EARTHQUAKE ENGINEERING & STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. (2013)

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/eqe.2374

Application of hybrid-simulation to fragility assessment of the

telescoping self-centering energy dissipative bracing system

Viswanath Kammula*,†, Jeffrey Erochko, Oh-Sung Kwon and Constantin Christopoulos

Department of Civil Engineering, University of Toronto, Toronto, Ontario, Canada

SUMMARY
Substructure hybrid simulation has been the subject of numerous investigations in recent years. The
simulation method allows for the assessment of the seismic performance of structures by representing
critical components with physical specimens and the rest of the structure with numerical models. In this
study the system level performance of a six-storey structure with telescoping self-centering energy dissipa-
tive (T-SCED) braces is validated through pseudo-dynamic (PsD) hybrid simulation. Fragility curves are
derived for the T-SCED system. This paper presents the configuration of the hybrid simulation, the newly
developed control software for PsD hybrid simulation, which can integrate generic hydraulic actuators into
PsD hybrid simulation, and the seismic performance of a structure equipped with T-SCED braces. The
experimental results show that the six-storey structure with T-SCED braces satisfies performance limits
specified in ASCE 41. Copyright © 2013 John Wiley & Sons, Ltd.

Received 16 January 2013; Revised 3 September 2013; Accepted 5 September 2013

KEY WORDS: hybrid simulation; self-centering systems; performance evaluation

1. INTRODUCTION

Earthquakes are paramount among the natural hazards impacting civil infrastructure. The past several
decades have witnessed a series of costly and damaging earthquakes. In order to be prepared for such
natural disasters, it is important to reasonably estimate, predict and mitigate the damage. The seismic
vulnerability curve or seismic fragility curve is one of the most important elements in the assessment of
damage and loss from earthquake events. The fragility curves represent the probabilities of structural
damage because of earthquakes as a function of ground motion intensities such as peak ground
acceleration (PGA), spectral acceleration (Sa) and spectral displacement (Sd). The fragility curves
vary with the type of structural system, characterization of earthquake excitation and the
performance levels chosen.
Buildings are designed to achieve a certain performance level depending on the hazard level.
However, the economic losses incurred in recent earthquakes in buildings have prompted an interest
in their performance above and beyond the traditional goal of life safety. Recent trends in
earthquake engineering have shown a shift from basic life safety performance level to more refined
performance levels. In performance based seismic design, these performance levels are evaluated
with corresponding anticipated levels of seismic hazard, and buildings are designed to satisfy
multiple performance objectives depending on the severity of the seismic excitation. For example,
the safety critical performance objective [1] mandates that the building should be fully operational
after a frequent earthquake and should not collapse even because of a rare seismic event with a

*Correspondence to: Viswanath Kammula, Department of Civil Engineering, University of Toronto, Toronto, Ontario,
Canada
†
E-mail: [email protected]

Copyright © 2013 John Wiley & Sons, Ltd.

 V. KAMMULA ET AL.

return period of 2500 years. Even though buildings are designed to meet multiple performance criteria,
they are designed to behave in the inelastic range during the design level earthquake acknowledging
the economic disadvantages of designing buildings to sustain earthquakes elastically. Therefore,
even though a structure designed using a performance based seismic design approach may not
collapse and will protect the lives of occupants, it is expected to undergo large inelastic cycles and
will likely have residual deformations after a design level earthquake event [2]. It has also been
found that above a certain level of residual drift (approximately 0.5%), it may not be economical to
repair a structure, and the structure would have to be demolished and replaced [3]. For ductile steel
structures, it has been shown that residual deformations are expected to exceed this threshold even
under design level earthquakes [4]. In order to mitigate the residual deformations and achieve higher
performance levels, self-centering systems have been developed in the past decade.
Self-centering systems allow for rapid recovery after major earthquake events by dissipating seismic
energy and containing a mechanism that restores a structure to its original zero-drift configuration even
after experiencing large inelastic cycles. Several self-centering systems have been developed in recent
years for application to building or bridge structures such as, for example, post-tensioned steel
moment frames [5–10], self-centering systems with shape memory alloys (SMA) [11–14], self-
centering braces [15–20], controlled rocking systems ([21–27]), and self-centering precast segmental
bridge bents ([28, 29]). The self-centering capability of these systems reduces or eliminates the
residual deformations after a seismic event, which facilitates the reoccupation of the facility
immediately and minimises both the monetary and logistic losses. Among these, the self-centering
energy dissipative (SCED) bracing system [18, 19] is a new self-centering system that has the added
advantage of being able to be used in the same way as other cross-braces such as the buckling-
restrained brace, avoiding the need for special connection details. Several studies have been
carried out in the past to investigate the performance of the novel SCED structural system
([18, 19, 30–33]). The objective of this study is to derive seismic fragility curves for the newly
developed SCED braced structural system.
To derive these fragility curves, hybrid tests were conducted on a full-scale single-storey steel frame
containing a new type of SCED brace called the telescoping SCED (T-SCED) brace. This T-SCED
brace is an extension of the SCED brace concept that allows for double the elongation capacity of
the original SCED brace [30, 31]. A detailed description of the design and behaviour of this brace
may be found in Erochko et al. [30, 31]; however, the design and behaviour of the T-SCED brace
will also be summarised briefly in a future section. The SCED brace concept has also been recently
extended by Chou and Chen [34], who have developed a similar system that relies on a dual-core
mechanism to extend the deformation capacity of a SCED brace. Because the T-SCED brace is
approximately functionally equivalent to the SCED brace from a behavioural standpoint, in this
study, it has been used to conduct a fragility analysis that applies to both SCED and T-SCED braces.
In general, fragility curves are derived based on empirical data from past earthquakes [35, 36],
expert’s opinions and analytical simulations [37–39]. Though vulnerability curves based on
observed damage from past earthquakes are the most realistic, they lack in generality that are
applicable to limited regions and often lack adequate data. Fragility curves based on expert’s
opinion include a large uncertainty as the damage assessment is based on the subjective observation
of field investigators and their seismic experience. The derivation of analytical fragility curve
requires a realistic analytical model of every component of structural system. So the fragility curve
accuracy would be compromised, if any of the component’s behaviour is not accurately modelled or
not realistic. To overcome this limitation of purely analytical derivations, in this study, fragility
curves are derived using hybrid (analytical-experimental) simulation.
The hybrid (analytical-experimental) simulation is carried out based on a step-by-step numerical
integration method to solve the governing equations of motion for a numerical structural model that
is formulated considering both the analytical and physical components of a structural system.
Contrary to a purely numerical simulation, the hybrid simulation method combines the physical
testing of a part of the structure (a substructure), with computer models of the remainder of the
structure. It provides complete picture of how earthquake events can affect large structures such as
buildings and bridges without having to physically test the entire structure. This enables civil
engineers to accurately and efficiently capture the effects that a substructure has on the overall

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

structure, while subjecting the substructure to the same forces and motions it would experience within a
complete structure. During the simulation, the physical substructures of the overall hybrid model are
tested in one or more laboratories using computer-controlled actuators, whereas the numerical
substructures are simultaneously analysed on one or more analytical platforms. Several researchers
have deployed hybrid simulation as a tool to assess the seismic performance of a variety of
structural systems. Some of the applications include, assessing the seismic performance of self-
centering moment-resisting frame and concentrically-braced frames [40]; performance evaluation of
zipper-braced steel frame [41, 42]; large-scale experimental verification of magnetorheological
damper [43]; performance assessment of a steel moment-resisting frames equipped with compressed
elastomeric dampers [44] and large-scale testing of a controlled rocking steel braced frames [25]. In
addition, some researchers have also used hybrid simulation to perform seismic fragility analysis [45].
This paper provides details of the development of fragility curves for the novel T-SCED bracing
system using the hybrid simulation technique. The following sections provide an overview on the
reference structure, the establishment of the hybrid simulation framework and the formulation of the
hybrid model. Then, the ground motions used in hybrid simulations and the simulation results are
discussed. The paper concludes with a summary of findings.

2. REFERENCE STRUCTURE

A regular six-storey office building that was previously studied numerically [32] was chosen for this
study. The building was assumed to be located in downtown Los Angeles, California where seismic
loads are expected to govern the design of the lateral force resisting system. The plan of the office
building consists of three 9 m bays in the north-south direction and five 9 m bays in the east-west
direction. The lateral force-resisting system in the east west (E-W) direction consists of two special
steel moment-resisting frames with three bays each (along the north and south edges of the
building). The lateral force-resisting system in the north-south (N-S) direction consists of two SCED
braced frames located in the center bays of the north-south frames as shown in Figure 1. The design
of these SCED braces is based on the T-SCED [30, 31] concept that will be explained briefly in the
Section 3. The building was designed according to ASCE 7–05 [46]. The plan and elevation of the
reference structure are shown in Figure 1. The design response spectrum of the building site is
shown in Figure 2. A response modification coefficient (R) factor of 7 was used for the SCED
system, as it was observed from the previous studies that a buckling restrained brace (BRB) and a
SCED designed for similar strengths have showed similar peak drifts [33].
An analytical model of one of the braced frames in the N-S direction, shown in Figure 3, was
developed in OpenSees [47]. Therefore, one half of the weight of the building was included in the
model. The masses were assumed to be lumped at the floor levels. The model includes all the
beams, columns and the braces that are part of the lateral force resisting system. In addition, leaning

Gravity columns
5@9144mm

Braced
frame
6@3658mm
3@9144mm

Moment resisting frame

9144 mm
(a) Plan (b) Elevation

Figure 1. Elevation of the six-storey self-centering energy dissipative structure.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

1.5

Spectral acceleration (g)

1

0.5

0
0 1 2 3 4

Figure 2. The design spectrum.

Master Nodes Rigid diaphragm constraint: the x-dir displacement of all

nodes at each floor are restrained to the Master node
71 73 74 75 76 77

61 63 64 65 66 67

51 53 54 55 56 57

41 43 44 45 46 47

31 33 34 35 36 37

21 23 24 25 26 27

Figure 3. Two-dimension analytical model of the six-storey self-centering energy dissipative braced frame.

columns were modelled to account for the P-Delta effects. These leaning columns were loaded
vertically with all of the dead and live loads that were not directly applied to the lateral load
resisting frame in the building. Because the same numerical model was used in the analytical
substructure of the hybrid model, described in a later section, P-Delta effects are also considered in
the hybrid simulations. The leaning columns were connected to the main frame by using a rigid
diaphragm constraint. The rigid diaphragm was modelled by constraining the horizontal
displacement of all nodes for a given floor to the first node of that floor, as shown in Figure 3.
Because the beam-column connections are pinned, as described in the next paragraph, the rigid
diaphragm assumption does not affect the global load resisting behaviour of the structure.
A rigid diaphragm assumption was also used to model the building floors to ensure the rigid body
translation of the floors. All columns were continuous and were pinned at the base. Hence, the
braces were the primary lateral force resisting elements. Column splices were modelled at every 2.5
storeys as rigid moment connections. Beam-to-column connections were modelled as pinned
connection. Hence, the beam elements were modelled as truss elements, as they contribute little to
the lateral resistance. All members were assigned rigid end offsets as applicable such that the length
of the offset at each end of the beam or column equals to half of the depth of the member that each
beam or column connects to. A bilinear hysteretic material model was assigned for the beam and
column members. The T-SCED braces were modelled using truss elements with the flag-shaped

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

self-centering material model, as shown in Figure 4. The 2D-model of the six-storey T-SCED structure
along with the node and element numbers is shown in Figure 3. For this steel building, inherent
damping was modelled using Rayleigh damping of 3% critical in modes 1 and 2.

3. TELESCOPING SCED

The original SCED brace, whose mechanics was explained in Christopoulos et al. [18, 19], consists of
two rigid longitudinal members, pretensioned tendons and energy dissipative mechanism comprising
either a friction device, viscous device, shape memory alloys, any combination of these or any source
of energy dissipation. The brace is assembled such that the tendons are in tension regardless of the
direction of motion resulting in an inherent self-centering capability and in a symmetric flag-shaped
hysteretic response, as shown in Figure 4. The deformation range over which the self-centering
capability of the SCED brace can be observed is limited by the length of the brace and by the choice
of the material used in the pretensioned tendons. For typical building storey geometries, this
deformation range generally corresponds to storey drifts of ±1.5% to ±2.0% times the storey height
[30, 31]. To further improve the deformation range over which the self-centering capability of the
SCED brace can be observed for high seismic applications, the SCED concept was extended by using
a T-SCED configuration that doubles the brace deformation over which full self-centering response
can be achieved [30, 31, 48]. This novel telescoping configuration has permitted storey drifts as high
as 4%, without any notable deterioration to the strength, stiffness and energy dissipation capacity.
The ultimate limit states of SCED braces are typically controlled by either rupture of the tendon or
connection failure, considering that the elements of the SCED brace are a capacity designed for the
expected maximum forces. Even though none of the T-SCEDs were tested up to these ultimate limit
states, it is worthwhile to note that any comparable conventional bracing member such as a BRB
would typically suffer significant damage and potentially the loss of all capacity at a similar drift
level (or even at a lower drift levels). The T-SCED would experience a more gradual rupture as the
tendons and friction interfaces are gradually broken and it would likely retain some load capacity
past 4% drift. This large deformation capacity with self-centering ability has enabled the
performance of hybrid simulations over a wide range of seismic inputs including those representing
the maximum credible earthquake (MCE) hazard level.

4. HYBRID SIMULATION

Because of the practical benefits of hybrid (numerical-experimental) simulation, the seismic

performance assessment of structures using the method has been actively used in recent years. The

Figure 4. Symmetric flag-shaped hysteresis.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

procedure for hybrid simulation involves the following: selection of a suitable simulation platform,
which can communicate and integrate the results from different components of the substructured
hybrid model; developing the substructured model of the structure comprising of both the physical
and experimental modules; and establishing proper communication among all the modules. UI-
SimCor [49, 50] was used as the main hybrid simulation platform, for these tests. UI-SimCor runs
the time integration scheme of the system. In UI-SimCor, all dynamic degrees of freedom and the
degrees of freedom (DOFs) of the nodes at the interface of substructures are included. These nodes
at the interface are termed as control nodes and the DOFs that are controlled for each node are terms
control DOFs in UI-SimCor. The program is capable of communicating with multiple software
packages including OpenSees, laboratory modules and data/image acquisition devices through
Ethernet network. The dynamic analysis is carried out in UI-SimCor where as all the experimental
and analytical components perform static experiment/analysis by imposing the target displacement
command received from the UI-SimCor, either using the actuator or by imposing the displacement
analytically. The target displacement is predicted with the α-Operator Splitting integration
scheme [51] based on measured displacements and forces at the previous step. The static analysis
and data/image acquisition modules are referred to as restoring force and auxiliary modules,
respectively. The complete capabilities of UI-SimCor with detailed examples can be found in Kwon
and Elnashai [50].
The second step in the preparation of the hybrid simulation was to develop the substructured model.
The substructured model for the present tests consisted of two restoring force modules. One of the two
restoring modules was the experimental module in which the first storey brace was represented as the
physical component. The second was the analytical module in which the remainder of the structure was
modelled. For the development of the hybrid model, one node per floor was assumed as the control
point. Each node has three DOFs, of which only the translational DOFs are controlled during the
experiment. Hence, each node has two control DOFs, one vertical DOF and one horizontal DOF.
This assumption is reasonable because all of the nodes for a given floor will have the same
horizontal displacement due to the rigid diaphragm constraint, and there is no moment transfer
because all the beam-column and brace-column connections are assumed to be pinned. The
rotational DOF in the analytical substructure accounts for moment transfer and potential hinging in
the columns. Hence, there are a total of six control point nodes and twelve control DOFs, one node
per floor and two DOFs per node. As previously mentioned, the masses were modelled in UI-
SimCor and were assigned to the control point nodes as shown in Figure 5.
The experimental substructure has one control point node with two control DOFs when the bottom
storey was represented as the physical specimen and two control point nodes with four control DOFs
when any other storey brace was represented as the physical specimen shown in Figure 5c. The

Modelled Not Modelled

a) Whole Model b) Masses in UI-SimCor c) Experimental d) Analytical

Figure 5. Hybrid model with different components when fifth storey was used as the physical specimen.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

analytical substructure in both cases has six control point nodes with eighteen DOFs. Because the
masses were assigned at the control point node, the DOFs associated with control point nodes would
also be dynamic DOFs. Hence, the UI-SimCor has twelve dynamic DOFs shown in Figure 5b.
The last step in the preparation of the hybrid simulation test was to establish flawless communication
among different components. Communication between UI-SimCor to OpenSees was established with
Network Interface for Console Applications. Network Interface for Console Applications is an
interface program developed for hybrid simulation with UI-SimCor. It can provide network interface
for OpenSees, Zeus-NL, Abaqus and other user-developed analysis applications. The laboratory
specimen was controlled through MTS controllers, which were not capable of receiving commands
over the network. So, to facilitate communication with the controller, hardware from National
Instruments was acquired, and a script was developed in LabVIEW that could receive a command
through the network from UI-SimCor, communicate the command to the MTS controller through
analogue voltage and then feedback both the measured forces and displacements to UI-SimCor. This
script is hereafter referred to as Network Interface for Controllers (NICON). Some of the important
functionalities of NICON included communication with UI-SimCor, ramp generation, displacement
limit and force limit checks, accommodation of any initial displacement and force offsets, and the
capability to perform coordinate transformations. Because the MTS controller controlled the laboratory
actuators in real time based on voltage signals from NICON, it was imperative to verify the
functionality of each component thoroughly to avoid any unexpected movement of the actuators. A
schematic showing the communication is given in Figure 6.
A picture of the experimental module is shown in Figure 7. A single-storey vertical steel frame with
a T-SCED brace was assembled. The two vertical columns were pinned at the base. The beam-to-
column connection was a bolted connection with slotted holes. The frame was controlled using two

V
NICON Interface O Laboratory
L
Command U Command
T MTS MTS
LABVIEW S NI Hardware Specimen
A Controller Actuator
Measurement B Measurement
G
E
NET
UI-SimCor WORK Restoring Force Module 1 (Experimental Module)
(Main Hybrid
Simulation Command
Platform) Measurement
NICA OpenSees
Restoring Force Module 2 (Analytical Module )

Command
ACTIA Webcams, Digital
Auxiliary Module 1
cameras

Figure 6. Communication flow among components of the hybrid simulation framework.

a) Picture of the Frame setup b) 3-D conceptual view of the setup (Erochko et al., 2012)

Figure 7. Illustration of the experimental setup.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

dynamic actuators mounted to the reaction wall. The laboratory set up was capable of controlling only
one DOF. However, as discussed in previous paragraphs, the experimental module has four control
DOFs, hence a DOF transformation was essential in order to carry out the hybrid simulation.
The experimental substructure of Figure 5c can be redrawn conveniently as shown in Figure 8. The
entire transformation procedure is illustrated in Figure 8. The ‘x’ and ‘y’ represents the global axes of
the member, and the ‘u’ represents the local (diagonal) axis of the member. Hence, the four global
coordinates (x1, y1, x2 and y2) are converted into two local coordinates (u1 and u2) using Equation
(1). For convenience, the command values received from UI-SimCor are subscripted with ‘rcvd’ and
the measured values sent to UI-SimCor are subscripted with ‘msd’.

x1rcvd
2 3
" #
u1rcvd λx λy 0 0 66 y1rcvd

 7
¼ (1)
7
6 7
u2rcvd 0 0 λx λy 4 x2rcvd 5
y2rcvd
where λx = cos θ and λy = sin θ and θ is the angle of inclination of the member
The lower node was assumed to be fixed and the displacement difference Δu′rcvd was applied at the
upper DOF, Equation (2).

Δurcvd ¼ u2rcvd u1rcvd

′
(2)
and Δu rcvd ¼ Δurcvd =λx

The actual displacement, the specimen experienced Δu′msd and the corresponding force ΔF′msd
were measured. The measured values were transformed back to the global coordinates as in the
succeeding text.
Δumsd ¼ Δu′ msd *λx
(3)
and ΔFmsd ¼ ΔF′ msd =λx

u1msd u1rcvd Fu1msd ΔFmsd





¼ and ¼ (4)
u2msd u1rcvd þ Δumsd Fu2msd ΔFmsd

Figure 8. Illustration of the coordinate transformation procedure.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

2   3
u1msd λy *y1rcvd
x1msd Fx1msd λx *Fu1msd
2 3 2 3 2 3
λx
6 7
6 7
y1msd 7 6 y1rcvd 7 6 λy *Fu1msd
Fy1msd 7
6 7 7 6 6 7
7¼6  7and6 7¼6 (5)
6 6 7 6 7
6  7
4 x2msd 5 6
6 u2msd λy *y2rcvd 7 7
4 Fx2msd 5 4 λx *Fu2msd 5
y2msd 4 λx 5 Fy2msd λy *Fu2msd
y2rcvd

The final transformed values in Equation (5) were used as the feedback to the UI-SimCor. This
entire transformation procedure was implemented in the interface program NICON to facilitate the
hybrid simulation. However, if the test setup allowed for multi-DOF, actuating several DOFs could
be controlled simultaneously.
Preliminary tests were performed to estimate the frame’s frictional resistance and the overall
slackness in the frame. It was found that the frictional resistance that was exterior to the brace itself
was less than 5% of the brace’s activation load [48]. Because this amount is small and a real system
would also likely to contain some minor resistance due to friction, this additional resistance was
neglected; however, the overall slackness in the system cannot be neglected. The slackness is
primarily due to the slotted connections at the beam-to-column. Because of this slackness, the brace
would not move as much as the frame. In other words, the received target command (TGD) from
the UI-SimCor was achieved in the frame but not in the brace. Hence, an external feedback loop
was implemented in NICON such that at every time step iterations were performed until the TGD
was achieved in the brace.

Figure 9. Flowchart illustrating iterative mechanism.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 Copyright © 2013 John Wiley & Sons, Ltd.

Table I. Ground motions used for experimental hybrid simulation.

Duration Anticipated drift
Test no. Excitation name Earthquake name Year Station name PGA (g) Scale factor (seconds) demand (%)
1 A-ING000 Whittier Narrows-01 1987 Inglewood—Union Oil 0.29 1.0 11.78 0.13
2 TCU071-N Chi-Chi, Taiwan-03 1999 TCU071 0.38 1.0 14.95 0.11
3 PEL180 San Fernando 1971 LA—Hollywood Store FF 0.17 1.0 12.99 0.25
4 LAC180 Northridge-01 1994 LA—City Terrace 0.32 1.0 23.70 0.39
5 A-BRD130 Whittier Narrows-01 1987 Brea Dam (Downstream) 0.31 1.0 10.62 0.31
6 AND250 Loma Prieta 1989 Anderson Dam 0.24 1.0 14.70 0.31
7 CHY014W Chi-Chi, Taiwan 1999 CHY014 0.23 1.0 40.10 0.48
8 PEL360 Northridge-01 1994 LA—Hollywood Store FF 0.36 1.0 14.84 0.58
9 ORR291 San Fernando 1971 Castaic—Old Ridge Route 0.27 1.0 16.43 0.67

V. KAMMULA ET AL.
10 ATAR090 Whittier Narrows-01 1987 Tarzana—Cedar Hill 0.64 1.1 11.14 0.62
11 G03090 Loma Prieta 1989 Gilroy Array #3 0.37 1.0 15.29 0.84
12 HDA165 Loma Prieta 1989 Hollister Diff. Array 0.27 0.8 18.48 0.8
13 HDA255 Loma Prieta 1989 Hollister Diff. 0.28 1.0 19.85 0.83
14 CHY080N Chi-Chi, Taiwan-03 1999 CHY080 0.22 1.0 13.17 0.76
15 CH10270 Loma Prieta 1989 Oakland—Outer Harbor Wharf 0.27 0.65 17.78 0.85
16 H-Z14090 Coalinga-01 1983 Parkfield—Zone 14 0.27 0.98 20.63 0.88
17 ORR090 Northridge-01 1994 Castaic—Old Ridge Route 0.57 1.0 15.00 1.45
18 HEC090 Hector Mine 1999 Hector 0.34 1.1 16.50 1.31
19 A01090 Loma Prieta 1989 Foster City—APEEL 1 0.30 1.12 20.00 1.3
20 G02090 Loma Prieta 1989 Gilroy Array #2 0.37 1.1 10.00 1.4
Earthquake Engng Struct. Dyn. (2013)

21 ORR360 Northridge-01 1994 Castaic—Old Ridge Route 0.41 0.8 20.00 1.1
22 ORR360 Northridge-01 1994 Castaic—Old Ridge Route 0.46 0.9 20.00 1.76
23 TAR360 Northridge-01 1994 Tarzana—Cedar Hill A 0.99 0.77 25.00 1.78
24 CHY028-E Chi-Chi, Taiwan 1999 CHY028 0.65 0.9 25.00 1.79
25 H-PVY045 Coalinga-01 1983 Pleasant Valley P.P.—yard 0.59 0.9 15.00 1.75
26 CHY028-E Chi-Chi, Taiwan 1999 CHY028 0.65 0.9 25.00 1.79
27 TAR090 Northridge-01 1994 Tarzana—Cedar Hill A 1.78 0.75 20.00 1.81
28 CHY028-N Chi-Chi, Taiwan 1999 CHY028 0.79 0.81 22.00 1.83
DOI: 10.1002/eqe

29 PET090 Cape Mendocino 1992 Petrolia 0.66 0.71 13.00 1.88

30 TCU074-E Chi-Chi, Taiwan 1999 TCU074 0.59 0.82 30.00 1.85
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

To facilitate this, in addition to the displacement measurement from actuators, the actual
deformation of brace, measured with linear variable differential transformer (LVDT), was also
taken and their absolute difference was calculated. Then, for each simulation step, that is, for
every new command received from UI-SimCor, depending on whether the difference was below
the allowable limit (0.1 mm), the frame was pushed further until the TGD was achieved in
the brace. However, because controlling based on the brace LVDT was perceived not to be the
safest method, additional checks to capture any LVDT malfunction were implemented. The
maximum slackness (smax) was quantified to be 6 mm, based on the initial tests. This was used
to limit the additional movement required in the frame and also to capture any LVDT
malfunction. This check was implemented at two levels, the first before calculating the command
for the next iteration and the second during the ramp stage. These additional checks helped to
prevent any unwanted movement of the actuators. The scheme is illustrated as a flowchart in
Figure 9.
For any ith simulation step, assuming n iterations are required to achieve the TGD in the brace and
the command of every iteration is executed in z increments of sinusoidal ramp, the following terms are
defined, which are used in the flow chart of Figure 9.

CmdActi;n is the Command to Actuator

TGDi = Δu′rcvd,i (as in Equation (2)) (6)
ΔTGDi ¼ TGDi TGDi 1

Hybrid simulations were carried out after verifying that every component of the simulation
framework and every functionality of NICON were working properly. Because the SCED brace
does not sustain any notable strength and stiffness degradation within a large range of
displacements, thirty hybrid simulation tests were performed on the same specimen each with a
different ground excitation.

5. GROUND MOTIONS

A total of 30 far-field ground motions were selected from the PEER NGA database [52]. Table I lists
the details of the ground motions that were used in the simulations to derive the fragility curves. The
PGA range for the final selected suite of ground motions, listed in Table I, is 0.2 g to 1.8 g with an
average of 0.5 g. Based on the preliminary analytical simulation results, the ground motions were
scaled to represent a wide range of intensities. Because the peak storey drifts are of primary
interest, the experimental simulation time was reduced by stopping the ground motion after the

2.25 - 2.5
Spectral Acceleration (g)

MCE Level
2.0 - 2.25
1.75 - 2.0
1.5 - 1.75
1.25 - 1.5 DBE Level

1.0 - 1.25
0.75 - 1.0
0.5 - 0.75
0.25 - 0.5
0.0 - 0.25
0 1 2 3 4 5 6
Frequency Count

Figure 10. Frequency of the ground motion for different spectral acceleration bins.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

peak drift had been reached. The experimental duration for each test is shown Table I. The
anticipated drift shown in the last column of Table I, is the expected drift demand on the
experimental specimen obtained from the analytical simulation.
The spectral accelerations of the selected ground motions at the fundamental period of the structure
are shown in Figure 10. The spectral accelerations are distributed from low intensity earthquake events
up to the MCE level, as shown in Figure 10 . In addition, this variation allows an extension of the
results to similar structures with slightly different fundamental periods.

Figure 11. Comparison of responses from hybrid simulation and analytical prediction.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

6. SIMULATION RESULTS

Simulation results from one of the hybrid simulation tests shown in Table I, Test case 27, are
presented in Figure 11. The scaled PGA of the applied excitation was 1.33 g, which represents a
high-intensity earthquake. The displacement time histories of the first and the sixth floor, and the
brace hysteresis are shown in Figure 11 b and c. On the same graphs, analytically predicted
responses are overlapped for comparison. From the graphs, it can be inferred that the
experimental response is in good agreement with the analytical prediction. The slight difference in
the response could be because of the difference in the brace hysteresis of the idealised numerical
model when compared to the experimental specimen as shown in Figure 11d. From the same
figure, it can also be observed that the experimental T-SCED hysteresis is symmetric in terms of
strength and stiffness.
The experimental response of the sixth floor shown in Figure 11c is from the hybrid model
where the first storey is represented as the physical component. It is speculated that the
magnitude of difference between the experimental and analytical response would vary with the
storey brace being considered as well as the storey brace selected as the physical component in
the hybrid model. For instance, from Figure 11, it can be observed that the magnitude of the
difference between the experimental result and the analytical prediction for the top storey node
shown in Figure 11c is less than that of the first floor node shown in Figure 11b. In addition, the
magnitude of difference would vary with the storey brace represented as the physical specimen.
A series of tests were performed to investigate this issue and was found that the magnitude of
variation in the response is not negligible. Detailed results are not included in this paper because
of limited space.
The acceleration time history responses of the first floor, for the same test, are shown in Figure 12.
The accelerations were obtained from numerical differentiation of the displacement time history.
From the figure, it can be clearly observed that the experimental response is in good agreement
with the analytical prediction. The maximum interstorey drift and the maximum nodal acceleration
among all storeys, from several different excitations, are shown in Figure 13 and Figure 14. The
design base earthquake (DBE) and MCE level spectral accelerations shown in Figure 13 and
Figure 14 are obtained from the design spectrum shown in Figure 2, corresponding to the

1.5 Analytical HST Experimental HST

Acceleration (g)

1.0
0.5
0.0
0 2 4 6 8 10 12 14
-0.5
-1.0
-1.5 Time (Sec)
a) 1st floor response

Figure 12. Comparison of acceleration responses from hybrid simulation and analytical prediction.
Peak Interstorey Drift (%)

2.5 Analytical HST Ex

Experimental HST
Collapse prevention limit
2
Life safety limit
1.5
1
DBE

0.5
MCE

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
Spectral Acceleration (g)

Figure 13. Graph showing the variation of interstorey drift with spectral acceleration.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

2 Analytical HST Experimental HST

E

acceleration (g)
Peak nodal
1.5
1
0.5

MCE
DBE
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
Spectral Acceleration (g)

Figure 14. Graph showing the variation of nodal acceleration with spectral acceleration.

fundamental period of the structure. In general, the experimental response tends to be lower by an
average of 18% and 7% than the analytical prediction for peak interstorey drift and peak nodal
acceleration, respectively. A few test cases with higher level of displacement difference are shown
in Figure 15.
It is expected that the differences between the experimental and numerical response results were
caused by the idealisation of the true experimental hysteretic curve in the numerical model by using
an idealised flag shape (bilinear representation) as shown in Figure 11d. The experimental specimen
has a relatively smooth hysteretic curve without a sharp transition between the elastic state and the
activated state. This is especially true when the deformation is small. Because the hysteresis loop in
the small displacement range forms a loop, the experimental specimen dissipates some seismic
energy during these small cycles, which may also have resulted in relatively lower response in
comparison with the analytically predicted response. The residual deformations caused due to this
energy dissipation are very small and are unlikely to have a significant effect on the global response
of the structure [53].
From the same graph, Figure 11d and e, it can also be observed that the experimental specimen
has relatively lower initial stiffness; however, the effect of this is not predominant on the global
response, as shown in Figure 11b and c. This is because of the fact that the displacement at
which the specimen activated is very low, approximately 2.5 mm and represents about 5% of the
peak response of 53 mm in the brace. Hence, the post yielding stiffness governs the global
response. Similar behaviour would be expected in other storeys if other storey braces were
physically tested.
In addition, the global response of the structure in a hybrid simulation is also dependent on the
number of braces that are tested experimentally. Given the differences between the numerical
model of the SCED braces and the experimental specimen, the response obtained using one
experimental specimen is better than the one obtained from a pure numerical simulation alone. To
increase the accuracy of substructure hybrid simulation, research projects on numerical model
updating method in hybrid simulation [54] and on element selection strategy [55] have also been
carried out.
Figure 13 shows that the structure meets the life safety performance criteria at DBE level and the
collapse prevention performance level at MCE level. The criteria of the DBE and MCE limit states
for braced steel frame systems are listed in Table II. In terms of acceleration, Figure 14 shows that
most of the data points lie between 0.5 g and 1.0 g even at the DBE level. The associated
performance limit states are listed in Table III. Figure 13 and Figure 14 clearly show that there are
not many data points corresponding to the more damaged limit states such as immediate occupancy,
even though the hazard is varied from the DBE level to MCE level. Hence, in case of SCED
system, the limits associated with these performance levels can be further relaxed. To generalise this
conclusion, however, it is necessary to run more comprehensive simulation considering record-to-
record variability at DBE and MCE levels.
Experimental and analytical seismic fragility curves are derived using the SAC FEMA (Federal
Emergency Management Agency) [56] procedure. In this method, the median seismic demand at a
given seismic intensity, Sa , is expressed as D ^ ¼ aðSa Þb . It is assumed that the seismic demand
follows the log-normal distribution, and the record-to-record variability is represented with
dispersion βD=S
^ a . The probability of exceeding the performance limit states defined in Table II and

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

Figure 15. Comparison of dynamic analysis for different ground excitations; earthquake name, year and
recorded station are shown in the brackets.

Table II. Structural performance levels and damage expressed in terms of drift (Source: [57]).
Structural performance levels
Elements Type Collapse prevention Life safety Immediate occupancy

Braced steel Primary Extensive yielding Many braces yield or Minor yielding or
frames and buckling of braces. buckle but do not buckling of braces.
Many braces and their totally fail. Many
connections may fail. connections may fail.
Secondary Same as primary Same as primary Same as primary
Drift 2% transient 1.5% transient 0.5% transient
or permanent 0.5% permanent negligible permanent

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

Table III. Structural performance levels and damage expressed in terms of accelerations.
Structural performance levels
Immediate
Elements Type Collapse prevention Life safety occupancy
Braced steel Primary Interior spaces completely Severe accelerations, Some furniture
frames in shambles, severe damage most furniture toppled over, tipping
to sensitive equipment some damage to
sensitive equipment
Acceleration Over 1.0 g 0.5g–1.0 g 0.0g–0.5 g

Table III is evaluated by considering the median performance limit states, C ^ , and their assumed
dispersions, βC^ . More specific details on the seismic fragility analysis procedure may be found in
Kammula [55]. The derived experimental and analytical fragility curves are compared in Figure 16
and Figure 17. In the figure, the horizontal axis represents the demand expressed in terms of
the spectral acceleration, and the vertical axis represents the probability of exceeding the specified
limit states.
Figure 16 and Figure 17 clearly shows that, in general, the analytical results are conservative. The
difference between the analytical and the experimental fragility curves is inversely proportional to
the drift limit state. Differences in modelling of the hysteresis, as explained in the previous

1.0
Probability of Exceedance

LS = 0.5%
0.8
Maximum Credible EQ

LS = 1.0%
Analysis
Design EQ

0.6
Hybrid
0.4
LS = 1.5%
0.2
LS = 2.0%

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Spectral Acceleration (g)

Figure 16. Seismic fragility curves in terms of maximum drift limits.

1.0 LS = 0.5 g
Analysis
Probability of Exceedance

0.8
Maximum Credible EQ

Hybrid
Design EQ

0.6

0.4 LS = 1.0 g

0.2

LS = 1.5 g
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Spectral Acceleration (g)

Figure 17. Seismic fragility curves in terms of maximum floor accelerations.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

section, could be one of the primary reasons. For a given hazard, the probability of reaching a certain
limit state is inversely proportional to the drift limit associated with it. The probability of reaching
the life safety performance limit even at MCE level hazard is less than 30% as shown in
Figure 16. In addition, the residual displacements were negligible after all hybrid simulations. In
terms of maximum floor acceleration, the probability of reaching acceleration limit of 1.0 g is less
than 40% even at MCE level hazard. These results signify the high performance of structures with
SCED braces.

7. CONCLUSIONS

The seismic performance of a six-storey structure with T-SCED brace lateral force resisting system was
assessed using hybrid simulation. The framework and the configuration used for the hybrid simulation
setup were first explained. One of the T-SCED braces in the six-storey structure was represented as the
physical specimen whereas the rest of the structure was modelled numerically. In total, 30 hybrid
simulations were carried out. These experiments have contributed to the understanding of the global
behaviour of the T-SCED system in addition to demonstrating the hybrid simulation technique itself.
The experiments have also contributed to the understanding of some of the practical intricacies of
hybrid simulation such as the iterative mechanism approach and the DOF transformation associated
with the testing process of the hybrid simulation. The following are the main observations from
these experiments.

• The hysteretic curve from both the experiment and analytical simulations are in good agreement
with only minor differences. The experimental specimen does not show transitions as sharp as the
analytical model from the elastic state to the nonlinear state.
• In general, the nodal displacement responses from both simulations match well. The maximum
response from the experimental results tends to be smaller than the maximum responses from
analytical predictions.
• Based on the criteria in ASCE 41–06, the structure meets the life safety performance criteria at the
design level hazard and the collapse prevention performance level at maximum credible seismic
hazard. To generalise this conclusion, though, it is necessary to run more comprehensive simulation
considering record-to-record variability at DBE and MCE level intensities.
• The structure mitigates residual deformations that would otherwise be expected in yielding brac-
ing systems.
• In this study, the T-SCED system was subjected to more than 30 earthquakes without sustaining
any notable damage or response deterioration.
• The fragility curves show that the performance of the six-storey T-SCED structure under investi-
gation is good both in terms of displacement and acceleration demand.
This particular research into seismic performance assessment of SCED braced structures using
hybrid simulation has provided a valuable insight in understanding both the hybrid simulation
procedure and the global behaviour of self-centering systems.

APPENDIX A: Design details of the six-storey structure

Table A-1. Parameters of the experimental telescoping self-centering energy dissipative specimen
Parameter Experimental specimen parameters
Initial stiffness k1 (kN/mm) 300
Post yielding stiffness ratio r 0.015
Post yielding stiffness k2 (kN/mm) 4.5
Activation load Pa (kN) 420
Ratio of forward to reverse activation force β 0.9
Yield displacement (mm) 1.4

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

Table A-2. Final details of the telescoping self-centering energy dissipatives used in the analytical model
Storey Brace activation load, Pa, (kN) No. of braces required (n) Revised fundamental period (seconds)

1 2553 6 0.51
2 2553 6
3 2214 5
4 1926 5
5 1483 4
6 937 2

Table A-3. Cross sectional properties (US customary unit)

Storey Column (C3) Column (C1) Column (C2) Column (C4) Column (C5) Girder

6 W12 × 45 W5 × 16 W12 × 106 W14 × 68 W14 × 176 W18 × 50

5 W12 × 45 W5 × 16 W12 × 106 W14 × 68 W14 × 176 W18 × 50
W12 × 96 W6 × 20 W14 × 159 W14 × 132 W14 × 257
4 W12 × 96 W6 × 20 W14 × 159 W14 × 132 W14 × 257 W18 × 50
3 W12 × 96 W6 × 20 W14 × 159 W14 × 132 W14 × 257 W18 × 50
W14 × 132 W8 × 28 W14 × 176 W14 × 211 W14 × 257
2 W14 × 132 W8 × 28 W14 × 176 W14 × 211 W14 × 257 W18 × 50
1 W14 × 132 W8 × 28 W14 × 176 W14 × 211 W14 × 257 W18 × 50

Table A-4. Dynamic characteristic of the structure

Variable Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6

Tn (secondss) 0.51 0.18 0.15 0.11 0.09 0.07

ω (rads-1) 12.48 34.77 40.79 61.31 65.76 90.18
Storey 1 0.12 0.60 0.22 3.17 0.59 1.52
Storey 2 0.26 1.00 1.87 4.34 0.52 0.81
Storey 3 0.43 1.09 1.18 2.47 0.16 1.50
Storey 4 0.62 0.74 2.03 1.45 0.77 0.95
Storey 5 0.82 0.01 0.62 3.65 0.42 2.14
Storey 6 1 1 1 1 1 1

ACKNOWLEDGEMENTS
This research was carried out with the financial support of the Natural Sciences and Engineering Research
Council of Canada (NSERC) and startup research funds from the Department of Civil Engineering at
University of Toronto. The support from the technical staff at the Structural Testing Facility at University
of Toronto is greatly acknowledged.

REFERENCES
1. SEAOC. Performance Based Seismic Engineering of Buildings. Vision 2000 committee, Structural Engineers
Association of California, Sacramento, California, 1995.
2. Pampanin S, Christopoulos C. Towards Performance-based seismic design of MDOF structures with explicit
consideration of residual deformations. ISET Journal of earthquake technology 2004; 41(1):53–73.
3. McCormick J, Aburano H, Ikenaga M, Nakashima M. Permissible residual deformation levels for building
structures considering both safety and human elements. Proc. 14th World Conf. Earthquake Engng, Beijing,
China, No. 05-06-0071, 2008.
4. Erochko J, Christopoulos C, Tremblay R. Residual Drift response of SMRFs and BRB frames in steel buildings
designed according to ASCE 7–05. Journal of Structural Engineering 2011; 137(5):589–599.
5. Ricles JM, Sause R, Garlock M, Zhao C. Posttensioned seismic-resistant connections for steel frames. Journal of
Structural Engineering 2001; 127(2):113–121.
6. Christopoulos C, Filiatrault A, Uang CM, Folz B. Post tensioned energy dissipating connections for moment-
resisting steel frames. Journal of Structural Engineering 2002; 128(9):1111–1120.
7. Rojas P, Ricles JM, Sause R. Seismic performance of post-tensioned steel moment resisting frames with friction
devices. Journal of Structural Engineering 2005; 131(4):529–540.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 HYBRID SIMULATION OF SELF-CENTERING ENERGY DISSIPATIVE BRACING SYSTEM

8. Kim H-J, Christopoulos C. Friction damped posttensioned self-centering steel moment- resisting frames. Journal of
Structural Engineering 2008; 134(11):1768–1779.
9. Herning G, Garlock MM, Ricles J, Sause R, Li J. An overview of self centering steel moment frames. Proceedings of
the 2009 Structures Congress, 2009; 1412–1420.
10. Chou C-C, Chen J-H. Seismic design and shake table tests of a steel post-tensioned self-centering moment frame with a
slab accommodating frame expansion. Earthquake Engineering and Structural Dynamics 2011; 40(11):1241–1261.
11. Graesser EJ, Cozzarelli FA. Shape-memory alloys as new materials for a seismic isolation. Journal of Engineering
Mechanics 1991; 117(11):2590–2608.
12. Dolce M, Cardone D, Marnetto R. Implementation and testing of passive control devices based on shape memory
alloys. Earthquake Engineering and Structural Dynamcis 2000; 29(7):945–968.
13. DesRoches R, McCormick J, Delemont M. Cyclic properties of superelastic shape memory alloy wires and bars.
Journal of Structural Engineering 2004; 130(1):38–46.
14. Wilson JC, Wesolowsky MJ. Shape memory alloys for seismic response modification: a state-of the-Art review.
Earthquake Spectra 2005; 21(2):569–601.
15. Nims DK, Richter PJ, Bachman RE. The Use of the energy dissipating restraint for seismic hazard mitigation.
Earthquake Spectra 1993; 9(3):467–489.
16. Tsopelas T, Constantinou MC. Experimental and analytical study of a system consisting of sliding bearings and fluid
restoring force/damping devices. NCEER-94-0014. Buffalo, NY: Dept. of Civ. Eng., State University of New York at
Buffalo, 1994.
17. Filiatrault A, Tremblay R, Kar R. Performance evaluation of friction spring seismic damper. Journal of Structural
Engineering 2000; 126(4):491–499.
18. Christopoulos C, Tremblay R, Erochko J, Choi H. Response of 2-d and 3-d buildings incorporating buckling-
restrained and self-centering bracing systems. Proceedings, Annual Conference-Canadian Society for Civil
Engineering 2008a; 4:2162–2171.
19. Christopoulos C, Tremblay R, Kim H-J, Lacerte M. Self-centering energy dissipative bracing system for the seismic
resistance of structures: development and validation. Journal of Structural Engineering 2008b; 134(1):96–107.
20. Zhu S, Zhang Y. Seismic analysis of concentrically braced frame systems with self- centering friction damping
braces. Journal of Structural Engineering 2008; 134(1):121–131.
21. Kurama Y, Pessiki S, Sause R, Lu L-W. Seismic behaviour and design of unbonded post tensioned precast concrete
walls. PCI Journal 1999; 44(3):72–89.
22. Palermo A, Pampanin S, Buchanan A, Newcombe M. Seismic Design of multi-storey buildings using laminated ve-
neer lumber (LVL). In Proceedings, 2005 NZSEE Conference, 11–13 March 2005 Wairakei, NZ. Wellington: New
Zealand Society for Earthquake Engineering, 2005.
23. Azuhata T, Midorikawa M, Ishihara T. Earthquake damage reduction of buildings by self-centering systems using
rocking mechanism. The 14th World Conference on Earthquake Engineering, 2008.
24. Toranzo LA, Restrepo JI, Mander JB, Carr AJ. Shake-table tests of confined-masonry rocking walls with supplemen-
tary hysteretic damping. Journal of Earthquake Engineering 2009; 13(6):882–898.
25. Eatherton M, Hajjar JF, Deierlein GG, Krawinkler H, Billington S, Ma X. Large-scale cyclic and hybrid simulation
testing of a controlled rocking steel braced frame system. 9th US National and 10th Canadian Conference on
Earthquake Engineering, Toronto, July 25, 2010.
26. Sause R, Ricles JM, Roke DA, Chancellor NB, Gonner NP. Large-scale experimental studies of damage-free self-
centering concentrically-braced frame under seismic loading. In ASCE Structures Congress, 2010.
27. Deierlein G, Krawinkler H, Ma X, Takeuchi T, Eatherton M, Kasai K, Hajjar J, Midorikawa M. Earthquake resilient
steel braced frames with controlled rocking and energy dissipating fuses. Steel Construction 2011; 4(3):171–175.
28. Elgawady MA, Sha’lan A. Seismic behaviour of self-centering precast segmental bridge bents. Journal of Bridge
Engineering 2011; 16(3):328–339.
29. Chou C-C, Chen Y-C. Cyclic tests of post-tensioned precast CFT segmental bridge columns with unbonded strands.
Earthquake Engineering and Structural Dynamics 2006; 35:159–175.
30. Erochko J, Christopoulos C, Tremblay R, Kim HJ. Shake table testing and numerical simulation of a self-centering energy-
dissipative braced frame. Earthquake Engineering and Structural Dynamics 2012; 42(11):1617–1635.
31. Erochko J, Christopoulos C, Tremblay R. Design and testing of an enhanced elongation telescoping self-
centering energy dissipative (T-SCED) brace. 15th World Conference on Earthquake Engineering, Lisbon,
Portugal, 2012.
32. Choi H, Erochko J, Constantin C, Tremblay R. Comparison of the seismic response of steel buildings incorporating
self-centering energy-dissipative braces, Buckling Restrained Braces and Moment-Resisting Frames. University of
Toronto, Report No. 05–2008, 2008.
33. Tremblay R, Lacerte M, Christopoulos C. Seismic response of multistory buildings with self centering energy
dissipative steel braces. Journal of Structural Engineering 2008; 134:1,108–120.
34. Chou C-C, Chen Y-C. Development and seismic performance of steel dual-core self-centering braces. 15th World
Conference on Earthquake Engineering, Lisbon, Portugal, September 24-28, 2012.
35. Orsini G. A model for buildings’ vulnerability assessment using the parameter less scale of seismic intensity (PSI).
Earthquake Spectra 1999; 15(3):463–483.
36. Rossetto T, Elnashai AS. Derivation of vulnerability functions for European-type RC structures based on observa-
tional data. Engineering Structures 2003; 25(10):1241–1263.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe
 V. KAMMULA ET AL.

37. Mosalam KM, Ayala G, White RN, Roth C. Seismic fragility of LRC frames with and without masonry infill walls.
Journal of Earthquake Engineering 1997; 1(4):693–720.
38. Reinhorn AM, Barron-Corvera R, Ayala AG. Spectral evaluation of seismic fragility of structures. Structural Safety
and Reliability (ICOSSAR 2001), 2001.
39. Chryssanthopoulos MK, Dymiotis C, Kappos AJ. Probabilistic evaluation of behaviour factors in EC8-designed R/C
frames. Engineering Structures 2000; 22(8):1028–1041.
40. Lin Y-C, Ricles J, Sause R, Seo C-Y. Earthquake simulations on a self-centering steel moment resisting frame with
Web friction devices. Proceedings of the 2009 Structures Congress, 2009; 1364–1373.
41. Yang TY, Stojadinovic B, Moehle J. Hybrid simulation of a zipper-braced steel frame under earthquake excitation.
Earthquake Engineering and Structural Dynamics 2009; 38(1):95–113.
42. Stavridis A, Shing P. Hybrid testing and modelling of a suspended zipper steel frame. Earthquake Engineering and
Structural Dynamics 2010; 39(2):187–209.
43. Christenson R, Lin YZ, Emmons A, Bass B. Large-scale experimental verification of semiactive control through real-
time hybrid simulation. Journal of Structural Engineering 2008; 134(4):522–534.
44. Karavasilis TL, Ricles JM, Sause R, Chen C. Experimental evaluation of the seismic performance of steel MRFs
with compressed elastomer dampers using large scale real-time hybrid simulation. Engineering Structures 2011;
33(6):1859–1869.
45. Lin S-L, Elnashai AS, Spencer BF, Hashash YMA, Fahnestock LA. An integrated earthquake impact assessment
system. Mid-America Earthquake (MAE) Center, Report No. 11/02, 2011.
46. ASCE. Minimum design loads for buildings and other structures, ANSI/SEI7-05 Including Supplement No.1.
American Society of Civil Engineers, 2006.
47. McKenna F, Fenves GL. The OpenSees command language manual, version 1.2. Pacific Earthquake Engineering
Research Center, University of California at Berkeley, 2001.
48. Erochko J. Improvements to the design and Use of post-tensioned self-centering energy dissipative (SCED) braces,
Doctoral Thesis, University of Toronto, Toronto, 2013.
49. Kwon O, Nakata N, Elnashai A. Technical note: a framework for multi-site distributed simulation and application to
complex structural systems. Journal of Earthquake Engineering 2005; 9(5):741–753.
50. Kwon O, Elnashai A. A framework for distributed analytical and hybrid simulations. Structural Engineering and
Mechanics 2008; 30(3):331–350.
51. Combescure D, Pegon P. α-Operator splitting time integration technique for pseudodynamic testing. Error propaga-
tion analysis. Soil Dynamics and Earthquake Engineering 1997; 16:427–443.
52. Chiou B, Darragh R, Gregor N, Silva W. NGA project strong-motion database. Earthquake Spectra 2008; 24(1):23–44
53. Eatherton MR, Hajjar JF. Residual drifts of self-centering systems including effects of ambient building resistance.
Earthquake Spectra 2011; 27(3):719–744.
54. Kwon O, Kammula V. Model updating method for substructure pseudo-dynamic hybrid simulation. Earthquake
Engineering and Structural Dynamics, 2013; 42(13):1971–1984. DOI: 10.1002/eqe.2307
55. Kammula, V. Application of hybrid simulation to fragility assessment of self-centering energy dissipative (SCED)
bracing system, Master Thesis, University of Toronto, Toronto, 2012.
56. Cornell CA, Jalayer F, Hamburger RO, Foutch DA. Probabilistic basis for 2000 SAC federal emergency manage-
ment agency steel moment frame guidelines. Journal of Structural Engineering 2002; 128:526–533.
57. ASCE (2007). Seismic rehabilitation of existing buildings, ASCE/SEI Standard 41–06, Reston, VA.

Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. (2013)
DOI: 10.1002/eqe