Engineering Structures 150 (2017) 599–621

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Modelling of masonry infilled RC frames subjected to cyclic loads: State

of the art review and modelling with OpenSees
Nurbaiah Mohammad Noh a,⇑, Laura Liberatore b, Fabrizio Mollaioli b, Solomon Tesfamariam a
a
School of Engineering, The University of British Columbia, Kelowna, BC, Canada
b
Department of Structural and Geotechnical Engineering, ‘‘Sapienza” University of Rome, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Reinforced concrete frames with unreinforced masonry infill walls represent a widely adopted building
Received 6 February 2017 system. During seismic events, infill walls, usually considered as non-structural elements, may signifi-
Revised 3 June 2017 cantly affect the characteristics of the system in terms of in-plane stiffness, strength, and energy dissipa-
Accepted 2 July 2017
tion capacity. The assessment of framed structures infilled with unreinforced masonry walls has been
investigated since the 1950s. Developing reliable numerical models of infill walls has become an impor-
tant issue since then. The analytical simplified model based on equivalent diagonal struts is often used to
Keywords:
assess the infilled frames response. The nonlinear behavior of the equivalent diagonal strut is usually
Masonry infill walls
In-plane behavior
described by constitutive laws that account for the stiffness, the strength and the hardening or softening
Macro-model behavior of the infill. The present work focuses on the evaluation of various parameters needed to define
Equivalent strut the monotonic and hysteretic response of infill walls modelled by equivalent struts. In order to select a
simple and reliable analytical model that suitable for representing the infill wall response, different strut
formulations and hysteretic models have been analysed in detail and used to reproduce several experi-
mental tests available in the literature. The numerical analyses are performed by means of the
OpenSees computer program. Three uniaxial material models available in OpenSees are used to assess
their capability in reproducing the experimental hysteretic response. Finally, from the comparison among
different models and between numerical and experimental results, suggestions are made to properly
model the in-plane non-linear response of infills.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction were conducted (e.g. [9–18]) to investigate the response of infill

Reinforced concrete (RC) frame buildings with unreinforced
masonry (URM) infill walls, henceforth denoted as masonry infilled
RC frame structures, are widely used in building constructions.
Damage observation from past earthquakes (e.g. [1–3]) show that
the seismic response of these buildings is strongly affected by
the presence of masonry infill walls. The URM infills may increase
lateral stiffness, strength, and energy dissipation capacity of the
bare RC frame [4–6]. The presence of the infill walls on the bare
frame might induce unexpected forces distribution and lead to
local collapse when it is subjected to cyclic loading [7]. These
effects depend on the infills geometrical distribution in plan and
elevation, infill mechanical properties, infill aspect ratio, configura-
tion of openings, and construction methods [6,8].
The seismic behavior of masonry infilled RC frame buildings has
been investigated since the 1950s. Many experimental studies
⇑ Corresponding author.
E-mail address: nurbaiah.mohammadnoh@alumni.ubc.ca (N. Mohammad Noh).
walls and their interaction with the surrounding RC frame. Concur-
rently, analytical studies have been carried out by researchers (e.g.
[4–6,19–25]) to model and evaluate the performance of these
buildings under both monotonic and cyclic loads. Although there
are numerous studies in the literature, the issue related to the
assessment and the modelling of masonry infilled RC frames
remain unresolved. Lack of a standardized guideline on how to
configure an infill in a global seismic analysis of infilled frames is
a main issue. Besides, the infill influence on the seismic behavior
of framed structures is widely recognized but how it is incorpo-
rated in the design process differs noticeably from one country to
another [26]. Nevertheless, masonry infill walls are often neglected
in the design process and analytical modelling, and are considered
as non-structural elements.
Reliable analytical methods are needed for the assessment of
infilled framed structures [6]. This task is rather challenging due
to the complex interaction between the frame and the infill, partic-
ularly in the nonlinear range. Moreover, the selection of the most
appropriate model is related to the analysis overall objectives:

http://dx.doi.org/10.1016/j.engstruct.2017.07.002
0141-0296/Ó 2017 Elsevier Ltd. All rights reserved.
600 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
the evaluation of local interaction effects requires the adoption of
detailed models whereas for the assessment of global response,
simpler models, capable of combining simplicity and accuracy
are more appropriate. In this regard, macro models, which involve
replacing the masonry infill with one or multiple equivalent struts,
offer a good compromise between efficiency and precision [6,21],
provided that proper rules are adopted for the definition of the
struts mechanical characteristics. As the masonry wall is a non-
homogenous anisotropic material, a valid constitutive law is
difficult to find. Previous studies highlighted that the application
of different equations to estimate the equivalent strut parameters
lead to noticeable differences in the infilled frames response [7].
Consequently, selecting adequate models for the analyses of
masonry infilled RC frames is imperative.
The aim of this study is to develop general guidelines for eval-
uating the parameters to define the monotonic and hysteretic
response of the masonry infill walls, specifically for the in-plane
lateral cyclic response. Outline of the paper is as follows. First, var-
ious modelling approaches available in the literature are reviewed.
Specific attention is devoted to the single-strut modelling method,
in which the many different formulations are discussed and com-
pared. Afterwards, numerical analyses of several one-story, one-
bay infilled RC frames tested experimentally are performed
through the Open System for Earthquake Engineering Simulation
(OpenSees) software package [27]. These analyses are dual-
purpose: (i) identify the most appropriate formulae for the defini-
tion of the strut constitutive law and (ii) verify the suitability of
numerous material models available in the OpenSees [27]. From
the comparison among different models and between numerical
and experimental results, suggestions are made to properly model
the in-plane non-linear response of masonry infilled frames.
2. Modelling approaches for masonry infilled RC frame
Comprehensive reviews on the experimental, analytical, and
numerical research works by Crisafulli et al. [28], Shing and Meh-
rabi [29], Asteris et al. [6,7], and Tarque et al. [23], categorized five
main failure mechanisms of the infilled-frames: (1) the corner
crushing (CC) mode; (2) the sliding shear (SS) mode; (3) the diag-
onal cracking (DK) mode; (4) the diagonal compression (DC) mode,
and (5) the frame failure (FF) mode. The first failure mode, CC,
occurs when a weak masonry infill panel is surrounded by strong
frame members and the connection between the infill and the
frame is weak. The second failure mechanism, SS, which is given
by the horizontal sliding through multiple bed joints at approxi-
mately mid-height of the panel, occurs when the strength of mor-
tar joints is low and the infill is surrounded by a strong frame. This
failure mode also introduces the short column behavior, in which
plastic hinges develop at mid-height of the columns. This can also
lead to the shear failure of the columns. The third failure mecha-
nism, DK, that takes place when the frame is more flexible than
the masonry infill, consists of a crack along the compressed diago-
nal of the masonry infill panel, which might be joined with a hor-
izontal crack at the mid height of the panel. The fourth mechanism,
DC, displays crushing at the centre of the masonry infill panel due
to high compression forces. Masonry crushing may also develop
when a slender flexible infill experiences large out-of-plane defor-
mations. The last mechanism, FF, consists of a purely flexural
mode, in which there is no separation between the infill and the
frame under a low load level. Under a moderate load level, if the
infill panel and bounding frame are not properly tied together,
the infill panel will separate partially from the frames and flexural
hinging at the top and bottom of the column will develop. This fail-
ure mode is particularly important when investigating the existing
structures, which in many cases display the frame weaknesses. The
general conclusion from these studies is that different failure
modes may develop depending on the infill and frame relative
stiffness.
The possible failure mechanisms described above are only for
solid infill walls (without openings). Previous studies also involved
experimental work on masonry infill walls with openings
[12,14,29–33]. Experimental results highlighted that masonry
infills wall with openings can still enhance the performance of
RC frames, in terms of resistance, stiffness, ductility, and energy
dissipation capacity, but to a lesser extent than a solid infill wall.
An experimental study by Kakaletsis and Karayannis [12,14]
reported that masonry infills with an opening crack and detach
from the bounding frame at an early stage before the column yield-
ing. The location of the openings also influences the global perfor-
mance [30,33]. Moreover, when the frames are only partially
infilled, short column effect consisting in a brittle shear failure
may develop. However, there are inconsistent evidence were accu-
mulated whether the openings lead to a more brittle or more duc-
tile behavior than the solid infill wall [34].
Theoretically, the knowledge of both the geometric properties
of the infill and of the cyclic behavior of the masonry material
should allow a straightforward modelling. The intricate infill-
frame interaction lead to diverse possible modelling assumptions
characterized by different computational efforts and accuracy
[35]. Different approaches have been developed and validated with
experimental test results for the modelling of masonry infills in
order to capture the different possible failure mechanisms that
take place in infilled RC frames. They are divided into two main
groups based on the modelling approach: micro-modelling tech-
nique e.g. [4,36] and macro-modelling technique [20,41]. A third
approach, which have similar precision with micro-modelling
technique, is known as the meso-modelling technique [7,23].
These approaches differ mainly in the degree of their infill panel
modelling detail, as described in the following sections.
2.1. Micro-modelling and meso-modelling
The micro- and meso-modelling approaches using nonlinear
finite element (FE) methods are able to provide an accurate
description of the interaction between the frame and infill at the
local level [4]. The micro-modelling approach includes two differ-
ent modelling strategies: the detailed and the simplified micro-
modelling [36]. The former is based on the use of continuum ele-
ments for masonry units and mortar joints, whereas the unit-
mortar interface is represented by discontinuous elements
(three-phase material) for the simulation of fracture behavior in
brick units and mortar joints [11,7,34]. The latter resorts to
expanded units represented by continuum elements and the dis-
continuous elements (two-phase material) for the mortar joints
and the unit-mortar interface [7,8]. The response of each element
is described through proper constitutive relations. Chiou et al.
[37] developed several plasticity-based continuous interface mod-
els to capture the shear sliding in the joint.
In the meso-modelling approach (one-phase material), units,
mortar, and their interface are smeared out and the masonry is
treated as a continuum through a homogenization process [7].
Contact, gap or spring elements can be considered for modelling
the interface infill-frame, allowing the separation of the infill from
the bare frame [38]. Smeared crack elements can be used to model
both reinforced concrete frames and brick infills. For instance, in
the analyses reported in [19,39], infill panels are modelled as a
homogeneous material before fracture, and the effect of mortar
joints is smeared out [29,40].
Overall, the above-mentioned methods are quite complex due
to a large amount of information required for the constitutive mod-
els and high computational effort involved which lead to delayed
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
assessment, particularly for large structures. The micro-modelling
approach is normally suitable for small buildings or for buildings
sub-assemblages and focuses on the assessment of the interaction
between bricks and mortar.
2.2. Macro-modelling
The complexity and computational effort required by the
micro-models led to the development of macro-models. Macro-
models use few elements and do not consider any distinction
between the masonry units and joints. This approach provides sim-
ple and efficient tools to describe the global response of masonry
infilled RC frames. The diagonal equivalent strut model, which con-
siders only the compressive force, is often adopted for the in-plane
seismic analyses of solid masonry infilled frames [20,41] and it has
been updated to account for the out-of-plane/in-plane interaction
[42]. The identification of the mechanical characteristics of the
equivalent strut depends on the actual system properties. The
macro-model should represent all the important properties of the
masonry wall including strength, stiffness and deformation
capacity.
Polyakov [43] first suggested that masonry infill walls subjected
to lateral loads could be represented by a diagonal no-tension strut
element with appropriate mechanical properties. The idea was
based on the observation that the infill wall separates from the sur-
rounding frame at a moderate drift level, so that a compression
diagonal path develops along two opposite corners of the wall.
Since then, many researchers have developed this model, by esti-
mating the width, the mechanical properties, the number, and con-
figuration of the equivalent struts. For the evaluation of the strut
parameters, e.g. the strut width, various predictive equations have
been proposed. For example, Holmes [44] continued the idea of a
single pin-jointed diagonal strut and proposed that the effective
width of the strut model should be 1/3 of the diagonal length.
Afterwards, numerous researchers continue to study the determi-
nation of the effective strut width considering other characteris-
tics, e.g. the relative stiffness parameter, kh, the uncracked or
cracked condition of the panel, etc., to control the stiffness and
strength of the masonry infill (detail of the proposed equation
can be found in Tarque et al. [23] and Asteris et al. [6]). The stiff-
ness is expressed as a function of the effective width of the strut,
whereas the maximum capacity is related to a predefined failure
mode. However, Amato et al. [45,46] introduced a new equation
for the stiffness of the equivalent strut which takes into account
also the vertical load transferred from the bounding frame to the
masonry infill walls. The influence of vertical load is integrated into
the equivalent strut width estimation.
A single diagonal strut in each direction, as shown in Fig. 1, can
be used to represent the behavior of the infill walls under lateral
loads in global effect. However, the single diagonal strut cannot
be used to model and describe in detail the complex behavior of
the RC frame with infill walls subjected to cyclic loading. Observa-
tion from previous studies highlighted that a single diagonal strut
cannot be used to adequately capture the bending moments and
shearing forces in the RC frame members [47,49,52,53]. To over-
come this drawback, the single strut model was improved by intro-
ducing multiple struts. However, the use of different numbers of
strut and strut configurations might lead to a different result of
the overall response of the infilled frames. The strut model was
also extended to infills with the openings according to two differ-
ent methods: (i) the use of several diagonal struts around the
opening as in [47,48]; and (ii) the application of proper coefficients
to modify the strut stiffness and strength [21,49–51].
Crisafulli [53] investigated the influence of the different multi-
ple strut configurations in the structural response of masonry
infilled frames. Afterwards, Crisafulli and Carr [54] developed a
601
new multi-strut macro-model including two parallel off-diagonal
struts and a special shear spring to account the diagonal tension
failure and shear mechanism in the infill walls (Fig. 2(a)). This
model, however, involves complicated calibration for the required
parameters, which are difficult to define in many cases.
El-Dakhakhni et al. [52] proposed an analytical method to esti-
mate the stiffness and the ultimate load capacity of concrete
masonry-infilled steel frames by using multiple struts model. The
proposed analytical method is based on the nonlinear behavior of
the steel frame and masonry panel. To estimate the bending
moments and shearing forces in the frame members, two addi-
tional off-diagonal struts are introduced at the points of maximum
moments in the beams and the columns, which approximately
occurred at the end of the contact lengths. The whole panel is,
therefore, represented by three struts, as shown in Fig. 2(b). In this
method, each strut is defined by force-deformation laws based on
the orthotropic behavior of the masonry panel. The use of a multi-
strut model also simplifies the modelling of the interaction
between different panels in multi-story buildings.
Sattar and Liel [55] proposed a FE-enhanced strut model
approach that can simulate the in-plane seismic response of struc-
tures up to ultimate failure. Two diagonal compression struts (in
each direction) are adopted to represent the behavior of the infill
wall (Fig. 2(c)). The proposed model also focuses on the modelling
of RC-frame elements in order to study the interaction between the
infill wall and the frame. However, the model is not able to capture
completely the infill-frame interaction that might lead to shear
failure at the bottom of the columns and near the beam-column
joints. To overcome the problem of capturing shear failure in the
column, Burton and Deierlein [22] developed a simple model by
utilizing a pair of inelastic compression only strut in each direction
(dual-strut model) (Fig. 3). The dual-strut model can capture the
infill wall strength and stiffness deterioration by using a peak-
oriented hysteretic model and the shear failure of the column.
Reinforced concrete columns and beam by elastic elements with
zero-length flexural plastic hinges combined with a rigid softening
shear degradation model. This approach can capture loss in the
axial (gravity) load carrying capacity due to column shear failure.
2.3. Single strut modelling
The simplest modelling option to simulate the behavior of infills
is using a single- or double-diagonal strut for monotonic or cyclic
loading, respectively [23]. In this study, the cyclic response of the
diagonal struts is reproduced by using different hysteretic materi-
als available in OpenSees [27] software. The main differences
among these material models are discussed and shown with exam-
ples. The goal of the present study is to select the simplest and
most accurate model, in terms of input parameters and character-
ization of hysteretic curves, respectively. The following section
briefly summarized the main important task that need to consider
during the equivalent strut calibration, which include selection of
the force-displacement relationship definition and cyclic behavior,
and the evaluation of input strut parameters (e.g. equivalent strut
width), the stiffness and the strength.
2.3.1. Backbone curve – Constitutive law
The selection of reliable constitutive models (force-
displacement relationships) to describe the equivalent strut behav-
ior are required in order to conduct nonlinear analyses of masonry
infilled RC frames. Numerous constitutive hysteretic models or
semi-empirical relationships have been proposed, mostly based
on experimental parametric studies. The constitutive law for a sin-
gle equivalent strut can be described via a multi-linear relationship
[20,40,56–60] or via a combination of linear and nonlinear
branches [57,61,62]. However, many studies used the monotonic
602 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

Δ 3 4

Concentrated or Distributed
plasticity model

Truss element with uniaxial material (i.e.,

1 2
Concrete01, Hysteretic, and Pinching04)

Fig. 1. Schematic diagram of the equivalent strut model for infilled frames.

Beam-column
joint

hz Masonry strut

column
beam θ
Shear spring hz

(a) (b)

(c)
Fig. 2. Multiple strut model proposed by: (a) Crisafulli and Carr [53,54]; (b) El-Dakhakhni et al. [52]; and (c) Sattar and Liel [55].

constitutive model as the backbone curve for the cyclic response.
The constitutive law is used to represent the initial stiffness, the
strength and the panel post-peak behavior in compression. The
tensile strength is often neglected in the analyses. If the multiple
strut model is adopted in the analysis, the evaluation of the consti-
tutive law is necessary for each strut.
To investigate the effect of different strut constitutive models
available in the literature, three monotonic constitutive laws are
discussed: Liberatore and Decanini [59,60], Panagiotakos and Far-
dis [57], and Dolšek and Fajfar [41]. In this study, these models will
be tested and compared to experimental test data. The best-fit
model will be used to further investigate the suitability of various
hysteretic laws.
The model proposed by Liberatore and Decanini [59,60]
assumes a four branches skeleton curve of the lateral force-
displacement (Hm-d) relationship(Fig. 4(b)). The first segment (line
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 603

Beam flexural
hinge

Column flexural
Off-diagonal hinge and shear
compression-only spring in series
infill strut

Column flexural hinge,

Central shear and compression-only
compression-only axial spring in series
infill strut at column base

Fig. 3. Schematic diagram of typical infill frame model proposed by Burton and Deierlein [22].

(a) (b)
Fig. 4. In-plane lateral force-displacement envelope of the model proposed by: (a) Panagiotakos and Fardis [57] and (b) Liberatore and Decanini [59,60].

OF) represents the un-cracked phase up to the first cracking (Hmf at where Em is the elastic modulus of the infill, Ec is the elastic modu-
point F) of the infill. The second segment (line F-FC) of the envelope lus of concrete, h is the slope of the strut with respect to the hori-
corresponds to the post-cracking phase up to the development of zontal axis, tm and hm are the thickness and the height of the
the maximum strength (Hmfc). The descending third phase (line masonry panel, respectively, h is the story height, I is the moment
FC-R) defines the post-peak strength deterioration of the infill up of inertia of the columns. Parameters K1 and K2 depend on the
to the residual strength, Hmr, reached at the displacement ur. parameter kh, as shown in Table 1.
The stiffness at complete cracking stage (Kmfc) and the strength The lateral stiffness of the equivalent strut of Kmfc (in the hori-
(Hmfc), that are required for the proposed model, are estimated zontal direction) at the complete cracking stage is given by the fol-
according to [56,58]. All parameters are defined as functions of lowing relation:
geometric and mechanical characteristics of the infill and the sur-
rounding frame. The width of the equivalent strut bm, which is Em tm bm
K mfc ¼ cos2 h ð3Þ
needed for the stiffness calculation, is determined through the rel- dm
ative stiffness parameter kh proposed by Stafford-Smith [63] and
where dm is the length of the equivalent strut, the meaning of Em, tm,
by two constants K1 and K2 calibrated based on the experimental
bm and h are the same as mentioned above. Decanini et al. [56,58]
tests:

 calculate the strength of Hm value by considering four possible fail-
K1 ure modes and the corresponding failure stresses: (a) diagonal ten-
bm ¼ þ K2 d ð1Þ
kh sion, Hdt; (b) sliding shear, Hss; (c) corner crushing, Hcc; and (d)
diagonal compression, Hdc. The equivalent failure stresses for each
where kh is a non-dimensional parameter depending on the geo-
failure mode are calculated as:
metric and mechanical characteristics of the frame-infill system,
K1 and K2 are coefficients that change according to kh, and d is the
length of the equivalent strut. The parameter kh is defined by the Table 1
following expression: Coefficient value to calculate the properties of the strut [56].
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi kh < 3.14 3.14 < kh < 7.85 kh > 7.85
4 Em t m sinð2hÞ
kh ¼ h ð2Þ K1 1.3 0.707 0.47
4Ec Ihm K2 0.178 0.010 0.04
604 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

0:6sm0 þ 0:3r0 of the un-cracked panel; (b) post-elastic linear response, character-
Hdt ¼ ð4Þ
bm =d ized by a reduced value of stiffness due to the detachment of the
infill from the surrounding frame; (c) softening response of panel
ð1:2 sin h þ 0:45 cos hÞs0 þ 0:3r0 after the maximum force; and (d) attainment of the residual axial
Hss ¼ ð5Þ
bm =d strength at a given value of displacement. To fully define the back-
bone curve of the Panagiotakos and Fardis [57] model, the follow-
1:12 sin h cos h ing parameters are needed. The initial shear stiffness R1 of the
Hcc ¼ rm0 ð6Þ uncracked masonry wall can be evaluated as follows:
K 1 ðkh Þ0:12 þ K 2 ðkh Þ0:88
Gm tm lm
1:16 tan h R1 ¼ ð12Þ
Hdc ¼ rm0 ð7Þ hm
K 1 þ K 2 ðkh Þ
where tm, lm, and hm are the masonry panel thickness, length, and
where sm0 is the vertical compression strength of the masonry mea- height, respectively, while, Gm is the shear modulus of the masonry
sured on masonry specimens; r0 is the total vertical stress due to infill obtained in diagonal compression tests. The lateral cracking
gravity loads; rm0 is the shear strength obtained from diagonal strength Fy, assumed equal to:
compression tests; and s0 is the slide resistance in the joints mea- F y ¼ f ms t m lm ð13Þ
sured from tests on triplets or obtained from the empirical relation
s0 = 0.7 sm0. where fms is the masonry failure stress, as evaluated by the
Assuming that the infill strength corresponds to the minimum diagonal-compression tests. Then, the displacement at crack (yield-
value among the four-failure modes, the lateral strength of the ing) dy can be determined as:
equivalent strut (in the horizontal direction) is given by: Fy
dy ¼ ð14Þ
Hmfc ¼ ðHÞmin t m bm cos h ð8Þ R1

The stiffness of the uncracked infill K0, is evaluated using the elastic The axial post-cracking stiffness can be calculated using the follow-
stiffness of the frame (Kt0) and the infill stiffness at a complete ing expression:
cracking stage Kmfc as the following equation [59,60]: Em t m bm
R2 ¼ ð15Þ
K 0 ¼ 3K t0 þ 4K mfc ð9Þ dm
where Em is the Young modulus of masonry and the width bm com-
where
puted according to [64], as reported in [65]:
1 þ 6/ 12Ec Ip
K t0 ¼ ð10Þ bm ¼ 0:175ðkh hm Þ
0:4
dm ð16Þ
2 þ 3/ h3
where kh is defined as in Eq. (2)., and dm is calculated as follow:
and
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2
Ibeam h dm ¼ hm þ lm ð17Þ
/¼ ð11Þ
Icolumn l
The lateral maximum strength, Fm, assumed equal to:
Ibeam and Icolumn are the moments of inertia of the cross section of
F m ¼ 1:3F y ð18Þ
the beam and the column, Hmf takes the value around 80% of Hmfc,
and Hmr takes the value around 35% of Hmfc, according to the coeffi- and the displacement dm is corresponding to the maximum force:
cient value assumed in [59].
FmFy
The constitutive law proposed by Panagiotakos and Fardis [57] dm ¼ dy þ ð19Þ
R2
also consists of four segments. It was directly derived by compar-
ing the experimental results on bare and infilled RC structures. The stiffness of the softening branch R3 can be taken within the
Conceptually, the proposed model is based on four different states range from 0.5% to 10% of the initial stiffness R1. The lateral residual
of stress, as shown in Fig. 4(a): (a) initial elastic and shear behavior force Fr can be assumed equal to 5–10% of the maximum strength.

(a) (b)
Fig. 5. Hysteretic law proposed for the equivalent struts: (a) Klingner and Bertero [65]; (b) Cavaleri et al. [61,62].
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 605

Finally, the ultimate displacement dr is corresponding to the resid- opening). The slope of the softening response R3 is evaluated
ual force: through the model by Panagiotakos and Fardis [57],
FmFr described in the previous subsection.
dr ¼ dm þ ð20Þ
R3
2.3.2. Cyclic behavior
Dolšek and Fajfar [41] proposed an alternative model, which is the- The equivalent strut utilization to perform nonlinear analyses
oretically similar to the model proposed by Panagiotakos and Fardis requires a practical conceptual cyclic law model to simulate the
[57]. The only key difference between the two models is the hysteretic behavior of the masonry infills under cyclic loading.
assumption of the zero-residual strength (Fr = 0). This model Notably, the cyclic behavior of the masonry infilled frame is quite
requires four parameters as described below: difficult to capture. A set of rules is needed to account for the actual
behavior. Klingner and Bertero [65] proposed a model based on a
(1) The lateral cracking strength Fy is assumed to be equal to dynamic test. Fig. 5(a) shows the compression strength of envelope
60% of Fm. curve, which exhibits an exponential degradation beyond the peak
(2) The displacement at crack (yielding) dy can be determined strength. Cavaleri et al. [61,62] introduced some modification to
through simple analytical relationships similar in Eq. (14), the model developed by Klingner and Bertero [65] to enhance
which is the initial stiffness is defined as R1 in Eq. (12). the accuracy of the loading and unloading branches (Fig. 5(b)).
(3) The maximum lateral force Fm is defined according to the The cycles of the modified model are based on the envelope
proposal by Žarnić and Gostič [66]: strength curves of the strut. Besides, the model suggested by Cava-

qffiffiffiffiffiffiffiffiffiffiffiffiffiffi leri et al. [61,62] considers the pinching effect during the reloading
f ms t m lm by a zero value of the restoring force before reloading. The predic-
F m ¼ 0:818 1 þ C 2I þ 1 ð21Þ
CI tion of a cyclic behavior of the masonry infilled RC frames requires
three parameters (b, px, and py). These parameters are related to
where
the definition of the unloading and reloading branches.
lm Panagiotakos and Fardis [57] proposed a hysteretic interstory
C I ¼ 1:925 ð22Þ
hm shear force-drift relationship (Fig. 6), which has a multilinear curve
in monotonic loading, to represent the cracking, ultimate strength,
post-ultimate branch, and residual strength. The model requires
(4) The maximum horizontal displacement dm is defined as a
three empirical parameters to control the loading and unloading
percentage of the wall height (0.2% for plain walls, 0.15%
branches.
for walls with a window, and 0.10% for walls with a door
Crisafulli [53,54] developed a more detailed cyclic model that
can simulate different failure mechanisms. The masonry behavior
under cyclic loading is presented by different axial force-
relationships. The most significant novelty of the model proposed
by Crisafulli is the addition of a shear spring to account for the
shear behavior of the infill panel. This model, however, is quite
complex because it requires a very large number of parameters dif-
ficult to calibrate, especially when test results on the materials are
not available.
Liberatore [59] proposed a simple hysteretic model, which takes
into account the degradation of stiffness in the unloading branch
and the degradation of strength under displacement cycles of con-
stant amplitude through two parameters: a and b. The parameter a
controls the stiffness of the unloading branch, its value ranges from
0 to 1, for a = 1 the unloading stiffness is the same as the initial
one, as shown in Fig. 7(a). The parameter b enables the control of
reloading stiffness degradation according to the following expres-
sion (Fig. 7(b)):


Eh
Fig. 6. Hysteretic law for equivalent strut proposed by Panagiotakos and Fardis H00 ¼ H0 1  b  ð23Þ
[57]. E

(a) (b)
Fig. 7. Hysteretic law for equivalent strut proposed by Liberatore [59].
606 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

where Eh is the cyclic dissipated energy and E⁄ is the area enclosed ð1  ac Þðac hÞtm rc þ ðab lÞðt m sb Þ
Rcc ¼ ð30Þ
in the force-displacement envelope up to the attainment of the cos h
residual strength. The parameters calibration and the adopted
model validation were made through the comparison between 0:5hm t m f a
Rdc ¼ ð31Þ
numerical and test results of different types of masonry [59]. cos h
ft and c are the tensile strength of the infill material and load factor,
2.3.3. The infill material properties respectively, and h0 is depicted in Fig. 8. Moreover, ach and rc are the
Numerous studies proposed different models to estimate the contact lengths and contact stress between the column and infill,
stiffness and the strength of the diagonal strut. In the following, and abh and sb are the contact lengths and contact stress between
those adopted in this study are briefly discussed. beam and infill, respectively. fa is the permissible stress and leff is
the effective length of the strut (see Fig. 8),and is computed as:
2.3.3.1. Evaluation of the strut strength. Several studies [56,67–70]


suggested that the evaluation of the strut maximum compressive leff

f a ¼ 0:39f m 1  ð32Þ
strength should depend on the possible failure mechanisms of 40tm
the panel, which appeared in both experimental tests and actual In the model by FEMA 306 [67], the diagonal tension strength is cal-
behavior under seismic action. The final compressive strength of culated as in Eq. (25). The cracking strength of masonry infill is
the strut is selected based on the minimum possible resistance cal- assumed equal to 1/10 of the compressive strength. The sliding-
culated by specific equations as described below. Comprehensive shear capacity of the infill is determined as:
studies by Haldar et al. [71] and Moretti [72] reported that the typ-
ical failure mechanism of the masonry infill can be classified into ðs0 þ lry Þtm lm
Rss ¼ ð33Þ
four groups: (i) bed-joint sliding/sliding shear failure; (ii) diagonal cos h
tension cracking; (iii) diagonal compression failure; and (iv) corner whereas in FEMA 356 [73] the shear strength of the infill is calcu-
crushing of infill. Table 2 provides an overview of four models that lated as:
considers at least three different failure modes and that are fre-
quently cited in the literature. The models are adopted in this Rss ¼ Ani  f v ie ð34Þ
study to define an adequate strength of the diagonal strut.
Decanini and Fantin [56] suggested that the axial strength of the
strut based on the different failure mode can be determined using
Eq. (4)–(7). All parameters involved in these equations are defined
in the previous section (Section 2.3.1).
Paulay and Priestly [68] and Priestley and Calvi [69] defined the
axial force that can initiate a diagonal tension cracking (Rdt), bed-
joint sliding failure (Rss), and diagonal compression strength (Rdc)
as:
p
Rdt ¼ t m dm f m ð24Þ
2

0:03f m
Rss ¼   t m dm ð25Þ
1  0:3 hl

2
Rdc ¼ atm f m sec h ð26Þ
3
Fig. 8. Geometrical parameters used in the model proposed by Saneinejad and
where a, vertical contact length between infill and column, is Hobbs [70].
defined as
p
h
a¼ ð27Þ Table 3
2 k
Equations for the width of the equivalent strut.
Saneinejad and Hobbs [70] defined the strength of the strut in the Reference Equivalent width, bm
four common failure modes as:
Decanini and Fantin [56] (DF) As in Eq. (1)
pffiffiffi
Rdt ¼ 2 2t m hm f t cos h ð28Þ Paulay and Priestly [68] bm ¼ d4m
Priestly and Calvi [69] (PPC)
8
( Saneinejad and Hobbs [70] (SH) a s
< ð1ac Þðrc hÞð f cc Þþðab lÞð f cb Þ
cso tm dm bm ¼ min cos h
Rss ¼ min 10:45 tan h0 ð29Þ : f
hm ðf a Þ
0:5 c

0:83ctm dm cos h
FEMA 306 [67] bm ¼ 0:175ðkh Þ0:4 dm
where; kh as in Eq. (2)

Table 2
Failure modes of masonry infill.

Reference Diagonal tension cracking, Rdt Sliding shear failure, Rss Corner crushing of infills, Rcc Diagonal compression failure, Rdc
Decanini and Fantin [56] (DF) Yes Yes Yes Yes
Paulay and Priestly [68] Yes Yes No Yes
Priestly and Calvi [69] (PPC)
Saneinejad and Hobbs [70] (SH) Yes Yes Yes Yes
FEMA 306/356 [67,73] Yes Yes Yes No
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 607

fvie is the expected shear strength of infill wall and Ani is the area of
net-mortared section. The Ani is computed as:
Ani ¼ 0:6bm lm and f v ie ¼ 0:75ð0:75so þ ro Þ
Finally, the corner crushing failure, which is modified from the
equation provided by Stafford-Smith and Carter [74] expressed as:
Rcc ¼ bm t m f m90
where fm90 is the compressive strength measured in horizontal
direction.
2.3.3.2. Evaluation of the equivalent strut width. The width of the
strut, bm is important parameters in the equivalent strut modelling
method and various methods are proposed in the literature
[56,67–70](see summarized in Table 3).
3. Structural and non-structural modelling with OpenSees
This section provides in detail the procedure adopted to select
ð35Þ the most adequate models for both the RC members and the infill.
The simulations and calibrations were performed in OpenSees.
3.1. RC frame modelling
ð36Þ
The non-linear behavior of RC frame structures subjected to
cyclic loads can be characterized by the development of plastic
hinges close to the element ends, even though the interaction
between the infill and frame may lead to shear failure in the col-
umns. Numerical approaches need to be utilized to simulate the
various nonlinearities arising from the seismic cyclic demands.
To this aim, the concentrated- (lumped) and distributed- (fibre
based) plasticity models are the most accepted, due to their robust-
ness, flexibility and capability to simulate the global response of
framed structures.

Fig. 9. RC frame modelling using distributed plasticity elements.

Fig. 10. RC frame modelling using concentrated/lumped plasticity elements.

608 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

3.1.1. Fibre-based distributed-plasticity model
The fibre-based distributed plasticity approach, permits the
spread of the inelastic behavior by numerical integration over
the element cross-section and along the deformable region of the
member length [75]. This method requires the discretization of
the element into fibres, as shown in Fig. 9. For each element, five
integration points are usually considered. The nonlinear hysteretic
characteristics in the cross-sections are captured by defining the
uniaxial materials, i.e. unconfined concrete, confined concrete or
steel reinforcement. Several control parameters associated with
the hysteresis model need to be considered to establish the rule
for inelastic reversed cyclic loading. The numerical convergence
issue can be improved by increasing the number of fibres in each
cross-section. The fibre based-model captures the axial-moment
(P-M) interaction and can be used where the cracking and
tension-stiffening behaviors govern the response. However, this
model is not able to properly represent the deterioration phenom-
ena of the steel reinforcing bars due to buckling and low-cycling
fatigue.
the accelerated reloading stiffness deterioration. The model
requires six parameters: elastic stiffness (Ke), effective yield
strength (My), strain hardening ratio (Mc/My), pre-capping rotation
(hcap), post-capping rotation (hpc), and cyclic deterioration parame-
ter (k). Fig. 11 shows the moment-rotation law (backbone) accord-
ing to Ibarra et al. [76]. All the parameters are determinable using
the predictive equations and tools developed by Haselton et al.
[77], as shown in Table 4.
This simulation approach can be used for collapse prediction, to
captures deterioration of the steel reinforcing bars due to rebar
buckling and low cyclic fatigue and to records the strength and
stiffness deterioration to assess the global collapse. Overall, this
approach is simple, effective for interface effects, and computation-
ally efficient. However, it requires the moment-rotation relation-
ship, instead of stress-strain relationships and it is incapable of
capturing the P-M interaction, which is particularly important for
the columns.
3.2. Infill wall panel modelling

3.1.2. Lumped-plasticity modelling approach

In OpenSees, the equivalent diagonal struts are modelled using
The lumped plasticity model can be introduced using an elastic
truss elements. Several parameters are required to define the
beam-column with two zero-length elements at both ends
strength degradation, and loading and unloading rules. In this
(Fig. 10). The zero-length elements corresponds to a rotational
study, three material models available in OpenSees are adopted
hinge having a hysteretic response. The rotational hinge behavior
and compared: Concrete01 material, Hysteretic material, and Pinch-
is defined by uniaxial material models, which describe the
ing4 material. These models are able to capture the strength and
moment-rotation or force-deformation relationship.
stiffness degradation, the loading and unloading cycles and the
The most recent OpenSees implementation, the peak oriented
hysteretic model developed by Ibarra et al. (modIMKmodel) [76],
can be used to describe the behavior of the RC beam-column ele-
ment. This model captures the four modes of cyclic deterioration,
which includes the basic strength deterioration, the post-capping
strength deterioration, the unloading stiffness deterioration, and

Fig. 12. Concrete01 Material model in OpenSees.

Fig. 11. Monotonic behavior of component model developed by Ibarra et al. [76].

Table 4
Plastic hinge parameters in the model by Haselton et al. [77].

Model Parameter Predictive equations [77]

h i
K e ¼ EIyg
EI P 0:2
0:07 þ 0:59t þ 0:07 LHS
< 0:6
1:25ð0:89Þv ð0:91Þ0:01cunits f
Mc 0
c
My
hcap;pl v 0
0:13ð1 þ 0:55asl Þð0:13Þ ð0:02 þ 40qsh Þ0:65 ð0:57Þ0:01cunits f c

hpc 0:76ð0:31Þv ð0:02 þ 40qsh Þ1:02 6 0:10

k 170:7ð0:27Þv ð0:10Þs=d
0
where m, Ls =H, qsh , f’c, asl , and s/d is the axial load ratio ½P=Ag f c, the shear span ratio,
the lateral confinement ratio, the concrete strength, the slip term (=1), and the ratio
of transverse tie spacing to the column depth, respectively. Fig. 13. Hysteretic Material model in OpenSees.
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 609

earity of the stress-strain constitutive law. The parameters

required are: the maximum stress, the strain corresponding to
maximum stress, the ultimate stress and ultimate strain corre-
sponding to the ultimate stress, as shown in Fig. 12. In Concrete01
material, the tangent line of the first linear branch of the curve (Em)
is evaluated as two times the maximum stress divided by the max-
imum strain. The strength degradation of this material object is
degraded by the linear unloading/reloading stiffness.

3.2.2. Hysteretic material

The Hysteretic material is a uniaxial trilinear hysteretic material
object, which includes the pinching effect, the cyclic deterioration
due to deformations and dissipated energy, and the degradation of
the unloading stiffness based on ductility (Fig. 13). The parameters,
such as initial slope, maximum stress, failure strain, pinching factor
and degraded unloading stiffness are needed to study the cyclic
response of the infill wall. The pinching factor for strain during
reloading (PinchX), pinching factor for stress during reloading (Pin-
chY) and power used to determine the degraded unloading stiff-
ness based on ductility (Beta) are determined through the
Fig. 14. Piching4 Material model in OpenSees.
comparisons with experimental data. Values equal to 1, 0.1 and
0.5, are found for the three parameters. Damage parameters are
residual strength. Details of each material model are discussed not considered.
below.
3.2.3. Pinching4 material
3.2.1. Concrete01 material The Pinching4 material model simulates a pinched load-
The Concrete01 material is a zero-tensile strength material. It deformation response of structural components that exhibit degra-
requires only minimum data to define the features of the nonlin- dation in flexural strength and stiffness under cyclic loading
(Fig. 14). In the loading direction of interest (i.e. positive and/or
Table 5 negative), two parameters – rDisp/rForce – define the ratio of the
Geometric characteristic of RC frame in the Experiment 1 [17] and 2 [19]. deformation/force (in this case rotation/moment), in which the
Design Characteristics Experiment 1 Experiment 2 reloading begins at the maximum deformation/force demand of
1:3 2:3 the previous loading cycle in the loading direction of interest.
Length (cm)-axis to axis of the column 159 225 The parameter uForce defines the ratio of flexural strength devel-
Height (cm) 106 172.5 oped upon unloading from the negative/positive load to the maxi-
Cross-section of the column (cm) 15  15 15  15 mum flexural strength developed under monotonic loading. The
Cross-section of the beam (cm) 10  20 15  20
cyclic deterioration in unloading and reloading stiffness and flexu-
Longitudinal of reinforcement of the column 8Ø6 8Ø8
Tensile and compression reinforcement of 3Ø6 3Ø8 ral strength are controlled with gKLim, gDLim, and gFLim, respec-
the beam tively. The values of cyclic degradation parameters (gKLim,
Stirrups Ø2.7 Ø4 gDLim, and gFLim) and pinching parameters (rDispP, rDispN, rFor-
Spacing stirrups in critical regions (cm) 3.33 5
ceP, rForceN, uForceP and uForceN) need to be adjusted to simulate
Spacing stirrups in non-critical regions (cm) 6.67 10
the force-deformation history for selected specimens.

Table 6
Material properties in the experimental studies considered.

Material Compressive Strength (MPa) Shear Strength (MPa) Yield Stress (MPa) Ultimate Stress (MPa)
Concrete
Experiment 1 26.5
Experiment 2 33.1(M1)
28.3(M2)
Brick unit
Experiment 1 (6.3 cm) 6.0
Experiment 2 –
Mortar
Experiment 1 (Type SASTM) 10.7
Experiment 2 6.3
Masonry
Experiment 1 (Type S mortar) 4.2 0.33
Experiment 2 2.1 0.4
Longitudinal Steel bars
Experiment 1 (Ø6) 348 457
Experiment 2 (Ø8) 523 552
Transverse Steel bars
Experiment 1 (Ø2.7) 271 395
Experiment 2 (Ø4) 434 519
610 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

4. Numerical simulations with OpenSees: Selection and
calibration
The numerical analyses performed in this section are aimed at
calibrating and validating response of the different models with
experimental test data (both bare- and infilled-RC frames sub-
jected to in-plane cyclic loading.
4.1. Experimental data
Two sets of experimental tests were considered. The experi-
mental studies selected here were conducted by Stylianidis [17]
and Pires [9], denoted in this paper as Experiment 1 and Experi-
ment 2, respectively.
Stylianidis [17] performed quasi-static cyclic tests on 1/3
scaled single story-single bay bare and infilled RC frames. The
infill walls were made of perforated clay brick units with
dimension 190  63  90 mm and a 7.5 mm thick mortar. The
specimens FBN (bare frame) and F1N (masonry infilled frame)
from the first program were selected for the comparison. The
loading program is based on displacement controlled loading
cycles, with increasing displacements up to 30 mm. An axial
compressive load of 80 kN (14% of the column strength) was
applied on the columns.

Table 7
Material properties assigned to the structural element for the selected specimens.

Material Experiment ID f’c [MPa] Ec [MPa]

Concrete01 Experiment 1 26.5 4700 (f’c) 1/2

Experiment 2 M1 = 33.1
M2 = 28.3 – –
fy, [MPa] Es, [GPa] b
Steel01 Experiment 1 348 210 0.01
Experiment 2 434 0.006

Fig. 15. Bare frame, Model 1. Numerical and experimental results: (a) Experiment 1, and (b) Experiment 2.

Fig. 16. Bare frame, Model 2. Numerical and experimental results: (a) Experiment 1, and (b) Experiment 2.
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 611

Pires [9] tested 2/3 scaled single story-single bay bare (M1) and
infilled (M2) frames. The brick units had dimensions
300  200  150 mm with 1:4 mortar’s proportion. These speci-
mens were tested under alternated horizontal actions. The loading
program consisted in displacement controlled loading cycles that
imposed increasing displacements up to 100 mm, unless failure
of both columns was observed. A vertical force of 100 kN was
applied on both columns.
Tables 5 and 6 summarize the geometric characteristics and
mechanical properties adopted for the specimens considered in
this study.

4.2. Numerical model

The experiments were reproduced as two-dimensional models.

Fig. 17. Cumulative displacement dissipated energy: Experiment 1.
Two different models were developed for the RC elements (col-
umns and beam): (1) distributed plasticity model, and (2) lumped
plasticity model (plastic hinges). Equivalent diagonal struts, mod-
elled as truss elements, were used to study the influence of the
effect of the infill panel on the frame response. The stiffness degra-

Fig. 18. Contribution of the frame and the infill in the experimental tests: (a) Experiment 1, and (b) Experiment 2.

Fig. 19. Comparison of the equivalent strut backbone curves proposed by Liberatore and Decanini [59,60], Panagiotakos and Fardis [57] and Dolšek and Fajfar [41]: (a)
Experiment 1, and (b) Experiment 2.
612 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
dation of the equivalent strut in loading and unloading phases was
modelled using the three different models available in OpenSees:
(a) Concrete01 material, (b) Hysteretic material, and (c) Pinching4
material. Fig. 1 illustrates the modelling scheme of the masonry
infilled RC frame. The model properties were calibrated with the
experimental test results.
4.2.1. RC frame modelling: Model 1 and Model 2
The nonlinear behavior of the RC beam and columns in Model 1
was modelled by utilizing the distributed plasticity elements in the
first model. Particularly, both the columns and the beam and rep-
resented by the nonlinearBeamColumn element [27]. The cross-
sections are defined using the fibre discretization with distinct
fibres for steel reinforcement. Each element has five integration
points. Concrete material that includes unconfined (cover) and con-
fined (core) concrete is modelled using the Concrete01 material
model (Fig. 12). The steel reinforcing bars were modelled using a
Steel02 material, which is based on the uniaxial Giuffre-
Menegotto-Pinto steel material, with 1% strain hardening ratio.
Table 7 summarizes the mechanical properties adopted for the
structural materials. The confinement degree of the RC frame
members used is K = 1.3 for both experiments.
The model calibration is based on the comparison between
numerical and experimental results in terms of total base shear-
top displacement curves and cumulative energy dissipation. An
iterative approach was used, which consisted in adjusting the
Fig. 20. Force-displacement curves of masonry infilled RC frames obtained using various strength models: (a) Experiment 1, and (b) Experiment 2.
Table 8
Mechanical properties of the infill assigned to the equivalent strut.
Material Experimental ID
Masonry brick 1 (F1 N)
2 (M2)
a
The value of the parameters Em and s0 are not available – the value of Em was calculated based on 550 fm suggested by FEMA356, and s0 was been evaluated using
empirical equation s0 = 0.7 sm0.
Table 9
Concrete01 material parameters for the equivalent strut.
Experimental ID Dimensionless cross section, w/d
1 0.1868
2 0.2500
model parameters until the simulated hysteretic curves repro-
duced the experiment behavior appropriately.
Fig. 15 shows the comparison between the calibrated model
and the experimental results for both experiments. A good agree-
ment between the two is obtained. The pinching effect is captured
with acceptable accuracy, whereas the reloading stiffness of the
model of the real structure is less accurately predicted. The
strength of the numerical models slightly underestimates the
experimental strength, which the numerical value is 10.3% and
10.5%, slightly lower than the experimental values of Experiment
1 and 2, respectively.
Model 2 is based on the lumped plasticity model. RC beam and
columns are idealized using the elastic elements with zero-length
flexural plastic hinges at both ends, developed by Ibarra et al. [76].
The model requires six parameters as described in Section 3.1.2.
The plastic hinge parameters of the RC beam and columns were
determined using the semi-empirical equations developed by
Haselton et al. [77]. Fig. 16 shows the force-displacement compar-
ison between the experimental results and the numerical predic-
tions of Experiments 1 and 2. As depicted in Fig. 16, the
hysteretic curves obtained by the numerical analyses are in good
agreement with the experiments except for the pinching effect,
which is larger and slightly greater than those recorded in Experi-
ments 1 and 2, respectively. The level of underestimation of the
maximum force value obtained from the numerical model is 6.8%
and 5.2% for Experiments 1 and 2, respectively.
fm [MPa] Em [MPa] sm0 [MPa] s0 [MPa]
4.2 2.31a 0.33 0.231a
2.1 1.84 0.4 0.532
Strength envelope compressive strength-strain law values
rm [MPa] em rR [MPa] eR
1.881 0.00018 0.658 0.0124
1.044 0.00011 0.365 0.00028
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 613

Fig. 21. Force-displacement comparison of Experiment 1 using Concrete01 Material for masonry infill with: (a) Model 1 of RC frame, and (b) Model 2 of RC frame.

Fig. 22. Force-displacement comparison of Experiment 2 using Concrete01 Material for masonry infill with: (a) Model 1 of RC frame, and (b) Model 2 of RC frame.

Fig. 23. Force-displacement comparison of Experiment 1 using Hysteretic Material for masonry infill with: (a) RC frame Model 1, and (b) RC frame Model 2.
614 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
A more detailed understanding of the calibrated models perfor-
mance is attained by comparing the cumulative displacement
energy dissipated during the experimental tests and those evalu-
ated numerically. This comparison was conducted only for Experi-
ment 1, since raw data for Experiment 2 are not available. Fig. 17
shows that both Models 1 and 2 overestimate the dissipated
energy, being the average error equal to 8.2% and 14%, for Models
1 and 2, respectively. It is observed that the two models are cap-
able of capturing adequately the response of the bare frame. There-
fore, both models are used further to simulate the behavior of the
infilled frames.
4.2.2. Infill strut modelling (Concrete 01, Pinching4 and hysteretic
material)
In the numerical simulation of the test specimens, a single diag-
onal strut replaces the masonry infill wall in each diagonal direc-
tion (Fig. 1). No-tension truss elements with a multilinear
response, were used to simulate the behavior of the infill in terms
of strength and stiffness. The stiffness degradation in loading and
unloading phases was implemented using three different material
models (Concrete01 material, Hysteretic material, and Pinching4
material) available in OpenSees, as discussed in Section 3.2.
The initial step in this process was the identification of the con-
tribution of the masonry infill wall to the overall response recorded
in the experimental tests. To this aim, the net response of the infill
was determined by subtracting the response of the bare frame
from the total response of the infilled frame, as shown in Fig. 18
for both test results. Fig. 18 highlights the significant contribution
of the infill which is, however, dependent on the geometric and
mechanical properties of both the frame and the infill.
Next, the experimental force-deformation curves of the
masonry infill were compared with the three different backbone
curve models of the equivalent strut outlined in Section 2.3.1.
The key element of this approach was to verify and select the most
suitable constitutive model to reproduce the infill response. The
comparison is shown in Fig. 19, where the experimental and the
Fig. 24. Force-displacement comparison of Experiment 2 using Hysteretic Material for masonry infill with: (a) RC frame Model 1, and (b) RC frame Model 2.
Table 10
Cyclic deterioration parameters for the Pinching4 material hysteretic model.
Value rDispP rDispN rForceP
Lima et al. [35] 0.5 0.5 0.25
Calibrated_E1 0.8 0.8 0.1
Calibrated_E2 0.8 0.8 0.1
estimated monotonic curves are reported. The backbone curves
by Dolšek and Fajfar [41], and Panagiotakos and Fardis [57] give
higher value of the strength and a slightly greater value of the dis-
placement at the peak compared to the model by Decanini et al.
[56,60,62,63]. The post-peak branch is adequately predicted by
the model by Decanini et al. [56,60,62,63]. On the other hand,
the models by Dolšek and Fajfar [41] and Panagiotakos and Fardis
[57] show a steeper decrease of the strength. With reference to
Experiment 2, all models overestimate the initial elastic stiffness.
Overall, the backbone curve model proposed by Decanini et al.
[56,60,62,63] shows the best agreement with the backbone curve
of both experimental tests. Hence, this model is used in the subse-
quent analyses to outline the backbone curve of the struts. For this
model, the calibrated elastic stiffness K0, degrading stiffness K2,
elastic strength Hmf, and residual strength Hmr are:
K 0 ¼ 3K t0 þ 4K mfc
K 2 ¼ 0:02K mfc
ð37Þ
Hmf ¼ 0:8Hmfc
Hmr ¼ 0:35Hmfc
Finally, the width and the strength of the struts were calculated also
according to different models described in Section 2.3.3. As for the
axial strength of the equivalent diagonal strut, each model takes
into account various failure modes. The strength models by Pana-
giotakos and Fardis [57] and Dolšek and Fajfar [41] were not
included in this evaluation because these models were based on a
single failure mechanism (shear failure). Fig. 20 shows the compar-
ison between the experimental results and numerical results
obtained by estimating the strut width and strength by adopting
the equations by Decanini et al. [56,60], Saneinejad and Hobbs
[70], Paulay and Priestly [68] and Priestley and Calvi [69], and FEMA
306 [67]. It is noted that the models proposed by Paulay and Priestly
[68] and Priestley and Calvi [69] and Decanini et al. [56,60] provide
adequate estimate of the experimental results for both experimen-
tal tests. Concerning Experiment 1, the adoption of FEMA 306 [67]
rForceN uForceP uForceN gKLim gDLim gFLim
0.25 0.05 0.05 0.9 0.5 0.9
0.1 0.01 0.01 0.8 0.1 0.1
0.1 0.01 0.01 0.9 0.1 0.1
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 615

Fig. 25. Force-displacement comparison of Experiment 1 using Pinching4 Material for masonry infill with: (a) RC frame Model 1, and (b) RC frame Model 2.

Fig. 26. Force-displacement comparison of Experiment 2 using Pinching4 Material for masonry infill with: (a) RC frame Model 1, and (b) RC frame Model 2.

Fig. 27. Cumulative displacement dissipated energy, Comparison for Experiment 1: (a) Model 1, and (b) Model 2.
616 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

and Saneinejad and Hobbs [70] models provide a lower resistance.
For Experiment 2, only the model by Saneinejad and Hobbs [70]
noticeably underestimates the strength, whereas the other three
models provide a fairly good prediction of the observed experimen-
tal response. This is because of the model by Saneinejad and Hobbs
[70] used the permissible stress fa (Eq. (32)) instead of using the
compressive strength fm as applied in the other models.
Therefore, in this study the strength model proposed by Paulay
and Priestly [68] and Priestley and Calvi [69] is used combined
with the backbone curve by Liberatore and Decanini [60]. The basic
mechanical properties of the infill wall used in the model are pre-
sented in Table 8.
The model adopted for the diagonal strut was further imple-
mented in OpenSees by using the three different uniaxial material
models in order to reproduce the hysteretic response of the infills.
These material models are capable of capturing the cyclic strength
and stiffness degradation, which are generally observed in experi-
mental tests.
4.2.2.1. Strut model using Concrete01 material result. Table 9 shows
the value of the parameters required for the Concrete01 material. It
is worth mentioning that the actual strain at the peak stress was
decreased by half of its actual value in order to obtain a realistic
value of the elastic stiffness. The numerical results obtained using
the Concrete01 material are shown in Figs. 21 and 22 for Experi-
ments 1 and 2, respectively. Overall, good agreement between
experimental and numerical curves is obtained. The percentage
errors in the peak strength of masonry infilled-frame using the
Concrete01 material for the infill, combined with Models 1 and 2
for the frame members, is about 2.8% and 0.4%, respectively. How-
ever, some differences are noted in the unloading phases, which
are multi-linear in the numerical model. This indicates a limitation
of the Concrete01 material, which does not allow for a smooth
unloading response. Figs. 21(b) and 22(b) clearly shows that the
infill wall model combined with Model 2 gives a better perfor-
mance in terms of hysteretic loops shape, unloading and reloading
stiffness and peak strength.
4.2.2.2. Strut model using hysteretic material result. To define the
backbone curve of the Hysteretic material, three points in tension
and compression are required. However, a very small value (1%
of the compressive stress and strain) is assumed for the tension

Table 11
Geometry of the investigated specimens.

Specimen tpanel (mm) L (mm) H (mm) (bxh)column (mm) (bxh)beam (mm)

SU1 [15] 58.4 1099 629 108  165 165  140
Specimen S [13] 60 1350 900 150  150 100  200
Specimen IS [13] 52 1350 900 150  150 100  200
Specimen 4 [11] 92 2337 1556.5 178  178 152  229
Specimen 5 [11] 92 2337 1556.5 178  178 152  229

Table 12
Infill Frame properties from Experimental Test.

Testing Label Loading Lm (m) Hm (m) Lm/Hm Lateral strength (kN) Failure mechanism
Lateral Vertical
SU1 [15] Cyclic Yes 0.991 0.559 1.26 92.7 Bed joint sliding
Specimen S [13] Cyclic Yes 1.2 0.8 1.5 81.46 Diagonal tension failure
Specimen IS [13] Cyclic Yes 1.2 0.8 1.5 72.92 Sliding shear failure
Specimen 4 [11] Cyclic Yes 2.159 1.422 1.5 162.36 Sliding shear failure
Specimen 5 [11] Cyclic Yes 2.159 1.422 1.5 266.89 Diagonal tension failure

Table 13
Infill frame strength, comparison between model predictions and experimental results.

Reference Measured lateral strength (kN) Predicted lateral strength (kN)

DF SH FEMA 356/306 PPC
SU1 [15] 78.7 72 (0.92) 78 (0.99) 97 (1.23) 101 (1.28)
Specimen S [13] 40 21 (0.53) 16 (0.41) 23 (0.59) 27 (0.69)
Specimen IS [13] 28 43 (1.54) 48 (1.72) 42 (1.49) 47 (1.67)
Specimen 4 [11] 113 121 (1.07) 149 (1.32) 106 (0.93) 150 (1.32)
Specimen 5 [11] 196 139 (0.71) 163 (0.83) 153 (0.78) 172 (0.87)
NRMSE 0.18 0.23 0.17 0.15

() Ratio of predicted to measured lateral strength.

Table 14
Strut model parameters for Concrete01 material.

Parameters Experimental test

SU1 Specimen S Specimen IS Specimen 4 Specimen 5
rm 4.590 1.52356 3.00000 3.05916 3.49480
rR 1.610 0.53325 1.05000 1.07071 1.22318
em 0.000374 0.001153 0.000528 0.000262 0.000229
eR 0.025059 0.077255 0.035423 0.017545 0.015358
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621 617

part in order to reproduce a no-tension response. The parameters
defining the shape of the hysteretic cycles are: b, which controls
the stiffness degradation during unloading, PinchX and PinchY,
which control the pinching effect. These parameters are calibrated
by comparing the numerical analysis to the experimental tests. The
obtained values are equal to: 0.1, 0.9 and 0.1 for b, PinchX and Pin-
chY, respectively. Damage parameters are not considered. The cyc-
lic response is shown in Figs. 23 and 24. Compared to the
Concrete01 material, the Hysteretic material model provides a better
agreement with the experimental response, especially in the
unloading branches. For Experiments 1 and 2, the numerical model
is able to adequately reproduce the actual strength; even though
the adoption of the lumped plasticity (Model 2, Figs. 23
(b) and 24(b)) results in slightly greater strength values than those
obtained with the distributed plasticity model (Model 1, Figs. 23
(a) and 24(a)).
4.2.2.3. Strut model using Pinching4 material result. The Pinching4
material was used due to its capability of modelling nonlinearity
and degradation of stiffness and strength. The multi-linear envel-
ope curve was defined by eight parameters including cracking,
yielding, maximum and residual strength, and corresponding dis-
placement. The strength and stiffness of the equivalent strut in
tension are assumed very small, in order to obtain a no-tension
strut. Damage parameters are assumed to be zero. The pinching
coefficients calibrated with the experimental data are indicated
in Table 10, where the coefficients used by Lima et al. [35] for a
masonry made out of artificial blocks of expanded clay are also
reported.
The numerical results (Figs. 25 and 26) are similar to those
obtained with the Hysteretic material (Figs. 23 and 24). The model
can properly capture the shape of the hysteretic loops. The per-
centage errors in peak strength values using this material com-
bined with Model 1 (distributed plasticity for frame members)
are equal to 4.3% and 6.9% for Experiments 1 and 2, respectively.
Whereas, when adopting Model 2 for the frame (lamped plasticity)
the percentage errors in peak strength values are equal to 1.6% and
2.9% for Experiments 1 and 2, respectively, indicating a satisfactory
agreement between the experimental and the numerical results,
although differences are observed in the reloading phases.

Table 15
Strut model parameters for Pinching4 material.

Parameters Experimental test

SU1 Specimen S Specimen IS Specimen 4 Specimen 5
Positive backbone
ePf1 (kN/mm2) 0.3420 0.121885 0.240000 0.244733 0.279584
ePf2 (kN/mm2) 0.3844 0.137120 0.270000 0.275325 0.314532
ePf3 (kN/mm2) 0.4270 0.152356 0.300000 0.305916 0.349480
ePf4 (kN/mm2) 0.1490 0.053325 0.105000 0.107071 0.122318
ePd1 0.0000056 0.0000125 0.0000123 0.0000097 0.0000086
ePd2 0.0000179 0.0001215 0.0000590 0.0000310 0.0000272
ePd3 0.0000304 0.0002306 0.0001057 0.0000524 0.0000458
ePd4 0.0010170 0.0077255 0.0035423 0.0017545 0.0015358
Negative backbone
eNf1 (kN/mm2) 3.420 1.21885 2.40000 2.44733 2.79584
eNf2 (kN/mm2) 3.844 1.37120 2.70000 2.75325 3.14532
eNf3 (kN/mm2) 4.270 1.52356 3.00000 3.05916 3.49480
eNf4 (kN/mm2) 1.490 0.53325 1.05000 1.07071 1.22318
eNd1 0.000056 0.000125 0.000123 0.000097 0.000086
eNd2 0.000179 0.001215 0.000590 0.000310 0.000272
eNd3 0.000304 0.002306 0.001057 0.000524 0.000458
eNd4 0.010170 0.077255 0.035423 0.017545 0.015358
Pinching
rDispP 0.8 0.8 0.8 0.8 0.8
fForceP 0.1 0.1 0.1 0.1 0.1
uForceP 0.01 0.01 0.01 0.01 0.01
rDispN 0.8 0.8 0.8 0.8 0.8
fForceN 0.1 0.1 0.1 0.1 0.1
uForceN 0.01 0.01 0.01 0.01 0.01
Unloading stiffness degradation
gK1 0.8 0.9 0.8 0.8 0.8
gK2 0.7 0.3 0.7 0.7 0.7
gK3 0.7 0.4 0.7 0.7 0.7
gK4 0.7 0.3 0.7 0.7 0.7
gKLim 0.8 0.9 0.8 0.8 0.8
Reloading stiffness degradation
gD1 0 0 0 0 0
gD2 0 0 0 0 0
gD3 0 0 0 0 0
gD4 0 0 0 0 0
gDLim 0.1 0.1 0.1 0.1 0.1
Strength degradation
gF1 1 1 1 1 1
gF2 0 0 0 0 0
gF3 1 1 1 1 1
gF4 1 1 1 1 1
gFLim 0.1 0.1 0.1 0.1 0.1
Energy degradation (gE) 10 10 10 10 10
Damage type (Cycle/energy) Cycle Cycle Cycle Cycle Cycle
618 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
In general, each of the models succeeds in reproducing the
observed experimental behavior with acceptable precision. How-
ever, the observation of the hysteretic curve shapes highlights
some discrepancy in the reloading phases. As for the unloading
branches, the Hysteretic and Pinching4 materials provide a better
matching to the experimental plots, compared to the Concrete01
material model, which is not able to capture the unloading stiffness
slope. Moreover, the use of the Pinching4 material allows to accu-
rately estimate the maximum strength of the infilled-frames.
Finally, a comparison is made also in terms of dissipated energy
(Fig. 27), for Experiment 1. Concerning Model 1 (distributed plas-
ticity), the actual dissipated energy is underestimated until a dis-
placement level of 12 mm for Hysteretic and Pinching4 material
model, and 18 mm for Concrete01 material models, whereas is
overestimated in the following cycles (Fig. 27(a)). The models with
Concrete01 material, Hysteretic material, and Pinching4 material
reproduce fairly well for the experiment, being the average errors
equal to 24%, 22%, and 21%, respectively. For Model 2 (lumped plas-
ticity), a similar trend is observed. The models with Concrete01
material, Hysteretic material, and Pinching4 material overestimate
the cumulative displacement dissipated energy after a displace-
ment equal to 18 mm, 6 mm and 12 mm, respectively (Fig. 27
(b)) and the average errors are equal to 20.2%, 18.1%, and 17.8%,
respectively.
Fig. 28. Comparison between experimental and numerical results, specimen SU1
(Kyriakides [15]).
Fig. 29. Comparison between experimental and numerical results, experimental data by Kakaletsis and Karayannis [13]: (a) specimen S, and (b) specimen IS.
In summary, the assessment of the above mentioned values and
the comparison among all the force-displacement curves, indicate
that the Pinching4 material is slightly superior to reproduce the
cyclic response compared to Concrete01 and Hysteretic materials.
However, given that the Concrete01 material is the simplest model,
requiring a limited number of parameters, and it still gives reason-
able estimates of the overall response. Therefore, both materials
(Concrete01 and Pinching4) will be used in combination with the
lumped plasticity model for the frame members in the following
analyses.
5. Verification of the selected models
In order to evaluate the accuracy and ability of the calibrated
models, five additional experimental tests are selected and numer-
ically reproduced using the Concrete01 and Pinching4 Materials.
Tables 11 and 12 report the geometrical characteristics, the
mechanical properties, and the predicted failure mechanisms of
the infills.
Table 13 shows a comparison of the predicted infill strength
with the experimental results for each strength model (see Sec-
tion 4.2.2). Based on the normalized root mean square error
(NRMSE) shown in Table 13, no significant difference is detected
between the models. However, the approach used by Paulay and
Priestly [68] and Priestley and Calvi [69] (PPC) provide the best
estimate of the lateral strength (NRMSE = 0.15). The ratio of pre-
dicted to observed lateral strength for this model ranges from
0.69 to 1.67. The model by Saneinejad and Hobbs [70] offers the
least accurate prediction (NRMSE = 0.23), the ratio ranging
between 0.41 and 1.72. Although all the models produced accept-
able average errors in estimating the strength of specimens, the
model proposed by PPC is used to calculate the strength of the
infills. The parameters assigned to the struts were evaluated
according to the procedure described in Section 2.3.3. Tables 14
and 15 provide the parameter values considered for the analyses.
The results, in terms of force-displacements curves, are shown
in Figs. 28–30. The comparison between the results obtained from
the experimental tests by Kyriakides [15] (Fig. 28), Kakaletsis and
Karayannis [13] (Fig. 29), and Mehrabi et al. [11] (Fig. 30), and
those obtained with the numerical analyses indicates that the
adopted model is generally adequate to simulate the response of
the masonry infilled RC frames. Specifically, a satisfactory agree-
ment with the experimental data is obtained for tests, IS [13], Spec-
imens 4 and 5 [11], where the error in estimating the peak strength
ranges from 0.5% to 16%. Higher errors are observed in Fig. 29(a),
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
Fig. 30. Comparison between experimental and numerical results, experimental data by Mehrabi et al. [11]: (a) specimen 4, and (b) specimen 5.
for a specimen made of a masonry defined as ‘‘week” by the
authors, which shows an unexpected resistance compared to the
‘‘strong” masonry (Fig. 29(b)).
As expected, the Pinching4 material model shows a higher effi-
ciency in terms of unloading and reloading stiffness, strength
degradation, force-displacement effect, and pinching effect for all
experimental tests. Although the Concrete01 material model pro-
vides a lower accuracy at each cycle, it is still able to provide a good
prediction of the global response.
The single-strut model provides a good match to the experi-
mental results, and, notwithstanding its simplicity, is a worth tool
to be used in design practice.
6. Conclusions
This paper summarized the procedure adopted for the selection
of a suitable model to be used in the assessment of the global per-
formance of RC frames with URM infill walls subjected to in-plane
lateral loads. A brief review of possible failure mechanisms and dif-
ferent modelling approaches available in the literature is first pre-
sented. The definition of the equivalent strut constitutive laws and
the equations used to estimate the strut properties are also
described in detail.
Based on the different failure mechanisms of the infill wall, an
appropriate and a simple analytical model is identified. Specifi-
cally, the infill wall is modelled using a single strut element in both
the diagonal directions. The backbone curve proposed by Libera-
tore and Decanini [59,60] is adopted as it accurately represents
the force-displacement curves of the experimental tests used in
this study. The selected backbone curve model includes the stiff-
ness and strength degradation during the cyclic loading process.
In addition, based on a comparative evaluation, the strength of
the masonry infill wall is calculated using the Paulay and Priestly
[68] and Priestley and Calvi [69] model.
In the context of macro modelling for infill panel, three different
hysteretic models, available in OpenSees, are investigated. As for
the RC frame members, both the distributed plasticity model and
the lumped plasticity model were considered.
The calibration of the model is conducted by comparing the
results of the numerical analyses with two sets of experimental
tests concerning one-story one-bay infilled frames. The efficiency
of the calibrated numerical model was than verified with five addi-
tional experimental tests. The comparison between numerical and
experimental results are conducted based on the force-
displacement response and, whenever available, on the cumulative
619
dissipated energy. The general conclusions that arise from the
comparisons are as follow:
1. The selected model can adequately reproduce the response of
the masonry infilled RC frames.
2. The distributed plasticity model adopted for the RC members
(Model 1) showed its ability to capture the deterioration of
strength and stiffness. The average error of the cumulative dis-
sipated energy is 8.2% while the average error obtained with the
lumped plasticity model (Model 2) is 14%.
3. The Pinching4 material was found to be the most suitable to
reproduce the cyclic response of infills. However, a good com-
promise between simplicity and accuracy was obtained with
the Concrete01 material. The average error of the cumulative
energy dissipated is, on the average, equal to 20% and 17.8%
for Concrete01 and Pinching4 materials, respectively.
4. The comparison between five additional experimental tests and
the corresponding numerical models showed that the selected
models can appropriately reproduce the response of masonry
infilled RC frames. The application of the simplest model, i.e.
the Concrete01 material can be used for envelope analysis, while
a more complex model, i.e. the Pinching4 material is necessary
to increase the hysteretic behavior accuracy.
In conclusion, the present study successfully selected the best-
simplified macro-model to evaluate the global performance of
masonry infilled frames under in-plane cyclic loads. This model,
calibrated considering one-story one-bay frames, can be adopted
for the assessment of new or existing buildings subjected to seis-
mic loads. However, further studies are needed to include the
interaction between in-plane and out-of-plane forces in the
response of infill walls, especially with reference to collapse
conditions.
Acknowledgements
The first author wish to thank the financial support by Ministry
of Higher Education (MoHE) Malaysia and study leave provided by
Universiti Teknologi MARA (UiTM), Malaysia to undertake this
research work at The University of British Columbia. The second
and third authors gratefully acknowledge the Ministry of the
Instruction, University and Research of Italy (MIUR) for supporting
this research project. This work has been partially carried out
under the program ‘‘Dipartimento di Protezione Civile – Consorzio
RELUIS 2017, Area Tematica 1”.
620 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621

References [34] Zhai C, Kong J, Wang X, Chen Z. Experimental and finite element analytical
investigation of seismic behavior of full-scale masonry infilled RC frames. J
Earthq Eng 2016;20(7):1171–98.
[1] Decanini LD, Liberatore L, Mollaioli F, De Sortis A. Estimation of near-source
[35] Lima C, De Stefano G, Martinelli E. Seismic response of masonry infilled RC
ground motion and seismic behaviour of RC framed structures damaged by the
frames: practice-oriented models and open issues. Earthq Struct 2014;6
1999 Athens earthquake. J Earthq Eng 2005;9(5):609–35.
(4):409–36.
[2] Zhao B, Taucer F, Rossetto T. Field investigation on the performance of building
[36] Lourenço PJBB. Computational strategies for masonry structures PhD
structures during the 12 May 2008 Wenchuan earthquake in China. Eng Struct
Thesis. Portugal: University of Porto; 1996.
2009;31(8):1707–23.
[37] Chiou Y-J, Tzeng J-C, Liou Y-W. Experimental and analytical study of masonry
[3] Decanini LD, Liberatore L, Mollaioli F. Damage potential of the 2009L’Aquila,
infilled frames. J Struct Eng 1999;125(10):1109–17.
Italy, earthquake. J Earthq Tsunami 2012;6(3):1250032.
[38] Liberatore L, Bruno M, AlShawa O, Pasca M, Sorrentino L. Finite-discrete
[4] Campione G, Cavaleri L, Macaluso G, Amato G, Di Trapani F. Evaluation of
element modelling of masonry infill walls subjected to out-of-plane loads. VII
infilled frames: an updated in-plane-stiffness macro-model considering the
European congress on computational methods in applied sciences and
effects of vertical loads. Bull Earthq Eng 2015;13:2265–81.
engineering (ECCOMAS) 2016;3:5219–29.
[5] Martinelli E, Lima C, De Stefano G. A simplified procedure for Nonlinear Static
[39] Attard MM, Nappi A, Tin-Loi F. Modeling fracture in masonry. J Struct Eng
analysis of masonry infilled RC frames. Eng Struct 2015;101:591–608.
2007;133(10):1385–92.
[6] Asteris PG, Antoniou ST, Sophianopoulos DS, Chrysostomou CZ. Mathematical
[40] Stavridis A, Shing PB. Finite-element modeling of nonlinear behavior of
macromodeling of infilled frames: state of the art. J Struct Eng
masonry-infilled RC frames. J Struct Eng 2010;136(3):285–96.
2011;137:1508–17.
[41] Dolšek M, Fajfar P. The effect of masonry infills on the seismic response of a
[7] Asteris PG, Cotsovos DM, Chrysostomou CZ, Mohebkhah A, Al-Chaar GK.
four-storey reinforced concrete frame - a deterministic assessment. Eng Struct
Mathematical micromodeling of infilled frames: state of the art. Eng Struct
2008;30(7):1991–2001.
2013;56:1905–21.
[42] Asteris PG, Cavaleri L, Di Trapani F, Tsaris AK. Numerical modelling of out-of-
[8] Skafida S, Koutas L, Bousias SN. Analytical modeling of masonry infilled RC
plane response of infilled frames: State of the art and future challenges for the
frames and verification with experimental data. J Struct 2014;2014:1–17.
equivalent strut macromodels. Eng Struct 2017;132:110–22.
[9] Pires F. Influence of masonry walls over the behaviour of reinforced concrete
[43] Polyakov SV. On the interaction between masonry filler walls and enclosing
frames under horizontal actions. PhD Thesis. Portugal: National Laboratory in
frame when loaded in the plane of the wall. Transl Earthq Eng 1960;2
Civil Engineering; 1990 [in Portuguses].
(3):36–42.
[10] Pires F, Carvalho EC. The behaviour of infilled reinforced concrete frames
[44] Holmes M. Steel frames with brickwork and concrete infilling. ICE Proc
under horizontal cyclic loading. In: Proceedings of the 10th World Conference
1961;19(4):473–8.
on Earthquake Engineering, vol. 6, p. 3419–3422; 1992.
[45] Amato G, Cavaleri L, Fossetti M, Papia M. Infilled frames: influence of vertical
[11] Mehrabi AB, Shing PB, Schuller MP, Noland JL. Experimental evaluation of
loads on the equivalent diagonal strut model. In: 14th World Conference on
masonry-infilled RC frames. J Struct Eng 1996;122(3):228–37.
Earthquake Engineering (WCEE); 2008.
[12] Kakaletsis DJ, Karayannis CG. Experimental investigation of infilled R/C frames
[46] Amato G, Cavaleri L, Fossetti M, Papia M. An updated model of equivalent
with eccentric openings. Struct Eng Mech 2007;26(3):231–50.
diagonal strut for infill panels. In: Proceedings of Eurocode 2009, vol. 6; 2009.
[13] Kakaletsis DJ, Karayannis CG. Influence of masonry strength and openings on
p. 119–28.
infilled R/C frames under cycling loading. J Earthq Eng 2008;12(2):197–221.
[47] Thiruvengadam V. On the natural frequencies of infilled frames. Earthq Eng
[14] Kakaletsis DJ, Karayannis CG. Experimental investigation of infilled reinforced
Struct Dyn 1985;13(3):401–19.
concrete frames with openings. ACI Struct J 2009;106(2):132–41.
[48] Hamburger RO, Chakradeo AS, Methodology for seismic capacity evaluation of
[15] Kyriakides MA. Seismic retrofit of unreinforced masonry infills in non-ductile
steel-frame buildings with infill unreinforced masonry. In: Proceedings of the
reinforced concrete frames using engineered cementitious composites. PhD
National Earthquake Conference, Memphis, Tennesse, U.S;1993. p. 173–82.
Thesis. Stanford University.
[49] Al-Chaar G. Evaluating strength and stiffness of unreinforced masonry infill
[16] Karayannis CG, Favvata MJ, Kakaletsis DJ. Seismic behaviour of infilled and
structures. Rep. No. ERDC/CERL TR-02-1, U.S. Army Corps of Engineers,
pilotis RC frame structures with beam-column joint degradation effect. Eng
Champaign, IL.
Struct 2011;33(10):2821–31.
[50] Asteris PG. Lateral stiffness of brick masonry infilled plane frames. J Struct Eng
[17] Stylianidis KC. Experimental investigation of masonry infilled RC frames. Open
2003;129(8):1071–9.
Constr Build Technol J 2012;6(1). 194-12.
[51] Decanini LD, Liberatore L, Mollaioli F. Strength and stiffness reduction factors
[18] Sigmund V, Penava D. Experimental study of masonry infilled R/C frames with
for infilled frames with openings. Earthq Eng Eng Vib 2014;13(3):437–54.
opening. In: 15th World Conference on Earthquake Engineering (WCEE); 2012.
[52] El-Dakhakhni WW, Hamid AA, Elgaaly M. Strength and stiffness prediction of
[19] Schmidt T. An approach of modeling masonry infilled frames by the F.E.
masonry infill panels. In: Proceedings of 13th World Conference on
method and a modified equivalent strut method. Darmstadt Concrete, Annu J
Earthquake Engineering (WCEE); 2004.
Concr Concr Struct., Ger Darmstadt Univ 1989;4:184–95.
[53] Crisafulli FJ. Seismic behaviour of reinforced concrete structures with masonry
[20] Decanini LD, Mollaioli F, Mura A Saragoni R. Seismic performance of masonry
infills PhD Thesis. University of Canterbury; 1997.
infilled R/C frames. In: 13th World Conference on Earthquake Engineering;
[54] Crisafulli FJ, Carr AJ. Proposed macro-model for the analysis of infilled frame
2004.
structures. Bull New Zeal Soc Earthq Eng 2007;40(2):69–77.
[21] Mondal G, Jain SK. Lateral stiffness of masonry infilled reinforced concrete (RC)
[55] Sattar S, Liel AB. Seismic performance of nonductile reinforced concrete frames
frames with central opening. Earthq Spectra 2008;24(3):701–23.
with masonry infill walls—I: development of a strut model enhanced by finite
[22] Burton H, Deierlein G. Simulation of seismic collapse in non-ductile reinforced
element models. Earthq Spectra 2016;32(2):795–818.
concrete frame buildings with masonry infills. J Struct Eng 2013;140(8).
[56] Decanini LD, Fantin GE. Modelos simplificados de la mampostería
[23] Tarque N, Candido L, Camata G, Spacone E. Masonry infilled frame structures :
incluida en porticos. Características de rigidez y resistencia lateral en
state-of-the-art review of numerical modelling. Earthquakes Struct 2015;8
astado límite. In: Jornadas Argentinas de Ingenierìa Estructural; 1987.
(1):225–51.
p. 81736 [in Spanish].
[24] Liberatore L, Mollaioli F. Influence of masonry infill modelling on the seismic
[57] Panagiotakos TB, Fardis MN. Seismic response of infilled RC frames structures.
response of reinforced concrete frames masonry infill modelling. Civil-Comp
In: 11th world conference on earthquake engineering; 1996. p. 238. ISBN: 0 08
Proc 2015;108:1–17.
042822 3.
[25] Bârnaure M, Ghita A-M, Stoica DN. Seismic performance of masonry-infilled
[58] Bertoldi SH, Decanini LD, Gavarini C. Telai tamponati soggetti ad azioni
RC frames. Urban Arhit Constr 2016;7(3):229–38.
sismiche, un modello semplificato: confronto sperimentale e numerico. In: Atti
[26] Kaushik HB, Rai DC, Jain SK. Code approaches to seismic design of masonry-
del 6 Convegno Nazionale ANIDIS, vol. 2; 1993. p. 815–24 [in Italian].
infilled reinforced concrete frames: a state-of-the-art review. Earthq Spectra
[59] Liberatore L.Approcci innovativi in termini di energia e di spostamento per la
2006;22(4):961–83.
valutazione della risposta sismica di strutture a più gradi di libertà. PhD
[27] Mazzoni S, McKenna F, Scott MH, Fenves GL. OpenSees command language
Dissertation. University of Rome La Sapienza Rome; 2001 [in Italian].
manual. Pacific Earthquake Engineering Research (PEER) Center; 2006.
[60] Liberatore L, Decanini LD. Effect of infills on the seismic response of high-rise
[28] Crisafulli FJ, Carr AJ, Park R. Analytical modelling of infilled frame structures - a
RC buildings designed as bare according to Eurocode 8. Ing Sismica 2011;28
general review. Bull New Zeal Natl Soc Earthqu Eng 2000;33(1):30–47.
(3):7–23.
[29] Shing PB, Mehrabi AB. Behaviour and analysis of masonry infilled frames. Prog
[61] Cavaleri L, Fossetti M, Papia M. Infilled frames: developments in the
Struct Eng Mater 2002;4(3):320–31.
evaluation of cyclic behaviour under lateral loads. Struct Eng Mech 2005;21
[30] Al-Chaar G, Lamb GE, Issa MA. Effect of openings on structural performance of
(4):469–94.
unreinforced masonry infilled frames. In: large-scale structural testing. ACI
[62] Cavaleri L, Di Trapani F. Cyclic response of masonry infilled RC frames:
Spec. Publ, vol. 211; 2003. p. 247–62.
experimental results and simplified modeling. Soil Dyn Earthq Eng
[31] Stavridis A. Analytical and experimental study of seismic performance of
2014;65:224–42.
reinforced concrete frames infilled with masonry walls. PhD Thesis. University
[63] Stafford Smith B. Methods for predicting the lateral stiffness and strength of
of California, San Diego.
multi-storey infilled frames. Build Sci 1967;2(3):247–57.
[32] Tasnimi AA, Mohebkhah A. Investigation on the behavior of brick-infilled steel
[64] Mainstone RJ. Supplementary note on the stiffness and strengths of infilled
frames with openings, experimental and analytical approaches. Eng Struct
frames. Build. Res. Establishment, London, England; 1974.
2011;33(3):968–80.
[65] Klingner RE, Bertero VV. Earthquake resistance of infilled frames. J Struct Div
[33] Mosalam KM, White RN, Gergely P. Static response of infilled frames using
1978;104(6):973–89.
quasi-static experimentation. J Struct Eng 1997;123(11):1462–9.
 N. Mohammad Noh et al. / Engineering Structures 150 (2017) 599–621
[66] Žarnić R, Gostič S. Masonry infilled frames as an effective structural sub-
assemblage. In: Proceedings of the international workshop on seismic design
methodologies for the next generation of codes; 1997.
[67] FEMA 306. Evaluation of earthquake damaged concrete and masonry wall
buildings: basic procedures manual, Washington DC; 2000.
[68] Paulay T, Priestley MJN. Seismic design of reinforced concrete and masonry
buildings. New York: John Wiley; 1992, ISBN 9780471549154.
[69] Priestley MJN, Calvi GM. Towards a capacity-design assessment procedure for
reinforced concrete frames. Earthq Spectra 1991;7(3):413–37.
[70] Saneinejad A, Hobbs B. Inelastic design of infilled frames. J Struct Eng
1995;121(4):634–50.
[71] Haldar P, Singh Y, Paul DK. Identification of seismic failure modes of URM
infilled RC frame buildings. Eng Fail Anal 2013;33:97–118.
[72] Moretti ML. Seismic design of masonry and reinforced concrete infilled
frames: a comprehensive overview. Am J Eng Appl Sci 2015;8(4):748–66.
621
[73] FEMA 356. Prestandard and commentary for the seismic rehabilitation of
buildings, Washington DC; 2000.
[74] Stafford Smith B, Carter C. A method of analysis for infilled frames. Proc Inst
Civ Eng 1969;44(1):31–48.
[75] Spacone E, Filippou FC, Taucer FF. Fibre beam-column model for non-linear
analysis of R/C frames: Part I. Formulation. Earthq Eng Struct Dyn 1996;25
(7):711–26.
[76] Ibarra LF, Medina RA, Krawinkler H. Hysteretic models that incorporate
strength and stiffness deterioration. Earthq Eng Struct Dyn 2005;34
(12):1489–511.
[77] Haselton CB, Liel AB, Taylor Lange S, Deierlein GG. Beam-column element
model calibrated for predicting flexural response leading to global collapse of
RC frame buildings PEER Report, 2007/03. Berkeley: University of California;
2008.