Engineering Structures 271 (2022) 114873

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Vibration mitigation of long-span bridges with damped outriggers

Lin Chen a , Zhanhang Liu a , Satish Nagarajaiah d,e , Limin Sun a,b,c ,∗, Lin Zhao a,b , Wei Cui a
a
Department of Bridge Engineering, Tongji University, Shanghai 200092, PR China
b
State Key Laboratory for Disaster Reduction of Civil Engineering, Tongji University, Shanghai 200092, PR China
c
Shanghai Qi Zhi Institute, Shanghai 200092, PR China
d
Department of Civil and Environmental Engineering, Rice University, 6100 Main Street, Houston, TX 77005, USA
e Department of Mechanical Engineering, Rice University, 6100 Main Street, Houston, TX 77005, USA

ARTICLE INFO ABSTRACT

Keywords: Long-span bridges have low vibration frequencies and small intrinsic damping, and hence are subjected to
Long-span bridges large-amplitude vibrations, e.g., vortex-induced vibrations (VIVs), which jeopardize the serviceability and
Damped outrigger safety of the bridge structures. This study proposes a novel concept to supplement damping to such bridges
Damping
for multimode vibration mitigation, by damping rotations of the bridge girder during vertical and torsional
Vortex-induced vibrations
vibrations instead of translational displacements that have been previously the focus of relevant studies. A
Complex modal analysis
salient advantage of the concept is that bridge girder rotations near its ends are relatively large during
vibrations enabling adequate supplemental damping and also convenient implementation. The concept is
realized by a damped outrigger, consisting of a stiff outrigger rigidly connected to the bridge girder and
horizontal dampers connecting the end of the outrigger to a bridge tower or a pier. A tensioned beam with
rotational constraints is first used to demonstrate damping effects of damped outriggers on suspension bridges.
Considering practical parameter ranges, the maximal damping ratio provided by one damped outrigger to
a specific mode is about 1.0%. Influences of the relative importance of axial tension and bending stiffness
and outrigger stiffness are studied. Furthermore, finite element analyses of an existing suspension bridge with
damped outriggers are conducted to verify the damping performance. Numerical results show that by installing
one damped outrigger, 6 out of 7 modes that have been involved in VIV events of the bridge can achieve a
damping ratio of approximately 1.0% or above. By installing damped outriggers at the two towers, more
modes can achieve a relatively large damping ratio, suggesting excellent performance of damped outriggers
in suppressing multimode vibrations of long-span bridges.

1. Introduction Yingwuzhou Bridge [6] in China, and the Jindo bridge in Korea [7], to
name but a few [1]. Therefore, suppression of such bridge vibrations
Long-span bridges have low vibration frequencies and small struc- has been a topical area during the last decades [8,9].
tural damping, and hence are subjected to large-amplitude vibrations Aerodynamic countermeasures are among the most widely used
that jeopardize the serviceability and safety of the bridge structures.
methods to prevent harmful vibrations and reduce vibration amplitude
Typical mechanisms of wind-induced bridge vibrations include aeroe-
of bridges [1,8,9]. By changing the shape and configuration of the
lastic instability featured by significant interactions between structural
motion and aerodynamic force (e.g., galloping and flutter), vortex- bridge cross-section, particularly near section corners, aerodynamic
induced vibrations (VIVs), and buffeting related to wind pressure fluc- countermeasures reduce wind velocity of separated flow and eliminate
tuations on the bridge [1]. Among those vibration mechanisms, VIVs of self-excitation forces on the bridge components [1]. They are effective
observable amplitude can occur under low wind speed particularly for in increasing flutter wind speed and mitigating VIVs, and typical exam-
long-span bridges whose vibration frequencies and inherent damping ples include the use of slotted box girder and sectional modifications
are extremely low. For example, in May 2020, the Humen Bridge in by adding wind nose, fairing, guide vanes, flaps and so on [1,10].
China experienced VIVs, resulting in the bridge being closed to traffic Alternative aerodynamic countermeasures have also been developed
for ten days [2,3]. The VIVs can happen in different bridge modes and verified, such as inclined railings [11], sealed traffic barriers [12],
under varied wind conditions [4]. Indeed, they have been observed
passive and active flow control [9,13]. Owing to the complexity of
on a number of bridges, including the Xihoumen Bridge [5] and the

∗ Corresponding author at: Department of Bridge Engineering, Tongji University, Shanghai 200092, PR China.
E-mail addresses: [email protected] (L. Chen), [email protected] (S. Nagarajaiah), [email protected] (L. Sun).

https://doi.org/10.1016/j.engstruct.2022.114873
Received 7 January 2022; Received in revised form 24 July 2022; Accepted 19 August 2022
Available online 15 September 2022
0141-0296/© 2022 Elsevier Ltd. All rights reserved.
L. Chen et al. Engineering Structures 271 (2022) 114873

fluid–structure interaction, the performance of aerodynamic counter- stiffness leads to damping degradation [32]. Multiple damped out-
measures in bridge vibration mitigation is often assessed by wind tunnel riggers at different locations along the building have been further
tests on scaled bridge models or high-fidelity simulations via compu- studied [33]. Optimization design of damped outrigger parameters and
tational fluid dynamics [14]. The performance is difficult to quantify locations has been pursued for tall buildings under seismic, wind, and
and it could be sensitive to the adopted aerodynamic parameters [15] other stochastic loads [34–36]. Recent studies have indicated that the
and the assumed structural damping in tests and simulations [3]. performance of damped outriggers can be enhanced by incorporat-
Besides, the cost of adding aerodynamic countermeasures can be high, ing inerters [37–39], negative stiffness devices [40–42], or combining
particularly when the bridge is already in operation [1]. with buckling braces [43]. More importantly, damped outriggers have
Structural damping is a key factor influencing the onset and am- already been implemented in the design of several iconic buildings,
plitude of bridge vibrations and it is small particularly for long-span including the Stantec tower in Canada [44] and the St. Shangari Palace
bridges, often less than 0.5% according to field measurements [3]. in Philippines [29].
Hence, another type of countermeasures to mitigate bridge vibrations For short-span bridges, there exist studies on using arms extended
is adding dampers or absorbers to dissipate vibration energy. During from bridge girders to apply moments for structural response reduc-
bridge vibrations, the girder has relative large displacements near the tion [45–48]. To the best knowledge of the authors, this is the first time
mid-span where the energy dissipation devices are expected to be that the damped outrigger is introduced between the bridge girder and
installed to perform adequately. The installation of typical dampers, a bridge tower or a pier in long-span bridges and its damping effect is
as two-terminal devices, requires a support which is not available investigated in detail. The aim of this study is three-folds:
near the bridge mid-span. In contrast, the tuned mass damper (TMD),
as a one-terminal absorber, is thus more frequently used for bridge i. to present the concept of damped outrigger for bridges, and
girder vibration mitigation. It has been implemented on a number of discuss its advantages and features;
bridges [1,16]. Note that each bridge VIV event is often dominated by ii. to investigate the damping effect and influential factors of the
a single-mode vibration [4]. However, during long-term operation, the proposed damping system using a simplified model of a suspen-
bridge can vibrate in different modes under varied wind conditions, sion bridge with damped outriggers; and
as indicated from field measurements of the Xihoumen Bridge [17–20] iii. to verify the damping effects using a full finite element model of
and the Humen Bridge [2,4], while the vibration mitigation effect of a a suspension bridge with damped outriggers.
TMD is sensitive to frequency tuning. Hence, multiple TMDs distributed
at different locations along the bridge span are desired targeting multi-
The remainder of this paper is structured as follows. Section 2
ple vulnerable structural modes [21,22]. As the mass ratio of the TMD
presents the concept of damped outriggers for bridge vibration mit-
relative to the bridge modal mass increases, the device performs better
igation. Subsequently, a simplified method is used in Section 3 to
in vibration mitigation, while the static load on the bridge increases
demonstrate damping effects of outriggers on bridges and the influen-
and more installation space is required. It is thus of practical interest to
tial factors. In Section 4, a finite element model of an existing large-
increase the mass efficiency. The inerter, which can amplify the inertia
force by rotating mass and gear mechanisms, has been combined with span suspension bridge with damped outriggers is used to quantify the
a TMD [23], leading to the so-called tuned mass damper inerter, for damping effects. Concluding remarks are provided in Section 5.
mitigating bridge VIVs, and its superior performances as compared to
traditional TMDs have been demonstrated by theoretical and numerical 2. The concept of damped outriggers for long-span bridges
analyses [24]. Apart from TMDs, other damping systems have also been
proposed in the literature, e.g., vertical dampers between bridge girder Long-span bridges when vibrate often have relatively large displace-
and towers for suspension bridges whose girder is continuous and ment amplitudes in the main span. Interestingly, by comparing the
not vertically constrained at the bridge tower [25,26], and connecting deformed shapes of the girder at two time instances, during the VIV
bridge main cables with towers using viscous dampers [27]. event of the Humen Bridge in May 2020 [2], it is clear that the bridge
Vibration mitigation of long-span bridges is still a challenging issue. girder rotates considerably at the position close to the bridge tower, as
Effective and implementable countermeasures are strongly desired. shown in Fig. 1. Large rotation amplitudes of the bridge girder near
Therefore, this study proposes a novel concept for suppressing bridge a bridge tower or a bridge pier can be seen from mode shapes of
vibrations, by damping rotations of the bridge girder in vertical bending many long-span bridges, including girder bridges [16] and suspension
or torsional modes instead of translational displacements that have bridges [4,10,19]. Based on such observations, this study proposes to
been previously the focus of relevant studies. A salient feature of the supply damping to bridge girder based on the girder rotation near a
concept is that bridge girder experiences relatively large rotations near bridge tower or pier, as demonstrated in Fig. 2. A stiff outrigger is
its ends, as illustrated in Fig. 1, allowing effective energy dissipation rigidly mounted on the girder. Dampers are installed horizontally to
by devices installed between the girder and a tower or a pier. This connect the end of the outrigger to a bridge tower or pier. During
makes the strategy practical for both new and existing bridges. The vibrations of the bridge girder, the outrigger transforms girder rotations
concept is realized by a damped outrigger system, which consists of a to translational displacements at the damper installation positions and
stiff outrigger rigidly connected to the bridge girder to transform girder the dampers are thus forced to deform to dissipate energy. The damped
rotations to horizontal displacements at the tip of the outrigger and outrigger system together behaves as a rotational damper to the bridge
horizontal dampers connecting the outrigger to a bridge tower or a pier girder. Meantime, the outrigger also amplifies the damping force as
for energy dissipation. such the damper size can be reduced.
It is acknowledged that the concept of damped outrigger is orig- The features of the proposed damping system are discussed in
inated in structural response reduction of high-rise buildings, first comparison with the previous suppression techniques for long-span
proposed in Refs. [28–30]. Stiff outriggers wherein are rigidly con- bridges, listed as follows.
nected to the core tube of a building and viscous dampers are installed
vertically connecting the tip of the outriggers to perimeter columns. It • In the damped outrigger, the dampers are installed in the hor-
is theoretically shown that significant damping can be supplemented izontal direction along the longitudinal axis of the bridge. The
to the first several structural modes and the damping is dependent of system is compatible with the longitudinal movements of the
the outrigger location [31]. The axial stiffness of perimeter columns bridge girder under vehicles, earthquake, thermal effects and so
critically influences the achievable damping and a decrease in the on.

2
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 1. Deck inclination change observed during VIVs of the Humen Bridge.

Fig. 2. The concept of vibration mitigation of long-span bridges with damped outriggers.

• The traditional longitudinal dampers installed between the bridge 3. Damping effect analyses using a simplified method
girder and bridge towers for reducing seismic responses of long-
span bridges may no longer be required. In other words, the To demonstrate damping effects of the proposed damped outriggers
damped outrigger may be designed to reduce both wind-induced in Fig. 2 on long-span bridges, suspension bridges which are partic-
vibrations and seismic responses. ularly vulnerable to VIVs are considered. A tensioned beam model is
• As the damped outrigger is installed near the tower, its weight is first used along with a simplified damping analysis method.
efficiently transmitted to the bridge tower through bridge deck,
suspenders and the main cables. Hence, no large vertical deforma- 3.1. A tensioned beam model
tion is caused in contrast to a TMD installed near the mid-span.
Besides, bearings can be added between the girder and the tower Fig. 3 shows a suspension bridge with a main span and a side span is
to reduce the vertical deformation and the tension of suspenders considered. Assuming rigid suspenders [3,49], the cables and the girder
near the damped outrigger and would not significantly decrease are the modeled as a tensioned beam considering the tension (𝐻) in
the supplement damping. the main cables and the bending stiffness of the girder (𝐸𝐼). Mass per
• One damped outrigger can supplement damping to multiple unit length of the beam is denoted by 𝑚. The length of the tensioned
modes including both vertical bending modes and torsional modes beam (𝐿) is the total length of the bridge main span and side span. The
damped outrigger is installed at a distance of 𝑙𝑜 from the left end of the
of bridge girders, as contrast to a TMD which is normally designed
beam. The beam is simply supported at its two ends.
aiming at a particular mode.
The lateral vibration of a tensioned beam with rotational constraints
• The damped outriggers can be installed between each tower and
are described by the following governing equation [50],
the bridge girder and also between the piers and the bridge girder,
to supplement larger damping to more structural modes. 𝜕4 𝑣 𝜕2 𝑣 𝜕2 𝑣 ∑ d
𝐸𝐼 −𝐻 +𝑚 + 𝑀𝑜𝑖 𝛿(𝑥 − 𝑙𝑜𝑖 ) = 0, (1)
• As the damped outrigger is installed near the end of the bridge 𝜕𝑥4 𝜕𝑥2 𝜕𝑡2 𝑖
d𝑥
girder, the installation has minimal effects on the traffic. It can where 𝑣(𝑥, 𝑡) denotes the lateral displacement of the beam with 𝑡 =
be inspected and maintained with the bridge open to traffic. time and 𝑥 = axial coordinate originated from the left beam end. In the
• However, a stiff and relatively long outrigger is required to add to preceding equation, 𝛿( ) is the Dirac delta function, 𝑀𝑜𝑖 is the moment
the bridge girder, and the bridge girder needs to be strengthened provided by the 𝑖th damped outrigger, and 𝑙𝑜𝑖 denotes the distance
near the installation point to effectively transmit rotation. The between the left beam end and the 𝑖th damped outrigger. The preceding
stokes of the dampers are also expected to be large considering equation can be rewritten in the following dimensionless form
both vertical vibrations and longitudinal motions of the bridge 𝜕4 𝑣 𝜕2 𝑣 𝜕2 𝑣 ∑ d
− 𝛾2 + 𝜋2𝛾 2 + 𝑀 𝑜𝑖 𝛿(𝑥 − 𝑙𝑜𝑖 ) = 0. (2)
girder. 𝜕𝑥4 𝜕𝑥2 2 d𝑥
𝜕𝑡 𝑖

3
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 3. A tensioned beam model to simulate a suspension bridge with damped outriggers.

The dimensionless variables are defined as 53]. Herein, an approximate method based on frequency shift [54] is
𝑣 𝑥 𝑙 𝐿 used. The tensioned beam is first discretized using the finite differ-
𝑣 = , 𝑥 = , 𝑡 = 𝜔01 𝑡, 𝑙𝑜𝑖 = 𝑜𝑖 , 𝛿(𝑥 − 𝑙𝑜𝑖 ) = 𝐿𝛿(𝑥 − 𝑙𝑜𝑖 ), 𝑀 𝑜𝑖 = 𝑀 ,
𝐿 𝐿 𝐿 𝐸𝐼 𝑜𝑖 ence method [55], leading to a 𝑁-degree-of-freedom system, and the
√
𝜋 𝐻 governing equation is given as
𝜔01 = . (3)
𝐿 𝑚
𝐌𝐪̈ + 𝐂𝐪̇ + 𝐊𝐪 = 𝟎, (8)
Note that 𝛾 represents relative importance of axial stiffness to bending
stiffness of a tensioned beam [51] as where 𝐌, 𝐂 and 𝐊 are the 𝑁 ×𝑁 mass, damping, and stiffness matrices,
√
𝐻𝐿2 𝐪 is the 𝑁 × 1 vector of nodal displacements, and 𝟎 is the vector of
𝛾= . (4)
𝐸𝐼 zeros. They are defined in Appendix. Once the coefficient matrices are
For a beam without tension, 𝛾 → 0 and for a cable without bending determined, eigenanalysis can be performed to obtain system frequen-
stiffness 𝛾 → ∞. The characteristic frequency 𝜔01 is the circular fre- cies and mode shapes. In the subsequent analyses, the tensioned beam
quency of the first mode of a cable of length 𝐿, tension 𝐻, distributed is divided into 500 elements if not otherwise specified.
mass 𝑚, and without any attachments. Correspondingly, frequency of In the case of small frequency shift induced by the attached
the beam, denoted by 𝜔, is normalized as dampers, complex frequency of the system can be approximated [54]
√ √
𝜔 𝜔𝐿 𝑚 𝜔 𝑚𝐿4 as
𝜔̂ = = ̃ with 𝜔̃ =
= 𝜋𝛾 𝜔, , (5)
𝜔01 𝜋 𝐻 𝜋2 𝐸𝐼 i𝜂𝑛
𝜔̂ 𝑛 = 𝜔̂ 0𝑛 + 𝛥𝜔̂ 𝑛 ≈ 𝜔̂ 0𝑛 + 𝛥𝜔̂ ∞𝑛 (9)
where 𝜔̃ is the frequency when normalized by the first circular fre- 1 + i𝜂𝑛
quency of a Bernoulli beam with bending stiffness 𝐸𝐼, total length 𝐿, √
and mass per unit length 𝑚 in the case of pinned–pinned supports. where i = −1, 𝜔̂ 0𝑛 denotes the 𝑛th mode frequency when damper
In Fig. 3, when the outrigger is rigid, the damping moment is given coefficient 𝑐𝑜𝑖 = 0, 𝛥𝜔̂ 𝑛 is the frequency shift induced by dampers,
as and in this limit 𝑐𝑜𝑖 = ∞ the frequency shift is denoted by 𝛥𝜔̂ ∞𝑛 . The
parameter 𝜂𝑛 associated with damper coefficient is defined [54] as
d𝜃𝑜𝑖 𝑟2 𝑐𝑜𝑖 𝜔01 𝐿 d𝜃𝑜𝑖 d𝜃 𝑐̃𝑜𝑖 𝑟2
𝑀𝑜𝑖 = 𝑟2𝑜𝑖 𝑐𝑜𝑖 , 𝑀 𝑜𝑖 = 𝑜𝑖 = 𝜋𝛾 𝑐̃𝑜𝑖 𝑜𝑖 , with 𝑐̃𝑜𝑖 = √ 𝑜𝑖 , ∑ 2
d𝑡 𝐸𝐼 d𝑡 d𝑡 𝐿 𝐸𝐼𝑚 𝑖 𝑐𝑜𝑖 𝜌𝑛𝑖
𝜂𝑛 = , (10)
2𝛥𝜔∞𝑛
(6)
where 𝜌𝑛𝑖 denotes the differential displacement of the 𝑛th undamped
where 𝑟𝑜𝑖 is the vertical length of the outrigger, 𝑐𝑜𝑖 is the damper
mode shape at the 𝑖th damped outrigger, see Eq. (A.14) in Appendix.
coefficient, 𝜃𝑜𝑖 = 𝜕𝑣∕𝜕𝑥|𝑥=𝑙𝑜𝑖 is the beam rotation at the damped
Correspondingly, damping ratio 𝜁𝑛 can be computed as
outrigger, and 𝑐̃𝑜𝑖 denotes the dimensionless damper coefficient.
For the case of a flexible outrigger with an end lateral stiffness of ℑ(𝜔̂ 𝑛 ) ℑ(𝜔̃ 𝑛 )
𝑘𝑜𝑖 and when 𝑐𝑜𝑖 = ∞, the moment provided by the outrigger is given 𝜁𝑛 = = , (11)
|𝜔̂ 𝑛 | |𝜔̃ 𝑛 |
as
where ℑ( ) and | | denote the imaginary part and modulus of a complex-
𝑟2𝑜𝑖 𝑘𝑜𝑖 𝐿 𝑟2𝑜𝑖 𝑘𝑜𝑖 𝐿
𝑀𝑜𝑖 = 𝑟2𝑜𝑖 𝑘𝑜𝑖 𝜃𝑜𝑖 , 𝑀 𝑜𝑖 = 𝜃𝑜𝑖 = 𝑘̃ 𝑜𝑖 𝜃𝑜𝑖 , with 𝑘̃ 𝑜𝑖 = , (7) valued number.
𝐸𝐼 𝐸𝐼
Using the approximate formula Eq. (9), the loci of the complex
where 𝑘̃ 𝑜𝑖 is the dimensionless stiffness of the outrigger.
frequency for damper coefficient increasing from 0 to infinitely large
3.2. An approximate method for complex frequency analyses forms a semi-circle on the complex plane [54], as shown in Fig. 4(a).
Correspondingly, the damping curve is shown in Fig. 4(b), where the
Dynamic analyses of tensioned beams with attachments, as shown in maximal damping ratio is approximately 𝜁𝑛,max = 0.5𝛥𝜔̂ ∞𝑛 ∕𝜔̂ 0𝑛 with
opt
Fig. 3, can be performed using the dynamic stiffness method [51]. Par- optimal damper parameter 𝜂𝑛 = 1, and the optimal damper coefficient
opt
ticularly, beams with rotational dampers have been studied in Refs. [52, is denoted by 𝑐̃𝑜 .

4
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 4. Approximate frequency evolution and damping variation of a tension beam with a damped outrigger.

Table 1
Structural parameters of typical suspension bridges.
No. Suspension bridge Main Side Tension 𝐸𝐼 Length 𝛾
span (m) span (m) 𝐻 (kN) (kN m2 ) 𝐿 (m)
1 The Xihoumen Bridge [17] 1650 578 5.31 × 105 4.43 × 108 2228 77.14
2 The Runyang Yangtze River Bridge [49] 1490 – 2.74 × 105 4.36 × 108 1490 37.35
3 The Great Belt Bridge [56,57] 1624 535 × 2 3.88 × 105 6.97 × 108 2694 63.56
4 The Jiangyin Yangtze River Bridge [58] 1385 – 9.60 × 105 3.62 × 108 1385 71.32
5 The Fatih Sultan Mehamet Bridge [59,60] 1090 – 3.62 × 105 3.55 × 108 1090 34.81
6 The Yangsigang Yangtze River Bridge [61,62] 1700 – 9.95 × 105 5.15 × 108 1700 74.72

Fig. 5. Mode shapes of the first eight modes of a tensioned beam with pinned–pinned supports and 𝛾 = 50. Note that the displacement and rotation are normalized separately.

3.3. Damping effect the index 𝑖 is thus omitted. A rigid outrigger is assumed, i.e., 𝑘̃ 𝑜 = ∞.
The two real frequencies are solved using eigenanalysis of the discrete
In long-term operation and under different wind conditions, a model as in Eq. (8) considering pinned–pinned end condition and fixed–
long-span bridge can experience vibrations in the first several modes, pinned (𝑐̃𝑜 = ∞) end condition respectively. Subsequently, complex
e.g., VIVs of the Humen Bridge [2] and the Xihoumen Bridge [18]. In frequencies of the first eight modes are evaluated using Eq. (9) for
this section, the first eight modes of a tensioned beam are considered 𝜂𝑛 varied from 0 to a very large value. The corresponding damper
in damping analyses. coefficient 𝑐̃𝑜𝑖 is computed using Eq. (10).
For parametric study, Table 1 lists the values of girder bending Fig. 6(a) shows the frequency loci of the first eight modes and
stiffness (𝐸𝐼), total horizontal tension (𝐻) of main cables, and total Fig. 6(b) plots the damping curves. The frequency evolution is similar
length (𝐿) including both the main span and side spans of typical long- to that of a cable with a lateral damper near its one end [63]. The
span suspension bridges. Simplifying the bridges as a tensioned beam complex frequencies solved using the analytical method [63] are also
with the bending stiffness 𝐸𝐼, axial force 𝐻 and total length 𝐿, the presented in the figure for comparison. Apparently, the frequency
values of 𝛾 are computed using Eq. (4) and also listed in the table. It variations and damping curves for all the modes are similar, and can
is seen that 𝛾 is in the range of 30 to 80. Therefore, in the following be well approximated by using the two real frequencies with 𝑐̃𝑜 = 0 and
analyses, 𝛾 = 50 is adopted in most of the cases and the influence of 𝑐̃𝑜 = ∞ [64,65].
variation in 𝛾 will be discussed. Fig. 6 shows that when the damped outrigger is installed at an end
Fig. 5 shows the first eight mode shapes of a tensioned beam with of the tensioned beam and when 𝛾 = 50, the maximal supplemental
pinned–pinned supports and 𝛾 = 50. In the figure, the rotation corre- damping for a mode is approximately 1.0%. Such amount of damping is
sponding to the mode shape along the beam is also shown. It is seen large as compared to the inherent damping of long-span bridges [3]. For
that the rotations near beam supports have the maximal amplitudes and a higher mode, the maximal damping is slightly increased, while the
opt
this applies to each mode. These features make it efficient to dissipate optimal damper coefficient 𝑐̃𝑜 is decreased with respect to increasing
multimode vibration energy of the beam using rotational dampers at mode number. In this case, when the design target is set to increase
the ends. the lower bound of the damping ratios of the first eight modes, the
opt
optimal damper coefficient is around 𝑐̃𝑜 = 0.12 and all the eight
3.3.1. One damped outrigger modes have supplemental damping larger than 0.6%, as indicated by
A single damped outrigger is installed at the left end of the tensioned the intersection point (nonzero damping) of the damping curves of the
beam, i.e., 𝑙𝑜 = 0. As one damped outrigger is considered in this section, first and eighth modes in Fig. 6(b).

5
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 6. Complex frequency and damping of a tensioned beam (𝛾 = 50) with a damped and rigid outrigger at 𝑙̄𝑜 = 0: (a) frequency loci with respect to increasing damper coefficient
𝑐̃𝑜 , where the circles denote the origins (𝑐̃𝑜 = 0) and the crosses denote the terminals (𝑐̃𝑜 = ∞) of the frequency loci; (b) damping curves. Thinner lines are exact solutions obtained
using analytical method in the literature.

Fig. 7. Influences of 𝛾 on the maximal attainable damping and the corresponding optimal damper coefficient when 𝑙̄𝑜 = 0 and the outrigger is rigid: (a) maximal damping versus
𝛾; (b) optimal damper coefficient versus 𝛾.

3.3.2. Influence of 𝛾 maximal damping is reduced to 60% of the maximal damping. Fig. 8(b)
Consider that a damped outrigger installed at 𝑙𝑜 = 0, and the shows that as the outrigger stiffness decreases, the optimal damper
parameter 𝛾 varied between 10 and 100 according to Table 1. For each coefficients are decreased. Those variations of damping and damper
𝛾, the maximal damping and corresponding optimal damper coefficients coefficients along with varying outrigger stiffness are similar for each
are computed using Eqs. (9) and (10). They are plotted versus 𝛾 in mode when the damped outrigger is close to a beam end. Assume that
Fig. 7. Fig. 7(a) suggests that as the value of 𝛾 is increased, the maximal the outrigger is a cantilever with uniform bending stiffness 𝐸𝐼𝑜 , and
attainable damping that can be achieved by a rotational damper is hence 𝑘𝑜 = 3𝐸𝐼𝑜 ∕𝑟3𝑜 . According to the definitions in Eq. (7), one finds
decreased monotonically, as the tensioned beam behaves more like a 𝐸𝐼𝑜 = 𝑘̃ 𝑜 𝐸𝐼𝑟𝑜 ∕(3𝐿). When 𝑘̃ 𝑜 = 1000 and for a bridge span 𝐿 = 1000 and
cable when 𝛾 is large. It is also seen that when 𝛾 ≤ 30 the maximal outrigger length 𝑟𝑜 = 10, 𝐸𝐼𝑜 = 10𝐸𝐼∕3, which is feasible in practice.
damping decreases when the mode order is increased while the maxi-
mal damping is almost the same for each mode when 𝛾 > 30. Fig. 7(b) 3.3.4. Discussions
shows that the optimal damper coefficient for each mode is almost In this section, a tensioned beam with one damped outrigger at one
unchanged as the value of 𝛾 increases. end is analyzed using an approximate method considering parameter
ranges of realistic suspension bridges. The damping effect is found
3.3.3. Influence of outrigger stiffness promising. The damping effect could be further improved by adding
According to the studies on damped outriggers for tall buildings and damped outriggers at both ends, and at the girder–pylon connection
dampers for stay cables, the stiffness of damper support is critical to in case that the bridge girder is continuous at the pylon. According
the damping effects [32,66]. Therefore, the influence of the support to previous studies on beam-like structures with near-end dampers,
stiffness of damped outriggers for bridges is examined in this section. when two dampers are installed respectively close to the two ends,
In this case, the undamped frequencies are computed using the same the damping effects can be simply superposed [54,65,67]. Note that
method as previously discussed. In the limit case when the damper when the damper is installed at an intermediate location, the frequency
coefficient is infinitely large, the stiffness matrix is modified according variations can be quite large [51] and hence some modes can even
to stiffness and location of the outrigger, as given in Eq. (A.16). attain a quite large damping ratio. Besides, as the damped outriggers
Consider that the damped outrigger is installed at 𝑙𝑜 = 0 and 𝛾 = 50. are installed near the beam ends, the boundary conditions can largely
Maximal damping ratios and optimal damper coefficients are obtained affect the damping performance [67].
using the approximate method for outrigger stiffness 𝑘̃ 𝑜 increasing from
10 to 105 . They are plotted with respect to 𝑘̃ 𝑜 in Fig. 8. It is seen 4. Numerical studies
that when 𝑘̃ 𝑜 is larger than 5 × 103 the maximal damping is stabilized
and thus the support can be considered as rigid. When the outrigger 4.1. Bridge description
stiffness 𝑘̃ 𝑜 is reduced to 200 the maximal damping is decreased to 80%
of the maximal damping corresponding to the case when the support To further verify the damping effects, the Xihoumen Bridge, a sea
is rigid. Furthermore, when the outrigger stiffness is around 100, the crossing bridge located in Zhoushan City, Zhejiang Province, China, is

6
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 8. Influences of outrigger stiffness (𝑘̃ 𝑜 ) on the maximal attainable damping and the corresponding optimal damper coefficient when 𝛾 = 50 and 𝑙̄𝑜 = 0: (a) maximal damping
versus outrigger stiffness; (b) optimal damper coefficient versus outrigger stiffness.

modeled for damping analyses in this section. It is a suspension bridge Table 2

with a slotted twin-box girder. The main span is of 1650 m and the side Computed frequencies of vertical bending modes of the bridge.

span of 578 m (north-side). The details of the bridge can be found in Mode Field measurement/monitoring Computed

a number of previous studies, e.g. Ref. [17], as also shown in Fig. 9. no. Frequency (Hz) Damping ratio (%) VIVs observed [19] frequency (Hz)
The bridge experiences VIVs under wind loading, according to Ref. [19] 1* 0.095 0.57% – 0.0934
and others. Hence, it is of practical importance to investigate vibration 2 – – – 0.1002
mitigation techniques for this bridge. 3 0.133 0.52% – 0.1324
4 – – – 0.1782
5 0.183 0.50% Y 0.1840
4.2. Finite element model 6 0.230 0.51% Y 0.2286
7 – – – 0.2599
Fig. 9 demonstrates the finite element model (FEM) of the bridge, 8 0.276 0.39% Y 0.2728
where the main cables and suspenders are modeled as link elements 9 0.327 0.42% Y 0.3231
10 0.378 – Y 0.3731
with pretension forces and the towers are modeled using beam ele-
11 – – – 0.3969
ments. The twin-box girder is modeled as two main beams intercon- 12 0.435 0.37% Y 0.4266
nected through cross beams, all using beam elements. The model has 13 0.493 0.32% Y 0.4805
1882 nodes and 2900 elements in total. The main cables are anchored 14 – – – 0.5293
at both ends and connected to the towers at the top without sliding. *It is noted that when no dampers are placed between the girder and the tower/pier, the
The two main beams are fixed at the north end and the south tower in 1st mode involves large amplitude vibrations of the girder in the longitudinal direction,
the transverse and vertical directions. At the north tower, the beams are and the frequency is 0.1106. With a small amount damping provided by the damped
outriggers, the vibration in the longitudinal direction is fully suppressed and this mode
restricted in the transverse direction, and their rotations with respect
is dominated by girder vertical vibration, and the frequency is 0.0934.
to the longitudinal axis of the bridge are restricted at both towers. In
the longitudinal direction, horizontal dampers have been designed to
connect the girder to the north tower and the north pier. In this study,
as the damped outriggers also suppress the longitudinal bridge motion, Dynamic analyses of the bridge without any damped outriggers is
the dampers in the original design are not considered any longer. first carried out. According to Ref. [19] and Ref. [18], VIVs in the
For the bridge, damped outriggers can be installed at three posi- modes with frequencies of 0.183 Hz, 0.230 Hz, 0.276 Hz, 0.327 Hz,
tions, i.e., the north tower, south tower, and the north anchor. They 0.378 Hz, 0.435 Hz and 0.493 Hz have been observed, and hence
are indexed as damped outrigger 1, 2 and 3 respectively. For each vertical bending modes of the bridge with frequencies less than 0.53 Hz
damped outrigger, two vertical rigid arms are respectively connected to are all considered. The extracted modal frequencies and mode shapes
the junction points of the two main beams and the cross-beam closest of the first 15 vertical vibration modes are illustrated in Fig. 10. Corre-
to a tower or the north anchor. The rigid outrigger is of 15 m high and spondingly, the frequencies are compared with measured frequencies if
the horizontal distance of the outrigger and the tower/anchor depends they are available in the literature [17–19] in Table 2, and consistence
on the location of the cross-beam, as shown in Fig. 9. The damper is is found, suggesting the accuracy of the FEM. The damping ratios
placed horizontally in the longitudinal direction of the bridge and the estimated from field measurement/monitoring data are also listed in
other end of the damper is connected to a tower or fixed in the case of the table. Note that the measurements were conducted when the bridge
damped outrigger 3. has dampers placed at the north tower and north anchor to restrict
longitudinal girder motion. In general, those dampers have marginal
4.3. Dynamic analyses damping effects on vertical vibrations, as seen from the table. The
structural modes that have been observed in VIV events are marked
Before dynamic analyses, static analysis is performed using the FEM with Y in the table. Girder rotation amplitudes in those modes are also
of the bridge under self-weight. By this means, the stiffness of the shown Fig. 10, which shows that the rotation amplitudes at the damped
bridge is determined at the static equilibrium. The equation govern- outrigger locations are considerably large in most modes.
ing free vibrations of the bridge with damped outriggers is generally
expressed as 4.4. Damping provided by damped outriggers
𝐌𝐪̈ + 𝐂𝐪̇ + 𝐊𝐪 + 𝐟𝑜 = 𝟎, (12)
For given damped outrigger installation positions and nonzero
where 𝐌, 𝐊 and 𝐂 are respectively the structural mass, stiffness and damper coefficients, the computed frequencies are complex-valued
damping matrices, 𝐪 denotes the generalized coordinates, and 𝐟𝑜 rep- numbers and then through using Eq. (11) the damping can be obtained.
resents the force provided by the damped outriggers. In the following, By varying the damper coefficients from zero to a fairly large number,
𝐂 = 𝟎 is considered assuming zero inherent damping of the bridge. variations of bridge frequencies and damping ratios can be appreciated

7
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 9. The finite element model of the Xihoumen bridge with damped outriggers installed at the three possible positions.

Fig. 10. Mode shapes and corresponding girder rotation of the first 15 vertical modes of the bridge without damped outriggers. The first mode involves a large longitudinal
vibration that is fully suppressed by a small amount of damping and the frequency reduces to 0.0934 Hz, see Table 2. Circles indicate damped outrigger locations.

and thus the maximal achievable damping for each mode can be vibration mode becomes dominated by girder vertical bending, and
determined. In the following, the bridge with different numbers of hence the maximal damping is quite large and the frequency variation
damped outriggers are discussed respectively. is obvious. The 7th mode, as seen in Fig. 10, is dominated by vibrations
in the side span which has a relatively small length as compared to the
4.4.1. Damped outrigger 1 main span and hence the maximal damping is also large. Fig. 11(a)
When only damped outrigger 1 is installed, variations of the fre- represents the frequency variations of this case. The variations are
quencies and damping ratios for the first 14 vertical modes are illus- pretty small and hence will not bee shown in the following cases.
trated in Fig. 11. Unsurprisingly, the frequency variations for most From Fig. 11(b–d), it is seen that for each mode there is one
modes are quite small and correspondingly the attainable damping is damper coefficient that can maximize the attainable modal damping,
generally not large, except for the 1st and the 7th mode. For the 1st and for different modes the maximal damping and corresponding opti-
mode, the vibration has a relatively large amplitude of the girder in mal damper coefficients vary considerably. These variations are mainly
the longitudinal direction and with the damped outrigger installed the attributed to the location of the outrigger and supporting effect of the

8
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 11. Dynamic characteristics of the bridge with damped outrigger 1: (a) frequency variations; (b–d) damping curves.

Fig. 12. Damping curves of the bridge with damped outrigger 2.

tower at damped outrigger 1. In this case, maximal attainable damping the case when damped outrigger 1 is installed, generally more modes
ratios of 7 modes, i.e., those in Fig. 11(b,d), are larger than 0.5%. In can achieve relatively high damping when the damped outrigger 2 is
contrast, the supplemental damping to the 4th, 5th, and 14th modes is installed, similar as observed in Section 3.3 based on the simplified
quite small, with damping ratio smaller than 0.2%, as in Fig. 11(c). A model. However, variations in maximal damping attained in different
comparison of the mode shapes in Fig. 10 shows the modes that can modes are larger, mainly because the presence of the intermediate
achieve a larger damping have mode shapes with a relatively large support (north tower) and also the longitudinal motion of the girder
slope at damped outrigger 1, and vice versa. that has not been considered by the model in Section 3. In general, the
maximal damping is around 1.0% for most modes, which is roughly
4.4.2. Damped outrigger 2 close to that analyzed using the simplified model, see Section 3.3. Note
The damping curves of the bridge with damped outrigger 2 are the parameter 𝛾 for the Xihoumen Bridge is 77.14 and in Section 3.3
plotted in Fig. 12. The damping for four modes are less than 0.1%, 𝛾 = 50 is adopted in most cases.
i.e., the 3rd, 5th, 7th, and 11th modes. It is seen from Fig. 10 that the The damping effect can also be estimated using Eq. (9) to avoid
mode shapes of the 5th, 7th and 11th modes are relatively flat at the complex eigenanalysis. Table 3 compares the maximal damping ra-
south tower and hence the supplied damping by rotational dampers tios obtained from the two real frequencies and those computed by
is small. For the 3rd mode, the mode shape has relatively large slope complex eigenanalysis. In many modes, Eq. (9) gives quite accurate
while the damper deformation could be decreased by the longitudinal estimates, e.g., the 9th to 14th modes. However, vibrations of the
motion of the girder and hence the supplemental damping is extremely bridge are coupled in three directions and the bridge has close-spaced
small. Fig. 12 shows that all the other modes can attain a maximal mode frequencies. Hence, in some modes the accuracy of the estimated
damping larger than 0.5% and particularly these modes in Fig. 12(a) damping is low. For examples, the clamped frequencies of the 1st,
have maximally achievable damping larger than 1.0%. Compared to 4th to 6th modes are decreased as compared to the corresponding

9
L. Chen et al. Engineering Structures 271 (2022) 114873

Table 3
Comparison of maximal modal damping ratios computed using Eq. (9) and from complex eigenanalysis using the FEM when
the bridge is installed with damped outrigger 2.
Mode Undamped Clamped Frequency Estimated Exact
no. Frequency (Hz) Frequency (Hz) difference (10−2 Hz) 𝜁𝑛,max (%) 𝜁𝑛,max (%)
1 0.1106 0.0948 −1.58 7.14 10.53
2 0.1002 0.1016 0.14 0.70 0.83
3 0.1324 0.1324 0.00 0.00 0.01
4 0.1782 0.1757 −0.25 0.70 0.75
5 0.1840 0.1838 −0.02 0.05 0.06
6 0.2286 0.2148 −1.38 3.02 4.86
7 0.2599 0.2600 0.01 0.02 0.01
8 0.2728 0.2833 1.05 1.92 2.26
9 0.3231 0.3326 0.95 1.47 1.60
10 0.3730 0.3819 0.89 1.19 1.26
11 0.3969 0.3971 0.02 0.03 0.03
12 0.4266 0.4359 0.93 1.09 1.13
13 0.4805 0.4895 0.90 0.94 0.98
14 0.5293 0.5341 0.48 0.45 0.47

Fig. 13. Damping curves of the bridge with damped outrigger 3.

undamped frequency. Therefore, in this section, complex eigenanalyses 4.4.5. On practical design
are performed to obtain damping ratios of the bridge. As shown above, when damped outrigger is installed at one position,
it is preferred to be installed at the south tower, i.e., damped outrigger
4.4.3. Damped outrigger 3 2. From Fig. 12, the damper coefficient is determined according to the
When damped outrigger 3 is installed, the supplemental damping intersection of the damping curves corresponding to the 8th and 12th
to modes with relatively large vibration amplitudes in the side span is
modes, i.e., 𝑐𝑜 = 0.32 × 105 kN s/m. Then, the damping ratios provided
generally high and vice versa, as seen in Fig. 13. Also, the modes with
to the 14 modes are listed in Table 4, where 9 out 14 modes attain a
relatively large amplitudes in longitudinal motion can also attain high
damping ratio larger than 0.5% and 7 modes have damping ratio close
damping, i.e., the 1st mode. Particularly, for the 7th and 11th modes
to or larger than 1.0%. The results confirm the multimode vibration
where the bridge only vibrates in the side span (see Fig. 10), the damp-
ing ratios are larger than 3.0%; while for modes with only vibrations suppression effect. Particularly, among the 7 modes that have been
in the main span, the damped outrigger has nearly no damping effects, involved in the observed VIV events, 6 modes attain a damping ratio
e.g., the 4th and 9th modes. In this case, the performance of the damped of 1.0% or larger except for the 5th mode. Such amount of damping is
outrigger near bridge girder end is different from that in Fig. 6 mainly found able to effectively suppress bridge VIVs according to wind tunnel
because of the intermediate supporting effect of the north tower. test [7].
Considering the installation of both damped outrigger 1 and 2 with
4.4.4. Damped outriggers 1 & 2 𝑐𝑜1 = 𝑐02 , the damper coefficient is determined roughly from Fig. 14(b)
Both damped outriggers 1 and 2 are installed in this case. Corre- according to the intersection of the damping curves corresponding to
spondingly, the damping curves for the 14 modes plotted in Fig. 14, the 6th and 12th modes, i.e., 𝑐𝑜1 = 𝑐02 = 0.36×105 kN s/m. The damping
with the damper coefficients considered equal, i.e., 𝑐𝑜1 = 𝑐𝑜2 . By ratios for the 14 modes are listed in Table 4. It is seen that 11 modes
comparing Fig. 14 with Figs. 11 and 12, an interaction between the have achieved a damping over 0.5% and damping of 8 modes is 1.0%
damped outriggers installed at the two positions is seen. On the one or above. It is also seen that damping of some modes is reduced as
hand, the maximal damping is increased from below to over 1.0% for
compared to case with only damped outrigger 2 installed. It is noted
the 3rd and 14th modes, when two damped outriggers are installed. On
that in both cases, the damping of the 5th mode is small because the
the other hand, the maximal damping ratios of the 8th and 9th modes
corresponding mode shape is flat at all the damped outrigger positions,
are decreased as compared to the case when only damped outrigger 2
as shown in Fig. 10.
is installed.
It is expected to balance the multimode damping effects of the two It can be concluded that by adding more damped outriggers, more
damped outriggers by adjusting the ratio between the two damper modes can achieve relatively large damping, while the parameters
coefficients. Furthermore, when three damped outriggers are installed, of the system need to be carefully optimized. The required damper
the multimode damping performance may further be improved. The op- coefficients, e.g., 𝑐𝑜1 = 0.32 × 105 kN s/m, is relatively large, while it
timization of damped outrigger parameters desires an extensive study is achievable by combining regular dampers. Besides, the length of the
and will be a topic of future study. outrigger can be increased as such the damper size is reduced.

10
L. Chen et al. Engineering Structures 271 (2022) 114873

Fig. 14. Damping curves of the bridge with damped outrigger 1 and damped outrigger 2 (𝑐𝑜1 = 𝑐𝑜2 ).

Table 4
Multimode damping of the bridge with damped outriggers.
Mode VIVs Frequency Structural damping Supplemental damping ratio (%)
no. observed [19] (Hz) ratio (%) Damped outrigger 2 Damped outrigger 1 & 2
1 – 0.0934 0.57 3.65 1.60
2 – 0.1013 – 0.61 0.37
3 – 0.1321 0.50 0.01 0.96
4 – 0.1782 – 0.61 0.31
5 Y 0.1840 0.50 0.06 0.04
6 Y 0.2286 0.51 1.13 1.03
7 – 0.2599 – 0.00 0.66
8 Y 0.2728 0.39 1.08 0.96
9 Y 0.3231 0.42 1.27 1.15
10 Y 0.3731 – 1.17 2.04
11 – 0.3969 – 0.02 1.32
12 Y 0.4267 0.37 1.12 1.03
13 Y 0.4806 0.32 0.98 1.85
14 – 0.5293 – 0.44 1.26

5. Conclusions • For a suspension bridge of 1650 m, by installing one damped out-

rigger at the bridge tower, 7 out of the first 14 vertical modes can
This study proposes a novel vibration mitigation strategy for long- achieve a minimum damping ratio of 0.98%, exhibiting excellent
span bridges, which uses damped outriggers installed between the performance in multimode vibration suppression.
bridge girder and a tower/pier to supply damping. It is novel in
The performance of the proposed damped outriggers can be sensitive to
the sense that rotations of bridge girder are constrained for energy
factors that can reduce rotations of the girder at the tower or the pier
dissipation while in the previous studies the focus is placed on trans-
in bridge girder vibrations, e.g., rotational constraints. Its performance
lational displacements. The main advantage is that the bridge girder
also needs experimental validation. These aspects will be investigated
rotations are relatively large when close to a tower or a pier and
in future.
hence the dampers installed on the bridge tower/pier are sufficiently
effective, rendering the strategy implementable in practice. Based on
analytical studies using a simplified model of a suspension bridge with CRediT authorship contribution statement
damped outriggers and numerical studies using a finite element model
of an existing suspension bridge with damped outriggers, the following Lin Chen: Conceptualization, Software, Methodology, Investiga-
conclusions can be drawn. tion, Writing – original draft. Zhanhang Liu: Software, Methodology,
Investigation, Writing – original draft. Satish Nagarajaiah: Concep-
• The proposed damped outriggers are effective in supplemental
tualization, Funding acquisition, Writing – review & editing. Limin
damping to multiple modes of long-span bridges.
Sun: Conceptualization, Investigation, Supervision, Writing – review &
• The damping effect of a damped outrigger on a bridge increases
editing. Lin Zhao: Resources, Methodology, Writing – review & editing.
when the stiffness of the outrigger increases and when the relative
Wei Cui: Resources, Methodology, Writing – review & editing.
bending stiffness of the girder as compared to the axial stiffness
of the main cables increases.
• When two or more damped outriggers are installed at different Declaration of competing interest
locations of a suspension bridge with a main span and side spans,
their parameters need to be optimized with the interaction effect The authors declare that they have no known competing finan-
considered. A combination of damped outriggers can increase the cial interests or personal relationships that could have appeared to
number of modes that can attain a relatively high damping but influence the work reported in this paper.
the damping of particular modes can be decreased.
• For a suspension bridge with a main span and a side span,
when only one damped outrigger is installed, the best multimode Data availability
damping performance is achieved when the damped outrigger is
installed at the girder end near a tower. Data will be made available on request.

11
L. Chen et al. Engineering Structures 271 (2022) 114873

[ ]⊤
Acknowledgments where θ = 𝜃0 𝜃1 ⋯ 𝜃𝑗 ⋯ 𝜃𝑁 𝜃𝑁+1 and

This study was partly supported by the National Natural Science ⎡−2 2 0⎤
⎢−1 0 1 ⎥
Foundation of China (Grant No. 52278314), and by the State Key ⎢ ⎥
1 ⎢ −1 0 1 ⎥.
Laboratory of Disaster Reduction in Civil Engineering of China (Grant 𝐓= (A.12)
no. SLDRCE13-MB-01), which is gratefully acknowledged. 2𝑎 ⎢ ⋱ ⋱ ⋱ ⎥
⎢ −1 0 1⎥⎥
⎢
⎣0 −2 2⎦
Appendix. Finite difference model of tensioned beams
The undamped eigenvector of the system described in Eq. (8) is
For frequency and damping analyses of the tensioned beam with denoted by 𝐮0𝑛 and normalized by 𝐮⊤ 𝐌𝐮0𝑛 = 1. Furthermore, the
0𝑛
damped outrigger, the finite difference model formulated in Ref. [55] vector is augmented to include the end nodes, i.e.,
is used. The tensioned beam is discretized into a total of 𝑁 +1 elements
with a uniform length of 𝑎 = 𝐿∕(𝑁 + 1). The nodes are indexed by 𝐮̃ 0𝑛 = [0 𝐮⊤
0𝑛
0]⊤ . (A.13)
𝑗 = 0, … , 𝑁 + 1. The left and right ends correspond to 𝑗 = 0 and Thus, the differential rotation 𝜌𝑛𝑖 is given [54] as
𝑗 = 𝑁 + 1, respectively. Correspondingly, an 𝑁-degree-of-freedom ( )
discrete system is obtained. Inherent damping is ignored. Without any 𝜌𝑛𝑖 = 𝐰⊤
𝑜𝑖 𝐓̃
𝐮0𝑛 , (A.14)
external dampers, mass and stiffness matrices of the system are given
where the (𝑁 + 1) × 1 vector 𝐰𝑜𝑖 indicates the location of the damped
as
outrigger. For example, when the damped outrigger is installed at the
⎡ 𝑚1 0 ⎤ left end,
⎢ 𝑚2 ⎥
⎢ ⎥ 𝐰𝑜 = [1 0 0 ⋯ 0]⊤ . (A.15)
𝑚3
𝐌=⎢ ⎥,
⎢ ⋱ ⎥ In the case of a flexible outriggers with 𝑐𝑜 = ∞, the stiffness matrix
⎢ 𝑚𝑁−1 ⎥
⎢ ⎥ 𝐊 is modified. For example, when one damped outrigger is installed at
⎣0 𝑚𝑁 ⎦ the left beam end,
⎡𝑄 𝑈 𝑊 0⎤
⎢𝐷 ⎥ 𝑀𝑜 = 𝑟2𝑜 𝑘𝑜 𝜃0 ≈ 𝑟2𝑜 𝑘𝑜 𝑣1 ∕𝑎. (A.16)
𝑆 𝑈 𝑊
⎢ ⎥
𝑉 𝐷 𝑆 𝑈 ⋱
𝐊=⎢ ⎥, (A.1) The stiffness matrix thus becomes
⎢ 𝑉 𝐷 ⋱ ⋱ 𝑊⎥
⎢ ⎡𝑟2 𝑘𝑜 𝑣1 ∕𝑎3 ⎤
𝑈 ⎥⎥
0 ⋯ 0
⎢ ⋱ ⋱ 𝑆 ⎢ 𝑜 ⎥
0 0 ⋯ 0
⎣0 𝑉 𝐷 𝑇⎦ 𝐊=𝐊+⎢
̃ ⎥. (A.17)
⎢ ⋮ ⋮ ⋱ ⋮ ⎥
In the mass matrix, 𝑚𝑗 = mass per unit length at node 𝑗. In the stiffness ⎢ 0 0 ⋯ 0 ⎥
⎣ ⎦
matrix,
( ) 2𝐻
⎧ 1 −5𝐸𝐼 fixed at 𝑗 = 1,
𝑗+1 + 22𝐸𝐼𝑗 − 3𝐸𝐼𝑗−1 +
⎪ 2𝑎4 𝑎2 References
𝑄=⎨ 1 ( ) 2𝐻 (A.2)
⎪ 4 −3𝐸𝐼𝑗+1 + 18𝐸𝐼𝑗 − 5𝐸𝐼𝑗−1 + 2 pinned at 𝑗 = 1,
⎩ 2𝑎 𝑎 [1] Fujino Y, Siringoringo D. Vibration mechanisms and controls of long-span
bridges: a review. Struct Eng Int 2013;23(3):248–68.
( )
⎧ 1 −3𝐸𝐼 2𝐻
fixed at 𝑗 = 𝑁, [2] Zhao L, Cui W, Shen X, Xu S, Ding Y, Ge Y. A fast on-site measure-analyze-
𝑗+1 + 22𝐸𝐼𝑗 − 5𝐸𝐼𝑗−1 +
⎪ 2𝑎4 𝑎2 suppress response to control vortex-induced-vibration of a long-span bridge.
𝑇 =⎨ 1 ( ) 2𝐻 Structure 2022;35:192–201.
⎪ 4 −5𝐸𝐼𝑗+1 + 18𝐸𝐼𝑗 − 3𝐸𝐼𝑗−1 + pinned at 𝑗 = 𝑁, [3] Dan D, Yu X, Han F, Xu B. Research on dynamic behavior and traffic manage-
⎩ 2𝑎 𝑎2
ment decision-making of suspension bridge after vortex-induced vibration event.
(A.3) Struct Health Monit 2021;14759217211011582.
[4] Ge Y, Zhao L, Cao J. Case study of vortex-induced vibration and mitiga-
1 ( ) 2𝐻 tion mechanism for a long-span suspension bridge. J Wind Eng Ind Aerodyn
𝑆= −2𝐸𝐼𝑗+1 + 10𝐸𝐼𝑗 − 2𝐸𝐼𝑗−1 + , (A.4) 2022;220:104866.
𝑎4 𝑎2 [5] Zhu Q, Xu YL, Zhu L, Li H. Vortex-induced vibration analysis of long-span bridges
1 ( ) 𝐻 with twin-box decks under non-uniformly distributed turbulent winds. J Wind
𝐷= 2𝐸𝐼𝑗+1 − 6𝐸𝐼𝑗 − , (A.5) Eng Ind Aerodyn 2018;172:31–41.
𝑎4 𝑎2
[6] Zhao H-W, Ding Y-L, Li A-Q, Liu X-W, Chen B, Lu J. Evaluation and early
1 ( ) 𝐻 warning of vortex-induced vibration of existed long-span suspension bridge using
𝑈= −6𝐸𝐼𝑗 + 2𝐸𝐼𝑗−1 − , (A.6)
𝑎4 𝑎2 multisource monitoring data. J Perform Constr Facil 2021;35(3):04021007.
1 ( ) [7] Seo J-W, Kim H-K, Park J, Kim K-T, Kim G-N. Interference effect on vortex-
𝑉 = 𝐸𝐼𝑗+1 − 2𝐸𝐼𝑗 − 𝐸𝐼𝑗−1 , and (A.7) induced vibration in a parallel twin cable-stayed bridge. J Wind Eng Ind Aerodyn
2𝑎4 2013;116:7–20.
1 ( ) [8] Ge Y, Zhao L, Xu K. Review and reflection on vortex-induced vibration of
𝑊 = 𝐸𝐼𝑗+1 + 2𝐸𝐼𝑗 − 𝐸𝐼𝑗−1 . (A.8) main girders of long-span bridges. China J Highw Transp 2019;32(10):1–18, in
2𝑎 4
Chinese.
The nodal displacement vector is defined as [9] Gao D, Deng Z, Yang W, Chen W. Review of the excitation mechanism and aero-
dynamic flow control of vortex-induced vibration of the main girder for long-span
𝐪 = [𝑣1 𝑣2 ⋯ 𝑣𝑗 ⋯ 𝑣𝑁 ]⊤ . (A.9) bridges: A vortex-dynamics approach. J Fluids Struct 2021;105:103348.
[10] Larsen A, Esdahl S, Andersen JE, Vejrum T. Storebælt suspension bridge–vortex
As lateral displacements are constrained at the beam ends, an aug- shedding excitation and mitigation by guide vanes. J Wind Eng Ind Aerodyn
mented displacement vector is defined as 2000;88(2–3):283–96.
[11] Xin D, Zhan J, Zhang H, Ou J. Control of vortex-induced vibration of a long-span
𝐪̃ = [0 𝐪⊤ 0]⊤ . (A.10) bridge by inclined railings. J Bridge Eng 2021;26(12):04021093.
[12] Bai H, Ji N, Xu G, Li J. An alternative aerodynamic mitigation measure for
Using central differences at intermediate nodes and forward/backward improving bridge flutter and vortex induced vibration (VIV) stability: Sealed
differences at the boundary points, beam rotations can be computed traffic barrier. J Wind Eng Ind Aerodyn 2020;206:104302.
from the nodal displacements by [13] Yang W-H, Chen W-L, Li H. Suppression of vortex-induced vibration of single-box
girder with various angles of attack by self-issuing jet method. J Fluids Struct
θ = 𝐓𝐪,
̃ (A.11) 2020;96:103017.

12
L. Chen et al. Engineering Structures 271 (2022) 114873

[14] Sarwar MW, Ishihara T. Numerical study on suppression of vortex-induced [40] Wang M, Nagarajaiah S, Sun F-F. Dynamic characteristics and responses
vibrations of box girder bridge section by aerodynamic countermeasures. J Wind of damped outrigger tall buildings using negative stiffness. J Struct Eng
Eng Ind Aerodyn 2010;98(12):701–11. 2020;146(12):04020273.
[15] Larsen A. Aerodynamic aspects of the final design of the 1624 m suspension [41] Wang M, Nagarajaiah S, Sun F-F. Optimal design of supplemental negative
bridge across the great belt. J Wind Eng Ind Aerodyn 1993;48(2–3):261–85. stiffness damped outrigger system for high-rise buildings resisting multi-hazard
[16] Fujino Y, Yoshida Y. Wind-induced vibration and control of trans-tokyo bay of winds and earthquakes. J Wind Eng Ind Aerodyn 2021;218:104761.
crossing bridge. J Struct Eng 2002;128(8):1012–25. [42] Sun F-F, Wang M, Nagarajaiah S. Multi-objective optimal design and seismic per-
[17] Li H, Laima S, Ou J, Zhao X, Zhou W, Yu Y, Li N, Liu Z. Investigation of vortex- formance of negative stiffness damped outrigger structures considering damping
induced vibration of a suspension bridge with two separated steel box girders cost. Eng Struct 2021;229:111615.
based on field measurements. Eng Struct 2011;33(6):1894–907. [43] Xing L, Zhou Y, Huang W. Seismic optimization analysis of high-rise buildings
[18] Li H, Laima S, Zhang Q, Li N, Liu Z. Field monitoring and validation of vortex- with a buckling-restrained brace outrigger system. Eng Struct 2020;220:110959.
induced vibrations of a long-span suspension bridge. J Wind Eng Ind Aerodyn [44] Shoraka MB, Woo D, Archila M, Flynn L. Stantec tower: Shear wall and outrigger
2014;124:54–67. design. Struct Eng Int 2017;27(1):49–53.
[19] Li S, Laima S, Li H. Data-driven modeling of vortex-induced vibration of a [45] Neff Patten W, Sun J, Li G, Kuehn J, Song G. Field test of an intelligent
long-span suspension bridge using decision tree learning and support vector stiffener for bridges at the I-35 walnut creek bridge. Earthq Eng Struct Dyn
regression. J Wind Eng Ind Aerodyn 2018;172:196–211. 1999;28(2):109–26.
[20] Huang Z, Li Y, Hua X, Chen Z, Wen Q. Automatic identification of [46] Gastineau AJ, Wojtkiewicz SF, Schultz AE. Lifetime extension of a realistic model
bridge vortex-induced vibration using random decrement method. Appl Sci of an in-service bridge through a response modification approach. J Eng Mech
2019;9(10):2049. 2013;139(12):1681–7.
[21] Larsen A, Poulin S. Vortex-shedding excitation of box-girder bridges and [47] Zoltowski K, Banas A, Binczyk M, Kalitowski P. Control of the bridge span
mitigation. Struct Eng Int 2005;15(4):258–63. vibration with high coefficient passive damper. Theoretical consideration and
[22] Huang Z, Chen Z. Application of multiple tuned mass damper for higher-order application. Eng Struct 2022;254:113781.
vortex-induced vibration of suspension bridge with steel box girder. J Vib Eng [48] Nagarajaiah S, Chen L, Wang M. Adaptive stiffness structures with dampers:
2013;26(6):908–14, in Chinese. Seismic and wind response reduction using passive negative stiffness and inerter
[23] Xu K, Bi K, Han Q, Li X, Du X. Using tuned mass damper inerter to mitigate systems. J Struct Eng 2022;148(11):04022179.
vortex-induced vibration of long-span bridges: Analytical study. Eng Struct [49] Han F, Deng Z, Dan D. Vertical vibrations of suspension bridges: a review and
2019;182:101–11. a new method. Arch Comput Methods Eng 2021;28(3):1591–610.
[24] Xu K, Bi K, Ge Y, Zhao L, Han Q, Du X. Performance evaluation of inerter-based [50] Chen Y, McFarland D, Spencer Jr B, Bergman L. Exact solution of free vibration
dampers for vortex-induced vibration control of long-span bridges: A comparative of a uniform tensioned beam combined with both lateral and rotational linear
study. Struct Control Health Monit 2020;27(6):e2529. sub-systems. J Sound Vib 2015;341:206–21.
[25] Hua X, Huang Z, Chen Z. Multi-mode vertical vortex-induced vibration of suspen- [51] Main JA, Jones NP. Vibration of tensioned beams with intermediate damper. I:
sion bridges and control strategy. China J Highw Transp 2019;32(10):115–24, Formulation, influence of damper location. J Eng Mech 2007;133(4):369–78.
in Chinese. [52] Oliveto G, Santini A, Tripodi E. Complex modal analysis of a flexural vibrating
[26] Chen Z, Hua X, Niu H, Huang Z, Zhang H, Chen J. Technological innovations beam with viscous end conditions. J Sound Vib 1997;200(3):327–45.
in eddy current damping and its application in civil engineering. China J Highw [53] Krenk S. Complex modes and frequencies in damped structural vibrations. J
Transp 2020;33(11):83–100, in Chinese. Sound Vib 2004;270(4–5):981–96.
[27] Møller RN, Krenk S, Svendsen MN. Damping system for long-span suspension [54] Main JA, Krenk S. Efficiency and tuning of viscous dampers on discrete systems.
bridges. Struct Control Health Monit 2019;26(12):e2448. J Sound Vib 2005;286(1–2):97–122.
[28] Smith RJ, Willford MR. The damped outrigger concept for tall buildings. Struct [55] Mehrabi AB, Tabatabai H. Unified finite difference formulation for free vibration
Des Tall Spec Build 2007;16(4):501–17. of cables. J Struct Eng 1998;124(11):1313–22.
[29] Infanti S, Robinson J, Smith R. Viscous dampers for high-rise buildings. In: [56] Karoumi R. Some modeling aspects in the nonlinear finite element analysis of
Proceedings of the 14th world conference on earthquake engineering, Vol. 14. cable supported bridges. Comput Struct 1999;71(4):397–412.
International Association for Earthquake Engineering; 2008, p. 1–8. [57] Kim K-S, Lee HS. Analysis of target configurations under dead loads for
[30] Gamaliel R. Frequency-based response of damped outrigger systems for tall cable-supported bridges. Comput Struct 2001;79(29–30):2681–92.
buildings (Ph.D. thesis), Massachusetts Institute of Technology; 2008. [58] Xiang H. Advanced theory of bridge structures. China Communications Press;
[31] Chen Y, McFarland D, Wang Z, Spencer Jr B, Bergman L. Analysis of tall 2013.
buildings with damped outriggers. J Struct Eng 2010;136(11):1435–43. [59] Dumanoglu A, Brownjohn J, Severn R. Seismic analysis of the fatih sul-
[32] Tan P, Fang C, Zhou F. Dynamic characteristics of a novel damped outrigger tan mehmet (second bosporus) suspension bridge. Earthq Eng Struct Dyn
system. Earthq Eng Eng Vib 2014;13(2):293–304. 1992;21(10):881–906.
[33] Tan P, Fang C, Chang C, Spencer B, Zhou F. Dynamic characteristics of novel [60] Kilic S, Raatschen H-J, Körfgen B, Apaydin N, Astaneh-Asl A. FE model of the
energy dissipation systems with damped outriggers. Eng Struct 2015;98:128–40. fatih sultan mehmet suspension bridge using thin shell finite elements. Arab J
[34] Fang C, Spencer Jr BF, Xu J, Tan P, Zhou F. Optimization of damped outrigger Sci Eng 2017;42(3):1103–16.
systems subject to stochastic excitation. Eng Struct 2019;191:280–91. [61] Deng X, Xie X, Xu G. Unified catenary calculation method for main cable of
[35] Das S, Tesfamariam S. Optimization of SMA based damped outrigger structure Wuhan Yangsigang Changjiang River Bridge in completion state. Bridge Constr
under uncertainty. Eng Struct 2020;222:111074. 2018;48(4):79–83, in Chinese.
[36] Zhou Y, Xing L. Seismic performance evaluation of a viscous damper- [62] Li T, Liu Z. An improved continuum model for determining the behavior of
outrigger system based on response spectrum analysis. Soil Dyn Earthq Eng suspension bridges during construction. Autom Constr 2021;127:103715.
2021;142:106553. [63] Chen L, Sun L, Nagarajaiah S. Cable vibration control with both lateral
[37] Ishii M, Kazama H, Miyazaki K, Murakami K. Application of tuned viscous mass and rotational dampers attached at an intermediate location. J Sound Vib
damper to super-high-rise buildings. In: Proceedings of the sixth world conference 2016;377:38–57.
of structural control and health monitoring. International Center for Numerical [64] Krenk S, Høgsberg JR. Damping of cables by a transverse force. J Eng Mech
Methods in Engineering (CIMNE); 2014. 2005;131(4):340–8.
[38] Asai T, Watanabe Y. Outrigger tuned inertial mass electromagnetic transduc- [65] Høgsberg JR, Krenk S. Linear control strategies for damping of flexible structures.
ers for high-rise buildings subject to long period earthquakes. Eng Struct J Sound Vib 2006;293(1–2):59–77.
2017;153:404–10. [66] Sun L, Chen L. Free vibrations of a taut cable with a general viscoelastic damper
[39] Liu L, Tan P, Ma H, Yan W, Zhou F. A novel energy dissipation outrigger system modeled by fractional derivatives. J Sound Vib 2015;335:19–33.
with rotational inertia damper. Adv Struct Eng 2018;21(12):1865–78. [67] Impollonia N, Ricciardi G, Saitta F. Dynamic behavior of stay cables with
rotational dampers. J Eng Mech 2010;136(6):697–709.

13