You are on page 1of 10

J. Inst. Eng. India Ser.

A (May–July 2013) 94(2):99–108


DOI 10.1007/s40030-013-0042-y

ORIGINAL CONTRIBUTION

Equivalent Strut Method for the Modelling of Masonry Infill


Walls in the Nonlinear Static Analysis of Buildings
A. Adukadukam • A. K. Sengupta

Received: 1 June 2012 / Accepted: 10 October 2013 / Published online: 30 October 2013
 The Institution of Engineers (India) 2013

Abstract In the seismic analysis of a framed building Keywords Equivalent strut  Infill wall 
with masonry infill walls, it is necessary to model the effect Masonry wall  Nonlinear static analysis 
of the walls on the lateral stiffness, strength and ductility of Pushover analysis  Seismic analysis
the building. The equivalent strut method is convenient for
modelling the walls in a large building. However, an List of symbols
appropriate axial load versus deformation relationship for B Breadth of section
the strut is required in a nonlinear static method of seismic D Depth of section
analysis, such as the pushover analysis. The present study d Length of the equivalent strut, equal to the
proposes a nonlinear axial hinge property for the strut, with diagonal length of the infilled panel
suitable performance levels. First, the equivalent strut fa Allowable compressive strength along the diagonal
method and the suitability of two approaches available in of a wall based on slenderness ratio
=
the literature for modelling the properties of the struts, are fbs Design bond shear strength between masonry units
briefly discussed. Next, the nonlinear axial load versus and mortar
deformation relationship is developed based on experi- h Height of centre-line of beam from top of footing
mental data compiled from the literature. The parabolic– h/ Clear height of the wall
plastic relationship is idealized as a tri-linear axial hinge l Centre-line to centre-line width of bay
property, so that it can be incorporated in commercial l/ Clear length of the wall
software for undertaking pushover analysis. Next, the use R Axial load in the equivalent strut
of the hinge property is demonstrated in the pushover Rc Load in the equivalent strut at local crushing of the
analyses of two framed reinforced concrete buildings. The corners of the infill wall
pushover curves based on the proposed hinge property Rs Load in the equivalent strut at shear cracking of the
shows improved modelling of the inelastic drifts of the infill wall
buildings. Although the modelling of a wall using a single Rdc Load in the equivalent strut at diagonal
strut has limitations, the proposed methodology is practical compression failure of the infill wall
for a pushover analysis of a building. Ru Ultimate strength of the equivalent strut
t Thickness of wall
ac Coefficient measuring the length of the wall in
contact with the column
d Axial deformation of the equivalent strut
A. Adukadukam
School of Mechanical and Building Sciences, Vellore Institute of e Axial strain in the equivalent strut
Technology, Vellore, India e0 Strain in the equivalent strut when the
strength is attained (R/Ru attains the
A. K. Sengupta (&)
value of 1) = 0.0025
Department of Civil Engineering, Indian Institute of Technology
Madras, Chennai 600036, India eu Strain in the equivalent strut till the strength is
e-mail: amlan@iitm.ac.in retained = 0.004

123
100 J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108

e 1, e 2 Strain in the equivalent strut at the diagonal member-level non-linear behaviours of the frame members
compression failure of a slender infill wall and the struts can be assembled to get the global non-linear
/ Diameter of bars response of a building under lateral forces, using the static
l Coefficient of internal friction between masonry pushover analysis or dynamic time history analysis. The
units and mortar results can further be used in a performance based analysis
h Angle of inclination of diagonal of the panel, with of a building, which quantifies damage in terms of per-
respect to horizontal formance levels, under a certain level of seismic hazard.
heff Angle of inclination of strut considering the length Since the present study investigated the modelling of the
of wall in contact with the column infill walls as equivalent struts in the pushover analysis of a
rc Average normal stress on wall in contact with building, the equivalent strut method is briefly described.
column
sb Average shear stress on wall in contact with beam
Equivalent Strut Method

Introduction Under the action of in-plane lateral load, an infilled frame


behaves like a solid cantilever with an initial bond at the
In a framed multi-storeyed building, the exterior enclosing frame–wall interface [11]. An infill wall which is not
walls and interior partition walls are considered to be infill physically connected to the bounding members by shear
walls if they are placed within the panels of the frames. keys or dowels, is classified as non-intergral. For such a
Traditionally, in a structural analysis of the building, the wall, after the bond is broken, there is separation between
infill walls are treated as a non-structural architectural the frame and the wall, at the diagonally opposite corners
component of the building. They are not explicitly mod- of the panel that are under tension. Compressive load is
elled in a computational analysis. For heavy masonry transferred to the wall near the other corners which are
walls, their weight is assigned on the supporting beams. under compression, through certain lengths of contact with
However, if there is no gap or negligible gap between the the bounding frame members. The wall then behaves like a
walls and the bounding frame members, the walls tend to strut under compression.
have a substantial influence on the structural behaviour of Holmes [6] proposed replacing the wall by an equivalent
the building during an earthquake. The modelling of infill pin-jointed strut along the diagonal under compression,
walls becomes imperative in the seismic analysis of a made up of the same material as that of the wall. The
framed building, for the following reasons. First, the walls thickness of the strut is same as that of the wall, but the
increase the lateral stiffness of the building, leading to width is equal to one-third of the length of the diagonal of
larger lateral loads. Second, irregular placement of the the infill. This is a minimalistic model, suitable for the
walls can generate plan or vertical irregularity. The shift in analysis of buildings with several infill walls in the storeys.
the centre of stiffness for a floor or presence of open A theoretical relationship between the width of the strut
ground storey, and the possibility of brittle shear failure of and a relative stiffness parameter of the frame—wall sys-
columns next to a partial height wall remain undetected if tem has been provided by the researchers [15]. The width
the effect of the walls is not modelled. Third, the of the strut was given in the form of charts for different
strengthening of an open ground storey by providing infill loads on the strut, and the aspect ratio of the infilled panel.
walls cannot be justified by analysis unless the effect of the An alternate equivalent strut model based on a plastic
walls is modelled. analysis was developed earlier [8]. An equivalent strut
Accurate numerical modelling of the nonlinear behav- method was proposed based on the ultimate load (UL) that
iour of masonry infilled frames under in-plane lateral loads, can be carried by the wall [12]. The aspect ratio of the
by the finite element method is difficult. The modelling infilled panel, interface stresses due to the frame–wall
involves consideration of the individual masonry units, interaction, and the flexural capacities of the bounding
mortar in the bed joints and the frame–wall interface, and beams and columns are accounted for in this method.
their appropriate constitutive models [10]. Developing such The failure of an infill wall in a frame under in-plane
a computational model is intensive and time consuming for lateral load can be categorized into the following three
the analysis of multi-storeyed buildings with multiple walls modes. The corresponding compressive loads in the
in each storey. Hence, researchers have investigated alter- equivalent strut are denoted in the brackets.
nate macro-scale modelling of the walls, based on the
concept of equivalent struts which simulates the composite • Local crushing (Rc) of the infill at one of the loaded
behaviour of an infilled frame. The modelling of the corners.

123
J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108 101

• Shear cracking (Rs) along the bed joints of the masonry analysis of a framed building with infill walls by the
units. equivalent static method (static analysis) or the response
• Diagonal compression failure (Rdc) due to instability of spectrum method (dynamic analysis), modelling of the
a slender infill wall. walls by the simpler EA approach is adequate. In these
The diagonal tensile cracking need not be considered as a analyses, the member forces are of primary interest. But in
failure mode, as higher load can be applied beyond tensile a pushover analysis (nonlinear static analysis) of a build-
cracking. ing, the UL approach is preferred. Here, the ultimate lateral
The expressions for the three failure loads are provided strength and the inelastic drift of the building under
as given below [12]: increasing lateral load are of primary interest. Since the UL
approach gives a lower strut width and hence lower stiff-
ð1  ac Þac thrc þ ab tlsb ness for a strut, it provides a more realistic estimate of the
Rc ¼ ð1Þ inelastic drift. However in a pushover analysis, in addition
cos h
= to the strut width and strength, the axial load versus
fbs tl= deformation behaviour of a strut is also required.
Rs ¼ ð2Þ
ð1  l tan heff Þ cos h The modelling of an infill wall by a single strut has
0:5htfa limitations. First, it is not suitable when there is a central
Rdc ¼ ð3Þ opening in the wall. An alternate is to use multi-strut
cos h
model. Of course for a large opening as is common with
The symbols used in the expressions are explained
windows, the strut effect of the wall is negligible, due to
under List of symbols. The lowest of Rc, Rs and Rdc is the
the drastic reduction in diagonal stiffness. Hence, the strut
strength (Ru) of the equivalent strut. The failure of a wall
need not be modelled. Second, the procedure cannot
primarily due to out-of-plane bending is not an effect of
explicitly model the yielding of the columns or beams at
the in-plane lateral load. The width of the strut is
the portions in contact with the wall [3]. It has been
calculated from Ru based on a design compressive
observed that a model consisting of three struts parallel to
strength for the masonry. A secant modulus of elasticity
the diagonal of the wall under compression, is effective in
is considered for the strut.
capturing the flexural behaviour of the bounding frame
members. However, the three struts model needs additional
effort in developing the computational model for a large
Investigation of Equivalent Strut Methods building. Hence, in the present study a single strut model
was adopted, that is expected to be appealing to the prac-
Essentially there are two approaches in modelling the ticing professionals to get the overall behaviour of a
properties of the equivalent strut for an infill wall. The building under a pushover analysis.
approach as proposed is the basis for a conventional linear
analysis of a frame under lateral load, and has been adopted
in this work [5, 15]. In this paper, this approach is referred Modelling of Axial Load against Deformation
to as the elastic analysis (EA) approach. The width of a Behaviour
strut is calculated based on the overall lateral stiffness of
the infilled frame. On the other hand, the approach as The axial load versus deformation behaviour of an equiv-
proposed [12] is based on the governing failure load. It is alent strut can be considered to be linear for convenience
applicable beyond the linear elastic range of the wall. In
this paper, this approach is referred to as the UL approach.
The procedure for calculating the width of a strut implicitly
considers the yielding of the bounding columns or beams Ru B, C
near the joints.
Axial load

A comparative study of both the approaches in deter-


mining the properties of the equivalent strut, and the
applicability of the two approaches in linear and nonlinear
analyses of a framed building, were reported by the D E
0.2Ru
researchers [14]. The EA approach is simpler in calcula-
tion. A higher strut width gives higher stiffness for a strut A
δu Axial deformation
and hence, higher lateral load (base shear) for a building.
Since the EA approach gives higher strut width, it is con- Fig. 1 Linear axial load against deformation curve for an equivalent
servative in estimating the base shear. In a linear seismic strut

123
102 J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108

Fig. 2 Typical specimen and l


the corresponding model
δ
Jack H

h R

(a) Test specimen (b) Computational model

Table 1 Properties of the frame members in the test specimens


Reference l, m h, m Section dimensions (B 9 D), mm 9 mm Reinforcement details (/), mm Concrete
cube strength,
Beam Column Beam Column MPa
Top Bottom

Achintya et al. [1] 1.45 1.15 135 9 250 135 9 250 2–20 2–20 4–20 15.00
Achyutha et al. [2] 1.38 0.75 100 9 125 100 9 125 2–8 2–8 6–8 58.60
Mehrabi et al. [9] 2.31 1.52 160 9 230 180 9 180 2–16 2–16 8–12 26.80

Table 2 Properties of infill


Reference t, mm l/ , m h/, m Compressive strength Bond shear
walls in the test specimens =
of masonry, MPa strength ðfbs Þ, MPa

Achintya et al. [1] 115 1.20 1.00 7.50 0.24


Achyutha et al. [2] 60 1.25 0.63 14.00 1.20
Mehrabi et al. [9] 100 2.13 1.42 15.00 0.48

(Fig. 1), under the presumption that the masonry units and lateral load versus drift tests of single-bay single-storey
mortar are quasi-brittle in nature, and the infill walls are infilled reinforced concrete (RC) frame specimens [1, 2, 9].
secondary members in a frame. However, the behaviour of For calculating the values of R, the test for each specimen
an infill wall under diagonal compression is nonlinear due was simulated by a computational model using the program
to the formation and development of multiple types of SAP 2000 NL [13] (Fig. 2). Each of the beam and columns
cracks. In the present study, emphasis was placed in was modelled by a frame element, with non-linear flexural
modelling the non-linear behaviour of an equivalent strut. hinge property near the joints, calculated based on the
The nonlinear behaviour can be incorporated in the com- section dimensions and the reinforcement (Table 1). The
putational model by assigning an axial ‘plastic hinge’ at the columns were considered to be fixed at the foundations. For
middle of the strut, with appropriate axial load versus a strut, the UL approach was used to calculate its width
deformation curve for the hinge. The curve is referred to as from the properties given in Table 2. The thickness of the
the hinge property. Although the approach is based on strut was kept same as that of the wall. At each step for the
lumped plasticity, the non-linear deformation is calculated applied load H, the value of R in the strut was noted from
over the length of the strut. In a performance based ana- the analysis of the model. The corresponding value of d
lysis, the performance levels such as immediate occupancy was calculated as the diagonal component of the recorded
(IO), life safety (LS) and collapse prevention (CP) need to drift.
be defined for the hinge property, to monitor the state of To get a common trend in the R versus d variations in a
damage in the strut. single plot for all the analysed specimens, the individual
In the present study, an axial load (R) versus deforma- values of R and d were normalised with respect to the cor-
tion (d) relationship was developed for the equivalent strut, responding calculated maximum load (Ru) and diagonal
based on the analyses of published results of monotonic length (d), respectively. The normalised axial load (R/Ru)

123
J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108 103

1.2 Original curve


R Linear approximation
1.0
Assumed curve beyond CP
0.8 Ru C CP
B LS D
R/Ru

0.6 Achintya et al., 1991 IO

0.4 Achyutha et al., 1994


Mehrabi et al., 1996
0.2 0.2Ru
Proposed curve A
0.0 0 0.5 u
0 0
0.000 0.001 0.002 ε0 0.003 0.004 0.005 (0.00125) (0.0025) (0.004)
ε
Fig. 4 Linear approximation of axial load against deformation curve
Fig. 3 Nonlinear axial load against deformation curve for an
equivalent strut
tangential slope of the parabolic curve reduces to half the
initial value. The LS performance level was selected at
versus deformation (d/d) data points are shown in Fig. 3. The e = e0 = 0.0025, the strain corresponding to the attaining
normalised deformation (d/d) is denoted as axial strain (e). of maximum load in the strut. The CP performance level
Using the above data points, a curve similar to that for was selected at e = eu = 0.004, beyond which the load
concrete under flexural compression was fitted. The curve carrying capacity of the strut is uncertain and is expected to
is parabolic till R/Ru attains unity at the strain e0 = 0.0026. drop. In the present study it was not attempted to trace the
For convenience, in the proposed equation the strain was post-ultimate descending branch of the axial load versus
taken as 0.0025, for normal strength masonry units. After deformation behaviour of the equivalent strut. A drop in
this R/Ru is equated to unity till eu = 0.004. strength was assumed at e = 0.004, representing the failure
The proposed equations for the axial load versus of the infill wall. A residual strength of 20 % of the peak
deformation curve (Fig. 3) are given as follows. For e \ e0, strength was assumed based on conventional practice, to
"     # avoid any numerical instability in the computation for
e e 2
R ¼ Ru 2  ð4aÞ pushover analysis due to zero post-peak stiffness value for
e0 e0 a member. The specific equations of the piece-wise linear
curve can be derived from the coordinates of the kinks in
For e0 \ e B eu,
the curve, as follows.
R ¼ Ru ð4bÞ
For e B 0.5e0,
The above equations are valid when the strength is  
3 e
governed by shear cracking or corner crushing failure modes. R ¼ Ru ð5aÞ
2 e0
In such a case Ru is the lower of Rc and Rs. These are the
modes of material failure. When the failure is by diagonal
compression, a sudden drop in the curve is expected to occur The slope of the curve represents the secant-stiffness of the
due to instability of the slender infill wall. This is a rare original curve up to the IO performance level.
failure mode which is limited to only slender walls. For 0.5e0 \ e B e0,
 
Ru e
R¼ 1þ ð5bÞ
2 e0
Simplified Relationships with Performance Levels

For a pushover analysis of a building, instead of using the The slope of the curve represents the secant-stiffness of the
nonlinear axial load versus deformation relationship, a original curve between the IO and LS performance levels.
simplified piece-wise linear relationship is proposed for the For e0 \ e B eu,
hinge property of the equivalent strut. This simplification is
R ¼ Ru ð5cÞ
necessary for the input of a hinge property in a computa-
tional model. For e [ eu,
Figure 4 shows the tri-linear (up to the ultimate state)
R ¼ 0:2Ru ð5dÞ
axial load versus deformation curve. The three perfor-
mance levels were decided in conjunction with the FEMA
356 recommendations. The IO performance level was Instead of a discrete jump in strength at the ultimate, a
selected at e = 0.5 e0 = 0.00125, the strain at which the gradual reduction can also be adopted. The above hinge

123
104 J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108

R
compression failure. All the performance levels IO, LS
and CP are defined at the same point corresponding to
C D
Ru B the failure load Rdc. The following equations represent
Diagonal compression failure the hinge property for the equivalent strut.
Rdc IO, LS, CP For e B e1,
 
e
R ¼ Rdc ð6aÞ
0.2Rdc A e1
0 1 0.5 0 0 u

(a) Rdc within AB For e [ e1,


R ¼ 0:2Rdc ð6bÞ
R
C D  
Rdc Rdc
B IO LS, CP Here, e1 ¼ 23 R u e0 .

Diagonal compression failure Case 2: 0.75Ru \ Rdc \ Ru


If Rdc is between 0.75Ru and Ru, there will be a drop in
0.2Rdc the load carrying capacity in the second linear region
A
(BC). The dotted line in Fig. 5b shows the drop due to
0 0.5 0 2 0 u
diagonal compression failure. The IO performance level
(b) Rdc within BC is defined at 0.5e0 as before. The LS and CP performance
levels are defined at the same point corresponding to the
Fig. 5 Axial hinge property for equivalent strut under diagonal failure load Rdc. The following equations represent the
compression failure, a Rdc within AB, b Rdc within BC
hinge property for the equivalent strut.
For e B 0.5e0,
property is applicable when the failure is governed by  
3 e
corner crushing or shear cracking. In the case of diagonal R ¼ Ru ð7aÞ
2 e0
compression failure, since Rdc is lower than Ru (lower of Rc
and Rs), a drop in the curve is modelled when the axial load
reaches Rdc, before attaining the strength governed by Rc or For 0.5e0 \ e B e2,
Rs . Ru  Rdc
R ¼ Rdc  ðe2  eÞ ð7bÞ
There can be two cases based on the value of Rdc e0  e2
compared to the load at IO performance level, which is
0.75 Ru
For e B e2,
Case 1: Rdc B 0.75Ru
R ¼ 0:2Rdc ð7cÞ
If Rdc is less than 0.75Ru, there will be a drop in the load
carrying capacity in the first linear region (AB). The  
dotted line in Fig. 5a shows the drop due to diagonal Here, e2 ¼ 2 RRdcu  1 e0 .

Fig. 6 Typical plan and 6m 6m 6m 6m


elevation of Building I

6m 3.6m

3.6m
6m

3.6m

6m 3.6m

(a) Floor plan (b) Frame elevation

123
J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108 105

6000

Base shear, kN
4000
Based on linear hinge property

Based on proposed hinge property


2000

0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
Roof displacement m

Fig. 8 Pushover curve along X-direction (Building I)


(a) Using linear hinge property
IO LS CP the constitutive relationships of the two materials. A bi-
linear shear hinge property was idealised based on the
initial shear stiffness and ultimate shear strength. For a
strut element, two alternatives of the axial hinge property
were used. First, a conventional linear hinge property was
used, as per the EA approach (Fig. 1). Here, all the three
performance levels were considered at the peak of the
curve. Next, the proposed (Eqs. 5–7) tri-linear hinge
property was used. The pushover curves using the two
types of hinge properties were compared to observe the
effect of modelling the hinge property of a strut on the
overall behaviour of the building.
In each pushover analysis of a computational model,
initially the gravity loads were applied in a force controlled
(b) Using proposed hinge property manner. Next, the lateral loads were applied in a dis-
placement controlled manner. The distribution of the lat-
Fig. 7 Formation of hinges in a typical frame of Building I eral force along the height of the building was calculated as
per a parabolic relationship [7]. A brief description of the
buildings and the corresponding results are presented.
Case Studies
Building I
Two multi-storeyed RC framed buildings were analysed to
study the effect of the proposed hinge property for an Figure 6 shows the typical floor plan and frame elevation
equivalent strut. One is a model regular building that was of the four storeyed building, using line diagram for each
designed for seismic loads (Building I). The other one is an frame member. The concrete slab is 120 mm thick at every
existing irregular building that was not designed to resist floor. Beams in all the floors are identical (230 mm
earthquake forces (Building II). For each building, a width 9 500 mm depth). The size of the corner columns is
computational model was developed using frame elements 400 9 350 mm2. All the other interior columns are
for the beams and columns, and strut elements for the infill 350 9 350 mm2. A raft foundation was assumed at the
walls. The diaphragm action of the slabs was modelled by base of the columns. All the frames are filled with masonry
assigning horizontal rigid diaphragm constraints. For a walls. The wall thicknesses are 230 and 120 mm for the
frame element, a tri-linear flexural hinge property near the exterior and interior walls, respectively. The compressive
joint was idealised from the moment versus curvature strength of concrete and the yield stress for the reinforcing
curve for the corresponding section of the member, and an steel were considered to be 20 and 415 MPa, respectively.
estimate of the length of plastic hinge for calculating The compressive strength of the masonry units (burnt clay
rotation. The moment versus curvature curve was devel- bricks) was taken as 3 MPa.
oped based on nonlinear mechanics of RC, satisfying the In the pushover analysis using linear hinge properties for
principles of equilibrium of forces at the section, compat- the struts, it was found that the formation of the hinges in
ibility of strains of concrete and the reinforcing bars, and the struts leads to the termination of the pushover analysis.

123
106 J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108

Fig. 9 Typical plan and


elevation of Building II

A A

3m

Stair case Lift core 25.86 3m

3m

3m

3m

19.56 m
(a) Floor plan (b) Sectional elevation along A-A

Figure 7a shows the formation of hinges in a typical inte- of each level, to accomodate the lift core and the staircase,
rior frame, using linear hinge properties for the struts. No two different diaphragm constraints were assigned at each
hinge has formed in the beams or columns. This result is level. The dimensions of the beams and columns vary as
not desirable, because in a pushover analysis the objective per the location. The columns are supported on isolated
is to study the behaviour of the building with the sequential footings. Some of the frame panels are infilled with con-
formation of hinges in the frame members, even beyond crete hollow block masonry units. Only the walls without
the failure of the struts. substantial openings were considered in the computational
Figure 7b shows the formations of hinges in the same model. The compressive strength of concrete and the yield
frame using the proposed hinge properties. The formation stress for the reinforcing steel were considered to be 25 and
of hinges in the struts was followed by the formation of 415 MPa, respectively. The compressive strength of the
hinges in the frame members. This is realistic and meets masonry units was taken as 1.65 MPa.
the objective of the pushover analysis. Figure 8 compares Similar to that in Building I, a pushover analysis using
the pushover curves using the two types of hinge properties linear hinge properties for the struts, got abruptly termi-
for the struts. The pushover curve based on linear hinge nated. The soft-storey behaviour was not reflected in the
property shows lower inelastic drift compared to that based analysis (Fig. 10a). However, when the analysis was
on the proposed hinge property. The curve based on the repeated using the proposed hinge properties, even after the
proposed hinge property reflects the effect of the modelled formation of hinges in the struts, there was formation of
nonlinearity in the equivalent struts, and provides better hinges in the columns which revealed the soft-storey
estimate of the deformation demand in each frame member behaviour of the building (Fig. 10b). Figure 11 compares
as per the state of hinge formed in it. Also, the inelastic the pushover curves using the two types of hinge properties
drift of the building leads to an appreciable lateral drift for the struts.
(P - D) effect. The analysis results can be further studied in terms of
the performances of the buildings and the frame members.
Building II Davis et al. [4] used the proposed hinge property for the
struts to analyse the performances of several open ground
Figure 9 shows a typical floor plan and sectional elevation storey buildings.
of an existing five-storeyed residential building. In the
absence of infill walls to accommodate parking of vehicles,
the ground storey is a soft and weak storey, generating a Conclusion
vertical irregularity in the building. As per the construction
drawings, the concrete slab is 120 mm thick at every floor. In a pushover analysis of a framed building, for modelling
Since there were large cut-outs near the centre of the slab the infill walls by equivalent struts, it is necessary to assign

123
J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108 107

an appropriate axial load versus deformation hinge prop-


erty to each strut. If a linear elastic relationship is used,
then the formation of hinges in the struts may lead to ter-
mination of the pushover analysis. This will not reflect an
appropriate nonlinear behaviour of the building. On the
contrary, when a nonlinear (simplified as piece-wise linear)
relationship is used, the formation of hinges in the struts is
followed by the formation of hinges in the frame members.
This is realistic and meets the objective of the pushover
analysis. Also, the pushover curve shows higher inelastic
drift due to the modelled nonlinearity in the equivalent
strut. Hence, it is recommended to use the proposed axial
load versus deformation hinge property for the equivalent
struts to model the effect of infill walls without openings.
The proposed hinge property can also be used as the
(a) Using linear hinge property backbone curve for modelling the hysteretic behaviour of
an equivalent strut in a nonlinear time history analysis of a
IO LS CP
building. Further refinement is possible for the descending
branch of the hinge property, based on deformation con-
trolled tests.

References

1. Achintya, P. Dayaratnam, S.K. Jain, Behaviour of brick infilled


RC frame under lateral load. Indian Concr. J. 65(9), 453–457
(1991)
2. H. Achyutha, S.S. Injaganeri, S. Satyanarayanan, C.S. Krishna-
moorthy, Inelastic behaviour of brick infilled reinforced concrete
frames. J. Struct. Eng. 21(2), 107–115 (1994)
3. P.G. Asteris, S.T. Antoniou, D.S. Sophianopoulos, C.Z. Chryso-
stomou, Mathematical macromodeling of infilled frames: state of
the art. J. Struct. Eng. 137(12), 1508–1517 (2011)
4. Davis, R., Menon, D. and Prasad, A. M., 2008, Evaluation of
magnification factors for open ground storey buildings using
nonlinear analyses, Proceedings, 14th World Conference on
Earthquake Engineering, Beijing, P. R. China, October 12–18
5. FEMA 356, 2000, Prestandard and Commentary for the Seismic
Rehabilitation of Buildings, Federal Emergency Management
(b) Using proposed hinge property Agency, USA
6. M. Holmes, Steel frames with brick and concrete infilling. Proc.
Fig. 10 Formation of hinges in the frame along A–A of Building II Inst Civil Eng. 19(4), 473–478 (1961)
7. IS 1893: 2002, Indian standard criteria for earthquake resistant
design of structures, Part—1. General provisions and buildings,
Bureau of Indian Standards, New Delhi
2000
8. T.C. Liauw, K.H. Kwan, Unified plastic analysis for infilled
frames. J. Struct Eng. 111(7), 427–1448 (1985)
9. A.B. Mehrabi, P.B. Shing, M.P. Schuller, J.L. Noland, Experi-
Base shear, kN

1500
mental evaluation of masonry-infilled RC frames. J. Struct. Eng.
122(3), 228–237 (1996)
1000 10. A.B. Mehrabi, P.B. Shing, Finite element modeling of masonry-
Based on linear strut hinge property
Based on proposed strut hinge property infilled RC frames. J. Struct. Eng. 123(5), 604–613 (1997)
500 11. Polyakov, S. V., 1963, Masonry in framed buildings, Gods-
udarstvenoe Isdatel’ stvo Literatury Po Stroidal stvui Architec-
ture, Moscow, Translated by G. L. Cairns, National Lending
0 Library for Science and Technology, U. K
0.00 0.02 0.04 0.06 0.08 12. A. Saneinejad, B. Hobbs, Inelastic design of infilled frames.
Roof displacement m J. Struct. Eng. 121(4), 634–650 (1995)
13. SAP, NL (Computers and structures incorporated, Berkeley,
Fig. 11 Pushover curve along X-direction (Building II) 2000)

123
108 J. Inst. Eng. India Ser. A (May–July 2013) 94(2):99–108

14. A.K. Sengupta, A. Asokan, Modelling of infill walls for nonlinear 15. B.S. Smith, C. Carter, A method of analysis for infilled frames.
static analysis of buildings subjected to seismic loads. Indian Proc. Inst. Civil Eng. 44(1), 31–48 (1969)
Concr. J. 82(2), 19–28 (2008)

123

You might also like