Vol.15, No.

2 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION June, 2016

Earthq Eng & Eng Vib (2016) 15: 283-295 DOI:10.1007/s11803-016-0322-4

Multiplication factor for open ground storey buildings – a reliability

based evaluation

Haran Pragalath D.C†, Avadhoot Bhosale†, Robin Davis P‡ and Pradip Sarkar§
Department of Civil Engineering, NIT Rourkela, Odisha 769008, India

Abstract: Open Ground Storey (OGS) framed buildings where the ground storey is kept open without infill walls, mainly
to facilitate parking, is increasing commonly in urban areas. However, vulnerability of this type of buildings has been exposed
in past earthquakes. OGS buildings are conventionally designed by a bare frame analysis that ignores the stiffness of the infill
walls present in the upper storeys, but doing so underestimates the inter-storey drift (ISD) and thereby the force demand in
the ground storey columns. Therefore, a multiplication factor (MF) is introduced in various international codes to estimate
the design forces (bending moments and shear forces) in the ground storey columns. This study focuses on the seismic
performance of typical OGS buildings designed by means of MFs. The probabilistic seismic demand models, fragility curves,
reliability and cost indices for various frame models including bare frames and fully infilled frames are developed. It is found
that the MF scheme suggested by the Israel code is better than other international codes in terms of reliability and cost.

Keywords: open ground storey; seismic risk; multiplication factor; fragility curves; reliability index

1 Introduction American Standard ASCE/SEI 7 (2010) and New

Zealand Code NZS 1170.5 (2004) prohibit this type of
Vehicle parking space for people living in residential extremely irregular buildings in seismically active areas.
apartments in populated cities is a matter of major However, EC 8 (2003), Indian Standard IS 1893 (2002),
concern. Hence, the trend has been to utilize the ground Israel Standard SI 413 (1995), Nepal National Building
storey of the building itself for parking. This leads to a Code NBC 201 (1995), Bulgarian Seismic Code
category of buildings with no infill walls in the ground (1987) and Costa Rica Code FCEACR (1986), among
storey to allow for free movement of vehicles. This type others, permit this type of buildings when appropriate
of buildings which are infilled in all upper storeys but multiplication factors (MFs) are used in their design.
no infill walls in the ground storey is called open ground In conventional design the stiffness contribution of
storey (OGS) building. They are also known as ‘pilotis’, infill walls present in the upper storeys of OGS framed
or ‘stilted buildings’. But the functional advantage comes buildings is ignored in structural simulation (referred as
with a price: from the seismic performance point of view ‘bare frame analysis’) mainly to avoid the complexity
such buildings have greater vulnerability. The stiffness in the modelling. Such bare frame analyses result in
and strength of the ground storey of such buildings are underestimation of the bending moments and shear forces
significantly lower compared to adjacent upper storeys in the open ground storey columns, and as compensation
and hence these buildings are also referred to as extreme a multiplication factor (MF) is suggested by various
soft-storey buildings. Past destructive earthquakes international codes and in published literatures. A review
have demonstrated the vulnerability of OGS buildings. is given in Kaushik et al. (2006) and the expression/
A number of OGS framed buildings have experienced values suggested by various codes are shown in the Table 1.
severe damage during the 2001 Bhuj earthquake (Fig. 1). It has been reported in the literature that expressions or
Collapse of this type of buildings was predominantly values of MF recommended in design codes are mostly
caused by the formation of soft-storey mechanism in the empirical (Kaushik et al., 2006) and they lack rational
ground storey columns. basis (Fardis and Panagiotakos, 1997; Fardis et al.,
International Building Code (ICC IBC-2012), 1999).
Studies were also reported on quantifying the MF
Correspondence to: Robin Davis P, Department of Civil
by using a rational approach such as the elastic energy
Engineering, NIT Rourkela, Odisha 769008, India approach (Scarlet, 1997), nonlinear dynamic analysis
Tel: +919583066760 (Hashmi and Madan, 2008) and capacity assessment
E-mail: [email protected] procedures (Favvata et al., 2013). However, an extensive
†
PhD Scholar; ‡Assistant Professor; §Associate Professor literature review on the seismic behavior of OGS
Received August 27, 2014; Accepted April 16, 2015 buildings revealed that the seismic risk associated with
284 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol.15

(a) (b)
Fig. 1 Open ground storey building during 2001 Bhuj earthquake (www.nicee.org)

Table 1 Multiplication factors as per various codes and design details of the ground storey columns

Code Criteria Expression for MF Range of MF

 VRW 
Drastic reduction of infill in 1  
 VED 
EC 8 (2003) 1.5 – 4.68
any storey

IS 1893 (2002) Ki 2.5 2.5

 0.7
K i 1

SI 413 (1995) Ki 0.6R 2.1 – 3.0

 0.7
K i 1 R = 3.5 for low ductility (including two adjacent storey)
R = 5 for medium ductility

Bulgarian Code (1987) Ki 3 3.0

 0.5
K i 1

Note: Ki - Lateral stiffness of ith storey considered, R - Response reduction factor, ∆VRW – strength of infill in the storey above,
∑VED – sum of design lateral force in the storey

code-designed OGS buildings has not been adequately the global response of building subjected to a specified
addressed and their reliability incorporating uncertainties ground motion. Nonlinear time history analyses (N-LTH)
in structural properties and earthquake loading remains of the building response are performed for different
to be quantified. The goal of the present study is earthquakes, and the response is expressed in terms of
therefore to evaluate, in a probabilistic framework, the maximum inter- storey drift at any storey of the building.
comparative performance (seismic risk and reliability) Third part is the calculation of limit state probabilities
of various multiplication factors (MFs) suggested by of attaining a series of (increasingly severe) limit states,
selected international building codes. LSi, through the expression given in Eq. (1):

2 Methodology P  LSi    P  LSi | A  a  P  A  a  (1)

a

Seismic risk of the OGS buildings will be evaluated The conditional probability in Eq. (1) is denoted as
in terms of fragility curves and reliability indices, using the seismic fragility, FR(x). This conditional probability,
a simplified method reported by Ellingwood (2001). explicitly stated, is the probability of meeting or
The methodology is comprised of three parts. First part exceeding a specified level of damage, LS, given a
is the identification of the seismic hazard, P [A = a], ground motion which has a certain level of intensity, a.
described by the annual probabilities of specific levels This conditional probability is often assumed to follow
of earthquake motion. The seismic hazard at a site is a two parameter lognormal probability distribution
usually represented through a seismic hazard curve, (Cornell et al., 2002; Song and Ellingwood, 1999).
GA(x) which is a plot of P [A = a] versus the level of A point estimate of the limit state probability of state
earthquake motion (a) expressed in terms of PGA (Peak i can be obtained by convolving the fragility FR(x) with
Ground Acceleration). Second part is the evaluation of the derivative of the seismic hazard curve, GA(x), thus
No.2 Haran Pragalath D.C et al.: Multiplication factor for open ground storey buildings – a reliability based evaluation 285

removing the conditioning on acceleration from Eq. (1)  S 

as follows:  ln D 
SC
P( D  C | IM)     (5)
  D|IM   c2
2 
dG A  
P  LSi    FR  x  dx (2)  
dx
where, D is the drift demand, C is the drift capacity
The parameters of the fragility-hazard interface must at chosen limit state, SD and SC are the median of the
be dimensionally consistent for the probability estimate demand and the chosen limit state (LS), respectively.
to be meaningful. The reliability index corresponding to The dispersion, βD/IM, of the inter-storey drift (di) from
the probability of failure can be found by the following the time history analysis can be calculated using Eq. (6),
standard equation as shown below:

 pf   1  P  LSi  (3)
 ln(d )  ln(aIM
2
i
b
) 
 D|IM  (6)
where, βpf is the reliability index and φ ( ) represents the N 2
standard normal distribution.
where a(IM)b represents the mean inter-storey drift and
2.1 Seismic hazard analysis N is the number of building models. Details of the mean
ISD are discussed in Section 2.2.1. Dispersion in limit
The seismic hazard (GA) at a building site is displayed state capacities recommended by different standards and
through a complementary cumulative distribution literatures are not in agreement. ATC 58 (2012) suggests
function. The hazard function is the annual frequency that the dispersion in capacity, βc , depends on the
of motion intensity at or above a given level, x, to the building type and construction quality, so accordingly
intensity. Elementary seismic hazard analysis shows 0.10, 0.25 and 0.40 are recommended depending on
that at moderate to large values of ground acceleration, the quality of construction being good, fair and poor,
there is a logarithmic linear relation between annual respectively. Several studies (Wen et al., 2004; Satter
maximum earthquake ground or spectral acceleration, and Liel, 2010; Raghunandan et al., 2012; Meslem and
and the probability, GA(a), that specific values of D′Ayala, 2013) show that the dispersion in capacity, βc,
acceleration are exceeded. This relationship implies that is dependent on the damage states and the type of frames
GA is described by following an equation suggested by (infilled or bare) also. The value s of βc suggested by
Ellingwood (2001): a recent study by Meslem and D′Ayala (2013) are used
herein and shown in Table 2.
A fragility curve is obtained for each limit state.
GA ( x)  1  exp[( x / u ) k ] (4) The methodology adopted in this study has been used
by many researchers (Nielson, 2005; Rajeev and
in which u and k are parameters of the distribution. Tesfamariam, 2012; Davis et al., 2010).
Parameter k defines the slope of the hazard curve which, 2.2.1 Probabilistic seismic demand model (PSDM)
in turn, is related to the coefficient of variation (COV) in The present study uses a closed-form log-normal
annual maximum peak acceleration. expression for fragility computation. This expression
The seismic hazard curve for the north-east region requires a continuous function between the seismic
(Manipur) of India, which is one of most seismically demand (EDP) and the intensity measure (IM), popularly
active regions, developed by Pallav et al. (2012) and
shown in Fig. 2, is selected for the present study.
1
2.2 Development of fragility curves
Annual mean rate
of exceedance

0.1
The fragility function represents the probability
of exceedance of a selected engineering demand 0.01
parameter (EDP) for a selected structural limit state
(LS) for a specific ground motion intensity measure 0.001
(IM). The fragility curve represents the cumulative
probability distribution that indicates the probability that 0.0001
a component/system damage will be at least a certain 0 0.2 0.4 0.6 0.8 1.0 1.2
Peak ground acceleration (g)
level, as a function of a particular intensity measure. The
seismic fragility, FR(x) can be expressed in closed form Fig. 2 Seismic hazard curve of north east region, India (Pallav
using the following equation, et al., 2012)
286 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol.15

Table 2 Dispersion in limit state capacity, βc the RC elements are carried out as per IS 456 (2000)
and detailed as per IS 13920 (1993). The earthquake
Damage state
Frame loading is considered as per Indian standards (IS 1893)
Immediate occupancy Near collapse to have similar design base shear for all the frames. As
Bare frame 0.15 0.26 the focus of the present study is on the values of MF
Infilled frame 0.18 0.38 and their implication on fragility and reliability, only
the MF factors are taken from the various codes. The
reinforcement design philosophy followed by various
known as Probabilistic Seismic Demand Model (PSDM), codes is only marginally different and therefore one of
which can be evaluated though cloud analysis of pairs the international codes, viz., Indian code, is chosen for
of values of the median inter-storey drift demand (EDP) that purpose.
and the corresponding peak ground acceleration (IM) In order to study the effect of MF values on the
obtained from N-LTH. As per 2000 SAC FEMA (Cornell probability of failure of OGS building, different MF
et al., 2002), the estimate of the median demand, EDP values suggested by selected international codes are
(SD), can be represented in a generalized form by a power considered. Fully infilled frames (FF) and bare frames
model as given in Eq. (7). (BF) are also considered in the study for comparison
which are designed without considering any MF (i.e.
MF = 1.0). The international codes considered in this
EDP  a  IM 
b
(7) study are EC 8 (2003), Indian code (IS 1893, 2002),
Israel code (SI 413, 1995) and Bulgarian code (Bulgarian
where, a and b are the regression coefficients of the seismic code, 1987). Figure 3 shows the building frame
PSDM. The parameters ‘a’ and ‘b’ can be calculated by configurations such as bare frame (BF), fully infilled
the cloud analysis as discussed. frame (FF), OGS frame with MF equal to unity (OGS)
2.2.2 Performance limit states and OGS frames designed with MF suggested by different
Performance limit states define the capacity of codes. The OGS frames are designated according to the
the structure to withstand different levels of damage. name of the code chosen for MF. The design details the
The median inter-storey drift (ISD) limit states for columns of the two, four and six storey frames are listed
RC moment resisting frame structures that define the in Table 3(a), 3(b) and 3(c), respectively.
capacity of the structure at various performance levels
(SC) are suggested by Ghobarah (2000) and ASCE/SEI
41-06 (2007). Transient drift limits for RC frames as per 4 Nonlinear modelling approach
ASCE/SEI 41-06 (2007) are considered in the present
study as 1% and 4% for light repairable damage (IO) A series of nonlinear time history analyses (N-LTH)
and near collapse (CP) performance level, respectively. of RC frames have been performed. The Open System
for Earthquake Engineering Simulation (OpenSEES)
Laboratory tool developed by McKenna et al. (2014)
3 Frames considered is used for all analyses. Force-based nonlinear beam-
column element that consider the spread of plasticity
The building frames considered in the present study along the element is used for modelling the beams
are designed for the highest seismic zone of India (Zone and columns. Formulation of the fiber-based element
V with PGA of 0.36 g) as per Indian standards IS 1893 is explained in Lee and Mosalam (2004). Kunnath
(2002) considering medium soil conditions (N-value (2007) has studied the sensitivity due to the number
in the range 10 to 30). The characteristic strength of of integration points in each element and suggested
concrete and steel are taken as 25 MPa and 415 MPa, the use of five integration points, a practice followed
respectively. The buildings are assumed to be symmetric in the present study. The material models for concrete,
in plan, and hence a single plane frame is considered reinforcement steel and the infill masonry wall used in
to be representative of the building along one direction. the study are discussed in the following sections.
Typical bay width and column height in this study
are 5 m and 3.2 m, respectively, as observed from the 4.1 Concrete and steel constitutive models
study of typical existing residential buildings. A frame
configuration of two bays with two, four and six storeys The concrete is modelled by considering the effect
are considered. The dead load of the slab (5 m × 5 m of confinement due to the special confining detailing
panel) including floor finishes is taken as 3.75 kN/m2 in the beams and columns using the Kent-Scott-Park
and live load as 3 kN/m2. The design base shear (VB) constitutive model (Kent and Park, 1971), and using
is calculated as per equivalent static method (IS 1893, Concrete01 uniaxial material model in the OpenSEES
2002). The structural analysis for all the vertical and library. The cover concrete is modelled as unconfined
lateral loads is carried out by ignoring the infill wall concrete. Young’s modulus (Ec) for both core and cover
strength and stiffness (conventional). The design of concrete is a function of compressive strength (viz.,
No.2 Haran Pragalath D.C et al.: Multiplication factor for open ground storey buildings – a reliability based evaluation 287

2-6 stories @ 3.2 m 2 bays @ 5 m 2 bays @ 5 m 2 bays @ 5 m

Beams

MF = 1.0
Columns

(a) BF (Bare frame) (b) FF (Fully infilled frame) (c) OGS (OGS frame)

2 bays @ 5 m 2 bays @ 5 m 2 bays @ 5 m 2 bays @ 5 m

2-6 stories @ 3.2 m

MF = 4.68
MF = 3.0
MF = 2.5

MF = 3.0
(d) IS (Indian code) (e) BLG (Bulgarian code) (f) SI (Israel code) (g) EC-8 (Eurcode)
Fig. 3 Configuration of the frames considered in the present study

Table 3(a) Column details for two storey frames

Reinforcement details
Frame Storey No. Width (mm) Depth (mm) Lateral ties
(uniformly distributed)

BF G&I 300 300 8 - ϕ20

FF G&I 300 300 8 - ϕ20 rectangular ties ϕ8 @ 175 mm c/c
throughout
OGS G&I 300 300 8 - ϕ20
IS G 425 425 8 - ϕ22
I 300 300 8 - ϕ20
rectangular ties ϕ16@ 85 mm c/c
BLG G 450 450 8 - ϕ25 for a distance of 535 mm from
I 300 300 8 - ϕ20 supports
SI G&I 450 450 8 - ϕ25
EC G 550 550 8 - ϕ25
I 300 300 8 - ϕ20
Note :ϕ- is the diameter of rebar in mm, G represents ground storey

on infill wall modelling and recommended a simple

5000 f ck ) as per IS 456 (2000).
Steel reinforcing bars are modeled using the uniaxial diagonal strut model which is numerically stable and
Giuffre-Menegotto-Pinto (Menegotto and Pinto, 1973) computationally efficient. The most recent infill wall
steel material model with an isotropic strain hardening model used by Celarec et al. (2012) is used in the present
ratio of 0.01 (Steel02 material). More details about study for its computational simplicity. This model is
reinforcement modelling used in the present study can originally proposed by Panagiotakos and Fardis (1996)
be found in Filippou et al. (1983). and Fardis (1996). The typical quadrilinear force-
displacement relationship of the diagonal struts (in
4.2 Infill wall modelling compression), measured in the axial direction is shown
in Fig. 4. In the present study, infill walls are modelled as
Many studies (Landi et al., 2012; Mosalam et al., equivalent diagonal single strut in both diagonals of each
1997; Al-chaar, 2002; etc.) have reported on the work bay using truss element available in OpenSEES. In order
288 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol.15

Table 3(b) Column details for four storey frames

Reinforcement details
Frame Storey No. Width (mm) Depth (mm) Lateral ties
(uniformly distributed)

BF G 350 350 8 - ϕ20

I, II & III 350 350 8 – ϕ20 rectangular ties ϕ8 @ 170 mm c/c
throughout
FF G 350 350 8 - ϕ20
I, II & III 350 350 8 – ϕ20
OGS1.0,1.0 G 350 350 8 - ϕ20
I, II & III 350 350 8 – ϕ20
IS G 475 475 12 - ϕ25 rectangular ties ϕ16 @ 85 mm c/c
for a distance of 540 mm from
I, II & III 350 350 8 – ϕ20
supports,
Bulgarian G 600 600 16 - ϕ25
I, II & III 350 350 8 – ϕ20
SI G&I 600 600 16 - ϕ25
II & III 350 350 8 - ϕ18
EC G 750 750 20 - ϕ32
I, II & III 350 350 8 - ϕ 18
Note :ϕ- is the diameter of rebar in mm, G represents ground storey

Table 3(c) Column details for six storey frames

Reinforcement details
Frame Storey No. Width (mm) Depth (mm) Lateral ties
(uniformly distributed)

BF G 450 450 8 - ϕ20

I to V 450 450 8 – ϕ20 rectangular ties ϕ8 @ 175 mm c/c
throughout
FF G 450 450 8 - ϕ20
I to V 450 450 8 – ϕ20
OGS G 450 450 8 - ϕ20
I to V 450 450 8 – ϕ20
IS G&I 650 650 12- ϕ32 rectangular ties ϕ16 @ 85 mm
II to V 450 450 8 – ϕ20 c/c for a distance of 535 mm from
supports,
BLG G 750 750 12 - ϕ32
I to V 450 450 8 – ϕ20
SI G&I 750 750 12 - ϕ32
II to V 450 450 8 – ϕ20
EC Ground 900 900 20 - ϕ32
I to V 450 450 8 – ϕ20
Note : Note :ϕ- is the diameter of rebar in mm, G represents ground storey

to model the strength and stiffness degardation of the

Fmax ‒ Maximum strength
infill walls in time history analysis, a pinching material Fmax Fcr ‒ Shear cracking strength
model is used which is implemented in OpenSEES by Fres ‒ Residual strength
Ibarra et al. (2005). The same model is used by many Fcr Ksec ‒ Secant stiffness
Force

other studies (Landi et al., 2012; Ravichandran and Kel ‒ Elastic stiffness
Klinger, 2012) on frames with infilled masonry. Ksec
Kel
5 Validation study Fres
Displacement
Validation of the modelling approach used in the Fig. 4 Force-displacement relationship of the diagonal struts
No.2 Haran Pragalath D.C et al.: Multiplication factor for open ground storey buildings – a reliability based evaluation 289

present study is carried out by conducting experiments 6 Selection of earthquake ground motion
on two reinforced concrete building frames (a bare
frame and an infilled frame) taken from the literature. Uncertainty in the seismic load is considered in the
The pseudo-dynamic experimental test conducted at study by the use of a suite of earthquake ground motions.
European Laboratory for Structural Assessment (ELSA Earthquake ground motions in a suite should be obtained
test laboratory) on four storey, three bay, RC full scale from past records of the region of interest. In this study,
frames (ICON frame – bare and infilled) are considered. far-field earthquakes are chosen since the building site is
The building frames are tested under two subsequent located at least 10 km away from faults (Ravichandran
unidirectional pseudo-dynamic loadings, first using and Klinger, 2012). Although India has experienced
the Acc-475 input motion and then the Acc-975 input several major earthquakes in the last few decades, the
motion. A detailed description of the test specimens, number of available earthquake records in Indian region
material properties and the loading schemes is given in is limited. In order to have a statistically sufficient
the SeismoStruct verification report. The test specimens number of ground motions a suite of synthetic ground
are modelled in OpenSEES as per the approach explained motions with sufficiently large numbers of samples
in Section 4. The nonlinear pseudo-dynamic time history may be required for the analyses. Therefore, a suite of
analysis for the record, Acc-475, is conducted and the top synthetic ground motions are generated in the present
displacement time histories are recorded. Figures 5 and study for nonlinear time history analysis of building
6 show the comparisons between the roof displacement models. The 22 pairs of far-field natural ground motion
histories obtained by the present computational study records given in FEMA P695 (2012) are modified to
and from the experimental study, for the bare and infilled match the design spectrum of Indian Standard IS 1893
frames, respectively. From these results it can be inferred (2002).
that the modelling approach adopted in the present study
can yield a reasonably accurate behavior of a bare frame
as well as infilled building frames for earthquake. This 7 Material uncertainty
modelling approach is found to be performing well for
the nonlinear dynamic analysis. A detailed validation Material properties for concrete, steel and infill wall
study of the same modelling approach is also discussed in structures are random in nature. Damping is another
in Haran Pragalath et al. (2015). source of uncertainty that can affect the dynamic response
of building significantly (Davenport and Carroll, 1986).
It is important to incorporate the uncertainties in all
possible material and modelling parameters in the
80
computational model to have a realistic representation
of the responses in probabilistic assessment.
Top displacement (mm)

60
40 The strength of concrete, reinforcement steel, infill
20 masonry and damping ratio are considered as random
0 variable as detailed in Table 4. Strength parameters for
-20
-40
concrete and steel are taken from Ranganathan (1999)
Experiment
-60 Present study
and shear strength of masonry infill is calculated from
-80 Agarwal and Thakkar (2001). Mean value for damping
0 2 4 6 8 10 12 14
Time (s) ratio for RC structures is assumed to be 5% as per IS-
Fig. 5 Comparison of roof displacement time histories for 1893 (2002) and coefficient of variation is taken from
ICON frame - Bare Davenport and Carroll (1986).
The techniques of random sampling are more
powerful and useful for performing probabilistic
analyses. However, in most cases the problems being
20 analyzed are extremely complex and the computation
time may be very long. As a result, the time needed to
Top displacement (mm)

10 perform hundreds or thousands of simulation may be

prohibitive. To overcome this problem McKay et al. (1979)
0 proposed an attractive alternative method in computer
experiments called as Latin Hypercube Sampling
-10 (LHS). Several authors (Ayyub and Lai, 1989; Iman and
Conover, 1980 and Tavares et al., 2012) have used LHS
Experiment
-20 Present study method successfully to consider uncertainties in materials
0 2 4 6 8 10 12 14
for developing fragility curves. This sampling scheme is
Time (s) used in the present study to incorporate uncertainties. 44
Fig. 6 Comparison of roof displacement history for ICON sets of random values of the parameters are generated
infilled frame as per Latin Hypercube sampling techniques. 44
 290 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol.15

Table 4 Details of random variables used

Random variables Mean COV (%) Distribution Remarks
Concrete compressive strength, fck 30.28 MPa 21.0 Normal Uncorrelated
Steel yield strength, fy 468.90 MPa 10.0 Normal Uncorrelated

Global damping ratio, ξ 5% 40.0 Normal Uncorrelated

Masonry shear strength, τc 0.20 MPa 12.0 Normal Uncorrelated

computational models are developed using these random elements are not considered in the numerical model. The
parameters for each of the selected building frames. beam column joint is assumed to be rigid.
The maximum ISD demands (EDP) for each storey
are monitored during N-LTH. The ISDs (maximum of
8 Development of PSDM all storeys) along with the corresponding PGAs (IM)
are plotted in a log-log graph. Each point in the plot
N-LTH analyses are carried out for all the 44 shown in Figs. 7‒9 represents a PGA value and the
computational models subjected to 22 pairs of selected corresponding maximum ISD for a particular frame. A
ground motions in order to obtain the median ISD demand power law (refer Eq. (7)) relationship for each frames
(EDP) for each frame. 22 pairs of ground motions are is fitted to the data using regression analysis, which
scaled linearly from 0.1 g to 1 g and each computational represents the PSDM model for the corresponding
model is analysed for a particular earthquake (randomly frame. The regression coefficients ‘a & b’, are found
selected) with a particular PGA. P–Δ effect is included for each frame and reported in Figs. 7-9. An insight of
in the N-LTH. Rayleigh damping is adopted in the the relative vulnerability of the frames can be obtained
numerical model and the damping matrices are estimated from the values of parameters ‘a & b’. It can be noted
from the 1st and 2nd natural frequencies. The Newmark that higher values of ‘a’ correspond to higher values of
direct integration method with the coefficients, α = 0.5, ISD and thereby higher vulnerability of the frame. With
β = 0.25 is used for N-LTH. The effects due to shear regard to two storey frames, the order of vulnerability
behavior, sliding effect and buckling of steel bars for RC increases perhaps in the order, FF (a = 1.10), EC (a =

Table 5 Relative vulnerability of the frames based on parameter ‘a’

Two storey frames Four storey frames Six storey frames
Frame ID value of a Frame ID value of a Frame ID value of a
FF 1.10 SI 2.90 SI 2.47
EC 1.29 FF 4.61 FF 3.70
SI 2.39 EC 5.10 BLG 3.73
BLG 2.72 BF 5.21 BF 4.64
IS 3.21 BLG 5.60 EC 4.91
OGS 6.23 IS 5.87 IS 5.52
BF 6.28 OGS 7.97 OGS 6.20

100
10
Cloud analysis

Cloud analysis
Constants

Constants

10
PSDM

PSDM
Best fit

Best fit
EDP (ISD, %)

EDP (ISD, %)

1
a b a b
1 BF 6.28 1.25 BF 5.21 1.21
FF 1.10 1.99 FF 4.61 2.19
OGS 6.23 1.21 OGS 7.97 1.22
IS 3.21 1.30 0.1 IS 5.87 2.03
BLG 2.72 1.24 BLG 5.60 2.05
0.1 SI 2.39 1.25 SI 2.90 2.11
EC 1.29 1.03 EC 5.10 2.18

0.01
0.01 0.1 1.0
0.1 1.0
IM (PGA, g)
PGA (g)
Fig. 7 PSDM model for two storeyed frames Fig. 8 PSDM Model for four storeyed frames
 No.2 Haran Pragalath D.C et al.: Multiplication factor for open ground storey buildings – a reliability based evaluation 291

10 of OGS is higher than that of FF. Due to the presence of

Cloud analysis
infill walls in the ground storey, the fully infilled frames

Constants
exhibit better performance. The maximum ISD at each

PSDM
Best fit
Interstorey drift, %

1 floor level for these frames for a single earthquake (at

a b a particular instant) is shown in Fig. 10(b). It is noted
BF 4.64 1.12 that the ISD at the ground storey of an OGS frame is
0.1
FF
OGS
3.70
6.20
1.67
1.21
significantly high compared to the other storeys. On
IS 5.52 1.73 the other hand, the maximum storey displacements for
BLG 3.73 1.50
SI 2.47 1.49 the FF frame is considerably less than that for the OGS
EC 4.91 1.69
frame. Higher values of ISD demands at the open ground
0.01
0.1 1.0 story are also reported by Favvata et al. (2013) in a
PGA (g) recent study using capacity assessment procedure. Infill
Fig. 9 PSDM Model for six storeyed frames walls present in the ground storey (FF) can enhance the
building performance many fold. In the present case,
infill walls present in the ground storey of the FF frame
1.29), SI (a = 2.39), BLG (a = 2.72), IS (a = 3.21), OGS reduce the exceedance probability of ISD by about 62%
(a = 6.23) and BF (a = 6.28). Table 5 gives a summary of at a PGA of 0.7g as compared with OGS frame (refer
the orders of relative vulnerabilities of the two, four and Fig. 10(a)). Performance comparison between FF and
six storey frames considered. OGS has been studied by various researchers (Davis et
al., 2010; Scarlet, 1997 and Dymiotis et al., 2001) and
they reported similar observations.
9 Fragility curves
9.2 Comparison of frames design with various MF
Fragility curves are developed for different
performance limit states as per the method discussed In order to study the effect of MF values on the
in Section 2.2 using the computed PSDM models (a performance of two, four and six storeyed frames,
& b), the dispersions of ISDs (βD/IM) and the dispersion the fragility curves for different performance levels
in capacity (βc). The performances of OGS buildings are compared as shown in Figs. 11-13. These figures
designed with MFs suggested by different codes are show that as the MF value applied in the ground storey
compared based on the corresponding fragility curves as increases the performance of the frames improves in
discussed in following sections. general and failure due to storey mechanism moves
from the open ground storey to the adjacent first storey.
9.1 Comparison of fragility curves of FF and OGS However, the fragility curves of SI frames (designed
using Israel code) show that when the MF is applied
In order to understand the behavior of an OGS frame in to both ground and first storeys (refer the four and six
comparison with a fully infilled frame the corresponding storeyed SI frames) the exceedance probability improves
fragility curves are compared. This is studied through significantly compared to other frames.
the four storeyed frame as shown in Fig. 10(a). It can
be seen from this figure that the exceedance probability

1.0

4 FF
Probability of being in/exceedance

0.8

OGS
3
Floor level

0.6

0.4
2
FF

0.2 OGS
1

0
0 0.5 1.0 1.5 2.0 2.5 0 1 2
PGA (g) ISD (%)
(a) Fragility curve (IO level) (b) Storey wise ISD profile (for a particular earthquake at an instant)
Fig. 10 Fragility curve and typical drift profile for four storey frame for BF, FF and OGS
292 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol.15

1.0 1.0
Probability of being in/exceedance

Probability of being in/exceedance

0.8 0.8

0.6 0.6
BF BF
FF FF
0.4 0.4
OGS OGS
IS IS
BLG BLG
0.2 0.2
SI SI
EC EC
0 0
0 0.5 1.0 0 1 2 3
PGA (g) PGA (g)
(a) at IO performance level (b) at CP performance level
Fig. 11 Fragility curves for two storey frames

1.0
1.0
Probability of being in/exceedance

Probability of being in/exceedance

0.8 0.8

0.6 0.6
BF BF
0.4 FF 0.4 FF
OGS OGS
IS IS
0.2 BLG 0.2 BLG
SI SI
EC EC
0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
PGA (g) PGA (g)
(a) at IO performance level (b) at CP performance level
Fig. 12 Fragility curves for four storey frames

1.0 1.0
Probability of being in/exceedance

Probability of being in/exceedance

0.8 0.8

0.6 0.6

BF BF
FF FF
0.4 0.4
OGS OGS
IS IS
BLG 0.2 BLG
0.2
SI SI
EC EC
0 0
0 0.5 1.0 0 1 2 3
PGA (g) PGA (g)
(a) at IO performance level (b) at CP performance level
Fig. 13 Fragility curves for six storey frames

10 Comparison of reliability and cost hazard curve using Eqs. (2) and (3). In the present
study, the hazard curve of North East India as shown
The reliability indices of the frames are calculated in Fig. 2 is chosen. Reliability index is calculated
for the purpose of quantifying their relative performance. for two performance objectives (PO-I and PO-II)
The reliability indices are estimated by combining such as Immediate Occupancy (IO) at Design Basis
the fragility curve for a particular limit state with the Earthquake (DBE) level which corresponds to a mean
No.2 Haran Pragalath D.C et al.: Multiplication factor for open ground storey buildings – a reliability based evaluation 293

annual probability of occurrence of 10% in 50 years always befitting their corresponding reliability indices.
and Collapse Prevention (CP) at Maximum Considered In the case of two storey frames, the EC frame
Earthquake (MCE) level which corresponds to a mean achieves greater reliability than the FF but its cost index
annual probability of occurrence of 2% in 50 years. is 1.4 times that of the latter. For four and six storey
Reliability indices are computed for each frame and frames, it can be seen than cost index for EC is almost
listed in Table 6 for PO-I and PO-II in decreasing order. the highest, but its reliability is lesser than that of FF. On
For two storey frames, the reliability index decreases the other hand the SI frame achieves a higher reliability
in the following order for the selected performance index than FF but with a cost index similar to that of EC.
objectives: EC, FF, SI, BLG, IS, BF, and OGS. As there This indicates that MF in both ground and first storeys
are only two storeys, higher MF in the ground storey (SI) is a better scheme for typical four and six storey
leads to better performance. Behavior of four and six open ground-storey frames for both reliability and cost.
storey frames are slightly different as can be observed
from their reliability indices. 11 Conclusions
The reliability index of four storey frames decreases
in the following order for the selected performance There is disparity in the value of MF and its scheme
objectives: SI, FF, EC, BLG, IS, BF and OGS. For six of application for design of open ground storey (OGS)
storey frames the order is as follows: SI, FF, BLG, EC, buildings proposed by various International codes.
IS, BF and OGS. The present effort attempts to study, in a probabilistic
The reliability index of OGS frames (two, four framework, the performance of typical two, four and
and six storeyed) at both PO-I and PO-II performance six storey OGS frames designed with MFs suggested by
objectives are invariably the lowest indicating extreme various international codes. Open ground storey frames
vulnerability. Frames BF and FF perform better than are “designed” with MF values suggested by IS 1893
OGS frames according to their reliability indices. (2002), Bulgarian Seismic Design Code (1987), SI
For four and six storey frames, IS, BLG and EC 413 (1995) and EC 8 (2003). The probabilistic seismic
have a reliability index that is lower than that of the demand models, fragility curves, reliability indices
corresponding FF due to large drifts occurring at the first (for a selected seismic hazard) and cost indices for all
storey. The reliability index obtained for frame SI is the the selected open ground storey frames are developed
highest for both four and six storey frames, which shows including bare frame (BF) and fully infilled frame (FF).
that MF on both ground and first storey (SI) provide The following major conclusions are drawn on the
better reliability (even more than that of the FF frame). basis of the present study:
It can be concluded from these results that MF in both  The open ground storey building designed with
ground and first storeys is perhaps a better scheme than multiplication factor of 1.0 (OGS) is found to be more
MF applied only in ground storey for typical four and six vulnerable than Bare Frame (BF) and Fully Infilled
storey frames. Frame (FF). Performance of FF is found to be superior
A higher value of MF will attract a higher cost of due to the presence of infill walls in all the storeys
construction. Cost associated with the resulting sections including the ground storey.
due to MF as per the various codes is computed and  The scheme of applying MF only to the ground
expressed in terms of a parameter ‘cost index’. The cost storey proposed by Indian, Bulgarian and Euro codes is
index is defined as the ratio of the cost of construction found to lead to satisfactory performance only for two
for the selected frame to that of a similar bare frame storey frames. This scheme is found to be not effective
(BF). Cost indices are computed for each frame and for four and six storey frames as these frames cannot
listed in the Table 6. This table shows that the cost of match the reliability of a corresponding fully infilled
construction associated with different MF values are not frame.

Table 6 Reliability and cost indices for the building frames

2-storey Frame 4-storey Frame 6-storey Frame
Reliability index Reliability index Reliability index
Frame ID Cost index Frame ID Cost index Frame ID Cost index
PO-I PO-II PO-I PO-II PO-I PO-II
EC 1.4 4.32 4.90 SI 2.3 2.75 3.14 SI 1.9 2.55 3.26
FF 1.0 3.35 3.72 FF 1.0 2.44 2.82 FF 1.0 2.30 2.88
SI 1.6 2.45 3.55 EC 2.5 2.38 2.78 BLG 1.4 2.18 2.85
BLG 1.3 2.32 3.43 BLG 1.6 2.27 2.72 EC 1.8 2.12 2.68
IS 1.1 2.21 3.23 IS 1.4 2.24 2.69 IS 1.0 2.08 2.68
BF 1.0 1.63 2.45 BF 1.0 1.72 2.60 BF 1.3 1.71 2.63
OGS 1.0 1.61 2.45 OGS 1.0 1.43 2.25 OGS 1.0 1.60 2.45
294 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol.15

 For four and six storey frames, the scheme of Quake Resistance — Part 1: General Rules, Seismic
applying MF to both the open ground storey and the Actions and Rules for Buildings, European Committee
adjacent first storey, as suggested by SI 413 (1995), is of Standardization, prEN1998-1, Brussels, Belgium.
found to be a better solution for both the reliability and Fardis MN (1996), “Experimental and Numerical
cost aspects. Investigations on the Seismic Response of RC Frames and
Recommendations for Code Provisions,” Laboratório
References Nacional de Engenharia Civil, Lisboa, Port.
Fardis MN, Negro P, Bousias SN and Colombo A (1999),
Agarwal P and Thakkar SK (2001), “A Comparative “Seismic Design of Open-storey Infilled RC Buildings,”
Study of Brick Masonry House Model under Quasi- Journal of Earthquake Engineering, 3(2): 173‒197.
Static and Dynamic Loading,” Journal of Earthquake Fardis MN and Panagiotakos TB (1997), “Seismic
Technology, 38(2‒4): 103‒122. Design and Response of Bare and Masonry-Infilled
Al-Chaar G (2002), “Evaluating Strength and Stiffness Reinforced Concrete Buildings, Part II: Infilled
of Unreinforced Masonry Infill Structures,” US Army Structures,” J. Earthquake Eng., 1(3): 475–503.
Corp of Engineers, Engineer Research and Development Favvata MJ, Naoum MC amd Karayannis CG (2013),
Centre. “Limit States of RC Structures with First Floor
ASCE/SEI 7 (2010), Minimum Design Loads for Irregularities,” Journal Structural Engineering and
Buildings and Other Structures, American Society of Mechanics, 47(6): 791‒818.
Civil Engineers, Reston, VA. FCEACR (1986), Seismic Code of Costa Rica, Federal
ASCE/SEI 41‒06 (2007), Seismic Rehabilitation of College of Engineers and Architects of Costa Rica, San
Existing Buildings, American Society of Civil Engineers, Jose, Costa Rica.
USA. FEMA P695 (2012), Quantification of Building Seismic
ATC 58 100% Draft (2012), Guidelines for Seismic Performance Factors, Applied Technology Council,
Performance Assessment of Buildings, Applied Redwood City: California for the Federal Emergency
Technology Council, Redwood City, CA. Management Agency, Washington, D.C.
Ayyub BM and Lai KL (1989), “Structural Reliability Filippou FC, Popov EP and Bertero VV (1983),
Assessment Using Latin Hypercube Sampling,” 5th “Effects of Bond Deterioration on Hysteretic Behavior
International Conference on Structural Safety and of Reinforced Concrete Joints,” Report EERC 83-19,
Reliability, ICOSSAR, San Francisco, California, Earthquake Engineering Research Center, University of
pp.1177‒1184. California, Berkeley.
Bulgarian Seismic Code (1987), Code for Design of Ghobarah A (2000), “Performance-based Design in
Buildings and Structures in Seismic Regions, Bulgarian Earthquake Engineering: State of Development,”
Academy of Science Committee of Territorial and Town Engineering Structures, 23: 878‒884.
System at the Council of Ministers, Sofia, Bulgaria. Haran Pragalath DC, Davis R and Sarkar P (2015),
Celarec D, Ricci P and Dolsek M (2012), “The Sensitivity “Reliability Evaluation of RC Frame by Two Major
of Seismic Response Parameters to the Uncertain Fragility Analysis Methods,” Asian Journal of Civil
Modeling Variables of Masonry-infilled Reinforced Engineering (Building and Housing), 16(1): 47‒66.
Concrete Frames,” Engineering Structures, 35:165–177. Hashmi AK and Madan A (2008), “Damage Forecast for
Cornell CA, Jalayer F, Hamburger RO, Foutch DA Masonry Infilled Reinforced Concrete Framed Buildings
(2002), “The Probabilistic Basis for the 2000 SAC/ Subjected to Earthquakes in India,” Current Science, 94:
FEMA Steel Moment Frame Guidelines,” Journal of 61‒73.
Structural Engineering, 128(4): 526‒533. Ibarra LF, Medina A and Krawinkler H (2005),
Davenport AG and Carroll H (1986), “Damping in Tall “Hysteretic Models that Incorporate Strength and
Buildings: Its Variability and Treatment in Design,” Stiffness Deterioration,” Earthquake Engineering and
ASCE Spring Convention, Seattle, USA, pp. 42‒57. Structural Dynamics, 34: 1489‒1511.
Davis PR, Padhy KT, Menon D and Prasad AM (2010), ICC IBC (2012), International Building Code,
“Seismic Fragility of Open Ground Storey Buildings in International Code Consortium, USA.
India,” 9th US National and 10th Canadian Conference Iman RL and Conover WJ (1980), “Small Sample
on Earthquake Engineering, Toronto, Paper no. 926. Sensitivity Analysis Techniques for Computer
Dymiotis C, Kappos JA, Chryssanthopoulos KM (2001), Models with an Application to Risk Assessment,”
“Seismic Reliability of Masonry-Infilled RC Frames,” Communications in Statistics – Theory and Methods,
Journal of Struct. Eng., 127: 296‒305. 9(17): 1749‒1874.
Ellingwood BR (2001), “Earthquake Risk Assessment IS 456 (2000), Indian Standard for Plain and Reinforced
of Building Structures,” Reliability Engineering and Concrete Code of Practice, Bureau of Indian Standards,
System Safety, 74: 251‒262. New Delhi.
Eurocode 8 (2003), Design of Structures for Earth- IS 1893 Part I (2002), Indian Standard Criteria for
No.2 Haran Pragalath D.C et al.: Multiplication factor for open ground storey buildings – a reliability based evaluation 295

Earthquake Resistant Design of Structures, Bureau of Kathmandu, Nepal.

Indian Standards, New Delhi. Nielson BG (2005), “Analytical Fragility Curves for
IS 13920 (1993), Ductile Detailing of Reinforced Highway Bridges in Moderate Seismic Zones,” Ph.D.
Concrete Structures Subjected to Seismic Forces - Code Thesis, Georgia Institute of Technology.
of Practice, Bureau of Indian Standards, New Delhi. NZS 1170.5 (2004), Structural Design Actions, Part
Kaushik HB, Rai DC, Jain SK (2006), “Code Approaches 5: Earthquake Actions -New Zealand-Commentary,
to Seismic Design of Masonry Infilled Reinforced Standards Association of New Zealand, Wellington,
Concrete Frames: A State-of-the-Art Review,” New Zealand.
Earthquake Spectra, 22(4): 961‒983. Pallav K, Raghukanth STG, Singh KD (2012),
Kent D and Park R (1971), “Flexural Mechanics with “Probabilistic Seismic Hazard Estimation of Manipur,
Confined Concrete,” J Struct Div, ASCE, 97(ST7): India,” Journal of Geophysics and Engineering, 9: 516–533.
1969‒90. Panagiotakos TB and Fardis MN (1996), “Seismic
Kunnath SK (2007), “Application of the PEER PBEE Response of Infilled RC Frame Structures,” Proceedings
Methodology to the I-880 Viaduct,” PEER 2006/10, of the 11th world Conference on Earthquake Engineering,
Pacific Earthquake Engineering Research Center, Acapulco, P-4 (225).
University of California, Berkeley. Raghunandan M, Liel AB, Ryu H, Luco N and Uma
Landi L, Diotallevi PP and Tardini A (2012), “Calibration SR (2012), “Aftershock Fragility Curves and Tagging
of an Equivalent Strut Model for the Nonlinear Seismic Assessment for a Mainshock-damaged Building, 15th
Analysis of Infilled RC Frames,” 15th World Conference World Conference on Earthquake Engineering, Lisbon,
on Earthquake Engineering, Lisboa, Portugal. Portugal.
Lee TH and Mosalam KM (2004), “Probabilistic Fiber Rajeev P and Tesfamariam S (2012), “Seismic Fragilities
Element Modeling of Reinforced Concrete Structures,” for Reinforced Concrete Buildings with Consideration
Computers and Structures, 82(27): 2285‒2299. of Irregularities, Structural Safety, 39:1‒3.
McKay M, Beckman R and Conover W (1979), “A Ranganathan R (1999), “Structural Reliability Analysis
Comparison of Three Methods for Selecting Values and Design,” Jaico Publishing House, Mumbai.
of Input Variables in the Analysis of Output from a Ravichandran SS and Klinger ER (2012), “Seismic
Computer Code,” Technometrics, 21: 239–245. Design Factors for Steel Moment Frames with Masonry
McKenna F, McGann C, Arduino P and Harmon JA Infills: Part 1,” Earthquake Spectra, 28(3): 1189‒1204.
(2014), “OpenSEES Laboratory,” <https://nees.org/ Sattar S and Liel AB (2010), “Seismic Performance
resources/openseeslab/>. of Reinforced Concrete Frame Structures with and
Menegotto M and Pinto PE (1973), “Method of Analysis without Masonry Infill Walls,” 9th U.S. National and
of Cyclically Loaded RC Plane Frames Including 10th Canadian Conference on Earthquake Engineering,
Changes in Geometry and Non-elastic Behavior Toronto, Canada.
of Elements under Normal Force and Bending,” Scarlet A (1997), “Design of Soft Stories-A Simplified
Preliminary Report, IABSE, 13: 15–22. Energy Approach,” Earthquake Spectra, 13: 305‒315.
Meslem A and D’Ayala D (2013), “Investigation into Seismosoft (2013), SeismoStruct v6.0 - Verification
Analytical Vulnerability Curves Derivation Aspects Report, Available from URL: www.seismosoft.com.
Considering Modelling Uncertainty for Infilled RC SI-413 (1995), Design Provisions for Earthquake
Buildings,” Proceeding of the 4th International Resistance of Structures, The Standards Institution of
Conference on Computational Methods in Structural Israel, Tel-Aviv, Israel.
Dynamics and Earthquake Engineering, Kos Island Greece.
Song J and Ellingwood BR (1999), “Seismic Reliability
Mosalam KM, Ayala G, White RN and Roth C of Special Moment Steel Frames with Welded
(1997), “Seismic Fragility of LRC Frames with and Connections: II,” Journal of Structural Engineering,
without Masonry Infill Walls,” Journal of Earthquake 125(4): 372.
Engineering, 1(4): 693‒720.
Tavares DH, Padgett JE, and Paultre P (2012), “Fragility
National Information Centre of Earthquake Engineering Curves of Typical as-Built Highway Bridges in Eastern
< http://www.nicee.org/> (March 30, 2015). Canada,” Engineering Structures, 40: 107‒118.
NBC-201 (1995), Nepal National Building Code for Wen YK, Ellingwood BR, Bracci J (2004),
Mandatory Rules of Thumb for Reinforced Concrete “Vulnerability Function Framework for Consequence-
Buildings with Masonry Infill, Ministry of Housing based Engineering,” Mid-America Earthquake (MAE)
and Physical Planning, Department of Buildings, Center Project DS-4 Report, Urbana, IL.