China-Japan Joint Seminar on Steel and Composite Bridges 65

Performance-Based Control Design of Steel Bridge Structures by

Introducing Energy-Dissipated Members

H.B. Ge, Z.Y. Chen

Dept. of Civil Eng., Nagoya University, Nagoya 464-8603, Japan
T. Usami
Dept. of Civil Eng., Meijo University, Nagoya, 468-8502, Japan

ABSTRACT: This paper aims to investigate the efficiency of hysteretic dampers incorpo-
rated in thin-walled steel bridge structures against major earthquakes. Hysteretic damper con-
cerned here is buckling-restrained brace (BRB), which dissipates energy through metallic
yield under axial loading. Inelastic dynamic analysis has been performed. It is found that such
a energy-dissipated member is effective on earthquake hazard mitigation.

1 INTRODUCTION

In recent years, various passive control technologies have been developed and used success-
fully into building structures to suppress structural vibrations induced by earthquakes. Among
them, the hysteretic damper, whose mechanism is through material yield to absorb earth-
quake-induced energy as well as increase effective damping, is the most popular one to be
employed due to its reliability and low cost. Although extensive experimental and analytical
researches have been conducted on building structures with such hysteretic dampers, the re-
searches on bridge structures remain infancy (Zahrai and Bruneau 1999; Usami et al. 2005).
Furthermore, efficiency of hysteretic dampers on seismic response reduction has been vali-
dated in comparison with the moment resisting steel structures (Yamaguchi et al. 2001) or
structures with conventional steel braces (Tena-Colunga and Vergara 1997), however, few
works have addressed the efficiency between various hysteretic dampers especially when they
are incorporated in bridge structures.

Three structural forms commonly used in bridge structures are employed as illustrative exam-
ples. They are one-level and multi-level bridge piers of portal frame type and arch bridges.
Fig. 1 shows an example of one level portal frame installed with a type of hysteretic dampers,
i.e., buckling-restrained brace (BRB). The main objective of this paper is to investigate the ef-
ficiency of the BRB member in aforementioned bridge structures through numerical experi-
ments. This investigation may provide useful messages for design of hysteretic dampers and
their further laboratory experimental research.

BRB

Fig. 1 Portal frame with hysteretic damping devices

66 China-Japan Joint Seminar on Steel and Composite Bridges

2 FUNDAMENTAL PRINCIPLES

Sketches of physical models for the controlled structure installed with BRB are shown in Fig.
2. As is seen from Fig. 2, it is assumed that the BRB member itself is an entire damping de-
vice. Consequently, basic properties of the damping device have the same meaning with those
of the BRB member (i.e., Fy,d = Fy,BRB, Kd = KBRB). The BRB member is then connected as a
whole with the main structure in the parallel manner. It is clear that the yield displacement of
the BRB member must be smaller than that of the main structure to ensure the BRB member
yields prior to the main structure during an earthquake. Namely, the design conditions of δy,d
< δy,f and δy,d = δy,fd must be satisfied.
M

Fy,BRB Fy,d
Fy,f Fy,d Fy,fd
KBRB Kd Kfd
Kf Kd

V V Damped frame

Fy,fd Damping device

Fy,BRB Fy,BRB
BRB Fy,f
Damping device Main frame

δy,BRB ∆ δy,fd δy, f ∆

δ y,d < δ y, f
δ y , fd = δ y , d
(a) BRB damping device (b) Damped frame
Fig. 2 Idealized SDOF system installed with BRB

To compare seismic demands quantitatively, the ratio of the damper’s yield strength to the
main structure’s lateral yield resistance, referred to as the strength ratio αF, and the ratio of the
damper’s initial stiffness to the main structure’s lateral stiffness, referred to as the stiffness ra-
tio αK, are generally used as design criteria (Ye and Ouyang 2000; Inoue and Kuwahara 1998;
Yamaguchi and Ashraf 2003).

3 DESIGN AND MODELING OF DAMPING DEVICES

3.1 Basic properties of buckling-restrained braces

To date, many types of BRB sections are used. Among them, a kind of BRB with a flat core
plate developed by Nagoya University (Kato et al. 2003; Usami et al. 2003) is used in this
study, whose cross section is shown in Fig. 3. Similar to the other BRBs, axial deformation of
such BRB is allowed only at the core plate, whose severe buckling is eliminated by the exte-
rior restraining members. The axial yield strength of BRB, Py,BRB, is given by
Py , BRB = σ y , BRB × ABRB (1)
where σy,BRB = tensile yield stress of steel material and ABRB = section area of the core plate.
 China-Japan Joint Seminar on Steel and Composite Bridges 67

δ BRB

T-shaped
restraining members FBRB
K BRB
l BRB
PBRB
PBRB
ABRB
Bolt Bolt
L
Core plate Unbonding material

Fig. 3 Buckling-restrained braces (BRB) Fig. 4 Schematic diagram for BRB deformation

σ
σy
E/60

E 2σy
2σ ε
Truss element

Fig. 5 BRB model Fig. 6 Hysteretic model with kinematic hardening rule

The geometric parameters and basic properties of BRB are illustrated in Fig. 4. Relationship
between these parameters can be expressed as follows,
2 2
2l BRB
F y , BRB = K BRB ⋅ δ y , BRB , δ y , BRB = σ y , BRB ⋅ , K BRB = ABRB ⋅ EL3
(2)
EL 2l BRB
where Fy,BRB, δy,BRB, and KBRB = the lateral yield strength, displacement, and elastic stiffness of
a pair of BRBs, respectively; lBRB = length of BRB; L = the structure span.

3.2 Hysteretic model and FE modeling of BRB

Fig. 5 shows the finite element model of a portal frame with BRB, referred to as BRB model
later. BRB is directly modeled by the truss element, only carrying axial tension and compres-
sion without local bucking. In this study, the core plates of the BRB employed, as shown in
Fig. 6, are made of SS400 steel. For such a BRB member, a bilinear stress-strain relationship
with kinematic hardening rule is suggested (Kato et al. 2003; Usami et al. 2003).

3.3 Design order and commentary for BRB in seismic design process

Knowing the predefined strength ratio αF and the stiffness ratio αK as well as corresponding
strength and stiffness of the main structure, Fy,BRB and KBRB can be determined in the same
manner as design of SPD devices. Then, the basic properties of BRB such as σy,BRB and ABRB
can be easily established from Eq. (2). Note that such design process will cause variation of
σy,BRB. In practical design, steel material is usually determined at the beginning, followed by
the establishment of δy,BRB. In other words, ABRB becomes the single adjustable parameter in
design of BRB damping devices.
68 China-Japan Joint Seminar on Steel and Composite Bridges

Lumped mass M/2 Lumped mass M/2 Dc 0 .5 Dc

S3 S4
Db

H = 12m
S2 S5
FA
0 .5 Db

t
(d)
S1 le S6
ts
t
L= 12m Bc

Db
・
・・Strengthened parts　 bs
ts
bs
・
・・Critical member segments Dc Bb

(a) (b) (c)

Fig. 7 One-level steel bridge piers of portal frame type: (a) layout; (b) pier section;
(c) girder section; and (d) details at strengthened corner

4 DESIGN AND ANALYTICAL MODELING OF STEEL BRIDGE STRUCTURES

4.1 Design and analytical modeling of one-level bridge piers

The analytical model to be investigated is one-level steel bridge piers of portal frame type,
which are widely used in Japan to support elevated expressways. The general layout and finite
element model are illustrated in Fig. 7(a). Since this type of portal frames is commonly heavy
burdened, the plates at the pier-girder connection parts should be strengthened by doubling
the plate thickness in order to avoid shear failure. The detailed size of strengthened parts is
shown in Fig. 7(d). The portal frame is designed in accordance with the Seismic Coefficient
Method (JRA 2002a; 2002b), assuming regional class A and ground class II. Detailed infor-
mation about this frame is summarized in the column titled by F1 in Table 1. In Table 1, the
basic properties are determined with a pushover analysis as presented in a previous study
(Usami et al. 2001), being somewhat analogous to those provided by Eurocode 8 (2003),
FEMA 356 (2000), and Chopra and Goel (2002).

Two-dimensional Timoshenko beam element of type B21 provided in the ABAQUS element
library is used to model the piers and girder, accounting for shear deformation. Each pier and
girder of the main frame is divided into 20 elements: 5 elements for each critical member seg-
ment, 2 elements for strengthened parts, and remains for the other parts. Here, critical member
segments are denoted by the dotted sections at the pier bases and adjacent to the rigid corners,
as shown in Fig. 7(a).

In order to trace the material cyclic behavior accurately, the modified two-surface model
(Shen et al. 1995) is employed. Rayleigh damping, which is usually utilized in dynamic
analysis, consists of mass proportional damping and stiffness proportional damping. Here,
mass proportional damping of 5 per cent is used whereas inherent stiffness proportional
damping is set as zero since it is neglectable if compared to the significant equivalent viscous
damping due to yielding of hysteretic damper.
 China-Japan Joint Seminar on Steel and Composite Bridges 69

4.2 Design and analytical modeling of multi-level bridge piers

Since highway bridge system is commonly composed of thin-walled members and a popular
format is low-rise (1~3 level) framing type, two-level and three-level portal frames are consid-
ered here. It is assumed that the two-level and three-level portal frames considered here have
girders and piers of the same section as the preceding one-level frame. Thus, each level has the
same yield inter-level force and level stiffness. The basic properties of multi-level portal frames,
denoted by F2 and F3, are listed in Table 1. The masses subjected on the top of columns are
the same in each layer.

4.3 Design and FE modeling of arch bridge

An upper-deck arch bridge mainly composed of reinforced concrete (RC) deck slab, steel gird-
ers and single span steel arch ribs is modeled. A total length of the bridge is 173.0 m, and a
two end-hinged steel arch has a span of 114.0 m with a rise at the crown of 16.87 m, giving a
rise-span ratio of 1/6.76. Fig. 8 shows the established FE model, where 3D Timoshenko beam
elements of type B31 provided in the ABAQUS program are employed to model the deck slab,
girders, columns (vertical members), arch ribs and transverse bracings while 3D truss elements
of type T3D2 for diagonals. In this analytical model, there are totally 884 elements with 462
nodes. More information of this bridge is referred to literatures by Usami (2003, 2005).

Table 1 Structural parameters and basic properties of the main frames

Main frame

Items F1 F2 F3
Steel grade SM490 SM490 SM490
Height, H (m) 12 12×2 12×3
Length, L (m) 12 12 12
Total deck mass, M (ton) 2042 2034 2895
in
Initial yield displacement, δ y,f (m) 0.078 0.134 0.178
in
Initial yield strength, V y,f (kN) 6758 5938 5129
Ultimate displacement, δu,f (m) 0.418 0.510 0.549
Ultimate strength, Vu,f (kN) 11836 10884 9557
Bi yield displacement, δbiy,f (m) 0.112 --- ---
Bi yield strength, Vbiy,f (kN) 9645 --- ---
Period, Tf (sec) 0.97 1.23 1.77

Pier Girder

Items Value Items Value

Width, Bc (mm) 2000 Width, Bb (mm) 2000
Depth, Dc (mm) 1000 Depth, Db (mm) 1000
Thickness, t (mm) 32 Thickness, t (mm) 32
Flange slenderness, Rfc 0.34 Flange slenderness, Rfb 0.34
Web slenderness, Rwc 0.34 Web slenderness, Rwc 0.78
Aspect ratio of flange, α fc 1.0 Aspect ratio of flange, α fb 0.5
Aspect ratio of web, α wc 0.5 Aspect ratio of web, α wb 0.83
70 China-Japan Joint Seminar on Steel and Composite Bridges

No. of flange sub-panels, nfc 2 No. of flange sub-panels, nfb 4

No. of web sub-panels, nwc 4 No. of web sub-panels, nwb 1
− −
Stiffener slenderness, λ sc 0.89 Stiffener slenderness, λ sb 0.72
Stiffener width, bs (mm) 140 Stiffener width, bs (mm) 140
Stiffener thickness, ts (mm) 32 Stiffener thickness, ts (mm) 32

Fig. 8 FE model of arch bridge

5 EFFICIENCY OF DAMPERS INSTALLED IN BRIDGE STRUCTURES

5.1 Efficiency of Dampers Installed in One-level Bridge Piers

300
200
100
Top diplacement, δ (mm)

0
-100 BRB BRB BRB
-200 Bare Bare Bare
-300
0 10 20 30 0 10 20 30 0 10 20 30
Time (sec)
(a) Top displacement responses
y

1
σ / σy

-1
BRB BRB BRB
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
ε / εy

(b) Stress-strain responses of dampers

αF = 0.5
αF = 1.0
αF = 1.5

BRB

0 10 20 30
Time (sec)
(c) Energy dissipated by dampers
Fig. 9 Responses from time-history analysis
 China-Japan Joint Seminar on Steel and Composite Bridges 71

In this study, the time-history analysis results for one-level bridge piers subjected to JRT-EW-
M accelerogram are illustrated in Fig. 9, according to the strength ratio, αF.

Seen from Fig. 9(a), the top displacement demands, δr,max, is sharply decreased from 2.3δy
(moderate damage) in Bare model to below 1.7δy (light damage) in BRB model, indicating
merits of installation of damping devices. Referring to Fig. 9(b), in the case of αF = 0.5, the
normalized axial deformation of BRB, εmax / εy)BRB, is 28.2. As αF is increased from 0.5 to 1.0, a
rapid decrease of deformation of dampers is observed, indicating a decrease in damage to the
main frame. Meanwhile, increasing αF from 1.0 to 1.5, the peak responses are improved a little.
When αF equals 1.5, the main frame remains in elastic range. This fact indicates that the main
frames can remain intact even under strong earthquake. Energies dissipated by the dampers are
shown in Fig. 9(c), and the stable energy dissipating capacity of the dampers is confirmed in all
the cases.

5.2 Efficiency of Dampers Installed in Multi-level Bridge Piers

Dynamic analysis results of MDOF systems are summarized in Table 2. Major findings ob-
tained from SDOF systems can be extended to MDOF systems. Distinguished from SDOF sys-
tem, MDOF systems have a particular problem of distribution of damping devices into the level.
Referring to the plastic energy dissipated by the dampers at each level as shown in Table 2, in
the case of F2-BRB-03, the energy dissipated averagely by damper in the first level and the
second level equal to 0.470 and 0.529, respectively. This case has shown suitable distribution
of stiffness and strength of damping devices in the MDOF frames. However, the case of F2-
BRB-05 exhibits the highly unequal distribution of plastic energy. The damper in the first level
dissipates 0.225 of the total plastic energy while in the second level it is 0.739. Although the
strength ratio increases from 0.3 to 0.5, F2-05 series produces larger strains and drifts than F2-
03 series. Non-uniform distribution of dissipated energy can be found in F3 series as well. The
optimum distribution of damping devices into bridge structures will be one of the important is-
sues in the future research.

Table 2 Summary of design and analysis results for MDOF frames with or without dampers
Time–history analysis results
Model ε a ) max δr,max Vb,max ε max  Ef E d ,1 Ed ,2 E d ,3 Ep
εy (m) 3
(×10 kN) ε y  BRB Ep Ep Ep Ep (×106J)

F2-Bare 27.8 0.531 11.1 – 1.000 – – – 26.7

F2-BRB-03 1.11 0.154 16.3 5.2 0.001 0.470 0.529 – 8.59

F2-BRB-05 1.65 0.169 23.1 4.6 0.006 0.255 0.739 – 6.67

F3-Bare 21.5 0.699 10.7 – 1.000 – – – 18.7

F3-BRB-03 1.38 0.289 19.4 9.4 0.002 0.193 0.484 0.321 11.6

Note: F2 and F3 stand for 2 and 3 DOF frames, respectively. BRB represents frames with BRB. The numbers (03 and 05)
represent the values of the strength ratio αF at each story. Ep represents the total plastic energy; Ef and Ed,i represent plastic

energy in the main frame and in the dampers of each story, respectively, where i = 1, 2, and 3.
72 China-Japan Joint Seminar on Steel and Composite Bridges

Side pier base (-29εy)

A2
端柱基部ひずみ（-29.0）

A1
G2

G1

Diagonal (-2.57ε
下横構ひずみ（ -2.57） y)
P2
P1 アーチリブひずみ（-2.51）
Arch rib (-2.51εy)

(a) Original model

Side pier base (-1.38εy)

A2
端柱基部ひずみ（-1.38）

A1
G2

G1

ブレース材ひずみ（
BRB (-7.85εy) -7.85）
P2
P1

(a) Upgrading model

Fig.10 Yielded members in the original and upgrading models

5.3 Efficiency of dampers installed in arch bridges

By far, only the seismic upgrading performance of arch bridge installed with BRB has been
investigated. Two schemes of replacing some diagonal braces with BRB are considered but
only one upgrading model is shown in Fig. 10(b), where bold lines represent the location of
BRB to be replaced. Details are referred to the literatures (Usami ed. 2003, 2005). Time-
history analyses are performed. Some findings are drawn as follows:

(1) Natural period of the modes in the longitudinal direction of the two Upgrading Models are
almost the same as the Original Model, while a little elongated for the periods of the trans-
verse modes, due to the fact that stiffness of those replaced members is lowered since section
area is reduced to 1/3~2/3 of original members.

(2) The maximum axial force demand at the side pier base is decreased from about Py in the
Original Model to lower than 0.5Py in the two Upgrading Models.

(3) In Upgrading Model 1, placement of BRB members only at side pier location does not de-
creases strain demand at the arch rib base, while in Upgrading Model 2, additional placement
of BRB members in arch ribs decreases the strain demand at the arch rib base to the elastic
range.

(4) In the two Upgrading Models, the maximum strain demand in BRB members is below the
available capacity value and stable behavior of BRBs is ensured.

6 SUMMARY AND CONCLUSIONS

This paper presented an investigation on the seismic upgrading efficiency of the buckling-
restrained braces (BRB) under major earthquakes. Three types of bridge structures are used as
illustrative examples. 2D and 3D finite element models are established and dynamic nonlinear
 China-Japan Joint Seminar on Steel and Composite Bridges 73

time-history analyses are performed. It has been shown that BRB members are effective on
dissipating earthquake-induced energy and eliminating or limiting damage to the main struc-
tures.

7 REFERENCES

[1] ABAQUS/Analysis user’s manual––version 6.4. (2003). ABAQUS, Inc., Pawtucket, R.I.
[2] Chen, Z. Y., Ge, H. B. and Usami, T. (2006) “Hysteretic model of stiffened shear panel dampers.”
Journal of Structural Engineering, ASCE, 132(3), 478–483.
[3] Chopra, A. K. and Goel, R. K. (2002) “A modal pushover analysis procedure for estimating seismic
demands for buildings.” Earthquake Engineering and Structural Dynamics 31, 561–582.
[4] Eurocode 8. (2003). British Standards Institution, Design of structures for earthquake resistance. Gen-
eral rules, seismic actions and rules for buildings, London.
[5] FEMA 356. (2000). Federal Emergency Management Agency, Prestandard and commentary for the
seismic rehabilitation of buildings, Washington DC.
[6] Inoue, K. and Kuwahara, S. (1998) “Optimum strength ratio of hysteretic damper,” Earthquake Engi-
neering and Structural Dynamics, 27, 577–588.
[7] JRA. (2002a). Design Specifications of Highway Bridges-Part V Seismic Design, Japan Road Asso-
ciation, Tokyo, Japan (in Japanese).
[8] JRA. (2002b). Design Specifications of Highway Bridges-Part II Steel Bridges, Japan Road Associa-
tion, Tokyo, Japan (in Japanese).
[9] JSSC. (1998). “Seismic response behaviour and design method for frame structures with hysteretic
dampers.” Japanese Society of Steel Construction Tokyo, Japan (in Japanese).
[10] Kato, M., Usami, T., Kasai, A. and Chusilp, P. (2003). “An experimental study on cyclic elasto-plastic
behavior of buckling restrained brace members,” Proceedings of the 6th Symposium on Ductility De-
sign Method for Bridges, Ductility Design Subcommittee, Earthquake Engineering Committee, JSCE,
Tokyo, Japan, pp. 345–350.
[11] Shen, C., Mamaghani, I. H. P., Mizuno, E., and Usami, T. (1995). “Cyclic behavior of structural steels.
II: Theory.” J. Eng. Mech., ASCE, 121(11), 1165–1172.
[12] Tanaka, K. (1998) “Design methodology for seismic response control of buildings using shear panel
dampers of ultra low yield strength steel.” Japan.
[13] Tena-Colunga, A. and Vergara, A. (1997) “Comparative study on the seismic retrofit of a mid-rise
steel building: steel bracing vs energy dissipation.” Earthquake Engineering and Structural Dynamics
26, 637–655.
[14] Usami, T. ed. (2003). Seismic Performance-Based Verification Procedures and Upgrading Measures
for Civil Engineering Steel Structures, Subcommittee for Seismic Design of Steel Bridges, Task
Committee of Performance-Based Seismic Design Methods for Steel Bridges, Tokyo.
[15] Usami, T., Kasai, A. and Kato, M. (2003). “Behavior of buckling-restrained brace members.” Behav-
iour of Steel Structures in Seismic Areas, STESSA 2003, Naples, Italy, 211–216.
[16] Usami, T., Lu, Z. H. and Ge, H. B. (2005). “A seismic upgrading method for steel arch bridges using
buckling-restrained braces.” Earthquake Engineering and Structural Dynamics 34, 471–496.
[17] Usami, T., Zheng, Y. and Ge, H. B. (2001). “Seismic design method for thin-walled steel frame struc-
tures.” Journal of Structural Engineering, ASCE, 127, 137–144.
[18] Yamaguchi, H. and Ashraf, E. A. (2003) “Effect of earthquake energy input characteristics on hyster-
etic damper efficiency.” Earthquake Engineering and Structural Dynamics, 32, 827–843.
[19] Yamaguchi, M., Yamada, S., Maeda, Y., Ogihara, M., Takeuchi, T., Narikawa, M., Nakashima, M.
and Wada, A. (2001). “Evaluation of seismic performance of partial frames using the shaking table
test-Seismic performance of moment resisting steel frame with damper: part 2,” Journal of Structural
and Construction Engineering, AIJ, 547, 153–160.
[20] Ye, L. P. and Ouyang, Y. F. (2000). “Dual seismic structure system and its parametric analysis.” En-
gineering Mechanics, 17 (2), 23–29.
[21] Zahrai, S. M. and Bruneau, M. (1999). “Ductile end-diaphragms for the seismic retrofit of slab-on-
girder steel bridges.” Journal of Structural Engineering, ASCE, 125(1), 71–80.