You are on page 1of 39

This article was downloaded by: [George Mason University]

On: 09 January 2015, At: 02:40

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

European Journal of Environmental and

Civil Engineering
Publication details, including instructions for authors and
subscription information:

Masonry infills and RC frames

interaction: literature overview and
state of the art of macromodeling
a a a
Fabio Di Trapani , Giuseppe Macaluso , Liborio Cavaleri &
Maurizio Papia
Department of Civil, Environmental, Aerospace and Materials
Engineering (DICAM), University of Palermo, 90128 Viale delle
Scienze, Palermo, Italy
Published online: 07 Jan 2015.
Click for updates

To cite this article: Fabio Di Trapani, Giuseppe Macaluso, Liborio Cavaleri & Maurizio Papia
(2015): Masonry infills and RC frames interaction: literature overview and state of the art
of macromodeling approach, European Journal of Environmental and Civil Engineering, DOI:

To link to this article:


Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at
Downloaded by [George Mason University] at 02:40 09 January 2015
European Journal of Environmental and Civil Engineering, 2015

Masonry infills and RC frames interaction: literature overview and

state of the art of macromodeling approach
Fabio Di Trapani*, Giuseppe Macaluso, Liborio Cavaleri and Maurizio Papia

Department of Civil, Environmental, Aerospace and Materials Engineering (DICAM), University

of Palermo, 90128 Viale delle Scienze, Palermo, Italy
(Received 1 July 2014; accepted 5 December 2014)
Downloaded by [George Mason University] at 02:40 09 January 2015

The issue of the influence of masonry infills within RC frames structures has been
widely investigated in the last decades by several researchers. The large interest
addressed to this topic depends on the actual observation that when in presence of
seismic events, the response of framed structures is strongly conditioned by the inter-
action with the infill walls, which however are considered as non-structural elements
and not included in the models. The influence of masonry infills role in structural
response is so much relevant to affect not only the overall strength and the stiffness
but it may also radically change the possible collapse mechanisms of the overall struc-
tural complex under the effect of strong ground motions. Infill panels may have a ben-
eficial effect on the structural response, being able in some cases to supply the lack of
resistance of structures to lateral actions, or an adverse contribution inducing unex-
pected and dangerous non-ductile collapse mechanisms. However, the studies carried
out on this topic have demonstrated that, independently from the beneficial or adverse
contribution of masonry infills on structural response, their presence cannot be
neglected in structural modelling both in design and verification phases. The paper
provides a large literature review regarding the modelling techniques developed in the
last decades, going from refined nonlinear FE micromodel approaches to simplified
equivalent single or multiple strut macromodels including also different technical code
statements. The reliability of these approaches is discussed highlighting advantages
and weakness points. Macromodelling approach is particularly pointed out since it
constitutes the most attractive technique to perform complex nonlinear analyses (static
and dynamic). A state of the art of the main issues regarding equivalent strut identifi-
cation (stiffness, constitutive law and cyclic behaviour) across scientific literature is
provided describing in detail noteworthy aspects of some approaches.
Keywords: infilled frames; masonry; interaction; reinforced concrete; macromodel;

1. Introduction
Reinforced concrete-framed structures, infilled with masonry panels, are widespread and
commonly employed in worldwide building traditions. The need to arrange infill walls
in framed structures naturally arises by the necessity to create a separation between
internal space of buildings and external environment. However, above all in the past,
masonry infills were always thought as secondary elements, and engineers did not intro-
duced them in their models or calculations just because infills did not have to play any
structural function.

*Corresponding author. Email:

© 2015 Taylor & Francis

2 F. Di Trapani et al.

It was only after significant seismic events that the observation of the damage for
these buildings has shown that interaction between masonry infills and frames had a
significant role in overall seismic response and capacity that was quite different case by
The topic of masonry infilled and RC frames interaction has a wide literature and is
studied for 50 years and is today not definitively assessed in all its aspects also because
the role of masonry infills may or may not be beneficial when seismic events occur.
Focusing the attention on the single infilled frame subjected to a lateral action, it is
anyway undeniable that masonry infills contribute with a strong stiffening effect that is
generally associated with a strength increasing and a reduction of lateral displacement
Even if the behaviour of a single infilled frame may be easy to predict, the
contribution given at the scale of an entire building is affected by several uncertainties
Downloaded by [George Mason University] at 02:40 09 January 2015

that cannot be assessed without performing detailed analyses.

Generally, the contribution of infills on the overall capacity of structures is strongly
dependant on regularity of their distribution in plan and over the height. A regular dis-
tribution of infills has a beneficial effect, especially for non-seismic designed buildings,
increasing their global bearing capacity and stiffness under lateral actions. On the other
hand, irregular distributions of panels may be really dangerous being the cause of a
potential anticipation of the collapse when seismic events occur. When infills are not
uniformly distributed in plan and mainly concentrated in some areas, their strong stiffen-
ing capacity may cause a significant shifting of the stiffness centroid. This fact produces
as first effect the modification of the actual dynamical modal properties with respect to
those expected during the design phases. The participating mass ratio is spread out on
higher modes not accounted and potentially dangerous. Moreover, the increasing of the
distance between the centre of mass and the centre of stiffness generates additional
torsional effects during the seismic event.
When irregular distribution of infills occurs over the height of buildings, it generally
produces strong differences of storeys strength and stiffness being the potential cause of
soft storey collapse mechanisms. In this case, the damage is concentrated only in the
storeys where infills are missing with fatal consequences (Figure 1).
Moreover, it should be also noticed that that even if the distribution of infills is reg-
ular in plan and over the height, the increase of stiffness causes higher restoring forces
that should be carried by the infill panels. This fact has a positive contribution to the
earthquake resistance until the force components that the panel transfers to the surround-
ing frame are compatible with the resistances of the RC members and joints. In fact, the
effects caused by the local interaction require the frame elements to have a bearing
capacity that can exceed design values to support the efforts increase transferred by the
infill. Especially in the case of low shear reinforced elements with no seismic detailing,
this may cause the local brittle shear collapses of the columns and even of the nodes
strongly compromising the overall capacity (Figure 2).
Other collapse mechanisms due to the infill–frame interaction occur when infills
present openings adjacent to the column in such a way the panel is shorter than the col-
umn itself. This fact modifies the design length of the column causing an unexpected
increase of local shear demand. In these cases, the columns undergo an anticipated
collapse that depends on the aspect ratio and on the free length of the column. If the
columns is squat shear failure occurs, otherwise the collapse is due a double plastic
hinge mechanism (Figure 3(a) and (b)).
European Journal of Environmental and Civil Engineering 3
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 1. Samples of collapses occurred for irregular distribution of infills over the height.

Despite the presence of masonry infill may produce an undesired behaviour of the
structure under seismic events, in many cases they become fundamental for the capacity
of structures to resist to earthquakes, especially when these are not seismically designed.
This has been demonstrated by the observation and evaluation of post-seismic damage.
An example is reported in Figure 4 for a RC building, where the plaster detachment
showed that the infills contributed to increase the lateral strength and stiffness of the
structure in the lower stories where the shear demand is maximum.
Independently from their positive or negative contribution on seismic response of
structures what is clear is that infill–frame interaction cannot be neglected in structural
models as usually is done in practical engineering. A really wide number of studies
have been provided by researchers from the entire world who proposed modelling strat-
egies which differ for the physical approach followed, complexity, aim and reliability
limits. Also different technical codes treat the question of modelling infill–frame interac-
tion even if their recommendations are not mandatory in many cases neither detailed
and clear rules are provided.
In the further sections, a general overview of the mechanical approaches available in
the literature is provided with a critical discussion on their reliability and practical appli-
cability, involving also different technical codes methods. Subsequently, literature equiv-
alent strut macromodels are deeply discussed and a state of the art of main structural
identification issues involving this structural approach is faced. In this section, some
approaches considered as reference are described in detail.
4 F. Di Trapani et al.
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 2. Effects of local infill–frame interaction: Sample of local shear collapse of column ends
and joints due to the interaction with the infills.

Figure 3. Frame-infill interaction effects: local collapse due to the different infill–column height:
(a) short column double plastic hinge mechanism; (b) short column shear failure.

2. Infilled frames modelling: a general literature overview on mechanical

The variability of the behaviour of the infilled frame systems makes not easy to find
mechanical models being computationally simple and able to capture all the involved
aspects. In the last decades, several researchers provided experimental and analytical
studies proposing modelling strategies to predict and introduce in practical technique the
above-discussed interaction effects. From a general point of view, the infill–frame
interaction has been substantially faced by two main approaches: macromodelling and
European Journal of Environmental and Civil Engineering 5
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 4. Positive contribution of infills to earthquake resistance of a RC framed structure.

micromodelling. The complexity of both these techniques moreover depends on the

typology of investigation that is needed (linear or nonlinear behaviour of the materials,
monotonic or cyclic loading).
The macromodel approach (Figure 5) is based on replacing masonry infills by means
of one (or more) equivalent pin-jointed struts for each infill. This technique is the most
frequently used in practice to perform linear/nonlinear static or dynamic analyses
because of its simplicity and a lower computational effort required. Most of the techni-
cal codes also suggest macromodelling approach for seismic assessment of infilled
framed structures.
Besides the attribution of geometrical dimensions of the cross section, the identifica-
tion of the equivalent diagonal strut requires the assignment of specified mechanical
characteristics for the strut depending on the properties of the actual system. Especially
for masonries constituting infills of existing buildings, the identification of the necessary
information is affected by a large uncertainty since this depend not only on masonry
properties (which are not easy themselves to determine) but also on manufacturing and

Figure 5. Macromodel approach: (a) actual system; (b) equivalent braced system.
6 F. Di Trapani et al.

local arrangement details. In a few words, the macromodel should summarise all this
aspects with a single strut able to account for strength, stiffness and damaging. The
number of information needed for the identification depends in fact strongly on the
assessment approach (e.g. linear or nonlinear analysis) which is necessary to carry out.
For a complete identification of the equivalent strut is necessary to determine at least:
(a) the initial stiffness, (b) the peak strength, (c) the constitutive law shape (monotonic
or cyclic).
The first studies on this topic are due to Holmes (1961), who worked with brick
masonry infilled steel frames. He proposed the empiric rule to replace the panel with an
equivalent diagonal strut, having cross section width w equal to 1/3 of the diagonal
length d. Afterwards, several other researchers proposed more detailed methods mainly
basing the identification of the equivalent width on the ratio between the elastic charac-
teristics of the infill and the surrounding frame (e.g. Dawe & Seah, 1989; Durrani &
Downloaded by [George Mason University] at 02:40 09 January 2015

Luo, 1994; Mainstone, 1971, 1974; Saneinejad & Hobbs, 1995; Stafford Smith, 1966;
Stafford Smith & Carter, 1969). In the more recent past, Papia, Cavaleri, and Fossetti,
(2003) proposed an identification technique that introduces the dependence of the strut
width not only on the stiffness ratios between frame and infill but also on the mechani-
cal elastic properties of the infill along the diagonal direction. The definition of the elas-
tic properties of masonry infill panels along a generic direction was subsequently treated
by Cavaleri, Di Trapani, Macaluso, and Colajanni (2014). Other authors (Amato,
Cavaleri, Fossetti, & Papia, 2008; Amato, Fossetti, Cavaleri, & Papia 2009) introduced
a further dependence of the stiffness of the equivalent strut on the vertical load transmit-
ted by the frame to the infill walls. An alternative method, based on a dynamic
structural identification strategy, was also introduced by Cavaleri and Papia (2003).
Regarding to the definition of the constitutive law for the equivalent strut, the study
by Panagiotakos and Fardis (1996), should be mentioned. In their paper, the authors
determine the yielding force, corresponding to the first cracking of the infill, considering
the tensile strength of the masonry evaluated by diagonal tests. In another study,
Bertoldi, Decanini, and Gavarini (1993) proposed to choose the maximum strength of
the strut depending on the possible failure mechanisms of the masonry panel.
The first experimental and analytical studies on the cyclic behaviour of infilled
frames structures refer to Klingner and Bertero (1978), who investigated on the effect of
cyclic loads by testing portions of multi-storey buildings and also provided a first hys-
teretic macromodel. Later, Doudoumis and Mitsopoulou (1986), introduced a cyclic hys-
teretic model providing for the first time an initial no-stress branch due to shrinkage of
contact zones. Experimental pseudo-dynamic tests on masonry infilled RC frames were
carried out by Mander and Nair (1994) and Mander, Nair, Wojtkowski, and Ma (1993)
and Mehrabi, Shing, Schuler, and Noland (1996) who also provided a cyclic law based
on the results of tested infilled frames specimens Figure 6.
Other hysteretic models were subsequently developed starting from different consider-
ations. Madan, Reinhorn, Mander, and Valles (1997) proposed a hysteretic single strut
model which included strength and stiffness reduction and pinching effect. In a more
recent study, Cavaleri, Fossetti, and Papia (2005) introduced a detailed force–displacement
law accounting for cyclic or monotonic behaviour of an equivalent strut providing a first
experimental calibration of the parameters involved.
After the development of several studies on single strut macromodels, some authors,
recognising the importance to account for the local shear stresses transferred by the
infills to the surrounding frame, introduced the possibility to use a multiple strut
configuration providing two or three struts. Among those studies, Crisafulli (1997),
European Journal of Environmental and Civil Engineering 7

Figure 6. Cyclic tests of infilled frames arranged with masonry panels having different strength
Downloaded by [George Mason University] at 02:40 09 January 2015

(Mehrabi et al., 1996)

investigated the influence of different multiple strut configurations in structural response.

In a further work, Crisafulli and Carr (2007), developed a detailed multi-strut macro-
model including in addition to classical truss elements, governed by axial compressive
laws, a special shear frictional strut to account for the vertical load influence on the
overall strength of the panel (Figure 7(a)).
Chrysostomou, Gergely, and Abel (2002) aimed to obtain the response of infilled
frames taking into account both stiffness and strength degradation of infills. They
proposed to model each infill panel using six compression only diagonal struts
(Figure 7(b)).
El-Dakhakhni, Elgaaly, and Hamid (2003) (Figure 8) proposed a 3 strut model
(having 1 concentric and 2 eccentric struts) in order to provide as much as possible a
realistic distribution of moment and shear forces on the frame elements. Simultaneously
to the development of macromodelling techniques, a substantially different approach has
been instead followed by other researchers who adopted an “exact representation” of
infills (micromodelling) to better reproduce frame–infill interaction. According to this

Figure 7. Multiple strut macromodels: (a) double strut macromodel with shear spring (Crisafulli
& Carr, 2007); (b) six-strut macromodel (Chrysostomou et al., 2002).
8 F. Di Trapani et al.

approach, the infill panel is modelled by means of planar shell finite elements while the
frame may be modelled by shell elements or beam elements. The definition of the inter-
face between the infill and the frame constitutes a quite sensitive question that was trea-
ted in different ways by the authors. Most of them used interface elements able to
reproduce frictional effects and frame–infill detachment in contact regions.
Such typology of approach, which is aimed at providing a more accurate response,
is able to capture well local interaction effects and frame global internal force distribu-
tion. However, the calibration of several sensitive parameters is required to obtain reli-
able results. The first studies referring to this approach are due to Mallick and Severn
(1967), who adopted shell elements to model the infill and beam element for the frame
(Figure 9) in order to evaluate the stiffening effects on one-storey/one-bay frames.
With same target Papia (1988), by means of a refined coupled FEM–BEM approach
(Figure 10) investigated on the modification of the overall stiffness with the variation of
Downloaded by [George Mason University] at 02:40 09 January 2015

the infill properties.

Also in Asteris (2003), the question of the lateral stiffness is pointed out. The atten-
tion is focused on the influence of the openings in masonry panels with the variation of
their extension and collocation (Figure 11).

Figure 8. Three-strut macromodel (El-Dakhakhni et al., 2003).

European Journal of Environmental and Civil Engineering 9

Figure 9. FEM idealisation proposed by Mallick and Severn (1967).

More complex numerical and computationally sophisticated nonlinear micromodels

Downloaded by [George Mason University] at 02:40 09 January 2015

were developed by Mehrabi and Shing (1997), and Shing and Mehrabi (2002), making
use of smeared cracking elements (Lofti & Shing, 1991) and discrete cracking elements
(Lofti & Shing, 1994) in order to capture the shear failure of reinforced concrete ele-
ment and the sliding of the masonry units through dilatant interface models (Figure 12).
Other authors (Giambanco, Fileccia Scimemi, & Spada, 2012; Spada, Giambanco, &
Rizzo, 2009) provided refined solutions in simulating masonry panel mechanics by the
use of a mesoscopic approach.
Based on a similar approach, Koutromanos, Stavridis, Shing, and Willam (2011)
provided a numerical/experimental comparison of infilled frames specimens (Figure 13),
tested with a quasi-static cyclical loading and a full scale frame tested on a shake table.
They demonstrated a good agreement of the developed technique with the experimental
results in terms of prediction of dynamic response and cracking pattern propagation on
the masonry infill and the frame.
FE micromodels have been chosen also by researchers who investigated the issue of
the local shear efforts requested in the nodal regions of frame members due to the inter-
action with infills under lateral loads. Fiore, Netti, and Monaco (2012) investigated the
influence of these forces through an elastic FE model deriving also and equivalent dou-
ble strut model to be used in practical case. Other authors (Cavaleri et al., 2013, 2014;
Doudoumis, 2007) used a mixed technique, adopting frame boundary elements for RC
members and shell elements for infill panels. In the latter work, the investigation was
performed with an nonlinear model and an analytical correlation between shear

Figure 10. FEM-BEM model for the infilled frame (Papia, 1988).
10 F. Di Trapani et al.
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 11. Evaluation of stiffness of infilled frames with openings through FE micromodeling
(Asteris, 2003).

Figure 12. Nonlinear numerical modelling of infilled frames (Mehrabi & Shing, 1997).

Figure 13. Nonlinear numerical modelling of infilled frames (Koutromanos et al., 2011).
European Journal of Environmental and Civil Engineering 11

distribution factors on frames and geometrical/mechanical characteristic of infill–frame

system was established.
The literature overview presented above underlines that the question of the assess-
ment of the seismic behaviour of infilled frames has been treated a long in the past and
continue nowadays to be a current issue presenting several still open questions. The
discussed models going from the most simple to the most complex always have the
common difficulty to be simultaneously accurate, computationally affordable and of
easy identification.
The single strut macromodels represent the easiest way to introduce the presence of
the infill panels in models in practical engineering. They are suitable for analysis of
complex structures and easy to identify. They are moreover able to provide a good
approximation despite their simplicity. Their limit mainly lies in the concentric configu-
ration of the equivalent strut which makes not possible to account for the shear trans-
Downloaded by [George Mason University] at 02:40 09 January 2015

mission in critical sections. Although this question may result less relevant for buildings
having sufficient shear reinforcement or weak infills, it becomes a quite sensitive issue
when non-seismically designed buildings are analysed. The multiple strut configurations
are able to overcome this problem but are affected by uncertainties arising for their
calibration, especially to perform nonlinear static or time history analyses.
Among the above-mentioned models, those including the cyclic behaviour often
depend in fact on a large number of parameters. In the case of multiple strut configura-
tions (double, triple pin-jointed struts or mixed axial and shear struts), the assessment of
a constitutive monotonic or cyclic law is needed for each strut.
Finally, FE micromodels represent the most accurate approach to capture the frame–
infill interaction being also the most similar to the real physic of the problem. Besides
the stiffening effects, micromodels are able to well represent through interface elements
complex issues such as the local frame–infill interaction, the sliding of the units along
mortar joints, the cracking propagation on infills and reinforced concrete elements.
Although this advantage they still present a double difficulty. The first one regards a
proper calibration that may be really difficult to provide, requiring, especially for nonlin-
ear cases, the knowledge of several parameters and a sufficient experience to handle this
kind of modelling. The second, and more relevant, is related to the high computational
effort needed. The application of micromodelling to complex structural system is in fact
nowadays prohibitive for the large time that is required, limiting the use of FE nonlinear
micromodels to simple case study. FE models represent anyway fundamental resource
for research purposes being suitable as reference models to develop simplified tech-
niques and to study problems which require a refined representation.
A summary chart of the infill–frame modelling techniques available in literature and
here discussed is below reported (Figure 14).

3. Infilled frames in technical codes

Also technical codes deal with infill–frame interaction. Eurocode 8 (2004), recommends
to consider infills in models when their contribution may significantly influence lateral
stiffness and strength. To consider possible planar irregularities in plan, it is stated to
include infill walls in models performing a sensitivity analysis regarding their position
and their properties.
Referring to a possible non-uniform distribution of infills over the structural height,
EC8 states that if more accurate models are not used an amplification of seismic action
should be considered by the magnification factor η that is:
12 F. Di Trapani et al.

Figure 14. Resume of numerical approaches for infilled frames modelling.

Downloaded by [George Mason University] at 02:40 09 January 2015

g¼1þ (1)
In Equation (1), ΔWRw is the strength reduction of the current storey with respect to the
upper infilled one while ΔWSd is the sum of the seismic shear forces acting at the top of
considered storey.
Despite EC8 often suggests to use reliable models, the code itself does not provide
any modelling approach as guide for designers in practical applications.
Similarly, Italian technical code (DM (Ministerial Decree), 2008) suggests to amplify
forces in potentially soft stories multiplying by a fixed magnification factor of 1.4. Also
here no modelling criteria for more complex analyses can be found in the code.
Conversely to Eurocode 8 and Italian code, the Federal Emergency Management
Agency (FEMA) code 356, 2000 provides a clearer definition of how to consider infills
in structural models. In particular the code gives the possibility to use a FE model for
the panel or, alternatively, the introduction of a diagonal braced strut having the same
thickness and Young modulus of the masonry infill and an equivalent width w
calculated using the expression provided by Mainstone (1974):
w ¼ 0:175 ðk1 h0 Þ d (2)
h′ being the height of the frame (between the centre-lines of the beams) and d is the
measure of the diagonal dimension (Figure 15). In Equation (2), the parameter λ1 has
the following expression:
Ed t sin 2h 4
k1 ¼ (3)
4 Ef Ic h

in which t is the thickness of the infill, h and l the height and length. The angle θ repre-
sents the slope of the strut and is evaluable as θ = atan(h/l), Ic is the moment of inertia
of the columns and Ed and Ef the Young’s modulus of the infill and of the concrete
frame, respectively.
In FEMA, the question of local interaction effects is also accounted. The code itself
specifies that beams and columns adjacent to infills should have sufficient strength to
support local shear effects arising from the infill–frame interaction in presence of lateral
actions. If more refined models are not used, the FEMA code states that flexural and
shear strength of beams and columns ends should exceed the internal forces evaluated
European Journal of Environmental and Civil Engineering 13

Figure 15. Geometrical features of infill–frame system considered in Equation (2).

Downloaded by [George Mason University] at 02:40 09 January 2015

by the application, at a specified length (Equations 4–5), of the horizontal and vertical
components of the axial force on the equivalent struts (Figure 16).
w h  lceff
lceff ¼ ; tan hc ¼ (4)
cos hc ‘

w h
lbeff ¼ ; tan hb ¼ (5)
sin hb ‘  lbeff

Further, it is required to verify also the shear request associated with a possible double
mechanism originating by the activation of plastic hinges at the ends of the reduced
lengths lceff and lbeff.

4. Major issues in structural identification of infill–frame systems by equivalent

strut models: A state of the art
4.1. The issue of the identification of the equivalent strut width
The first step in macromodelling approach is the assessment of the equivalent strut
width (w) to assign to the cross section in order to effectively reproduce the interaction
between infill and frames first of all regarding to stiffening effects.

Figure 16. Schemes for the evaluation of local effect in FEMA 356 code: (a) on columns; (b)
on beams.
14 F. Di Trapani et al.

Many factors influence the equivalent width to attribute to the strut and it is impor-
tant to underline that they are not simply related to the features of the masonry infill but
more properly on a ratio between infill properties and masonry properties. The stiffening
ratio between infill and frame plays a fundamental role, demonstrated by experimental
tests evidencing the correlation between the infill–frame interface contact length under
lateral loads and the relative stiffness. If the stiffness of the frame increases, the contact
length grows too increasing the stiffness of the system. Many studies therefore focused
their results on the determination of this length.
In 1961, Holmes observed the contact length of masonry infilled steel frames deter-
mining the first empiric rule for the equivalent strut: the replacement of the infill with a
pinned strut having the same thickness of the infill and a width equal to 1/3 of the
diagonal length (Equation (6)).
Downloaded by [George Mason University] at 02:40 09 January 2015

w 1
¼ (6)
d 3
In his work, Holmes proposed also a simple experimentally calibrated procedure to
determine the maximum load and the ultimate displacement of the system.
A first theoretical approach was provided by Stafford Smith (1966). After an experi-
mental investigation on diagonally loaded square-infilled steel frames, he developed the
idea of the strut suggested by Holmes providing an empirical curve for the evaluation
of its dimensions. The experimental and analytical investigations showed a certain anal-
ogy between the frame–infill contact length α and the behaviour of a beam on an elastic
foundation, so the definition of the dimensionless parameter
Ei t
kh0 ¼ h0 4 (7)
4Ef If h

was proposed in order to characterise the column-infill contact length and, consequently,
the stiffness of the system. A scheme of the system considered is reported in Figure 17.
In Equation (7), t and h are the thickness and the height of the infill, respectively; h′
is the height of the frame, measured between the centrelines of the beams; Ei is the
Young modulus of the masonry while Ef and If are the Young moduli of the material

Figure 17. Infill–frame interaction (Stafford Smith, 1966).

European Journal of Environmental and Civil Engineering 15

constituting the frame and the moment of inertia of the cross-sectional area of the frame
elements (beams and columns having the same dimensions).
The curve provided by Stafford Smith was based on experimental evidence and on
the results of several numerical investigations carried out by means of the finite differ-
ence method. It provides the dimensionless parameter w/d for a fixed value of λh′.
Referring to infilled frames subjected to vertical and lateral loads, Stafford Smith
observed an increase in the horizontal stiffness when a vertical load was applied as a
consequence of the increase of the length of contact of the beam on the infill, but no
parameters were inserted in order to account for this phenomenon.
Later, Stafford Smith and Carter (1969), extended the concepts developed before to
the case of rectangular frames, defining the parameter
Ei t sinð2hÞ
kh0 ¼ h0 4
Downloaded by [George Mason University] at 02:40 09 January 2015

4Ef Ic h
θ being the slope of the diagonal and Ic the moment of inertia of the columns. Different
curves w/d-λh′ have been defined with variation of the value of θ. Further, the influence
of the infill stress state along the diagonal direction was considered in the evaluation of
w. A set of curves was derived for different stress levels and fixed ratio l/h. By the com-
parison of these curves with the one provided in the previous study, it is not clear how
the former are related to the latter. Moreover, it should be observed that no analytical
form of the curves mentioned before is provided, so every comparison has to be per-
formed graphically. Further, Klingner and Bertero (1978), confirmed the validity of the
expression proposed by Mainstone (Equation (2)) experimentally working on infills con-
nected to the frame by means of proper reinforcement. Nevertheless, by the comparison
of the curves w/d-λh′ expressed by Equation (2) with the curves provided in
(Figure 18(a)), one concludes that a lower stiffness is obtained despite a stronger
frame–infill connection.
Durrani and Luo (1994), always basing on the results of Mainstone, proposed the
following analytical relationship for the evaluation of the width of the strut:

Figure 18. Comparison between w/d-λh′ curves: (a) Stafford Smith and Carter (1969) and
Klingner and Bertero (1978); (b) Stafford Smith (1966) and Durrani and Luo (1994).
16 F. Di Trapani et al.
w Ei t h04
¼ 0:32sin ð2hÞ
d m E c Ic h
6 I b h0
m ¼ 6 1 þ arctg (10)
p Ic ‘ 0
and Ib is the moment of inertia of the beam cross section. In Figure 18(b), the results
provided by Equation (9) for square-infilled frames with Ib =Ic ¼ 1. are shown and com-
pared with the ones given in Stafford Smith (1966). The two formulations show a good
An alternative approach was proposed by Saneinejad and Hobbs (1995), who con-
sidered the equilibrium of the scheme reported in Figure 19. The width of the equivalent
Downloaded by [George Mason University] at 02:40 09 January 2015

strut is still correlated with the length of the contact zones which depend on the plastic
moments of the RC elements. Analytical results were calibrated on experimental tests
by varying the frame strength. The procedure was supported by a detailed numerical FE
analyses. If results of this approach are compared in terms of initial lateral stiffness with
the ones obtainable by Mainstone (1974) and Stafford Smith (1966) significantly
different values of w/d are obtained.
This partial review of the experimental and analytical investigations shows that the
results obtained by different researchers are strongly influenced by the types of infill
and test, and this conclusion is confirmed by examining and comparing results of other
researches (e.g. Bertero & Brokken, 1983; Madan et al., 1997; Mehrabi & Shing, 1997;
Panagiotakos & Fardis, 1996; Valiasis, Stylianidis, & Penelis, 1993).
Based on this realisation, Papia et al. (2003) developed a procedure for modelling
the behaviour of infilled frames to be adapted to any particular situation that is here
described more in detail.
The identification of the equivalent pin-jointed strut is made by imposing the condi-
tion that the initial stiffness of the actual system (Figure 20(a)) is equal to the initial
stiffness of the equivalent braced frame (Figure 20(b)). It can be assumed that the

Figure 19. Equilibrium of the forced acting on the infilled frame system (Saneinejad & Hobbs,
European Journal of Environmental and Civil Engineering 17

Figure 20. Structural schemes considered in Papia et al. (2003): (a) FEM-BEM Model; (b)
braced frame with equivalent strut.
Downloaded by [George Mason University] at 02:40 09 January 2015

“exact” stiffness of the system in Figure 20(a) can be evaluated by a micromodel

approach, performed by adopting for the infill a FE discretisation while frame elements
are discretised in agreement with the Boundary Element Method. This method allows
an easy and reliable resolution of the contact problem in the regions in which frame and
infill transmit mutual compressive stresses. The shear stress in these regions is assumed
to be governed by the Coulomb friction law.
If the problem is first solved by means of the micromodelling approach and subse-
quently by means of the simplified scheme then, by imposing the equivalence of the
stiffness obtained from the two models, the width of the strut can be evaluated.
Denoting as D  i and Di the stiffness of the two schemes, the stiffness equivalence is
written as:
Di ¼ D (11)
The lateral stiffness of the system in Figure 20(b), equivalent to the scheme in
Figure 20(a), can be determined in the unknown w.
Considering a horizontal force applied to the schemes in Figure 21(b) and (c) pro-
ducing a unitary displacement in the middle of the beam, it can be easily demonstrated
that the equivalent stiffness of the braced system in Figure 21(a) is evaluable sum of the
single stiffness of the schemes in Figure 21(b) (Dd) and (c) (Df ) as:
 1 !
kd cos2 h Ef Ic I b h0
Di ¼ Df þ Dd ¼ þ 24 03 1  1:5 3 0 þ 2 (12)
1 þ kkdc sen2 h þ 14 kkd cos2 h
h Ic l

where the terms kd, kc and kb are expressed as:

Ed t w Ef Ac Ef Ab
kd ¼ ; kc ¼ ; kb ¼ (13)
d h0 ‘0
In Equations (12) and (13), Ed and Ef are, respectively, the Young moduli of the infill
along the diagonal direction and the Young modulus of the frame while Ac and Ab are
 i . is known, the
the cross-sectional areas of beams and columns. Since the stiffness D
equivalence in Equation (11) allows calculating the width w of the equivalent strut by
Equation (12). This operation can be done for any geometrical and mechanical
characteristic assignments of the infill–frame system.
By setting the parameter λ* as
18 F. Di Trapani et al.

Figure 21. Decomposition of later stiffness of macromodel considered in Papia et al. (2003).

Ed th0 h0 2 1 Ac ‘0
Downloaded by [George Mason University] at 02:40 09 January 2015

k ¼ þ (14)
Ef Ac ‘0 2 4 Ab h0
this defines the mechanical degree of coupling of the infill–frame system. The authors
using the described procedure determined numerically a suitable correlation between this
parameter and the dimensionless width w/d of the equivalent strut (Figure 22).
The numerical values of w/d obtained in such way can be fitted by the analytical
w c 1
¼ (15)
d z ðk Þb
In which, c and β depend on the Poisson’s ratio νd of the infill along the diagonal
direction according to the following expressions:
c ¼ 0; 249  0; 0116 md þ 0; 567 md 2
b ¼ 0; 146  0; 0073 md þ 0; 126 md 2
and the term z may take the values

1 if ‘=h ¼ 1
z ¼ (17)
1:125 if ‘=h ¼ 1:5
Equation (15) represents a useful tool for the direct estimation of the equivalent width
strut by simply assigning the geometrical and mechanical properties of the infill–frame
system. The width of the structure depends on the Young modulus Ed and on the
Poisson ration νd both calculated along the diagonal direction. In a further study,
Cavaleri et al. (2014) proposed a strategy for the evaluation of these values starting
from the elastic properties of the masonry panel on its principal direction by means of
the following relationships (Jones, 1998)
1 1 1 2v12 1
¼ ðcos hÞ4 þ  ðsen h cos hÞ2 þ ðsen hÞ4 (18)
Ed E1 G12 E1 E2
v12 1 1 1
md ¼ E d ððsen hÞ4 þ ðcos hÞ4  þ  ðsen h cos hÞ2 (19)
E1 E1 E2 G12
An updated expression of the Equation (15) was provided in where it is introduced the
equivalent strut stiffness also on the level of vertical load transferred from the frame to
the infill. The expression of w/d becomes as follows:
European Journal of Environmental and Civil Engineering 19
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 22. Numerical values of w/d with variation of λ* and fitting curves for aspect ratios
l/h = 1.0 and l/h = 1.5 (Papia et al., 2003).

w c 1
¼j (20)
d z ðk Þb
The coefficient κ depends on the amount of the vertical deformation of columns εv
produced by the total vertical load Fv acting at their top
j ¼ 1 þ ð18k þ 200Þev ; ev (21)
2Ac Ef

4.2. The issue of the definition of the constitutive law

When nonlinear analyses of infilled RC frame structures are needed, the assessment of a
complete constitutive law to attribute to equivalent strut is a fundamental step. Since the
equivalent strut represents an idealisation of a real physical problem (it does not actually
exist), its nonlinear capacity cannot be performed by a direct experimental approach but
derived observing the actual behaviour of an infill–frame system.
Several approaches have been proposed; some of them first provide the determina-
tion of a monotonic constitutive law to be used also as backbone curve for cyclic analy-
ses. Among those the ones here reported are considered in the opinion of the authors
the most suitable and flexible for more complex analyses. Panagiotakos and Fardis
(1996), by means of experimental cyclic tests on scale samples of frames with brick
infill panels defined a simplified tetra-linear relationship (Figure 23). If no residual resis-
tance is assumed, the segments are reduced to 3. The branches describe the initial shear
behaviour of the panel, the equivalent strut formation (after frame- infill detachment),
the softening response after the displacement Sm and a final constant residual resistance
(generally introduced to improve numerical stability).
The determination parameters necessary to define the different branches of the curve
are below described:
20 F. Di Trapani et al.

Figure 23. Force-Displacement relationship for the equivalent strut model proposed in Panagiotakos
and Fardis (1996).
Downloaded by [George Mason University] at 02:40 09 January 2015

 Initial stiffness K1

Gm t‘
K1 ¼ (22)
in which Gm is the tangential elastic modulus of the masonry infill, l, h and t are the
length, the height and the thickness of the panel, respectively;

 Yielding force Fy

Fy ¼ ftp t‘ (23)
in which ftp is the shear strength of the panel;

 Post yielding stiffness K2

Em tw
K2 ¼ (24)
Em being the mean Young modulus of masonry;

 Maximum force Fm assumed as 1.25Fy

 Displacement associate to the maximum force Sm

Fm  Fy
Sm ¼ Sy þ (25)

 Stiffness of the softening branch K3

To be assumed in the range:

0:005K1  K3  0:1K1 (26)
European Journal of Environmental and Civil Engineering 21

 Residual force Fr

To be assumed in the range:

0  Fr  0:1Fy (27)

 Displacement corresponding to the residual force Sr (or Su)

Fm  Fr
Sr ¼ Sm þ (28)
Downloaded by [George Mason University] at 02:40 09 January 2015

This relationship represents a real general and flexible instrument to describe the
nonlinear behaviour of the equivalent strut in order to reproduce one of the frame–infill
systems being this suitable for different ways of calibration. Some authors adopted this
law in their studies proposing different criteria to determine some parameters to have a
better agreement with their experimental results.
The constitutive law proposed by Bertoldi et al. (1993) was obtained by an experi-
mental campaign on 10 different frames having two bays with equal span and a variable
number of storeys (from 2 to 24). Pushover analyses were performed both for bare and
the infilled frames, identifying the equivalent strut by the equation
w K1
¼ þ K2 (29)
d kh

Table 1. Determination of parameters K1 and K2 ([15]).

λh < 3.14 3.14< λh < 7.85 λh > 7.85
K1 1.3 0.707 0.47
K1 −0.178 0.01 0.04

Figure 24. Force-Displacement relationship for the equivalent strut model considered in Bertoldi
et al. (1993).
22 F. Di Trapani et al.

where K1 and K2 values are given in Table 1 and the parameter λh is determined by the
expression of Mainstone. The Force–Displacement relationship of each equivalent strut
is reported in Figure 24.
The main parameters to be determined are Km and Fm, (maximum strength and stiff-
ness of the equivalent strut).
The stiffness Km can be calculated as follows:
Em wt
Km ¼ cos2 h (30)
The evaluation of the maximum strength Fm is performed considering four possible
collapse mechanisms of the panel representative of the mostly recognisable damage
modalities observing experiments: crushing at the centre of the panel, corners crushing,
units sliding and diagonal shear failure. Each of these mechanisms is associated with a
Downloaded by [George Mason University] at 02:40 09 January 2015

specified ultimate stress σw having a uniform distribution over the cross section and
over the length of the equivalent strut. The possible expressions for σw are below

 crushing at the centre of the panel

1:16rm0 tan h
rw1 ¼ (31)
K1 þ K2 kh

 corners crushing

1:12rm0 sin h cos h

rw2 ¼ (32)
K1 ðkhÞ0:12 þ K2 ðkhÞ0:88

 units sliding

ð1:2 sin h þ 0:45 cos hÞu þ 0:3rm0

rw3 ¼ (33)
þ K2

 diagonal shear failure

0:6sm0 þ 0:3r0
rw4 ¼ (34)
þ K2
European Journal of Environmental and Civil Engineering 23

In the above reported equations, rm0 is the normal compressive strength of the masonry
constituting the infill; sm0 is masonry shear strength; u is the sliding resistance of the
bed joints while r0 is the average normal stress acting on the panel. The horizontal
component of the corresponding critical force is associated with the minimum of those
strengths and is evaluable as follows:
Fm ¼ ðrw Þmin tw cos h (35)
This kind of approach is certainly more than others related to the mechanics of the
system considering all main possible failure modalities but a large validation, based on
several infilled frame typologies should be done to assess the actual applicability to
every system.
Other formulations have been proposed to define the constitutive law of the equiva-
lent strut. The two, here proposed, represent a good reference for applications. In partic-
Downloaded by [George Mason University] at 02:40 09 January 2015

ular, the determination of the equivalent strut strength constitutes one of the most
significant issues in defining the constitutive law since it does not depend simply on the
strength shear strength pane but more realistically on the infill–frame coupling and also
on the portion of vertical load carried by the infill. For this reason, a further relationship
for the evaluation of the maximum strength Fmax to attribute to the equivalent strut due
to Žarnic and Gostic (1997), is here given in order to provide a wider overview
Fmax ¼ 0:818 1 þ CI2 þ 1 (36)
CI being

CI ¼ 1:925 (37)
and the other symbols have the same significance of the ones used before.

4.3. The issue of the definition of a cyclic law

In order to extend the use of macromodels to perform nonlinear time history analyses,
several hysteretic models have been developed. The cyclic behaviour of an infilled
frame is not easy to capture, therefore the models that have been proposed from time to
time had to introduce more complex rules to account for the actual behaviour. Also in
this case, the definition of the cyclic law for the strut was determined observing the
results of experimental cyclic tests on the overall infill–frame system in order to perform
an indirect determination. The experimental evidence of several tests has shown that the
cyclic behaviour of an infilled frame has these general characteristics:

 both strength and stiffness degradations are recognised at each cycle depending
on the previous inelastic excursion;
 the cycles show a significant dissipation capacity of the infilled frame system with
respect to the bare ones if a brittle failure does not occur;
 the cycles are affected by pinching at the in proximity to the axes origins due to
the fact that the cracks occurred in the previous loading phase have to be closed
to regain strength and stiffness at the reloading.

One of the first hysteretic models to be used in macromodel approach is due to

Klingner and Bertero (1978). In this model, the infill wall is replaced by two struts
24 F. Di Trapani et al.
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 25. Hysteretic law for the equivalent strut proposed by Klingner and Bertero (1978).

acting in compression only. The initial stiffness is obtained considering the diagonal d
having a cross section of thickness equal to the actual one and width w calculated
according to Mainstone expression. The degradation of stiffness is then related to geo-
metrical and mechanical parameters of the system and is calibrated experimentally. The
envelope curve provides a softening branch beyond the peak strength.
In this model (Figure 25), it can be observed that: (1) the strut has a limited
strength; (2) the unloading stiffness in compression is equal to the elastic loading one;
(3) each reloading branch starts from axes origin and is linear with a slope that depends
on the previous maximum positive or negative displacement reached. The model has
been calibrated and provides good results, for infilled frames in which the panel is con-
nected to the surrounding frame by means of a distributed steel reinforcement. The

Figure 26. Hysteretic law for the equivalent strut proposed by Doudoumis and Mitsopoulou
European Journal of Environmental and Civil Engineering 25

major limitation is due to the fact that the model does not account for the damage
accumulation that produced by the shrinkage of the panel.
Doudoumis and Mitsopoulou (1986), to account for this fact, proposed a cyclic law
in which the strut is inactive in tension and also in compression until reaching a certain
level of deformation (Figure 26). However, the stiffness of the strut is defined through
an envelope curve and has constant loading-unloading slope that is not updated account-
ing for the history of deformation of the strut.
Panagiotakos and Fardis (1996), proposed a lateral force–lateral displacement law
for the strut showing a good agreement with the experimental results obtained on
infilled frames without connectors between frame and infill. In this model (Figure 27),
the initial stiffness depends on the masonry panel (geometry and shear modulus equal to
those of the material that constitutes the infills).
Madan et al. (1997) developed a hysteretic model with parameters calibrated to sim-
Downloaded by [George Mason University] at 02:40 09 January 2015

ulate the stiffness and strength degradation, as well as pinching effect. In the proposed
procedure, the masonry panel is replaced by struts having a strength envelope derived
by the procedure in Saneinejad and Hobbs (1995). The model is integrated by a hyster-
etic Bouc-Wen model. The calibration of the hysteresis parameters is performed on the
basis of experimental results. The decomposition of the model is illustrated in Figure 28.
The resulting model (Figure 28(d)) is composed by the sum of a Bouc-Wen hysteresis
model, a classic strength and stiffness degradation model and a slip lock model to
include the pinching effect.
One of the most complete cyclic models is due to Crisafulli (1997), updated in a
subsequent work by Crisafulli and Carr (2007).
The model is able to simulate the contribution of the infill walls with different levels
of accuracy, as a function of the input data available. The local effects due to the inter-
action between the panel and the frame may also be taken into account in a simplified
way by adopting the approach of two diagonal equivalent struts for each direction
(Figure 29(b)), which can be considered as an intermediate solution between a model
with three struts, more accurate but also more complex (Figure 29(c)), and the model
with a single strut (Figure 29(a)), more simple but also less accurate.

Figure 27. Hysteretic law for the equivalent strut proposed by Panagiotakos and Fardis (1996).
26 F. Di Trapani et al.
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 28. Hysteretic law for the equivalent strut proposed by Madan et al. (1997).

Figure 29. Possible multiple strut configurations considered in Crisafulli (1997): (a) single strut;
(b) double strut; (c) triple strut.

Figure 30. Hysteretic features considered in Crisafulli (1997): (a) normal compressive hysteric
law for masonry; (b) shear sprig; (c) shear spring hysteretic law.

The cyclic behaviour of masonry in compression (Figure 30(a)) is represented by

different hysteresis rules. The most significant characteristic of the model is the intro-
duction of a special shear spring (Figure 30(b)) to account for frictional effects between
mortar bed joints. It is assumed that the shear behaviour of mortar is linear elastic
European Journal of Environmental and Civil Engineering 27

before reaching the maximum shear strength in both loading and unloading phases
(Figure 30(c)). In the elastic branch, the shear stress τ is therefore obtained by the prod-
uct of the sliding deformation γ multiplied by the shear modulus of masonry Gm. The
shear strength depends on the limit bond value τ0, the friction coefficient μ and the com-
pression stress acting perpendicularly to the mortar joints. The values of μ and τ0 have
to be calibrated in order to characterise properly the actual capacity of masonry.
The model of Crisafulli is probably the most complete infill–frame macromodel ever
developed, being able to account directly for different failure mechanisms, frictional
effects and local infill–frame interaction. However, its use requires the calibration of a
wide quantity of parameters that are really difficult to determine in most of the cases.
Basing on original idea developed by Klingner and Bertero, but introducing several
modifications in order to improve the accuracy of loading and unloading branches with
respect to the experimental evidence, Cavaleri et al. (2005) provided a highly detailed
Downloaded by [George Mason University] at 02:40 09 January 2015

parametric cyclic model. In their formulation, the cycles are guided by a strength enve-
lope that corresponds to the ideal monotonic behaviour of the infilled frame. The main
feature of this model which makes it a really flexible instrument regards the possibility
to define a double slope for the first loading branch and for all the unloading branches
(Figure 31). Moreover, the model accounts for the pinching effect during the reloading
caused by the shrinkage of the panel with a zero force branch before regaining strength
and stiffness. The cycle shapes are updated at each cycle since the parameters have
memory of the previous displacement history. The model requires the calibration of 7

Figure 31. Hysteretic model for the equivalent strut in Cavaleri et al. (2005).
28 F. Di Trapani et al.

parameters to be used for the prediction of cyclic behaviour of the infilled frame. It
can be otherwise used for pushover analyses. In this case, the definition of only 3
parameters is needed.
Having as reference the Figure 31, the hysteretic law is divided into 8 branches. The
first three branches (OA), (AB) and (BC) represent the result monotonic loading. The
path OA identifies the linear elastic behaviour, while (OB) is the path experienced after
the first cracking. The path (BC) is a softening branch describing the velocity at which
the post-peak strength decays.
A double unloading slope is described by the braches (DF) and (FG), after a
horizontal no tensile path (MO) is experienced. The reloading occurs in two phases. In
the first phase (OM), due to the fact that the infill has shortened, no load can be carried
until a certain displacement. Beyond this point, the load carrying capacity is regained
through the path OB up to the reunification with the monotonic load pattern.
Downloaded by [George Mason University] at 02:40 09 January 2015

The predictive capacity of the model has been proved by the authors by means of a
comparison with experimental quasi-static cyclic tests carried out on infilled frames
infilled with different typologies of masonry and a first calibration of the parameters
involved have been also provided. Despite the flexibility of the model, that is suitable
to be used in every general case, it depends on several parameters needing a wide
experimental calibration. A simplified version of the above-described model has been
developed by Cavaleri and Di Trapani (2014a). The latter is based on the calibration of
the first 3 parameters α, β and ξ (which are not difficult to predict), to define the mono-
tonic behaviour and uses the simple pivot-model (Dowell, Seible, & Wilson, 1998) rules
for the determination of the hysteretic law.

4.4. The issue of openings in masonry infills: Implications in macromodelling

The presence of door or window openings within masonry infill walls produces a string
influence on lateral strength and stiffness with respect to the case of solid infills. From a
general point of view the overall damage mechanism, activated under lateral loads, un-
dergos a modification as function of the amplitude of the opening and its position within
the infill panel. A conspicuous number of studies concerning the effect of opening on
infilled frame systems is available in the literature, most of them are based on experi-
mental investigation or refined FE analyses. Among the earliest studies Papia (1988), by
means of coupled FEM-BEM approach evaluated the loss of stiffness due to the pres-
ence of centred homothetic openings with respect to the panel contour. He found that
that for a wide range of values, the stiffness have a linear decay when increasing the
amplitude of the opening. Basing on this approximately linear relationship, other authors
developed the idea to define a reduction factor for the equivalent strut width to account
the presence of the openings, maintaining the same modelling concept. Some important
studies in this direction are due to Asteris (2003), who determined, by means of a
refined FE discretisation, a stiffness reduction factor to be used in combination with the
expression provided by Mainstone for the identification of the equivalent strut width. A
similar approach was followed by Papia et al. (2003). The authors proposed the
following analytical expression for the reduction factor r as result of a deep numerical
investigation (Figure 32)
rðnÞ ¼ 1  2:17n4 þ 6:35n3  5:176n2 (38)
European Journal of Environmental and Civil Engineering 29
Downloaded by [George Mason University] at 02:40 09 January 2015

Figure 32. Determination of reduction facto r due to the presence of openings (Papia et al.,
2003): (a) FE Micromodel; (b) r(ξ) relationship.

where in the hypothesis of homothetic openings, the parameter ξ is evaluable as the

ratio between the length (or height) of the opening with respect to the length (or height)
of the panel.
An alternative expression for the reduction factor was proposed by Mondal and Jain
(2008). This time the extension of the opening with respect to the panel was considered in
terms of area ratio. The expression provided for the reduction factor ρw is the following
qw ¼ 1  2:6ac0 (39)
αc0 being the ratio between the area of the opening and the area of the infill. Finally,
the authors conclude their study stating that if the area of the opening is lower than 5%
of the area of the infill, its presence can be neglected, while if the area of opening is
greater than 40%, the presence of the infill can be neglected. An alternative relationship
to be used for steel infilled frames has been also formulated by Tasnimi and
Mohebkhahb (2011).
A wide experimental investigation on the topic has been carried out by Kakaletsis
and Karayannis (2009), who analysed the cyclic behaviour of several RC infilled frames
with openings having different sizes and positions (Figure 33). By means of the experi-
mental results obtained, the authors proposed a parametric model for the monotonic
curve to be used for the attribution of the equivalent strut nonlinear law. In recent study
provided by Decanini, Liberatore, and Mollaioli (2014), the authors considered a large
number (about 150) of literature experimental and numerical tests to derive some
numerical expressions for a reduction factor to be applied for the prediction of both

Figure 33. Experimental investigation on RC infilled frames with openings of different size and
position (Kakaletsis & Karayannis, 2009): (a) cyclic response; (b) cracking pattern.
30 F. Di Trapani et al.

strength and stiffness of the equivalent strut in presence of openings eventually

surrounded by a steel reinforcement.
Several other studies have been carried out on the topic. The majority of the authors
agree with the assumption that the seismic behaviour of the infill is really sensitive to
the opening dimensions and shape. In particular, if the openings are not centred, besides
the strength and stiffness also the collapse mechanisms change. On the other hand,
although many authors suggest the possibility of using a reduction factor for the deter-
mination of the effective equivalent strut width in order to consider the presence of the
openings, others believe that this is not strictly correct since the damage mechanisms
activated are substantially different.

4.5. Results of the use of equivalent strut in 2D or 3D models

Downloaded by [George Mason University] at 02:40 09 January 2015

Many studies regarding the identification of the equivalent strut are limited to the theo-
retical application to single storey – single bay frame. However, during an earthquake,
the interaction between infills and frames involves, however, the building in a more gen-
eral complex of events that depend on the distribution of the infills in plan and eleva-
tion. This means that despite the accuracy used for the identification of the equivalent
strut, the overall structural response is mainly dominated by the mechanism eventually
activated. In the early past, due to the rapid diffusion of nonlinear analysis techniques,
the author had the opportunity to implement their analytical approaches in larger struc-
tural models. The results of some significant studies in this direction are here illustrated.
Dolšek and Fajfar (2008), investigated on the effect of masonry infills on the seismic
response of a four-storey-reinforced concrete frame by pushover analysis (Figure 34).
They compared the responses of bare frames and infilled frame, the latter in the double
hypothesis of solid infills and infills with openings. The infills are modelled as equiva-
lent strut with a specific force displacement law. In the paper, it is clear that the damage
distribution intended localisation of the storey mechanism is quite different form a mod-
elling hypothesis to another. In particular, the presence of the infills tends to move the
damage to the lower stories. A second conclusion that can be drawn is that the presence

Figure 34. Pushover analysis comparison of different infilled frame models (Dolšek & Fajfar,
2008): (a) bare frame; (b) infilled frame with openings; (c) fully infilled frame.
European Journal of Environmental and Civil Engineering 31

of the infills in terms of strength and stiffness is significant up to a certain deformation.

After, the strength and stiffness contribution is rapidly lost. However, this conclusion
cannot be considered true in general but it depends on shear overstrength and ductility
capacity of the frame, as it can be observed in a subsequent study by Uva, Porco, and
Fiore (2012). Here, the authors investigate on the pushover response of the same struc-
tural model varying the strength of the infills. The results show that in the case of
strong infills, in spite of a significant increase of the strength the ultimate displacement
is significantly reduced. Similar conclusions are remarked also in a further paper by
Fiore, Porco, Raffaele, and Uva (2012) where the collapse modalities of two case stud-
ies, representative of low and mid-rise buildings, are investigated. In particular, it is
pointed out that the analysis of infilled frames structures carried out considering simpli-
fied 2D frames may neglect to consider the overall collapse mechanisms leading to
incorrect conclusions. A really comprehensive analysis of the uncertainties arising in the
Downloaded by [George Mason University] at 02:40 09 January 2015

assessment of parameters governing the infilled frames structures has been carried out
by Celarec, Ricci, and Dolšek (2012). Using as reference structural model of infilled
frame (Figure 35), several random variables are consider in the study, in particular the
strength of the infill and the elastic longitudinal and shear moduli. The results are pre-
sented in terms of tornado diagrams confirming that the uncertainty on the characteris-
tics of the masonry infills, in particular the infill strength, has the greatest impact on the
seismic response parameters. Considering the studies here cited and the conclusion of
others available in the literature, one can conclude that when infills are modelled by
means of equivalent struts within structural models, the collapse modality of the entire
structure radically changes with different possibilities that depend of the strength and
stiffness ratio between infill and frame, and the position of the infills. These may
include soft storey mechanisms, shear failure or collapse of columns due to the strong
axial load variation on them caused by the bracing action. For this reason, one can
conclude that modelling and analysis phases have to be handled carefully by designers.

4.6. Recent developments in macromodelling approach: Prediction of local shear

interaction effects and indirect estimation
In this section, two recent techniques of macromodelling are presented. They were
selected for the innovations proposed in their approach to the modelling questions. The
first regards the infill–frame interaction in nodal regions, which, as previously
mentioned, constitutes a quite relevant issue for modelling. The local shear demand, on

Figure 35. Response of an infilled frame due to uncertain parameters (Celarec et al., 2012): (a)
structural model; (b) pushover curves.
32 F. Di Trapani et al.

RC elements ends may exceed their capacity causing an unexpected brittle collapse of
the system. Multiple equivalent strut models can be used to predict the shear demand
although they require a not simple identification process. A recent strategy to circum-
vent the problem has been proposed by Cavaleri and Di Trapani (2014b). Setting a
multi-stage linear equivalence between single strut model and a special FE micromodel,
the authors found a way to correlate at each stage the axial load on the equivalent strut
and the shear demand on beam and column ends. The correlation is possible by means
of the parameter ψ, identifying the infilled frames properties, defined as
w ¼ k n fvom (40)
where k has been defined in Equation (14), n is the ratio between flexural strength of
beams and columns, and fv0m the shear strength of the panel. Once identified, the parameter
ψ allows us to determine the shear distribution coefficients (α) of the frame, representing
Downloaded by [George Mason University] at 02:40 09 January 2015

the ratio between the axial load on the strut and the actual shear on the critical sections
(Figure 36). The analytical ψ-α laws have been derived by means of a deep numerical
investigations. The technique is particularly suitable when the verification of the shear

Figure 36. Shear distribution coefficients versus ψ factor at l/h = 1 and l/h = 2. Numerical
analysis results and interpolating functions (Cavaleri & Di Trapani, 2014b).

Figure 37. Indirect estimation of the equivalent strut by Shing and Stavridis (2014): (a) simpli-
fied Force-Displacement law for and infilled frame; (b) comparison of the simplified curve with
an experimental test.
European Journal of Environmental and Civil Engineering 33

Figure 38. Indirect estimation of the equivalent strut by Shing and Stavridis (2014): (a) fibre
model pushover and simplified curve prediction; (b) comparison of the fibre model response with
an experimental shake table test.
Downloaded by [George Mason University] at 02:40 09 January 2015

capacity of RC members of complex models is needed, giving the possibility to use the
well-known single equivalent strut indirectly determine the actual shear distribution.
As second technique has been presented in a recent paper, Shing and Stavridis
(2014), who proposed an indirect determination of the mechanical properties to assign
to the equivalents strut. They started from the estimation of a ASCE 41-Type pushover
curve for the entire infilled frame (Figure 37(a)) to be predicted by means of simple
rules. The rules for the determination of the simplified law (Figure 34(b)) have been
experimentally derived by the authors. Subsequently, working with a fibre model devel-
oped in Opensees program, a concrete uniaxial law was attributed to the equivalent
strut. The latter was calibrated by a comparison with the pushover curve obtained from
the fibre strut model with the simplified curve predicted for the infilled frame
(Figure 38(a)). This technique allows both to assess monotonic and cyclic behaviour of
the strut by simply operating on the stress–strain law for the strut.
The reliability of the method has been validated comparing the results of the strut
fibre model with experimental shake table tests on full scale buildings (Figure 38(b)).

5. Conclusions
The infill–frame interaction has attracted the interest of researchers for the last 50 years
for its recognised relevance in structural response of RC-framed buildings. A large
quantity of studies and modelling approaches has been proposed by the scientific
community. The paper provided a critical discussion on the ways to model infill–frame
interaction giving firstly a general overview of the main techniques available in
literature. The latter substantially refers to two main methods, micromodelling and mac-
romodelling approach. Micromodelling approach is the closest to the actual physic of
the problem and allows a refined determination of structural stress state on the infill and
on RC members. It should be the preferable way to represent this interaction problem
but however a large computational effort is required for its application. Moreover, the
mechanical parameters needed to identify the FE micromodel are not always determin-
able in an easy way especially for nonlinear analyses. For this reason, this kind of
approach is nowadays not followed in practice but is fundamental as numerical support
to derive simplified models.
On the other hand in macromodelling approach, the substitution of a physical element
(namely a masonry infill) with a phenomenological element (namely a strut) constitutes a
more simple way to identify the infill–frame system requiring less parameters and a
34 F. Di Trapani et al.

significant lower computational effort. Despite these advantages, this approach lacks in
accuracy to reproduce the actual stress state on the frame elements especially those in con-
tact with the infills under lateral loads. This is due to the concentric disposition of the strut.
To circumvent this difficulty, multiple strut macromodels have been developed. They per-
mit the prediction of shear action on beam and columns with a moderate computational
effort but especially when nonlinear static or dynamic analyses are required the correct
calibration of each single strut becomes a relevant question.
For this reason, the final section of the paper has been dedicated to the single strut
macromodelling techniques illustrating a state of the art of the identification issues faced
by different authors. The most relevant issues have been treated summarising that:

 Equivalent strut width depends on the degree of coupling (geometrical and

mechanical) between the infill and the frame;
Downloaded by [George Mason University] at 02:40 09 January 2015

 Vertical loads transferred to the infills influence both stiffness and strength;
 The strength to assign to the strut depends on the panel shear strength but how-
ever is conditioned by the possible collapse mechanisms that can occur;
 A reliable representation of cyclic behaviour has to consider strength and stiffness
degradation and pinching. Moreover, a no-stress path at the load reversals should
be included to account for the panel’s shrinkage.
 The presence of the openings has to be carefully included in models being the
overall strength and stiffness of the system really sensitive to their amplitude and
 Local interaction effects may modify the collapse mechanisms if shear failure
occurs. For this reason, the shear demand has to be predicted and compared with
the capacity of the RC surrounding elements.
 Indirect estimation of the equivalent strut features is possible and represents a
strong alternative for the future developments.

It can be finally concluded that the infill–frame interaction macromodelling is

currently the most sustainable way (from a computational point of view) to represent
infilled frame interaction, allowing a sufficient predictive accuracy of the response in
presence of seismic events. The modelling aspects discussed in the paper significantly
influence the results when this technique is used; for this reason, the identification of
the equivalent strut properties should be treated with the due caution.

Amato, G., Cavaleri, L., Fossetti, M., & Papia, M. (2008). Infilled frames: Influence of vertical
loads on the equivalent diagonal strut model. XIV WCEE. Beijing.
Amato G., Fossetti M., Cavaleri L., & Papia M. (2009). An updated model of equivalent diagonal
strut for infill panels. Workshop: Quali Prospettive per l’Eurocodice 8 alla luce delle
esperienze italiane [Perspectives on Eurocode 8 basing on Italian experieces], Napoli.
Asteris, P. G. (2003). Lateral stiffness of brick masonry infilled plane frames. Journal of
Structural Engineering, 129, 1071–1079.
Bertero, V. V., & Brokken, S. (1983). Infills in seismic resistant building. Journal of Structural
Engineering (ASCE), 109, 1337–1361.
Bertoldi, S. H., Decanini, L. D., & Gavarini, C. (1993). Telai tamponati soggetti ad azioni sismi-
che, un modello semplificato: Confronto sperimentale e numerico [Infilled frames subjected to
seismic actions, a simplified model: Experimental and numerical comparison]. Atti del 6 Con-
vegno Nazionale ANIDIS [Proceedings of the 6th national conference on seismic engineering
ANIDIS] (pp. 815–824). Perugia.
European Journal of Environmental and Civil Engineering 35

Cavaleri, L., & Papia, M. (2003). A new dynamic identification technique: Application to the
evaluation of the equivalent strut for infilled frames. Engineering Structures, 25, 889–901.
Cavaleri, L., & Di Trapani, F. (2014a). Cyclic response of masonry infilled RC frames:
Experimental results and simplified modeling. Soil Dynamics and Earthquake Engineering,
65, 224–242.
Cavaleri, L., & Di Trapani, F. (2014b). Prediction of the additional shear action on frame mem-
bers due to infills. Bullettin of Earthquake Engineering. doi:10.1007/s10518-014-9668-z
Cavaleri, L., Di Trapani, F., Macaluso, G., & Colajanni, P. (2014). Definition of diagonal Poisson’s
ratio and elastic modulus for infill masonry walls. Materials and Structures, 47, 39–262.
Cavaleri L., Di Trapani F., & Papia M. (2013). Analysis of local shear effects in brick masonry
infilled RC frames. ECCOMAS Thematic Conference – COMPDYN 2013: 4th International
Conference on Computational Methods in Structural Dynamics and Earthquake Engineering
(pp. 2962–2977). Kos Island.
Cavaleri, L., Fossetti, M., & Papia, M. (2005). Infilled frames: Developments in the evaluation of
cyclic behaviour under lateral loads. Structural Engineering and Mechanics, 21, 469–494.
Downloaded by [George Mason University] at 02:40 09 January 2015

Celarec, D., Ricci, P., & Dolšek, M. (2012). The sensitivity of seismic response parameters to the
uncertain modelling variables of masonry-infilled reinforced concrete frames. Engineering
Structures, 35, 165–177.
Chrysostomou, C. Z., Gergely, P., & Abel, J. F. (2002). A six-strut model for nonlinear dynamic
analysis of steel infilled frames. International Journal of Structural Stability and Dynamics, 2,
Crisafulli F. J. (1997). Seismic behaviour of reinforced concrete structures with masonry infills
(PhD Thesis). University of Canterbury, Christchurch.
Crisafulli, F. J., & Carr, A. J. (2007). Proposed macro-model for the analysis of infilled frame
structures. Bulletin of the New Zealand Society of Earthquake Engineering, 40, 69–77.
DM (Ministerial Decree). (2008, January 14). Nuove norme tecniche per le costruzioni [New tech-
nical codes for constructions]. Rome: Ministry of Infrastructures and Transportations.
Dawe J. L., & Seah C. K. (1989). Analysis of concrete masonry infilled steel frames subjected to
in-plane loads. Proceedings of the 5th Canadian Masonry Symposium (pp. 329–340).
Vancouver, BC.
Decanini, L. D., Liberatore, L., & Mollaioli, F. (2014). Strength and stiffness reduction factors for in-
filled frames with openings. Earthquake Engineering and Engineering Vibration, 13, 437–454.
Dolšek, M., & Fajfar, P. (2008). The effect of masonry infills on the seismic response of a four-
storey reinforced concrete frame – A deterministic assessment. Engineering Structures, 30,
Doudoumis, I. N. (2007). Finite element modelling and investigation of the behaviour of elastic
infilled frames under monotonic loading. Engineering Structures, 29, 1004–1024.
Doudoumis, I. N., & Mitsopoulou E. N. (1986). Non-linear analysis of multi-storey infilled frames
for unilateral contact condition. Proceedings of 8th European Conference on Earthquake
Engineering (Vol. 3, pp. 63–70). Lisbon.
Dowell, R. K., Seible, F., & Wilson, E. L. (1998). Pivot hysteresis model for reinforced concrete
members. Structural Journal (ACI), 95, 607–617.
Durrani A. J., & Luo Y. H. (1994). Seismic retrofit of flat-slab buildings with masonry infill. Pro-
ceedings of the NCEER Workshop on Seismic Response of Masonry Infills. Report NCEER-
El-Dakhakhni, W., Elgaaly, M., & Hamid, A. (2003). Three-strut model for concrete masonry-in-
filled steel frames. Journal of Structural Engineering (ASCE), 192, 177–185.
Eurocode 8. (2004). Design of structures for earthquake resistance – Part 1: General rules,
seismic actions and rules for buildings. Brussels: CEN.
FEMA 356. (2000). Prestandard and commentary for the seismic rehabilitation of buildings.
Washington, DC: Federal Emergency Management.
Fiore, A., Netti, A., & Monaco, P. (2012). The influence of masonry infill on the seismic behav-
iour of RC frame buildings. Engineering Structures, 44, 133–145.
Fiore, A., Porco, F., Raffaele, D., & Uva, G. (2012). About the influence of the infill panels over
the collapse mechanisms actived under pushover analyses: Two case studies. Soil Dynamics
and Earthquake Engineering, 39, 11–22.
Giambanco, G., Fileccia Scimemi, G., & Spada, A. (2012). The interphase element.
Computational Mechanics, 50, 353–366.
36 F. Di Trapani et al.

Holmes M. (1961). Steel frames with brickwork and concrete infilling. Proceedings of Institution
of Civil Engineers. Paper No. 6501, 473–478. eISSN 1753-7789.
Jones, R. M. (1998). Mechanics of composite materials (2nd ed.). Ann Arbor, MI: Taylor and
Kakaletsis, D. J., & Karayannis, C. G. (2009). Experimental investigation of infilled reinforced
concrete frames with openings. ACI Structural Journal, 102, 132–141.
Klingner, R. E., & Bertero, V. V. (1978). Earthquake resistance of infilled frames. Journal of
Structural Engineering (ASCE), 104, 973–989.
Koutromanos, I., Stavridis, A., Shing, P. B., & Willam, K. (2011). Numerical modeling of
masonry-infilled RC frames subjected to seismic loads. Computers and Structures, 89,
Lofti, H. R., & Shing, P. B. (1991). An appraisal on smeared crack models for masonry shear wall
analysis. Computers and Structures, 41, 413–425.
Lofti, H. R., & Shing, P. B. (1994). Interface model applied to fracture of masonry structures.
Journal of Structural Engineering (ASCE), 120, 63–80.
Downloaded by [George Mason University] at 02:40 09 January 2015

Madan, A., Reinhorn, A. M., Mander, J. B., & Valles, R. E. (1997). Modeling of masonry infill
panels for structural analysis. Journal of Structural Engineering(ASCE), 123, 1295–1302.
Mainstone, R. J. (1971). On the stiffness and strength of infilled frames. Proceedings of Institute
of Civil Engineers, London, 7360S, 57–89.
Mainstone, R. J. (1974). Supplementary note on the stiffness and strength of infilled frames
Current Paper CP 13/74. UK: Building Research Station.
Mallick, D. V., & Severn, R. T. (1967). The behaviour of infilled frames under static loading.
Proceedings of Institute of Civil Engineers, 38, 639–656.
Mander, J. B., & Nair, B. (1994). Seismic resistance of brick infilled steel frames with and with-
out retrofit. The Masonry Journal, 12, 24–37.
Mander, J. B., Nair, B., Wojtkowski, K., & Ma, J. (1993). Experimental study on the seismic per-
formance of brick infilled steel frames with and without retrofit (Tech Rep No. 93-0001).
NCEER: Buffalo, NY.
Mehrabi, A. B., Shing, P. B., Schuler, M. P., & Noland, J. L. (1996). Experimental evaluation of
masonry-infilled RC frames. Journal of Structural Engineering (ASCE), 122, 228–237.
Mehrabi, A. B., & Shing, P. B. (1997). Finite element modelling of masonry-infilled RC frames.
Journal of Structural Engineering, 123, 604–613.
Mondal, G., & Jain, S. K. (2008). Lateral stiffness of masonry infilled reinforced concrete (RC)
frames with central opening. Earthquake Spectra, 24, 701–723.
Panagiotakos, T. B., & Fardis, M. N. (1996). Seismic response of infilled RC frames structures.
XXI WCEE. Acapulco.
Papia, M. (1988). Analysis of infilled frames using a coupled finite element and boundary element
solution scheme. International Journal for Numerical Methods in Engineering, 26, 731–742.
Papia, M., Cavaleri, L., & Fossetti, M. (2003). Infilled frames: Developments in the evaluation of
the stiffening effect of infills. Structural Engineering and Mechanics, 16, 675–693.
Saneinejad, A., & Hobbs, B. (1995). Inelastic design of infilled frames. Journal of Structural
Engineering, 121, 634–650.
Shing, P. B., & Mehrabi, A. B. (2002). Behaviour and analysis of masonry-infilled frames.
Progress in Structural Engineering and Materials, 4, 320–331.
Shing, P. B., & Stavridis, A. (2014). Analysis of seismic response of masonry-infilled RC frames
through collapse. ACI Structural Journal, Special Publication, 297, 7.
Spada, A., Giambanco, G., & Rizzo, P. (2009). Damage and plasticity at the interfaces in compos-
ite materials and structures. Computer Methods in Applied Mechanics and Engineering, 198,
Stafford Smith, B. (1966). Behaviour of the square infilled frames. Structural Division (ASCE),
92, 381–403.
Stafford Smith, B., & Carter, C. (1969). A method of analysis for infilled frames. ICE Proceed-
ings, Paper No. 7218. 31–48.
Tasnimi, A. A., & Mohebkhahb, A. (2011). Investigation on the behavior of brick-infilled steel
frames with openings, experimental and analytical approaches. Engineering Structures, 33,
Uva, G., Porco, F., & Fiore, A. (2012). Appraisal of masonry infill walls effect in the seismic
response of RC framed buildings: A case study. Engineering Structures, 34, 514–526.
European Journal of Environmental and Civil Engineering 37

Valiasis, T. N., Stylianidis, K. C., & Penelis, G. G. (1993). Hysteresis models for weak brick
masonry infills in R/C frames under lateral reversals. European Earthquake Engineering, 1,
Žarnić R., & Gostič S. (1997). Masonry infilled frames as an effective structural subassemblage.
In P. Fajfar & H. Krawinkler (Eds.), Seismic design methodologies for the next generation of
codes (pp. 335–346). Rotterdam: Balkema.
Downloaded by [George Mason University] at 02:40 09 January 2015