Journal of Sound and Vibration 421 (2018) 153e165

Contents lists available at ScienceDirect

Journal of Sound and Vibration

journal homepage: www.elsevier.com/locate/jsvi

Experimental and analytical study on vibration control effects

of eddy-current tuned mass dampers under seismic
excitations
Zheng Lu a, b, Biao Huang b, Qi Zhang b, Xilin Lu a, b, *
a
State Key Laboratory of Disaster Reduction in Civil Engineering, Tongji University, Shanghai, 200092, China
b
Research Institute of Structural Engineering and Disaster Reduction, Tongji University, Shanghai, 200092, China

a r t i c l e i n f o a b s t r a c t

Article history: Eddy-current tuned mass dampers (EC-TMDs) are non-contacting passive control devices
Received 18 August 2017 and are developed on the basis of conventional tuned mass dampers. They comprise a solid
Received in revised form 20 October 2017 mass, a stiffness element, and a damping element, wherein the damping mechanism
Accepted 23 October 2017
originates from eddy currents. By relative motion between a non-magnetic conductive
Available online 12 February 2018
metal and a permanent magnet in a dynamic system, a time-varying magnetic field is
induced in the conductor, thereby generating eddy currents. The eddy currents induce a
Keywords:
magnetic field with opposite polarity, causing repulsive forces, i.e., damping forces. This
Eddy-current tuned mass damper
Tuned mass damper
technology can overcome the drawbacks of conventional tuned mass dampers, such as
Passive control limited service life, deterioration of mechanical properties, and undesired additional
Shaking table test stiffness. The experimental and analytical study of this system installed on a multi-degree-
Analytical method of-freedom structure is presented in this paper. A series of shaking table tests were con-
Structural control ducted on a five-story steel-frame model with/without an EC-TMD to evaluate the effec-
tiveness and performance of the EC-TMD in suppressing the vibration of the model under
seismic excitations. The experimental results show that the EC-TMD can effectively reduce
the displacement response, acceleration response, interstory drift ratio, and maximum
strain of the columns under different earthquake excitations. Moreover, an analytical
method was proposed on the basis of electromagnetic and structural dynamic theories. A
comparison between the test and simulation results shows that the simulation method can
be used to estimate the response of structures with an EC-TMD under earthquake exci-
tations with acceptable accuracy.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction

Since Yao introduced the concept of vibration control in civil engineering in 1972 [1], relevant structural control theories
and methods have been significantly developed and have been proven to be economical and efficient in practice. The various
control strategies, proposed by several researchers worldwide, can be classified into active control, passive control, hybrid
control, and semi-active control, among which passive control is the prevalent strategy owing to its simplicity and lack of
reliance on additional energy input [2]. Among numerous passive control devices, the tuned mass damper (TMD), wherein

* Corresponding author. State Key Laboratory of Disaster Reduction in Civil Engineering, Tongji University, Shanghai, 200092, China.
E-mail addresses: [email protected] (Z. Lu), [email protected] (B. Huang), [email protected] (Q. Zhang), [email protected] (X. Lu).

https://doi.org/10.1016/j.jsv.2017.10.035
0022-460X/© 2018 Elsevier Ltd. All rights reserved.
154 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165

the input energy is dissipated by tuning the frequency, is widely used owing to its simple characteristics, convenient
installation, low cost, and favorable control effects at specific tuning frequencies [3].
The damping component of a TMD is mostly provided by conventional fluid viscous dampers. However, in practice, several
issues arise in viscous dampers. In particular, these dampers may leak over time, thus limiting the service life. The physical
properties may degrade in high-temperature environments, resulting in failure [4,5]. The stiffness of the overall system in-
creases because of the viscous dampers, which is unexpected for some structures. Moreover, it is difficult or rather expensive
to change the damping ratio of the TMD once the viscous dampers are installed. Consequently, some innovative damping
mechanisms have been developed to address the shortcomings of conventional TMD technology. For example, some re-
searchers replaced viscous damping with particle damping. However, the damping mechanism of particle dampers has not
been thoroughly understood because of their high nonlinearity [6e13]. Another simple and effective solution involves using
eddy current damping.
Eddy current damping is a non-contacting damping mechanism which is suitable for solving the aforementioned prob-
lems. Because the device in eddy current damping is usually made of metal, it hardly degrades in its life cycle, making it useful
for high-temperature applications. Owing to the non-contacting property, eddy current damping does not lead to an increase
in the overall stiffness of the structures; therefore, the stiffness characteristics of the structures remain unaffected.
Furthermore, the damping ratio can be easily adjusted by varying the air gap between the permanent magnet and the
conductor.
The damping mechanism via eddy currents has been reported by many researchers [14e18]. By the relative motion be-
tween a non-magnetic conductive metal and a permanent magnet or by changing the strength of the field magnets, a time-
varying magnetic field is induced in the conductor, thereby generating eddy currents (shown in Fig. 1). The eddy currents
induce another magnetic field with opposite polarity, thereby causing repulsive forces, i.e., damping forces. Because of the
electrical resistance of the conductor, the induced currents dissipate in the form of heat at the rate of I2R, where I and R
represent the current intensity and electrical resistance of the conductor, respectively. Therefore, the vibration energy of the
main structures is transferred into the conductor and is dissipated in the form of heat.
Eddy current damping is largely employed in the field of magnetic backing, and another application involves suppressing
the lateral vibration in rotor shafts [20], which was reviewed by Henry and Bae [16]. In the field of structural vibration control,
Matsuzaki et al. presented a vibration control system wherein the motion of a beam with an eddy current damper (ECD) could
be suppressed. They derived a theoretical solution for the beam and demonstrated the viability of the concept [21]. In another
study, an experiment was conducted to demonstrate the effectiveness of the ECD. Their results show that the electromagnetic
force is capable of damping the first few modes of vibration of the beam [22]. In another study, Zheng et al. analyzed the
effectiveness of an ECD in suppressing the free vibration response of a beam by numerical simulation and found that the
damping effect of the ECD varies with the amplitude of the beam. The larger the amplitude, the greater is the effective
damping [23]. Kwak et al. developed a new type of ECD to suppress the vibration of a beam. The ECD comprised a copper
plate, which was rigidly fixed onto the end of the beam, and two permanent magnets. The ECD was constructed and tested to
analyze the efficiency [15]. Later, Bae at el. developed the theoretical model of the ECD proposed by Kwak et al. and inves-
tigated the damping performance of a new ECD model. The experimental and simulation results demonstrate the potential of
the ECD in controlling the vibration of a cantilever beam [14]. Sodano et al. proposed a new damper configuration providing
more damping to the beam structure and developed an improved theoretical model with enhanced accuracy [24]. The

Fig. 1. Eddy current induced by the relative motion between a non-magnetic conductive metal and a permanent magnet [19].
 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165 155

aforementioned studies primarily concentrated on the evaluation and analysis of ECDs in suppressing the vibration of
cantilever beams with small-scale models.
Kienholz et al. combined the advantages of the ECD and tuned mass damper and developed an eddy-current tuned mass
damper (EC-TMD) to suppress the vibration of a solar sail array. They found that the magnetic damping techniques can
achieve high damping forces [25]. In another study, Larose et al. investigated the effectiveness of the EC-TMD in suppressing
the vibration of a small-scale bridge model by conducting wind tunnel tests. With the EC-TMD properly tuned, the global
vibration of the bridge model could be damped [26]. Bae et al. confirmed the effective damping of an EC-TMD in suppressing
the vibration of a beam model through both simulations and experiments [27]. In recent years, studies have been conducted
on large-scale models with EC-TMDs. Wang et al. studied the feasibility of a large-scale EC-TMD and found that the eddy
current damping was the dominant damping source. In addition, they demonstrated that the damping ratio could be easily
adjusted by varying the gap between the permanent magnets and the conductive plates [28]. Bae et al. proposed a simple yet
effective method for suppressing the vibration of a large beam structure using an EC-TMD without substantially increasing
the overall weight [29]. However, to improve the practical application of ECDs to engineering structures such as high-rise
buildings, it is necessary to analyze the damping performance in structural models by conducting large-scale tests.
In this study, the damping performance of an EC-TMD installed on top of a five-story steel-frame model was evaluated by
conducting an experimental study and determining an analytical solution. Shaking table tests conducted on the large-scale
five-story frame structure are first introduced. An analytical solution for the damping performance of the EC-TMD, derived on
the basis of electromagnetic and structural dynamic theories, is then discussed. Finally, a computational method is presented,
simulated, and validated.

2. Experimental design

2.1. Model design

The test model is a five-story steel-frame structure with a height of 5.48 m, comprising planes and columns, as shown in
Fig. 2(a). The columns were made of Q690 steel whereas the planes were made of Q345 steel. The dimensions of the columns
are 15 mm
180 mm
1060 mm and those of the planes are 30 mm
2000 mm
2000 mm. The total mass of the test model
is 6000 kg. The seismic excitations were inputted to the structure along the weak axis direction in the shaking table tests. The
natural vibration periods of the first five modes along the weak axis were 1.00, 0.35, 0.22, 0.17, and 0.15 s, respectively. Two
acceleration sensors and two displacement sensors were set up in each floor. The acceleration and displacement sensors are
numbered from 4 to 13 and from 76 to 85, respectively, as shown in Fig. 2(b). Four strain sensors were set up at the bottom of
the columns to monitor the maximum strain of the columns; the strain sensors were numbered from 101 to 104.

Fig. 2. Model configuration. (a) Image of the test model, and (b) sensor arrangement.
156 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165

2.2. Damper design

The EC-TMD system, manufactured by the Shanghai Research Institute of Materials, includes a supporting frame, four
cables, a tuning mass, a permanent magnet, and a copper plane, as shown in Fig. 3. The supporting frame was fixed on top of
the model. The lateral stiffness of the supporting frame was significantly greater than that of the model, thus ensuring that the
supporting frame moves along with the top of the model. The tuning mass with a permanent magnet was suspended from the
top of the supporting frame using four cables. Between the tuning mass and the steel frame model, the copper plane was fixed
using four small columns. To improve the vibration-reduction effects, the frequency of the EC-TMD should be designed as
close as possible to the first mode frequency of the primary structure. Hence, the length L of the suspended cables was
decided by the classical equation for a simple pendulum: L ¼ ðT=2pÞ2 $g ¼ 0:248 m. The EC-TMD is primarily used for
adjusting the damping properties of the auxiliary mass; however, it is not providing stiffness tuning, which is provided by the
coupling structural member.
In engineering practice, the mass of a tall building is considerable. Therefore, the mass ratio of the damper to the primary
structure should be small. For example, a TMD with a mass ratio of 1% was attached to the Shanghai Center Tower in China
[30], and a tuned liquid damper (TLD) with a mass ratio of 1% was attached to the Shin Yokohama Prince hotel in Japan [31].
On the basis of these practical cases, the mass ratio of the damper to the model structure in the tests was chosen to be
approximately 1%.

2.3. Excitations in shaking table tests

Fig. 4 shows the time histories and the corresponding power spectral density curves of three earthquake waves, including
the El Centro, Wenchuan, and Japan 311 waves. The earthquake waves were incorporated in the shaking table tests to evaluate
the damping performance of the EC-TMD system under different seismic conditions. The seismic excitations were inputted in
only the weak axis direction of the test model, with peak ground accelerations of 0.05, 0.10, and 0.15 g, respectively (g is the
acceleration due to gravity).

3. Response analysis

To evaluate the vibration control performance of the damper, the vibration damping effect is defined as follows.

the pðRMSÞ of structure without dampers  the pðRMSÞ of structure with an EC  TMD
vibration damping effect ¼
the pðRMSÞ of structure without dampers

100%
Here, the p and RMS represent the peak and root-mean-square values of the displacement/acceleration response at the top of
the model, respectively.

3.1. Displacement response

The RMS value of the displacement/acceleration is an indicator of the vibration energy and is usually employed to indicate
the random variable energy in random vibration. The peak value of the displacement/acceleration is of concern in engineering

Fig. 3. Damper configuration.

 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165 157

Fig. 4. Time histories of (a) El Centro, (b) Wenchuan, and (c) Japan 311 waves; Power spectral densities of (d) El Centro, (e) Wenchuan, and (f) Japan 311 waves.

design, because it is related to the safety of buildings and the comfort of the occupants. Table 1 lists the peak and RMS values
of the displacement responses at the roof of the test model in different cases. With regard to the vibration control effects in
reducing the displacement responses, the results show that (i) The response of the controlled model with an EC-TMD is lower
than that of the uncontrolled model, thus implying that the EC-TMD can dissipate vibration energy and provide effective
damping for the test model; (ii) The vibration-reduction effect in terms of the RMS values of the displacement responses is
significantly better than that in terms of the peak values, wherein the range of the former is 18.6e69.4% and that of the latter
is 10.4e34.9%; (iii) The vibration-reduction effect is different under different seismic inputs. The reduction in vibration is
favorable in the case of the El Centro wave; however, the damping effects are relatively poorer in the cases of the Wenchuan
and Japan 311 waves.
This may be because of the frequency characterization of the input earthquake excitations. Fig. 4 shows the excitations of
the El Centro, Wenchuan, and Japan 311 waves in the time and frequency domains. The frequency of the El Centro wave is
largely in the range of 0.5e2 Hz, which is near the natural frequency of the primary structure, whereas the differences be-
tween the dominant frequencies of the Wenchuan and Japan 311 waves and the fundamental frequency of the structure are
considerable. This frequency characterization makes the displacement response of the model in the case of the El Centro wave
to be significantly greater than that in the cases of the other inputs. Accordingly, in the case of the El Centro wave, the EC-TMD
dissipates more energy via mass tuning and eddy currents, thereby exhibiting a more favorable damping effect.

3.2. Acceleration response

Not only can the displacement be reduced, the acceleration can also be attenuated in most cases. Table 2 lists the peak and
RMS values of the acceleration responses at the roof of the test model in different cases. In all cases of the RMS values and
most cases of the peak values, the acceleration responses of the controlled model are lower than those of the uncontrolled
model. This shows that the EC-TMD can effectively dissipate the vibration energy of the test model. However, the vibration-
reduction performance of the model with the EC-TMD in the case of the Japan 311 wave is the worst, particularly for peak
ground accelerations of 0.05 and 0.15 g where the peak acceleration slightly increases.
This is because although the EC-TMD is a modified version of the conventional TMD and therefore is expected to exhibit
improved performance, the EC-TMD is still a linear damper, the performance of which is influenced by the frequency of the
excitation, similar to the conventional TMD. This is observed in both Table 1 (Section 3.1) and Table 2 (Section 3.2). In fact, the
key advantage of the EC-TMD is the non-contacting damping mechanism. Moreover, from Tables 1 and 2, the vibration
damping effect in terms of the RMS values is better than that in terms of the peak values in most cases. This is clarified by
analyzing the time history of the displacement/acceleration response in the following section.
158 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165

Table 1
Peak and RMS values of the displacement response at the roof of the test model.

Seismic input Peak input value (g) Without dampers With an EC-TMD Reduction effect

p (mm) RMS (mm) p (mm) RMS (mm) p (%) RMS (%)

El Centro 0.05 34.008 13.125 30.456 4.121 10.4 68.6
0.1 79.213 29.810 68.882 9.124 13.0 69.4
0.15 118.925 43.056 102.521 13.662 13.8 68.3
Wenchuan 0.05 8.924 1.806 7.048 1.470 21.0 18.6
0.1 18.537 4.082 14.738 3.219 20.5 21.1
0.15 27.305 6.262 22.082 4.788 19.1 23.5
Japan 311 0.05 4.475 1.415 3.449 0.996 22.9 29.6
0.1 10.842 4.089 8.797 2.517 18.9 38.5
0.15 22.438 8.273 14.602 4.250 34.9 48.6

3.3. Time histories and power spectral densities

The entire time history and power spectral density of the displacement/acceleration response are necessary for evaluating
the damage inflicted to the structures and can be regarded as indicators of the intensity of the earthquake ground motion to
some extent. Fig. 5 shows the displacementetime history and power spectral density curves at the roof level of the test model,
wherein the solid and dotted lines represent the controlled and uncontrolled models, respectively. From the time history
curves, it can be found that the vibration control effect is poor at the very beginning, and the controlled curves begin to decay
quickly thereafter. If the peak values are observed at the very beginning, the damping performance in terms of the peak values
will be poor; this is one of the reasons why the peak values of the responses is high in some cases and the vibration-reduction
effects in terms of the RMS values are better than those in terms of the peak values listed in Tables 1 and 2. At the beginning, it
takes some time for the EC-TMD to start moving and dissipate the vibration energy. Fig. 5(a) and (b) show that the peak values
of the test model in the cases of the El Centro and Wenchuan wave excitations are 118.9 and 27.3 mm, respectively. Therefore,
it was easier for the EC-TMD to create the damping effect in the case of the El Centro wave.
Fig. 5(c) and (d) show the corresponding power spectral density curves in the cases of the El Centro and Wenchuan waves,
respectively. Fig. 5(c) shows that the vibration energy is significantly high at the fundamental frequency of the test model.
Moreover, the peak value of the model with the EC-TMD is significantly lower than that without the dampers, thus indicating
that the EC-TMD has favorable damping effects at the fundamental frequency. Fig. 5(d) shows that the vibration energy is
significantly high at the first and second mode frequencies of the test model. Because of the additional mass of the EC-TMD,
the structural frequencies of the controlled model are slightly lower than those of the uncontrolled model. The peak values of
the model with the EC-TMD are significantly lower than those of the model without the dampers. This is because the EC-TMD
provides effective damping for the test model and suppresses the vibration. In most cases, the displacement responses of the
model with the EC-TMD are lower than those of the model without the dampers. This is observed not only in the time history
curves but also in the power spectral density curves.

3.4. Interstory drift ratio

The maximum interstory drift ratio of a building is an important index for structural design and should not exceed
specified ultimate values in the relevant design codes. Fig. 6 shows the maximum interstory drift ratio of the test model in the
case of the El Centro wave with different peak ground accelerations (0.05, 0.1, and 0.15 g). Although the amplitudes of the
excitations are different, the maximum interstory drift ratios with respect to the five stories of the test model are reduced
using the EC-TMD in all cases. The damping performance of the EC-TMD for a peak ground acceleration of 0.15 g is better than
those for peak ground accelerations of 0.05 and 0.1 g. This is because the vibration response of the test model in the former
case is greater than those in the other cases. Therefore, the EC-TMD moves more violently and dissipates more vibration
energy.

3.5. Maximum strain of the columns

The maximum strain of the columns is a parameter used to check whether the structure is in the elastic or elastoplastic
phase, in relation to the damage phenomena. Table 3 lists the maximum strains of the columns of the test model with/without
the EC-TMD for all cases. It can be seen that: (i) The maximum strain of the model in the case of the El Centro wave is the
highest, and the maximum strain in the case of the Japan 311 wave is the lowest, which coincides with the results of the
displacement and acceleration responses, given in Sections 3.1 and 3.2, respectively; (ii) The maximum strains of the columns
in all the cases do not exceed the yield strain of the Q690 steel columns, indicating that the test model remains in an elastic
phase during the shaking table tests; (iii) The maximum strain of the columns can be reduced using the EC-TMD.
 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165 159

Table 2
Peak and RMS values of the acceleration response at the roof of the test model.

Seismic input Peak input value (g) Without dampers With an EC-TMD Reduction effect

p (g) RMS (g) p (g) RMS (g) p (%) RMS (%)

El Centro 0.05 0.219 0.059 0.142 0.021 35.1 63.6
0.1 0.508 0.134 0.328 0.051 35.3 61.7
0.15 0.757 0.193 0.501 0.089 33.8 53.8
Wenchuan 0.05 0.201 0.047 0.186 0.033 7.4 30.2
0.1 0.450 0.108 0.407 0.071 9.4 34.2
0.15 0.665 0.161 0.631 0.113 5.2 29.9
Japan 311 0.05 0.108 0.034 0.115 0.022 6.7 36.2
0.1 0.271 0.080 0.261 0.053 3.9 33.3
0.15 0.458 0.141 0.490 0.097 6.8 31.6

Fig. 5. Time history curves of the displacement responses of the model in the cases of (a) El Centro, and (b) Wenchuan waves; Power spectral density curves of
the displacement responses of the model in the cases of (c) El Centro, and (d) Wenchuan waves.

The series of shaking table tests conducted on the model with/without the EC-TMD, along with the analysis of the
experimental results, prove that the EC-TMD performs effectively in suppressing the vibration of a large-scale primary
structure.
In the following section, a corresponding analytical/simulation solution is proposed to estimate the response of the
structures with the EC-TMD, and the simulation results are analyzed by comparing the experimental results.

4. Analytical/simulation solution

4.1. Analytical method

The magnetic field distribution with respect to the rectangular permanent magnet is first introduced, and subsequently,
the electromagnetic force is analyzed. The effects of the electromagnetic force can be reflected in the form of the eddy-current
damping coefficient, which is a function of the relative distance between the permanent magnet and the conductor. The
eddy-current damping coefficient can be then incorporated into the governing equations of the primary system.
160 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165

Fig. 6. Maximum interstory drift ratios of the test model in the case of the El Centro wave with peak ground accelerations of (a) 0.05, (b) 0.1, and (c) 0.15 g.

Table 3
Maximum strains of the columns in the test model.

Seismic input Peak input value (g) Uncontrolled (mε) Controlled (mε) Reduction effect (%)
El Centro 0.05 338 279 17.46
0.1 776 633 18.43
0.15 1163 958 17.63
Wenchuan 0.05 193 139 27.98
0.1 443 307 30.70
0.15 676 473 30.03
Japan 311 0.05 75 67 10.67
0.1 174 157 9.77
0.15 323 296 8.36

Gou et al. [32] derived analytical expressions to determine the magnetic field distribution of a rectangular permanent
magnet. To complete this analytical method, some key equations to calculate the magnetic field distribution are given and

Fig. 7. Coordinate system in a rectangular permanent magnet.

 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165 161

derived herein. The dimensions of the permanent magnet are assumed to be a mm (length)
b mm (width)
c mm (height),
and the direction of polarity of the magnet is along the z axis, as shown in Fig. 7.
The components of the magnetic field distribution along the three axes can be derived on the basis of the molecular
circulation model and the BioteSavart law.

Zh
N
Bx ¼ dBx ¼  ½Gða  x; y; zÞ þ Gða  x; b  y; zÞ  Gðx; y; zÞ  Gðx; b  y; zÞ (1)
2
0

Zh
N
By ¼ dBy ¼  ½Gðb  y; x; zÞ þ Gðb  y; a  x; zÞ  Gðy; x; zÞ  Gðy; a  x; zÞ (2)
2
0

Zh
Bz ¼ dBz
0
¼ N½Jðx; b  y; zÞ þ Jðy; a  x; zÞ þ Jða  x; b  y; zÞ þ Jðb  y; a  x; zÞ þ Jðb  y; x; zÞ þ Jðy; x; zÞ
þ Jða  x; y; zÞ þ Jðx; y; zÞ (3)

The auxiliary functions are defined as follows.

m0 J s
N¼ (4)
4p
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 

g21 þ g22 þ ðg3  z0 Þ2  g2  z0 ¼ h
Gðg1 ; g2 ; g3 Þ ¼ ln qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  (5)
z ¼0
g21 þ g22 þ ðg3  z0 Þ2 þ g2  0

2 3

j3  z0 
6j1 7 z 0 ¼ h
Jðj1 ; j2 ; j3 Þ ¼ arctan4 $qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi5 (6)
j2 j2 þ j2 þ ðj  z Þ2  z0 ¼ 0
1 2 3 0

where m0 ð¼ 4p
107 Þ and Js ð¼ 9:55
105 A=m2 Þ represent the magnetic permeability in the air and the equivalent current
density, respectively. The magnetic field distribution of the rectangular permanent magnet can be calculated using the
aforementioned equations. Wang et al. determined the magnetic field distribution along the z direction at the height of 5.5 cm
and found that the magnetic field is largely distributed on the projected area of the permanent magnet and decayed rapidly
outside this area [28]. Compared to the magnetic field distribution in the projected area, the magnetic field outside this area is
insignificant and therefore can be neglected, thus simplifying the analytical method and improving the efficiency of the
calculation.
When a conductive metal passes through a magnetic field, eddy current forces are created in the conductor. It is assumed
that the EC-TMD tunes along the y axis. The velocity of the EC-TMD is expressed as follows.

vTMD ¼ 0i þ vy j þ 0k: (7)

The current density J can be calculated as follows.

2 3
i j k
J ¼ sðv
BÞ ¼ s4 0 vy 0 5 ¼ svy ðBz i  Bx kÞ (8)
Bx By Bz

where s is the conductivity of the conductor in the magnetic field, and B is the magnetic field strength at the location of the
conductor.
Once the eddy currents are generated, they induce a magnetic field with opposite polarity as the primary field, causing a
repulsive electromagnetic force.
162 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165

2 3
Z Z i j k
6 7
F¼ J
BdV ¼ svy 4 Bz 0 Bx 5dV
Bx By Bz (9)
Z h   i
¼ svy Bx By i  B2x þ B2z j þ By Bz k dV

The electromagnetic force in the motion direction of the EC-TMD is Fy .

Z  
Fy ¼ svy B2x þ B2z dV (10)

According to the properties of the magnetic field distribution, the magnetic field is largely distributed in the projected area
of the permanent magnet, and the magnitudes of the magnetic fields are the same in this area. Therefore, Eq. (10) can be
simplified as follows.
 
Fy ¼ sdS B2x þ B2z vy (11)

where d and S are the thickness and projective area of the permanent magnet, respectively. Bx and Bz are the components of
the magnetic field strength at the position of the conductor and are decided by the relative distance between the conductor
and the permanent magnet.
From Eq. (11), it can be found that the electromagnetic force is proportional to the relative velocity between the permanent
magnet and the conductor; this force is the same as the linear damping force. Hence, the vibration control effect of the EC-
TMD can be considered equivalent to that of the linear damping coefficient, which is a function of the relative position be-
tween the conductor and the permanent magnet.
 
Ce ¼ sdS B2x þ B2z (12)

When an EC-TMD is present in a multi-degree-of-freedom system, the effect of the EC-TMD is equivalent to that of a
system with one spring and two dampers, as shown in Fig. 8, where Ct is the damping provided by the conventional TMD, and
Ce is the damping component due to the eddy currents, which depends on the relative position between the system and the
EC-TMD.
The governing equations for a multi-degree-of-freedom system with an EC-TMD can be written as follows.

MX€ þ C X_ þ KX ¼ Ex€g
M ¼ diag½M1 M2 …Mn Mt 
2 3
C1 þ C2 C2
6 C 7
6 2 C 2 þ C3 C3 7
6 7
C¼66 1 Cn 7
7
6 7
4 Cn Cn þ Ct þ Ce Ct  Ce 5
Ct  Ce Ct þ Ce
2 3
K1 þ K2 K2
6 K K2 þ K3 K3 7
6 2 7
6 7
K¼6
6 1 Kn 7
7
6 7
4 Kn Kn þ Kt Kt 5
Kt Kt
T
X ¼ ½X1 X2 /Xn Xt 
E ¼ ½  M1  M2 /  Mn  Mt T

where M, C, and K are the mass, damping, and stiffness matrices of the system with the EC-TMD, respectively, and Ex€g is the
inertial force matrix with respect to the ground acceleration. Xi is the relative displacement of the ith floor with respect to the
ground, and Mi ; Ci ; and Ki are the mass, damping, and stiffness values of the ith floor, respectively (i ¼ 1,2 …,n). Xt is the
relative displacement of the EC-TMD with respect to the ground; Mt ; Ct ;  and Kt are the mass, damping, and stiffness values of
the conventional TMD; and Ce is the time-varying damping coefficient of the EC-TMD with respect to the eddy currents.
 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165 163

Fig. 8. Simplified diagram of a system with an EC-TMD.

Fig. 9. Calculated and experimental displacementetime histories of the model in the cases of (a) El Centro, (b) Wenchuan waves (0.15 g), (c) zoom-in view of (a),
and (d) zoom-in view of (b).

Table 4
Comparison of the simulation and experimental results for the peak values of the displacement at the roof of the model with the EC-TMD.

Seismic input Peak input value (g) Simulation (mm) Experiment (mm) Error (%)
El Centro 0.05 32.158 30.456 5.3
0.1 69.524 68.882 0.9
0.15 106.682 102.521 3.9
Wenchuan 0.05 7.207 7.048 2.2
0.1 16.109 14.738 8.5
0.15 22.954 22.082 3.8
Japan 311 0.05 3.740 3.449 7.8
0.1 8.972 8.797 1.9
0.15 13.612 14.602 7.3
164 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165

4.2. Computational results

In Section 2, the shaking table tests conducted on the five-story steel-frame structure were discussed. The corresponding
computational simulations are presented in this section. The mass and stiffness matrices of the five-story model are indicated
by M s and K s , respectively.
2 3 2 3
M1 1100
6 7 6 7
6 M2 7 6 1100 7
6 7 6 7
Ms ¼ 6
6 M3 7¼6
7 6 1100 7 kg
7
6 7 6 7
4 M4 5 4 1100 5
M5 1227
2 3
K1 þ K2 K2
6 7
6 K2 K2 þ K3 K3 7
6 7
Ks ¼ 6
6 K3 K3 þ K4 K4 7
7
6 7
4 K4 K4 þ K5 K5 5
K5 K5
2 3
104 52
6 7
6 52 104 52 7
6 7
¼6
6 52 104 52 7
104 N=m
7
6 7
4 52 104 52 5
52 52

The damping ratios of the test structure in the first two structural modes are x1 (¼0.85%) and x2 (¼0.25%), respectively.
For the EC-TMD, the mass (Mt ) is 35 kg. Because the frequencies of the test model and the EC-TMD are equal, the stiffness
between the EC-TMD and the test model can be calculated as Kt ¼ 1381 ðN=mÞ. The damping coefficient with respect to the
eddy currents varies with the change in the relative distance between the conductor and the permanent magnet. Therefore, it
was necessary to calculate the damping coefficient of the EC-TMD for every step using the equations given in Section 4.1.
Fig. 9 shows the calculated and experimental displacementetime histories for the roof of the test model with the EC-TMD
in the cases of the El Centro and Wenchuan waves for a peak ground acceleration of 0.15 g. The calculated displacementetime
histories coincide with the experimental ones in most cases. Table 4 lists the comparison of the simulation results and the
corresponding experimental results for the peak displacement at the roof of the test model with the EC-TMD. Fig. 9 and Table
4 show that the simulation results are in good agreement with the experimental ones, thus demonstrating that the proposed
analytical method can be used to estimate the response of structures with an EC-TMD under earthquake excitations.

5. Conclusion

Compared to conventional TMDs, the EC-TMD with a non-contacting damping mechanism shows excellent performance
in several ways: 1) it does not degrade in its life cycle; 2) the damping performance remains largely unaffected even in high-
temperature environment; 3) the stiffness of the overall structure remains unaffected; 4) the damping ratio can be easily
adjusted by varying the air gap between the permanent magnet and the conductor.
A series of shaking table tests were conducted on a five-story steel-frame model. From the experimental results, the EC-
TMD was proven both feasible and effective in suppressing the vibration of large-scale structures. In terms of the displace-
ment and acceleration responses, the damping effects of the EC-TMD were found to be excellent, particularly in terms of the
RMS responses. Although the frequency characteristics of the El Centro, Wenchuan, and Japan 311 waves are significantly
different, the EC-TMD can be used to reduce the structural vibrations due to the seismic waves, demonstrating the effective
vibration-control performance of the EC-TMD. In addition, the damper helps in reducing the interstory drift ratio and the
maximum strain of the test model.
The relationship between the permanent magnet and the conductor is derived on the basis of the electromagnetic theory,
and the effect of the electromagnetic force is simplified to a time-varying damping coefficient. An analytical method is
proposed to evaluate the response of the structures with the EC-TMD. The analytical/simulation method was validated via the
shaking table test results, and the simulation accuracy was found to be reasonable.
The results of the experimental and analytical investigations demonstrate the effectiveness and excellent performance of
the EC-TMD in suppressing the vibration of a large-scale model. Therefore, this technology has a good potential for future
applications to engineering projects.
 Z. Lu et al. / Journal of Sound and Vibration 421 (2018) 153e165 165

Acknowledgement

Financial support from the National Natural Science Foundation of China (Grant No. 51478361), Science and Technology
Commission of Shanghai Municipality (STCSM, Grant No. 13231201600) are highly appreciated. Shanghai Research Institute of
Materials are acknowledged as well.

References

[1] J.T.P. Yao, Concept of structural control, J. Struct. Div. ASCE 98 (1972) 1567e1574.
[2] F. Casciati, L. Faravelli, A passive control device with SMA components: from the prototype to the model, Struct. Contr. Health Monit. 16 (2009)
751e765.
[3] A.J. Roffel, S. Narasimhan, Results from a full-scale study on the condition assessment of pendulum tuned mass dampers, J. Struct. Eng. 142 (2016),
https://doi.org/10.1061/(asce)st.1943-541x.0001339.
[4] Z. Lu, X. Chen, D. Zhang, K. Dai, Experimental and analytical study on the performance of particle tuned mass dampers under seismic excitation,
Earthq. Eng. Struct. Dynam. 46 (2017) 697e714.
[5] Z. Lu, X. Lu, W. Lu, S.F. Masri, Shaking table test of the effects of multi-unit particle dampers attached to an MDOF system under earthquake excitation,
Earthq. Eng. Struct. Dynam. 41 (2012) 987e1000.
[6] Z. Lu, X. Lu, H. Jiang, S.F. Masri, Discrete element method simulation and experimental validation of particle damper system, Eng. Comput. 31 (2014)
810e823.
[7] Z. Lu, X. Lu, W. Lu, S.F. Masri, Experimental studies of the effects of buffered particle dampers attached to a multi-degree-of-freedom system under
dynamic loads, J. Sound Vib. 331 (2012) 2007e2022.
[8] Z. Lu, X.L. Lu, W.S. Lu, S.F. Masri, Shaking table test of the effects of multi-unit particle dampers attached to an MDOF system under earthquake
excitation, Earthq. Eng. Struct. Dynam. 41 (2012) 987e1000.
[9] Z. Lu, X.L. Lu, S.F. Masri, Studies of the performance of particle dampers under dynamic loads, J. Sound Vib. 329 (2010) 5415e5433.
[10] Z. Lu, S.F. Masri, X. Lu, Parametric studies of the performance of particle dampers under harmonic excitation, Struct. Contr. Health Monit. 18 (2011)
79e98.
[11] Z. Lu, D.C. Wang, S.F. Masri, X.L. Lu, An experimental study of vibration control of wind-excited high-rise buildings using particle tuned mass dampers,
Smart Struct. Syst. 18 (2016) 93e115.
[12] Z. Lu, D.C. Wang, Y. Zhou, Experimental parametric study on wind-induced vibration control of particle tuned mass damper on a benchmark high-rise
building, Struct. Des. Tall Special Build. 26 (2017), https://doi.org/10.1002/tal.1359.
[13] Z. Lu, Z. Wang, S.F. Masri, X. Lu, Particle impact dampers: past, present, and future, Struct. Contr. Health Monit. (2017), https://doi.org/10.1002/stc.2058.
[14] J.S. Bae, M.K. Kwak, D.J. Inman, Vibration suppression of a cantilever beam using eddy current damper, J. Sound Vib. 284 (2005) 805e824.
[15] M.K. Kwak, M.I. Lee, S. Heo, Vibration suppression using eddy current damper, Transac. Kor. Soc. Noise Vib. Eng. 13 (2003) 760e766.
[16] H.A. Sodano, B. Jae-Sung, Eddy current damping in structures, Shock Vib. Digest 36 (2004) 469e478.
[17] M. Mehrtash, M.B. Khamesee, Modeling and analysis of eddy-current damping effect in horizontal motions for a high-precision magnetic navigation
platform, IEEE Trans. Magn. 49 (2013) 4801e4810.
[18] Q. Pan, T. He, D.H. Xiao, X.D. Liu, Design and damping analysis of a new eddy current damper for aerospace applications, Lat. Am. J. Solid. Struct. 13
(2016) 1997e2011.
[19] X.L. Lu, Q. Zhang, D.G. Weng, Z.G. Zhou, S.S. Wang, S.A. Mahin, S.W. Ding, F. Qian, Improving performance of a super tall building using a new eddy-
current tuned mass damper, Struct. Contr. Health Monit. 24 (2017), https://doi.org/10.1002/stc.1882.
[20] J.G. Detoni, Q. Cui, N. Amati, A. Tonoli, Modeling and evaluation of damping coefficient of eddy current dampers in rotordynamic applications, J. Sound
Vib. 373 (2016) 52e65.
[21] Y. Matsuzaki, D. Ishikubo, T. Kamita, T. Ikeda, Vibration control system using electromagnetic forces, J. Intell. Mater. Syst. Struct. 8 (1997) 751e756.
[22] Y. Matsuzaki, T. Ikeda, A. Nae, T. Sasaki, Electromagnetic forces for a new vibration control system: experimental verification, Smart Mater. Struct. 9
(2000) 127e131.
[23] X.J. Zheng, Y.H. Zhou, K. Miya, An analysis of variable magnetic damping of a cantilever beam-plate with end coils in transverse magnetic fields, Fusion
Eng. Des. 55 (2001) 457e465.
[24] H.A. Sodano, J.-S. Bae, D.J. Inman, W.K. Belvin, Improved concept and model of eddy current damper, J. Vib. .Acoust. Transac. ASME 128 (2006)
294e302.
[25] D.A. Kienholz, S.C. Pendleton, K.E. Richards Jr., D.R. Morgenthaler, Demonstration of solar array vibration suppression, in: 1994 North American
Conference on Smart Structures and Materials, International Society for Optics and Photonics, 1994, pp. 59e72.
[26] G.L. Larose, A. Larsen, E. Svensson, Modeling of tuned mass dampers for wind-tunnel tests on a full-bridge aeroelastic model, J. Wind Eng. Ind. Aerod.
54 (1995) 427e437.
[27] J.-S. Bae, J.-H. Hwang, J.-H. Roh, J.-H. Kim, M.-S. Yi, J.H. Lim, Vibration suppression of a cantilever beam using magnetically tuned-mass-damper, J.
Sound Vib. 331 (2012) 5669e5684.
[28] Z. Wang, Z. Chen, J. Wang, Feasibility study of a large-scale tuned mass damper with eddy current damping mechanism, Earthq. Eng. Eng. Vib. 11
(2012) 391e401.
[29] J.-S. Bae, J.-H. Hwang, D.-G. Kwag, J. Park, D.J. Inman, Vibration suppression of a large beam structure using tuned mass damper and eddy current
damping, Shock Vib. (2014), https://doi.org/10.1155/2014/893914.
[30] X. Lu, J. Chen, Parameter optimization and structural design of tuned mass damper for Shanghai centre tower, Struct. Des. Tall Special Build. 20 (2011)
453e471.
[31] Y. Tamura, K. Fujii, T. Ohtsuki, T. Wakahara, R. Kohsaka, Effectiveness of tuned liquid dampers under wind excitation, Eng. Struct. 17 (1995) 609e621.
[32] X.F. Guo, Y. Yang, X.J. Zheng, Analytic expression of magnetic field distribution of rectangular permanent magnets, Appl. Math. Mech. 25 (2004)
297e306. English Edition.