EESD-Hsiao Etal 2013

EARTHQUAKE ENGINEERING & STRUCTURAL DYNAMICS
Earthquake Engng Struct. Dyn. 2013; 42:1547–1564

Published online 14 February 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/eqe.2286
Evaluation of the response modification coefficient and collapse

potential of special concentrically braced frames
Po-Chien Hsiao*,†, Dawn E. Lehman and Charles W. Roeder

Department of Civil and Environmental Engineering, University of Washington, Seattle, WA 98195-2700, USA
ABSTRACT
Special concentrically braced frames (SCBFs) are commonly used for seismic design of buildings. Their
large elastic stiffness and strength efficiently sustains the seismic demands during smaller, more frequent
earthquakes. During large, infrequent earthquakes, SCBFs exhibit highly nonlinear behavior due to brace
buckling and yielding and the inelastic behavior induced by secondary deformation of the framing system.
These response modes reduce the system demands relative to an elastic system without supplemental damping
using a response modification coefficient, commonly termed the R factor. More recently, procedures put forth in
FEMAP695 have been made to quantify the R factor through a formalized procedure that accounts for collapse
potential. The primary objective of the research in this paper was to evaluate the approach for SCBFs. An
improved model for SCBFs that permits simulation of brace fracture was used to conduct response history
analyses. A series of three-story, nine-story and 20-story SCBFs were designed and evaluated. Initially, the
FEMAP695 method was conducted to estimate collapse and the corresponding R factor. An alternate procedure
for scaling the multiple acceleration records to the seismic design hazard was also evaluated. The results
show significant variation between the two methods. Of the three variations of buildings studied, the largest
vulnerability was identified for the three-story building. To achieve a consistent margin of safety against
collapse, a significantly lower R factor is required for the low-rise SCBFs (three-story), whereas the mid-rise
and high-rise SCBFs (nine-story and 20-story) may continue to use the current value of 6, as provided in
ASCE-07. Copyright © 2013 John Wiley & Sons, Ltd.
Received 14 February 2012; Revised 30 October 2012; Accepted 15 January 2013
KEY WORDS: braced frames; performance evaluation; response modification coefficient; brace fracture;
incremental dynamic analysis
1. INTRODUCTION
Special concentrically braced frames (SCBFs) are commonly used as the primary lateral-load resisting
system in buildings. These frames have beams, columns and braces concentrically joined at
workpoints. In the design office, the system is initially designed as a vertical truss system, and the
initial lateral stiffness and resistance of the frame is provided through the axial response of all members
(beams, columns and braces). Conventional design uses reduced seismic loads. In larger, less frequent
earthquakes, inelastic action in the brace is expected. This inelastic behavior is primarily governed by
the severe buckling, post-buckling and tensile yield deformations of the brace; secondary yield
mechanisms may occur in the gusset plate connection, beams and columns.
To assure that the system can achieve these large inelastic deformations without premature failure,
current AISC SCBF design requires capacity design of the beams, columns and connections to meet
the tensile and compressive capacity of the brace [1]. Local and global slenderness limits are applied
to the brace to assure that the brace can sustain multiple large drift cycles.
*Correspondence to: Po-Chien Hsiao, Department of Civil and Environmental Engineering, University of Washington,
Seattle, WA 98195-2700, USA.
†
E-mail: pchsiao@uw.edu
Copyright © 2013 John Wiley & Sons, Ltd.

1548 P.-C. HSIAO, D. E. LEHMAN AND C. W. ROEDER
Currently, there are few design restrictions on or understanding of the gusset plate, which leads to
considerable variation in the gusset plate connections designed for SCBF systems. Prior research has
shown that different gusset plate designs influence the seismic performance of the braced frame and
that poorly designed gusset plates can severely decrease the drift capacity of an SCBF [2]. These
findings have led to a new, balanced design procedure that provides a method for consistent gusset
plate design and increases the ductility of the braced frame system beyond that achievable with the
current AISC SCBF seismic design criteria [3, 4] including a new elliptical clearance model [2].
The method was developed and supported by a coordinated experimental and analytical research
program on SCBF systems. In all, more than 30 single-story, single-bay frames were tested and
analyzed [5–8]. A multi-story experimental program, including 3 two-story and three-story frames
[4, 9], was also conducted to investigate application of the design recommendations developed from
the single-story research, to validate a design method for midspan gusset plate connections, and to
evaluate the distribution of inelastic deformation between different stories.
Design of SCBFs depends on several seismic performance factors. In particular, the response
modification coefficient (R factor), which is used to reduce the elastic seismic design loads to those
used for member design, significantly influences the design and, therefore, would be expected to
influence the performance of SCBFs. R factors were originally derived in an ATC-3-06 report [10],
with the historic values determined from engineering judgment and qualitative comparisons with the
known response capabilities of a relatively few, well-understood seismic-force-resisting systems. In
general, these R values were also related to the Rw values used in the Uniform Building Code for
allowable strength seismic design. In the most recent provisions, the National Earthquake Hazards
Reduction Program (NEHRP) Recommended Provisions for Seismic Regulations for New Buildings
and Other Structures [11] provided the R factors for more than 75 different seismic-force-resisting
systems including SCBFs. However, the R values were somewhat arbitrarily assigned without
appropriate verification of their seismic response characteristics. It has become clear that a more
rational assessment of the R value is needed for all seismic resisting systems, including SCBFs.
In modern design, large, nonlinear inelastic deformations are expected during large seismic loads and
the system must be detailed to sustain these demands. However, the seismic design forces must be large
enough to result in sufficient strength and stiffness to meet the elastic demands in lesser events. The
seismic design forces are determined by reducing the expected elastic seismic forces by the R value.
An analytical study was undertaken to evaluate the seismic performance and R values for SCBF
systems. SCBF systems were designed using newly proposed seismic design procedures [4], and
then nonlinear dynamic analysis were performed using practical, validated models as described in
the succeeding text. The improved models of SCBFs combined with ATC-58 fragility curves allow
accurate prediction of all performance limit states, such as no repair, possible brace replacement,
brace fracture, and potential collapse, in the performance evaluations.
The impact of the R value was studied by evaluating the full seismic performance, including
collapse. A series of SCBF buildings were designed using a range of R values. The building designs
were assessed with two methods. The first method used the FEMA P695 incremental dynamic
analysis (IDA) procedure [12] to estimate collapse potential and to propose R values that
consistently achieved the collapse prevention performance limit state. The second method included a
suite of 20 acceleration records for the 10% and 2% in 50 years seismic hazards that were scaled to
the seismic hazard level for the building site, and the predicted system performance was used to
propose appropriate R factors of SCBF structural systems. The deflection amplification factor, Cd,
and system overstrength factor, Ωo, were also evaluated as part of this nonlinear dynamic analysis.
2. DESIGN OF THE MODEL BUILDINGS
A series of three-story, nine-story and 20-story idealized SCBF buildings were designed using the
equivalent lateral force procedure. The buildings were based on the model buildings in the SAC
Steel project, adopting the basic floor plan, story height, and gravity loads [13]. Appropriate
modifications were made to translate the buildings to SCBFs. Figure 1 shows the typical elevations
and floor plans of the different buildings considered here.
Copyright © 2013 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2013; 42:1547–1564
DOI: 10.1002/eqe
EVALUATION OF THE R FACTOR AND COLLAPSE POTENTIAL OF SCBFS 1549
Figure 1. Elevations and floor plans of (a) three-story, (b) nine-story and (c) 20-story buildings.
A primary study parameter was the impact of variation in the R value. Several different R values
were evaluated, and include the following: (i) R of 3, 4, 5, and 6 for the three-story buildings; (ii) R
of 3, 6, 7 and 8 for the nine-story buildings; and (iii) R of 3, 4, 6, and 8 for the 20-story buildings.
Current codes employed an R value of 6 for SCBF systems.
The buildings were designed using the NEHRP recommended provisions [11]. All of the buildings
were designed for a location in Seattle, WA, using a Seismic Design Category D with soil Site Class C.
The three-story buildings had 4 6 bays with identical story height, whereas the nine-story and
20-story buildings have 5 5 and 4 4 bays, respectively and with taller bottom stories to reflect
typical midrise construction, as shown in Figure 1. The 20-story frame exceeds the height limit of
48.8 m (160 ft) for steel SCBF as per NEHRP [11], and the design requires peer review, which was
not considered by this study. Therefore, the designed 20-story buildings might be different in
practice. However, it should be noted that this research used larger seismic design forces than the
minimum required by the specification, and it is unlikely that the seismic performance of the
buildings will be better than that predicted here
The braced bays were situated along the perimeter of the buildings in a symmetric plan configuration
using a multi-level X-bracing configuration, as illustrated in the figure. Table I summarizes the total
gravity loads for individual floors and the roof (lighter loads were applied to the roof) for each building
configuration. The buildings were designed to meet the NEHRP design spectrum using the United
States Geological Survey’s (USGS) mapped spectral acceleration values at 2% in 50 years hazard level
for the location of Seattle (updated in 2008). All buildings are designed with equivalent lateral force
procedures. The 20-story buildings would ordinarily be designed using the response spectrum
procedure. The equivalent lateral force method was used to permit direct comparison of all building
heights. However, response spectra analyses were performed, and they showed that the equivalent static
analysis resulted in larger member design forces and different distribution of forces over the frame than
the response spectrum method. The following values were used: SS (mapped spectral acceleration
parameter at short periods) of 1.4 g and S1 (mapped spectral acceleration parameter at a period of 1 s) of
DOI: 10.1002/eqe
Table I. Design parameters and periods.

Gravity loads (kN)
Model buildings R Floor Roof CS T (s) T1 (s)
Three-story 3 9379 9011 0.390 0.31 0.34

4 0.293 0.36
5 0.234 0.37
6 0.195 0.38
Nine-story 3 9721 9317 0.258 0.73 0.69
6 0.129 0.84
7 0.111 0.88
8 0.097 0.94
20-Story 3 2952 2610 0.144 1.31 1.49
4 0.108 1.66
6 0.072 1.85
8 0.054 2.01
0.53 g. The site class effect was included and the damped spectra were adjusted to 2% of critical damping
by applying bS and b1 factors of 0.8. The resulting design spectral acceleration parameters were SDS of
1.17 g and SD1 of 0.57 g. The buildings were considered to be general office buildings, and an
occupancy importance factor, I, of 1.0 was used. The redundancy factor of 1.0 was used for all model
buildings; whereas higher values of the redundancy factor would lead to heavier braces and stiffer
structures, and would potentially increase the seismic design forces for load combinations including
gravity load.
In the design of the three-story, nine-story and 20-story buildings, the braces were all rectangular
hollow structural sections and satisfied the AISC seismic compactness criteria [14]. The gusset plate
connections were rectangular and designed using two different methods to compare current with the
proposed balanced design procedure [3]. All of the beam-to-column connections of the frames were
designed as welded-flange welded-web connections. Framing members were designed to develop the
expected capacity of the braces. Table I shows the seismic response coefficient, CS, design periods,
T, and the fundamental periods obtained from modal analysis, T1, for all the model buildings. The
slenderness ratios of the braces varied with the nine-story building having the lowest slenderness
ratio and the 20-story building having the largest slenderness values. This slenderness ratio is one of
several parameters affecting brace fracture.
These designs resulted in very different amounts of steel. To present the variation of the steel weight
with various design R values, which directly influence the construction cost of the buildings, total steel
weights of the lateral-load resisting frames in single direction were calculated and normalized by the
total design gravity loads of the buildings. Figure 2 shows the differences of the normalized steel
weights for buildings using R = 3 and 8 compared with R = 6. The comparison shows that reducing
the design R value (R = 3 here) for the three-story (low-rise) buildings had relative lower impact on
the construction cost than the nine-story and 20-story (mid-rise and high-rise) buildings.
3. SYSTEM MODELING INCLUDING EFFECT OF GRAVITY FRAMES
To assess the seismic performance of SCBFs and the R value in a rational manner, comprehensive nonlinear
dynamic analysis models are required. These models must include every yield mechanism and failure mode
that impacts the seismic response of the frame and must be capable of simulating the response beyond initial
fracture. These modeling requirements have been met for CBFs and proposed in prior studies. Detailed
information on the modeling techniques used is provided in the following references [15, 16].
All of the braced frame systems of the buildings in the study were modeled using the proposed
simulation model of SCBFs, which consisted of the proposed spring models of gusset plate
connection, rigid links and line-elements [15], and the brace-fracture model [16] to enable accurate
simulation of the SCBF systems to well beyond brace fracture.
DOI: 10.1002/eqe
Figure 2. Variations of the normalized steel weight of buildings using R = 3 and 8 compared with that
of R = 6.
The gravity loads and second-order (P-delta) effects were included in the simulation by employing a
leaning column connected to the frame by rigid links, as illustrated in Figure 3. Analyses were
performed with and without contributions of the gravity framing to the lateral stiffness of the
system, to study their impact. When this secondary resistance was included, nonlinear springs
simulating the combined rotational strength and stiffness of gravity beam-column connections were
placed between the rigid link and the leaning column. Half of the building gravity loads at each
floor level were supported by the leaning column, because each building model used two seismic
frames in each direction to resist the lateral loads, as shown in Figure 1.
The properties of the nonlinear spring were based upon Liu’s model of shear-plate connections with
composite slabs [17], and the springs were simulated using the Pinching4 material model in OpenSees.
Each spring element had the number of the gravity bays within a half building, modeled using the total
moment strength of all of the contributing single shear-plate connections (for example a multiplier of
12 was used for the three-story model buildings to simulate the contribution of half of the total number
of gravity connections). The rigid beams used in the gravity frame model were supported by rollers,
which were slaved to the nodes at the middle of the frame in the lateral direction at each level.
When this secondary resistance was excluded, the springs were replaced by pins, and therefore, no
rotational stiffness was included.
The base of the leaning column was pinned. Wide-flange sections W10 49 and W12 65 were
selected for the gravity columns of the three-story and nine-story buildings throughout the height,
Figure 3. Illustration of the simulation of P-delta effect and the lateral load resisting contribution of
gravity frames.
DOI: 10.1002/eqe
respectively. The gravity columns of the 20-story buildings varied: the smallest wide flange section was
a W8 31 and the largest wide-flange section was a W14 90. To represent the initial axial and bending
stiffness and strength of the gravity columns, a particular cross section of the leaning column was adopted
with a cross-sectional area, moment of inertia and plastic moment capacity equal to the total of the gravity
columns in half the building (for example 11.5 columns total for the three-story model buildings). The
resulting models led to the fact that the slenderness ratio of the leaning column for the three-story
buildings was much lower than that for the nine-story and 20-story buildings, whereas that of the 20-
story buildings was slightly larger than that of the nine-story buildings. However, on the basis of the
analytical predictions, this variation did not have a great impact on the results.
To simulate the earthquake characteristics of the ground motions in Seattle, the SAC ground motion
record sets assembled for Seattle (each set consists of 20 acceleration records) [13, 18] were used in
this investigation. All of the ground motions were applied using a time step of 0.005 s. To represent the
different seismic hazards, the records were appropriately scaled using the fundamental period of the
structure and the periods resulting from the probable damage states. The details and the consequences
of the scaling procedure are described later, in the section on multi-performance level evaluation.
The effects of the gravity framing were investigated first, to determine if it is necessary for inclusion
in the analyses. The results are shown in Figure 4. This figure uses the maximum story displacement
without the lateral resistance of the gravity framing, MSDw/oG, normalized to the maximum story
displacement with the lateral resistance with gravity framing, MSDw/GF to illustrate the impact of the
gravity frame on the computed seismic response. Three building models are considered, including
three-story, nine-story and 20-story buildings. All three models were designed using an R value of 6
and a set of 20 ground motions scaled to 2% in 50 years seismic hazard level.
The figure shows that the gravity frames can impact the computed response. The effect was significant
on the low-rise buildings (three-story), and relatively insignificant on the mid-rise (nine-story) and
high-rise (20-story) buildings. For consistent comparisons, this effect of gravity frames is included in
the rest of the performance evaluations in the study.
4. COLLAPSE ASSESSMENT USING FEMA P695 METHODOLOGY
The FEMA P695 IDA method was used to estimate collapse potential, as related to the R factors and
the four idealized buildings. The following provides an overview of the methodology and its
application to this research project. Further information is found in the reference report (ATC-63).
In the IDA and P695 methodologies, considered buildings are grouped into ‘archetypes’ that are
expected to have similar performance characteristics. Archetype structures should be designed with a
selected range of structural geometry and design parameters including structural configurations, seismic
design categories, fundamental periods, and so on. These archetypes were assembled into ‘performance
groups’ reflecting major changes in structural behaviors within the archetype design space. Given the
Figure 4. Difference of seismic responses between the buildings with and without the gravity frame effect.
DOI: 10.1002/eqe
archetype performance group development, nonlinear models are developed and nonlinear analysis and
performance evaluation of the system are completed.
The FEMA P695 analysis method for evaluation of collapse potential was used in the research. Figure 5
illustrates the relations between the important design parameters used in the methodology. The figure
shows three important design states. The elastic response is based on the assumption that 100% of the
effective seismic weight of the structure, W, participates in fundamental mode of the system with a
period of T1. The second is that the ratio of the mean spectra acceleration corresponding to the
maximum considered earthquake (MCE), SMT, to the seismic response coefficient, CS (V/W, the ratio of
the design base shear and the weight of the structure) is defined as 1.5 times of the R value. The third
and final design concept is that the median spectral acceleration at collapse initiation is ŜCT, which is
predicted using the IDA method.
The primary objective of the method is to evaluate the collapse potential of an archetype for a given
R value. This evaluation compares the computed adjusted collapse margin ratio (ACMR) with the
acceptable collapse margin ratio, which is based on the collapse margin ratio (CMR). The following
paragraphs briefly quantify each.
The CMR is defined as the ratio of the median 5%-damped spectral acceleration of the collapse level
ground motions, ŜCT, to the SMT at the fundamental period of the structure, T1, of the MCE seismic
hazard for the site. This ratio is shown in Figure 5. The IDA procedure is used to determine ŜCT.
The method accounts for the frequency content (spectral shape) of the ground motions, and a
consistent set of ground motion variation parameters is defined for different sites, hazard levels and
structural periods [19], where ŜCT is determined using the 44 selected acceleration records, which
satisfy the ground motion selection criteria. To account for statistical variation in the records, the
CMR is multiplied by a simplified spectral shape factor (SSF) to obtain the ACMR. The SSF values
are specified in the report, and depend on the structural period and the building ductility capacity.
Finally, the archetype is evaluated by comparing the calculated ACMR to an acceptable ACMR,
which was specified according to the uncertainty factors of structural system, including the quality
of design requirements, test data, numerical modeling, and a prescribed set of ground motions. By
the FEMA P695 procedure, the acceptable AMCR should reflect conditional probability of collapse
of 20% for individual archetype (AMCR20%). The ACMR of an individual archetype needed to be
greater than the corresponding acceptable AMCR to pass the trial. If the determined ACMR was
less than the acceptable ACMR, the R value of the performance group does not meet the
performance requirements of collapse prevention.
The three-story and 20-story buildings were evaluated using the FEMA P695 IDA procedure with
design spectral acceleration, SDS, of 1.17 g, which is slightly greater than the value of 1.0 g suggested
by the methodology. For the research, the SAC ground motion set for 2% in 50 years seismic hazard in
Seattle was used in place of the 44 ground motion records provided by the methodology. Originally, the
SAC ground motion records had been scaled to fit the USGS mapped spectral acceleration values at
Figure 5. Illustration of the R-factor as defined by the FEMA P695 methodology [12].
DOI: 10.1002/eqe
multiple period points [13]. For this research, all of the SAC acceleration records were scaled back to
original ground motion records prior to the IDA process by simply dividing the SAC records by their
scaling factors, which were from 0.94 to 10.04 with a median of 2.34, used for scaling the SAC ground
motions given in the SAC report [13]. The original records were subsequently increased until collapse
was predicted for each record following the IDA process.
Two methods for scaling the intensity of ground motions were investigated. The first method,
designated Method M1, collectively increased the entire set of ground motion records upon the
median spectral acceleration of the record set at the fundamental period, so-called record set
intensity (RSI), ST(T1), rather than using different intensities for each record. Using this method, the
records were scaled as a group in proportional amounts, and the value of ŜCT was defined as the RSI
of the group when 50% of the records predicted failure.
The second method, designated as method M2, scaled each record individually, using different
intensities for each record at the period of T1, Sa(T1). Using this method, ŜCT was defined as median at
the period T1 of the response spectra that resulted in predicted collapse by the IDA procedure [20].
Figure 6(a) and (b) shows the resulting IDA curves using the methods M1 and M2, respectively. It is
apparent that scaling the ground motions in different ways significantly changed the distribution of the
resulting IDA curves, which results in different CMRs for the same set of ground motions.
In an actual building, collapse will depend on the resistance of the gravity frame as well as the residual
resistance of the braced frames. Braced frames dramatically lose most of their resistance and sustain
increasing story drift after prediction of the brace fracture. In many IDA evaluations, the collapse
prediction is based on flattening of the IDA responses curves, such as those shown in Figure 6. The
results show that some brace fracture curves flatten at relatively low spectral accelerations whereas
others do not flatten at all. This is partly caused by the irregular characteristics of acceleration response
spectra and the large increases or decreases that may occur with modest changes in period.
In this research, to account for these and other uncertainties, a potential collapse limit state corresponding
to 5% drift was selected for two reasons. First, the nonlinear analyses were deemed to have sufficient
accuracy to approximately 5% story drift. Second, experimentation indicates that most beam–column
connections and columns do not retain sufficient resistance beyond this drift level. Tests, including
SCBF gusset plate connections, indicate that they retain their basic moment resisting connection integrity
well beyond brace fracture, but this integrity cannot be assured for drift levels larger than 5% [2]. This
limiting story drift capacity was used to determine ŜCT. In Figure 6(a) and (b), ŜCT was determined when
the analyses for 10 out of 20 ground motions (50%) had a story drift greater than 5%.
The results shown in Figure. 6 were compared with the acceptable ACMR20% values in FEMA P695
to evaluate the collapse resistance of each system, as shown in Table II. The P695 method requires an
estimate of the system ductility and the structural fundamental period to determine the acceptable
ACMR20%. On the basis of the prior experimental results [4–8], SCBF usually buckled around a story
drift of 0.3% and fractured brace around a story drift of 2.5%, which led to the story-drift ductility of
8.3. The corresponding SSF values, required to scale to CMR to determine the ACMR, were 1.4 and
Figure 6. Incremental dynamic analysis results for the three-story building (R = 6) using (a) M1 and (b) M2
scaling methods with collapse story drift capacity of 5% radians.
DOI: 10.1002/eqe
Table II. Summary of evaluations of the three-story and 20-story buildings using FEMA P695 incremental
dynamic analysis procedure.
Scaling methods M1 [ST(T)] M2 [Sa(T)]
Model buildings 3-Story 20-Story 3-Story 20-Story

R factors 6 3 6 3 6 3 6 3
SMT (g) 1.76 1.76 0.65 0.65 1.76 1.76 0.65 0.65
ŜCT (g) 1.88 2.8 0.55 0.82 2.5 3.35 0.65 0.93
CMR 1.07 1.60 0.85 1.26 1.42 1.91 1.00 1.43
SSF 1.4 1.4 1.65 1.65 1.4 1.4 1.65 1.65
ACMR 1.50 2.23 1.39 2.08 1.99 2.67 1.65 2.36
Accep. ACMR20% 1.73 1.73 1.73 1.73 1.73 1.73 1.73 1.73
Pass/Fail Fail Pass Fail Pass Pass Pass Fail Pass
CMR, collapse margin ratio; SSF, spectral shape factor; ACMR, adjusted collapse margin ratio.
1.65 for the three-story and 20-story buildings, respectively. (These values are for structures having the
ductility greater than 8). An acceptable ACMR20% of 1.73 was determined, in which the three-story
and 20-story buildings were regarded as an archetype and the quality ratings for design requirements,
test data, and numerical modeling were all quantified as ‘B-Good’. The resulting total uncertainty factor
was 0.65.
Table II summarizes the parameters and the results of the evaluations for the three-story and 20-story
buildings with design R values of 6 and 3. As expected, decreasing the R-factor increases the CMR for
all systems studied.
However, the results are not consistent between the two methods. Using Method 1 (as defined in
FEMA P695), both of the systems designed using an R-factor of 3 ‘pass’ the evaluation criteria,
whereas the values of ACMR for the 20-story buildings were smaller than those for three-story
buildings. This result suggests that SCBF systems meeting current design will fail. In contrast, using
scaling method M2, only the 20-story building designed with R = 6 fails to meet the acceptable
ACMR20%. The M1 scaling method suggested that the R factor of 3 is appropriate for both of
three-story and 20-story buildings. The M2 scaling method suggests that R factor of 6 is appropriate
for three-story SCBFs, whereas an R value of 3 is needed for 20-story SCBFs. These contrasting
results suggest significant ambiguity with the proposed method.
Finally, the analysis for both scaling methods M1 and M2 predicted greater potential for collapse
with the 20-story SCBF than for the three-story SCBF. These results are contrary to findings from
prior earthquakes and other studies, in which low rise SCBFs have been deemed more vulnerable [21].
5. MULTI-PERFORMANCE LEVEL EVALUATION PROCEDURE
The results of the P695 analyses revealed ambiguities in the method. To study these further and potentially
eliminate them, an alternate evaluation approach for SCBFs was investigated. This procedure directly
scaled each ground motion to the seismic hazard for the site (M2) with enhancements described later.
In addition to investigating the collapse limit state, the resulting multiple damage limit states of the
SCBFs were examined and considered. Therefore, the range of system performance limit states was
evaluated (e.g., brace buckling, fracture) for each acceleration record and was then considered in the
evaluation of the R factors, as opposed to only the collapse potential. Again, potential collapse of the
structure was assumed if any floor level of the structural system developed a 5% story drift.
The 20 SAC Seattle ground motions were scaled to the 2/50 hazard level (MCE level). In addition,
the seismic performance at the 10/50 hazard level was evaluated, using SAC 10% in 50 years records
for Seattle [13]. The target spectra for each hazard level followed the NEHRP recommended provisions
[11] using the corresponding USGS mapped spectral acceleration values representative of different
hazard levels, 10/50 and 2/50. Without the reduction of 2/3, the 2/50 target spectrum is 1.5 times
that of the design spectrum, whereas the 10/50 target spectrum is slightly smaller than the design
spectrum, as shown in Figure 7(a).
DOI: 10.1002/eqe
Figure 7. (a) The adopted target and design spectra at Seattle; (b) illustrations of the corresponding frames
for the three periods, T1, T2 and T3; (c) the scaled median acceleration spectra and the corresponding target
spectra for the three-story, (d) nine-story, and (e) 20-story special concentrically braced frame buildings.
In addition to the three-story and 20-story SCBFs idealized buildings, an idealized nine-story
SCBFs was investigated. A larger number of R values were evaluated for each system.
The evaluation of the P695 procedure revealed ambiguities associated with scaling ground motions
and results that contradict the normal expected performance of the SCBF system [21, 22]. The research
team considered various scaling approaches that have been used to evaluate the seismic performance of
structures. Prior studies have scaled records using a single period, multiple periods, or a period range
[13]. Multi-period scaling was deemed important for SCBFs, because SCBFs are initially stiff, and
significant shifts in structural period are expected beyond brace buckling and again beyond brace
fracture. Therefore, an approach of scaling the ground motions to match the design response spectra
prior to and beyond the brace fracture was used.
This conclusion was substantiated by a preliminary study. All of the idealized buildings were analyzed
using earthquake records scaled to the first mode only. These analyses demonstrated that significant brace
damage (buckling, yielding, and eventually fracture) was sustained by the three-story and nine-story
building models, but much more limited brace damage in the 20-story building models. Conversely, the
20-story building models exhibited significant higher mode effects, and these higher modes must be
included in the scaling procedure.
As a result, two different scaling approaches were used. For the structures primarily governed by the
fundamental mode prior to damage, that is, the three-story and nine-story buildings, the following
approach was used. For the 10/50 target spectrum, minimal nonlinearity was observed from the
preliminary analysis and those ground motions were scaled to meet the 10/50 target spectrum using
only the first (fundamental) period (T1) of the structure (each ground motion was scaled
independently, and for the structure, that is the number of stories plus R factor, of interest.).
The analyses indicated that structural damage occurred at the 2/50-hazard level, typically in the form
of brace fracture at the critical story. Therefore, multi-period scaling was used, including the
DOI: 10.1002/eqe
fundamental period (T1) and two additional periods for the more severe damage states of the structure.
The second period of interest (T2) corresponded to loss of braces in one-direction. The third period
of interest (T3) corresponded to loss of all braces at the critical story used. These two states are
illustrated in Figure 7(b). The dynamic analyses indicated that the critical story was the bottom story
for the three-story and nine-story building models.
Using this method, the following expressions of scaling factor, f10/50 and f2/50, were developed to
scale each ground motion for each hazard level.
Sa1;t
f10=50 ¼ (1)
Sa1;g
Sa1;t Sa2;t Sa3;t

f2=50 ¼ w1 þ w2 þ w3 (2)
Sa1;g Sa2;g Sa3;g
The expression is a weighted function of the spectra accelerations at the periods of interest. In the
expression, Sai,t is the target elastic spectral (acceleration) values corresponding to each Ti period as
described previously (i = 1 to 3), Sai,g is the spectral (acceleration) value of the ground motion at the
period Ti, and wi is the weight for the Ti period points. The weights were 0.55, 0.35 and 0.1 for T1,
T2 and T3, respectively. Using this scaling approach, the median scaled acceleration spectra at 10/50
and 2/50 hazard levels shown in Figure 7(c) and (d) were developed comparing with the
corresponding target spectra for the three-story and nine-story SCBF buildings, respectively. The
values of three periods, T1–T3, are indicated in the figures. The plots show that the median scaled
(acceleration) spectra matches the target spectra on average between T1 and T3 period points, which
is the region of primary variation in the period.
The preliminary analysis of the 20-story building models indicated significant contributions of the
second mode in top and bottom stories with the first mode governing the response of the remainder
of the building. To include the second mode effects, a modified scaling approach was adopted to
scale the ground motions for use in the 20-story building analysis, basing the scaling function on the
first two modal periods for both of the hazard levels. Note that because significant damage was not
observed in high-rise buildings, the larger periods resulting from brace damage were not included as
they had been in the three-story and nine-story scaling methods. The resulting expression, given in
Equation (3), was used to scale the ground motions for both of the 10/50 and 2/50 events.
Sam2;t Sam1;t
fm;10=50 and fm;2=50 ¼ wm2 þ wm1 (3)
Sam2;g Sam1;g
In the expression, Sami,t is the target design elastic spectral acceleration value for the ith modal
period (i = 1 and 2), Sami,g is the spectral acceleration value of the ground motions at the ith modal
period. For the second expression, weights of wm2 = 0.6 and wm1 = 0.4 were used. Figure 7(e) shows
the median scaled acceleration spectra for 10/50 and 2/50 hazard levels and the corresponding target
spectra for the 20-story SCBF building. The values of the first two mode periods, TM1 and TM2, used
for scaling were indicated in the figure.
A single set of the scaled ground motion records were used for the analyses, which were scaled on
the basis of the computed periods of the three-story, nine-story and 20-story buildings. In all cases, the
computed periods for the buildings designed using an R value of 6 were used for the other R-value
building design with the assumption that variation of the periods due to different R factors had a
minor effect of the scaled ground motions.
The scaled 2% and 10% in 50 years (2/50 and 10/50) ground motions were used for the performance
evaluation of SCBFs, which varied with building height. Figure 8(a)–(c) shows the average maximum
story drifts (MSD) computed for the three-story, nine-story and 20-story SCBFs for the 10/50 and 2/50
hazard levels, respectively. As noted in the figure, a range of R values were studied, and included R = 3
DOI: 10.1002/eqe
Figure 8. Seismic responses of the (a) three-story, (b) nine-story and (c) 20-story buildings with various R
factors at 10/50 and 2/50 hazard levels.
and 6 for the three-story building models, and R = 3, 6, and 8 for the nine-story and 20-story
building models.
The deformation concentrated in the bottom story for the three-story and nine-story frames, which
always occurred after brace fracture at that story. The results show that the three-story SCBFs designed
with R = 6 had much more severe deformation and damage than the nine-story and 20-story SCBF
building models designed using the same response modification factor. In the 10/50 events, the average
MSDs vary between 0.3% and 0.5%, and brace fracture was not predicted. During the 2/50
earthquakes, brace fracture initiated between drifts of 1.5% to 2%. As mentioned previously, brace
fracture resulted in severe concentration of deformation. The average MSDs varied from 1% to 3% for
the three-story buildings for R values of 3 to 6, whereas the MSDs varied from 1% to 2% radians for
nine-story buildings with R values of 3 to 8. These results show that brace fractures occurred in the
three-story building models, with a lesser number for nine-story building models.
The response of the 20-story building models differed significantly from the three-story and
nine-story building models in two ways. First, the upper stories sustained some of the largest
deformations. Response to the 2/50 ground motions indicates concentration of story drift, and
inelastic deformation occurred in both the bottom and upper stories due to the brace fracture.
Further, the distribution of deformation along the height of the building depends on the R values
used in the frame design. Finally, the resulting peak MSDs, which varied between 1.0% to 1.3%,
indicated minimal brace fracture in these structures. These results suggest an increase in structural
performance with building height, which is contrary to the results from the P695 methodology
discussed earlier.
The average maximum story drift distributions shown in Figure 8 provide interesting and valuable
information, but they do not specifically quantify performance. The analytical models used to simulate
SCBFs, discussed in the prior section, are capable of predicting both the global seismic behavior as
well as the local performance including brace buckling capacity, out-of-plane brace displacement
and the onset of brace fracture. This model allows more comprehensive evaluation of the system
performance at all performance levels.
Several performance limit states were considered here including the following: (i) brace buckling;
(ii) possible replacement of the brace; (iii) brace fracture; and (iv) potential collapse of the building.
The brace buckling (and yielding) and fracture behaviors were computed directly. Requirements for
possible replacement of the brace were defined as follows: (i) the maximum out-of-plane
displacement of the braces exceeded three times the brace depth or (ii) prediction of brace fracture.
These damage states are based upon definitions developed in a prior study on fragility curves of
CBFs [23]. The same story drift limit of 5% was adopted as the potential collapse criteria, because
experiments show that braced frames have considerable inelastic deformation capacity after brace
fracture due to the gusset plate connections. Therefore, basing collapse solely on brace fracture is
unduly conservative. Experiments also show that these connections are unlikely to be reliable at
drifts greater than 5%. The authors have greater confidence in the predicted damage states with this
alternate evaluation procedure than with the IDA procedure, because the damage states are based
upon experimental observation rather than on unverified analytical procedure.
DOI: 10.1002/eqe
All analytical responses were assembled to determine the occurrence probability of the four
performance limit states. Figure 9(a)–(c) shows the probabilities of those damage limit states for the
three-story, nine-story and 20-story buildings for each of the studied R factors at the 2/50 hazard
level (MCE level), respectively.
As expected, significant reductions in R factors effectively reduced the probabilities of all damage
limit states beyond the brace buckling for all story-height buildings. Clearly, brace buckling will
occur in the 2%-in-50 year event, even for very low R values. Yet the results show that it is not the
R factor alone to impact system performance. Clearly, the number of stories impacts performance,
with the 20-story building sustaining significantly less damage than the three-story building for a
given R factor. The reduced damage of the 20-story building might partially result from the use of
equivalent lateral force procedure in design. These results suggest that a smaller R factor may be
needed for consistent performance of all SCBF buildings. This clearly contradicts the prediction
from the P695 IDA evaluation.
Figure 10 shows similar damage level comparisons for the 10/50 hazard level. Figure 7(a) shows
that the 10/50 hazard level is very close to the design spectrum (2/3 of the MCE). Results show that
the probability of brace buckling was high, but little additional damage was predicted regardless of
the R value.
6. DEFLECTION AMPLIFICATION AND OVERSTRENGTH FACTORS
The deflection amplification, Cd, and overstrength, Ωo, factors were also computed for each of the
building models. The results of those analyses are presented here.
The deflection amplification factor is commonly considered the ratio of the maximum displacement
and the yield displacement. Here, it was calculated as the ratio of the maximum dynamic responses
at the 2/50 hazard level to the elastic deformations under the design equivalent lateral forces.
Figure 11(a)–(c) shows the mean and standard deviation of the deflection amplification factors for
Figure 9. Probability of damage at the 2/50 hazard level with various R values: (a) three-story special
concentrically braced frame (SCBF); (b) nine-story SCBF, and (c) 20-story SCBFs.
DOI: 10.1002/eqe
Figure 10. Probability of damage at the 10/50 hazard level with various R values: (a) three-story special
concentrically braced frame (SCBF); (b) nine-story SCBF, and (c) 20-story SCBFs.
Figure 11. The means and standard deviations of the deflection amplification factors for (a) three-story,
(b) nine-story, and (c) 20-story buildings with various R factors.
DOI: 10.1002/eqe
three-story, nine-story and 20-story buildings, respectively. In all cases, the bottom story always had
larger value than the upper stories, although the deviations for a given story and among the stories
are less for the taller structures.
The values of Cd vary with R factor, with a decrease in both the mean and standard deviation. The
current value of Cd = 5 for SCBFs appears appropriate for the upper stories of SCBF frames, but
underestimates the inelastic story drift of the bottom stories by a considerable amount for all but the
tallest structures.
The overstrength factor is used to define the required resistance of certain capacity-controlled
components (as opposed to deformation-controlled components) to prevent failure. In this research
study, the overstrength factor was computed as the ratio of the maximum story shear computed from
the dynamic analysis corresponding to the 2/50 hazard level to the design story shear computed
using the equivalent lateral force method.
Figure 12(a)–(c) shows the mean and the standard deviation values of the overstrength factor for
three-story, nine-story and 20-story buildings, respectively. The overstrength values also decrease
with a decrease in the R factors. It is likely that the lateral overstrength is induced by the difference
between the design and computed periods as shown in Table I.
For the three-story building model, the overstrength values were similar over the stories. For the
nine-story and 20-story building model, the overstrength factors in the lower stories were similar to
the values computed in the three-story building models and increased in upper stories. The increased
values in the upper stories likely resulted from selection of a less efficient brace cross section,
resulting in larger design overstrength involved. With an R factor of 3 for low-rise buildings, and an
R factor of 6 for the mid-rise and high-rise structures (as well as neglecting the increase of the
overstrength at the upper stories for mid-rise and high-rise buildings), an overstrength value of 2 is
recommended, which agrees with current design code.
Figure 12. The means and standard deviations of the overstrength factors for (a) three-story, (b) nine-story,
and (c) 20-story buildings with various R factors.
DOI: 10.1002/eqe
7. COMPARISON OF THE RESULTS USING THE PROPOSED AND FEMA P695 PROCEDURES
Comparison of the FEMA P695 methodology with the alternate method described here raises several
major concerns about accuracy and reliability of the FEMA P695 procedure for evaluation of braced
frames. In particular
(1) The probability of the potential collapse estimated by the P695 methodology was significantly
higher for the 20-story building models than for the three-story building models. In stark contrast,
the alternate analysis procedure clearly showed that the three-story frames had much larger
story drift demands with increased structural damage and brace fracture (for the hazard level
corresponding to the 2%-in-50-year event). The later finding is consistent with findings from
earthquake reconnaissance [22].
(2) The FEMA P695 procedure would suggest that a relatively smaller R value is required for the
20-story (high-rise) buildings compared with three-story building. This is clearly different than
the large body of existing knowledge of earthquake engineering [21, 22], in direct opposition to
the alternate analysis performed here.
(3) The FEMA P695 procedure is sensitive to the scaling methods used in IDA evaluation as shown
by the different results achieved from scaling methods M1 and M2 in Table II. This sensitivity
increases the ambiguity of the method.
The results show that of the two evaluation methods, the proposed evaluation procedure is
more reliable, for two reasons. First, the FEMA P695 focuses on an imperfect collapse
assessment, which is difficult to validate. The alternate method evaluates validated multiple limit
states to capture a wide range of performance states that are as follows: (i) documented by
experimental observations; (ii) important to the building owner and occupants; and (iii) ignored
by the P695 method.
Second, the alternate method provides a rational mechanism for scaling the earthquake ground
motions to meet both the seismic hazard level and expected structural performance. Higher mode
effects are considered in the evaluation of the high-rise buildings; effects of structural damage on the
dynamic response are incorporated as appropriate.
Both methods used a modeling approach, which employed the most accurate modeling of braced
frame behaviors used to date, because the models were validated to simulate the full range of
behaviors and damage states, including brace fracture observed in experiments. This final point
emphasizes the primary limitation with the P695 procedure. Focusing solely on the stability or
toppling aspects of collapse does not provide a reliable method to provide consistent performance
among building archetypes. Collapse prediction is a highly uncertain response mechanism, difficult
to fully simulate in a nonlinear dynamic response analysis, and its validation is not possible with the
current experimentation. The P695 procedure extends this uncertain prediction with statistical
predictions that imply great accuracy and reliability of the prediction. Engineers and building
owners are more concerned by the probability of brace fracture, reparability of the structure, and
other structural performance states that impact building performance. The alternate procedure
provides a mechanism to estimate structural performance for SCBF systems and provide a consistent
level of performance using the R value method.
8. SUMMARY AND CONCLUSIONS
To investigate the impact and appropriate values for R factors of SCBF systems, three-story, nine-story and
20-story SCBF buildings, which were designed on the basis of the equivalent lateral force procedure
with various R factors, were evaluated by nonlinear dynamic analysis. All of the buildings were
idealized multi-story structures designed for the seismic hazards of Seattle, Washington. The buildings
in the study were designed to account for the lower damping of steel structures (2%-damping) by
amplifying the design base shear by 1.25 (=1/0.8), which is not required in NEHRP (2003), but 2%
damping is smaller than that typically measured in steel buildings. As a result, the resulting design R
values, with the consideration of this amplification, as adopted in the study, would be 0.8 times those
DOI: 10.1002/eqe
obtained with the normal 5% damped spectrum. Two methods were used to evaluate the appropriate R
factor values as well as their impact on the seismic performance and design parameters for SCBFs.
Initially, the P695 methodology was used. Finally, an alternative performance-based method was
developed and used to evaluate the SCBF models.
The ground motions selected for analysis were from two bins, including the 2% and 10% in 50 years
Seattle SAC ground motions. Both methods used state-of-the-art line-element modeling, which were
experimentally validated. The modeling approach was also used to accurately predict the yielding
and buckling behavior of the brace, the post-buckling behavior of the systems, the onset of brace
fracture, and the post-fracture behavior of the system.
The research was carried out by first evaluating a pair of three-story and 20-story SCBFs designed
with R factors of 3 and 6 using the methodology of FEMA P695 (ATC-63). These results were
compared with the alternative evaluation process, which used an appropriate scaling method of
ground motions to meet the higher mode and performance of a range of SCBF archetypes, including
a series of SCBF buildings (three-story, nine-story and 20-story buildings).
The results were compared with the P695 evaluated results to determine appropriate design
parameters for SCBFs, including R, Cd, and Ωo The conclusions of the modeling requirements,
ground motion scaling procedures, effectiveness of the P695 procedure and the proposed
procedure follow.
• The contribution of gravity frames to the lateral resistance had considerable impact on and reduces
the seismic responses for the low-rise buildings (three-story buildings) and needs to be considered
in the performance evaluation. For mid-rise and high-rise buildings (nine-story and 20-story
buildings in the study), the effect of the gravity frame is relatively small.
• The evaluation of SCBFs using the FEMA P695 methodology with the IDA procedure to assess
collapse potential led to uncertainty of estimating the R factors. This uncertainty results from the
variations of the resulting IDA curves by different scaling approaches of ground motions as well
as the difficulty in predicting collapse.
• In the alternate performance evaluation procedure, the proposed scaling method effectively
reflects the characteristics of the structural systems to best represent the ground motion intensity
at certain hazard levels. For first-mode-governed structures such as the low-rise and mid-rise
buildings, the proposed scaling method provided a reasonable fit of the ground motion response
spectra to the design response spectra for the variation of structural periods due to brace fracture.
For the taller buildings, which are mainly governed by the second mode, the scaling method
assures the ground motion response spectra reasonably matching the design response spectra at
the first two mode structural periods.
• Seismic performance evaluation, particularly collapse, entails many variabilities and uncertainties.
The alternative evaluation procedure is much better able to address these uncertainties and provide
a better picture of SCBF performance than the FEMA P695 method.
• By using the accurate analytical models, including fracture prediction and proper scaling approach
of ground motion records, the seismic performance of the SCBFs can be assessed at multiple
damage levels for multiple hazard levels. This fulfills the objectives of performance based earthquake
engineering and provides a mechanism for a more consistent design among building archetypes.
Solely focusing on collapse does not fulfill this objective or provide this consistency.
• The analytical results show that the appropriate R factors actually varied with the number of
stories. The results suggest that if consistent safety against collapse is to be achieved, low-rise
SCBFs (three-story) require smaller R value for low-rise buildings. An R factor of 3 is suggested.
The results showed acceptable results for the mid-rise and high-rise SCBF buildings (nine-story
and 20-story) designed with an R-value of 6.
• The results from all of the dynamic analyses were compiled revealing that both the deflection
amplification and overstrength factors varied with the R factors. The current value of the deflection
amplification, Cd, (Cd = 5 for SCBFs) is appropriate for the upper stories of SCBF frames, but the
value underestimates the inelastic story drift of the bottom stories for all but the tallest structures.
Using the proposed R factors of 3 (low-rise SCBFs) and 6 (mid-rise SCBFs), an overstrength factor
of 2 was deemed appropriate, which agrees with current design code.
DOI: 10.1002/eqe
ACKNOWLEDGEMENTS
This research was funded by the National Science Foundation under grants CMS-0619161, NEESR-SG
International Hybrid Simulation of Tomorrow’s Braced Frames. Supplemental funding was provided
by AISC. Advice and guidance was provided by Stephen Mahin, Keh-Chyuan Tsai, Tom Schlafly
(AISC), Tim Fraser, Walterio Lopez (Rutherford and Chekene), Larry Muir (CIVES) and Rafael Sabelli
(Walter P. Moore). The advice and financial support of these individuals and institutions is greatly appreciated.
REFERENCES
1. AISC. (2005a). Manual of Steel Construction, Load and Resistance Factor Design, (3rd edn), American Institute of
Steel Construction: Chicago, IL.
2. Lehman D.E., Roeder CW. Improved Seismic Design of Concentrically Braced Frames and Gusset Plate Connections.
ASCE, 2008.
3. Roeder CW, Lumpkin E.J., Lehman DE. Balanced design procedure for special concentrically braced frame
connections. Elsevier, Journal of Constructional Steel Research 2011a; 67(11)1760–72.
4. Lumpkin, E.J., Hsiao, P-C, Roeder, C.W., Lehman, D.E., Tsai, C-Y, Wu, A-C, Wei, C-Y, Tsai, K-C. Investigation of
the seismic response of multi-story braced frames. Journal of Constructional Steel Research 2012; 77: 131–144.
5. Johnson S. Improved seismic performance of special concentrically braced frames. MS thesis, University. of
Washington, Seattle, 2005.
6. Herman D. Further improvements on and understanding of SCBF systems. MS thesis, University of Washington,
Seattle, 2006.
7. Kotulka BA. Analysis for a design guide on gusset plates used in special concentrically braced frames. MS thesis,
University of Washington, Seattle, 2007.
8. Powell JA. Evaluation of special concentrically braced frames for improved seismic performance and constructability.
MS thesis, University of Washington, Seattle, 2010.
9. Roeder, C.W., Lehman, D.E., Clark, K, Powell, J., Yoo, J-H, Tsai, K-C, Lin, C-H, and Wei, C-Y. Influence of gusset
plate connection and braces on the seismic performance of x-braced frames. Earthquake Engineering and Structural
Dynamics 2011b; 40(4):355–74.
10. ATC-3-06 report. Tentative Provisions for the Development of Seismic Regulations for Buildings. ATC, 1978.
11. FEMA 450. NEHRP recommended provisions for seismic regulations for new buildings and other structures. FEMA
450-1/2003 Edition, Part 1: Provisions, Federal Emergency Management Agency, Washington, D.C, 2004a.
12. FEMA P695. Quantification of building seismic performance factors FEMA P695 ATC-63 project report. Federal
Emergency Management Agency, Washington, D.C., 2008.
13. FEMA 355C. State of the art report on systems performance of steel moment frames subject to earthquake ground
shaking. Federal Emergency Management Agency, Washington, D.C., 2000.
14. AISC. Seismic provisions for structural steel buildings. ANSI/AISC Standard 341-05, American Institute of Steel
Construction, Chicago, IL, 2005b.
15. Hsiao P-C, Lehman D.E., Roeder C.W. Improved analytical model for special concentrically braced frames. Journal
of Constructional Steel Research 2012a; 73:80–94.
16. Hsiao P-C, Lehman D.E., Roeder C.W. A model to simulate special concentrically braced frames beyond brace
fracture. Earthquake Engineering and Structural Dynamics 2012b; 42: 183–200.
17. Liu J, Astaneh-Asl A. Moment-rotation parameters for composite shear tab connections. Journal of structural
engineering, ASCE, 2004; 130(9):1371–1380.
18. Somerville P, Smith N, Punyamurthula S, Sun J. Development of ground motion time histories for phase 2 of the
FEMA/SAC steel project. SAC Background Document, Report No. SAC/BD-97/04, 1997.
19. Baker JW, Cornell CA. 2006. Spectral shape, epsilon and record selection. Earthquake Engineering. & Structural
Dynamics 34(10):1193–1217.
20. FEMA 355F. State of Art report on performance prediction and evaluation of steel moment-frame buildings.
Federal Emergency Management Agency, Washington, D.C., 2000.
21. Chen C-H, Mahin S. 2010. Seismic collapse performance of concentrically steel braced frames. Proceedings of the
Structures Congress and 19th Analysis and Computation Specialty Conference, Orlando, Florida.
22. AIJ. Preliminary reconnaissance report of the 1995 Hyogoken-Nanbu earthquake. Architectural Institute of
Japan, 1995.
23. ATC-58. Fragility curves for concentrically braced steel frames with buckling braces. Applied Technology
Council, 2009.
DOI: 10.1002/eqe

EESD-Hsiao Etal 2013

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EESD-Hsiao Etal 2013

Uploaded by

Copyright:

Available Formats

EARTHQUAKE ENGINEERING & STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2013; 42:1547–1564

Evaluation of the response modiﬁcation coefﬁcient and collapse

Po-Chien Hsiao*,†, Dawn E. Lehman and Charles W. Roeder

Received 14 February 2012; Revised 30 October 2012; Accepted 15 January 2013

Copyright © 2013 John Wiley & Sons, Ltd.

2. DESIGN OF THE MODEL BUILDINGS

Table I. Design parameters and periods.

Three-story 3 9379 9011 0.390 0.31 0.34

3. SYSTEM MODELING INCLUDING EFFECT OF GRAVITY FRAMES

4. COLLAPSE ASSESSMENT USING FEMA P695 METHODOLOGY

Model buildings 3-Story 20-Story 3-Story 20-Story

5. MULTI-PERFORMANCE LEVEL EVALUATION PROCEDURE

Sa1;t Sa2;t Sa3;t

6. DEFLECTION AMPLIFICATION AND OVERSTRENGTH FACTORS

8. SUMMARY AND CONCLUSIONS

You might also like