Engineering Structures 106 (2016) 154165

Contents lists available at ScienceDirect

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

Explicit modeling of damping of a single-layer latticed dome

with an isolation system subjected to earthquake ground motions
Huidong Zhang a, Qinghua Han b,, Yuanfeng Wang c, Yan Lu b
a

School of Civil Engineering, Tianjin Chengjian University, Tianjin 300384, PR China

School of Civil Engineering, Tianjin University, Tianjin 300072, PR China
c
School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, PR China
b

a r t i c l e

i n f o

Article history:
Received 25 November 2014
Revised 15 October 2015
Accepted 17 October 2015
Available online 11 November 2015
Keywords:
Single-layer latticed dome
Material damping
Damping at joints
Isolation bearing
Seismic performance

a b s t r a c t
The non-uniformly distributed material damping in single-layer latticed domes subjected to earthquake
ground motions has been ignored in engineering practice, and the structural damping of bearings and
joints has been modeled at a structural level in previous studies. In this paper, an explicit method for
modeling the material damping and structural damping using a finite element method is proposed.
The proposed method includes important characteristics for modeling single-layer latticed domes. The
steel material damping is directly taken into account using RambergOsgood material model (powerlaw) with hysteretic damping; the structural damping at bearings is modeled based on the bearing type.
The ball joints of domes tend to be simplified as nodes in the finite element method without considering
the actual geometric shapes and sizes of the balls for convenience. In this paper, the energy dissipation at
joints and modeling methods for damping are proposed and discussed in terms of the joint type in domes.
To illustrate the proposed method for computation, a typical single-layer latticed dome with base
isolation bearings subjected to six near-fault earthquake ground motions is selected as an example.
The dynamic demands of this single-layer latticed dome are analyzed using the simulation technology
proposed in this paper. The effects of key parameters of the dome on the damping forces, which are of
interest to many practitioners, are presented and discussed. Compared with the previous modeling methods for damping, the proposed method can model the damping effects of domes with a higher fidelity,
which eliminates the unrealistically high forces generated with conventional modeling methods.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
To reasonably evaluate the seismic dynamic performance of a
structure, the sources of energy dissipation of the structure must
be carefully considered, and a thorough understanding of the
damping properties of these sources is needed. However, these
damping properties are not only very complex, but also obscure
[1].
In structures, damping appears in several forms, which may be
broadly categorized as external and internal [2]. External sources
mainly come from active control systems such as the dampers used
to reduce the vibration of a structure and boundary effects such as
the loss of energy at the bearings through either friction or transmission into the supporting structure. The other form of external
damping is aero-acoustic effects caused by moving through a fluid
Corresponding author.
E-mail address: qhhan@tju.edu.cn (Q. Han).
http://dx.doi.org/10.1016/j.engstruct.2015.10.027
0141-0296/ 2015 Elsevier Ltd. All rights reserved.

such as air. Internal sources include the internal friction of materials and the friction at the joints in a structure [3]. These frictions
include [2]: macro-slip, micro-slip, structure and metal defects,
and visco-elastic characteristics, which cause different force
deformation (or stressstrain) relationships in a structure (or
material) during loading and unloading, even within the elastic
range. The area enclosed by the forcedeformation (or stress
strain) curve (hysteresis loop) serves as an indication of the
amount of energy dissipated. In fact, all damping is ultimately
caused by frictional effects, which may take place at different
scales [4]. As stated by Charney [5], a better solution is to utilize
a simple friction or hysteretic model to simulate these effects, such
as the BoucWen model. However, although the energy dissipation
mechanisms of some sources are known and their damping properties can be determined by experiments, these sources have often
not been considered in the previous dynamic analyses because of
the absence of an acceptable mathematical model to estimate
damping forces [6]. To accurately estimate the damping forces in

155

H. Zhang et al. / Engineering Structures 106 (2016) 154165

a structure, some non-viscous damping models were also developed by Adhikari [7], Woodhouse [8] and Val and Segal [9]. Thus,
it is recommended that the sensitivity of the calculated parameters
of a structure to the damping model formulations be investigated
[10].
For steel domes, a typical reported viscous damping ratio is
about 12% [11]. Because the damping ratio is small and it is difficult to be exactly quantified due to the estimation of structural
damping depending on a wide range of factors, the damping ratio
usually was not accurately considered in previous dynamic analyses. In respect to damping model, a simple damping model such as
linear viscous damping can be assumed without much concern.
The use of a viscous damping model tends to represent the damping effect at a global scale [12], and the current viscous damping
models are used based on the assumption of a small amplitude
(also called zero amplitude) [13]. Therefore, Rayleigh damping
is widely used within the elastic range of a structure. Rayleigh
damping models used in inelastic seismic analyses have also been
presented [1417]. In these studies, Rayleigh damping model was
modified to eliminate the unrealistically high damping force. However, Rayleigh damping still lacks physical evidence, and there is
no guarantee that the actual damping forces are properly modeled.
As shown in Petrini et al. [18], the initial-stiffness proportional
damping is inappropriate for an inelastic dynamic analysis, and
tangent-stiffness viscous damping appears to be more appropriate.
Jehel et al. [19] considers that it is easier to control the viscous
damping ratio by the use of Rayleigh damping based on the tangent stiffness in inelastic dynamic analyses. Some researchers have
suggested eliminating the stiffness proportional damping term and
only specifying a value for the mass proportional damping term
[10]. At present, there is no clear consensus on how to resolve this
issue. Nevertheless, the reliability of the results predicted by a nonlinear dynamic analysis is strongly dependent on the selected
damping model. Extensive research indicates [7,20,21] that the
structural damping observed in practice is a strong function of
the displacement and a weak function of the frequency. Therefore,
a damping model based on the displacement should be the most
accurate. However, the damping model with this characteristic is
almost never used for spatial structures in a dynamic analysis
during earthquake engineering because of its complex dynamic
equations. In dynamic equations, it should be noted that only
the non-modeled energy dissipation needs to be considered in
the damping term. Otherwise, the damping sources should be
explicitly simulated [10].
In this paper, the damping mechanisms of a single-layer latticed
dome are analyzed and discussed. A dynamic analytical procedure
for modeling damping is proposed in OpenSEES. This procedure
includes important features that were not considered in the previous dynamic analyses. The damping forces of particular interest to
many practitioners are given and discussed.

pertaining to the transition between the elastic and plastic stages

of the stressstrain curve, which is also called the hardening exponent. a usually has a value of about 5 or greater. Based on numerous
experiments and calculations, the hardening ratio a = 8 is widely
used for a common steel material. It has also become a standard
practice to determine the value of a using the 0.01% and 0.2% proof
stresses [23],

ln20
lnr0:2 =r0:01

The stressstrain relationship described in Eq. (1) does not obey

Hookes law, even within the elastic range, and the strain is the
sum of the elastic and plastic components. As a result, the nonlinear material energy dissipation within the elastic range and the
material damping can be captured by hysteretic loops. The
stressstrain curves based on the RO equation with different
hardening ratios a are illustrated in Fig. 1. It can be noted in
Fig. 1 that the material strength increases with an increase in the
hardening ratio in the elastic phase, while in the plastic phase it
decreases with an increase in the hardening ratio.
2.2. Cyclic stressstrain behavior of material and steel material
damping within elastic range
The stressstrain hysteresis curve of the steel material for a = 8
is shown in Fig. 2. Here, the elastic modulus and yield strength of
the steel material are 200 GPa and 207 MPa, respectively. In
Fig. 2, it can be observed that the steel material exhibits a relationship between the stress and strain that is not elastic, even at a
stress well below the yield point. Thus, the stressstrain relationships during loading and unloading are different and lead to the
hysteresis loops due to the internal friction of the steel material.
These hysteresis loops represent the energy dissipation of the steel
material, and its energy dissipation capacity can be described by
means of the equivalent damping ratios under the different material stresses,

n D=4pW

400

a

250

225

200

Stress /MPa

Stress /MPa

The RO stressstrain relationship is a power law relation

between the stress and strain. The following RO equation is the
usual representation of the non-linear behavior of a steel material
[22].

r
e p
rp
E

=7
=10
=13

350
300

2.1. RambergOsgood (RO) steel material model

where D determines the amount of energy dissipated per unit volume of the material during one cycle of oscillation, W represents the
maximum elastic strain energy, and n is the material damping ratio.
For example, the maximum stress is equal to about 175 MPa, and
the damping ratio of the steel material is about 0.08, as illustrated
in Fig. 3. The viscous damping ratios under different stress levels
are evaluated and shown in Fig. 4. These results indicate that the
steel material damping ratio depends on the stress and on the

2. Steel material damping

150
100

where e is the strain, r is the stress, E is the initial elastic modulus, p

is the given level of plastic strain (typically 0.2%), rp is the proof
stress corresponding to the plastic strain p, and a is a parameter

0.2=207MPa

175

=7
=10
=13

150
125

50

200

0.001

0.002

0.003

0.004

0.005

Strain

0.00

0.02

0.04

0.06

0.08

0.10

Strain
Fig. 1. Stressstrain curves based on RO model with different hardening ratios a.

156

H. Zhang et al. / Engineering Structures 106 (2016) 154165

Stress /MPa

400

=8

Ramberg-Osgood steel material

300
200
100
0
0.000

-0.004

-0.008

0.004

0.008

Strain

-100

The energy dissipation of material

is captured by the loops.

-200

by Lazan [24]. But it should be particularly noted that the material

damping ratio of a member or structure is far below this value
because the stress in a member or structure is non-uniformly distributed and it does not simultaneously reach the high stress in
all parts of a member or structure. Generally, the average material
damping ratio of a steel structure is about 0.10.5% under a low
stress level [9,24]. However, the damping effect for a steel material
has often been ignored by the use of the bilinear material model or
simplified and implicitly modeled by means of a frequencydependent damping model at a structural level in the previous
dynamic analyses. A reasonable method for modeling the material
damping due to hysteretic behavior is to directly take into account
the damping property by means of a material constitutive model
with a hysteretic characteristic.

3. Discussion of damping within joints

Fig. 2. Stressstrain hysteresis curves.

3.1. Joints with slip characteristics

200

Stress /MPa

Hystertic hoop

150
100

0.0005

0.0010

Strain

-100

D=SB-C-B=87949 N*m/m

-150

W=SOAB=87500 N*m/m3

-200

(-175MPa,-0.001)

0O
0.0000
-50

-0.0005

=8

50

A
-0.0010

(157MPa,0.001)

=D/(4W)=0.080

-250
Fig. 3. Calculation of material damping ratio.

Material damping ratio

0.12

=7
=10
=13

0.10
0.08
0.06
0.04
0.02
0.00
60

80

100

120

140

160

180

200

220

Maximum stress /MPa

Fig. 4. Material damping ratios under different exponents and stress levels.

deformation demand, and increases with the stress instead of the

frequency. It can also be observed that the parameter a controls
the energy dissipation capacity, and the damping ratio quickly
decreases with an increase in a for a given stress. In the current
study, the maximal material damping ratio within the elastic range
can reach 10% or more because of the highly nonlinear behavior of
steel material. These results show that the material damping ratio
quickly increases at a higher stress. This effect has been reported

In steel structures, there are two types of joints: welded joints

and joints with slip characteristics such as bolted joints. The damping ratio of a bolted steel structure is usually between 0.01 and
0.025. Joints with slip characteristics exhibit two types of motion
during vibration: micro-slip and macro-slip [25]. Micro-slip is a
phenomenon that occurs between contacting surfaces when a
shear load that is smaller than that necessary to produce macroslip is applied. When the load continues to increase until relative
motion occurs between the contacting surfaces, the macro-slip
occurs. Beards [26] performed a series of experiments, which
showed that the damping due to interfacial slip in bolted joints
was much larger than the material damping. The energy dissipation mechanism of the joint is a complex process that is largely
influenced by the interface pressure. In order to directly simulate
this energy dissipation, modeling techniques for a joint with slip
characteristics have been proposed [27].
In this paper, a modeling method for a joint with slip characteristics is developed using an explicit modeling technique to capture
energy dissipation. As shown in Fig. 5(a), the uniaxial ElasticPPGap
material (Elastic-Perfectly Plastic Gap Material) has a defined shear
forcedeformation relationship to model a gap and shear bolt. It
should be noted that the relationship between the shear force
due to friction and slip distance is defined by means of the uniaxial
Steel01 material model to model the micro-slip and macro-slip of
the connection; the uniaxial Steel01 material is a uniaxial bilinear
steel material object with kinematic hardening and optional isotropic hardening in OpenSEES. The mechanical model with slip characteristics is given in Fig. 5(b). The connection consists of three
parallel springs. In the mechanical model, a parallel material made
up of these materials is constructed and assigned to the zeroLength
element. The hysteresis loops for this type of connection subjected
to cyclic loading are given in Fig. 6. Here, the assumed parameters
are used only to illustrate that the proposed method is able to
model the hysteretic behaviors of a connection with slip characteristics. The results show that when the macro-slip demand is
reached another additional phase due to the bolt shear is added
to the hysteretic loop with the parallelogram, and the shear force
is increased as a result of the shear resistance effect from the bolt.
Therefore, the energy dissipation capacity or damping performance
of the connection mainly depends on the sliding friction effect
between the interfaces. It has been reported that the damping
value of the joints with this mechanism accounts for approximately 90% of the total damping of a structure [2].
The above method is used in an attempt to model the actual
physics taking place within the joint on a meso-scale. Although
the explicit method can be applied to engineering, it is very

157

H. Zhang et al. / Engineering Structures 106 (2016) 154165

(a)

(b)
gap

Curves 1 and 2 are modeled by macro-slip

Steel01 material
micro-slip

Kbolt

Failure

4
Sheared bolt (Hooked
by member or plate)

Hook

Kbolt
Ps

Ps

3
Gap

symmetry

friction

Kinit

Curves 3 and 4 are modeled by

ElasticPPGap material

150

Failure x

100
80

120

Shear bolt or pin

micro-slip
40
macro-slip
60
20

0
-0.003 -0.002 -0.001-200.000
-40
-60
-80
-100

Moment versus Rotation

Hysteresis loops

Moment [kN*m]

Load, Ps /kN

Fig. 5. Mechanical model of bolt or pin connection.

Displacement /m
0.001 0.002 0.003 0.004 0.005

Static friction coefficient, =0.1

Sliding friction load, P=17.5kN
Diameter of bolt, d=0.024m
Yielding strength of bolt, f y=345MPa
Gap, G=0.002m

90
60
30
0
0.00

Assumed shear failure when shear

displacement of bolt Sd=0.0045m;

0.03

0.06

0.09

0.12

Rotation [rad]
Fig. 6. Hysteresis loops and their shapes.
Fig. 7. Moment versus rotation curve [29].

complicated when there are a lot of joints in a three dimensional

structure. An alternative approach is to look at the effect of the
joint on the overall dynamics of the structure. In this approach,
the microphysics needs not to be considered in detail, but rather
the overall dynamical effect of the joint needs to be examined.
The mass proportional damping can be used for modeling the
damping mechanism with slip characteristics [28]. The mass
damping coefficient of a joint, am, can be expressed as follows:

am 2na x

in which x is the fundamental circular frequency of a structure, and

na is the friction damping ratio of a joint. Based on experimental
tests, this approach has been shown to give an acceptable result
[28].
3.2. Welded joints
In large-span single-layer latticed domes, welded joints are
commonly used in engineering practice. Welded joints are much
stiffer, and may thus have smaller damping ratios than other types
of joints. For a welded steel structure, a damping ratio between
0.005 and 0.015 is typical [2]. When the construction quality is
good, these joints usually provide little friction damping between
the steel tube and ball. However, a finite element node tends to
be used to represent the zone of the joint to simplify the geometry
modeling. As a result, the energy absorption of the joint due to
deformation is not captured by the simplified mechanical model.
In this section, based on the previous literature [17], a direct damping modeling method is discussed to consider the added damping
of a welded joint.
In a single-layer latticed dome, the energy dissipation or
absorption mainly depends on the bending deformations of the

members and the welded balls. Kim et al. [29] investigated the
moment-resisting capacity of a welded ball joint. The moment
rotation curve is shown in Fig. 7. It can be seen that the welded
zone is stiff rather than infinitely stiff. Therefore, when the member deforms as a result of bending, the welding-ball also deforms.
The real welded ballmemberball group and its simplified
mechanical model for dynamic analysis are illustrated in Fig. 8.
Here, the shear effect of the steel tube is ignored, and the end
moments of the member are assumed to be equal.
For each group, Ka and Ke represent the elastic bending stiffnesses of the ball and member, respectively. It is assumed that ha
and he are the rotations of the welding ball and member, respectively, and Wa and We represent the energy dissipations from the
balls and member, respectively. The ratio of the elastic bending
stiffness of the ball joint, Ka, to the elastic member stiffness Ke is
defined as n = Ka/Ke. The total energy dissipation WT within the
elastic range can be expressed as follows [17]:

W T Mha Mhe K a h2a K e h2e

Ball M

n1
K e h2e
n

M Ball

Ball-Member-Ball group in engineering practice

M

Simplified mechanical model in analysis

Fig. 8. Ballmemberball group and simplified mechanical model.

158

H. Zhang et al. / Engineering Structures 106 (2016) 154165

where the energy dissipation that is not captured, Wa, is equal to 1/

n  (Kehe2). In the simplified mechanical model, when stiffnessproportional damping is used the energy dissipation, We, is given
as follows:

W e K e h2e

bK e h_ e dhe

where b is the stiffness-proportional damping coefficient of a member, he is the end rotation of the member and h_ e is the velocity of
rotation. According to Eq. (5), the added energy dissipation that is
not captured Wa is

Wa

1
K e h2e
n

b0 K e h_ e dhe

where b0 is the added stiffness-proportional damping coefficient for

a member. Through Eqs. (6) and (7), the following expression can be
found:

b0

1
b
n

Although the above analysis process is derived with a twodimensional ballmemberball group, the result can also be
applied in a three-dimensional case.
According to the above analyses, the equivalent viscous damping ratios of the member, the welded ball joints, and the ball
memberball group (n1, n2, and n) have the following relation:

n1 n2 n

The stiffness-proportional damping coefficient of the group, b0,

is equal to (b + b0 ). According to Eq. (9),

b0

10

4.1. Modeling method of HDRB

An HDRB consists of multiple elastomeric layers made of special
dissipative rubber compounds, separated by steel sheets embedded in the rubber layer, and can be manufactured in a rectangular
or circular shape. Fig. 9 shows a schematic diagram of an HDRB.
The nonlinear behaviors of the bearings are modeled by means of
a hysteresis model (KikuchiAikenHDR material) developed by
Kikuchi and Aiken [33] in OpenSEES. A mechanical model of the
bearing is illustrated in Fig. 10. This mechanical model consists
of three components: the upper steel plate, rubber layers and
sheets, and lower steel plate, which are modeled by zeroLength elements in OpenSEES. The influence of the axial load on the shear
behavior is not taken into account.
4.2. Modeling method of LRB
An LRB consists of alternating layers of steel and rubber, which
provide horizontal flexibility. To add damping, a hole is created in
the center and filled with lead. In an LRB, the energy is dissipated
by the rubber and inner lead core. The LRB makes use of the elasticplastic properties of lead to dampen the cyclical loading from a
seismic source [34]. This helps to dissipate the energy in the form
of heat as the crystalline structure of the lead breaks down and
reforms continuously during the load cycles. A schematic diagram

According to Eqs. (8)(10), the following relationships exist:

n2

1
n
n1

iNode

11

From Eq. (11), it can be observed that the added damping value
of the welded joints accounts for approximately 1/(n + 1) of the
total damping of the group. A larger value for n will result in a
smaller damping ratio from the joints. For example, when n = 4,
the damping accounts for 20% of the total damping. The proportion
of the damping from a welded joint in the total damping is far
below that of a joint with slip characteristics.
4. Modeling of isolation bearings
At present, seismic isolation is enjoying substantial growth
because many researches have shown that isolation bearings can
remarkably improve the seismic performance of the structures
subjected to seismic loads [30,31]. On one hand, significant seismic
energy may be dissipated in the isolators by hysteresis to reduce
the horizontal deformation demand; on the other hand, the fundamental period can be elongated by providing a base isolation material, which will prevent damage to a structure during an
earthquake. A successful use of earthquake base isolation in a
single-layer latticed dome has been reported in the literature [32].
The energy dissipation or damping capacity at bearings is
dependent on the type of bearing. It is clear that the damping of
bearings with isolation systems is only local rather than being
evenly distributed in the overall structure. In order to simulate
the absorbing energy of supplemental damping, the bearings
should be regarded as structural components, and their nonlinear
behaviors should be directly taken into account, therefore, the
use of Rayleigh damping should be avoided for bearing elements

jNode
x

Steel Plates

Steel Sheet

mNode
nNode
D
D0
Fig. 9. Schematic diagram of HDRB.

iNode
Steel Plate
(zeroLength Element)
jNode
Rubber
(zeroLength Element)
mNode
Steel Plate
(zeroLength Element)
nNode
Fig. 10. Mechanical model of bearing.

H0

n
n1 n1
n

n1
b
n

because the direct modeling method can adequately predict the

damping effects from bearings. In this study, two seismic isolation
technologies are considered: a high damping rubber bearing
(HDRB) and lead rubber bearing (LRB). Their damping performances are compared and discussed. The radiation damping at
the boundary is not considered in this paper.

159

H. Zhang et al. / Engineering Structures 106 (2016) 154165

of an LRB is presented in Fig. 11. The inner steel sheets not only
provide the bearing with a good vertical load capacity but also
properly confine the lead core. The KikuchiAikenLRB material in
OpenSEES is adopted to model the nonlinear behaviors of the
LRB. The modeling method is like that shown in Fig. 10.
5. Damping modeling for single-layer latticed domes

Member

Joint

Material Damping:R-O model

Sliding:

or

m = 2a

Welded:

' =

r
be
em
om
dt
e
d
Ad

in which

Welded:

= 21 /

1
n

' =

or

Sliding:

m = 2a

Fig. 12. Proposed modeling method for damping in single-layer latticed domes.

This paper proposes a method to model some damping mechanisms in single-layer latticed domes. There are three main damping sources in a single-layer latticed dome: material damping,
joint damping, and the damping at bearings.
In this paper, the steel material damping is explicitly captured
by means of an RO material model with hysteretic characteristics,
while the nonlinear behaviors of bearings are directly modeled
based on the types of bearings. Thus, some energy dissipation
can be directly considered by hysteretic behaviors.
For the damping mechanisms at joints, they were commonly
simplified as a finite element node for a joint zone in previous
dynamic analyses due to its geometry complexity. In such a case,
the energy dissipation of the zones could not be captured.
However, the energy dissipation can be considered by means of
the proposed methods in Sections 3.1 and 3.2.
The proposed method for damping modeling is illustrated in
Fig. 12. This modeling method can also be applied to other
structures. The method is implemented by using TCL (Tool Control
Language) for OpenSEES in TCLEditor context.
6. Illustrative example
6.1. Model description
A typical single-layer latticed dome with welded ball joints is
selected as an example to present the proposed modeling procedure for damping in Section 3.2 of the paper, as shown in Fig. 13.
Its span length is 50 m, and the height is 10 m. The members used
are steel tubes with a diameter of 180 mm and thickness of 5 mm.
The parameters of the steel material are listed in Table 1. The uniform roof load is set at 180 kg/m2, which is assumed to be concentrated at the joints as masses. The equivalent viscous damping of a
member is defined as 1%. According to Eq. (11), the total equivalent
damping ratio of the ballmemberball is equal to 1%  (1 + 1/n).
Because the viscous damping model is frequency-dependent and
assumed to be uniformly distributed in a structure, the equivalent
damping ratio of the structure is also equal to 1%  (1 + 1/n). The
stiffness with a time variant property (tangent stiffness) is adopted

Attachment Plates

Joint

to construct the damping matrix in the current study. This can

eliminate the unrealistically high viscous damping force when
the structure softens as a result of yield or instability.
All the analyses are performed using OpenSEES (OpenSEES
2013). Two base isolation bearings, HDRB and LRB, are adopted.
The bottom of every bearing is fully fixed. The inner rubber
material is Bridgestone HDR-X0.6 (standard compressive stress)
for HDRB [35]. In LRB, the ratio of the pre-yield stiffness to postyield stiffness of an LRB is difficult to determine. This ratio depends
on the material of the LRB, construction, and analytical model of
the isolated building. However, the international standard [36]
suggests that this value should be 1015, and the value is set at
13 in this paper. The selected material parameters for LRB are
listed in Table 2. These material parameters are selected from the
literature [37]. In the dome, the diameter and height of the
bearings are set as 0.6 m and 0.2 m, respectively. The diameter of
the lead core is 0.15 m.
The Displacement-based beam-column element with distributed
plasticity and a linear curvature distribution is used to model the
nonlinear behaviors of members, and the geometric nonlinearity
is also considered. Compared to a lumped plasticity model, the
material damping in the distributed plasticity model can be captured by fiber section elements, and the distributed plasticity
model has good computational efficiency [10]. In this paper, each
member is modeled as one element because only the macrodynamic demands are considered. Through the modal analysis, it
is observed that the first frequencies of the dome with HDRBs
and LRBs are 1.64 Hz and 2.45 Hz, respectively. The latter is stiffer.
6.2. Damping performance of isolation bearings
The shear hysteretic behaviors of the HDRB and LRB are modeled using the proposed method in Sections 4.1 and 4.2, and their
damping performances are evaluated. Fig. 14 shows the hysteretic
loops of an HDRB with a diameter D = 0.5 m and different rubber
heights under cyclic loads. It is observed that the energy
dissipation of the HDRB mainly depends on the rubber material,

Steel Sheet
Rubber Layers

Lead Plug
Force

d
D

d
D

D0

D0
Fig. 11. Schematic diagram of LRB.

H0

160

H. Zhang et al. / Engineering Structures 106 (2016) 154165

B

Top

H=10m

Top
Y

D=50m
Elevation
S2
S1

Plane

Fig. 13. A single-layer latticed dome.

Table 1
Steel material parameters.
Parameters

Value

Elastic modulus, E
Yield strength, fy
Hardening exponent, a
Yield offset

200 GPa
207 MPa
8
0.002

Table 2
Material parameters of LRB [37].
Gr (MPa)
Rubber
Lead

observed in Fig. 14 that the bearing with the minimum height

has the maximum energy dissipation under the same lateral deformation among these bearings. However, this bearing has the smallest damping ratios in Fig. 15, the reason is that the calculation of
equivalent damping ratio depends on not only the energy dissipation of the bearing but the maximum elastic strain energy according to Eq. (3) and Fig. 3.
Fig. 16 shows the hysteretic loops of bearings with different
lead core diameters. These results indicate that the diameter of
the lead core has a significant influence on the energy dissipation

rpb (MPa)

Gp

0.24

8.33

0.588

0.22

0.45

and the HDRB is obviously characterized as visco-elastic. When the

height H of the rubber is equal to 0.1 m, the bearing presents highly
nonlinear behaviors as the lateral deformation increases. The
shapes and areas of the hysteresis loops show that the HDRB has
a good energy dissipation capacity. However, as the height of the
rubber material increases, both the lateral stiffness and energy dissipation capacity decrease, and the hysteretic loops become
remarkably different. Fig. 15 shows the results for the equivalent
damping ratios of HDRBs. It can be noted that the damping ratios
are very high and linearly decrease with an increase of the lateral
deformation of the bearing, and a bearing with a larger height of
rubber material has a larger equivalent damping ratio. It is also

-0.4

-0.3

-0.2

0.16
0.14

D=0.5m

0.10

0.0

0.1

0
0.0

H=0.15m
D=0.5m

-200

0.1

0.2

0.3

0.4

Lateral deformation /m

0.4

600
400
200
0

-0.4

-0.3

-0.2

-0.1
-200

-400

-400

-600

-600

Fig. 14. Shear hysteresis curves of HDRBs.

0.3

Fig. 15. Damping ratios of HDRBs.

400

-0.1

0.2

Lateral deformation /m

d=0.1m
d=0.15m
d=0.2m

200

H=0.1m
H=0.15m
H=0.3m

0.12

Shear force /kN

D=0.5m

0.18

600

Shear force /kN

H=0.1m
H=0.15m
H=0.3m

Damping ratio

0.20

0.0

0.1

0.2

0.3

0.4

Lateral deformation /m

Fig. 16. Hysteretic loops of LRBs.

161

H. Zhang et al. / Engineering Structures 106 (2016) 154165

0.39

No. of Joint (Section B-B)

0

0.36

Mean of vertical peak responses /m

Damping ratio

0.33
0.30
0.27
0.24
0.21

H=0.15m
D=0.5m

0.18
0.15
0.12

d=0.1m
d=0.15m
d=0.2m

0.1

0.2

0.3

0.4

Lateral deformation /m

0.080

10 11 12

with HDRBs (A-A)

with LRBs (A-A)
with HDRBs (B-B)
with LRBs (B-B)

0.076
0.072
0.068
0.064
0.060
0.056
0.052
0.048

The ratio,n,is equal to 3.

No. of Joint (Section A-A)

Fig. 17. Damping ratios of LRBs.

Fig. 18. Mean demands of vertical peak responses of joints.

and the energy dissipation performance of the bearing can be

effectively improved as a result of the high absorption capacity
of the lead core. Although the energy dissipation increases with
the diameter of lead core, and the lateral shear performance is
improved, there seems to be little changed in the equivalent lateral
stiffness. This feature of an LRB was also reported by Feng et al.
[37]. Fig. 17 shows the equivalent damping ratios of LRBs with different lead core diameters. It should be noted that the LRB has a
larger damping ratio than the HDRB, and the equivalent damping
ratio decreases with an increase in the lateral deformation for a
given lead core diameter. It can also be seen that increasing the
diameter of the lead core can obviously increase the equivalent
damping ratio of the bearing.
6.3. Selected earthquake ground motions
Several seismologists have suggested that base-isolated buildings are vulnerable to large pulse-like ground motions generated
at near-fault locations [38,39]. Such base-isolated structures might
perform poorly because of large isolator displacements due to
long-period pulses associated with the near-fault motion. This
has led to considerable interest by practitioners. In this paper,
the dynamic demands of a base-isolated structure are obtained
under six near-fault earthquake ground motions. The peak ground
accelerations and velocities of the typical earthquake motions
selected in this paper are given in Table 3 [40]. After the gravity
loads are applied to the structure, each earthquake record is
applied in all three directions.
6.4. Results and discussions
For single-layer lattice domes or large-span space structures,
the vertical dynamic demands tend to control the structural
stability [41,42]. Therefore, the important vertical dynamic
demands are discussed here, including the vertical dynamic
responses and vertical damping forces of joints.

6.4.1. Mean demands of peak vertical responses of joints

The mean demands of the vertical peak responses of the internal joints on sections AA and BB (see Fig. 13) of the structure
subjected to six earthquake ground motions are shown in Fig. 18.
Here, the stiffness ratio n in Eq. (8) is set to 3. In this paper, the
mean demand is calculated using the following expression:

Dm

6
X
Di =6

12

where Dm is the mean of the vertical peak demands of all the earthquakes, and Di is the peak demand under the ith earthquake ground
motion in Table 3. In this figure, it is observed that, compared with
the dome with HDRBs, the peak responses of the dome with LRBs
decrease by around 25%. These results also show that the peak
deformation of the joint at the top of the dome is generally less than
that of other joints, while the joints at the 3rd and 4th rings have a
relatively larger vibration. In engineering practice, it should be
noted that although single-layer latticed domes that use the HDRBs
or LRBs can inhibit relative horizontal deformations, large vertical
deformations may also occur because their geometric shapes cause
them to lose vertical stability if the bearings lack sufficient lateral
stiffness.
6.4.2. Shear-deformation demands of isolation bearings
Figs. 19 and 20 show the results for the shear-deformation
demands of bearing P in the X and Y directions under the Northridge ground motion (Rinaldi station, 1994). For the HDRB, the
peak shear and deformation of the bearing can reach around
217 kN and 0.24 m in the X direction, respectively. These values
are around 108 kN and 0.076 m in the Y direction, respectively.
For LRB, the peak shear and deformation are around 186 kN and
0.055 m in the X direction and around 150 kN and 0.022 m in the
Y direction, respectively. It can be seen that, because it has less lateral stiffness, the HDRB experiences larger deformations than the

Table 3
Peak accelerations of six near-fault motions [40].
Near-fault earthquake motions

Imperial Valley, California (Array # 5), October 15, 1979

Imperial Valley, California (Array # 7), October 15, 1979
Northridge, California (Newhall station), January 17, 1994
Northridge, California (Rinaldi station), January 17, 1994
Northridge, California (Sylmar station), January 17, 1994
Landers, California (Lucerene Valley station), June 28, 1992

PGA (g)

PGV (m/s)

0.238
0.132
0.455
0.838
0.828
0.721

0.239
0.192
0.325
0.472
0.493
0.785

0.079
0.06
0.29
0.852
0.377
0.818

0.107
0.05
0.928
1.661
1.175
0.976

0.133
0.122
0.674
0.73
0.746
0.319

0.008
0.008
0.372
0.507
0.243
0.459

Duration (s)

Scale factor

19
17
25
14
40
45

1
1
1
1
1
1

H. Zhang et al. / Engineering Structures 106 (2016) 154165

Shear force /kN

162

250

Table 4
Energy dissipated at bearing P under Northridge ground motion (Rinaldi station,
1994).

HDRB
LRB

200
150
100
50

-0.10

-100

0.05

0.10

0.15

0.20

0.25

HDRB
Direction:
X

Direction:
Y

Direction:
X

Direction:
Y

Ed (kN m)
Equivalent damping
ratio

66.127
0.205

19.682
0.224

24.012
0.360

14.050
0.355

The ratio,n,is equal to 3.

Northridge (Rinaldi station, 1994)
Direction: X

-150
-200

Shear force /kN

Fig. 19. Sheardeformation curves of bearing P (direction: X).

150

HDRB
LRB

100

-0.006

-0.003

Dome with HDRBs

Dome with LRBs

20

Section S1

15
10
5

0
0.000

-0.02

50
0
0.00
-50
-100
-150

0.02

0.04
0.06
0.08
Lateral deformation /m

The ratio,n,is equal to 3.

Northridge (Rinaldi station, 1994)
Direction: Y

LRB and this leads to a larger energy dissipation. Under near-fault

earthquake motions, the peak shear force and horizontal peak
deformation tend to be large as a result of the impact effects of
these motions. However, the unexpectedly large horizontal deformation of a bearing may be unfavorable for the horizontal stability
of a dome, and an appropriate lateral stiffness is needed for a bearing to resist the base shear induced by an earthquake.
In this paper, the energy dissipation of the bearings can be compared based on the above hysteretic curves. The energy dissipation
of a bearing can be described by means of the area enclosed by the
hysteretic curves, as shown in Eq. (13).

u2

u1

F s udu 

n1 
X
ui ui1 
Fs
ui1  ui
2
i1

0.009

0.012

The ratio,n,is equal to 5.

Northridge, Rinaldi station, 1994

Fig. 21. Momentcurvature (Mu) curves of section S1.

Fig. 20. Sheardeformation curves of bearing P (direction: Y).

Ed

0.006

Section curvature /(1/m)

Joint moment /(kN*m)

-0.04

0.003

-5
-10

-0.06

LRB

Lateral deformation /m
Joint moment /(kN*m)

-0.15

0
-0.05 0.00
-50

Energy dissipated at
bearing

13

where Ed is the energy dissipation of a bearing caused by the shear

force, u is the shear displacement, and Fs(u) is the shear force produced by the lateral displacement of the bearing. The energy dissipated is listed in Table 4. Here, in order to evaluate the average
damping ratios of the different types of bearings, the energy dissipation in Table 4 is translated to the energy dissipated in one cycle
of a bearing subjected to a static load, and is approximately
described by using an equivalent damping ratio. Through this
equivalence, it is seen that the equivalent damping ratios of the
two bearings are between 0.2 and 0.4, with the damping of the
LRB apparently higher than that of the HDRB.
6.4.3. Momentcurvature (Mu) demands of end sections of members
The momentcurvature curves of the end sections S1 and S2
(see Fig. 13) of the typical members of the structure subjected to

-0.004

-0.002

15

Dome with HDRBs

Dome with LRBs

10 Section S2

0
0.000
-5

-10

0.002

0.004

0.006

Section curvature /(1/m)

The ratio,n,is equal to 5.
Northridge, Rinaldi station, 1994

Fig. 22. Momentcurvature (Mu) curves of section S2.

the original Northridge motion (Rinaldi station, 1994) are presented in Figs. 21 and 22. Here, the stiffness ratio is set to 5.
For the dome with HDRBs, the maximal end moment and curvature are around 20 kN m and 0.011/m in section S1 and around
10 kN m and 0.005/m in section S2, respectively, while these
demands are around 21 kN m and 0.013/m in section S1 and
around 13 kN m and 0.006/m in section S2 for the dome with LRBs,
respectively. Thus, the values for the dome with LRBs are slightly
larger than those for the dome with HDRBs. This occurs because,
as discussed earlier, the lateral stiffness of the dome with HDRBs
is less than that of the dome with LRBs. As a result, the dome with
HDRBs has more flexibility, and the internal forces of the members
are released and reduced to a certain degree.
As seen in Figs. 21 and 22, on the whole, the moment and curvature demands of the end sections of the members in the dome
with HDRBs are within the elastic range, while the end sections
of the members in the dome with LRBs have unrecoverable residual curvatures. It should be noted that the bending-resistance performance of a member strongly depends on its axial load, which

163

H. Zhang et al. / Engineering Structures 106 (2016) 154165

7. Parametric analysis
The previous research tended to focus on the response, stress,
and stability of domes. While in recent years, there has been interest in the quantification of the damping force in a structure, it has
rarely been reported in the previous literatures for a dome and is
not sufficiently understood by practitioners. In the following sections, the effects of several key parameters on the damping forces
under the above six earthquake ground motions are analyzed and
discussed.
7.1. Effects of ball joint stiffness on viscous damping forces of joints

Mean of maximum vertical damping forces /kN

According to Eq. (8), the ratio of the ball joint stiffness to elastic
member stiffness has a significant effect on the tangent stiffnessproportional coefficient. Here, the effects of different ratios on
the peak damping forces of joints are discussed. The ratios n = 1,
3, and 5 are selected. Figs. 23 and 24 present the results in relation
to the mean values of the vertical peak damping forces of joints

No. of Joint (Section B-B)

2.8

10

12

14

n=1
n=3
n=5

n=1
n=3
n=5

2.4

2.0

Section A-A
Section B-B

1.6

Dome with HDRBs

1.2
0.8
0.4
1

No. of Joint (Section A-A)

Mean of maximum vertical damping forces /kN

Fig. 23. Damping forces with different stiffness ratios (HDRB).

No. of Joint (Section B-B)

0

10

12

14

3.2
n=1
n=3
n=5

2.8

n=1
n=3
n=5

2.4

Section A-A
Section B-B

2.0

Dome with LRBs

1.6
1.2
0.8
0.4
1

No. of Joint (Section A-A)

Fig. 24. Damping forces with different stiffness ratios (LRB).

with different stiffness ratios under the above six near-fault earthquake motions. It can be seen that these damping forces of the
joints are around 0.43 kN. The mean values of all damping forces
in Figs. 23 and 24 are 0.9 kN and 1.0 kN, respectively. The mean
damping force of the dome with LRBs is slightly larger. As a consequence, the deformation of the dome with LRBs, as shown in
Fig. 18, is less than that of the dome with HDRBs. It should be noted
that there is a pronounced increase in the damping forces with a
decrease in the stiffness ratio.
In dynamic equations, when the stiffness-proportional damping
model is used, the damping force is proportional not only to the
stiffness-proportional damping coefficient but also to the stiffness
and velocity demand. According to Eq. (8), a decrease in the stiffness ratio n leads to an increase in the stiffness-proportional
damping coefficient b0 , which finally results in higher damping
forces, indicating that in these domes the damping force strongly
depends on the stiffness-proportional coefficient. Moreover, it
can also be observed that the use of different types of bearings
results in different distributions of the damping forces in the dome.
7.2. Effects of selection of damping models on viscous damping forces
of joints
In this paper, the damping forces of joints are compared
between the proposed explicit stiffness-proportional damping
model in Section 3.2 and the conventional Rayleigh damping
model. Here, the formulation of Rayleigh damping used was modified by Hall [16], with the goal of achieving a near-constant value
of damping for all modes, and Hall considered that this was consistent with much of the field data. For a given single-layer latticed
dome, when the viscous damping model is used, the damping
ratios are fairly constant due to its small difference in first several
natural frequencies [43]. In the damping model, the parameter, R,
that defines the range the damping ratios is set to 1.3. Figs. 25
and 26 show the mean values of the vertical peak damping forces
of joints with different damping models under the above six earthquake motions. As shown in these figures, the Rayleigh damping
model leads to higher viscous damping forces. Compared to the
Rayleigh damping model, where the damping matrix consists of
the mass part and stiffness part, the updated tangent stiffnessproportional damping model leads to lower damping forces
because of the elimination of the mass-proportional damping component. A mean difference of about 60% in the damping forces
between two damping models is observed.
In a dome, because damping forces that are too high can greatly
inhibit the real peak response of the structure, the use of viscous

Mean of maximum vertical damping forces /kN

decreases with the increase of the axial load. However, a member

in a dome subjected to earthquake ground motions tends to be
subjected to a varying axial force instead of a constant axial force.
As a consequence, the members in a dome may present complex
hysteretic properties.

No. of Joint (Section B-B)

1.5

10

12

14

Rayleigh damping
Stiffness-proportional damping
Rayleigh damping
Stiffness-proportional damping

1.4
1.3
1.2

Section A-A
Section B-B

1.1

Dome with HDRBs

n=3

1.0
0.9
0.8
0.7
0.6
0.5
1

No. of Joint (Section A-A)

Fig. 25. Damping forces with different damping models (HDRB).

H. Zhang et al. / Engineering Structures 106 (2016) 154165

Mean of maximum vertical damping forces /kN

164

No. of Joint (Section B-B)

2.0

10

12

14

Rayleigh damping
Stiffness-proportional damping
Rayleigh damping
Stiffness-proportional damping

1.8
1.6
1.4

7.3. Effects of roof loads on viscous damping forces of joints

Roof loads of 90, 135 and 180 kg/m2 are used to illustrate the
effects of roof loads on the viscous damping forces of joints. Figs. 27
and 28 show the mean values of the vertical peak viscous damping
forces of joints under different roof loads. It is seen that the
increase in these damping forces with an increase in the roof load
is remarkable. Because the natural frequencies decrease of the
dome with an increase in the roof load, according to the equation
b0 = 2n2/x, this leads to a larger stiffness-proportional damping
coefficient. As previously analyzed, the viscous damping forces
strongly depend on the stiffness-proportional coefficient in these
domes. Thus, a larger roof load will have a higher damping force.

Section A-A
Section B-B

1.2
1.0
Dome with LRBs
n=3

0.8
0.6
1

No. of Joint (Section A-A)

Mean of maximum vertical damping forces /kN

Fig. 26. Damping forces with different damping models (LRB).

No. of Joint (Section B-B)

0

8
2

0.8
0.7

10

12

90kg/m
2
135kg/m
2
180kg/m

Section A-A
Section B-B

0.6

14

90kg/m
2
135kg/m
2
180kg/m

HDRB
n=3

0.5
0.4
0.3
0.2

No. of Joint (Section A-A)

Mean of maximum vertical damping forces /kN

Fig. 27. Damping forces under different roof loads (HDRB).

No. of Joint (Section B-B)

0

8
2

1.0

12

14

90kg/m
2
135kg/m
2
180kg/m

1.2

10
90kg/m
2
135kg/m
2
180kg/m

LRB
n=3

Section A-A
Section B-B

0.8
0.6
0.4
0.2
1

for the dynamic analysis of a dome subjected to earthquake ground

motions.

8. Conclusions
In this paper, some additional important features of damping
were considered to analyze the dynamic demands of a typical
dome subjected to earthquake ground motions. According to the
above results and discussions, the following conclusions can be
obtained.
(1) The typical models of a steel material using elasticplastic
(bilinear) or other empirically derived models for the
stressstrain behavior are unable to capture the material
damping within the elastic range. Unlike other typical material constitutive models, the RO material model is able to
capture the material damping. However, the material damping has been ignored or described by means of a frequencydependent damping model at a structure level in the previous dynamic analyses. This is not reasonable because the
material damping depends on material stress instead of frequency and is not uniformly distributed in a structure.
(2) The damping mechanisms within a real structure are still
not fully understood because of their complex nature. At
present no acceptable mathematical model for estimating
the damping forces has been available. Despite the urgent
need for damping estimates in design, there is still no unique
technique for accurately estimating damping forces. As
noted earlier, a good alternative is to use a hysteretic-type
model to explicitly take into account the damping effects
in a structure, instead of the excessive use of viscous
damping.
(3) In this paper, through the analyses for some damping
sources in a dome with welded joints, the material damping
and damping at bearings were directly modeled, and the
small amount of damping from welded joints was modeled
by means of the stiffness-proportional damping model with
a time variant stiffness property proposed in Section 3.2 to
reduce the viscous damping forces as much as possible.
These results will help engineers to get a better understanding of damping in domes in engineering design.

No. of Joint (Section A-A)

Fig. 28. Damping forces under different roof loads (LRB).

damping models that generate unrealistically high damping forces,

such as Rayleigh damping, can result in an underestimation of the
peak displacement demands, collapse potential, and peak strength
demands of the structure. Therefore, some energy sources should
be carefully considered when Rayleigh damping model is used

Acknowledgements
The authors would like to gratefully acknowledge the financial
support provided by the National Natural Science Foundation of
China (Grant Nos.: 51108301 and 51308386), the National Science
Fund for Distinguished Young Scholars (Grant No. 51525803) and
the Research Fund for the Doctoral Program of Higher Education
of China (Grant No.: 20130032110044).

H. Zhang et al. / Engineering Structures 106 (2016) 154165

References
[1] Escandar A et al. Mechanical properties and fuzzy modeling of high-damping
rubber with thermal effects. J Mater Civ Eng 2007;19(5):42836.
[2] Beards CF. Structural vibration: analysis and damping. Oxford: ButterworthHeinemann; 1996.
[3] Sheen RL. Experimental measurement of material damping for space
structures in simulated zero-G (Rept no. AFIT/CI/NR-83-84T). OH, U.S.: Air
Force Inst. of Tech., Wright-Patterson AFB; 1984.
[4] Kareem A, Gurley K. Damping in structures: its evaluation and treatment of
uncertainty. J Wind Eng Ind Aerodyn 1996;59(23):13157.
[5] Charney FA. Unintended consequences of modeling damping in structures. J
Struct Eng 2008;134(4):58192.
[6] Adhikari S. Damping modelling using generalized proportional damping. J
Sound Vib 2006;293(12):15670.
[7] Adhikari S. Dynamics of nonviscously damped linear systems. J Eng Mech
2002;128(3):32839.
[8] Woodhouse J. Linear damping models for structural vibration. J Sound Vib
1998;215(3):54769.
[9] Val DV, Segal F. Effect of damping model on pre-yielding earthquake response
of structures. Eng Struct 2005;27(14):196880.
[10] Deierlein GG, Reinhorn AM, Willford MR. Nonlinear structural analysis for
seismic design: a guide for practicing engineers (NEHRP Seismic Design
Technical Brief No, 4). National Institute of Standards and Technology, U.S.
Department of Commerce; 2010.
[11] Tatemichi I, Hatato T, Yoshimichi A, Fujiwara S. Vibration tests on full-size
suspen-dome structure. Int J Space Struct 1997;12(34):21724.
[12] Chopra AK. Dynamics of structures. New Jersey: Prentice Hall; 1995.
[13] Clough RW, Penzien J. Dynamics of structures. New York: McGraw-Hill; 1993.
[14] Zhang HD, Wang YF. Energy-based numerical evaluation for seismic
performance of a high-rise steel building. Steel Compos Struct 2012;13
(6):50119.
[15] Zhang HD, Han QH. A numerical investigation of seismic performance of large
span single-layer latticed domes with semi-rigid joints. Struct Eng Mech
2013;48(1):5775.
[16] Hall JF. Problems encountered from the use (or misuse) of Rayleigh damping.
Earthquake Eng Struct Dynam 2006;35(5):52545.
[17] Zareian F, Medina RA. A practical method for proper modeling of structural
damping in inelastic plane structural systems. Comput Struct 2010;88(1
2):4553.
[18] Petrini L, Maggi C, Priestley MN, Calvi GM. Experimental verification of viscous
damping modeling for inelastic time history analyzes. J Earthquake Eng
2008;12(S1):12545.
[19] Jehel P, Lger P, Ibrahimbegovic A. Initial versus tangent stiffness-based
Rayleigh damping in inelastic time history seismic analyses. Earthquake Eng
Struct Dynam 2013;43(3):46784.
[20] Bandstra JP. Comparison of equivalent viscous damping and nonlinear
damping in discrete and continuous vibrating systems. J Vib, Acoust Stress
Reliab Des 1983;105(3):38292.
[21] Baburaj V, Matsuzaki Y. A study on the material damping of thin angle-ply
laminated plates. J Sound Vib 1994;172(3):4159.
[22] Ramberg W, Osgood WR. Description of stressstrain curves by three
parameters. Technical note No. 902. Washington, DC: National advisory
Committee for Aeronautics; 1943.

165

[23] Rasmussen Kim JR. Full-range stressstrain curves for stainless steel alloys. J
Constr Steel Res 2003;59(1):4761.
[24] Lazan BJ. Damping of materials and members in structural
mechanics. London: Pergamon Press; 1968.
[25] Groper M. Microslip and macroslip in bolted joints. Exp Mech 1985;25
(2):1714.
[26] Beards CF. The damping of structural vibration by controlled interfacial slip in
joints. J Vib, Acoust Stress Reliab Des 1983;105(3):36973.
[27] Guo T, Song L, Zhang G. Numerical simulation of the seismic behavior of selfcentering steel beam-column connections with bottom flange friction devices.
Earthquake Eng Eng Vib 2011;10(2):22938.
[28] Bingham JG, Crookston JR, Dutson JD, Ferney BD, Ferney GD, Rowsell EA. The
Joint Damping Experiment (JDX), vol. 4781. National Aeronautics and Space
Administration, Langley Research Center; 1997.
[29] Kim YJ, Lee YH, Kim H. Bending test of welded joints for single-layer latticed
domes. Int J Steel Struct 2008;8(4):35767.
[30] Buchanan AH, Bull D, Dhakal R, MacRae G, Palermo A, Pampanin S. Base
isolation and damage-resistant technologies for improved seismic
performance of buildings. Christchurch, New Zealand: Department of Civil
and Natural Resources Engineering, University of Canterbury; 2011.
[31] Bhuiyan AR, Alam R, Haque N. Seismic performance assessment of a
continuous highway bridge seismically isolated by lead rubber bearings.
Asian Trans Eng 2012;2(3):6271.
[32] Kato S, Nakazawa S, Uchikoshi M, Ueki T, Osugi F. Earthquake response of
domes implemented by hysteresis dampers for earthquake isolation. In:
Proceedings of IASS, Sydney, Australia; 1998.
[33] Kikuchi M, Aiken I. An analytical hysteresis model for elastomeric seismic
isolation bearings. Earthquake Eng Struct Dynam 1997;26(2):21531.
[34] Ryan KL, Kelly JM, Chopra AK. Formulation and implementation of a lead
rubber bearing model including material and geometric nonlinearities. In:
Structures congress (ASCE); 2006.
[35] Yamamoto M, Minewaki S, Higashino M, Hamaguchi H, Kyuke H, Sone T, et al.
Performance tests of full size rubber bearings for isolated super high-rise
buildings. In: JSSI 15th anniversary international symposium on seismic
response controlled buildings for sustainable society, Tokyo; 2009.
[36] ISO/TC 45/SC 4. ISO 22762-3: elastomeric seismic-protection isolators-Part 3:
application for buildings-specifications. International Organization for
Standardization; 2005.
[37] Feng D, Chen C, Liu W, Tanaka K. A performance test study on Chinese G4 lead
rubber bearings. In: Proceedings 13th world conference on earthquake
engineering, Vancouver, British Columbia, Canada; 2004.
[38] Heaton TH, Hall JF, Wald DJ, Halling MW. Response of high-rise and baseisolated buildings to a hypothetical Mw 7.0 blind thrust earthquake. Science
1995;267(5195):20611.
[39] Hall JF, Heaton TH, Halling MW, Wald DJ. Near-source ground motion and its
effects on flexible buildings. Earthquake Spectra 1995;11(4):569605.
[40] Jangid RS. Optimum leadrubber isolation bearings for near-fault motions. Eng
Struct 2007;29(10):250313.
[41] Yamada S, Takeuchi A, Tada Y, Tsutsumi K. Imperfection-sensitive overall
buckling of single-layer lattice domes. J Eng Mech 2001;127(4):3826.
[42] Han QH, Xu Y, et al. Failure mechanism of steel arch trusses: shaking table
testing and FEM analysis. Eng Struct 2015;82(1):18698.
[43] Zhang HD, Wang YF, Han QH. Nonlinear material loss factors of single-layer
latticed domes subjected to earthquake ground motions. J Struct Eng 2015;141
(7):0401418.