JFE TECHNICAL REPORT

No. 14 (Dec. 2009)

Seismic Response Control for High-Rise Buildings

Using Energy-Dissipation Devices†
KAMURA Hisaya*1 NANBA Takayuki*2 OKI Koji*3 FUNABA Taku*4

Abstract: examples of the application of JFE vibration dampers to

Application of energy dissipation devices is reason- high-rise buildings.
able and cost effective to maintain main structural mem-
bers in elastic state for high-rise buildings. This paper 2. Vibration Dampers and
discusses required energy dissipation performance for Structural Design of High-Rise Buildings
the longperiod ground motion on M7 class earthquake
2.1 Recent Trend in
and the ability of JFE hysteretic energy dissipation
Structural Design of High-Rise Buildings
devices.
The structural design of a high-rise building reduces
1. Introduction the plasticization and input energy during an earthquake
by assigning a relatively large elastic limit to the main
Seismologists have recently warned of the strong frame and lengthening the natural period at a safety
likelihood that Japan will suffer an M7-class large earth- limit, respectively, by means of use of relatively smaller
quake at some point in the next three decades. If this section. Through these steps, we can perform exami-
happens, it will be particularly important to maintain the nations in pursuit of both economical rationality and
building functions of high-rise buildings after the quake1). seismic safety4). To meet these conditions, architects
JFE Steel has reached the commercial stage in the tend to rely on high-tensile steel materials. This can be
development of extra-mild steels such as an ultralow- disadvantageous, however, as the decreased stiffness
yield-point steel (JFE-LY100) and low-yield-point steel of frames made from high-tensile steel materials tends
(JFE-LY225) for hysteretic dampers, as well as three to compromise occupant comfort during strong winds.
types of vibration dampers: a buckling-restraint brace Vibration damping structures using energy-dissipation
type, a wall-panel type, and a stud-panel type2). We devices are often adopted to meet safety requirements
have also developed hysteretic and visco-elastic hybrid during earthquakes while improving economic efficiency
damper, though not yet to the point of commercializa- and occupant comfort during strong winds.
tion. This paper describes important points to keep in According to the results of hearings at major design
mind in the structural design of vibration damping struc- offices and structural design departments of general con-
tures applied to recent high-rise buildings, and outlines tractors, energy-dissipation devices of various kinds are
the structural performance of the vibration dampers now installed in almost all high-rise, steel-frame build-
developed at JFE Steel. We also evaluate the perfor- ings of 60 m or more in height, and hysteretic dampers
mance of vibration dampers installed in high-rise build- are rated as the best energy-dissipation devices in terms
ings against long-period earthquake motions3), and give of cost performance. Steel materials such as low-yield-

†
Originally published in JFE GIHO No. 21 (Aug. 2008), p. 31–41

*1 *3
Dr. Eng., Staff Deputy Manager,
Senior Researcher General Manager, Construction Engineering Services Dept.,
Civil Engineering Res. Dept., Construction Materials & Services Center,
Steel Res. Lab., JFE Steel
JFE Steel

*2 *4
Senior Researcher Deputy Manager, Staff,
Civil Engineering Res. Dept., Market Development Sec.,
Steel Res. Lab., Building Structure Engineering (East) Dept.,
JFE Steel JFE Engineering

26
 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

point steel are used in 70% of energy dissipation mem- of the whole system when the damper portion starts to
bers in dampers. LY225, a grade with relatively small dissipate energy. The ratio of the elastic stiffness KD
strain hardening and strain rate dependence, accounts of the damper portion to the elastic stiffness KF of the
for about 90% of the steel used for dampers. main frame is referred to as the “stiffness ratio k.” The
Among hysteretic dampers, the buckling-restraint stiffness ratio k expresses the contributional ratio of the
brace type is the most frequently used. The wall-panel shear force of the damper portion and the frame in the
type is often constrained by building plans. Though elastic region. In the calculation of KD, we need to con-
the yield strength of the wall-panel type can be eas- sider a deformation component due to the axial expan-
ily increased, the wall-panel structure cannot easily sion and contraction of a column adjacent to the damper
provide openings. The stud-panel type, on the other portion.
hand, provide openings readily. Yet as to be described The condition under which the damper portion of a
later, stiffness decreases due to the effect of the bend- hysteretic damper yields prior to the main flame, i.e.,
ing deformation of the members supporting the damper the condition under which a hysteretic damper holds,
and the beam members attached. In spite of this, the is δDy < δFy. Hence, the contributional ratio of the yield
last three years have seen the increasing adoption of the strength β of the damper portion must satisfy the follow-
stud-panel type in residential RC high-rise buildings, ing equation5):
structures that have relatively rigid beam compared with
steel structures and must be designed with passages and
other types of openings. The buckling-restraint brace
type seems to be studied as a hysteretic damper with
balanced properties from these standpoints. where Uβ is an upper limit value of the contributional
ratio of yield strength β of the damper portion.
2.2 Design Method for
The following has been proposed as the optimum
an Effective-Moment-Resistant Frame
value for βopt of the contributional ratio of yield strength
with a Hysteretic Damper
β of the damper portion, on the precondition that the
Figure 1 shows the restoring characteristics of a plastic deformation of the main frame is slight6):
moment-resistant frame (MRF) with a hysteretic damper.
An MRF with a hysteretic damper is divided into a main
frame consisting of columns and beams, and a damper
portion consisting of a damper with connecting and sup-
porting members. The shearing springs replacing the The following has been proposed as the range in
main frame and the damper portion are presumed to which a damper portion can be expected to confer an
have the restoring characteristics of a complete elasto- effective hysteresis damping effect7):
plastic type. The ordinate of the restoring characteristics
of an MRF with a hysteretic damper shown in Fig. 1
represents the story shear force Q, and the abscissa rep-
resents the inter-story displacement δ. β is an index of
yield shear force of the damper and expresses the con- The results of seismic response analyses have veri-
tributional ratio of the damper portion to the maximum fied that the displacement response becomes minimal
story shear force Qu of the whole system. ψ, or the “trig- when β is in the vicinity of βopt, and that the hysteresis
ger level coefficient,” is an index of the story shear force damping effect of the damper portion tends to decrease
gradually when β exceeds Uβ 8,9).
MRF with hysteretic damper The values of Eqs. (1) to (3) are shown in Fig. 2.
Qu Because the stiffness ratio k is preferably in the range
KF+KD= (1+k) KF
Frame KF of 0.5 to 2, the contributional ratios of yield strength for
1 1
QFy= (1−β) Qu each story can clearly be set in the range of 0.1 to 0.3
Q

Qy=ψQu KF Damper according to the stiffness ratio, k.

QDy=βQu 1
Because the damper yields earlier than the main
KD=kKF
1 frame, we also need to consider the yield strength
δDv δFy increase due to strain hardening after plasticization.
δ The upper limit value of damper steel materials is
ψ:Trigger level coefficient assumed to be the median of a specified maximum yield
β:Strength ratio of the damper
k:Stiffness rato of the damper strength, i.e., 250 N/mm2 for LY100 and 350 N/mm2 for
Fig. 1 Restoring characteristics of MRF with hysteretic damper LY225. The material standard strengths of LY100 and

JFE TECHNICAL REPORT No. 14 (Dec. 2009) 27

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

QDyi
1.0

Shear force
0.8 KDi
S KDi
0.6 U β
β

β
U opt
Story drift
0.4
βopt δyi
S M δyi
0.2 1
Uβ δyi
2.8
0 Fig. 3 Elastic stiffness of hysteretic damper system
0.5 1 1.5 2
k
LD LD
Fig. 2 Relationship of β To k δ
S yi

ABi δ
S yi
LY225 can be considered 88 N/mm2 and 225 N/mm2, ABRi i-th story
respectively, hence yield strength increases of about Hi LBRi
2.8 times and 1.6 times can be expected for a damper λLBRi
using LY100 and a damper using LY225, respectively. i-th story
In consideration of the yield strength increase of LY100, 2LD ξABRi
Fig. 2 also plots the values obtained by multiplying the
(a) Shear deformation component
upper limit value of the contributional ratio of damper
yield strength by 1/2.8. This value takes on numerical
values relatively close to βopt in the range up to k = 2. ρi δ
M yi
ui
That is, the damper will not lose the hysteresis damp-
Mδyi ρi
ing effect early if the contributional ratio of the yield i
strength β of the damper is set at a value less than βopt NDi i-th story
ρi−1
according to k. NDi
ρi
From the foregoing, we might assume that if the i-th story
damper yield strength on each story is set at a value of ρi−1
i−1
not more than βopt, we would not need to consider a yield
strength increase of the damper. With a damper steel (b) Bending deformation component
material with relatively small strain hardening, such as Fig. 4 Deformation components of hysyteretic damper system
LY225, the yield strength increase due to strain harden-
ing has only a small effect even when the contributional mation component Sδyi during the damper yielding of
ratio β of the damper yield strength is set at a value in the i-th story and the bending deformation component
the vicinity of Uβ. Mδyi due to the axial expansion and contraction of col-
umns of a damper-installed span as examples. QDyi is the
2.3 Elastic Stiffness
yield shear force of the damper system. Hi denotes the
of Hysteretic Damper System
story height. If Mδyi is ignored, the elastic stiffness of the
As described above, the design of a moment-resistant damper system becomes SKDi. In actuality, however, the
frame with a hysteretic damper requires that we appro- apparent stiffness decreases to MKDi due to Mδyi.
priately set the yield strength and stiffness of a hysteretic As shown in Fig. 4, the elastic stiffness SKDi related
damper and supporting members (a “damper system”) to Sδyi in the case of a buckling-restraint brace is
on each story in the main frame. The yield strength of expressed by the following equations, in consideration
a damper system can be evaluated in a simple manner of the axial expansion and contraction of the brace and
as a shear force acting on the damper system when the beam:
damper yields. Yet in the case of a high-rise building,
the stiffness of the damper system is greatly affected by
the total bending deformation due to the expansion and
contraction of columns supporting the damper. This calls
for great care in evaluating the stiffness of the damper
system.
A method for analytically evaluating the effect is
described below.
Figure 3 shows the elastic stiffness KDi of the where, LBRi and λLBRi denote the total brace length of
damper system of an i-th story by taking the shear defor- the i-th story and the nominal damper length without the

28 JFE TECHNICAL REPORT No. 14 (Dec. 2009)

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

Without bending deformation

length of the elastic connecting section, respectively; With bending deformation
ABRi and ξLBRi denote the sectional area of the brace Story Story
and the sectional area of the connecting section, respec- 8 B08 8 C08
eq
tively; A BRi denotes the equivalent sectional area of the 7 7
brace; 2LD denotes the length of the span over which the 6 6
damper is installed; ABi denotes the sectional area of the 5 5
upper floor beam of the i-th story; and E denotes Young’s 4 4
modulus. 3 3
As shown in Fig. 4(b), a relative rotational angle of 2 2 KD kN/mm
KD kN/mm
ρi is generated in the floor beams of the i-th story. Mδyi is 1 1
0 300 0 300
expressed as follows using this ρi:
(a) BRB (b) Stud-panel type
Fig. 5 Elastic stiffness distribution of hysteretic damper
system

An additional axial force NDi is applied to the col- tem of the buckling-restraint brace is two or three times
umn, and the column expands and contracts as shown in as high as that of the stud-panel type. Even in the eight-
Fig. 4(b). The amount of expansion and contraction ui story building, the apparent stiffness in the upper part of
of the column of the i-th story is expressed by the fol- the building decreases to less than one-half of that when
lowing equations: the effect of the total bending deformation is ignored.
We thus find that the buckling-restraint brace is greatly
influenced by the total bending deformation due to the
axial expansion and contraction of the column, and that
the stud-panel type damper is greatly influenced by the
elastic deformation of the supporting members of shear
panels and attached beams.

3. JFE’s Vibration Damping Devices

where Aci denotes the sectional area of the column adja-
3.1 Structural Types of Hysteretic Damper
cent to the damper of the i-th story.
Using ui of Eq. (7), we obtain the relative rotational JFE produces hysteretic vibration dampers of the
angle ρi from the following equation: three structural types (the brace type, the assembled
stud-panel type, and the wall type) shown in Fig. 6.
Extra-mild steels are used in the plasticized parts of each
(SN400 grade steel is also used in the brace type).

3.2 Restoring Force and

Therefore, MKDi is expressed as follows:
Fatigue Characteristics
of Shear-Yielding Type Dampers
In shear-yielding type vibration dampers (the wall
type and the assembled stud-panel type), the steel grade
and width-thickness ratio of the damper steel used in the
We obtain the elastic stiffness KDi of the damper sys- panel part both have influences on the hysteretic charac-
tem of the i-th story from Eqs. (4) and (10). In the case teristics and amount of energy dissipation of the damper.
of an inverted K type, where buckling-restraint braces To ensure sufficient energy dissipation characteristics
intersect each other at one point on a lower beam of the
i-th story, we can make the calculations by changing i + 1
of Eq. (8) to i.
Figure 5 shows the stiffness of a damper system of
an eight-story frame designed using buckling-restraint
braces and stud-panel type dampers in two cases: once
with the total bending deformation taken into account,
and once with the total bending deformation ignored. As
the figure clearly shows, the stiffness of the damper sys- Fig. 6 JFE Hysteretic Damper

JFE TECHNICAL REPORT No. 14 (Dec. 2009) 29

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

during an earthquake, we need to grasp the fatigue char- Experiment Experiment

Model
acteristics associated with the small-amplitude cyclic 300
Culculate
τMAX⫽1.5τ y 600
loading generated by wind loads and the large-amplitude

Shear stress (N/mm2)

200 0 G1
cyclic loading during large earthquakes.
100 G1 400
We confirmed the performance of the shear-yielding

Eη
0
type vibration dampers by performing two types of tests.
⫺100 200
First, we performed the loading test with progressively
higher loads using three parameters: the steel grade, ⫺200
0.02205 G1
the width-thickness ratio, and the loading rate. Second, ⫺300 0
⫺0.06 ⫺0.03 0 0.03 0.06 100 200 300 400
we performed the low-cycle fatigue test performed Shear drift angle (rad) η
using two parameters: the steel grade and the loading
(a) Hysteresis loops (b) Cumulative ductility factor
amplitude. To test the performance of the shear panel,
Fig. 8 Comparison of model for experiment (LY225)
we applied a load via displacement control of a 150-t
actuator with the loading device shown in Fig. 7. For
0.1
the load test with progressively higher loads, the wave-

Shear drift angle (rad)

LY100
form of the dynamic test was a sine wave of 2 Hz and Δγt LY160
the waveform of the static test was a triangular wave of Δγp LY235
0.5 mm/s. After amplifying the amplitude proportion- 0.01
ally from 1/800 rad to 6/100 rad, the amplitude of 6/100
rad was repeated until cracks passed through the panel Δγe
part. In the low-cycle fatigue test, cyclic loads were 0.001
101 102 103 104 105
applied until the yield strength decreased to 95% of the
Number of cycle
maximum yield strength, or until cracks passed through.
Fig. 9 Relation of fatigue life
The restoring characteristics and fatigue characteris-
tics obtained from the results of the series of tests are
described below. Figure 9 shows the relation of the number of cycles
Based on the results of the static loading test with to the total loading amplitude of each steel grade in the
progressively higher loads, we tried to model the hyster- results of the fatigue test for the shear panel. Inciden-
esis for each steel grade and each width-thickness ratio. tally, the number of cycles N adopted is the number of
Though Ramberg-Osgood type models have generally cycles obtained when the yield strength decreased to
been proposed, we used tri-linear type modeling in this 95% of the maximum yield strength. Both the elastic
study, in consideration of the universality and the ease strain Δγe and plastic strain Δγp can be approximated by
of design. Figure 8(a) shows modeling with LY225, as a straight line, and the Manson-Coffin rule holds. The
an example. Errors appear because the rise gradient of test results for LY100 were obtained from the assembled
the experiment decreases as the loading amplitude rises. type, hence the plastic strain tends to rise to somewhat
Yet loops quite similar to each other are plotted. Fur- high levels at low amplitudes as a result of the difference
thermore, as shown in Fig. 8(b), the experiment and the in constraint conditions. On the whole, however, LY225
calculation are almost in agreement at low amplitudes, has a long fatigue life, exhibiting a tendency about the
in terms of the cumulative ductility factor. The energy same as that observed in the material test results. With
dissipation of the hysteresis model is evaluated at about repetitive application with a loading amplitude of not
90%, and the value tends to fall further as the shear drift more than about 0.01 rad, the fracture mode provides
angle increases. weld cracks, hence we ultimately find little difference
between the steel grades.
Displacement The fatigue life curve is expressed by the following
control equation for each steel grade:

Pin

Load beam 150 t Actuater

Test specimen

Fig. 7 Setup of shear panel test

30 JFE TECHNICAL REPORT No. 14 (Dec. 2009)

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

Axial member 300

(Tube) E2
200 σd
σy

Stress (N/mm2)
100
E1
Constraint 0
Axial E2=E1/55
member (Tube) −100
member (Flat-bar) σy
−200 σd
−300
−2 −1 0 1 2 3
Constraint Axial strain (%)
member (RHS)
Fig. 11 Evaluation of hysteretic loop model
Fig. 10 Buckling restraint braces by tube

0.5 1.0 1.5 (%)

300
3.3 Restoring Force and Fatigue Characteristics

Cumulative ductility factor

Experimental value
of Brace-Type Dampers 250 Calculated value
200

Various strategies entailing the use of RC members, 150

steel tube concrete, shapes, etc. have been proposed for 100
the buckling-restraint method for the axial members of 50
the brace10). Two points are to be considered for the con- 0
struction and design of buckling-restraint braces: Cycle
(1) The yield strength and stiffness are sufficient to pre- Fig. 12 Comparison of cumulative ductility factor
vent the buckling of the axial member.
(2) The plastic contraction allowance of the axial mem- 10
ber is ensured, and insulation is used to prevent the Material
Flat-bar
friction and adherence of the axial member and the Double-tube
Strain amplitude (%)

restraining member.
Any buckling-restraint method for the axial member
of the brace may be adopted, provided that both of these 1
points are ensured. The steel tube JFE uses as the brace-
type damper serves as a buckling-constraint member that
requires no special insulation (Fig. 10). The two points
described above are ensured by appropriately setting
0.1
the width-thickness ratio, diameter-thickness ratio, and 1 10 100 1 000 10 000 100 000 1 000 000
clearance of the axial member and constraint steel tube. Number of cycle
The authors has conducted the following experiments Fig. 13 Fatigue characteristic of BRB
to grasp the conditions necessary to ensure that this
brace will exhibit sufficient performance as a damping loops and the fatigue characteristics. To obtain as simple
member: an expression as possible, we perform the modeling
(1) Cyclic loading experiment with a brace as a single using a tri-linear model in which the yield strength (σy)
member with variation in the following parameters: and the yield strength after strain hardening (σd) provide
the mechanical properties of the axial member, the break points. Figure 11 shows a hysteresis loop model
slenderness ratio of the auxiliary steel pipe, the of LY225. Figure 12 compares experimental and calcu-
width-thickness ratio of the axial member, the diame- lated values of the changes in the cumulative ductility
ter-thickness ratio of the auxiliary steel tube, and the factor resulting from the strain dissipation energy. The
clearance between the axial member and the auxiliary calculated and experimental values for the cumulative
steel tube ductility factor correspond well, though the former are a
(2) Partial frame experiment to grasp the applicability little lower than the latter. This model therefore appears
to an actual structure and the hysteresis characteris- to sufficiently serve its purpose.
tics as a moment restraint frame with a brace Figure 13 shows results of a fatigue rest conducted
(3) High-speed loading experiment with the actual seis- on test specimens using LY100 and LY225 as axial
mic ground motions considered members and materials. The axial members have the
(4) Fatigue characteristics experiment sectional shapes of flat bars and circular tubes. The
In this section we describe the modeling of hysteresis ordinate in the figure denotes the total strain range, and

JFE TECHNICAL REPORT No. 14 (Dec. 2009) 31

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

the number of cycles adopted is the number of cycles

obtained when the tension-side peak load decreases to
95% of the load in a stable state.
The figure also shows the results of flat bar (LY100)
by the following equation as an example of a fatigue life

38 400
161 000
curve:

38 400

GL
38 400 unit: mm
We thus find that if the buckling-constraint member
of a brace type damper is appropriately designed, the Fig. 14 Analytical model
prescribed deformability will be obtained irrespective of
the axial sectional shape. 600 OSA_NS
h⫽10% WOS_EW
500 NAGOYA_EW
TOMA_NS
4. Evaluation of the Performance of Dampers 400

VE (cm/s)
against Long-Period Seismic Ground Motions 300

4.1 Long-Period Seismic Ground Motions 200

by Ocean-Trench Earthquakes 100

0
1 2 3 4 5 6 7 8
Ocean-trench massive earthquakes are characterized Period (s)
by an abundance of long-period components and long Fig. 15 Energy spectrum
duration times. Their epicenters are positioned off of
Tokai, Tonannkai, Nannkai, etc. These earthquakes pose arranged in an X shape (Fig. 14). The story height is
serious threats to high-rise buildings and like structures 5 m for the first floor and 4 m for every floor above it.
with long natural periods. The dampers for these struc- The planar shape is the same for all stories. The mem-
tures must therefore perform especially well. ber construction consists of 600 × 600 mm columns
In this section we evaluate whether the above- with box-shaped sections with plate thicknesses of 35
described tube-in-tube type brace can be expected to to 55 mm, and girders 700 to 800 mm in depth and 200
meet the performance requirements in the event of a to 250 mm in width with H-shaped sections with flange
massive ocean trench earthquake. We begin by preparing thicknesses of 14 to 28 mm. The damping constant is
a model of a high-rise building of a type thought to be 2%. The degree of yield stress of the axial member of
greatly influenced by massive ocean-trench earthquakes, the brace is 100 N/mm2 and the distribution conforms
then perform a time history response analysis with the to an optimum distribution of the yield shear force coef-
OSA-NS waves (simulated waves)11) of an ocean-trench ficient based on the damper yield strength of the first
earthquake and determine the required performance of story. We set the primary natural period of a model of
the damper based on our analytical results. Next, we the frame alone at T1 = 4.64 s. Figure 15 shows the
conduct a dynamic experiment in which the structural energy spectra of long-period seismic ground motions
response waves obtained by the above-described time after references 3) and 11). We use the OSA-NS wave in
history response analysis are imposed on the tube-in- this analysis.
tube type brace. We derive the actual performance from
4.2.2 Analysis results
the results of the experiment and compare it with the
calculated required performance. We perform a time history response analysis using
the yield shear force coefficient sαy1 of the damping
4.2 Required Performance
brace of the first story as a parameter. Figure 16 shows
of Hysteretic Dampers
acceleration-converted energy input VE, and Fig. 17
shows the ratio of the energy dissipation of the damp-
4.2.1 Input seismic waves
ing brace to the input energy. From Fig. 16 we obtain
and analytical model
VE = 200 to 280 cm/s, and this energy input agrees sub-
The analytical model was a 40-story building of a stantially with the VE shown in Fig. 15. From Fig. 17
steel construction, with analytical model damping braces we obtain sαy1 = 12 × 10−3 as optimum value, where the

32 JFE TECHNICAL REPORT No. 14 (Dec. 2009)

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

400 the yield strength Ny is Nyd = 1 658.2 kN. If we select

300
the axial-load-carrying tube from the 100 N/mm2 class
low-yield-point steel tubes with a slenderness ratio λ of
VE(cm/s) 200 120 or less and a diameter-thickness ratio of 15, the size
of the tube available from among standard JFE sizes is
100
φ 298 × 19.9. On the assumption that both ends are pro-
vided with a 1-m connecting section, however, we set
0 5 10 15 20
α (⫻10⫺3)
s y1
the axial-load-carrying tube length at 5.5 m.
Fig. 16 VE vs. sαy1 For the analysis results, the maximum value ηdmax of
η was 658.6 kN·m and that of Wpdmax was 491.5 kN·m at
−3
sαy1 = 12 × 10 .
0.4
Energy dissipation ratio

4.3 Performance of the Hysteretic Dampers

0.3

0.2 4.3.1 Outline of experiment

0.1 The test specimen used in the experiment is a tube-
in-tube type buckling-restraint brace (inner-tube restraint
0 5 10 15 20 type) in which both ends are pin connected. A 100 N/mm2
s αy1 (⫻10⫺3)
class low-yield-point steel tube (JFE-LY100) with a
Fig. 17 Energy dissipation ratio vs. sαy1
diameter-thickness ratio of 15 is used as the axial-load-
carrying tube, and an ordinary steel tube is used as the
constraint tube. Details of the test specimen are shown
40 MRF
s a y1=0.012 in Fig. 20.
The load is imposed by repeatedly inputting the input
30
wave described below until the test specimen is bro-
MRF ken. For comparison with the required capacity under
Story

20
With an increasingly severe condition for response, we per-
damper formed a time-history-response analysis with a model
10
of a frame alone and prepared the input wave based on
0 the horizontal displacement response wave of the 7th
5 10 15 story with the story drift angle set at maximum. First,
Story drift angle, R (×10−3)
we converted the horizontal displacement response wave
Fig. 18 Distribution of story drift angle

800

600

400
η

200

0
5 10 15 20
αy1 (⫻10⫺3)
s
Fig. 20 Testing specimen
Fig. 19 η vs. sαy1

value at which the energy dissipation ratio is regarded as Command strain %

1.5
local maximum12). Figure 18 shows the story distribu-
(%)
Strain

tion of the maximum story drift angle R. When sαy1 = 0

12 × 10−3, R is within 1/100 even in the seventh story, 21.5
where the maximum drift angle is the largest. Figure 19 1.5 Measured strain % 1st test
(%)
Strain

shows the cumulative ductility factor η13) of the damp- 0

ing brace of the seventh story. The value of sαy1 declines
21.5
as the value of η increases, and ηd = 186.3 when sαy1 is 0 100 200 300 400 500
optimal. The cumulative plastic strain energy Wp per Time (s)
damping brace at this point is Wpd = 834.1 kN·m, and Fig. 21 Time history of strain of hysteretic damper

JFE TECHNICAL REPORT No. 14 (Dec. 2009) 33

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

into an axial displacement wave of the damping brace 200

Stress (N/mm2)
and corrected for the different lengths of the axial-load- 100
carrying tubes of the test specimen and brace. The actu- 0
ally inputted wave (the first wave) is shown in Fig. 21.
⫺100
The amplitude ratio on the ordinate, a value obtained by
⫺200
dividing the input relative displacement by the length of
the axial-load-carrying tube, is equivalent to the strain ⫺1.5 ⫺1.0 ⫺0.5 0.0 0.5 1.0 1.5
Strain %
amount of the axial member of the brace. The experi-
Fig. 22 Stress-strain relation
mental design is clearly appropriate when we consider
the significant effect of the high-amplitude ratios on the
energy dissipation of the damper. Figure 22 shows the stress-strain relation of the
first wave. Here we find that stable hysteresis loops
4.3.2 Experimental results and discussion
are drawn even though the maximum stress intensity
Loading was performed 10 times in all. We observed increases due to the effect of strain hardening. We cal-
a decrease in the yield strength and fractures during the culate the cumulative ductility factor η of the test speci-
tenth loading, and therefor decided to end the experi- men by adding the values obtained each time, as shown
ment. in Table 1. The result becomes ηc = 16 081.1. Wpc at
Photo 1 shows the condition of the final fracture. this point is 908.4 kN·m, and Nyc is 124.4 kN. To keep
Buckling occurred at an end portion of the brace test the assumptions conservative, however, we exclude the
specimen, and a crack occurred in the valley portion of value that led to the fracture in the tenth loading. ηc is
the buckling, leading to the fracture. about 86.3 times the value of ηd found in the analysis,
Table 1 shows results of the loading performed each and the tube-in-tube brace in question has sufficient
time. The cumulative ductility factor η decreases a little performance. For ηdmax also, we provide an allowance of
as the number loads increases. We speculate that this approximately 24.4 times. And for the maximum VE, of
was the result of a decrease in the stiffness of the brace 550 cm/s assumed from Fig. 15, ηdmax is 2432.2 and ηc
caused by an accumulation of local buckling in the end has a margin of approximately 6.6 times.
portion of the test specimen.

5. Energy-Dissipation Devices
for Both Wind-Resistant and
Earthquake-Resistant Purposes

5.1 Out Live of

Test Speciment and Experiment

This section describes the performance evalua-

Photo 1 Failure mode
tion of a hybrid damper of the partial damping wall
type now being developed at the JFE Steel (Fig. 23,
Table 1 Testing results
Photo 2). This hybrid damper is composed of a visco-
Maximum stress Dissipation elastic damper designed for wind response connected
(N/mm2) Cumulated
Number energy
ductility factor, η
in a series-parallel manner with a hysteretic panel
plus minus (kNm)
1 214 201 105 1 863
2 222 204 103 1 829 Kp⫽34.62 kN/mm
Kp/Kpcal⫽1.031
3 224 206 102 1 813
950

σy⫽225 N/mm2 19
t⫽6 mm
100

4 228 206 102 1 799

180
80

P5
25

5 226 206 101 1 783

480

H-450⫻200⫻12⫻19 t⫽5 mm
785
375
3 010

(SN490B)
6 228 206 100 1 774
580

K⫽12.16 kN/mm
P12
2 060

K/Kcal⫽0.991 High damping rabber

7 227 206 99 1 757 Kc⫽12.16 kN/mm
205

3 750cm2⫻2-layers
Kc/Kccal⫽0.996
P19
8 228 207 98 1 740 1 000

9 231 206 97 1 724

1 950
10 236 203 92 1 628
Sum. 1 000 17 709 Fig. 23 Hysteretic and visco-elastic hybrid damper

34 JFE TECHNICAL REPORT No. 14 (Dec. 2009)

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

the damping constant at small amplitudes of 2 to 3%.

(b) Figure 25 shows the damping ratio and constant of
the visco-elastic damper. The ratio of the deformation
of the high-damping rubber to the damper displace-
ment (relative displacement between the horizontal
members and the lower beam) is 75 to 90%. Though
the stiffness of the high-damping rubber is almost
the same as the stiffness obtained with the evaluation
equation in reference15), we use a value corrected for
each cycle in the evaluation of the performance of the
hybrid damper, to adjust for the slight cycle depen-
Photo 2 Testing set-up of hybrid damper
dence. The damping constant is approximately 0.3.
(c) Figure 26 shows examples of hysteresis loops of the
damper effective in seismic response14). The hysteretic hybrid damper at small amplitudes. The deformation
panel damper is composed of two studs and a horizon- of the high-damping rubber is approximately 60%
tal member (both: H-450 × 200 × 12 × 19, SN490B9), of the story drift and the contributional ratio of yield
with part of the web formed from a low-yield-point steel strength is approximately 25%. The damping constant
(t = 6 mm, σy = 225 N/mm2). The visco-elastic damper
is formed by stacking high-damping rubber (t = 5 mm)
made by Yokohama Rubber Company, Ltd. in two lay- 100

Damping ratio, Kve (kN/mm)

80
ers. We call this hybrid damper a “series-parallel” type 60
for two reasons: first, the damper is made up of two 40
20 Calculated
types of dampers; second, the visco-elastic damper is stiffness
0
connected in series to a low-yield-point panel via a hori-
0.3
zontal member and connected in parallel to the studs.
0.2
We tested this hybrid damper by subjecting it to 0.3Hz
0.1 1.0Hz
dynamic loading at 0.3 and 1 Hz as shown in Photo 2.
The amplitude used in the experiment is divided into 0 1 2 3 4 5 6
small amplitudes (assumed for wind response) and large Displasement (mm)
amplitudes (assumed for seismic response), with a story Fig. 25 Damping ratio-displacement relation of visco-elastoc
drift angle R of 1/500 (δ = 6 mm) serving as the bound- damper
ary. 0.6
0.3Hz, δ⫽4mm
0.4
5.2 Experimental and Discussion
0.2
(a) Figure 24 shows the restoring characteristics of the
Q/Qp

0.0
low-yield-steel panel of the hysteretic panel damper.
⫺0.2
The maximum displacement Rpmax of the panel is
1/16.9 rad and the yield strength decreases due to ⫺0.4 Conbined
Viscoelastic
buckling. All parts of the studs and horizontal mem- ⫺0.6
⫺0.15 ⫺0.10 ⫺0.05 0 0.05 0.10 0.15
bers remain within the elastic range even after the
R (10 ⫺2 rad)
panel buckles, except for the parts in the vicinity of
Fig. 26 Hysteresis loops of hybrid damper (Small amplitude)

Hysteresis only, η⫽230.2

Combined, η⫽231.2 1.0
1.50 Yielding
1.25 0.8
Hysteresis damper
Energy rate

1.00 0.6 +Horizontal

Q/Qp

0.75
0.4
0.50
0.2
0.25 Visco-elastic
damper+Stud
0 2 46 8 10 12 0 0.2 0.4 0.6 0.8 1.0
δ/δp R (10−2 rad)
Fig. 24 Q-δ relation of hysteretic shear panel damper Fig. 27 Energy rate of hybrid damper due to story drift angle

JFE TECHNICAL REPORT No. 14 (Dec. 2009) 35

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

1.5
1 Hz, increase
1.0

0.5 Hysteresis damper

Q/Qp Hysteresis damper

3 010
0

3 010
⫺0.5 Visco-elastic damper
Visco-elastic damper
⫺1.0
Condined
Calculate
⫺1.5
⫺1.5 ⫺1.0 ⫺0.5 0 0.5 1.0 1.5
R (10⫺2 rad) unit: mm
Fig. 28 Hysteresis loops of hybrid damper (Large amplitude) (a)Small amplitube (wind) (b)Large amplitube (earthquake)

of the hybrid damper is 6 to 7%. At R = 1/500 or less, Fig. 29 Energy dissipation mechanism of hybrid damper
the hysteretic panel damper is elastic and mainly the
visco-elastic damper dissipates energy (Fig. 27). As shown in Fig. 29, the energy dissipation in the
(d) Figure 28 shows hysteresis loops of the hybrid series-parallel hybrid damper is performed mainly by
damper at large amplitudes. Rpmax = 1/77.9 rad. The the visco-elastic damper at small amplitudes and by the
performance of the panel is the same as with the hysteretic panel damper at large amplitudes. Hence, the
hysteretic panel damper alone shown in Fig. 24. series-parallel hybrid damper is effective for both wind
The contributional ratio of energy dissipation of the response and seismic response. And at large amplitudes
panel increases with increasing amplitude (Fig. 27). we can suppress the deformation of the visco-elastic
The deformation of the high-damping rubber is sup- damper to less than the elastic limit of the studs and
pressed to not more than 10 mm (γ = 200%), and no improve the stiffness per unit area by making the visco-
decrease in performance is observed. At R = 1/100, elastic damper thin.
the energy dissipation capacity of the hysteretic panel
6. Concluding Remarks
damper increases by about 8% due to the effect of the
visco-elastic damper. In this paper we described the trend in the applica-

Table 2 Example of buildings installed JFE hysteretic dampers

Number Structural Material Number of

No. Use Damper type Site
of story type for dampers dampers
1 Office 12-story Steel LY100 118 BRB Tokyo
2 Office 17-story Steel LY100 128 BRB Tokyo
3 Office 28-story Steel LY100 156 BRB Tokyo
4 Office 11-story Steel LY100 32 BRB Tokyo
5 Office 26-story Steel LY100 340 BRB Tokyo
6 Residence 40-story RC LY100 80 Stud-panel Tokyo
7 Complex 25-story Steel LY225 380 BRB Tokyo
8 Complex 13-story Steel LY225 112 BRB Tokyo
9 Office ̶ Steel LY225 208 BRB Okayama
10 Residence 31-story RC LY225 70 Stud-panel Tokyo
11 Residence 30-story RC LY225 72 Stud-panel Tokyo
12 Office 39-story Steel LY225 168 BRB Tokyo
13 Residence 30-story RC LY160 444 Stud-panel Kanagawa
14 Residence 58-story RC LY225, etc. 1 152 Stud-panel Tokyo
15 Residence 59-story RC LY225 16 Stud-panel Tokyo
16 Office 16-story Steel LY225 85 BRB Osaka
17 Residence 40-story RC LY225 92 Stud-panel Tokyo
18 Manufacturing ̶ Steel LY225, etc. 70 BRB Tokyo
19 Office 23-story Steel LY100 28 BRB Tokyo
20 Retrofits 8-story Steel LY225 ̶ BRB Tokyo
BRB : Buckling restraint brace

36 JFE TECHNICAL REPORT No. 14 (Dec. 2009)

 Seismic Response Control for High-Rise Buildings Using Energy-Dissipation Devices

for long-period seismic ground motions described in

Chapter 4. The authors would like to extend their thanks
to the persons concerned.

References
1) Headquarters for Earthquake Research Promotion. “National
Seismic Hazard Maps for Japan (2007).” 2007.
2) e.g., Ito, S.; Kamura, H.; Shimokawa, H.; Katayama, T.;
Hirota, M.; Ueki, T. Hysteretic Energy Dissipation Devices
with Low Yield Strength Steel for Passive Structural Control.
NKK Technical Report. 2000, no. 170, p. 67–74.
3) Tsunoda, M.; Ishii, T.; Miyagawa, K.; Kitamura H. Evalu-
ation On Performance Of Buckling-Restrained Tube-In-
Tube Energy Dissipative Braces for the Long-Period Ground
Motion. Summaries of Technical Papers of Annual Meeting
Architectural Institute of Japan. C-1, 2006-07, p. 895–896.
4) Japanese Society of Steel Construction. Guidelines for
Dynamic Seismic Design of Steel Frames Using CFT Col-
Photo 3 Example of buildings installed JFE hysteretic dampers umns. JSSC Technical Report. 2006, no. 76.
5) Inoue, K. Seismic Design of Structures with Hysteretic
Damper, Proceedings of Symposium on a new course of Seis-
tion of damping structures to recent high-rise buildings, mic Design. Architectural Institute of Japan. 1995, p. 95–111.
6) Building Research Institute and Japan Iron and Steel Fed-
important points to keep in mind in the structural design eration. Seismic Design of Steel Structures with Hysteretic
of damping structures, and an outline of the structural Damper. 2002.
performance of JFE-developed vibration dampers. 7) Inoue, K.; Ono, S. Optimum Strength Ratio of Hysteretic
Damper and Design Strength of Frames. Journal of Structural
Further, we discussed the required and actual energy- Engineering. 1995, vol. 41B, p. 9–15.
dissipation performance of vibration dampers installed 8) Kamura, H.; Inoue, K.; Kuwahara, S.; Ogawa, K. Modeling
in high-rise buildings. of the Moment Resistant Frame with Hysteretic Damper to
Fishbone-Shaped Frame for the Response Analysis. Journal of
Lastly, Table 2 and Photo 3 show representative Structural and Construction Engineering. Transactions of AIJ.
track records of the vibration dampers introduced in 2002-12, no. 562, p. 151–158.
this paper, as observed over the past five years. As the 9) Ogawa, K. Residual Deformation of Steel Frames with Hys-
teretic Dampers. Journal of structural and construction engi-
table clearly illustrates, many of JFE’s vibration damp- neering. Transactions of AIJ. 2001-01, no. 539, p. 151–158.
ers are adopted in RC high-rise buildings, as well as in 10) e.g., Maeda, Y.; Kamura, H.; Takeuchi, T.; Iwata, M.; Wada A.
steel-frame high-rise buildings. And the adoption of JFE Fatigue Properties of Practical-Scale Unbonded Braces (Part 1,
2). Summaries of technical papers of Annual Meeting Archi-
vibration dampers in production facilities other than tectural Institute of Japan. C-1, 1999, p. 813–814.
high-rise buildings is also increasing. Many predictions 11) Kawabe, H.; Kamae, K.; Irikura, K. Strong Motion Simula-
of damage due to massive earthquakes, such as inland tion of Hypothetical Nankai Earthquake Using Characteristic
Source Model. Proceedings of the 2002 Fall Meeting of Seis-
earthquakes in the Tokyo Metropolitan Area, have been mological Society of Japan. A31.
disclosed to the public in recent years. This is one of the 12) Atsumi, T.; Uramoto, H.; Ishii, M.; Kitamura, H. Vertical Dis-
reasons why seismic dampers are expected to become tribution of Hysteretic and Velocity-Dependent Dampers for
High-Rise Steel Structure. Summaries of Technical Papers of
more widespread in various building applications in the Annual Meeting Architectural Institute of Japan. B-2, 2006,
years to come. JFE will take part in this trend by devel- p. 763–766.
oping new vibration damping techniques to meet social 13) Akiyama, H. Seismic Design of Buildings Based on Energy
Balance. Gihodo Shuppan. 1999-11.
needs in the future, as well. 14) Hirota, M.; Katayama, T.; Kamura, H. Study on Shear Wall
Damper connected H-Section Members Using Low Yield
Portions of the results of a research project conducted Strength Steel (Part 5‒7). Summaries of Technical Papers of
Annual Meeting Architectural Institute of Japan. C-1, 1998,
jointly with the Kitamura Laboratory, Structural Engi- p. 801–806.
neering, Department of Architecture, Faculty of Science 15) Ishikawa, R.; Suwa, H.; Gokan, S.; Suzuki, T. A Study of
and Technology, Tokyo University of Science are used Mechanical Model for Viscoelastic Damper (Part 1, 2). Sum-
maries of Technical Papers of Annual Meeting Architectural
in the study on the performance evaluation of dampers Institute of Japan. B-2, 1999, p. 959–962.

JFE TECHNICAL REPORT No. 14 (Dec. 2009) 37