1

Bulletin of the New Zealand Society for Earthquake Engineering, Vol. 54, No. 1, March 2021

SEISMIC COLLAPSE SAFETY OF REINFORCED

CONCRETE MOMENT RESISTING FRAMES
WITH/WITHOUT BEAM-COLUMN JOINT DETAILING

Naveed Ahmad1, Muhammad Rizwan2, Muhammad Ashraf3,

Akhtar Naeem Khan3 and Qaisar Ali3
(Submitted September 2019; Reviewed January 2020; Accepted November 2020)

ABSTRACT
FEMA-P695 procedure was applied for seismic collapse safety evaluation of reinforced concrete moment
resisting frames with/without beam-column joint detailing common in Pakistan. The deficient frame lacks
shear reinforcement in joints and uses concrete of low compressive strength. Shake-table tests were performed
on 1:3 reduced scale two-story models, to understand the progressive inelastic response of chosen frames and
calibrate the inelastic finite-element based models. The seismic design factors i.e. response modification
coefficient, overstrength, ductility, and displacement amplification factors (R, 0, Rμ, Cd) were quantified.
Response modification factor R = 7.05 was obtained for the frame with beam-column joint detailing while R
= 5.30 was obtained for the deficient frame. The corresponding deflection amplification factor Cd/R was
found equal to 0.82 and 1.03, respectively. A suite of design spectrum compatible accelerograms was obtained
from PEER strong ground motions for incremental dynamic analysis of numerical models. Collapse fragility
functions were developed using a probabilistic nonlinear dynamic reliability-based method. The collapse
margin ratio (CMR) was calculated as the ratio of seismic intensity corresponding to the 50th percentile
collapse probability to the seismic intensity corresponding to the MCE level ground motions. It was critically
compared with the acceptable CMR (i.e. the CMR computed with reference to a seismic intensity
corresponding to the 10% collapse probability instead of MCE level ground motions). Frame with shear
reinforcement in beam-column joints has achieved CMR 11% higher than the acceptable thus passing the
criterion. However, the deficient frame achieved CMR 29% less than the conforming frame. This confirms
the efficacy of beam-column joint detailing in reducing collapse risk.

INTRODUCTION and ductility factor Rμ, i.e. R = 0  Rμ [7, 8]. The overstrength
factor is calculated as the ratio of the maximum strength of a
The seismic design procedures given in code use site-specific
fully yielded structure to the design base shear calculated
5% damped elastic acceleration response spectrum for
following the seismic code (Figure 1). Uang and Maarouf [9]
calculating spectral acceleration at the fundamental period of
analysed a six-story RC moment frame, subjected to the 1989
structures, which is reduced by response modification factor R
Loma Prieta earthquake, and found 0 = 1.90. An overstrength
for computing the reduced level base shear force [1]. For
example, the ASCE 7-16 [2] specifies R = 8.0 for a special 0 = 2.20 was suggested by Hwang and Shinozuka [10] based
moment-resisting RC frames to reduce the elastic base shear on the analysis carried out on a four-story RC intermediate
demand on the structure. The reduced base shear force is then frame located in seismic Zone 2 as per the Uniform Building
distributed over the height of the structure for calculating lateral Code (UBC). Further findings on the overstrength factor can be
force at floor levels. The elastic deformation demand on found in relatively more recent research [11-15]. Mwafy and
structure, obtained under reduced level static lateral forces, is Elnashai [11] performed nonlinear static pushover analysis and
modified with the deflection amplification factor Cd to calculate response history analysis of eight-story reinforced concrete
inelastic deflection demand. The Cd factor is taken equal to R irregular frames, twelve-story regular frames, and eight-story
as suggested in the EC8 [3] and NBCC [4], or a fraction of R as regular frame-wall structures and found the corresponding
suggested in the ASCE 7-16 [2]. overstrength factors varies from 2.10 to 2.60, 2.14 to 3.04, and
2.30 to 3.86, respectively. Likewise, reinforced concrete frames
Response modification factor R was initially proposed based on and frame-wall structures studied by Elnashai and Mwafy [12]
the fact that well-detailed structures sustain large inelastic exhibited an overstrength factor above 2.0. The response
deformations under lateral loads without collapse and develop history analysis gave a higher overstrength factor in comparison
lateral strength above their design strength [5]. Generally, the R to the nonlinear static pushover analysis procedure. Massumi et
factor intends to take into account the energy dissipation, al. [13] investigated reinforced concrete frames from single-
overstrength, and ductility capacity of the structure [6]. It is the story to ten-stories and found that low-rise buildings had higher
ratio of the elastic force VE, that would be developed in structure overstrength in comparison to the high-rise buildings while the
for design basis earthquake if the structure remains entirely number of bays did not affect the overstrength significantly.
linearly elastic, to the design base shear V calculated following Using the ultimate stress level design following the North
the seismic code (Figure 1). Relating structural damping with American Building Codes, the overstrength factor reduced from
response modification factor is not straightforward, therefore, it 3.30 to 1.70 for buildings with single-story to ten-stories,
is often expressed as a function of structure’s overstrength 0 respectively. Reinforced concrete frames designed and

1 Corresponding Author, Associate Professor, Department of Civil Engineering, UET Peshawar, Peshawar, naveed.ahmad@uetpeshawar.edu.pk
2 Lecturer, Department of Civil Engineering, UET Peshawar, Peshawar.
3 Professor, Department of Civil Engineering, UET Peshawar, Peshawar.
2

analysed following the Iranian Standard 2800 have exhibited dependent on the structural ductility. For example, a code-
overstrength factors from 2.63 to 2.359 for six-story to nine- conforming frame structure having an effective/yield period
stories, respectively [14]. The ductility factor Rμ component of greater than 0.70 sec and has global ductility of 3.0, the
R has been related to the translation ductility ratio of structures corresponding Sp = 0.70 and k = 3 for the frame. This will result
and was dependent also on the structural period. In the current in a total reduction factor of 4.30 for reducing the horizontal
New Zealand Standards for Structural Design Actions action coefficient. Various analytical models have been
NZS1170.5-2004 [16], the response modification factor, developed for calculating the Rμ factor for structures [17-21].
represented by k, is related to structural ductility ratio and All the previous researches have shown the dependence of the
period for reducing the elastic horizontal action coefficient. R factor on the structural system type, materials, load path,
This standard included structural performance factor Sp to structural inelastic mechanisms, and the type of loading.
further reduce horizontal action for structure, which is

Figure 1: Illustration of seismic response parameters (R, 0, Cd), defined in NEHRP recommended provisions (FEMA, 2004).

The inelastic deflection demand on a structure is computed by The majority of the previous researches did not study the effect
amplifying the elastic deflection demand with a deflection of the shear hinging mechanism on the seismic response
amplification factor Cd (Figure 1). The elastic roof deflection parameters, which is likely to occur during earthquakes in
(δE/R) is obtained through an elastic analysis of the frame under reinforced concrete moment resisting frames lacking shear
reduced level static lateral forces. The value of Cd is equal to R reinforcement in beam-column joints. Rizwan et al. [27] and
assuming the equal displacement principle, such as adopted in Ahmad et al. [28] performed shake table tests on two-story
EC8 or may be taken as a fraction of R i.e. Cd = 0.6875R reinforced concrete deficient frames (i.e. lacks shear
suggested in the ASCE 7-16 for special moment-resisting RC reinforcement in beam-column joint and uses low strength
frames. The ratio of Cd/R = 1.0 has been supported by much concrete). This caused shear hinges in beam-column joints of
past research conducted on a single-degree of freedom systems the frame under simulated ground motions that resulted in the
[19, 22, 23]. However, the distinct ratio has been suggested for reduction of lateral stiffness, strength, and translational
the multi-degree of freedom systems depending on the displacement ductility ratio of the frame. The present research
structural system (i.e. RC or steel frame), the desired deflection transforms this effect into the measure of collapse probability
(i.e. maximum roof drift or maximum inter-story drift), and the of frame for safety evaluation following the procedure given in
simulated lateral loads (i.e. response history analysis or modal FEMA-P695. The present research investigated a two-story
spectrum analysis or equivalent static force analysis) [9, 24-26]. reinforced concrete moment resisting frame that lacks shear
For example, a value of Cd/R = 0.7 to 0.9 has been proposed for reinforcement in beam-column joints and uses low strength
calculating the roof drift ratio while Cd/R = 1.0 to 1.50 has been concrete in comparison to the moment-resisting frame
proposed for calculating inter-story drift [9]. Unlike, Abou- conforming to the seismic design requirements. This will
Elfath and Elhout [25] found Cd/R = 1.0 for calculation of both signify the importance of good quality materials and beam-
maximum roof drift and inter-story drift demands. The column joint detailing in reducing the seismic collapse
NZS1170.5 suggests multiplying the elastic displacement by a probability of the moment-resisting frame.
scale factor equal to the structural ductility factor.
It is worth mentioning, the R factors given in the seismic code Research Program
for structures are derived on the empirical basis and qualitative In this research, the FEMA-P695 methodology [32] was
judgments. Its use is rational for well-detailed structures but adopted to quantify the building performance parameters and
may overestimate the seismic safety of structures that possess obtain the collapse margin ratio (CMR) for safety evaluation of
construction deficiencies due to unregulated constructions. As deficient RC frame and frame conforming to the design
observed, the response modification factor R may reduce even requirements. Shake table tests were conducted on 1:3 reduced
by 60 percent for RC moment frames incorporating scale two-story modern RC frames. The deficient frame
construction deficiencies e.g. low strength concrete, reduction incorporated construction deficiencies typically found in the
in longitudinal and transverse reinforcement, and lack of shear modern constructions of RC SMRFs/IMRFs in Pakistan. These
reinforcement in beam-column joints [27, 28]. Moreover, the lack shear reinforcement in beam-column joints and use
recent 2010 Chile and 2011 Christchurch earthquakes have concrete of low compressive strength. The models were tested
revealed the impending vulnerability of structures designed to through multiple-level excitations using the 1994 Northridge
modern codes; resulting in significant economic losses in the earthquake accelerogram to study the inelastic behaviour of
design basis earthquakes and observed with collapses under frames and calibrate the inelastic finite element based
earthquake ground motions higher than design basis earthquake numerical models. Moreover, the experimentally measured
[29-31]. displacement and acceleration response histories were analysed
 3

to derive frames’ base shear-displacement capacity curve. This few adjustments to be made to the fragility functions. This
was used to quantify seismic performance factors i.e. response included spectral shape factor (SSF) to account for the
modification coefficient, overstrength, and ductility factors (R, uncertainties in seismic demand due to the distinct spectral
0, Rμ). Generally, elastic analysis of finite element models is shape of ground motions for rare earthquakes [32]. The
performed under reduced static lateral forces for calculating the methodology also suggests four types of uncertainties,
elastic roof deflection demand. However, in the present including record-to-record variability, design requirements-
research, the elastic displacement (δE/R) of structures under related uncertainty, test data-related uncertainty, modeling-
reduced level base shear was directly obtained from the related uncertainty, to be included in the development of the
experimentally derived capacity curve. A suite of design collapse fragility function of the structure.
spectrum compatible accelerograms was retrieved from the
Haselton et al. [35, 36] evaluated code conforming RC SMRF
PEER strong ground motions, which were used for the inelastic
buildings using the collapse assessment method that has been
analysis of calibrated finite element based numerical models
incorporated into the FEMA P695/ATC-63 procedure. It was
prepared in the SeismoStruct program to obtain the structures’
found that the buildings designed to recent code provisions i.e.
inelastic roof deflection demand (δ). The ratio of inelastic to
ASCE 7-05 and ACI 318-05, have acceptable collapse safety,
elastic roof deflection demand gave an estimate of deflection
thus, able to pass the method and are deemed to have acceptable
amplification factor Cd (Cd = δR/δE). Collapse fragility
collapse safety. However, the provisions of ASCE 7-05 were
functions were developed for structures using a probabilistic
not appropriate for some tall buildings (20-story), where the
nonlinear dynamic reliability-based method [33]. The collapse
safety margin was about 31% less than the acceptable.
fragility functions were interpreted to obtain the seismic
Alternatively, these structures designed to the requirements of
intensity (SCT) at the 50th percentile of collapse, which was
ASCE 7-02 qualified to achieve the acceptable collapse. The
divided by the corresponding seismic intensity (SMT) for the
acceptable collapse probability was based on the collapse risk
maximum considered earthquake (MCE) ground motions. The
of archetype designs conforming to the current building code
MCE level ground motions were taken equal to 3/2 of design
provisions. Liel et al. [37] investigated non-ductile moment
basis earthquake (DBE) ground motions, as suggested in
frames, representative of the mid-1970s in California and
ASCE-7-16. For example, in present study it was computed as
designed to the seismic design procedure of 1967 UBC. This
3/2  0.40g = 0.60g). The ratio of SCT to SMT (i.e. 0.60g) gave
has shown 40 times higher mean annual frequency of collapse
an estimate of collapse margin ratio CMR (FEMA-P695, 2009).
in comparison to the code-conforming frames. The present
The calculated CMR was adjusted to take into account possible
study extends the methodology for the seismic collapse safety
sources of uncertainties in seismic hazard and structure’s
evaluation of a deficient frame (i.e. lacks shear reinforcement
response, giving adjusted collapse margin ratio (ACMR). This
in beam-column joints and uses concrete with low compressive
was critically compared with the acceptable ACMR10%
strength) in comparison to the moment-resisting frame
(ACMR10% corresponds to the adjusted collapse margin ratio
conforming to the design requirements. The considered
computed with reference to a seismic intensity corresponding
deficient moment resisting frame is abundantly found in the
10% percent collapse probability instead of using intensity
modern building stock of Pakistan.
corresponding MCE level ground), to assess the seismic safety
of both deficient frame and frame conforming to the design
requirements.

FEMA-P695 FOR CALCULATING COLLAPSE

MARGIN RATIO (CMR)
In the FEMA-P695 procedure, the structure seismic
performance factors are related to the MCE level ground
motions (Figure 2). The structure capacity curve is presented in
terms of spectral acceleration and spectral displacement.
Depending on the seismic code, the MCE level ground motions
may be defined directly e.g. as given in the IBC [34], or it can
be obtained by amplifying the DBE level ground motions by a
factor of 1.50 [34]. The methodology expresses the safety of
structure in terms of collapse margin ratio (CMR), as given in
Equation 1. It is calculated as the ratio of ground motion capable
to cause collapse to the ground motion corresponding to the
maximum considered earthquake (MCE).

ŜCT Figure 2: Seismic performance factors, as defined in the

CMR  (1) FEMA-P695 Methodology for computation of collapse
SMT
margin ratio.
where SCT is the spectral acceleration of ground motion causing
the structure to collapse. It is defined as the seismic intensity SELECTED REINFORCED CONCRETE FRAMES
corresponding to the 50th percentile of collapse probability of The selected test model considered in the present research is a
structure. SMT is the elastic spectral acceleration at the 1:3 reduced scale two-story frame, representative of modern
fundamental period T of structure for MCE ground motions. earthquake-resistant reinforced concrete moment-resisting
Equation 1 can be expressed also in terms of the spectral frames in a high seismic zone of Pakistan (Figure 3). The use of
displacement demands. The CMR refers to the amount of beams having dimensions 457 mm x 305 mm and columns
increase required in the MCE ground motions to achieve the having dimensions 305 mm x 305 mm, reinforced with grade
structure’s collapse probability of 50%. 60 steel bars, are common in modern low-rise frames in
Pakistan. The total seismic weight of the frame calculated is
The CMR is largely dependent on the seismic design procedure
equal to 633 kN, comprising dead load and 50 percent portion
(codes, standards), structure analysis, and construction of
of the live load. The basic material properties considered for the
structures. In addition to this, the FEMA P695 has suggested
design of the as-built frame are contained in Table 1. Such
4

frames are analyzed and designed following the static force prepared in finite-element based program SAP 2000 for modal
procedure given in the Building Code of Pakistan-Seismic analysis and calculation of design level forces (axial, shear, and
Provisions and the recommendations given in the ACI-318-05 bending) in beam/column members. Table 2 reports the basic
[38] for a special moment-resisting frame (SMRF). It is worth dynamic properties of the as-built prototype frame exhibiting
mentioning that the width of the column is not sufficient to the first modal period of the frame is 90% larger than the code
develop beam reinforcing into column joint using a standard permissible maximum period, and the first mode contributes
hook, therefore, longitudinal bars slip and pullout is common in 84% mass only.
such frames [27, 28]. A representative 2D numerical model was

Figure 3: Geometric and reinforcement details of reinforced concrete moment resisting frame. The deficient model was similar
but lacking ties in joints.

Table 1: Material properties considered for the design of as-built prototype reinforced concrete frame.

Compressive Young’s
Poisson’s Yield Strength Maximum Strength
Property Strength Modulus
Ratio (MPa) (MPa)
(MPa) (MPa)

Concrete 21 21.52 GPa 0.20 - -

Re-bars - 200 GPa - 414 621

Table 2: Dynamic properties of as-built prototype reinforced concrete frame.

Vibration Period
Effective Seismic Modal
Seismic Weight
Property Weight Participation (Sec.)
(kN) Factor
(kN)
Modal Max. Allowed

Value 633 528 1.19 0.80 0.42

The design level base shear force VD was calculated using factor. The considered moment-resisting frame was analysed
Equation 2: and designed for VD = 0.11W  70 kN. This was distributed
between the floors, considering the linear deflected shape of the
CV I model, such as 2/3rd was applied at the roof level and 1/3rd was
VD  W (2) applied at the first-floor level. Under the design level lateral
RT forces, the first-floor beam experienced a maximum bending
moment demand of 94 kN-m. Therefore, the beam was
where CV is the seismic coefficient, W is the seismic weight of
reinforced with 3#6 top/bottom longitudinal steel bars, which
the structure, I is the importance factor (which is 1.0 for
has a nominal moment capacity of 96 kN-m. The corresponding
standard occupancy structures), R is the response modification
 5

maximum moment experienced at the base of the ground-story Following the simple model idealization, all the linear
column was 70 kN-m. Therefore, the column was reinforced dimensions of beam, column, and slab and diameter of the steel
with 8#6 longitudinal steel bars. The columns and beams were re-bars were reduced by a scale factor SL = 3. Concrete for the
provided with #3 stirrups at a spacing equal to 76 mm. 1:3 reduced scale model was prepared with a mix proportion of
Beam/column members’ dimensions and reinforcements were cement, sand, and coarse aggregate having a maximum size of
kept the same on both the ground and first stories. Moreover, 3/8 in. (9 mm) to respect the aggregate scaling requirements for
the code recommended beam-to-column moment capacity concrete preparation for the test model. The ACI concrete mix
/ design procedure was followed for the preparation of concrete
ratio , where Mb is the moment capacity of beam and Mc
with compressive strength of 3000 psi (21 MPa) and 2000 psi
is the total moment capacity of columns (both below and above) (14MPa) for design conforming and deficient frames
meeting at a joint. This was found equal to 0.70 at first-floor respectively. A mix proportion of 1:1.80:1.60 (cement: sand:
connection, which is less than 1.0, thus, it is anticipated to aggregate) with a water-to-cement ratio equal to 0.48 was used
ensure capacity protection of top ends of columns (i.e. to avoid to achieve 3000 psi (21 MPa) and a mix proportion of
plastic hinge at the columns’ top end). Because code allows 1:3.50:2.87 (cement: sand: aggregate) with a water-to-cement
plastic-hinges to form at the base of ground-story columns, a ratio equal to 0.80 was used to achieve 2000 psi (14 MPa).
simple static equilibrium consideration will require the
columns’ top-end flexure capacity to be greater than 0.75 times Initially, special steel formworks were designed and prepared
the beam flexure capacity to avoid plastic hinging at the top end for all the components including the model base pad (footing),
of ground-story columns. In the present case, the ratio of flexure columns, and slab with provisions for in-plane and transverse
capacity of the top end of ground-story columns to the beam beams. The construction sequence included the preparation of
flexure capacity is 0.86, which is about 15% higher than the reinforced concrete base pads 22 in. width x 15 in. depth x 8 ft.
static equilibrium requirements. length (559 x 381 x 2439 mm) for both the models one after the
other, which were cured for 14 days with moist bags. The
EXPERIMENTAL PROGRAM column longitudinal steel #2 rebars ( 6.35 mm) were extended
into the pad and were bent with a standard hook. The concrete
Construction of 1:3 Reduced Scale Test Models strength of the base pad was kept equal to 5000 psi (34.50 MPa)
to secure rigidity of the base pad and avoid deflection under
The unidirectional seismic simulator Shake Table-1 with bending and shear actions. It was followed by the construction
footprints of 5 x 5 ft. (1.5 x 1.5 m) at the Earthquake of columns, construction of in-plane and transverse beams, and
Engineering Center of UET Peshawar has a payload capacity of slab monolithically for each of the models in series, which were
8 tons but can be reliably operated for shake table testing of cured for 14 days. Similarly, the next story columns, beams, and
structures up to weight, not more than 5 tons. Furthermore, the slab were constructed and cured. Both the models were left for
shake table can be operated with a transient motion with a 28 days to attain the design specified strength.
maximum acceleration of ±1.1g, the maximum velocity of
±1.1m/sec, and maximum displacement of ±125mm. The It is worth to mention that the model and prototype uses
considered frame was scaled down by a scale factor SL = 3 to essentially the type of the same materials (i.e. concrete and steel
reproduce the test model for shake table testing. For simplicity re-bars), which have similar stress-strain behavior and material
reason and because scaling stress-strain properties of both density (unit weight). The reduced scale test models were
concrete and steel re-bar materials for model construction are subjected to gravity and seismic mass less than the required
quite demanding and costly, a simple model idealization was mass following the similitude requirements. Therefore, the test
considered similar to the frame tested by Ahmad et al [28]. In model was provisioned with additional floor masses (1200 kg
simple model idealization, the material stress-strain properties for each floor) to meet the requirements of gravity and dynamic
essentially remain the same for both the prototype and model. mass simulation [39, 40], as reported in Figure 4.

Figure 4: Simulation of additional floor mass for test models.

6

Test Setup and Instrumentations of Test Models allowed the transducers to record 12 inches (305 mm) lateral
movement of the model with reference to the fixed steel
At the testing laboratory, the model was lifted through a 20-ton
reference frame. Figure 5 shows the final test model with the
overhead crane and placed on the tabletop of the seismic
instrumentation scheme for recording model response.
simulator. The model was secured firmly using 18 steel bolts of
½ inch (13 mm) diameter. The over-hanged portion of the base
pad was placed on a specially fabricated roller support, 20 ton Overhead crane support
Displacement
transducers 3
comprised of a 4-leg steel stool that was provided with 4#8 Accelerometer 3
(425mm) steel rods to allow the model lateral movement
during testing. The test model was supported by a wooden
scaffolding to facilitate placing and mounting of floor Load + Addi onal Mass
Displacement
transducers 2
additional masses (steel blocks), without disturbing the model.
The test model was instrumented with six accelerometers Accelerometer 2

having a maximum capacity of 10g and three displacement

transducers having a maximum capacity of 24 inches (610 mm).
Fixed frame Reference frame
Two uniaxial accelerometers (front and back) were installed on
each floor and base pad to record the in-plane response
acceleration of the test model. For in-plane lateral displacement
measurement, a fixed steel reference frame was erected in-lined Shake Table Accelerometer 1
with the model. The displacement transducers were mounted on Roller Support
Displacement
the reference frame while the transducer string was stretched by transducers 1

a half-length of 12 inches (305 mm) and attached to each floor

and base pad, keeping the table positioned at mid-way. This Figure 5: Test model instrumentation for the recording of
displacement and acceleration response histories.

(a) Northridge 1994 Acceleration Time History

(b) 5% Damped Acceleration Response Spectrum (c) 5% Damped Displacement Response Spectrum
Figure 6: Selected acceleration time histories of Northridge-1994 earthquake for shake table testing of models.

Input Excitation and Testing Protocols maximum displacement of 90 mm, and can approximately
excite the structure symmetrically in both positive/negative
A natural acceleration time history record of the 1994 directions. This acceleration time history has the additional
Northridge earthquake (horizontal component, 090 CDMG advantage of linearly scaling up to 1.0g without exceeding the
Station 24278 - PEER strong motion database) was selected as maximum displacement and velocity limits of the seismic
an input excitation after careful analysis of several simulator. Figure 6 reports the acceleration time history, 5%
accelerograms compatible with the regional tectonics i.e. active damped elastic acceleration response spectrum and
shallow crustal earthquakes, thrust fault mechanism and being displacement response spectrum for the selected earthquake
recorded at the source-to-distance in the intermediate-field record. After the shake table self-check run for system
condition. The chosen accelerogram has a maximum adjustment, the selected acceleration time history was applied
acceleration of 0.57g, a maximum velocity of 518 mm/sec, and
 7

to the test model with multiple excitations with linearly 130% of design motions. This indicates the overstrength of
increasing shaking amplitude (i.e. linearly scaling of materials in the plastic hinges of the beam at the top story. The
acceleration amplitudes). The scaling factors were chosen equal model was found in the incipient collapse state after this run,
to [0.05, 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80, 0.90, 1.00, however, it was able to resist base motion with shaking intensity
1.30], to push the structure from elastic to inelastic and near equal to 1.06g. The test model also tilted in the out-of-plane
collapse state. This testing scheme is similar to the incremental direction, therefore, further testing was not feasible. Figure 7
dynamic analysis of a model for a given acceleration time shows the observed damages of the design conforming model
history. Both the models were tested progressively and their under significant test runs.
damage behavior was observed after every run, the tests were
concluded when the test model was found in the incipient Deficient Frame Model
collapse state. After the self-check, this model was pushed laterally to a roof
drift of about 1.75% in the first significant run. The model was
Observed Damage Behavior of Test Models observed with slight cracks at the base of ground-floor columns.
Flexure cracks were observed also at the base of first-floor
Design Conforming Frame Model columns. The model was observed with diagonal cracks in the
The shake-table performs an auto-run to understand the joint panels at both the first-floor and second-floor levels.
dynamic characteristics of the model and develop a system Horizontal flexure cracks and slight vertical cracks were
transfer function so that the system can simulate base motion at observed also at the beam-ends at both the first-floor and
the tabletop following the input acceleration time history to the second-floor. Upon increasing the intensity of input excitation,
controller. Therefore, this model was first subjected to a self- the model experienced a roof drift of 2.57% under 30% run.
check excitation that pushed the structure laterally to about During this run, the existing damages in the model were further
1.88% roof drift. This happened under the seismic simulator’s exacerbated. Flexure cracks at the base of both ground-floor and
automatic run before subjecting the structure to multiple first-floor columns were further widened. Similarly, horizontal
excitations. The shaking intensity of this excitation was found and vertical flexure cracks in the beam at the first-floor level
to have a maximum acceleration of 0.60g. During this run, the further widened. Damages in the beam-column joint panels at
model was observed with significant flexure cracks in the first- the first-floor level were further exacerbated. Extreme damages
floor beam due to the flexure yielding of reinforcing steel and were observed in beam-column joints on the first story; a
the formation of the plastic mechanism at the beam-ends. Slight concrete chunk was about to detach from the joint panels. Upon
vertical cracks were observed in the beam at the beam-column further increasing the amplitude of input excitation, the model
interface, which was expected due to the strain distribution was laterally pushed to a roof drift of about 4.77%. The model
across the beam depth and beam longitudinal steel re-bars during this run was observed with concrete spalling at the top
slip/pullout. The longitudinal re-bars slip in similar beams was and bottom ends of both the ground-story and first-story
observed also during quasi-static cyclic tests performed on full- columns. Damages in the joint panels at the first-floor level
scale beams [41]. Flexure cracks were also observed at the base further increased, and it was observed with cover concrete
of ground floor columns. Slight flexure cracks were also spalling. Joint panels at the second-floor level were severely
observed in the second-floor beam. The test runs were damaged, and it was observed with concrete cover detachment
performed following the testing protocols, however, none of the and spalling. The model after this run was found in the incipient
tests caused any further damage to the model, except the test collapse state. Figure 8 shows the observed damages of the
runs with scaling factor equal to 1.0 (100% intensity) and 1.30 deficient frame model under significant test runs.
(130% intensity). Upon subjecting the model to input excitation In comparison to the design conforming frame, the deficient
of 100% intensity that resulted in peak input acceleration of frame was observed to deform laterally to larger roof drift under
0.62g, the damages in the model remained fairly the same. The similar input excitations. This indicates relatively lower lateral
previous damages in the model were further exacerbated upon stiffness of the deficient frame. The damage evolution has
subjecting the model to input excitation with an intensity of shown that damages in the deficient model were more specific
130% of input excitation. This resulted in a peak input to the beam-column joint region than the beam/column
acceleration of 1.06g. During this run, the model was observed members. It was because the deficient model lacks confining
with concrete spalling and core crushing at the base and top ties in beam-column joints and uses concrete with lower
ends of the ground-story columns due to excessive compressive compressive strength. Furthermore, the joint panels were
strain demand at the column sections. It is worth mentioning damaged under less shear demand (in transferring beam
/
that the beam-to-column moment capacity ratio was moments to columns) due to the lower principal tensile strength
less than 1.0. Moreover, the moment capacity of the ground- of the joint panel. Since the joint principal strength capacity
story column section at the top end was 15% higher than the primarily depends on the strength of core concrete that is related
static equilibrium requirements that ensure plastic hinges in to the compressive strength of concrete [42, 43].
beams and at the base of ground-story columns. The plastic Figure 9 reports the damage severity of the top story beam-
hinging at the top end of ground-story columns indicates the column joint under the last test run. The frame with joint
overstrength of materials in plastic hinges of the beam and detailing resisted base motions with a shaking intensity of
dynamic magnification of moment due to higher modes. Minor 1.06g, although the joints received slight damages due to
spalling was also observed at the base of first-story columns, material overstrength in plastic hinges of the beam, the joints
however, this was primarily due to the cover concrete crushing. still had the potential to resist shear. Unlike, the deficient frame
Due to proper anchorage of column longitudinal steel in the without joint detailing experienced extensive damages, cover
base pad, bar slip or pullout was not observed. Additionally, the concrete spalling and core concrete cracking/crushing, and was
model was observed with severe diagonally cracks in the found on the verge of collapse. This indicates the benefit of
transverse beam at the first-floor due to the twisting of the beam beam-column joints detailing in increasing the lateral resistance
that has induced torsion in the transverse beams. Slight diagonal of the moment-resisting frame.
cracks were observed in the beam-column joint region of the
top-story. Despite the shear reinforcement in the beam-column
joint, the joint incurred slight damages under base motions of
8

Figure 7: Damage observed in design conforming frame under significant test runs.

Figure 8: Damage observed in the deficient frame under significant test runs.
 9

Design Conforming Model, under 1.06g Deficient Model, under 0.69g

Figure 9: Damage incurred by beam-column joints under the last test run.

SEISMIC RESPONSE PARAMETERS (R, 0, R, Cd)

Vmax  g 
 2   max T,T1 
2
 y,eff  C0
Development of Force – Displacement Capacity Curves W  4  (3)

The data recorded through accelerometers and displacement where C0 relates the fundamental displacement to roof
transducers were corrected for the instrument coefficients, to displacement, Vmax/W is the maximum base shear resistance
convert the recorded voltage to accelerations (in terms of “g” normalized by weight of the structure, T is the fundamental
value) and displacement (in terms of “inch”). The floor period obtained using the code specified equation and T1 is the
acceleration and displacement response histories recorded modal period for the first mode and g is the acceleration due to
during each test run were analyzed for baseline correction and gravity, equals g = 9.81 m/sec2. T1 may likely be larger than T
filtering using SeismoSignal, to remove the unnecessary noise since T1 is based on the actual model properties of the structure
from the actual signal. A linear type baseline correction and a while T is calculated using the empirical formula given in the
Butterworth filter with Bandpass filter configuration with a code. This is confirmed by many experimental and numerical
frequency range of 0.10 Hz to 25 Hz were employed. The floor studies [44-46].
displacements were further corrected with the base pad
displacement time history, by subtracting the base pad 300
displacement from the floor displacements to obtain the floor Design Conforming Model
displacement time histories relative to the base of the model. 250
Base Shear Force (kN)

The lateral force-deformation capacity curves for the 200

corresponding prototype structures were calculated by first
transforming the model recoded data to the prototype using the 150
conversion factors (model-to-prototype) as per the similitude
requirements: the floor displacement was multiplied by a scale 100
factor SL = 3 and the floor forces were multiplied by a factor of Deficient Model
SL2 = 32 = 9. The floor accelerations were first multiplied by the 50
floor masses (including the additional block mass, self-weight
of slab and beams on the floor, and half mass of the columns 0
0 1 2 3 4 5 6
above/below the floor), to calculate the floor inertial forces. The Roof Drift (%)
floor inertial forces were summed to calculate the base shear
force. For each test run the maximum roof displacement and the Figure 10: Experimentally derived force-displacement
corresponding maximum base shear force were identified and capacity curves for two-story design conforming and
correlated to obtain the base shear force and displacement deficient RC frames.
capacity envelope curve. Figure 10 shows the capacity curves
obtained for both the design conforming and deficient frame The FEMA-P695 has suggested using the ASCE/SEI 41-06
models. It can be observed that the lateral stiffness, lateral load equation, i.e. Equation 4 given below, to calculate C0:
resistance, and ultimate displacement capacity of the deficient
frame were reduced due to the construction deficiencies. The N
reduction in lateral resistance is relatively more pronounced in m  x 1,x
comparison to stiffness and ultimate displacement capacity, C0  1,r 1
N
which is due to the joint shear hinging mechanism. (4)
m  x
2
1,x
1
Elasto-Plastic Idealization of Capacity Curves
The capacity curves were idealized as elastoplastic following where mx is the floor mass at x level, ø1,x is the ordinate of the
the suggestion of FEMA-P695. This involved the identification fundamental mode at floor levels, and ø1,r is the corresponding
of maximum base shear resistance of structure (Vmax) and the ordinate of the fundamental mode at the roof level. This
ultimate displacement capacity (u) corresponding to the base requires the modal analysis of the elastic frame model,
shear force (0.8Vmax), and computation of effective yield roof however, it can be approximated also using the code specified
displacement (y, eff) using Equation 3. lateral force distribution pattern, as it is a reasonable
consideration for low-rise structures.
10

In the present case, the design base shear is known, the and deficient models respectively. This suggests a reduction of
corresponding elastic displacement can be identified from the 25% in the response modification factor of the structure due to
experimental capacity curve, which can be used to calculate the the inclusion of construction deficiencies. It is worth
effective stiffness. The effective yield displacement can be mentioning that the derived response modification factors in the
calculated by dividing Vmax by the effective stiffness. Figure 11 present study are distinct from the response modification factors
presents the derived elastoplastic curves for both the design (R = 7.54 for the conforming model and R = 3.70 for the
conforming and deficient RC frames. A marginal extension of deficient model) derived by Rizwan et al. [27] for the same
the capacity curves was performed to reach the 0.80Vmax for frames. It is because Rizwan et al. [27] have used the energy-
calculating u. Both the conforming and deficient frames balance rule to idealize the capacity curves, which differ from
achieved the maximum lateral resistance approximately at the the idealization suggested by FEMA-P695.
same lateral displacement demand. However, the ultimate
displacement capacity of the deficient frame is less than the Deflection Amplification Factor
conforming frame, which is due to the shear hinging mechanism
of the deficient frame model. The same is not true for ductility, Elastic Deflection under Design Level Forces
as the deficient frame has exhibited a higher ductility ratio in The experimental force-displacement capacity curves were
comparison to the conforming frame. analyzed to obtain the elastic displacement (E/R) of frames
corresponding to the design level base shear (V). An elastic
Response Modification Factor displacement E/R = 20 mm and E/R = 26 mm were obtained
The experimental elastoplastic capacity curves were interpreted for conforming and deficient frames respectively. Under the
to calculate the seismic response parameters, particularly the same level of reduced base shear force, the deficient frame
overstrength factor 0 and response modification factor R. The deflects 30% more than the compliant frame.
overstrength factor was calculated as the ratio of Vmax to V, Inelastic Deflection through Response History Analysis
which gave 0 =3.61 for design conforming frame and 0 =
2.10 for the deficient frame. This shows a reduction of 42% in The inelastic deflection  under the design basis earthquake
0 due to the considered construction deficiencies i.e. lack of ground motions was calculated through inelastic response
confining ties in beam-column joints and using concrete with history analysis of frames. First, a representative finite element
low compressive strength. The structure response modification model was prepared in SeismoStruct (Figure 12) simulating the
factor can be approximated as, R = VE/V = VE/Vy  Vy/V = R  observed inelastic mechanisms of considered frames. Under
shake table tests, the design conforming frame was observed
0. The ductility factor R of the structure was calculated using
with beam yielding, and flexure cracking at the base of ground-
the ductility-dependent response modification factor, as
floor columns for design basis earthquake (Figure 8). In
proposed by Newmark and Hall [17], and given in Equation 5
SeismoStruct, the flexure mechanism of the frame is directly
through Equation 7.
modeled through a fiber-based beam element with distributed
For Short Period, T < 0.20 sec plasticity [47], employing force-based formulation [48, 49], to
simulate the geometric nonlinearity and material inelasticity.
R  1 (5) The beam/column members were modeled using an inelastic
force-based flexure beam-type 3D element (Figure 13). The
For Intermediate Period, 0.20 sec < T < 0.50 sec member sectional stress-strain state is obtained through the
integration of the nonlinear uniaxial material response of the
individual fibers (i.e. unconfined and confined concrete and
R  2 1 (6) steel fibers in which the section has been subdivided). The
flexure frame element is capable of modeling geometric
For Long Period, T > 0.50 sec nonlinearity and material inelasticity of beam/column members
under cyclic deformation and fully accounts for the spread of
R   (7) inelasticity along the member length and across the section
depth. The fiber-section modeling of columns accurately
where R is the ductility factor,  is the structural transitional simulates the lateral response degradation due to the variation
ductility ratio. For the considered frames Teff was found equal of axial loads in comparison to lumped plasticity modeling [50].
to 0.64 sec and 0.74 sec for design conforming and deficient The section was subdivided into a total of 100 fibers. Although
frames respectively, using the classical fundamental period any appropriate number of section fibers from 100 to 400 may
formula i.e. Teff = 2 (m/keff). This seems relatively less than the be chosen, selecting larger fibers are beneficial for response
fundamental period obtained using the modal analysis history analysis of structure undergoing higher nonlinearities to
technique in the SAP2000 program (Table 2). This is because avoid non-convergence issues. However, fewer fibers will
the experimental model also consisted of a slab that contributed facilitate a large number of analyses. A total of 5 integration
to the stiffness and strength of the model. The value of keff was sections were defined over the element, which is sufficient for
obtained from the experimentally derived capacity curves accurately modeling the softening behavior of force-based
(Figures 11 and 12). elements [51]. Experimental tests on full-scale beams have
The Building Code of Pakistan – Seismic Provisions (BCP-SP, revealed the contribution of longitudinal re-bars slip to member
2007) allows only a maximum of 2.50% story drift for the code deformation [41]. This was modeled through the assignment of
conforming frames subjected to design basis earthquake ground moment-rotational lumped plasticity hinges at the beam ends,
motions. This suggests ductility  for the computation of R to simulate the member deformation due to re-bar extension and
should be based on the maximum displacement of 183 mm and fixed-end rotation. The elastic-hardening moment-rotation
not the ultimate displacement corresponding to the incipient constitutive relationship obtained from experimental tests on
collapse state of the structure. For these considerations, R = full-scale beams [41] was used in the present study. Due to the
special reinforcement detailing of beam/column members, the
1.95 was obtained for design conforming frame and R = 2.54
modeling ignored the shear stress-strain response and
was obtained for deficient frame. The conforming frame
longitudinal re-bars buckling. The observed experimental
ductility factor obtained is relatively lower, which is due to the
response has confirmed the flexure yielding of beam/column
hardening effect in the model that resulted in relatively larger
members that was followed by concrete cover spalling and core
effective yield displacement. However, the total response
concrete crushing.
modification factor R is equal to 7.05 and 5.30 for conforming
 11

300

Vmax = 253 kN
250

0.8Vmax = 203 kN
Base Shear Force (kN)

200

150

100
V = 70 kN

50

yield disp = 94 mm ultimate disp = 400 mm

0
0 50 100 150 200 250 300 350 400 450
Roof Displacement (mm)

(Design Conforming Model)

300

250
Base Shear Force (kN)

200

Vmax = 146 kN
150

0.8Vmax = 117 kN

100
V = 70 kN

50

yield disp = 72 mm ultimate disp = 380 mm

0
0 50 100 150 200 250 300 350 400 450
Roof Displacement (mm)

(Deficient Model)

Figure 11: Elasto-plastic idealization of experimental force-displacement capacity curves using the FEMA-P695 procedure.

In the case of deficient frames, damages were observed also in rise to principal tensile and principal compressive stresses,
the beam-column joints, causing shear hinging (Figure 9). The which upon exceedance can result in cracking and damage of
beam-column joint modeling included idealizing joint panel joint panel zones. The proposed modeling technique idealizes
with stiff elastic flexure beam type elements provisioned with a the joint panel through a lumped plasticity moment-rotation
zero-length link element at the joint center that connects the spring and beam-column elements through the inelastic bending
joint horizontal element with the vertical elements through a element, respecting the global equilibrium, thereby simulating
rotational spring, connecting both beam and column to the joint the shear and moment in beam-column members and inelastic
through a common rotational spring. deformation in joint. This mechanism was modeled through a
lumped plasticity shear simulation hinge; a rotational spring
Under lateral loading, as shown in Figure 14, a joint panel in assigned with equivalent moment-rotation constitutive
RC frames is subjected to moment and shear at the beam-joint relationship that relates the joint shear with moment capacity of
and column-joint interfaces, besides the column axial load due spring [52, 53], as given in Equation 8 and 9 and described in
to gravity and lateral loads. Horizontal shear develops in the Figure 14.
joint panel zone under the influence of external loads that give
12

1 joint panel region. The joint peak shear strength was calculated
M j   jh A jh using Equation 10.
1 b j / Lb 1 (8)

jd Lc
 j  MPa   t tt t  JI 
0.15
 BI 
0.30
f 
c
' 0.75
(10)
A jh  b j  h j (9)
where t is the in-plane geometry parameter, which is 1.0 for
where Mj is the rotational spring moment capacity, jh is the
interior, 0.7 for the exterior, and 0.4 for knee connection; t is
joint shear strength corresponding to the diagonal tensile
the joint eccentricity parameter equal to (1-e/bc)0.67; t is the
strength of joint, Ajh is the joint shear area, bj is the joint panel
out-of-plane geometry parameter, which is 1.0 for
width and hj is the joint panel depth, Lb is the total length of the
subassemblies with zero or one transverse beam and 1.2 for
beam on left and right side of the joint between the contra-
flexure points, Lc is the total length of the column above and subassemblies with two transverse beams; t is an adjusting
below the joint between the contra-flexure points; jd is the factor to set the overall average of the ratio, it is equal to 1.31;
internal moment arm for the corresponding moment at the beam fc’ is the concrete compressive strength; JI = (j  fyj)/fc’ is the
ends (Figure 14). joint transverse reinforcement index, where j is the volumetric
joint transverse reinforcement ratio in the direction of loading
and fyj is the yield stress of joint transverse reinforcement; BI =
(b  fyb)/fc’ is the beam reinforcement index, where b is the
beam reinforcement ratio and fyb is the yield stress of beam
reinforcement. The shear deformation corresponding to the
peak shear strength model proposed by Kim and LaFave [55]
was calculated using Equation 11.
1.75
 j 
 j  rad.   t  t t  t BI  JI 
0.30
0.10
 ' (11)
 fc 

where t = (JPRU)2.1 is the parameter for describing in-plane

geometry, where JP represents the ratio of the number of not-
free in-plane surfaces around a joint panel to the total number
of in-plane surfaces of the joint panel, to consider possible
changes in joint shear strength according to in-plane geometry;
JP is 1.0 for interior connections, 0.75 for exterior connections,
and 0.5 for knee connections; t is the joint eccentricity
parameter equal to (1-e/bc)-0.6, which is 1.0 for no eccentricity;
t is the out-of-plane geometry parameter, which is 1.0 for
subassemblies with zero or one transverse beam and 1.4 for
Figure 12: Numerical model for the considered frame. subassemblies with two transverse beams; t = 0.0055 is an
adjusting factor to set the overall average of the ratio.
The joint shear hinge was assigned with multi-linear pinching The above shear-deformation models, Equations 10 and 11,
hysteretic behavior [54] with shear and deformation limits give an estimate of maximum shear and corresponding
obtained using the empirical model of [55]. This modeling deformation in the joint panel, shear-deformation for the other
technique was similar to the nonlinear modeling proposed limit states can be obtained using the proposed empirically
earlier by authors Ahmad et al. [56], which was further derived factors given in Table 3. Peak shear strength of 245 kN
extended for generalization to model RC frames with/without and 137 kN was calculated for conforming and deficient frames
beam-column joints, and also it was validated in predicting the respectively. The corresponding maximum deformation is
displacement response history of the frame for seismic 0.0024 and 0.0037 for conforming and deficient frames
excitation. respectively. Putting the limit state shear stresses in Equation 8
The joint shear strength-shear deformability model proposed by provides an estimate of the corresponding limit state moments
Kim and LaFave [55] largely depends on the geometry of the for the shear simulation hinge (Figure 15).
joint, compressive strength of concrete, longitudinal
reinforcement of the beam, and transverse reinforcements in the

Figure 13: Inelastic modelling of FE based frame element. The section is subdivided into unconfined and confined concrete and
steel fibers. Each fiber is assigned with an appropriate stress-strain relationship.
 13

Table 3: Limit state shear and deformation for joint panel [55] used for the constitutive relationship of shear-hinge simulation
spring (Figure 15). The peak values are obtained using Equations 10 and 11.

Cracking Yielding Maximum

Parameters
cr Std. y Std. max Std.

(MPa) 0.442 x j(max) 0.299 0.890 x j(max) 0.154 j(max) 0.153

(rad.) 0.0197 x j(max) 0.437 0.362 x j(max) 0.420 j(max) 0.410

where,
Tb represents the tension force due to the extension of the longitudinal rebar of the beam
Cb represents the compression force at the compressed toe of the beam
Mb represents the resulting moment in the beam
Vb represents the resulting shear force in the beam
Mc represents the resulting moment in columns
Vc represents the resulting shear force in columns
Pc represents the axial load on columns

Figure 14: Modelling and idealization of joint shear damage through moment-ration spring [53].

Figure 15: Constitutive relationship for joint shear simulation hinge. Limit state moments-rotations are computed for the
corresponding shear/deformation was calculated using Equations 10 and 11.
14

The modeling technique was tested and validated against the and fault mechanism, and also, the average spectrum closely
shake table tests conducted on the considered frames. Figure 16 matches the design spectrum. Figure 17 reports the comparison
presents the comparison of numerically predicted to the of the design spectrum and the average acceleration spectrum
experimentally observed roof displacement response of frames. of selected accelerograms. The mean spectrum of acceleration
The finite element models reasonably predicted the records is higher than the design spectrum for short periods but
displacement response time histories, exhibiting a similar trend the spectrum is matching for intermediate and long periods,
of displacement demand. The models predicted peak roof drift which is important for selected frames. The average spectrum
for conforming frame with error “(Numerical - Experimental)/ has PGA equal to 0.58g.
Experimental  100” about 3.14%, indicating slight The models were analyzed under the design basis of earthquake
overprediction, which is conservative for assessment purposes.
ground motions. Inelastic peak roof drift demand  = 1.61 mm
However, the error increased to 9.10% in the case of the
(Std. Dev. = 0.64) was obtained for the conforming model while
deficient model, which is due to the more complex behavior of
the deficient model exhibiting beam yielding, re-bar slips, and  = 1.99 mm (Std. Dev. = 0.67) was obtained for the deficient
joint shear hinging. The model was also capable to predict the model. This gave Cd = R/E = 5.78 for conforming model and
local damage mechanisms. Cd = 5.48 for the deficient model. Thus, Cd/R = 0.82 for design
conforming model and Cd/R = 1.03 for the deficient model. The
After the development of the finite element models, the models aforementioned calculated values are based on the median
were analyzed to design basis earthquake ground motions. A estimate of the roof drift demand. Comparing the derived
suite of spectrum compatible acceleration records was obtained deflection amplification factors with the ASCE-7-10
from the PEER strong ground motions database (Table 4). The recommendations of 0.6875R, an increase of 19% was observed
records were scaled and matched to the design spectrum, using in the deflection amplification factor for the conforming model
SeismoSelect that included a search engine to select records while an increase of 50.50% was observed in the deflection
from the online databases meeting the requirements of the amplification factor for the deficient model.
designer about the magnitude, source-to-site distance, site class,

Table 4: Ground motion records obtained from the PEER ground motions database. PGA* is a PGA of scaled record.

GM Record RSN Year Event Station MW PGA( g)*

1 63 1971 San Fernando, USA Fairmont Dam 6.61 0.69
2 125 1976 Friuli, Italy Tolmezzo 6.50 0.41
3 336 1983 Coalinga, USA Parkfield – Fault Zone 11 6.36 0.33
4 830 1992 Cape Mendocino, USA Shelter Cove Airport 7.01 0.64
5 952 1994 Northridge, USA Beverly Hills – 12520 Mulhol 6.69 0.68
6 1642 1991 Sierra Madre, USA Cogswell Dam – Right Abutment 5.61 0.75
7 2385 1999 Chi-Chi, Taiwan TCU071 5.90 0.66
8 4213 2004 Niigata, Japan NIG023 6.63 0.72
9 4455 1979 Montenegro, Yugo. Herceg Novi – O.S.D. Paviviv 7.10 0.44
10 4841 2007 Niigata, Japan Joetsu Yasuzukaku Yasusuka 6.80 0.44
11 5474 2008 Iwate, Japan AKT019 6.90 0.65

COLLAPSE MARGIN RATIO (CMR) reliability method (FORM) framework [58]. The collapse
probability (Pf) of the frame was calculated using Equation 12.
Collapse fragility functions were developed for both code-
compliant and deficient frames using the probabilistic nonlinear
dynamic reliability-based method (NDRM) of Ahmad et al. Pf    RI  (12)
[33]. The method involved a response history analysis of
numerical models for several earthquake motions [57]. The where  is the standard normal cumulative distribution function
selected acceleration records were anchored to common PGA and RI is the reliability index, calculated through FORM
and linearly scaled from 0.20g to 2.36g. The considered scaling approximation and using Equations 13 to 15.
also included design level ground motions i.e. PGA = 0.40g,
and MCE level ground motions i.e. PGA = 0.60g. Due to the C  D (13)
non-availability of reliable data from local sources, the present RI 
study approximated the MCE level ground motions as 3/2 of
 C2   D2
design level ground motions as proposed in ASCE 7-16. The
drift demands were obtained for various intensity levels (Figure C  LN  C   0.5 C2 (14)
18). The ultimate roof displacement capacity of both
conforming and deficient frames is shown to identify the D  LN   D   0.5 D2 (15)
records exceeding the structural capacity. The curves exhibit
lower uncertainties in the case of the elastic response of
where  is the median estimate of displacement capacity (C)
structures that increased with the onset of damage in the
structures. The level of uncertainties in drift demands is and displacement demand (D),  is the mean estimate of
relatively higher for the deficient frame due to its complex displacement capacity (C) and displacement demand (D),  is
damage mechanisms (i.e. yielding of beam/column members the logarithmic standard deviation of displacement capacity
and joint shear hinge). The drift demands for each intensity (C) and displacement demand (D).
level were combined with the drift capacity in a first-order
 15

200

150

100
Roof Displacement (mm)

50

‐50

‐100

‐150 Shake Table Test

SeismoStruct
‐200
15 20 25 30 35 40 45 50 55 60
Time (sec)

(a) Design Conforming Frame

200

150

100
Roof Displacement (mm)

50

‐50

‐100

‐150 Shake Table Test

SeismoStruct
‐200
23 28 33 38 43 48 53
Time (sec)

(b) Deficient Frame

Figure 16: Comparison of numerically predicted to experimentally observed roof displacement response.

The ultimate displacement capacity of models was obtained intensity measure. However, realizing the fact that for linear
from the experimentally idealized elastoplastic capacity curves scaling of accelerograms for IDA, a constant offset remains
(Figure 11). This gives ultimate displacement capacity u = 400 between PGA and SA(T). Therefore, the use of PGA as the
mm (roof drift = 5.58%) for the conforming model and u = 380 seismic intensity will not cause any significant difference in the
mm (roof drift = 5.25%) for the deficient model. These values fragility functions. Figures 19 and 20 show the collapse fragility
were assumed to define C for code-compliant and deficient functions derived for the conforming and deficient frames
frames respectively. Depending on the completeness and respectively.
reliability of test data, the FEMA-P695 suggests test data The algorithm developed by Baker [59] was used to perform
uncertainty of 0.10 to 0.50. Therefore, a value of 0.50 was taken fitting to the data, and also to extend the incomplete fragility to
for C. The values of D and D were obtained through the desired level of collapse probability is achieved. Median
incremental dynamic analysis of models. intensity of 2.16 with a logarithmic standard deviation of 0.75
The collapse probability obtained was correlated with the was obtained for the conforming frame while the median
seismic intensity for each target level of input excitation. The intensity of 1.50 with a logarithmic standard deviation of 0.70
FEMA-P695 has suggested the use of spectral acceleration was obtained for the deficient frame. Seismic intensity
SA(T) at the fundamental period of the structure as the seismic corresponding to the 50th percentile of the collapse of frames
16

was identified giving IMCT = 2.16 for the conforming frame and fragility functions already included uncertainties due to record-
IMCT = 1.50 for the deficient frame. The calculated CMR = to-record variability and test data related uncertainties (RTR,
IMCT/ IMMCE = 2.16/0.60 = 3.60 for the conforming frame and TD). It is interesting to note that the calculated uncertainty of
CMR = 1.50/0.60 = 2.50 for the deficient frame were obtained. collapse fragility till this is 0.75, which is in agreement with the
“(2RTR + 2TD)0.50 = 0.7” for RTR = 0.50 and TD =0.50, which
2.00 are maximum values suggested in FEMA-P695. It is worth
Mean Spectrum
mentioning that the larger record-to-record randomness is also
due to the type of scaling i.e. spectral matching of
Spectral Acceleration (g)

1.50 Design Spectrum

accelerograms, in comparison to the cloud analysis and
different scaling/matching techniques [57, 60]. Although the
1.00 obtained value is relatively higher, it is believed to be
conservative for seismic performance assessment of frames.
0.50 Nevertheless, the use of larger ground motions will provide a
more accurate estimate of randomness. The additional
uncertainty due to modeling-related and design requirements-
0.00 related collapse uncertainties was considered. Efforts were
0 1 2 3 4
made to carefully model the inelastic behavior of frames for
Time Period (sec)
simulating the design and observed damage behavior of the
model as closely as possible. Therefore, a medium level of
Figure 17: Comparison of average acceleration spectrum of
uncertainty of 0.35 was considered for DR and MDL, resulting
linearly scaled acceleration records to the design spectrum.
in the total system uncertainty of Total = 0.90. The total system
collapse uncertainty was used to adjust the collapse fragility of
frames.
2.50
Moreover, as the structural models were analysed using the
2.00 design spectrum compatible accelerograms, the FEMA-P695
suggested spectral shape factor was used for adjusting fragility
1.50 functions using Equation 17.
PGA (g)

1.00 ACMR  SSF  CMR (17)

Analysis exceeding
0.50 ultimate drift capacity where ACMR is the adjusted collapse margin ratio and SSF is
the spectral shape factor, which is dependent on the ductility
0.00 demand and period of structures [61]. The SSF takes into
0.00 2.00 4.00 6.00 8.00 10.00
account the effects in seismic response due to the distinct
Roof Drift Demand (%) spectral shape of rare ground motions compared to the design
ground motions. Using the FEMA-P695 suggested tables, SSF
(Design Conforming Frame)
= 1.10 for the conforming and SSF = 1.11 for the deficient
frames were obtained. Figure 21 shows the final adjusted
2.50 fragility functions for the conforming and the deficient frames.
Because of this, the collapse margin of frames updated to
2.00 ACMR = 4.00 and 2.78 for the conforming and the deficient
frames respectively.
1.50
PGA (g)

The FEMA-P695 also suggests calculating the acceptable

1.00 collapse margin ratio of ACMR10%. The intensity at the 10th
percentile collapse probability was identified, which was used
0.50 Analysis exceeding instead of intensity corresponding to MCE ground motions, to
ultimate drift capacity
calculate CMR10%. This gave CMR10% =3.24 and CMR10% =
0.00 3.20 for conforming and deficient frames respectively. This
0.00 2.00 4.00 6.00 8.00 10.00
consideration makes the evaluator accept a 10% collapse
Roof Drift (%) probability of structure for the MCE level ground motions.
CMR10% was also corrected with SSF to obtain ACMR10%. This
(Deficient Frame) gave ACMR10% equal to 3.60 for the conforming frame and 3.54
Figure 18: Roof drift demands obtained for frames using for the deficient frame.
selected records scaled to multiple-levels. The acceptable ACMR10% was compared with the ACMR
obtained relative to MCE level ground motions. It has been
A logarithmic standard deviation of 0.75 was obtained for observed that the conforming frame ACMR is greater than the
fragility functions. However, the FEMA-P695 suggested acceptable ACMR10% (ACMR/ACMR10% = 4.00/3.60 = 1.11),
uncertainties in the collapse were also considered to obtain the demonstrating the safety of the conforming frame. However, it
total system level uncertainty using Equation 16. has been observed that the deficient frame ACMR is less than
the acceptable ACMR10% (ACMR/ACMR10% = 2.78/3.54 = 0.79),
(16) which is 21% less than the required ratio of 1.0, demonstrating
Total   RTR  TD   DR   MDL the vulnerability of the deficient frame. Relatively comparing
the collapse margin ratio ACMR/ACMR10% of the code-
where Total is the total system collapse uncertainty, RTR is the compliant frame to deficient frame, a reduction of about 29%
record-to-record collapse uncertainty, DR is the design “(0.79 – 1.11)/1.11 × 100” was observed. This vulnerability of
requirements-related collapse uncertainty, TD is the test data- deficient frame is due to the improper construction i.e. lacking
related collapse uncertainty and MDL is the modeling-related confining ties in beam-column joints and have concrete with
collapse uncertainty. It is worth to mention that the derived low compressive strength.
 17

0.70

0.60

0.50
Probability of Exceedance

0.40

0.30

0.20

Seismic Intensity,
IMCT = 2.16g
0.10
MCE Ground Motions
Seismic Intensity,
IMMCE = 0.6g
0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0

Intensity Measure, PGA(g)

Figure 19: Derived collapse fragility function for conforming frame.

0.80

0.70

0.60
Probability of Exceedance

0.50

0.40

0.30

0.20
Seismic Intensity,
MCE Ground Motions
IMCT = 1.50g
Seismic Intensity,
0.10
IMMCE = 0.6g

0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0

Intensity Measure, PGA(g)

Figure 20: Derived collapse fragility function for the deficient frame.

CONCLUSIONS requirements of the ACI-318-05 achieved collapse margin ratio

larger than the acceptable, as ACMR = 4.00 is larger than
The present research adopted the FEMA-P695 methodology for ACMR10 = 3.60. Although the margin exceeds the acceptable
the seismic performance assessment of both design conforming collapse margin ratio by 11%, this increase is not very high, and
and deficient low-rise (two-story) RC moment frames. It was therefore, considering slightly conservative seismic response
evaluated through the computation of the collapse margin ratio modification factor e.g. R = 8.0 as suggested in ASCE 7-16, will
to quantify the margin between the seismic intensity capable of further ensure the safety of the structure. The prevailing
causing a 50% collapse probability relative to the MCE level deficiencies (i.e. lacking confining ties in beam-column joints
ground motions. The MCE ground motions were taken as 3/2 and have concrete of low compressive strength) in the existing
of design basis earthquake ground motions. Moreover, 10% of modern RC frame structures in Pakistan, which are due to the
collapse probability was considered acceptable under MCE lack of awareness and negligence in construction, reduces the
ground motions. The conforming model analysed to the static structural seismic performance factors.
force procedures of UBC-97 using R = 8.5 and designed to the
18

0.80

0.70

Deficient RC Frame
0.60 Conforming RC Frame
Probability of Exceedance

0.50

0.40

0.30

0.20

0.10

0.47 0.666 1.50 2.16

0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
Intensity Measure, PGA(g)

Figure 21: Adjusted collapse fragility functions for code-compliant and deficient frames. The value of acceptable ACMR for both
the conforming frame and the deficient frame is also shown.

This consequently reduced the collapse margin ratio of the REFERENCES

deficient frame by about 29% in comparison to the conforming
frame. Realizing the fact that the use of low strength concrete 1 ATC (1978). “Tentative Provisions for the Development of
and lack of ties in beam-column joints is not uncommon in the Seismic Regulations for Buildings”. Report No. ATC-3-06,
modern building stock in most of the developing countries, Applied Technology Council (ATC), Redwood City, CA,
therefore, further reducing R factor will be needed at the design USA, 505 pp.
stage to reduce the risk of structures. As the considered 2 ASCE 7-16 (2016). “Minimum Design Loads for Buildings
deficiencies reduced the response modification factor by 25%, and Other Structures”. American Society of Civil
therefore, a response modification factor equal to 5.0 is Engineers/Structural Engineering Institute, Reston, VA,
tentatively proposed to ensure the safety of the considered USA, 822 pp.
deficient frame structure. This study also highlights the need for 3 EC8 (2004). “Design of Structures for Earthquake
suggesting confidence factors/reliability factors in seismic code Resistance, Part 1: General Rules, Seismic Actions and
to permit design professionals to choose conservative values in Rules for Buildings. EN 1998-1”. European Committee for
the situation when 100 percent execution of code specified Standardization, Brussels, Belgium, 229 pp.
designs in the field is a challenge. The findings can be used also
4 NBCC (2010). “National Building Code of Canada”.
to raise awareness and support developing policies for
National Research Council of Canada, Ottawa, Ontario,
implementing structure-specific risk mitigation plans for
1203 pp.
reducing the potential risk of the modern existing building
stock. 5 ATC (1995). “Structural Response Modification Factors”.
ATC-19, Applied Technology Council: Redwood City, CA,
The findings presented herein are based on the experimental USA, 64 pp.
and numerical studies performed on two-story RC frames
6 NEHRP (2000). “Part 2: Commentary – Recommended
having one-bay, which represents the least redundant structural
Provisions for Seismic Regulations for New Buildings and
system [62]. The above findings may not be directly applicable
Other Structures”. Building Seismic Safety Council,
to frames with more vertical lines of columns for resisting
Washington, DC, 460 pp.
lateral load. For simplicity, only 11 acceleration time histories
were selected, however, the IDA procedure can be performed 7 Whittaker A, Hart G and Rojahn C (1999). “Seismic
with a limited number of ground motions for the derivation of response modification factors”. ASCE Journal of Structural
fragility functions through linear scaling of accelerograms to Engineering, 125 (4): 438-444.
multiple intensities levels. It is worth mentioning, this results in https://doi.org/10.1061/(ASCE)0733-
relatively lower uncertainties in response quantities. Moreover, 9445(1999)125:4(438)
the selected ground motions are relevant for intermediate/far- 8 Kappos AJ (1999). “Evaluation of behaviour factors on the
field conditions, the near-field conditions are not addressed. basis of ductility and overstrength studies”. Engineering
Structures, 9(9): 823-835. https://doi.org/10.1016/S0141-
ACKNOWLEDGMENTS 0296(98)00050-9
The authors are grateful to the anonymous reviewers for 9 Uang CM and Maarouf A (1993). “Safety and economy
carefully reviewing the manuscript and providing constructive considerations of UBC seismic force reduction factors”.
remarks that improved the quality of the final manuscript. Proceedings of the 1993 National Earthquake Conference,
 19

Central United States Earthquake Consortium, May 2-5, 24 Krawinkler H (1996) “Pushover analysis: why, how, when,
Memphis, USA, 121-130. and when not to use it”. Proceedings of the 65th Annual
10 Hwang H and Shinozuka M (1994). “Effect of large Convention of the Structural Engineers Association of
earthquakes on the design of buildings in eastern United California, Maui, Hawaii, pp. 17-36.
States”. Proceedings of the 5th US National Conference on 25 Abou-Elfath H and Elhout E (2019). “Evaluating the code
Earthquake Engineering, Earthquake Engineering approaches for estimating the seismic drifts of steel frame
Research Institute (EERI), July 10-14, Oakland, CA, USA, buildings designed under variable levels of seismicity”.
223-231. Bulletin of Earthquake Engineering, 17(7): 4169-4191.
11 Mwafy AM and Elnashai AS (2002). “Calibration of force https://doi.org/10.1007/s10518-019-00634-z
reduction factors of RC buildings”. Journal of Earthquake 26 Mollaioli F, Mura A and Decanini LD (2007). “Assessment
Engineering, 6(2): 239-273. of the deformation demand in multi-storey frames”. Journal
https://doi.org/10.1080/13632460209350416 of Seismology and Earthquake Engineering, 8(4): 203-219.
12 Elnashai AS and Mwafy AM (2002). “Overstrength and 27 Rizwan M, Ahmad N and Khan AN (2018). “Seismic
force reduction factors of multistory reinforced-concrete performance of compliant and noncompliant special
buildings. The Structural Design of Tall and Special moment-resisting reinforced concrete frames”. ACI
Buildings, 11(5): 329-351. https://doi.org/10.1002/tal.204 Structural Journal, 115(4): 1063-1073.
13 Massumi A, Tasnimim AA and Saatcioglu M (2004). https://doi.org/10.14359/51702063
“Prediction of seismic overstrength in concrete moment 28 Ahmad N, Shahzad A, Rizwan M, Khan AN, Ali SM,
resisting frames using incremental static and dynamic Ashraf M, Naseer A, Ali Q, and Alam B (2019). “Seismic
analysis”. Proceedings of the Thirteenth World Conference performance assessment of non-compliant SMRF
on Earthquake Engineering, August 1-6, Vancouver, BC, reinforced concrete frame: shake-table test study”. Journal
Canada, Paper No. 2826. of Earthquake Engineering, 23(3): 444-462.
14 Mohammadi R, Massumi A and Mashkat-Dini A (2015). https://doi.org/10.1080/13632469.2017.1326426
“Structural reliability index versus behavior factor in RC 29 Westenenk B, de la Llera JC, Jünemann R, Hube MA, Besa
frames with equal lateral strength”. Earthquakes and JJ, Lüders C, Inaudi JA, Riddell R and Jordán R (2013).
Structures, 8(5): 995-1016. “Analysis and interpretation of the seismic response of RC
http://dx.doi.org/10.12989/eas.2015.8.5.995 buildings in Concepcion during the February 27, 2010,
15 Aydemir ME and Aydemir C (2016). “Overstrength factors Chile earthquake”. Bulletin of Earthquake Engineering,
for SDOF and MDOF systems with soil structure 11(1): 69-91. https://doi.org/10.1007/s10518-012-9404-5
interaction”. Earthquake and Structures, 10(6): 1273-1289. 30 Kam WY, Pampanin S, Dhakal RP, Gavin H and Roeder
http://dx.doi.org/10.12989/eas.2016.10.6.1273 CW (2010). “Seismic performance of reinforced concrete
16 Standards New Zealand (2004). “NZS1170.5: Structural buildings in the September 2010 Darfield (Canterbury)
Design Actions. Part 5: Earthquake Actions ‐ New earthquakes”. Bulletin of the New Zealand Society for
Zealand”. Standards New Zealand, Wellington, 76 pp. Earthquake Engineering, 43(4): 340-350.
https://www.standards.govt.nz/sponsored- https://doi.org/10.5459/bnzsee.43.4.340-350
standards/building-standards/NZS1170-5 31 Kam WY and Pampanin S (2011). “The seismic
17 Newmark NM and Hall WJ (1982). “Earthquake Spectra performance of RC buildings in the 22 February 2011
and Design”. EERI Monograph Series, Earthquake Christchurch earthquake”. Structural Concrete, 12(4): 223-
Engineering Research Institute (EERI), Oakland, CA, USA, 233. http://hdl.handle.net/10092/9006
103 pp. 32 FEMA (2009). “FEMA P695: Quantification of Building
18 Krawinkler H and Nassar AA (1992). “Seismic design Seismic Performance Factors”. Federal Emergency
based on ductility and cumulative damage demand and Management Agency (FEMA), Washington, DC, USA, 378
capacities”. In Fajfar P and Krawinkler H (Eds.) Nonlinear pp.
Seismic Analysis and Design of Reinforced Concrete 33 Ahmad N, Ali Q, Crowley H and Pinho R (2014).
Buildings. Elsevier Applied Science, New York, USA. “Earthquake loss estimation of residential buildings in
19 Miranda E and Bertero VV (1994). “Evaluation of strength Pakistan”. Natural Hazards, 73(3): 1889-1955.
reduction factors for earthquake resistant design”. https://doi.org/10.1007/s11069-014-1174-8
Earthquake Spectra, 10(2): 357-379. 34 IBC (2019). “The 2018 International Building Code”.
https://doi.org/10.1193/1.1585778 International Code Council, ICC Publications, IL, USA,
20 Vidic T, Fajfar P and Fischinger M (1994). “Consistent 728 pp.
inelastic design spectra: strength and displacement”. 35 Haselton CB, Liel AB, Deierlein GG, Dean BS and Chou
Earthquake Engineering and Structural Dynamics, 23(5): JH (2011). “Seismic collapse safety of reinforced concrete
507-521. https://doi.org/10.1002/eqe.4290230504 buildings. Part I: Assessment of ductile moment frames”.
21 Borzi B and Elnashai AS (2000). “Refined force reduction Journal of Structural Engineering, 137(4): 481-491.
factors for seismic design”. Engineering Structures, 22(10): https://doi.org/10.1061/(ASCE)ST.1943-541X.0000318
1244-1260. https://doi.org/10.1016/S0141-0296(99)00075- 36 Haselton CB, Liel AB and Deierlein GG (2008). “Example
9 evaluation of the ATC-63 methodology for reinforced
22 Miranda E (2001). “Estimation of inelastic deformation concrete special moment frame buildings”. Proceedings of
demands of SDOF systems”. Journal of Structural the ASCE SEI Structures Congress, Vancouver, Canada.
Engineering, 127(9): 1005-1012. 37 Liel AB, Haselton CB and Deierlein GG (2011). “Seismic
https://doi.org/10.1061/(ASCE)0733- collapse safety of reinforced concrete buildings. ii:
9445(2001)127:9(1005) comparative assessment of nonductile and ductile moment
23 Veletsos AS and Newmark NM (1960). “Effect of inelastic frames”. Journal of Structural Engineering, 137(4): 492-
behavior on the response of simple systems to earthquake 502.
motions”. Proceedings of the 2nd World Conference on https://doi.org/10.1061/(ASCE)ST.1943-541X.0000275
Earthquake Engineering, Tokyo, Japan, 2: 895-912. 38 ACI (2005). “ACI 318: Building Code Requirements for
20

Structural Concrete”. American Concrete Institute (ACI), buildings”. Bulletin of the New Zealand Society for
Farmington Hills, MI, USA, 430 pp. Earthquake Engineering, 48(3): 157-169.
39 Quintana-Gallo P, Pampanin S, Carr AJ and Bonelli P https://doi.org/10.5459/bnzsee.48.3.157-169
(2010). “Shake table tests of under designed RC frames for 51 Calabrese A, Almeida JP and Pinho R (2010). “Numerical
the seismic retrofit of buildings – design and similitude issues in distributed inelasticity modelling of RC frame
requirements of the benchmark specimen”. Proceedings of elements for seismic analysis”. Journal of Earthquake
the Annual Conference of the New Zealand Society of Engineering, 14(S1): 38-68.
Earthquake Engineering, March 26-28, Wellington, NZ, https://doi.org/10.1080/13632461003651869
Paper No. 39. 52 Celik OC and Ellingwood BR (2008). “Modelling beam-
40 Morcarz P and Krawinkler H (1981). “Theory and column joints in fragility assessment of gravity load
Application of Experimental Model Analysis in Earthquake designed reinforced concrete frames”. Journal of
Engineering”. Technical Report, Report No. 50, John Earthquake Engineering, 12(3): 357-381.
Blume Earthquake Engineering Center, Department of https://doi.org/10.1080/13632460701457215
Civil and Environmental Engineering, Stanford University, 53 Alath S and Kunnath SK (1995). “Modeling inelastic shear
USA, 263 pp. deformations in RC beam-column joints”. Proceedings of
41 Ahmad N, Masoudi M and Salawdeh S (2020). “Cyclic the 10th Conference on Engineering Mechanics, University
response and modelling of special moment resisting beams of Colorado at Boulder, Boulder, Colorado.
exhibiting fixed-end rotation”. Bulletin of Earthquake 54 Sivaselvan M and Reinhorn AM (2001). “Hysteretic
Engineering. https://doi.org/10.1007/s10518-020-00987-w models for deteriorating inelastic structures”. ASCE
42 Pampanin S, Calvi GM and Moratti M (2002). “Seismic Journal of Engineering Mechanics, 126(6): 633-640.
behavior of R.C. beam-column joints designed for gravity https://doi.org/10.1061/(ASCE)0733-
only”. Proceedings of the 12th European Conference on 9399(2000)126:6(633)
Earthquake Engineering, September 9-13, London, UK, 55 Kim J and LaFave M (2012). “A simplified approach to
Paper No. 726. joint shear behavior prediction of RC beam-column
43 Priestley MJN (1997). “Displacement-based seismic connections”. Earthquake Spectra, 28(3): 1071-1096.
assessment of reinforced concrete buildings”. Journal of https://doi.org/10.1193/1.4000064
Earthquake Engineering, 1(1): 157-192. 56 Ahmad N, Shahzad A, Ali Q, Rizwan M and Khan AN
https://doi.org/10.1080/13632469708962365 (2018). “Seismic fragility functions for code compliant and
44 Ahmad ME, Ahmad N, Pervez S, Iqbal A, Khan AZ, Rahim non-compliant RC SMRF structures in Pakistan”. Bulletin
ME, Hassan W, Umer K and Khan K (2019). “Seismic of Earthquake Engineering, 16(10): 4675-4703.
performance evaluation of modern bare and masonry- https://doi.org/10.1007/s10518-018-0377-x
infilled RC SMRF structures”. Advances in Civil 57 Vamvatsikos D and Cornell C (2002). “Incremental
Engineering, 2019(Article ID 6572465): 15 pp. dynamic analysis”. Earthquake Engineering and Structural
https://doi.org/10.1155/2019/6572465 Dynamics, 31(3): 491-514. https://doi.org/10.1002/eqe.141
45 Crowley H and Pinho R (2004). “Period-height relationship 58 Der Kiureghian A (2005). “First- and second-order
for existing European reinforced concrete buildings”. reliability methods” Chapter 14 in Engineering Design
Journal of Earthquake Engineering, 8(1): 93-119. Reliability Handbook. Editors: Nikolaidis E, Ghiocel DM
https://doi.org/10.1080/13632460409350522 and Singhal S, CRC Press LLC.
46 Masi A and Vona M (2010). “Experimental and numerical 59 Baker JW (2015). “Efficient analytical fragility function
evaluation of the fundamental period of undamaged and fitting using dynamic structural analysis”. Earthquake
damaged RC framed buildings”. Bulletin of Earthquake Spectra, 31(1): 579-599.
Engineering, 8(3): 643-656. https://doi.org/10.1193/021113EQS025M
https://doi.org/10.1007/s10518-009-9136-3
60 Koopaee ME, Dhakal RP and MacRae G (2017). “Effect of
47 Pinho R (2007). “Nonlinear dynamic analysis of structures ground motion selection methods on seismic collapse
subjected to seismic actions” Page 63-89 in Advanced fragility of RC frame buildings”. Earthquake Engineering
Earthquake Engineering Analysis. Editor: Pecker A, and Structural Dynamics, 46(11): 1875-1892.
Springer. https://doi.org/10.1002/eqe.2891
48 Neuenhofer A and Filippou FC (1997). “Evaluation of 61 Baker JW and Cornel CA (2006). “Spectral shape, epsilon,
nonlinear frame finite-element models”. Journal of and record selection”. Earthquake Engineering and
Structural Engineering - ASCE, 123(7): 958-966. Structural Dynamics, 35(9): 1077-1095.
https://doi.org/10.1061/(ASCE)0733- https://doi.org/10.1002/eqe.571
9445(1997)123:7(958)
62 Tena-Colunga A and Cortes-Benitez JA (2015).
49 Spacone E, Ciampi V and Filippou FC (1996). “Mixed “Assessment of redundancy factor for the seismic design of
formulation of nonlinear beam finite element”, Computers special moment resisting reinforced concrete frames”. Latin
and Structures, 58(1): 71-83. https://doi.org/10.1016/0045- American Journal of Solids and Structures, 12(12): 2330-
7949(95)00103-N 2350. http://dx.doi.org/10.1590/1679-78251800
50 Koopaee ME, Dhakal RP and MacRae G (2015).
“Analytical simulation of seismic collapse of RC frame