Brick and Block Masonry – Trends, Innovations and Challenges – Modena, da Porto & Valluzzi (Eds)

© 2016 Taylor & Francis Group, London, ISBN 978-1-138-02999-6

Experimental investigation of confined masonry infill walls

J.M. Leal G. & J.J. Pérez Gavilán

Instituto de Ingeniería, UNAM, Distrito Federal, México

J.H. Castorena G. & J.I. Velázquez D.

Facultad de Ingeniería, UAS, Sinaloa, México

ABSTRACT:  An experimental investigation of the behavior of confined masonry walls used as infill
walls was conducted. The main variables used were the size of the frame columns, the use of the confin-
ing elements and the use of horizontal reinforcement. Six specimens scaled 1:2 were tested in pairs, each
specimen in the pair having identical characteristics except that in one case the frame columns represented
those of a six floor building, while in the other, columns of a three storey one. The first pair, did not have
any confinement elements, in the second pair tie-columns and tie-beams were added while in the third pair
horizontal reinforcement was included. The specimens were subjected to pseudo-static increasing cycles
of lateral deformation. The effect of the variables is described in terms of strength and displacement
capacity. Sliding was the dominant failure mode of the infill walls.

1  INTRODUCTION obtained from an analysis of the structure without

walls.
Concrete frames Infilled with masonry walls are Analytical models based on the concept of the
designed considering the frame as the main resist- equivalent diagonal strut that considers the struc-
ant element. It is usually considered the contribu- ture as an equivalent braced frame system with a
tion of the wall as a resistance reserve. However, diagonal compression strut replacing the infill wall,
this consideration does not always lead to safer provide an accurate prediction of the global behav-
designs. Passed seismic experiences have left the ior of the system (Madia & Parsekian 2011). In
learning that the infill walls lead beneficial and these analytical models, numerous empirical equa-
adverse effects to the behavior of the structure, tions are employed, through which, researchers
depending on the physical and mechanical charac- have tried to relate the mechanical and geometrical
teristics of its components. Hamburger & Meyer properties of infilled frames with some structural
(2006) and Murty & Jain (2000) noted beneficial parameters, such as stiffness, lateral strength and
effects of frames infilled with masonry walls dur- the contact length between the frame and the infill
ing the 1906 San Francisco earthquake and mod- panel. The structural parameters above mentioned
erate earthquakes in India, respectively. On the are obtained as a function of the dimensionless
other hand, Saatcioglu et al. (2001) noted the poor parameters proposed by Stafford-Smith (1969)
performance of reinforced concrete infilled frames, and Bazan (1980), which express the relative stiff-
the primary structural system used in Turkey, dur- ness between the wall and the frame.
ing the Kocaeli 1999 earthquake (Ravichandra & The behavior of infilled frames subjected to lat-
Klinger 2011). eral loads has been related to the relative stiffness
At global level, infill walls significantly increase of the wall and the frame. If infill panels are too
stiffness and strength of the structure and improve robust, they could induce premature failure of the
the system’s capacity for energy dissipation. At frame. Ravichandra & Klinger (2011) concluded,
the local level, along the contact length, between “If the infill strength ratio (the ratio of story shear
wall and frame, shear and compressive stresses are strength of infills to the story shear strength of
generated. The length of the interface depends on bare frame) is less than about 0.35, the presence
the relative stiffness of the masonry panel and the of infills does not change the failure mechanism,
frame. Normal stresses are highly concentrated at which involves hinging in beams and at column
the corners. This concentration causes a signifi- bases”. However, “When the infill strength ratio
cant increase in the bending moment acting on the reaches about 0.35, the presence of the infill begins
frame (Crisafully 1997). Therefore, the presence of to change the failure mechanism of the frame, from
infill walls transforms the internal force diagrams hinging in beams and at column bases, to story

2319
mechanisms involving column hinging at multiple The edifications had the same configuration
levels of the lower stories”. in plant, however, with the goal to obtain differ-
Then, the relative stiffness between the wall and ent column sizes, they had different height: three
the frame, as much as the reinforcement condi- and six storeys. The buildings had three bays of
tions have an important influence on the behavior 7  meters in each direction and the inter storey
of infilled frames. height was kept constant equal to 3 m. Frames A
In most cases infill walls consist on unreinforced and D (Fig.  1) were infilled with masonry walls
masonry walls, because it is considered that there and, in direction X, concrete walls restrict the hori-
is no need for additional confinement as the frame zontal displacement of the edification, in frames 1
provides it. Another reason for not using this sys- and 4 (Fig. 1).
tem is the added cost of the confining elements The specimens were designed based on the pro-
and the difficulty of constructing the upper tie- totype buildings. Due to space limitation in the
beam that in theory should be in contact with the laboratory, the specimens were scaled 1:2. The
frame’s beam. main criteria for the scaling were to preserve stress.
However, it is well known that unreinforced The percent of longitudinal reinforcements was
masonry walls have very limited capacity for distor- preserved in the model.
tion and their strength degrades rapidly after the For concrete and mortar, maximum size of the
first diagonal cracks due to tension appear. Con- gravel and sand was scaled and their compressive
fined masonry, is an alternative that considerably strength was preserved. Scaled masonry units were
increase the displacement capacity of the walls. In manufactured especially for this investigation. The
addition, tie-columns also provide a considerable materials and manufacturing processes were the
increase in the out of plane stability of the walls. same as those employed for real units.
Confined masonry infill walls allows the use The six specimens were built and tested in pairs.
of horizontal reinforcement as it can be properly Each specimen in the pair had identical charac-
anchored in the tie-columns. teristics except for the size of the frame members,
Similar expressions are provided by the Canadian since in one case they represented the frame ele-
Standard (CSA S304.1–04 2004), the New Zealand ments of the six floor building while in the other
standard (NZS 4230 2004) and the Mexican code they correspond to those of the three storey one.
(NTCM 2004) to determine the contribution of In the first pair of specimens, no confining
horizontal reinforcement to lateral strength. They elements and no horizontal reinforcement were
are used indifferently for shear walls or infill walls. provided (specimens MD6NSR and MD3NSR,
However, Flores (2014) concluded, after testing corresponding to six and three story buildings,
four concrete frames infilled with clay masonry respectively). In the second pair tie-columns
walls with increasing percent of horizontal rein- and tie-beams were added to the infill walls.
forcement, that the lateral strength of the system The dimensions of these concrete elements were
do not increase with the percent of horizontal 65 × 100 mm (MD6N and MD3N). Finally for the
reinforcement.
In this study six specimens, scaled 1:2 were
tested. The variables in study were the size of the
frame columns, the use of the confining elements
and the use of horizontal reinforcement. The
results indicate that all variables have a significant
effect on the capacity of distortion and resistance
of the system.

2  Experimental Program

2.1  Tested models

Two prototype buildings were analyzed and
designed according to Mexican code (NTCC,
2004). The analysis of buildings were made with
static lateral forces varying linearly with height,
using a seismic coefficient equal to 0.40 and a
seismic reduction factor R  =  3 (in Mexican code
Q = 3). The frames were designed so that they were
capable to resist 80% of the total seismic forces,
without contribution of the walls. Figure 1.  Floor plan of the prototype building.

2320
third pair, horizontal reinforcements anchored in
the tie-columns were added to the infill walls, that
consisted in a 4 mm diameter bars every six courses
(MD6NRH and MD3NRH).
Figures 2–3 illustrate the reinforcing details and
dimensions of the specimen’s frame corresponding
to the six and three story buildings, respectively.
The reinforcement in columns and beam consisted
of 12 and 10 mm diameter longitudinal bars with
6 mm diameter stirrups.
In the columns, reinforcement percent ratios
were 2.4% and 3.3%. Plastic hinges could develop
at the end of the frame members so that the sepa-
ration of the stirrups was reduced in these zones.
Infill walls were made of brick units with dimen-
sions 65 × 25 × 125 mm and mortar joints 7 mm
thick. The overall dimensions of the panel cor-
responding to the six story building were 1.35  m
height and 3.25 m long, which represents an aspect
ratio L equal to 0.415. The global dimensions of
the panel corresponding to the six story building
were 1.375  m height and 3.325 long, having an Figure 3.  Reinforcing details and dimensions of frame
aspect ratio equal to 0.414. Reinforcement details corresponding to three story building.
and dimensions of the infill walls are shown in
Figure 4.

2.2  Material properties

Concrete strength and elastic modus were obtained
through standard cylinder tests. Masonry’s prop-
erties including elastic modulus, compressive
strength and diagonal compression strength were
investigated through the corresponding tests using

Figure  4.  Reinforcing details and dimensions of infill

wall corresponding to the six story buildings.

standard Mexican tests (NMX-C-464-ONNCCE

2011).
In Table 1 average values of the results of mate-
rial tests are shown.

2.3  Test setup

Lateral load was applied using two hydraulic actu-
ators as shown in Figures 5 and 7. The hydraulic
actuators were pin-connected to a steel beam that
was bolted to the frame beam so that the lateral
load was transferred uniformly to the frame, simu-
lating a seismic load.
The vertical load was applied directly to the
frame columns as shown in Figure 6. The hydrau-
lic actuators reacted against a steel beam anchored
Figure 2.  Reinforcing details and dimensions of frame to the foundation with tensors on each side of the
corresponding to six story building. wall.

2321
Table 1.  Properties of materials.

Material Mechanical Properties Value MPa

Concrete Compressive strength, 19.7

frame elements
Young’s Modulus, 19823
frame elements
Compressive strength, 15.4
confinement elements
Young’s Modulus, 17397
confinement elements
Masonry Compressive strength 3.4
Diagonal compressive 0.5
strength
Elastic modulus 640.8
Units Compressive strength 9.4
Mortar Compressive strength 15.0

Figure 5.  Arrangement for the application of the lateral

load, plan view.

2.4  Load sequence

Lateral load was applied according to the test pro-
Figure 6.  Setup for the application of the vertical load,
cedure described in appendix A of the NTCM-
lateral view.
2004.
Initially the vertical load was applied and kept
constant through the test. The vertical load was
24.52 × 103 kN and 12.02 × 103 kN on each column
3 tEST RESULSTS
of the specimen corresponding to the six and three
story buildings, respectively.
3.1  Hysteretic curves
The lateral load was applied as follows: the first
two cycles applied a load up to 25% of the esti- The hysteretic curves are presented in Figures 8–13.
mated cracking load, the following two cycles up The first specimen tested was MD6N. For that wall,
to 50% of the estimated cracking load and two once the drift reached 0.02 a new cycle was tried
more up 100% of the cracking load. Subsequently, searching for failure. However, the LVDT used to
the tests were displacement controlled. Drifts with measure the lateral drift exhausted its maximum
increments of 0.002 were applied doing two cycles distance, reaching a drift of 0.0335. Then specimen
in each increment. was unloaded (lateral load only) and the device
When the specimen reached a distortion of 0.02 was changed for a new one with longer capacity.
the distortion increment changed to 0.004. This The test restarted with the negative branch up to
change is not specified in the NTCM code, as it 0.0335 to preserve the symmetry of the test. Then
was designed mainly for confined masonry walls it was decided that the test should continue nor-
not for infill walls. mally with drift increments of 0.004.

2322
 Figure 9.  Hysteric curve of the specimen MD3NSR.

Figure 10.  Hysteric curve of the specimen MD6N.

Figure 7.  Experimental setup, lateral view.

Figure 11.  Hysteric curve of the specimen MD3N.

Figure 8.  Hysteric curve of the specimen MD6NSR.

In the case of specimen MD6NRH, there was

an oil leak in one of the lateral load actuators.
That explains the cycle anomalies about a drift of
0.025. The problem was fixed and the test contin- Figure 12.  Hysteric curve of the specimen MD6NRH.
ued normally.
A remarkable stability of the cycles and a slow
strength degradation was observed in all cases. At earlier stage of the tests, separation between
Pinching of the unloading branch is also present the panel and frame was observed, as has been
although is more pronounced for the specimens reported previously. As the tests were progressing,
with smaller sized columns. The loading paths the separation between these elements increased
intersected the vertical axis, approximately, on the considerably. The maximum separation between
same level of lateral load. column and wall was approximately 10 mm.

2323
 Because the foundation and the beam to transfer compromised, but because no out of plane forces
the lateral load were very stiff, gaps between foun- were applied there is no evidence of this in the hys-
dation and panel and panel and specimen’s beam teretic curves.
were very thin, difficult to observe. Envelopes for specimens corresponding to six and
The specimens without reinforcing elements three storey buildings are shown in Figures 15–16
developed well-defined first inclined cracks. Later, respectively. It is observed that the horizontal
inclined cracks and horizontal sliding defined the reinforcement had a greater contribution to maxi-
failure mode. The specimens with tie-beams and mum load in specimens with a greater wall/frame
tie-columns developed cracking on the panel more stiffness ratio. In both figures, it is observed that
distributed than the specimens without confining the confinement elements do not increase signifi-
elements. The specimens with horizontal reinforce- cantly the lateral strength of the system; however,
ment also had more distributed cracking. In this the maximum load is reached at higher drift than
pair of specimens was difficult to identify the first those without confining elements, maintaining the
crack. integrity of the wall.
In the specimens with confining members, the In Figure 17, the comparison of envelope curves
shear strength of the central tie-column had an between specimens with different wall to frame
important role. In these specimens, the ultimate stiffness ratios are shown. It is observed that this
lateral load strength was reached when the central
tie-column failed when a crack crossed it.
In all specimens, combination of sliding of the
infill wall and diagonal tension in the connection
of beam-column of the frame was the dominant
failure mode of the system. The final crack pattern
of specimen MD3N is shown in Figure 14.

3.2  Envelope curves

The loading-drift curves presented similar proper-
ties and they can be characterized in four stages.
At low levels of lateral displacements, masonry
and frame acted as a monolithic composite struc- Figure 14.  Crack pattern in specimen MD3N.
tural system. When separation between wall and
frame occurred, the wall acted as a diagonal strut
and, due to its high stiffness, the masonry infill
stiffened the flexible frame and also increased its
initial strength. In most specimens, inclined crack
appear first followed later by horizontal cracks, the
lateral stiffness of the system reduced gradually
until the maximum shear strength was attained.
The last stage corresponded to descending branch
produced due the softening of the material, char-
acterized by gradual crushing of the masonry.
It is very important to note that although the in
Figure  15.  Envelope curves of specimen corresponding
plane strength of the wall degrade slowly, it was to six story building.
evident that the lateral stability of the walls were

Figure 16.  Envelope curves of specimen corresponding

Figure 13.  Hysteric curve of the specimen MD3NRH. to three story building.

2324
parameter had relevant influence in resistance to Confining elements do not increase the lateral
lateral load, since the specimen corresponding to load strength; however, maximum load is reached
six story building had a lateral load strength 2.1 at higher drift than those without confining ele-
times greater than the specimen corresponding to ments, maintaining the integrity of the wall for a
three story building. In specimens with a smaller larger lateral displacement.
wall to frame stiffness ratio, the length of contact
between wall and frame was larger, therefore the
contact forces were better distributed. Acknowledgements
The Table  2  summarizes the values of initial
stiffness (Ki), maximum resistance to lateral load This investigation was partially funded with by
(Vmax) and drift at maximum lateral load (∆max) CONACYT that provided the Phd scholarship
developed for different specimens. and additional funds provided through the gradu-
ate program of UNAM. The tests were carried out
in the Structures Lab of the Autonomous Univer-
4  CONCLUSIONS sity of Sinaloa (UAS).

The contribution of horizontal reinforcement on REFERENCES

lateral strength of the system depends of the wall/
frame stiffness ratio. Horizontal reinforcement Bazán, Z.T.E. 1980. Muros de Mampostería ante Cargas
contributed to increase the deformation capacity Laterales. PhD Thesis. National Autonomous Univer-
of the system and the influence of the reinforce- sity of Mexico.
ment is independent of the wall to frame stiffness CSA Standards 2004: CSA S304.1-04, Design of masonry
ratio. structures, 2004.
The wall to frame stiffness ratio has an impor- Crisafulli, F.J. 1997. Seismic Behavior of Reinforced
tant effect on the shear strength of the system. In Concrete Structures with Masonry Infills. PhD Tesis,
specimens with a smaller wall to frame stiffness University of Carterbury. New Zealand.
Flores, L.E. 2014. Ensaye de Muros Diafragma de
ratio, the length of contact between wall and frame Mampostería con Diferente Cuantía de Refuerzo
was larger, therefore the contact forces were better Horizontal. XIX Congreso Nacional de Ingeniería
distributed. Estructural. Puerto Vallarta. Jalisco. México.
Hamburger, R.O & Meyer J.D. 2006. The Performance of
Steel Frames building with infill masonry walls in the
1906 San Francisco Earthquake, Earthquake Spectra.
Vol. 22. No S2. S43–S67.
Madia, F.C. & Parsekian, G.A. 2011. Modeling a Rein-
forced Concrete Building Frame with Infill walls, 11th
North American Masonry Conference, Minneapolis,
U.S.A.
Murty, C.V.R. & Jain, S.K. 2000. Beneficial influence
of masonry infill walls on seismic performance of
RC frame buildings. Proceedings of the 12th World
Conference on Earthquake Engineering. Auckland,
New Zealand.
NMX-C-464-ONNCCE 2011. Determinación de la
resistencia a compresión y módulo de elasticidad de
Figure 17.  Envelope curves with different stiffness ratio.
pilas de mampostería de arcilla o de concreto. Organ-
ismo Nacional de Normalización y Certificación de la
Construcción y Edificación, México, D.F.
Table  2.  Summary of main parameters of envelope NTCC 2004. Normas Técnicas Complementarias para el
curves. Diseño y Construcción de Estructuras de Concreto.
Gaceta oficial del Distrito Federal, I (103 BIS): 88–194.
Vmax (kN) Δmax (mm/mm) México D.F.
NTCM 2004. Normas Técnicas Complementarias para el
Specimen Ki (kN/mm) (+) (-) (+) (-) Diseño y Construcción de Estructuras de Mampost-
ería. Gaceta oficial del Distrito Federal. I (103 BIS):
MD6NSR 117.75 267.7 275.7 0.0242 0.0201 4–53, México, D.F.
MD6N 123.88 311.4 297.6 0.0335 0.0331 NZS Standard 2004: NZS 4230:2004, Design of rein-
MD6NRH 115.14 313.7 302.7 0.0240 0.0240 forced concrete masonry structures, 2004.
MD3NSR 19.68 162.6 158.2 0.0100 0.0100 Ravichandra, S.S. & Klinger, R.E. 2011. Seismic Design
MD3N 103.06 148.2 147.4 0.0241 0.0243 Factors for Steel Moment Frames with Masonry
MD3NRH 103.73 194.0 181.9 0.0241 0.0241 Infills: Part 1 and 2. 11th North American Masonry
Conference. Minneapolis, U.S.A.

2325
Saatcioglu, M., Mitchell, D., Tinawi, R., Garnerd, N., Stafford-Smith, B. 1962. Lateral stiffness of infilled
Gillies, J., Ghobarh, A., Anderson D., L. & Lau D. 2001, frames. ASCE Journal of Structural Division, Vol. 88,
The August 17, 1999, Kocaeli (Turkey) earthquake- No. ST6, pp. 183–199.
damage to structures, Canadian Journal of Civil
Engineering, Vol 28, pp. 715–737.

2326