13th World Conference on Earthquake Engineering

Vancouver, B.C., Canada

August 1-6, 2004
Paper No. 2175

PSEUDO-DYNAMIC TEST OF A FULL-SCALE CFT/BRB FRAME:

PART 2 - CONSTRUCTION AND TESTING
Chui-Hsin CHEN 1, Po-Chien HSIAO2, Jiun-Wei LAI 3,
Min-Lang LIN 4, Yuan-Tao WENG5 and Keh-Chyuan TSAI6

SUMMARY
This is the second in a three-part paper describing a pseudo-dynamic test program for a full-scale threestory three-bay CFT/BRB frame, conducted through international collaboration between researchers in
Taiwan and the United States. While the first and third companion papers describe the performance-based
design details and the validation of simulation models of the frame respectively, this paper focuses on the
construction and experimental testing of the specimen. A full-scale braced frame with concrete filled tube
(CFT) columns and buckling restrained braces (BRB) was tested in October of 2003 to study the seismic
performance of such system. Three different types of moment connections, namely through beam, external
diaphragm and bolted types, varying from floor to floor have been designed and constructed for the
exterior beam-to-column joints. Furthermore, three brace types, including the single-cored, double-cored
and all-metal BRBs were installed in the three different stories. The test frame was loaded using pseudodynamic procedures to simulate the structural responses under ground motions corresponding to
earthquake hazards for a high-seismicity site with 50%, 10%, and 2% chance of exceedance in 50 years.
Nonlinear dynamic analyses illustrate that the peak story drift is likely to reach 0.025 radian after applying
the 2/50 design earthquake on the frame specimen. Further, the experimental responses of the specimen
also demonstrate great agreement with the predictions. In addition to the experimental investigations, an
internet testing platform entitled Internet-based Simulations for Earthquake Engineering (ISEE), launched
in the National Center for Research on Earthquake Engineering as a prototype of Internet-based

, Assistant Research Fellow, National Center for Research on Earthquake Engineering, 200, Sec.3,
Hsinhai Rd., Taipei 10658, Taiwan, ROC, Email: chchen@ncree.gov.tw
2
Graduate student, National Taiwan University, Email: r91521221@ntu.edu.tw
3
Research Assistant, National Center for Research on Earthquake Engineering, 200, Sec.3, Hsinhai Rd.,
Taipei 10658, Taiwan, ROC, Email: jwlai@ncree.gov.tw
4
Associate Research Fellow, National Center for Research on Earthquake Engineering, 200, Sec.3,
Hsinhai Rd., Taipei 10658, Taiwan, ROC, Email: mllin@ncree.gov.tw
5
Postdoctoral Researcher, National Center for Research on Earthquake Engineering, 200, Sec.3, Hsinhai
Rd., Taipei 10658, Taiwan, ROC, Email: d88521003@ms88.ntu.edu.tw
6
Director, National Center for Research on Earthquake Engineering, 200, Sec.3, Hsinhai Rd., Taipei
10658, Taiwan, ROC, Email: kctsai@ncree.gov.tw

cooperative structural experimental environment is described. Preliminary experimental results illustrate

that CFT/BRBF performs extremely well after the application of design earthquake load effects.
INTRODUCTION
Through international collaboration between researchers in Taiwan, Japan, and the United States, a fullscale 3-story 3-bay RC column and steel beam RCS composite moment frame was tested in October of
2002 in the structural laboratory of National Center for Research on Earthquake Engineering (NCREE) in
October 2002 [1]. In the year 2003, a full-scale 3-story 3-bay CFT column with the buckling restrained
braced composite frame (CFT/BRBF) specimen has been tested in October in a similar manner. The 3story prototype structure is designed for a highly seismic location either in Taiwan or United States. The
typical bay width is 7m and typical story height is 4m. The total height of the frame, including the footing,
is about 13m. The 2150mm wide concrete slab is adopted to develop the composite action of the beams.
Measuring 12 meters tall and 21 meters long, the specimen is among the largest frame tests of its type ever
conducted. The frame has been tested using the pseudo-dynamic test procedures applying input ground
motions obtained from the 1999 Chi-Chi and 1989 Loma Prieta earthquakes, scaled to represent 50%,
10% (DBE), and 2% (MCE) in 50 years seismic hazard levels. Following the pseudo-dynamic tests, since
none of the brace was fractured, quasi-static loads have been applied to cyclically push the frame to large
inter-story drifts up to the failure of the braces. It provides valuable data to validate possible failure
mechanism and analytical models for large deformation responses. Being the largest and most realistic
composite CFT/BRB frame ever tested in a laboratory, the test provides a unique data set to verify both
computer simulation models and seismic performance of CFT/BRB frames. This experiment also provides
great opportunities to explore international collaboration and data archiving envisioned for the Networked
Earthquake Engineering Simulation (NEES) or the Internet-based Simulations for Earthquake Engineering
(ISEE) [2] research programs launched in USA and Taiwan, respectively. This paper summarizes the
displacement-based seismic design procedures adopted in the design of CFT/BRB frame specimen.
During the planning stage, extensive nonlinear dynamic analyses were also carried out in order to ensure
the possible seismic demands would not exceed the force and displacement limits of the test facility. The
analytical predictions were broadcasted along with the real time experimental results during each test.
This paper describes the experimental and analytical results and evaluates the seismic performance of the
frame specimen.
EXPERIMENTAL SETUP
The 3-story CFT/BRB frame shown in Fig. 1 is employed in this experimental research. The prototype
three-story building consists of 6-bay by 4-bay in plane. In the two identical prototype CFT/BRB frames,
only the two exterior beam-to-column joints in each floor are moment connections, all other beam-tocolumn connections are assumed not to transfer any bending moment. The braces are installed in the
center bay. Square CFT columns are chosen for the two exterior columns while the center two columns are
circular CFTs. Story seismic mass is 31.83 ton for the 1st and 2nd floors, 25.03 ton for the 3rd floor for each
CFT/BRB frame (half of the building). All steel is A572 GR50 with yield stress of 350 MPa and the
nominal compression stress of the infilled concrete in the CFT columns is 35 MPa. The adopted
displacement-based seismic design (DSD) procedures [3], [4] assume that the CFT/BRB frame specimen
vibrates essentially in a single mode. The DSD details of the specimen can be found in the reference [5],
[6]. The final selections of structural members are given in Table 1. It considers the actual material
coupon strength as given in Table 2. The supporting beams above the braces satisfy the capacity design
principal considering the strain-hardening braces and an unbalanced vertical load resulted from the
difference of the peak BRB compressive and tensile strengths. The fundamental vibration period is about
0.68 second. Three different types of moment connections, namely through beam, external diaphragm and
bolted end plate types, varying from the first floor to the third floor were fabricated for the exterior beam-

to-column connections. The details of the moment connections are schematically given in Fig. 2. In this
full-scale test, being subjected to the large force demand induced by axial force of columns and braces, the
footings need to be strengthened by some measures. The details of the exterior and interior footings are
shown in Fig. 3 and 4 respectively. Anchor bolts and stiffeners were employed to rise the capacity of the
footings. Three types of BRBs, including the single-cored, double-cored and the all-metal BRBs, were
adopted in three different floors. In particular, two single-cored unbonded braces (UBs), each consisting
of a steel flat plate in the core, were donated by Nippon Steel Company and installed in the second story.
Each UB end to gusset connection uses 8 splice plates and 16-24mm F10T bolts. The two BRBs
installed in the third story are double-cored constructed using cement motar infilled in two rectangular
tubes [7] while the BRBs in the first story are also double-cored but fabricated with all-metal detachable
features [8]. Each end of the double-cored BRB is connected to a gusset plate using 6- and 10-24mm
F10T bolts at the third and first floor, respectively. No stiffener was installed at the free edges of any
gusset before the testing.
Table 1 Selection of member sizes and grades
Member
Location
Beam (mm)
Brace (cm2)

Beam Sizes and Core Cross Sectional Area of Braces (A572 GR50)
1FL
2FL
3FL
H4562011017
H450200914
H400200813
30
25
15

Dimension of Columns (A572 GR50) unit : mm

CFTs: C1: Tube: 3509, C2: Pipe: 4004009

Table 2 Material strength

3FL

Positions of Sampling
Beam
Flange

2FL

3BRB
Beam

1FL

2UBB
Beam

fy (MPa)
372

fu (MPa)
468

426
373
414
482
397
370
354
421
374
543

493
483
503
538
545
486
485
534
488
584

Web
core steel material
Flange
Web
core steel material
Flange
Web
core steel material
Steel
Steel
f c =35 MPa

1BRB
C1(Tube-400-9)
C2(Pipe-400-9)
Concrete

3@2.33m

2@3.5m

3@2.33m
b1

b1

C2

b1

b1

b1

b1

b1

b1

C1

3B2

3B3

2B1

BRB3
2B2

2B3

1B1

BRB2
1B2

1B3

BRB1
3@7m

Fig.1 Floor framing plan and elevation

b1

b1

b1

b1

b1

b1

b1

3B1

b1

b1

3@4m

Plane Frame

C2

C1

6@7m

2.15 m

4@7m

b1

Steel
(A572Gr.50)

b
st

nd

rd

Fig.2 Moment connection details (a, b, c for 1 , 2 and 3 floor, respectively)

Fig.3 Details of exterior footings

Fig.4 Details of interior footings

TESTING TECHNIQUESISEE
Internet-based Simulations for Earthquake Engineering (ISEE) had been developed by the researchers in
NCREE as a prototype of Internet-based cooperative structural experimental environment. During the fullscale CFT/BRB frame test, ISEE was applied to allow remote participants witness the real time video
images of the specimen in the laboratory as well as the digital response data through the Internet. The
schematic of the ISEE framework configured for this study is shown in Fig. 5. The approach utilized in
this experiment within the context of ISEE is a platform for Networked Structural Experiments (PNSE)
[2]. The primary purpose of PNSE is to organize the control systems for an Internet testing. The PNSE
focused on the core works directly related to the progression of networked collaborative experiments.
Three types of modules on the PNSE, namely the PNSE server, Command Generation Module (CGM),
and the Facility Control Modules (FCM), are connected. Furthermore, it employs socket operations to
utilize the Transmission Control Protocol / Internet Protocol (TCP/IP) suit to ensure high interoperability
between heterogeneous network and working environments. When the target displacement was

satisfactorily imposed during a pseudo-dynamic test, the measured response data (about 500 channels) of
the test specimen were sent back to the PNSE and stored in the Database for the real-time Internet
distribution. In the mean time, the target displacement for the next time step was calculated in CGM and
then sent to the FCM to command the actuators to move. The web-casting of the digital response data has
been done through an independent Web Server. Meanwhile, the Real Video Server broadcasted the video
images to allow data viewers witness the experimental responses in the laboratory.
ISEE
Internet-based
Simulations on PNSE Server
Platform for Networked
Earthquake
Structural Experiments
Engineering

Database

Data
Viewer

Web Server

FCM

CGM

Facility Control Module

Command Generation
Module

Video Stream
Video Encoder
Server
Static
Video
Data logger
SD

1A- 001

3 actuators

3 actuators

4 actuators

sensor

Fig. 5 ISEE framework employed in the pseudo-dynamic test

EXPERIMENTAL PREPARATION
Inelastic static and dynamic time history analyses were conducted using PISA3D [9] and OpenSees (Open
System for Earthquake Engineering Simulation), developed at National Taiwan University and Pacific
Earthquake Engineering Research Center (PEER), respectively. In the PISA3D model, all BRBs were
modeled using the two-surface plastic (isotropic and kinematic) strain hardening truss element. All the
beam members were modeled using the bi-linear beam-column elements. The CFT columns were modeled
using the three-parameter (stiffness, strength and pinching degradation) degrading beam-column element.
The P-M curves of the CFT columns were constructed using EC4 [10] before multi-linearized. In the
OpenSees model, all the CFT columns and steel beams of the frame were represented using the flexibilitybased nonlinear beam-column element with discretized fiber section model. All the BRBs were modeled
using the truss element with bilinear isotropic strain hardening. A leaning column has been introduced in
both the PISA3D and the OpenSees models in order to simulate the 2nd order effects developed in the
gravity columns. Two ground acceleration records, 1999 Chi-Chi Taiwan and 1989 Loma Prieta
Earthquakes as shown in Fig. 6 scaled to various hazard levels were utilized in the nonlinear time history
analysis. Four earthquake load effects of three different hazard levels, including 50/50 (using record
TCU082), 10/50 (using record LP89g04NS), 2/50 (using record TCU082), another 10/50 (using record
LP89g04NS) were scheduled for the frame test (http://cft-brbf.ncree.gov.tw). Analytical results suggested
that the arrangement of 4, 3 and 3 actuators, each having 980kN force and 50cm stroke capacities, was
appropriate for the 1st, 2nd and 3rd floor, respectively. Each footing was anchored down using four 69mm
diameter post-tension bars to the strong floor. The predicted floor displacement and story shear time
histories were webcasted before, during and after the test (http://cft-brbf.ncree.gov.tw).

Original Ground Acceleration Time History

0.22
0.18
0.14
0.1
0.06
0.02
-0.02
-0.06
-0.1
-0.14
-0.18
-0.22

TCU082EW , PGA = 0.217g

LP89g04NS , PGA = 0.212g

Ground Acceleration (g)

Ground Acceleration (g)

Original Ground Acceleration Time History

0.22
0.18
0.14
0.1
0.06
0.02
-0.02
-0.06
-0.1
-0.14
-0.18
-0.22
0

10

15

20
25
Time (sec)

30

35

40

45

10

15

20
25
Time (sec)

30

35

40

45

Fig. 6 Original ground accelerations used in test (before scaling)

EXPERIMENTAL OBSERVATIONS
As noted above, four earthquake ground accelerations scaled to three different PGAs were planned for the
PDT of the CFT/BRB frame specimen. However, in the Test No. 1, due to the buckling of the gusset plate
occurred at the brace to beam connection in the first story, stiffeners were added at the free edges of all the
gusset plates underneath the floor beams. A total of six PDTs were conducted before the final cyclic
loading test.
Phase-1 Tests

Test No. 1 (50/50, TCU082EW, PGA=0.276, October 3)

The PDT was stopped at the time step of 9.25 seconds without finding any unexpected structural or
nonstructural damage. Some bangs resulted from the slippage of the bolted beam connections were
experienced. However, at the time step of 12.3 second, the out of plane buckling of gusset plate (Photo 1)
at the north BRB-to-beam connection was detected at the first story. It was decided to install stiffeners at
the free edges of the three gusset plates underneath the floor beams as shown in Fig. 7 before further tests.
The north BRB in the first story was removed from the frame and the buckled end was heat straightened.
After the buckled gusset at the first story was straightened, five stiffeners each of 12 mm thick and 200
mm wide were welded to the gusset free edges and the web of the supporting beam (Fig. 7) at all floors.
The straightened top end of the first story north BRB was reversed and connected to the column to
foundation joint. The repair of the gusset, the installation of the stiffeners and the reinstallation of the
north BRB at the first story were completed before October 5. The analytical predictions and the
experimental responses broadcasted through the Internet are given in Fig. 8.

Fig. 7 Added stiffeners and rib after the Test No.1

60

Test No.1
R oof Disp. (m m )

40
20

OpenSees
PISA3D
EXP.

0
-20
-40
-60
0

12

16

Time (sec)

Fig. 8 Roof displacement time history in Phase 1

Test No.1

Photo 1 Out of plane buckling of gusset plate in

Phase 1 Test No.1

Test No. 2 (50/50, TCU082EW, PGA=0.276g, October 5)

Test resumed using the same ground accelerations as that for Test No. 1 in reversed direction on October
5, the analyses predicted the experimental roof displacements satisfactorily as shown in Fig. 9. The
experimental peak story drift reached 0.0052, 0.0043, 0.0038 radians respectively. Some minor flaking of
the white wash near the beam-to-column moment connection was observed. All the six beam-to-column
moment connections appeared to remain essentially elastic. More but still scattered bangs resulted from
the bolt slippages were evident during the test.

Test No. 3 (10/50, LP89g04NS, PGA=0.426g, October 5)

The subsequent PDT continued on the same day using the ground accelerations and the intensity as
planned. Fig. 10 indicates that the analyses satisfactorily predicted the roof responses of the specimen.
During the tests, the concrete slab near the two brace-to-column joints was crushed. More frequent bangs
resulted from the bolt slippage were evident during the test. Some additional but minor flaking of white
wash near the second floor north beam-to-column moment connection was observed.
80

80

Test No.2

OpenSees
PISA3D
EXP.

40

R o o f D isp . (m m )

R o o f D isp . (m m )

Test No.2

-40

-80

10

20

30

Time (sec)

40

-40

-80

50

OpenSees
PISA3D
EXP.

40

10

20

30

40

50

Time (sec)

Fig. 9 Roof displacement time history in Phase 1 Fig. 10 Roof displacement time history in Phase 1
Test No.2
Test No.3

Test No. 4 (2/50, TCU082EW, PGA=0.622g, October 6)

Using the strongest earthquake load effects in this series of the PDT, test No. 4 was started and continued
smoothly as shown in Fig. 11 for the roof displacement time history. At the time step of about 5.0 second,
the bangs started to present. The PDT was stopped at the time step of 12.54 second in order to photograph
the experimental observations. Cracks on the top of concrete foundation near the gusset plate for the south

BRB-to-column joint were observed. Moreover, it was observed that the north BRB on the first floor was
bent slightly near the north BRB-to-column joint where heat straightening had applied after the Test No.1.
It was decided to stop this test and to add the stiffeners on the free edges of the gussets at the two brace-tocolumn base connections (Photo 2) before further testing. One pair of angles was installed bracing the
stiffener to the two anchoring steel blocks in order to prevent further out of plane deformation of the
slightly-bent section of the north BRB.
150

Test No.4
R oo f D isp. (m m )

100

OpenSees
PISA3D
EXP.

50
0
-50
-100
-150
0

12

16

Time (sec)

Fig. 11 Roof displacement time history in Phase 1

Test No.4

Photo 2 Stiffeners at the two brace-to-column

base connections after Phase 1 Test No.4

Test No. 5 (2/50, TCU082EW, PGA=0.622g, October 7)

4000

300

Test No.5

Test No.5

OpenSees
PISA3D
EXP.

200

2000
Load drop
and recover

-2000

R oof Disp. (m m )

Base Shear (kN)

The tests resumed on the Oct. 7 by applying the same earthquake accelerations as that for Test No. 4.
More frequent bangs resulted from the bolt slippage were evident during the test. At the time step of about
27.0 second, a loud but relatively low sound was experienced. In addition, a sudden actuator load drop
was also observed. However, the test was continued with the specimen strength recovered (Fig. 12). The
test was successfully completed and the experimental time history responses were accurately predicted by
the two analytical models (Fig. 13). After the test was completed, it was found that the concrete footing
supporting the first story south BRB to column base was cracked at the corner (Photo 3). It has resulted in
the prestress loss in the tie downs and the subsequent slippage of the footing by 15mm toward north. The
peak story drifts reached about 0.024, 0.02 and 0.012 radians for the first, second and third stories,
respectively.

100

-100

-4000
-200

0
200
Roof Disp. (mm)

-200

10

20

30

40

50

Time (sec)

Fig. 12 Roof Displacement versus base shear in Fig. 13 Roof displacement time history in Phase 1
Phase 1 Test No. 5
Test No.5

200

R o o f D is p . (m m )

Test No.6
OpenSees
PISA3D
EXP.

100

-100

-200

10

20

30

40

50

Time (sec)

Fig. 14 Roof displacement time history in Phase 1 Photo 3 The concrete footing supporting the first
Test No.6
story south BRB to column base was cracked at
the corner after Phase 1 Test No.5

Test No. 6 (10/50, LP89g04NS, PGA=0.426g, October 7)

After un-tighten the four tie downs of the footing, the footing was pull back toward south direction by
11mm using an actuator force of about 1000kN applied at the 2nd floor. A steel cross beam was placed
over the cracked side of the footing before re-applying the prestress in the four tie downs. The test was
resumed and completed with the experimental responses accurately predicted (Fig. 14).
Test No. 7 (Cyclic Increasing Story Drifts from 0.01 to 0.0375 radian, Oct. 8)
After stiffening the gussets and the application of six earthquake load effects, the peak story drift demand
reached 2.5% at the second floor during the Test 4. However, it appears that the beam-to-CFT column
moment connections have undergone relative minor nonlinear deformations and all UBs and BRBs have
not been damaged. High-mode buckling of south BRB (all metal construction) at the first story was
visible. It was decided that cyclic increasing uniform inter-story drifts (4 cycles 0.01, 4 cycles of 0.0125, 4
cycles of 0.02, 4 cycles of 0.025 and 2 cycles of 0.0375 radians) be imposed on the specimen to examine
the failure modes. At the 4th 0.02 radian cycle, the gusset buckled out-of-plane at the brace-to-column
joint for 3BRBS (the south BRB in the 3rd story). The 1BRBS high-mode had also buckled severely. On
the first 0.025 radian cycle, the top end of the two UBs in the 2nd story had been found fractured and
deformed the supporting beam laterally at the gusset to the 3rd floor beam connection. The cyclic tested
was therefore stopped.
Phase-2 Tests :
Repairing the Buckled Gussets and Replacing the BRBs for the Phase-2 Tests
Since the scheduled PDT and cyclic tests were completed with failures only in bracing components
including the BRBs, UBs and the gusset plates, it was decided that Phase-2 tests be conducted after
repairing the damaged components. Phase-2 tests not only allowed to make the best use of the 3-story, 3bay frame but also aimed to investigate the performance of the stiffened gussets plates and the new BRBs.
Before the Phase-2 tests, the laterally buckled gusset under the 3rd floor beam had been removed before
installing a new one as shown in Photo 4 to Photo 6. In addition, stiffeners were welded at the free edges
of the gusset at all the brace to column joints (Photo 7 and Photo 8). Six new BRBs, two all metal double
cored construction for the 1st story and four concrete filled double cored for the 2nd and 3rd stories, have
been installed. The core areas of the new BRBs are 25.9cm2, 21.6cm2 and 12.9cm2 in the 1st, the 2nd and
the 3rd stories respectively. Tests have been successfully conducted on October 31 and November 1.

Photo 4 Buckling of the gusset

under the 3rd floor beam after
Phase 1 tests.

Photo 5 Removal of the gusset

under the 3rd floor beam before
Phase 2 tests.

Photo 6 Added stiffeners and

kickers at the gusset under the 3rd
floor before Phase 2 tests.

Photo 7 Buckling of the gusset at the

Photo 8 Added stiffeners at the free edges of the gusset at
brace to column joint after Phase 1 tests.
the brace to column joint before Phase 2 tests.
Test No. 1 (10/50, TCU082EW, PGA=0.53, Oct. 31)
The roof displacement time history is shown in Fig. 15. The test was stopped at the time step of 8.05
second to allow periodical observations. Some concrete cracks were found near the north brace-to-column
connection on the footing and on the concrete slab of the 2nd floor. At the time step of 22.76 second, minor
cracks were also found on the concrete slab of the 3rd floor around the north brace-to-column connection.
The BRBs performed very well in dissipating energy throughout the PDT. Meanwhile, all the moment
connections were found only yielded without fracture as they were after the Test No.7 in Phase-1 tests.
The maximum drift ratio occurred in the 1st story is about 2.4% radians.
300

Brace Shear (kN)

2000

R oof D isp. (mm )

200
100
0
-100

Phase2, Test No.1

OpenSees
PISA 3D
EXP .

-200
-300
0

10

1000

LP89g04NS (2/50)
Test No.2
0.61g

0
-1000

1FL
-2000
-4000-2000

15

20

25

30

35

Time (sec)

Fig. 15 Roof displacement time history in Phase 2

Test No.1

2000 4000

Story Shear (kN)

Fig. 16 The story shear versus the shear
contributed by the BRB axial force on the 1st
floor in Phase 2 Test No. 2

Test No. 2 (2/50, LP89g04NS, PGA=0.61, Nov. 1)

Throughout the tests, the stiffening of the gussets appeared to be very effective in improving the stability
of the BRBs. Only some concrete slabs were found crushed around the brace-to-column connections. The
BRBs were found well-functioning and dissipating enormous amount of energy. It was found during the
test that the BRBs resisted the majority of the story shear when the story drift was small as shown in Fig.
16. Even when the BRBs have undergone large inelastic deformations, it still resisted about 50% of the
story shear. In this PDT, the maximum drift ratio occurred in the 2nd story is about 2.6% radians.
Throughout the whole test, the structural damage of the CFT/BRB frame was not found severe. The roof
displacement time history is shown in Fig. 17.
300

Roof Disp. (m m)

200
100
0
-100

Phase2, Test No.2

Open Sees
PISA3D
EXP.

-200
-300
0

10

15

20

25

Time (sec)

Fig. 17 Roof displacement time history in Phase 2 Photo 9 Uplift of exterior CFT column in the 1st
Test No.2
cycle of 0.0375 radian in Phase 2 Test No. 3
Test No. 3 (Cyclic increasing the story drifts from 0.0125 to 0.0375 radian, Nov. 1)
Because the BRBs had undergone through two strong earthquake PDTs, and the beams and the columns
had also experienced several earthquake PDTs and the Phase-1 cyclic tests, additional cyclic loading with
increasing uniform inter-story drifts were applied to gain insight of the ultimate failure modes of the
specimen. Unlike the Phase-1 cyclic test, 4 cycles of 0.01 radian were removed. The adopted inter-story
drift history ranged form 0.0125 to 0.0375 radians. At the 4th 0.025 radian cycle, the steel tube of the two
exterior CFT columns were found fractured on the column-to-footing interfaces. As the test went on, the
observed conditions were getting more and more serious. At the 1st cycle of 0.0375 radians, two exterior
CFT columns were completely broken and uplifted from the footings as shown in Photo 9. The test was
finally stopped after the 2nd 0.0375 radian cycle. It was also found that the axial deformation of the north
BRBs in the 1st story was not symmetric to the mid point of the braces. The deformations appeared to
concentrate in only one end of the all-metal BRB in the last few loading cycles.
All the key test results are available through the web site (http://cft-brbf.ncree.gov.tw). Test results have
confirmed that welding the stiffeners at the free edges of the gusset are extremely effective in preventing
the buckling to the gusset.
EXPERIMENTAL RESULTS AND DISCUSSION
Preliminary experimental results are discussed herein. For the further analytical simulations of the test
frame, please refer to the companion paper [11].
Phase-1 tests
In order to obtain the strain gauge reading and axial force relationship of the brace members, there are
eight uniaxial strain gauges arranged in the both ends of each brace. Before the installation of the braces,
the elastic member test was conducted for each UB and BRB. The linear relationship between the axial
forces and strain readings was computed from the elastic member tests, therefore the corresponding axial
force reading can be found during the PDTs immediately. In Test No. 2, braces started to experience
plastic deformation at the first and the second floor. By computing the linear elastic relationships between

the axial force and core displacement of the braces, it is confirmed that the initial stiffness of the BRBs of
the specimen in Table 3 is rather close to the design value. This suggests that the unbonding mechanism
of the BRBs is effective in these BRBs. After the application of six earthquake effects in Phase 1 tests, it
is found that the UBs and BRBs performed rather satisfactorily without evident failure. Figs. 18 to 20
depict the hysteresis of the BRBs and UBs from the first floor to the third floor respectively during Test
No. 5. Furthermore, the BRBs dissipated most of hysterestic energy absorbed by the structure in different
levels of earthquake intensities (Figs. 21-23). In each case, the energy dissipated by the north BRB is
almost the same as that by the south BRBs in each floor suggesting the accurate transformation of the
strain gauge readings into the brace axial forces. In the meantime, the ultimate story drift of the BRB
composite frame was controlled rather effectively under the effects of the design earthquakes. In Test
No.3 (10/50), the story drift in the three stories never reached 0.02 radian (Fig.24). And in Test No. 5
(2/50), the peak story drift of the three stories were smaller than 0.025 radian (Fig.25). The end rotation
of the BRBs in the first story is defined in Fig. 26. The relationship of and the story drift in Test No.5 is
depicted in Fig. 27. It is obvious that the demand of rotation imposed on the brace end is great and seems
to increase with the story drift. It is suggested that this phenomenon should be considered in the BRB
component tests and in BRB connection design. Also, the further researches on the end rotation demand
of BRBs are required.
Table 3 Effective stiffness of BRBs or UBs in Phase-1 tests

Axial Force (103kN)

87.99
181.36
184.32

TCU082EW (2/50)
0.622g

Py

0
-1

Test No.5

1BRBN
-2
-60 -40 -20

20 40 60

Displacement (mm)

87.35
191.10
185.90
2

Axial Force (103kN)

91.49
183
192.96

3BRB
2UBB
1BRB

Analysis
N (kN/mm)
S (kN/mm)

TCU082EW (2/50)
0.622g

Py

0
-1

Test No.5

2UBBN
-2
-60 -40 -20

20 40 60

Displacement (mm)

error (%)

87.35
191.10
185.90
2

Axial Force (103kN)

Experiment
N (kN/mm)
S (kN/mm)

4.7
4.2
3.8

0.7
5.1
0.8

TCU082EW (2/50)
0.622g

Py

0
-1

Test No.5

3BRBN
-2
-60 -40 -20

20 40 60

Displacement (mm)

Fig. 18 Hysteresis of 1BRBN in

Phase 1 Test No. 5

Fig. 19 Hysteresis of 2UBBN in

Phase 1 Test No. 5

Fig. 20 Hysteresis of 3BRBN in

Phase 1 Test No. 5

Fig. 21 Dissipated energy in

Phase 1 Test No. 2 (50/50)

Fig. 22 Dissipated energy in

Phase 1 Test No. 3 (10/50)

Fig. 23 Dissipated energy in

Phase 1 Test No. 5 (2/50)

Level 10/50
PGA=0.426g

Level 2/50
PGA=0.622g
2

EXP
PISA3D
OpenSees

Floor

Floor

EXP
PISA3D
OpenSees

0
-2 -1.5 -1 -0.5 0

0.5

1.5

0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5

Story Drift (radian %)

Fig. 24 Peak story drift in Phase 1 Test No. 3
(10/50)

Story Drift (radian %)

Fig. 25 Peak story drift in Phase 1 Test No. 5
(2/50)

Phase-2 tests
The BRBs performed excellently in dissipating energy in both Test No. 1(10/50) and Test No. 2(2/50).
They dissipated most of the energy in each story as depicted in Fig. 28 and 29. The similar phenomenon
observed in Phase-1 tests and Phase-2 tests is that the energy dissipated by the BRBs in the lower story
accounts for more portion of the energy dissipated by the corresponding story. With the satisfying seismic
performance of the BRBs, the story drift were well controlled during the tests. The peak story drift in
Phase-2 tests is shown in Fig. 30. The fact that the positive peak story drift in Test No. 1(10/50) is slightly
over the designed value might be resulted from the abrupt and drastic responses of the Chi-Chi earthquake
record. Nevertheless, the peak story drift in the event of LP2/50 shows great agreement with the designed
value.

N-Brace
S-Brace
Braces
Story

GPA=0.53g

500

0
1FL

2FL

3FL

Fig. 28 Dissipated energy in

Phase 2 Test No. 1 (10/50)

N-Brace
S-Brace
Braces
Story

Test No.2

1500

LP89g04NS (2/50)
GPA=0.62g

1000

Phase II

Floor

Test No.1

1000

2000

TCU082EW (10/50)

Dissipated Energy (k N-m)

Dissipated Energy (kN-m)

1500

CC10/50
LP2/50

500
0
1FL

2FL

3FL

Fig. 29 Dissipated energy in

Phase 2 Test No. 2 (2/50)

0
-3

-2
-1
0
1
2
Story Drift (radian %)

Fig. 30 Peak story drift

in Phase 2 tests

Top End Rotation (radian)

Top End

0.03

0.015

TCU082EW (2/50) 1BRB

Test No.5
GPA=0.622g

-0.015

Rotation

-0.03
-0.03 -0.015

0.015

0.03

Story Drift (radian)

Fig. 26 Definition of top brace end rotation

Fig. 27 Top end rotation vs. Story drift in Test No. 5

CONCLUSIONS
Based on the test and analytical results, summary and conclusions are made as follows:

Test results confirm that the global dynamic responses of the 3-story 3-bay CFT-BRB frame specimen
can be satisfactorily predicted using both PISA3D and OpenSees.

The peak story drift reached 0.025 radian in Phase 1 tests after applying the 2/50 design earthquake on
the specimen. It appears that the DSD procedure adopted in the design of the specimen is effective in
limiting the ultimate story drift under the effects of the design earthquake.

CFT/BRBF performed extremely well after the application of six earthquake load effects. Also,
according to the free vibration tests conducted after each earthquake pseudo dynamic test, the stiffness
and damping of the specimen only changed slightly.

Stiffeners added along the free edges of the gusset plate are effective in preventing out-of-plane
instability of the brace-to-column connections. However, it also introduces flexural demands on the
BRBs. Further researches are required to study the BRB end connections.

All the moment connections survived all the Phase-1 and Phase-2 tests without failure as shown in
Photo 10. The BRBs effectively control the story drift and reduce the nonlinear demand imposed on
these moment connections.

Tests confirmed that the PNSE architecture implemented for the ISEE is very effective in
disseminating real time test results through the Internet.

Photo 10 Yield of moment connections after PDTs

ACKNOWLEDGEMENTS
The National Science Council of Taiwan provided the financial support for this experimental research
program. Nippon Steel Company donated two unbonded braces which have been installed in the 2nd floor
of the frame specimen. The laboratory supports provided by the NCREE are very much appreciated. In
particular, Mr. Kung-Juin Wang successfully implemented and executed the pseudo dynamic test
software. Mr. Wei-Chung Cheng constructed the web servers for on-line data and video display. Also,
valuable suggestions provided by many US, Japanese and Taiwanese researchers on this joint effort are
gratefully acknowledged. The detailed list of participants is given in the web site: http://cftbrbf.ncree.gov.tw

REFERENCES
1.

2.

3.

4.
5.

6.

7.

8.
9.

10.
11

Chen CH, Cordova P, Lai WC, Deierlein GG, Tsai KC, Pseudo-Dynamic Test of a Full-Scale RCS
Frame: Part 1- Design, Construction and Testing, Proceedings of the 13th World Conference on
Earthquake Engineering, Vancouver, Canada. Paper no. 2178, 2004.
Wang SJ, Wang KJ, Yang YS, Cheng WC, Yeh CC, Tsai KC. Networked Pseudo-dynamic Testing
Part II : Application Protocol Approach, Proceedings of the 13th World Conference on Earthquake
Engineering, Vancouver, Canada. Paper no. 1548, 2004.
Loeding S, Kowalsky MJ, Priestley MJN, Displacement-based Design Methodology Applied to
R.C. Building Frames, Structural Systems Research Report SSRP 98/08, Structures Division,
University of California, San Diego, 1998, 269 pp.
Medhekar MS, Kennedy DJL, Displacement-Based Seismic Design of Buildings-Theory,
Engineering Structures, Vol.22, No.3, 2000, 201-209.
Tsai KC, Weng YT, Lin SL, Goel S, Pseudo-Dynamic Test of a Full-Scale CFT/BRB Frame: Part
1 Performance Based Specimen Design, Proceedings of the 13th World Conference on
Earthquake Engineering, Vancouver, Canada. Paper no. 750, 2004.
Weng YT, A Study of Multi-mode Seismic Performance-based Evaluation and Displacementbased Design Procedures, Ph.D. Thesis, Supervised by Prof. Keh-Chyuan Tsai, Department of
Civil Engineering, National Taiwan University, Taipei, Taiwan 2003.
Tsai KC, Hwang YC, Weng CS, Shirai T, Nakamura H, Experimental Tests of Large Scale
Buckling Restrained Braces and Frames, Proceedings, Passive Control Symposium 2002, Tokyo
Institute of Technology, Tokyo, December 2002.
Tsai KC, Lin SL, A Study of All Metal and Detachable Buckling Restrained Braces, Center for
Earthquake Engineering Research, National Taiwan University, Report No. CEER/R92-08, 2003.
Tsai KC, Lin BC, User Manual for the Platform and Visualization of Inelastic Structural Analysis
of 3D Systems PISA3D and VISA3D, Center for Earthquake Engineering Research, National
Taiwan University, Report No. CEER/R92-07, 2003.
CEN, Design of Composite Steel and Concrete Structures-Part 1-1: General Rules and Rules for
Buildings, EuroCommittee for Standardization, Brussels, Belguim, 1992.
Lin ML, Weng YT, Tsai KC, Hsiao PC, Chen CH, Lai JW, Pseudo-Dynamic Test of a Full-Scale
CFT/BRB Frame: Part 3 Analysis and Performance Evaluation, Proceedings of the 13th World
Conference on Earthquake Engineering, Vancouver, Canada. Paper no. 2173, 2004.