Structures 55 (2023) 169–184

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

Theoretical and experimental research on damping performance of

suspended multiple mass pendulums damper
Haoding Sun, Haoxiang He *, Yang Cheng, Shitao Cheng
Beijing Key Lab of Earthquake Engineering and Structural Retrofit, Beijing University of Technology, Beijing 100124, China

A R T I C L E I N F O A B S T R A C T

Keywords: Conventional suspended mass pendulum damper (SMPD) include suspended single mass pendulum damper
Tuned vibration damping (SSMPD), tandem suspended multiple mass pendulums damper without springs (TSMP), and parallel suspended
Suspended mass pendulum multiple mass pendulums damper without springs (PSMP), these types of SMPD have the advantages such as
Multi-pendulums
clear mechanism, simple construction, and good damping effect, but they have a narrow tuning frequency band
Shaking table tests
Multiple high-rise structure
and the pendulum length needs to be set too long for low-frequency super tall structures and too short for high-
frequency tall structures, limiting their application in practical engineering. Because of this, the paper proposes
connecting the spring to the mass pendulum, forming a tandem suspended multiple mass pendulums damper
with spring (TSMP-S) and a parallel suspended multiple mass pendulums damper with spring (PSMP-S), and
establishing structural system dynamics equations and performing the dynamic characterization of the TSMP-S
and PSMP-S, obtain their frequency resolved solutions to explain the damping mechanism. In order to study the
vibration-damping performance of these two types of SMPD, through numerical simulations to compare the
damping effect of the conventional SMPD, TSMP-S, and PSMP-S on the controlled structures, and a series of
shaking table tests on controlled and uncontrolled structures to verify the effectiveness of SSMPD, TSMP, PSMP,
TSMP-S and PSMP-S for damping control of controlled structures under conventional earthquake and earthquake
at four types of sites. Numerical simulations and shaking table test results in both show that the TSMP-S has the
best damping effect, with the advantages of increasing the tuning frequency band, achieving multi-order vi­
bration control, and flexible adjustment of the pendulum length, which is suitable for actual engineering.

1. Introduction Given the advantages of SMPD and its good damping effect, re­
searchers have conducted more intensive theoretical and experimental
The conventional tuned damping control devices [1–4] include the studies on the damping mechanism of SMPD. In terms of theoretical
tuned mass damper (TMD), the tuned liquid damper (TLD), and the research, Li et al. [7–8] established a theoretical mechanical model of
suspended mass pendulum damper (SMPD), etc. TMD consists of a SMPD in a multi-layer structure and studied the effects of various pa­
damper, a spring, and a mass block, and is usually suspended or sup­ rameters such as different suspension methods of the suspended
ported on the controlled structure, the damping effect is significant pendulum and the pendulum-to-structure mass ratio μ on the structural
when its self-oscillation frequency is the same as the main frequency of damping control, and the results showed that with the increase of the
the structure and is therefore widely used [5]. TLD is a tuning device pendulum-to-structure mass ratio μ, the damping effect was better.
equipped with liquid [6], which is simple, tunable, and low-cost. SMPD Kourakis [9] and Tuan et al. [10] established the finite element
refers to suspending the mass pendulum on the main structure and the modeling of the Taipei 101 Tower with SMPD, and carried out dynamic
suspension pendulum can swing, so that it is close to the self-oscillation time-history analysis under earthquake and wind loads. The results
frequency of the structure to achieve tuned damping and thus reduce the showed that SMPD can effectively reduce the wind vibration and seismic
dynamic response of the structure. Compared with TMD and TLD, SMPD response of the Taipei 101 Tower. Roffel et al. [11–13] proposed a
has the advantages such as obvious damping effect, simple construction, procedure for state evaluation of existing pendulum TMDs by studying
clear mechanism and easy arrangement, so it has broad engineering the dynamic response of a nonlinear pendulum TMD with three-
application prospects. dimensional motion coupled to a flexible multi-degree-of-freedom

* Corresponding author.
E-mail address: [email protected] (H. He).

https://doi.org/10.1016/j.istruc.2023.06.048
Received 4 May 2023; Received in revised form 9 June 2023; Accepted 9 June 2023
Available online 14 June 2023
2352-0124/© 2023 Institution of Structural Engineers. Published by Elsevier Ltd. All rights reserved.
H. Sun et al. Structures 55 (2023) 169–184

structure. Pasala et al. [14] propose an adaptive suspended mass control period of the main structure of the Taipei 101 building is about
pendulum with a pendulum length that can be adjusted in real time 6.8 s, and the theoretical pendulum length of the suspended pendulum is
using a shape memory alloy (SMA) wire actuator. Experimental and 11.5 m, which requires about 4 layers of space and is more difficult to
theoretical results show that this adaptive pendulum can quickly adjust implement. In summary, the traditional SMPD is difficult to be applied
the pendulum length to match the dominant frequency of the structure, directly in engineering practice, so the engineering feasibility of SMPD
and the damping efficiency is greatly improved compared to that of a needs to be solved urgently.
conventional suspended mass pendulum damper. Huang et al. [15–16] To solve the problem of the narrow frequency band of SMPD, Yan et
established coupled dynamic equations for a single-degree-of-freedom al [23] realized SMPD to control the multi-directional vibration of the
structural space-suspended mass pendulum model and numerically sign tower of Henan Art Center in China by connecting multiple addi­
analyzed the vibration characteristics and performance of the tional springs to one mass block, and the experimental results showed
pendulum, and the results showed that the suspended mass pendulum that the device could significantly reduce the wind-driven response of
has a significant weakening effect on the seismic response of trans­ the tower, and at the same time, the frequency of the dampers could be
mission towers. Huo et al. [17] proposed a passive adaptive suspension easily adjusted by swapping the spring stiffness and position, but due to
pendulum and studied its damping effect on single-layer mesh-shell the setting of only one mass block, it could not realize multi-order and
structures under different horizontal and vertical seismic effects, and the multi-vibration control of the tower. Lu et al. [24] and Miguel et al. [25]
results showed that the damping effect of this adaptive suspension proposed to use the suspended double mass pendulums damper system
pendulum was significantly better than that of the conventional SMPD at to control the vibration of the structure. Lu et al. proposed to use the
large pendulum angles. double pendulum system to control the wind vibration of the Shanghai
In terms of experimental research, Setareh et al. [18–19] established center tower, China’s tallest building, and conducted parameter design
an equivalent single-degree-of-freedom model of the pendulum TMD and theoretical analysis. The results show that the space occupancy rate
and verified the possibility of its tuned damping through theory and of the double pendulum system is further optimized compared with the
experiment, and the semi-active magnetorheological device was applied single pendulum system, and the vibration reduction effect is better.
to the pendulum TMD system to control the excessive vibration of the Miguel also chose to use the double pendulum system to control the
floor, and the good damping effect of this semi-active pendulum damper wind vibration of the steel frame structure. The results show that
was proved by theory and test. Wang et al. [20] proposed a self- compared with the single pendulum, the suspended double pendulum
adjustable variable pendulum tuned mass damper (SAVP-TMD), which can achieve the multi-order frequency tuning of the structure, and the
can be adjusted by a microcontroller for the pendulum length, making damping effect is better. However, the above two double pendulum
the SAVP-TMD self-adaptive to the main frequency of the structure. Both systems have the problem of the long pendulum length used for the low
theory and tests had shown that this form of pendulum tuned mass fundamental frequency structure resulting that the engineering feasi­
damper had better vibration control effect than the traditional sus­ bility is low. Roffel et al. [26] proposed an adaptive suspended mass
pended pendulum. Eason et al. [21] proposed a tandem-shaped sus­ pendulum, this type of pendulum can adjust the damping value of the
pended mass pendulum with adjustable pendulum length to match its pendulum through adjustable air dampers on the tuning frame, thus
intrinsic frequency to the main frequency of the structure, which had effectively adjusting the natural frequency of the pendulum, but this
been proven both experimentally and theoretically that it can success­ form of pendulum design is more complex and has poor engineering
fully prevented the transition of the response from the low amplitude applicability. Gerges et al [27] replaced the spring and damping part of
solution to the high amplitude solution or return the response from the the traditional TMD with a steel cable spring to form a suspended steel
high amplitude solution to the low amplitude solution, thereby pro­ cable spring SMPD, and the experimental study showed that the problem
tecting the primary structure. Sun et al. [22] used a nonlinear tuned of the arrow frequency of the pendulum damper could be well improved
mass damper (NTMD) in parallel with an adaptive-length pendulum by this method.
tuned mass damper (APTMD). The experiments indicated that the To solve the problem on the feasibility of the engineering pendulum
combination of the two TMDs can greatly attenuate high amplitude length, Qin et al [28] proposed a friction-type SMPD, which can realize
resonance branch phenomena and can achieve adaptive structural the SMPD stiffness and pendulum length adjustment and improve the
principal frequency by varying the pendulum length. It provided an stability of the mass block in the swinging process by setting the disc
effective and feasible solution for vibration reduction control of non- spring and friction damping piece, but the pendulum length adjustment
linear structures. length range of this SMPD is narrow and does not have flexibility. Kurino
All the above studies have proved that SMPD has a good damping et al [29] proposed a suspended pendulum type TMD, when the TMD
effect and its damping performance can be further enhanced by semi- reaches a specific speed, a specific pressure within the fluid in the high-
active control, nonlinear stiffness modification and adaptive suspen­ function oil damper with a unique hydraulic circuit causes part of the
sion, etc. However, it should be noted that SMPD has deficiencies such as built-in valve to open or close, thus actively controlling the swing
a narrow tuning band, too long tuning pendulum length in long-period amplitude of the TMD. Although this method can adjust the swing length
structure, and too short pendulum length in short-period structure, and and thus meet the engineering availability, the operation method is
SMPD damping studies mainly focus on SSMPD with only one concen­ difficult and not easy to use in engineering practice. Huber [30] pro­
trated mass and no additional stiffness, the period equation of SSMPD is posed a double mass pendulums box type tuned damper with additional
√̅̅̅̅̅̅
T = 2π l/g, where l is the length of the pendulum and g is the accel­ rubber damping to give the double mass pendulums box type The range
eration of gravity. The period of the SSMPD should be made the same as of motion of TMD provides a limit to the traditional suspension
the main structure period by adjusting the pendulum length of the pendulum to solve the problems such as large installation space with the
SSMPD, to control the vibration of the structure. But the pendulum double pendulum length, and it has been applied. The above two ways of
length is the only parameter to control the period of the structure, so the combining a suspended mass pendulum and TMD can partially solve the
single pendulum is poorly tunable. For example, the optimal pendulum problem of pendulum length, However, a further in-depth research is
lengths of SSMPD are only 0.062 m, 0.248 m, 0.559 m, and 0.993 m for needed on the practical feasibility and effectiveness.
the structures with first periods of 0.5 s, 1.0 s, 1.5 s, and 2 s, respectively, In summary, although certain results have been achieved in
which shows that SSMPD is difficult to be directly applied to actual low enhancing the vibration control effect of SMPD, the problem of
and mid-rise buildings due to the limited pendulum length. On the other improving the multi-order frequency modulation ability and pendulum
hand, the use of space will be affected due to the excessive pendulum length of SMPD is still at the stage of theoretical investigation, and there
length for the super high-rise building structure. For example, the is no mature theoretical basis, and reliable test results to guide the actual
design and application of SMPD, and there is an urgent demand to

170
H. Sun et al. Structures 55 (2023) 169–184

l1 l1
kp1
l2 l2
m1 kp2 m1
kp1 l1
l2 l1
kp2 l2
m2 m2
m2 m2
m1
m1

(a) TSDP-S (b) PSDP-S (c) TSDP (d) PSDP

Fig. 1. Sketch of four suspended multi-pendulum damper.

analyze the mechanism needs to be analyzed and experimentally veri­ as follows

fied. In view of this, based on the existing research results, this paper ⎡ ⎤
(ma1 + ma2 )g ma2 g ma2 g
puts forward the problem of flexibly adjusting the stiffness of SMPD by [ ]{ } { }
ma1 0 ẍa1 ⎢ ka1 + la1
+
la2
−
la2 ⎥ xa1
connecting the series or parallel suspended multi-mass pendulums with +⎢
⎣
⎥
0 ma2 ẍa2 ma2 g ma2 g ⎦ xa2
the structural wall through the spring, so as to achieve multi-level fre­ − ka2 +
la2 la2
quency adjustment and improve the problem of the length of the { }
pendulum. By constructing mechanical models of various types of =
0
SMPD, the dynamic performance of various types of suspended pendu­ 0
lums is analyzed theoretically, and the dynamic equations of various (1)
types of suspended pendulum structural systems are established. It in­
dicates that the suspension multi-pendulum dampers with springs have where ma1 and ma2 are the masses of the TSDP-S, la1 and la2 are the boom
multi-order tuning capability and improved the problem of the lengths of the TSDP-S, ka1 and ka2 are the stiffnesses of the additional
pendulum length, and have obvious superiority for structural damping springs of the TSDP-S, and xa1 and xa2 are the relative displacements of
control. the two pendulums of the TSDP-S to the ground, respectively.
From the above equation, the two-order frequency analytic expres­
2. Analysis of the dynamic characteristics of each type of SMPD sion of the TSDP-S is obtained as
⎧ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
⎪ √
√I ± I 2 − 4m m I
Combined with the existing research, different forms of suspended ⎪
⎪
⎪
⎪ 1 √ a1 a1 a1 a2 a2
⎪ fa =
mass multi-pendulum with additional springs are listed in this paper, as ⎪
⎪
⎪ 2π 2ma1 ma2
⎪
shown in Fig. 1 (this paper will discuss in detail the double mass pen­ ⎨
(ma1 + ma2 ) ma1 ma2 + m2a2
dulums damper as an example of the multiple mass pendulums damper ⎪
⎪ Ia1 = ka2 ma1 + ka1 ma2 + g+ g
⎪ l la2
form, the rest of the multiple mass pendulums damper can follow the ⎪
⎪
⎪
⎪
a1
⎪
theoretical model of the double mass pendulums damper). The mass ⎪
⎪ m g
⎩ Ia2 = ka1 ka2 + ka1 a2 + ka2 a1
(m + ma2 )g (ma1 + ma2 )ma2 2
+ g + ka2
ma2 g
block is connected to the structural floor through a rigid boom and la2 la1 la1 la2 la2
universal hinge support. The structural model of the SMPD with two (2)
mass blocks acting in series is shown in Fig. 1(a) and Fig. 1(c), where the
If ka1 = ka2 = 0, ma1 = ma2 = m, and la1 = la2 = l, the first two orders
mass blocks are connected to the structural floor through a rigid boom
of frequency analytic expressions for the TSDP (as in Fig. 1(c)) are ob­
and universal hinge support. The double mass pendulums with addi­
tained as follows:
tional springs are named tandem suspended double mass pendulums
⎧ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√̅̅̅ ̅
damper with spring (TSDP-S), and the double mass pendulums without ⎪
⎪
⎪ 1 (2 − 2)
additional springs are named tandem suspended double mass pendu­ ⎪
⎨ fc1 = 2π
⎪ g
l
lums damper (TSDP). The difference between the two SMPDs is the √ ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ (3)
TSDP-S is connected to the structural wall through the spring, and the ⎪
⎪ √̅̅̅ ̅
⎪
⎪ 1 (2 + 2 )
⎪
connection method is a bolt-anchored connection. By reasonably setting ⎩ fc2 =
2π l
g
the tuned mass size m of the TSDP-S mass block, the length l of each rigid
boom section, and the spring stiffness k, the first two self-oscillation It can be seen that the parameters affecting the self-oscillation fre­
periods of the TSDP-S can be made close to the first two fundamental quency of TSDP-S include the additional spring stiffness, the pendulum
periods of the structure, thus realizing the tuned vibration reduction of length and the mass of the mass block, through reasonable adjustment of
the structure at multiple frequencies and reducing the structural dy­ the above three elements, the multi-order frequency of TSDP-S can be
namic response; TSDP is not equipped with the spring, therefore it does effectively adjusted to achieve multi-order tuned damping of the struc­
not have the ability to adjust the stiffness and lacks the flexibility to ture. However, the factors affect the self-oscillation frequency of the
regulate its first two orders of self-oscillation period. TSDP only include the length of the pendulum and the mass of the mass
The structural model of the SMPD with two mass blocks acting in block. Due to the lack of the tuning effect of the additional spring
parallel is shown in Fig. 1(b) and Fig. 1(d). The double mass pendulums stiffness, the damping frequency range of the TSDP is narrower than that
with additional springs are named parallel suspended double mass of the TSDP-S.
pendulums damper with spring (PSDP-S), and the double mass pendu­ The dynamics model of the PSDP-S is shown in Fig. 1(b), and the
lums without additional springs are named parallel suspended double corresponding free dynamic equations of the system are as follows
mass pendulums damper (PSDP). The difference between the two ⎡m g
b1
⎤
SMPDs is the same as TSDP-S and TSDP mentioned above and will not be [ ]{ } − kb1 0 { } { }
mb1 0 ẍb1 ⎢ l ⎥ xb1 0
(4)
b1
given unnecessary details. 0 mb2 ẍb2
+⎢
⎣ mb2 g
⎥
⎦ xb2 = 0
The free dynamic equation of the TSDP-S system shown in Fig. 1(a) is 0
lb2
− kb2

171
H. Sun et al. Structures 55 (2023) 169–184

Fig. 2. Schematic diagram of structural systems with four types of the SMPD.

where mb1 and mb2 are the masses of the two masses of the PSDP-S, mass blocks of the PSDP-S to the ground, respectively.
respectively, lb1 and lb2 are the lengths of the pendulum of the PSDP-S, The analytical expression for the first two orders of frequency of the
kb1 and kb2 are the spring stiffness connected to the two mass blocks in PSDP-S can be obtained as
the PSDP-S, and xb1 and xb2 are the relative displacements of the two

1 n-1 n
1 n-1 n
1 n-1 n

1 n-1 n

1 n-1 n

1
1
p1

p1

p1 2 2
p2

p2

p2

(a) Equivalent mechanical model of TSDP-S structural system

1 n-1 n
1 n-1 n
1 n-1 n

1 n-1 n

1 n-1 n

2 p2
1
p1 p2
1 2
p1 p2

p1

(b) Equivalent mechanical model of PSDP-S structural system

Fig. 3. Equivalent mechanical model for TSDP-S and PSDP-S controlled structural systems.

172
H. Sun et al. Structures 55 (2023) 169–184

⎧ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
⎪ √ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ 3. Damage constitutive model of multi-scale modified concrete
⎪
⎪ √ 2
⎪
⎪
⎪ 1 √Ib1 ± I b1 − 4mb1 mb2 Ib2
⎪ fb =
⎪
⎪
⎨ 2π 2mb1 mb2 In order to evaluate the damping performance of various types of
mb1 mb2 g mb1 mb2 g (5) SMPD more reasonably and fully study their dynamic characteristics and
⎪
⎪ Ib1 = + − mb1 kb2 − mb2 kb1 damping mechanisms acting together with the controlled structure, we
⎪
⎪ lb2 lb1
⎪
⎪
⎪
⎪ m g m g m m take the frame structure with TSDP-S, PSDP-S, TSDP, and PSDP installed
⎪
⎩ Ib2 = kb1 kb2 − kb1 b2 − kb2 b1 + b1 b2 g2
lb2 lb1 lb1 lb2 on the top floor respectively as the research object and establishes the
structural dynamic equations. The overall simplified model of the
If kb1 = kb2 = 0, mb1 = mb2 = m, and lb1 = lb2 = l, the first two order structure with the four types of SMPD is shown in Fig. 2. The equivalent
frequency analytic expressions of the PSDP (as in Fig. 1(d)) are obtained mechanical model of TSDP-S and PSDP-S controlled structural system is
as follows shown in Fig. 3. In the figure, mi, ki, and ci are the mass, stiffness, and
√̅̅̅
1 g damper of each layer of the structure, kp1 and kp2 are the additional
fd1 = fd2 = (6) spring stiffness of the SMPD, mp1 and mp2 are the masses of the sus­
2π l
pended pendulum mass blocks, cp1 and cp2 are the damping between the
According to the above equation, it can be seen that the parameters pendulums and the springs, l1 and l2 are the pendulum lengths, θ1(t) and
affecting the self-oscillation frequency of the PSDP-S involve the addi­ θ2(t) are the angle of pendulum swing, Fi(t) is the external load acting on
tional spring stiffness, the pendulum length, and the mass of the mass each floor of the controlled structure. When the structure vibrates under
block. By reasonably adjusting the above three elements, the multi-order seismic action, the mass blocks of each type of SMPD fitted at the top of
frequency of the PSDP-S can be effectively adjusted to achieve the multi- its floors will oscillate with the vibration of the structure.
order tuned damping of the structure. While the frequency of the PSDP is The dynamic equation of the structural system equipped with the
the same as that of the SSMPD, due to the simultaneous shaking of the four types of SMPD in which the mass block oscillates reciprocally with
two pendulums, when limiting the pendulum mass, the tuning damping the action of the earthquake is given by
of the structure can be achieved by adjusting the pendulum length of the
two pendulums, which can realize multi-order frequency control MÜ + CU̇ + KU = − MÜg (t) (7)
compared to the SSMPD.
In summary, the frequencies of TSDP-S and PSDP-S are related to the where M, C, and K are the mass matrices, damping matrix, and stiffness
mass of the mass block, the pendulum length, and the spring stiffness. matrix of the structure, respectively. U is the relative displacement
Compared with the traditional SSMPD, TSDP, and PSDP, the TSDP-S and matrix of each floor and two mass blocks to the ground and Üg (t) is the
PSDP-S can achieve the same or close to the self-oscillation frequency of ground vibration acceleration input. The overall stiffness matrices of the
each order with the previous orders of the controlled structure by structure fitted with the TSDP-S and PSDP-S are as follows
optimally adjusting the above three elements, which can increase the
frequency range of its tuned damping and adjust the pendulum length
flexibly at the same time and is more suitable for engineering practical
applications.

⎡ ⎤
k1 + k2 − k2 ⋯ 0
⎢ ⎥
⎢ ⎥
⎢ − k2 k2 + k3 − k3 ⋯ 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ − k3 ⋱ ⋱ ⋮ ⎥
⎢ ⎥
⎢ mp1 g mp2 g ⎥
⎢ − kp2 ⎥
Ka = ⎢ 0 − kn− 1 kn +
l1
+
l2
+ kp1 + kp2 − kp1 ⎥ (8)
⎢ ⎥
⎢ ⎥
⎢ (mp1 + mp2 )g (mp1 + mp2 )g mp2 g mp2 g ⎥
⎢ ⋮ ⋮ − kp1 − kp1 + + − ⎥
⎢ lp1 l1 l2 l2 ⎥
⎢ ⎥
⎢ mp2 g mp2 g ⎥
⎣ ⎦
0 ⋯ − kp2 − kp2 +
l2 l2

⎡ ⎤
k1 + k2 − k2 ⋯ 0
⎢ ⎥
⎢ − k k2 + k3 − k3 ⋯ 0 ⎥
⎢ 2 ⎥
⎢ ⎥
⎢ − k3 ⋱ ⋱ ⋮ ⎥
⎢ ⎥
⎢ ⎥
⎢ mp1 g mp2 g mp1 g mp2 g ⎥
Kb = ⎢
⎢
0 − kn− 1 kn + + − kp1 − kp2 − + kp1 − + kp2 ⎥
⎥ (9)
⎢ l1 l2 l1 l2 ⎥
⎢ mp1 g mp1 g ⎥
⎢ ⋮ ⋮ ⋮ − + kp1 − kp1 0 ⎥
⎢ l1 l1 ⎥
⎢ ⎥
⎢ ⎥
⎣ mp2 g mp2 g ⎦
0 0 ⋯ − + kp2 0 − kp2
l2 l2

173
H. Sun et al. Structures 55 (2023) 169–184

Table 1 Table 2
Earthquake Information. Parameters of 5 types of SMPD.
Earthquake Time Station magnitude PGA (g) Types of Length Length Spring Spring
SMPD l1(m) l2(m) stiffness stiffness
Imperial Valley 1940 El Centro 6.5 0.27
kp1(N/m) kp2(N/m)
Cape Mendocino 1992 Petrolia 7.2 0.59
SSMPD 0.26 – – –
TSDP-S 1.50 1.50 2.70 × 105 6.02 × 106
Each mass matrix can be expressed as M = diag[m1, m2, …, mn, mp1, PSDP-S 1.00 1.00 4.84 × 105 5.77 × 106
TSDP 0.24 0.08
mp2], where diag denotes the diagonal matrix. The damping matrix Cs of – –
PSDP 0.28 0.03 – –
the controlled structure can be expressed in terms of the Rayleigh
damping matrix, the damping values of each type of suspended mass
pendulum dampers cpi = 2mpiωpiζpi, mpi is the mass of each mass block in mass and stiffness of each layer is selected as the controlled structure in
each type of suspended mass pendulum dampers, ωpi is the circular this study, and five types of SMPD, namely SSMPD, TSDP-S, PSDP-S,
frequency of each suspended mass pendulum dampers, where the TSDP, and PSDP, are used to control the frame structure respectively.
damping ratio ζpi can be found according to the optimal damping ratio The mass and stiffness units of the 10-layer frame structure are first
ζpo proposed by Den [31], i.e. extracted by SAP2000 finite element software, and then the mass and
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ stiffness matrices are put into the Matlab theoretical model for numer­
3μ ical simulation and analysis. The damping effect of each type of
ζpo = (10)
8(1 + μ)3 controlled structures under the action of El Centro and Cale seismic
waves is compared and analyzed through numerical simulations to
where μ is the mass ratio of each suspended mass pendulum dampers to verify that the suspended multi-pendulum equipped with springs has the
the controlled structure. The total damping matrix can be expressed as: capability of multi-order tuned damping. The specific information on
[ ] seismic waves as shown in Table 1.
Cs + cpi − cpi
C= (11) As shown in Fig. 4, the controlled structure is a 10-layer frame
− cpi cpi
structure, with each layer height of 4 m, plane size of 30 m × 30 m, the
{ }T { }T
Defining X = U U̇ and Y = U Ü as the state vector of the mass of each layer of 4 × 105 kg, and stiffness of each layer of 5.5 × 105
kN/m. All types of SMPD are installed on the top layer of the structure,
system, Eq. (7) can be expressed in the form of the following state and the total mass of all five types of SMPD is taken as 2% of the total
equation mass of the structure. The first two order periods of the structure are
Ẋ = AX + BÜge (t)Y = CX + DÜge (t) (12) 1.134 s and 0.357 s respectively, and the main control vibration type is
the first-order mode. By adjusting the length of the pendulum, the self-
[ ] [ ]
0 I I 0 oscillation period of the SSMPD is adjusted to around 1.134 s, and the
where A = , C = , B = D = first two order periods of the TSDP and PSDP are adjusted to 1.134 s and
− M1K − M1Cs − M1K − M1C
[
0
] 0.357 s. By adjusting the length of the pendulum and the spring stiffness,
. Based on the ground motion acceleration excitation vector, the the first two order periods of the TSDP-S and PSDP-S are adjusted to
− I
coupled dynamic response time of the structure and each type of SMPD 1.134 s and 0.357 s. In order to meet the actual needs of the project, the
can be calculated by solving the system state-space equations. length of the pendulum rigid boom and spring stiffness of various types
of SMPDs are calculated as shown in Table 2.
4. Comprehensive evaluation of concrete performance based on The dynamic time analysis of the structure under the action of El
modified damage variable Centro and Cale seismic waves is carried out, and the peak acceleration
values of the earthquake waves are adjusted to 0.4 g according to the
In order to compare and analyze the damping effect of various types seismic design requirements. The comparison graphs of the structural
of suspended multiple mass pendulums damper-multiple layer structure top displacement and acceleration time-history curves of the controlled
systems subjected to earthquake and the ability of multi-order tuned structure are shown in Fig. 5. It is evident that all five types of SMPD can
damping, a reinforced concrete 10-layer frame structure with the same effectively work on tuning the damping of the controlled structure,
among which, the damping effect of TSDP-S is obviously better,

Fig. 4. Model diagram of 10-layer frame structure.

174
H. Sun et al. Structures 55 (2023) 169–184

Fig. 5. The time history curve of top layer of the structure.

followed by TSDP. The damping advantage of PSDP-S is not fully dis­ the pendulum, but its damping effect is only inferior to the TSDP-S. The
played, and the damping effect of the remaining three types of tradi­ damping effect of the series-shaped suspended double mass pendulums
tional SMPD is close, but its ability to flexibly adjust the pendulum with two mass blocks in series is better than that of the parallel-shaped
length is not possessed by the remaining three types of traditional SMPD. suspended double mass pendulums with two mass blocks in parallel and
Fig. 6 (a) shows the displacement time-history comparison of the the SSMPD, and the damping effect will be further improved with the
TSDP-S upper and bottom pendulums under the El Centro wave. The additional spring. Although the damping advantage of the PSDP-S is not
result shows that the displacement amplitude of the upper pendulum is manifested, its ability to flexibly adjust the length of the pendulum is not
significantly greater than that of the bottom pendulum, but the vibration possessed by the rest of the SMPD without springs, so it is more suitable
frequency of the bottom pendulum is higher than that of the upper for practical engineering applications.
pendulum. The two pendulums synergistically play the role of multi-
order tuning and damping. Fig. 6(b) shows the displacement time-
history comparison of the PSDP-S left and right pendulums under El
Centro wave. The result shows that the displacement amplitude of the
left pendulum is significantly greater than that of the right pendulum,
but the frequency of the right pendulum is higher than that of the left
pendulum. The two pendulums also play the role of multi-order tuning
and damping.
Fig. 7 shows the comparison of the peak inter-layer displacement
(Dpi) and the sum of absolute values of interlayer displacement at each
time point (Dsa) of the controlled structure under the control of various
types of SMPD respectively under the El Centro wave. The calculation
formula is shown as follows
Dpi = max|ui (t)| (13)

∑
n
Dsa = |ui (t)| (14)
i=1

where ui is the dynamic response of step i, and n is the total steps of

dynamic response.
The results further show that the TSDP-S has the best performance,
while the TSDP does not have the ability to flexibly adjust the length of Fig. 7. Layer drift comparison of the structure.

Fig. 6. Displacement time-history curve of pendulum.

175
H. Sun et al. Structures 55 (2023) 169–184

Fig. 8. Power spectral density curve of the acceleration response at the top of the structure.

To verify the ability of each SMPD for multi-order tuned damping, motions are selected for each class of site conditions. The different site
the power spectrum density curves of the top-level acceleration response conditions ground motions are classified based on the average shear
of the structure under the action of El Centro and Cale seismic waves are wave velocity vs30 of the soil layer within 30 m of the surface depth. The
selected for comparison, as shown in Fig. 8. The peak power spectra and
envelope area of the uncontrolled structure is much larger than those of
all types of SMPD structures. The peak power spectrum curve and en­
velope area of the SSMPD are closer to PSDP-S and PSDP, while the
TSDP-S and TSDP structures are relatively smaller, which proves that the
structure equipped with tandem-shaped suspended double mass pen­
dulums has a superior damping performance from the perspective of
energy dissipation. Compared with the suspended mass pendulum
damper without additional springs, the peak of the power spectra and
the envelope area of the structure of the series and parallel-shaped
double mass pendulums damper with additional springs are smaller.
For the comparison of the peak of the power spectrum and the envelope
area at the second-order frequency, the series and parallel-shaped
double mass pendulums damper with additional springs are signifi­
cantly smaller, which adequately indicates that the suspended mass
pendulum damper with additional springs can more effectively improve
the damping capacity of the structure and play a multi-stage tuned
damping effect.
In order to verify the damping performance of various types of SMPD
more comprehensively, various seismic waves of different site types are
selected to carry out dynamic time-history analysis of the controlled
structure. According to the American design code ASCE/SEI 7–16 [32],
the site conditions are divided into sites B, C, D, and E, and five ground Fig. 9. Average acceleration response spectrum with different site types.

Table 3
Information of different site conditions ground motions.
Type Number Name Station Year Vs30(m/s) Mw PGA(g)

Site B W1 San Fernando Pasadena-Old Seismo Lab 1971 969.07 6.6 0.20
W2 Tabas Iran Tabas 1978 766.77 7.4 0.86
W3 Loma Prieta Piedmont Jr High School Grounds 1989 895.36 6.9 0.08
W4 Landers Lucerne 1992 1369.00 7.3 0.78
W5 Northridge-01 Vasquez Rocks Park 1994 996.43 6.7 0.14
Site C W6 San Fernando Santa Felita Dam 1971 389.00 6.6 0.15
W7 Livermore-02 APEEL 3E Hayward CSUH 1980 517.06 5.4 0.06
W8 Kern County Taft Lincoln School 1952 385.43 7.4 0.16
W9 Mammoth Lakes-03 Long Valley Dam (L Abut) 1980 537.16 5.9 0.08
W10 Imperial Valley-06 Cerro Prieto 1979 471.53 6.5 0.17
Site D W11 El Mayor-Cucapah Chihuahua 2010 242.05 7.2 0.25
W12 Friuli Italy-01 Codroipo 1976 249.28 6.5 0.06
W13 Imperial Valley-06 Delta 1979 242.05 6.5 0.24
W14 Victoria Mexico Chihuahua 1980 242.05 6.3 0.27
W15 Superstition Hills-02 Brawley Airport 1987 208.71 6.5 0.14
Site E W16 Loma Prieta Foster City-APEEL 1 1989 116.35 6.9 0.26
W17 Iwate Japan AKT015 2008 135.4 6.9 0.06
W18 Loma Prieta Foster City-Menhaden Court 1989 126.40 6.9 0.11
W19 Niigata Japan FKS020 2004 133.05 6.6 0.04
W20 Niigata Japan NIG011 2004 149.97 6.6 0.06

176
H. Sun et al. Structures 55 (2023) 169–184

Fig. 10. Average damping rate for four types of sites.

basic information of different site conditions ground motions is shown in response. The energy damping ratio (Dre) is defined as the ratio of the
Table 3, and the average response spectrum are shown in Fig. 9. difference between the response envelope area of the uncontrolled and
The damping rate of peak (Drp) is defined as the ratio of the difference controlled structures to the response envelope area of the uncontrolled
between the maximum value of uncontrolled and controlled structural structure. See Eq. (15) and (16) for the calculation of the two damping
response and the maximum value of the uncontrolled structural rates.

500 500
5040

50 390 50
5 5
A A
400

550

500
6

10
4050

5
400

L-shaped angle
50

42 42
5
500

Connected to the Bolt hole

shaking table

(a) Elevation drawing (b) Plan and A-A section drawing

Fig. 11. Test frame details/mm.

177
H. Sun et al. Structures 55 (2023) 169–184

Fig. 12. Details of the design of the SMPD connection to the frame.

max|Su (t)| − max|Sc (t)| dimensions are 500 mm (length) × 500 mm (width) × 6 mm (thickness);
Drp = (15)
max|Su (t)| the cross-sectional dimensions of the column are 50 mm (length) × 5
mm (width), and the distance between each column is 390 mm. The
∑n ∑n
i=1 |Sui (t)| − i=1 |Sci (t)| height of the first floor is 500 mm, and the height of the remaining floors
Dre = ∑n (16)
i=1 |Sui (t)| is 400 mm. The columns and floor slabs are connected by L-shaped
angles and M6 bolts (the M6 bolt’s outer thread diameter is 6 mm with
Adjust the PGA of each seismic wave to 0.4 g and the comparison
shear stress of 24.38 Mpa, tensile strength of 400 Mpa, and shear
results of the average damping rate for different forms of SMPD under
strength of 200 Mpa). The limit walls are added between the second and
various types of ground motion are shown in Fig. 10.
third floors of the frame to meet the connection needs of the springs and
Note: AD-DRP is the average peak displacement damping rate; AD-
walls in the suspended mass pendulum damper with additional springs,
DRE is the average energy displacement damping rate; AA-DRP is the
and the walls are made of acrylic panels with dimensions of 550 mm
average peak acceleration damping rate; AA-DRE is the average energy
(length) × 550 mm (width) × 10 mm (thickness). The detailed drawing
acceleration damping rate.
of the test frame is shown in Fig. 11.
It can be seen from Fig. 10 that the average damping rate of the
The mass block is connected to the top floor of the frame by a live
TSDP-S is the highest under the action of seismic waves in various types
knotted screw with holes and a closed-end screw hook with M6 bolts, the
of sites, the TSDP is followed, and the damping effect of the parallel-
connection can be rotated 180 degrees, and when the structure vibrates,
shaped suspended double mass pendulums damper is worse compared
the mass block can swing back and forth with the live knotted screw with
with it and slightly equal to the damping effect of the SSMPD, but all five
holes, one end of the spring is fixed to the mass block, and the other end
types of SMPD can achieve the damping effect on the controlled struc­
is connected to the wall by the open screw hook. The detailed design of
ture. From the above conclusions, it can be concluded that the damping
the connection between the SMPD and the frame structure is shown in
effect of a tandem suspended mass pendulum damper is better, but
Fig. 12. The SMPDs are arranged in the middle of the floor to prevent the
considering the application in practical engineering, the suspended mass
controlled structure from twisting during the vibration.
pendulum damper with additional springs is more valuable for engi­
The overall test model of the three-layer steel frame with five types of
neering application because the suspended mass pendulum damper
SMPD installed on the top floor respectively is shown in Fig. 13. Fig. 14
without spring is limited by the floor height and cannot adjust its length
shows the vibration states of the controlled structure driven by the
of pendulum rigid boom flexibly to achieve the optimal damping effect.
shaking table during the test. The test equipment in the figure is a one-
The diagrams also makes it evident that the SMPD has a higher damping
way earthquake simulation shaker, and two acceleration sensors and
rate for the controlled structure for sites B and C, so it more suitable for
one laser displacement meter are evenly arranged along the vibration
these types of sites. After a comprehensive comparison, the series-
direction on the shaking table and each floor respectively to measure the
shaped suspended double mass pendulums damper has the best damp­
dynamic responses of the shaking table and the structure. The masses of
ing effect under seismic waves in four types of sites, and the TSDP-S has
the first to third floor of the steel frame are 24.47 kg, 23.78 kg, and
better engineering application value because it has the ability to adjust
23.78 kg. The first to third order frequencies of the three-layer steel
the pendulum length flexibly.
frame that was excited by white noise are shown in Fig. 15, and the value
is 3.97 Hz, 14.01 Hz, and 23.20 Hz, respectively. The structural damping
5. Test analysis for various types of SMPD ratio is 0.38% based on the free attenuation vibration method [33]. The
white noise excitation is again input to the three-layer steel frame once
In order to further verify the vibration damping performance of all the test cases have been finished. It is discovered that the structural
various types of SMPD on multi-layer structures, a three-layer steel frequency and damping are essentially unaltered, proving that there was
frame test model equipped with SSMPD, TSDP-S, PSDP-S, TSDP, and no elastic–plastic deformation of the frame during the test. The specific
PSDP is designed and tested on the shaking table to compare and analyze information of the steel frame model is shown in Table 4, and the
the dynamic response of the structure equipped with various types of pendulum length of each SMPD and the spring stiffness values are shown
SMPD under ground motions. in Table 5.

5.1. Experiment overview 5.2. Experimental results and analysis

The test frame is a three-layer steel structure model, the main frame To test the accuracy of the theoretical method, the time history re­
of the structural model equipped with each SMPD has the same di­ sults of the top dynamic response of the uncontrolled structure and the
mensions. All components are made of steel with yield strength 345 structure equipped with TSDP-S under simple harmonic excitation and
MPa, elastic modulus 206 GPa, and density 7850 kg/m3. The floor El Centro seismic waves are analyzed. The theoretical calculation of the

178
H. Sun et al. Structures 55 (2023) 169–184

Fig. 13. Shaking table test model diagram.

structural damping ratio is set at the same level as the experimentally and the shaking table test, the traditional dimensionless indexes root
measured structural damping ratio, which is taken as 0.38%. The TSDP- mean square error (RMSE) and mean absolute percentage error (MAPE)
S damping ratio is obtained according to Eq. (10), ζp1 = ζp2 = 0.36%, and is used to evaluate the error between the theoretical and test dynamic
the results are shown in Fig. 16 and Fig. 17. It can be found that the analysis data of the uncontrolled structure, and the related equations are
theoretical values are in high agreement with the shaking test values and shown in Eq. (17) and Eq. (18).
the theoretical analysis has good accuracy.
In order to quantify the deviation between the theoretical calculation

179
H. Sun et al. Structures 55 (2023) 169–184

(a) Structure equipped (b) Structure equipped (c) Structure equipped (d) Structure equipped
with TSDP-S with PSDP-S with TSDP with PSDP
Fig. 14. Diagram of test process.

Table 5
5 types of SMPD parameter information.
Types length of length of Spring Spring The The
of pendulum pendulum stiffness stiffness mass of mass of
SMPD l1(mm) l2(mm) kp1(N/ kp2(N/ mass mass
f mm) mm) block block
m1(kg) m2(kg)

SSMPD 14 – – – 0.96 –
TSDP-S 100 100 0.207 3.366 0.48 0.48
PSDP-S 100 100 3.690 0.300 0.48 0.48
TSDP 10 10 – – 0.48 0.48
PSDP 10 10 – – 0.48 0.48

f to the true value and is used to measure the accuracy of the measure­
ment, the smaller the value of the root mean square error, the higher the
accuracy of the measurement. MAPE is a statistical indicator of the ac­
curacy of the model, and the smaller the value of MAPE, the more ac­
f
curate the theoretical results. Table 6 shows the results of the
comparison of the error of the dynamic analysis data for the uncon­
trolled structure and controlled structure under each excitation, and the
results show that the theoretical calculation results fit well with the
shaking table test results.
Fig. 15. Frequency response curve of uncontrolled structure. The response history comparison of the top floor of the test structure
with various types of SMPD under the earthquake waves of El Centro
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ and Cale is shown in Fig. 18. It is clear that the structures equipped with
1∑ n
various types of SMPD all have a good damping effect under the seismic
RMSE = |yi − ̂y i |2 (17)
n i=1 waves, among which the damping effect of the TSDP-S is the best and the
damping effect of the TSDP is slightly worse, which fully proves the
n ⃒ ⃒
1∑ ⃒y i − ̂
y i ⃒⃒ conclusion obtained from the theoretical analysis that the tandem-
MAPE = ⃒ × 100% (18)
n i=1 ⃒ yi ⃒ shaped suspended double mass pendulums damper has a more supe­
rior damping performance and the effectiveness of the spring tuning
where n is the sample length, yi is the sample true value, and ŷi is the stiffness.
predicted value. The power spectra of the top acceleration response of the test
The RMSE characterises the degree of curve fit of the measured value structure equipped with various types of SMPD under the action of El

Table 4
Steel frame model information sheet.
Type of frame Floor dimensions/mm (length × width × Column cross section/ Column height/ Mass/ First-order Second-order
thickness) mm mm kg cycle/s cycle/s

Three-layer steel 500 × 500 × 6 50 × 8 400/300/300 57.35 0.252 0.0714

Frame

180
H. Sun et al. Structures 55 (2023) 169–184

Fig. 16. Comparison of theoretical and experimental values for uncontrolled structures.

Fig. 17. Comparison of theoretical and experimental values for controlled structures.

Table 6
Comparison of errors in theoretical calculations and shaking table test results.
Motivation Types of the structure Type of analysis Displacement Acceleration

Peak (m × 10-3) RMSE MAPE(%) Peak (m/s2) RMSE MAPE(%)

El Centro uncontrolled structures Theoretical calculations 6.9410 0.5437 1.0665 0.5643 0.0435 2.1342
shaking table test 6.2159 0.5096
El Centro controlled structures Theoretical calculations 5.6508 1.0705 2.4381 0.4496 0.1446 7.0302
shaking table test 4.6009 0.4665
3.9 Hz simple harmonic excitation uncontrolled structures Theoretical calculations 5.2042 1.2151 3.4774 – – –
shaking table test 5.3372 –
5.0 Hz simple harmonic excitation uncontrolled structures Theoretical calculations 9.0952 0.9575 1.2098 – – –
shaking table test 9.2399 –

Centro and Cale ground motions are compared, as shown in Fig. 19. The consistent with the results obtained from the theoretical analysis and
results show that all types of SMPD have good damping effect for the further proves the accuracy of the mechanical mechanism.
first-order vibration of the controlled structure, and the acceleration To further verify that the various types of SMPD have good damping
power attenuation at the first two frequencies of the TSDP-S control effect and robustness on the structure, the various types of site waves
system is the most obvious, i.e., the TSDP-S is most effective in con­ selected in the previous section are used as shaking table excitation, and
trolling the first two orders of frequency of the structure and achieves the average peak damping rate and average energy damping rate of
the double-order tuned damping. The acceleration power attenuation at different forms of SMPD under various types of site ground shaking were
the first two frequencies of the PSDP-S control system is also obvious, finally obtained, and the comparison results are shown in Fig. 20. It can
but its damping effect is not satisfactory. It can be proved that the sus­ be seen that the damping rate of TSDP-S is the highest under various
pended mass pendulum damper with additional spring can more effec­ types of site waves, followed by TSDP, and the damping rate of SSMPD is
tively achieve multi-order tuned damping, but the tuned damping effect basically the same as that of PSDP-S and PSDP, and this result is basically
of the parallel-shaped suspended mass pendulum is not as good as the consistent with the theoretical value in the previous section, i.e., the
tandem-shaped suspended mass pendulum form. The structures equip­ damping performance of TSDP-S is the best, followed by TSDP, and the
ped with the remaining types of SMPD also have attenuated energy in damping performance of the remaining three types of SMPD is not much
the frequency domain under the first two orders of vibration relative to different. In sites B and C, the damping rates of the SMPD are higher than
the uncontrolled structure, but not as much as the TSDP-S, which is other sites, which is consistent with theoretical results, thus, this type of

181
H. Sun et al. Structures 55 (2023) 169–184

·
·

(a) The top layer displacement time-history curve (b) The top layer acceleration time-history curve
of the structure under El Centro wave of the structure under El Centro wave

· ·

(c) The top layer displacement time-history curve (d) The top layer acceleration time-history curve
of the structure under Cale wave of the structure under Cale wave
Fig. 18. Comparison of the top layer dynamic time-history curve of each model.

Fig. 19. Power spectral density curve of the acceleration response at the top of the experimental structure.

damper is more suitable for sites B and C. 6. Conclusions

The test results further verify the effectiveness of various types of
SMPD for damping effect of the multilayer structures, and verify the The traditional type of suspended mass pendulum damper has good
feasibility of the SMPD equipped with additional springs for multi-order damping performance, but it has the problems such as inflexible
tuning control of structures, the test results is consistent with the theo­ adjustment of pendulum length and narrow damping frequency band, so
retical derivation of this paper. Due to the limitation of the pendulum it is greatly restricted in the practical application of engineering. In this
length, the SMPD equipped with additional springs are more suitable for study, a series and parallel suspended mass pendulum with additional
practical engineering applications, while the pendulum length of the springs is proposed, which can achieve the optimal damping effect by
SMPD without additional springs are too short and are not applicable in adjusting the stiffness of the springs, the mass of the mass block, and the
practical engineering applications, although they also have a certain length of the pendulum rigid boom, and realize the tuned damping
damping effect. control of multi-order vibration pattern, which is more suitable for

182
H. Sun et al. Structures 55 (2023) 169–184

Fig. 20. Average damping rate of four types of sites for experimental structures.

engineering practice because the pendulum length can be flexibly shaped suspended double mass pendulums is equal to that of the
adjusted. Five different types of SMPD are selected for theoretical and SSMPD because the control performance and damping effect are low due
experimental studies, and the dynamic equations of various types of to the lack of coupling effect between the two pendulums, but the PSDP-
suspended double mass pendulums are established and analyzed for S is more suitable for practical engineering applications because of its
their dynamic characteristics, and their frequency analytical solutions ability to flexibly adjust the pendulum length, increase the tuning fre­
are obtained to explain their damping mechanisms. On this basis, the quency band and multi-order vibration tuning damping.
overall dynamic equations of the structure with each type of suspended (2) The multi-layer frame shaking table test further verified the
double mass pendulums are established. In order to verify the damping effectiveness of various types of SMPD on the damping control of the
performance of the series and parallel suspended double mass pendu­ multi-layer structure, and the TSDP-S had the best damping effect and
lums with additional springs and their multi-order vibration tuning could effectively control the first two order cycles of the structure. The
damping capability, a 10-layer frame model is established for the dy­ theoretical values are in good agreement with the experimental values,
namic time course theoretical analysis, and a 3-layer steel frame struc­ which fully verify the rationality and accuracy of the numerical analysis
ture is constructed for shaking table model tests, and the following of the dynamics equations of each type of SMPD proposed in this paper.
conclusions are obtained after comprehensive comparison by analyzing (3) For structures affected by multi-order vibration patterns, the
the dynamic time course response and power spectrum density under damping effect of the two tandem-shaped suspended multiple mass
the action of different seismic waves: pendulums damper is obvious and can be selected in preference without
(1) The series and parallel suspended multiple mass pendulums considering the limitation of pendulum length. In engineering practice,
damper with additional springs can achieve tuned damping of the the TSDP-S and PSDP-S are more suitable for multi-layer damping
controlled structure by adjusting multiple parameters such as the spring control due to their more flexible ability to adjust the pendulum length.
stiffness, the mass of the mass block, and the pendulum length, which Compared to the other site categories, the damping rate of the controlled
can flexibly adjust its pendulum length and achieve multi-order vibra­ structure is higher for the 5 types of SMPD in sites B and C than the other
tion tuning control of the structure. Theoretical analysis shows that sites, i.e. the SMPD are more suitable for use in sites B and C. Consid­
among the five types of SMPD, the damping efficiency of the two ering all aspects, the TSDP-S has the best engineering application
tandem-shaped suspended multiple mass pendulums damper is higher, prospect among the five types of SMPD because it can flexibly adjust the
but the damping effect of the TSDP-S is better because it can flexibly pendulum length to solve the engineering limitation problem, broaden
adjust the stiffness by springs. The damping effect of the two parallel- the tuning frequency band, realize multi-order tuning damping, and

183
H. Sun et al. Structures 55 (2023) 169–184

achieve the best damping effect. [12] Roffel AJ, Narasimhan S. Extended Kalman filter for modal identification of
structures equipped with a pendulum tuned mass damper. J Sound Vib 2014;333
How to use multiple types of dampers in combination with sus­
(23):6038–56. https://doi.org/10.1016/j.jsv.2014.06.030.
pended multiple mass pendulums damper so as to optimize their tuned [13] Roffel AJ, Narasimhan S. Results from a full-scale study on the condition
damping capability to the maximum extent still needs to be further assessment of pendulum tuned mass dampers. J Struct Eng 2016;142(1):15–24.
studied. Optimized design of parameters such as the pendulum length, https://doi.org/10.1061/(ASCE)ST.1943-541X.0001339.
[14] Pasala DTR, Nagarajaiah S. Adaptive-length pendulum smart tuned mass damper
the mass of the mass block and the spring stiffness based on engineering using shape-memory-alloy wire for tuning period in real time. Smart Struct Syst
reality also needs to be further carried out. 2014;13(2):203–17.
[15] Huang C, Huo L-S, Gao H-G, Li H-N. Control performance of suspended mass
pendulum with the consideration of out-of-plane vibrations. Struct Control Hlth
Declaration of Competing Interest 2018;25(9):e2217.
[16] Huang C, Huo LS, Li HN. Equivalent linearization and parameter optimization of
isochronous-pendulum tuned mass damper. Adv Struct Eng 2023;26(2):234–57.
The authors declare that they have no known competing financial https://doi.org/10.1177/13694332221120698.
interests or personal relationships that could have appeared to influence [17] Huo LS, Zhang YW, Huang C, Hongnan LI. Research on vibration control of large-
span spatial structures with a passive adaptive suspended mass pendulum.
the work reported in this paper. J Disaster Prevention and Mitigation Eng 2021;41(05):968–76.
[18] Setareh M, Ritchey JK, Baxter AJ, Murray TM. Pendulum tuned mass dampers for
Acknowledgments floor vibration control. J Perform Constr Fac 2006;20(1):64–73.
[19] Setareh M, Ritchey JK, Murray TM, Koo J-H, Ahmadian M. Semiactive tuned mass
damper for floor vibration control. J Struct Eng 2007;133(2):242–50.
This work was partially supported by National Natural Science [20] Wang LK, Shi WX, Zhou Y. Study on self-adjustable variable pendulum tuned mass
Foundation of China [Grant No. 51878017]. damper. Struct Des Tall Spec 2019;28(1):e1561.
[21] Eason RP, Sun C, Dick AJ, Nagarajaiah S. Steady-state response attenuation of a
linear oscillator-nonlinear absorber system by using an adjustable-length
References pendulum in series: numerical and experimental results. J Sound Vib 2015;344:
332–44.
[1] Weber H, Kaczmarczyk S, Iwankiewicz R. Non-linear dynamic response of a cable [22] Sun C, Nagarajaiah S, Dick AJ. Experimental investigation of vibration attenuation
system with a tuned mass damper to stochastic base excitation via equivalent using nonlinear tuned mass damper and pendulum tuned mass damper in parallel.
linearization technique. Meccanica 2020;55:2413–22. https://doi.org/10.1007/ Nonlinear Dynam 2014;78(a):2699–715. https://doi.org/10.1007/s11071-014-
s11012-020-01169-3. 1619-3.
[2] Ashasi-Sorkhabi A, Malekghasemi H, Ghaemmaghami A, Mercan O. Experimental [23] Yan WM, Ji JB, Jiang HG, et al. A new type pendulous TMD and its application on a
investigations of tuned liquid damper-structure interactions in resonance tower for wind-induced vibration control. J Build Struct 2010;31(2):55–60.
considering multiple parameters. J Sound Vib 2017;388:141–53. [24] Lu XL, Chen JR. Parameter optimization and structural design of tuned mass
[3] Ontiveros-Perez SP, Miguel LFF. Reliability-based optimum design of multiple damper for Shanghai centre tower. Struct Des Tall Spec 2011;20(4):453–71.
tuned mass dampers for minimization of the probability of failure of buildings https://doi.org/10.1002/tal.649.
under earthquakes. Structures 2022;42:144–59. https://doi.org/10.1016/j. [25] Miguel LFF, Lopez RH, Torii AJ, Beck AT. Reliability-based optimization of
istruc.2022.06.015. multiple Folded Pendulum TMDs through Efficient Global Optimization. Eng Struct
[4] Ershadbakhsh AM, Ghorbani-Tanha AK, Fallahi R. A novel active tendon pendulum 2022;266:114524. https://doi.org/10.1016/j.engstruct.2022.114524.
tuned mass damper and its application in transient vibration control. Structures [26] Roffel AJ, Lourenco R, Narasimhan S, Yarusevych S. Adaptive compensation for
2023;47:2273–80. https://doi.org/10.1016/j.istruc.2022.12.037. detuning in pendulum tuned mass dampers. J Struct Eng 2011;137(2):242–51.
[5] Hoang N, Fujino Y, Warnitchai P. Optimal tuned mass damper for seismic [27] Gerges RR, Vickery BJ. Parametric experimental study of wire rope spring tuned
applications and practical design formulas. Eng Struct 2008;30(3):707–15. https:// mass dampers. J Wind Eng Ind Aerod 2003;91(12):1363–85. https://doi.org/
doi.org/10.1016/j.engstruct.2007.05.007. 10.1016/j.jweia.2003.09.038.
[6] Tait MJ. Modelling and preliminary design of a structure-TLD system. Eng Struct [28] Qin L, Yan WM, Li YX. Design of frictional pendulum TMD and its wind control
2008;30(10):2644–55. https://doi.org/10.1016/j.engstruct.2008.02.017. effectiveness. Earthq Eng Eng Vib 2009;29(05):153–7.
[7] Li HN, Song BY. Seismic response reduction for tall buildings by suspended mass [29] Hori Y, Kurino H, Kurokawa Y. Development of large tuned mass damper with
pendulums. Earthq Eng Eng Vib 1995;15(4):55–61. stroke control system for seismic upgrading of existing high-rise building. Int J
[8] Li HN. Study of vibration damping properties of Pendulum-structure system. Eng High-Rise Build 2016;5(3):167–76. https://doi.org/10.21022/
Mech 1996;13(3):242–50. IJHRB.2016.5.3.167.
[9] Kourakis I. Structural systems and tuned mass dampers of super-tall buildings: case [30] Huber P. Seismic damping systems with TMDs for high rise buildings. Santiago:
study of Taipei 101. Massachusetts Institute of Technology 2007. 16th World Conference on Earthquake Engineering 2017; 41.
[10] Tuan AY, Shang GQ. Vibration control in a 101-storey building using a tuned mass [31] Den Hartog. JP. Mechanical vibrations. New York: 4th ed., McGraw-Hill. 1956.
damper. J Appl Sci Eng 2014;17(2):141–56. https://doi.org/10.6180/ [32] ASCE/SEI 7-16. Minimum design loads and associated criteria for buildings and
jase.2014.17.2.05. other structures. American Society of Civil Engineers 2017.
[11] Roffel AJ, Narasimhan S, Haskett T. Performance of pendulum tuned mass dampers [33] Lu Z, Chen XY, Zhou Y. An equivalent method for optimization of particle tuned
in reducing the responses of flexible structures. J Struct Eng 2013;139(12):35–8. mass damper based on experimental parametric study. J Sound Vib 2018;49:
https://doi.org/10.1061/(ASCE)ST.1943-541X.0000797. 571–84. https://doi.org/10.1016/j.jsv.2017.05.048.

184