solutions due to the external
The general solutions of the Laplace equation inside the tank can be
excitation as Eq. (20) are solved as:
expressed as:
∞ Aω2
xn = (p sin ωt − 2ξn ωn ω cos ωt )
φ = Aωx cos ωt + A ∑ (D1n cos ωt + D2n sin ωt )sin an x cosh an z pn2 + (2ξn ωn ω)2 n (31)
n=0

(22) where ωn = kn/ mn , and ξn = cn/2mn ωn denotes its damping ratio. The
resultant force F exerted by the equivalent system can be simply ob-
The first term on the right-hand side of Eq. (22) means the rigid- tained as (Clough and Penzien, 1993):
body motion of tank and the summation part represents the relative
∞ ∞
liquid motion in the tank. Substituting Eq. (22) into Eq. (12) gives the ⎛ ⎞ Amn ω4
F = ⎜mf + ∑ mn⎟ x¨g + ∑ (−pn sin ωt
following relation:
⎝ n=0 ⎠ n=0 pn2 + (2ξn ωn ω)2
∞
+ 2ξn ωn ω cos ωt ) (32)
− ω3cosωt⋅x + ∑ [(−ω2D1n + γαωD2n) cosωt + (−γαωD1n − ω2D2n) sinωt
n=0
Because the horizontal forces resulted from fluid and equivalent
g model are equal, the coefficients in Eqs. (29) and (32) must be the same.
+ (D1n cosωt + D2n sinωt ) an tanhan h⎤ sinan x coshan h = 0
γ ⎥ (23) In each corresponding modal response, the following relation should be
⎦
maintained as:
Since x is an odd function, its Fourier expansion can be expressed as
follows: pn ω4 8γρ p ω4
mn = tanhan h 2 n 2
∞
pn2 + (2ξn ωn ω) 2 3
an L pn + q (33)
4(−1)n
x= ∑ sinan x
2ξn ωn ω5
n=0
an2 L (24) 8γρ qω4
mn = tanhan h 2
pn2 + (2ξn ωn ω) 2 3
an L pn + q2 (34)
Substituting Eq. (24) into Eq. (23), two sets of equations must be
satisfied: The equivalent damping ratio of mode n is obtained from Eq. (34)

57
W.-H. Tsao, W.-S. Hwang Ocean Engineering 167 (2018) 55–64

Fig. 2. The equivalent mechanical system representing response of actual liquid in porous media.

divided by Eq. (33): substituting Eq. (35) into Eq. (33) as follows:

γα γμ γ 8γρ
ξn = = cothan h mn = tanhan h
2ωn 2κρ an g (35) an3 L (36)

Then the equivalent moving masses can be determined by The equivalent stiffness can be obtained by:

58
W.-H. Tsao, W.-S. Hwang Ocean Engineering 167 (2018) 55–64

kn = mn ωn2 =
8ρg
tanh2an h in Fig. 3 for two different amounts of water. The first-mode frequencies
an2 L (37) of both cases increase rapidly when γ < 0.4, which means they are
sensitive when small porosity is employed. Since TLD is usually de-
By comparing the rigid-body responses of Eq. (29) and Eq. (32), the
signed for a long-period structure to resist its gust loading, a shallow-
fixed mass can be obtained as:
water tank with large porosity would be preferred. If the tank is oc-
∞
cupied with only few solids like γ > 0.8, its first-mode frequency is
mf = γρLh − ∑ mn nearly the same as the first-mode frequency of the pure water tank. This
n=0 (38)
is the reason why most researchers ignore the frequency shift if only
in which γρLh denotes the total water mass per unit width. few damping devices are installed. However, when the porosity is be-
According to Eq. (11), the wave elevation can be expressed as: tween 0.4 and 0.8, to ascertain the natural frequency of the TLD will be
∞ necessary in application.
A
η (x , t ) = ∑ γω
(D1n sin ωt − D2n cos ωt ) an sin an x sinh an h
n=0 (39) For a case γ h = 0.1L, the damping ratios of porous tanks with dif-
Therefore wave elevation of the nth mode on the right-side wall ferent porosity and permeability in Eq. (35) are shown in Fig. 4. For
( x = L/2 ) in Eq. (39) becomes: the same permeability, the small porosity results in less damping.
For the same porosity, the damping ratio increases as the perme-
4 Aω2 ability decreases. However, the permeability should satisfy Eq. (16)
ηn = tanh an h 2 (p sin ωt − q cos ωt )
γan L pn + q2 n (40) to ensure the porous TLD be an underdamped system. Therefore,
large porosity and small permeability are usually preferred. The
By comparing Eq. (40) with Eq. (31), the modal wave elevations on
optimal damping ratio is generally about 5%–10% in practice, hence
the right-side wall can be expressed in terms of the modal equivalent
the available porosity and permeability would be within the area
displacement x n as:
surrounded by two curves of ξ0 = 5% and ξ0 = 10% as shown in
4 Fig. 4. By combining porosity and permeability, the natural fre-
ηn = tanhan h⋅x n
γan L (41) quency and the damping ratio of porous TLD can be adjusted in a
flexible way.
For the case, h = 0.1L, γ = 0.9 and κ = 2.5·10−6 m2 (α =0.4 s−1),
4. Examples and discussions the free-vibration frequencies of the first three modes are 3.22, 8.70
and 12.53 rad/s, respectively, and the first three equivalent
Since the oscillation frequencies and damping ratios have been damping ratios are 5.6%, 2.1% and 1.4%, respectively. To observe
derived in the last section, the dynamic characteristics of fluid inside the free-vibration response of sloshing inside the porous tank, the
rectangular porous tank will be discussed in this section. The perfor- free surface is set to be a stationary inclined line η (x , 0) = 0.02x as
mances of porous TLDs will also be analyzed by the equivalent me- the initial condition. Since the initial shape of free surface is an odd
chanical model. function, the wave elevation only consists of odd-mode shapes as
well. According to Eq. (11), this wave elevation is:
4.1. Example 1: dynamic characteristics of rectangular TLD with porous
∞
media 4(−1)n −αγt αγ
η (x , t ) = ∑ 0.02 2 2
sin an x e 2 ⎛ωdn cos ωdn t +
2
sin ωdn t ⎞
n=0
ωdn an L ⎝ ⎠
In this example, the dynamic viscosity of water μ = 10−3 Pa-s, and
(42)
density ρ = 1000 kg/m3 are used. The relation between the first-mode
natural frequency and porosity is expressed as Eq. (19), and it is shown The analytic results of wave elevations on the right wall (x = L/2)

Fig. 3. The relation between the first-mode natural frequency and porosity for different amounts of water.

59
W.-H. Tsao, W.-S. Hwang Ocean Engineering 167 (2018) 55–64

Fig. 4. The damping ratios of porous tanks with respect to the porosity and permeability for γ h = 0.1L.

Fig. 5. Wave elevations on the right-side wall of γ = 0.9 and α = 0.4 (1st mode; 。sum of first 5 modes; sum of first 20 modes).

are shown in Fig. 5. It shows just little difference in wave elevation x¨ x˙

⎡ ms + mf ⎤ ⎧ s ⎫ ⎡ cs + ∑ cn −c0 −cn ⎤ ⎧ s ⎫
between the first mode and the summation of the first 20 modes, while ⎢ m0 ⎥ ⎪ x¨0 ⎪ + ⎢ −c0 c0 ⎥ ⎪ x˙ 0 ⎪
the summation of the first 5 modes is nearly the same as that of the first ⎢ ⋱ ⎥⎨ ⋮ ⎬ ⎢ ⋱ ⎥⎨ ⋮ ⎬
20 modes. ⎢ mn ⎥ ⎪ ⎪ ⎢ −cn cn ⎥ ⎪ ⎪
⎣ ⎦ ⎩ x¨n ⎭ ⎣ ⎦ ⎩ x˙ n ⎭
Since the resultant force exerted by the equivalent system is related
to the equivalent masses as shown in Eq. (32), the percentage of ⎡ ks + ∑ k n − k 0 −kn ⎤ ⎧ xs ⎫ ⎧ f (t ) ⎫
⎢ −k 0 k 0
⎥⎪ x0 ⎪ ⎪ 0 ⎪
equivalent masses to the total water mass implies the contribution of +⎢ ⎥ ⋮ =
each mode. The ratios between the summation of the first N modal ⎢ ⋱ ⎥⎨ ⎬ ⎨ ⋮ ⎬
⎪ xn ⎪ ⎪
equivalent masses and the total water mass are shown in Fig. 6. It in- ⎢
⎣
− k n k n ⎥
⎦ ⎩ ⎭ ⎩ 0 ⎪ ⎭ (43)
creases with water depth due to larger percentage of fixed mass. For
shallow-water tanks, the ratios are still larger than 80% even only the where ms , cs , ks and xs are the mass, damping coefficient, stiffness and
first mode is considered. This gives the reason that the first-mode re- displacement of main structure, respectively, and f is the external force
sponse account for the major part of the response. In fact, more than applied to it. Since the energy dissipation mechanism of porous TLD is
99% contribution comes from the first five modes. similar to that of TMD, two parameters, tuned-frequency ratio ft and
optimal damping ratio of damper ξopt , should be particularly addressed.
The tuned-frequency is the particular frequency of external force that
4.2. Example 2: SDOF system with rectangular porous TLD subjected to makes the damper resonate out of phase with the structural motion,
harmonic forces while the optimal damping ratio results the structure in the least re-
sponse. When a harmonic force is applied on an undamped structure,
For a porous TLD attached to a single-degree-of-freedom (SDOF) the tuned-frequency and optimal damping ratios of TMD can be ex-
system as shown in Fig. 7(a), its equivalent model is shown in Fig. 7(b). pressed as following relations, respectively (Den Hartog, 1956):
The equation of motion for structure-porous TLD system can be sim-
1
plified as: ft =
1 + Rm (44)

60
W.-H. Tsao, W.-S. Hwang
Fig. 6. The ratios between the summation of the first N modal equivalent masses and the total water mass for different water depths.
3Rm
ξopt =
8(1 + Rm) (45)
where Rm is the ratio between the masses of TMD and main structure. It
is given as m 0 / ms in this example. In general, the mass ratio is given in
advance, so the tuned-frequency and damping ratios are determined
immediately. The optimal parameters for other different types of ex-
citation or damped structures can be referred to the text book (Soong
and Dargush, 1997). To analyze porous TLD by the equivalent model,
the tuned-frequency ratio and damping ratio may not be exactly the
same as the results of Eqs. (44) and (45), however, they are close.
This example is carried out for demonstration. Different values of
porosity, permeability and tank dimensions are used to show the re-
ducing performance of structural vibration. A SDOF system connected
with a rectangular porous TLD shown in Fig. 7(a) is subjected to a
harmonic force f (t ) = F0 sin ωt , where F0 = 2 N. The main structure
has ms = 400 kg, ks = 4130 N/m, and cs = 10 kg/s, therefore its natural
Fig. 7. An SDOF system with porous TLD: (a) Real problem, (b) Analytic model.
61
Ocean Engineering 167 (2018) 55–64
frequency ωs = 3.21 rad/s and damping ratio ξs = 0.4%. The forcing
frequency ratio β is defined as:
ω
β=
ωs (46)
The attached TLD has its length L = 100 cm and width B = 10 cm.
The total water mass m w = 8 kg. The density of porous media is
ρp = 1050 kg/m3, so the additional weight ρp (1 − γ ) LBh is taken into
account as part of main structure. The design procedure can be simply
addressed in the following steps:
1. As the total water mass is given, the relation of porosity and depth
can be determined from the curves in Fig. 8.
2. From every set of porosity and depth one can obtain its first-mode
equivalent mass, therefore the tuning frequency ratio of porous TLD
can be determined by Eq. (44).
3. The intersection point of two curves resulted from mass and natural
W.-H. Tsao, W.-S. Hwang Ocean Engineering 167 (2018) 55–64

Fig. 8. Total water mass and first-mode natural frequency of porous TLDs with respect to the porosity and water depth (L = 10B = 1 m).

Fig. 9. The first-mode equivalent damping ratio of porous TLDs with respect to γ / ω0 and permeability (L = 10B = 1 m).

frequency conditions as shown in Fig. 8 will indicate the expected
value of porosity and water depth. Therefore the porous TLD has its
first-mode natural frequency to match the tuning frequency.
4. The optimal damping ratio can be obtained by Eq. (45). To make the
first-mode damping ratio equals the optimal one, the permeability of
porous media can be easily found in Fig. 9 with respect to γ / ω0 .
In this way, the water depth, porosity, permeability can be solved.
For comparison, another tank with length L = 120 cm (while width
remains the same) is used to show the alternative design parameters
that can also gain optimal performance. All the parameters of both TLDs
are shown in Table 1.
In this example, the moving masses of the first five modes are
considered in the equivalent model for convenience. A TMD with the
same mass, which is 8 kg, is used for comparison. The normalized fre-
quency responses of steady-state structural displacements by two

Table 1
Parameters of two different optimal porous TLDs.
L (cm) B (cm) γ κ (10−6 m2) h (cm) m0 (kg) ω0 (rad/s) f0 (ω0 / ωs ) ξ0 (%) Rm0 (%) ft ξopt (%)

TLD-1 100 10 0.87 1.81 9.2 6.31 3.16 0.985 7.61 1.57 0.985 7.61
TLD-2 120 10 0.67 1.39 10.0 6.34 3.15 0.986 7.60 1.57 0.985 7.60

62
W.-H. Tsao, W.-S. Hwang
Fig. 10. Normalized structural displacements by two different optimal porous TLDs and TMD.
Fig. 11. Normalized wave elevations in two different porous TLDs.
optimal porous TLDs and TMD are shown in Fig. 10. Two peak values of
the frequency response by both porous TLDs are almost equivalent,
which implies the frequencies of porous TLDs are practically tuned to
that of the main structure. Therefore, to determine the tuned-frequency
ratio by giving Rm as the ratio between the first-mode moving mass and
the mass of main structure in Eq. (44) could be applicable. Overall, the
design procedure could be flexible and feasible in practice.
In Fig. 10, both the porous TLD and TMD show excellent ability in
vibrational control. These two TLDs are almost equally effective and
they can reduce the maximum response under 9% of that of the un-
controlled structure. The TMD is 13% better than the porous ones. This
could be inevitable because the effective mass in TLD is always less than
the mass of TMD when their total masses are equal. However, the
porous TLD is still worth to be studied and applied in practice because
of its lower cost.
The normalized frequency responses of the steady-state wave am-
plitude for both TLDs shown in Fig. 11 are no more than 0.15. The wave
elevation of small-porosity TLD (γ =0.67) is higher than that of large-
porosity one (γ =0.87), which can be explained by Eq. (41). Large wave
63
Ocean Engineering 167 (2018) 55–64
response may occur for small-porosity TLD.
5. Conclusions
In this paper, the equivalent mechanical system developed can
completely account for the linear dynamic response of fluid in a rec-
tangular TLD with porous media. The equivalent masses, frequencies
and damping ratios are derived so that the dynamic analysis for porous
TLD can be easily carried out without using any empirical formula. In
our cases, the summation of the first five modal responses would be
enough to replace the entire analytic solution of fluid. Furthermore, it
would be acceptable and convenient to determine the tuned-frequency
and optimal damping ratio of porous TLD according to the first-mode
equivalent mass.
In the application of vibration control, once the tuned-frequency and
optimal damping ratio of porous TLD are determined, the accurate de-
sign parameters can be guaranteed by giving appropriate combinations
of tank dimensions, porosity and permeability. Examples have shown the
porosity = 0.4–0.8 would be better but the permeability depends on the
W.-H. Tsao, W.-S. Hwang
tank dimensions, porosity and the objective damping ratio (i.e. the op-
timal damping ratio in general). In addition, the design procedure of
porous TLD has also been demonstrated. Therefore the design procedure
of porous TLD could be simple, flexible and feasible. The detail of its
practical application will be further studied in the future.
Acknowledgements
The authors sincerely appreciate the financial support sponsored by
the National Science Council of the Republic of China under the Grant
105-2221-E-002-108-.
References
Abramson, H.N., 1966. The Dynamic Behavior of Liquids in Moving Containers with
Application to Space Vehicle Technology. SP-106. NASA, U.S.A.
Abbaspour, M., Hassanabad, M.G., 2010. Comparing sloshing phenomena in a rectangular
container with and without a porous medium using explicit nonlinear 2-D BEM-FDM.
Scientia Iranica Trans. B-Mech. Eng. 17, 93–101.
Amano, K., Iwano, R., Sibata, Y., 1993. Three-dimensional analysis method for sloshing
behavior and its application to FBRs. Nucl. Eng. Des. 140, 297–308.
Beck, J.L., 1972. Convection in a box of porous material saturated with fluid. Phys. Fluids
15, 1377–1383.
Biot, M.A., 1962. Mechanics of deformation and acoustic propagation in porous media. J.
Appl. Phys. 33, 1482–1489.
Bauer, H.F., 1984. Oscillations of immiscible liquids in a rectangular container: a new
damper for excited structures. J. Sound Vib. 93, 117–133.
Chen, S.R., Chang, C.C., Cai, C.S., 2008. Study on stability improvement of suspension
bridge with high-sided vehicles under wind using tuned-liquid-damper. J. Vib. Contr.
14, 711–730.
Chen, Y.H., Hwang, W.S., Tsao, W.H., 2017. Nonlinear sloshing analysis by regularized
boundary integral method. J. Eng. Mech. 143 040170046–1.
Chen, Y.H., Hwang, W.S., Tsao, W.H., 2018. Nonlinear dynamic characteristics of rec-
tangular and cylindrical TLD's. J. Eng. Mech. 144, 06018004.
Chang, P.M., Lou, J.Y.K., Lutes, L.D., 1998. Model identification and control of a tuned
liquid damper. Eng. Struct. 20, 155–163.
Crowley, S., Porter, R., 2012. An analysis of screen arrangement for tuned liquid damper.
J. Fluid Struct. 34, 291–309.
Clough, R.W., Penzien, J., 1993. Dynamics of Structures. McGraw-Hill, New York.
Chang, J.C.H., Soong, T.T., 1980. Structural control using active tuned mass dampers. J.
Eng. Mech. 106, 1091–1098.
Chwang, A.T., Wang, K.H., 1984. Nonlinear impulsive force on an accelerating container.
J. Fluid Eng. 106, 233–240.
Den Hartog, J.P., 1956. Mechanical Vibrations, fourth ed. McGraw-Hill, New York.
Dodge, F.T., 2000. The New Dynamic Behavior of Liquids in Moving Containers. Update
of NASA SP-106. South west Research Institute, SanAntonio, Texas.
Fujino, Y., Pacheco, B.M., Chaiseri, P., Sun, L.M., 1988. Parametric studies on tuned li-
quid damper (TLD) using circular containers by free-oscillation experiments. Struct.
Eng./Earthquake Eng. JSCE Proc. 5, 381–391.
Faltinsen, O.M., Timokha, A.N., 2011. Natural sloshing frequencies and modes in a rec-
tangular tank with a slat type screen. J. Sound Vib. 330, 1490–1503.
Graham, E.W., Rodriguez, A.M., 1952. The characteristics of fuel motion which affect
airplane dynamics. J. Appl. Mech. 19, 381–388.
Housner, G.W., 1957. Dynamic pressure on accelerated containers. Bull. Seismol. Soc.
Am. 47, 15–35.
Huang, L.H., 1991. The inertial effect of a finite thickness porous wavemaker. J. Hydraul.
Res. 29, 417–432.
Housner, G.W., Bergman, L.A., Caughey, T.K., Chassiakos, A.G., Claus, R.O., Masri, S.F.,
Skelton, R.E., Soong, T.T., Spencer, B.F., Yao, J.T.P., 1997. Structural control: past,
present, and future. J. Eng. Mech. 123, 897–971.
Hrovat, D., Barak, P., Robins, M., 1983. Semi-active versus passive or active tuned mass
dampers for structural control. J. Eng. Mech. 109, 691–701.
Huang, L.H., Chao, H.I., 1992. Reflection and transmission of water wave by porous
64
Ocean Engineering 167 (2018) 55–64
breakwater. J. Waterw. Port, Coast. Ocean Eng. 118.
Ha, M., Cheong, C.L., 2016. Pitch motion mitigation of spar-type floating substructure for
offshore wind turbine using multilayer tuned liquid damper. Ocean Eng. 116,
157–164.
Huang, L.H., Hsieh, P.C., Chang, G.Z., 1993. Study on porous wave makers. J. Eng. Mech.
119.
Hsu, H.J., Huang, L.H., Hsieh, P.C., 2005. Oblique impact of water waves on thin porous
walls. J. Eng. Mech. 131.
Ibrahim, R.A., 2005. Liquid Sloshing Dynamic: Theory and Applications. Cambridge
University Press, Cambridge.
Ibrahim, R.A., Pilipchuk, V.N., Ikeda, T., 2001. Recent advances in liquid sloshing dy-
namics. Appl. Mech. Rev. 54, 133–199.
Jin, H., Liu, Y., Li, H.J., 2014. Experimental study on sloshing in a tank with an inner
horizontal perforated plate. Ocean Eng. 82, 75–84.
Jin, Q., Li, X., Sun, N., Zhou, J., Guan, J., 2007. Experimental and numerical study on
tuned liquid dampers for controlling earthquake response of jacket offshore platform.
Mar. Struct. 20, 238–254.
Ju, Y.K., Yoon, S.W., Kim, S.D., 2004. Experimental evaluation of a tuned liquid damper
system. Struct. Build. 157, 251–262.
Kumar, A., Sinhamahapatra, K.P., 2016. Dynamics of rectangular tank with perforated
vertical baffle. Ocean Eng. 126, 384–401.
Li, Y.C., Di, Q.S., Gong, Y.Q., 2012. Equivalent mechanical models of sloshing fluid in
arbitrary-section aqueducts. Earthq. Eng. Struct. Dynam. 41, 1069–1087.
Liu, D., Lin, P., 2008. A numerical study of three-dimensional liquid sloshing in tanks. J.
Comput. Phys. 227, 3921–3939.
Li, Y.C., Wang, J.T., 2012. A supplementary exact solution of equivalent mechanical
model for sloshing fluid in a rectangular tank. J. Fluid Struct. 31, 147–151.
Maravani, M., Hamed, M.S., 2011. Numerical modeling of sloshing motion in a tuned
liquid damper outfitted with a submerged slat screen. Int. J. Numer. Meth. Fluid. 65,
834–855.
Modi, V.J., Welt, F., 1988. Damping of wind induced oscillations through liquid sloshing.
J. Wind Eng. Ind. Aerod. 30, 85–94.
Nield, A.D., Bejan, A., 2013. Convection in Porous Media, fourth ed. Springer, New York.
Nayak, S.K., Biswal, K.C., 2015. Fluid damping in rectangular tank fitted with various
internal objects-an experimental investigation. Ocean Eng. 108, 552–562.
Okamoto, T., Kawahara, M., 1992. Two-dimensional sloshing analysis by the arbitrary
Lagrange-Eulerian finite element method. Structural Engineering and Earthquake
Engineering 8, 207–216.
Peek, R., Jennings, P.C., 1988. Simplified analysis of unanchored tanks. J. Earthquake
Eng. Struct. Dyn. 16, 1073–1085.
Ruiz, R.O., Lopez-Garcia, D., Taflanidis, A.A., 2016. Modeling and experimental valida-
tion of a new type of tuned liquid damper. Acta Mech. 227, 3275–3294.
Soong, T.T., Dargush, G.F., 1997. Passive Energy Dissipation Systems in Structural
Engineering. Wiley, New York.
Tait, M.J., 2008. Modelling and preliminary design of a structure-TLD system. Eng. Struct.
30, 2644–2655.
Tait, M.J., Deng, X., 2010. The performance of structure-tuned liquid damper systems
with different tank geometries. Struct. Contr. Health Monit. 17, 254–277.
Tait, M.J., El Damatty, A.A., Isyumov, N., 2005. An investigation of tuned liquid dampers
equipped with damping screens under 2D excitation. Earthq. Eng. Struct. Dynam. 34,
719–735.
Tamura, Y., Fujii, K., Ohtsuki, K., Wakahara, T., Kohsaka, R., 1995. Effectiveness of tuned
liquid dampers under wind excitation. Eng. Struct. 17, 609–621.
Veletsos, A.S., 1984. Seismic response and design of liquid storage tanks. Guidelines for
the Seismic Design of Oil and Gas Pipeline Systems 255–370.
Wu, N.J., Hsiao, S.C., Wu, H.L., 2016. Mesh-free simulation of liquid sloshing subjected to
harmonic excitations. Eng. Anal. Bound. Elem. 64, 90–100.
Wakahara, T., Ohyama, T., Fujii, K., 1992. Suppression of wind-induced vibration of a tall
building using tuned liquid damper. J. Wind Eng. Ind. Aerod. 43, 1895–1906.
Warnitchai, P., Pinkaew, T., 1998. Modelling of liquid sloshing in rectangular tanks with
flow-dampening devices. Eng. Struct. 20, 593–600.
You, K.P., Kim, Y.M., Yang, C.M., Hong, D.P., 2007. Increasing damping ratios in tuned
liquid damper using damping bars. Key Eng. Mater. 353–358, 2652–2655.
Zahrai, S.M., Abbasi, S., Samali, B., Vrcelj, Z., 2012. Experimental investigation of uti-
lizing TLD with baffles in a scaled down 5-story benchmark building. J. Fluid Struct.
28, 194–210.