Construction and Building Materials 164 (2018) 70–83

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Experimental study of the in-plane cyclic behaviour of masonry walls

strengthened by composite materials
Reboul Nadège a,⇑, Mesticou Zyed b, Si Larbi Amir b, Ferrier Emmanuel a
a
Université Claude Bernard Lyon 1, LMC2, IUT Lyon 1, 82 boulevard Niels Bohr, 69100 Villeurbanne, France
b
Univ Lyon, Ecole Nationale d’Ingénieurs de Saint-Etienne (ENISE), Laboratoire de Tribologie et de Dynamique des Systèmes (LTDS), UMR 5513, 58 rue Jean Parot, 42000
Saint-Etienne, France

h i g h l i g h t s

Experimental characterization of masonry walls under cyclic in-plane solicitations.

Analyses of the damage and failure modes of the different reinforced walls.

Comparative study based on performance parameters.

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

Article history: This study presents the results of an experimental campaign on masonry walls to characterize their beha-
Received 8 February 2017 viour under in-plane solicitations. A series of five walls has been built, including one reference wall and
Received in revised form 16 December 2017 four walls reinforced with externally bonded composites that are either FRP-composites (fibre reinforced
Accepted 27 December 2017
polymer) or TRC-composites (textile reinforced mortar). In the present work, strengthening materials
were not applied over the entire masonry surface: walls were reinforced only with vertical composite
sheets, spaced to reduce the amount of composite materials used for the reinforcement. Each wall was
Keywords:
tested under a vertical compression load and cyclic reversed horizontal displacements until failure.
Masonry
Strengthening
This article presents the results in terms of load displacement curves, analyses the damage and failure
FRP modes of the different walls and provides a comparative study based on performance parameters, such
TRC as load capacity, ductility, stiffness degradation, energy dissipation capacity to highlight advantages and
Cyclic tests limitations of the different strengthening materials.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Strengthening of masonry structures requires the attention of a
growing community of researchers because more stringent seismic
design requirements have been recently adopted in many coun-
tries. Numerous (also recent) earthquakes have highlighted the
vulnerability of masonry structures that show substantial damage
(if not failure) when they are subjected to seismic actions. The fail-
ure modes of unreinforced masonry walls subjected to in-plane
loading are diagonal cracking, sliding and toe crushing.
One of the methods most commonly used over the past several
years to upgrade masonry structures is the application of exter-
nally bonded composite materials. The use of Fibre Reinforced
⇑ Corresponding author.
E-mail addresses: nadege.reboul@univ-lyon1.fr (N. Reboul), zyed.mesticou@
enise.fr (Z. Mesticou), amir.si-larbi@enise.fr (A. Si Larbi), emmanuel.ferrier@
univ-lyon1.fr (E. Ferrier).
Polymers (FRPs) has now become popular [1,2] due to their light
weight and their high level of mechanical performance (which also
implies significant costs). However, the potential of FRPs is usually
far from being fully exploited when they are applied on masonry
materials with very low tensile and shear strength. Furthermore,
despite the benefits they provide, FRP materials suffer from the
undesirable effects of organic binders (epoxy resins): poor beha-
viour at high temperatures (their maximum service temperature
must be lower than the glass transition temperature of the resin),
difficulties with application under wet conditions (in the absence
of any special formulation, resin should not be applied to wet
surfaces), strong environmental impact (non-removability, for
instance) and potential health risks for persons performing their
application. To this list of limits might also be added their compat-
ibility problems with some substrates and their lack of water
vapour permeability, which is likely to generate condensation
problems.

https://doi.org/10.1016/j.conbuildmat.2017.12.215
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83
An alternative solution involves replacing the epoxy resins by
inorganic matrices and focussing on Textile Reinforced Concrete
(TRC). Indeed, these composite materials that resemble a fine-
grained cementitious matrix with a fabric reinforcement have
recently been developed [3,4,5]. Referred to as either Fibre Rein-
forced Cementitious Matrix (FRCM) or Textile Reinforced Mortar
or Concrete/Cement (TRM or TRC), these materials overcome most
of the limitations of FRPs (they are particularly compatible with
masonry substrates, easy to apply – even under wet conditions,
more eco-friendly,...).
Previous studies have concluded that the use of these materials
can be an effective method to strengthen masonry walls subjected
to in-plane actions. Diagonal compression tests have been carried
out on tuff masonry panels strengthened over their whole surface
and on both sides by a glass-TRC [6,7] or by a carbon-TRC [8]. In all
cases, the reinforced panels exhibit a better shear strength and a
better ductility: for symmetrically strengthened walls with one
fabric layer, the average shear strength is amplified by a factor
equal to 1.75 [6] (from 0.24 MPa for the unreinforced walls to
0.42 MPa for the strengthened walls) or 1.94 [7] (from 1.06 MPa
to 2.06 MPa) with glass fabrics and 5.45 [8] (from 0.055 MPa to
0.3 MPa) with carbon fabrics. The latter author notes that TRC
strengthening methods provide similar increments in shear capac-
ity if normalized according to the calibrated reinforcement ratio.
The study of Papanicolaou et al. [9] has an innovative nature for
two main reasons. First, they adopted the shear-compression test
to investigate the in-plane response of masonry walls. This test is
more complex than the diagonal compression test, but the result-
ing stress state is more similar to the state created by a seismic
action. Secondly, the test is conducted under cyclic reversed
quasi-static loading conditions to simulate earthquake-type load-
ing. In this study, they have analysed the behaviour of clay brick
walls strengthened on both faces by a TRC made up of carbon fibre
Fig. 1. Unreinforced masonry wall and strengthening schemes.
71
rovings connected by a polypropylene grid. When composite mate-
rials are applied only on one side, the asymmetry leads to signifi-
cant out-of-plane displacements. To analyze separately the
effects of in-plane and out-of-plane solicitations, it is standard
experimental practice in adopting symmetric strengthening con-
figurations even if this does not correspond to field applications.
In this study, a cement-lime based matrix was used to help to reg-
ulate the internal humidity of the clay brick masonry. For size
effects, authors select enough bricks to get a mechanical response
of a full size specimen. TRC-reinforced walls appear to be at least
65–70% as resistant as FRP-reinforced walls with identical fibre
configurations, and they are more ductile; the gain in deformability
can be up to 30%. Koutas et al. [10] explored the behaviour of
masonry structures with larger dimensions: nearly full-scale
three-storey frames, made up of perforated fired clay bricks and
retrofitted with carbon TRC over the entire surface of the infills,
are loaded until failure under cyclic loading conditions. Again,
the TRC-retrofitting method leads to an enhanced global response
of the masonry infills both in terms of lateral strength and defor-
mation capacity.
In all of these previous studies, TRCs are demonstrated to be
effective in strengthening masonry structures when they are
applied on both sides over the entire masonry surface, for all the
tested substrates, for all the tested in-plane loading conditions
and for all the tested textile fabrics. The aim of the present study
is to reduce the amount of composite materials used in the
strengthening scheme (making it more realistic) and to experimen-
tally assess the effectiveness of these «optimized» (meaning more
materials efficient) strengthening configurations. As the dimen-
sions of composite reinforcements decrease, their efficiency is
more conditioned by a correct connection with the masonry.
Therefore, the addition of anchorages was performed in the present
study and their contribution will be discussed.
72 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83
Table 1
Characteristics of the composite materials.
Composite CFRP GFRP
1-fold 2 folds
Real thickness [mm] – 1.5 ± 0.1
Mass per unit area [g/cm2] – 600
Design thickness [mm] 0.48 0.23
Ultimate tensile load nu [kN/cm] 8.16 1.75 ± 0.3
Ultimate strain eu [lstrain] 16,000 14,000 ± 1000
Real ultimate stress ru [MPa] – 115.8 ± 28.3
Net ultimate stress [MPa] – 755.2 ± 184.6
Given by the manufacturer Direct tensile tests on 3
specimens (27.5 mm wide)
The results concerning the quasi-static in-plane response of
hollow concrete block masonry walls, strengthened only by verti-
cal TRC sheets, are presented in [11]. The present study is devoted
to the experimental investigation of the in-plane response of
masonry walls that are partially reinforced by composite materials.
A comparison will be made between TRC effectiveness and FRP
effectiveness. Five masonry walls, four of which were reinforced,
were subjected to shear-compression tests under reversed cyclic
quasi-static loads. The analysis of the results should lead to a better
understanding of the effects of partial TRC or FRP retrofitting on
the behaviour of masonry walls.
2. Experimental program
2.1. Masonry walls and material properties
Five masonry walls were built with hollow concrete blocks. This
type of masonry units have been chosen because they are com-
monly used in France to build the loadbearing structure of individ-
ual houses. The walls are considered to be slender walls for which
vertical strengthening is the most efficient way to strengthen the
walls. Originally measuring 500  200  75 mm3, these blocks
were cut in half lengthwise to have walls with dimensions suitable
for the laboratory tests while maintaining a representative number
of vertical joints. The mortar mix proportions used in this study
were cement (CEM I):sand 1:3 with a water-cement ratio of 0.5.
Each masonry wall was laid in a running bond pattern with mortar
joints approximately 10 mm thick. All wall specimens were 1260
mm in height, 1030 mm wide and 75 mm thick (dimensions are
given on Fig. 1(a)).
Each wall is built on a reinforced concrete footing and is topped
by a horizontal reinforced concrete beam that distributes the lat-
eral and vertical loads as uniformly as possible.
Four walls have been reinforced: two have been externally
strengthened with Fibre Reinforced Polymers whereas Textile
Reinforced Concrete was used to reinforce the other two walls.
The idea is not to oppose one type of composite with another:
the objective is not to exactly substitute FRP with TRC. Main objec-
tive is to propose plausible strengthening configurations, to show
their feasibility and to study their effectiveness. The widths and
thicknesses of all specimens were changed depending on the con-
figurations to take into account the differences in axial stiffness of
all different selected strengthening materials, increasing the width
when glass fibres or TRC were used and reducing the width when
carbon fibre were used in FRP.
2.1.1. FRP materials and FRP strengthening configurations
In this study, two types of FRPs, differing in their textile fabric,
have been used as reinforcement for the masonry walls. The first
material consists of a bidirectional fabric of carbon fibres, embed-
ded in an epoxy matrix (CFRP). The strength of the fabric is concen-
trated in the direction of the warp with 70% of the fibres along that
CTRC GTRC
1-fold 2 folds
3 ± 0.3 6 ± 0.5
370 410
0.21 0.16
0.9 ± 0.06 0.7 ± 0.06
46,000 ± 4000 7500 ± 1500
29.4 ± 2.2 12.6 ± 1.7
420.0 ± 31.4 472.5 ± 63.8
Direct tensile tests on 6 Direct tensile tests on 6
specimens (100 mm wide) specimens (100 mm wide)
direction and 30% of the fibres in the weft direction. The second
type of FRP is a bidirectional fabric of glass fibres combined with
an epoxy-resin (GFRP). Table 1 provides the mass per unit area
and the design thickness, derived as the ratio between the mass
per unit area of the fabric and the bulk density of the strengthening
material [12], for the different fabrics.
The CFRP-UW wall (the Unconfined masonry Wall reinforced
with CFRP; Fig. 1(b)) has been strengthened symmetrically by
applying a 70 mm wide CFRP strip on both sides of each of the
faces of the wall.
The GFRP-UW wall (the Unconfined masonry Wall reinforced
with GFRP; Fig. 1(c)) has been symmetrically strengthened by
applying two superposed 400 mm wide GFRP sheets on both sides
of each of the faces of the wall.
2.1.2. TRC materials and TRC strengthening configurations
As in the case of FRP materials, the influence of fibre type was
investigated for TRC composites by using two different fibre-rein
forced-cementitious-matrix materials:
- a bidirectional fabric of carbon fibres embedded in a cementi-
tious matrix (CTRC). This fabric is balanced between the warp
and the weft directions. The mesh size and the weight of the
textile are 13 mm  13 mm and 370 g/m2, respectively.
- a bidirectional fabric of glass fibres embedded in a cementitious
matrix (GTRC). Contamine [13] has formulated this composite
so that it combines both high mechanical performance (an ade-
quate fibre-matrix bond) and good workability (compatible
with an in-situ application of this strengthening material).
In both cases, the matrix is a ready-to-use thixotropic mortar,
containing synthetic fibres, based on Portland cement with 0.4 of
water to cement ratio and graded sand with a maximum grain size
of 0.8 mm. This mortar is characterized by a 28-days compression
strength equal to at least 42 MPa (provider data).
The CTRC-UW wall (the Unconfined masonry Wall reinforced
with CTRC; Fig. 1(d)) has been symmetrically strengthened by
applying a 100 mm wide CTRC strip on both sides of each of the
faces of the wall.
The GTRC-UW wall (the Unconfined masonry Wall reinforced
with GTRC; Fig. 1(e)) has been symmetrically strengthened by
applying two superposed 200 mm wide GFRP strips on both sides
and at the centre of each of the faces of the wall.
All reinforcement sheets, regardless of whether they are in FRP
or in TRC, are anchored in the reinforced concrete foundation via
monodirectional carbon fibres. An anchor consists of two parts: a
dowel, inserted in a hole drilled through the concrete foundation
with a diameter of 16 mm and depth of 150 mm, and a fan part
which ensures the stress transfer. The aim is to better exploit the
potential of the reinforcements by avoiding premature delamina-
tion at strip ends.
 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83 73

Fig. 2. Response of the strengthening materials. (a) Load-strain curves and (b) Net stress-strain curves.

Table 1 indicates the characteristics of the different strengthen-

ing materials, obtained either by the manufacturer or by mono-
tonic tensile tests at the laboratory. To compare these materials,
it has been chosen to present these results in terms of load versus
strain curves (Fig. 2(a)) because the efficiency of reinforcements
depends on the load they are likely to sustain and because
strengthening dimensions (width and thickness) differ depending
on the configurations. The tensile load is calculated here as the
force that can sustain a composite material that has the same
thickness as the strengthening material on the masonry wall and
with width equal to the sum of the widths of all the strengthening
strips, applied on half of the wall (a priori tensile zone). To better
understand the behaviour of each strengthening material, Fig. 2
(b) presents their strain-net stress curve where net stress is calcu-
lated as the load divided by the cross-section area of the textile.
Fig. 2 underlines on the one hand the differences between the
mechanical behaviours of TRC and FRP: FRPs are linear elastic until
failure whereas TRCs exhibit a nonlinear behaviour, directly corre-
lated to the multi-cracking of these materials under tensile load- (a) Experimental test setup.
ing. On the other hand, this figure illustrates that, by varying
materials and the number of textile layers, the reinforcement com-
posites used in this study vary in the load they can withstand but
also in the flexibility they have. The aim is to assess the impacts of
these material characteristics on the global behaviour of the rein-
forced masonry walls.
The different strengthening configurations have been obtained
with regards to different parameters such as the requirement to
improve the mechanical behaviour of masonry structures, the need
to limit the costs (and to reduce the quantities of strengthening
materials) and the development of strengthening schemes with a
priori significant dissipation capacity according to previous works
based on monotonic loadings (in the case of TRC materials [11]). In
this respect and taking into account the incomplete understanding (b) Loading sequence.
of masonry walls reinforced by TRC-composites (a fortiori with hol-
low concrete blocks and with a mechanical anchorage of composite Fig. 3. Test setup and loading protocol.
bands), it seems inappropriate to adopt the traditional calculation
solicitations. In the present work, there could be no question of fro-
approaches, clearly developed for FRP-materials, in order to design
zen and incisive prioritization, but this is a flexible comparative
TRC-strengthenings and to promote an ultimately biased compar-
evaluation, wisely established, by prioritizing a phenomenological
ison between the two types of strengthening materials. Similarly,
analysis and insisting on the differentiated behavioural
a design by imposing a same material cost seems inadvisable in a rig-
mechanisms.
orous comparative perspective, since an emerging material (such as
TRC) is going to be more expensive at this stage of its development,
which could no longer be valid in case of a massive use of these mate- 2.2. Test setup and loading protocol
rials. Therefore, it is considered that the comparisons made are rel-
evant elements which lead to a better understanding of the The behaviour of the walls was investigated using the loading
behaviours (global and local) of masonry walls under in-plane cyclic setup shown in Fig. 3(a). The reinforced concrete footings, on
74 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83

Fig. 4. Instrumentation. (a). Location of displacement transducers and (b), (c), (d) and (e) strain gauges along the centreline of the composite sheets for CFRP-UW, GFRP-UW,
CTRC-UW and GTRC-UW walls respectively.

which walls have been built, are anchored to the strong floor
through a set of threaded rods with the purpose of preventing
any possible slippage or lifting due to a flexural mechanism. The
compression load, which depicts the weight of the upper floors
in a building, is gradually applied with a 500 kN capacity actuator
according to a force-controlled loading sequence. This compression
load is constant (and equals 1500 daN (i.e., 0.2 MPa, approximately
6% of the masonry compression strength)) for the duration of the
test.
Then, all walls have been subjected to an alternate cyclic
loading under displacement control by using a 500 kN hydraulic
actuator, which is connected on one side to two threaded steel rods
 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83
Fig. 5. Unreinforced wall. (a) Failure mode and (b) Load-displacement curve.
with diameter of 16 mm embedded in the upper horizontal rein-
forced concrete beam and on the other side to a reaction wall with
500 kN capacity. A frame, equipped with four threaded rods, envel-
ops the upper reinforced concrete beam to ensure a right displace-
ment control when the actuator «pulls on» the wall. The loading
cycle consists of a series of reversed displacement cycles of
increasing amplitude. For each increment, three cycles at the same
amplitude are applied (Fig. 3(b)). Previous studies carried out on
reinforced concrete structures [14] have indicated that, under
quasi-static loading conditions, loading protocols with repeated
cycles and increasing amplitude lead to smaller cyclic envelopes.
This type of loading protocol results in a greater loss of strength
in the post-peak zone. Therefore, the loading protocol adopted in
the present study is a priori a safe option to assess the deformation
capacity of the masonry structures, which is an important criterion
for the seismic design.
Displacement transducers have been employed to assess the
boundary conditions: to this end, transducers have been set up
on both sides of the reinforced concrete footing (LVDT 2 and LVDT
3 on Fig. 4(a)) to ensure that walls do not slip. The transducers
LVDT 4 and LVDT 5 provide information on rotation conditions
on the top of the wall. To investigate the global behaviour of the
walls, the walls are also instrumented with a laser sensor to mea-
sure the deflection at the top of the wall, and the displacement
transducer LVDT 1 returns the displacement halfway up the wall.
Data collection is complemented thanks to a series of strain gauges
that are glued along the centreline of the composite sheets on their
outer surface. They are aligned with the vertical direction. Their
number and their location differ depending on the strengthening
configurations. Fig. 4(b)–(e) provide information on their layout.
3. Behaviour of masonry walls – failure modes and hysteretic
curves
Fig. 5 shows the failure modes and illustrates the lateral load
versus top displacement curves obtained for all the tested walls.
3.1. Behaviour of the unreinforced masonry wall
A shear-flexural failure mode is observed for the unreinforced
masonry wall (Fig. 5(a)). First signs of cracking appear during load-
ing in the direction of the pull with a crack formation along the first
75
mortar bed joint. This crack then tilts at approximately 25 degrees
and spreads throughout the first layer of concrete blocks. Then,
cracking is observed during loading in the push direction at the
second mortar bed joint, and the crack propagates until it reaches
the cracks that already exist. At that time, the wall tends to slip in
the push direction, which is reflected in the hysteretic curve, by a
more significant energy dissipation effect and by higher residual
displacements than the displacements in the pull direction. In this
latter direction, the unreinforced wall is still subjected to a bending
loading with flexural cracks that extend along the first layer of con-
crete blocks. This damage mechanism achieves a lower energy dis-
sipation with narrower cycles. Considering all of these
observations, despite great care to construct the walls and to per-
form the test, damage mechanisms are not symmetrical. The dam-
age mechanisms depend heavily on heterogeneities (initial or
induced) that are consubstantial with masonry walls and on the
large variability in the characteristics of the constituent materials.
3.2. Behaviour of the FRP-reinforced walls
The damage of the CFRP-reinforced wall (Fig. 6(a)) is initiated
through a cracking around the carbon anchorage during loading
in the push direction. No deficiency of the anchorage has been evi-
denced. Thereafter, a crack appears along the diagonal in the pull
direction – the crack is tilted at an angle of approximately 60
degrees and is concentrated in the unreinforced zone, pointing
out that the reinforcements and the anchorage can influence crack
progression. Whatever the loading direction, blocks in compres-
sion at the wall toe show signs of crushing.
The GFRP-reinforced wall has an asymmetrical hysteretic curve
(Fig. 6(d)), which is justified by a malfunction of the loading setup:
for this wall, tested first (the order in which walls are presented is
indeed not the order in which tests were carried out), the system
used in the pull direction has not proven itself to be effective in
achieving the expected displacements: in this case, the hydraulic
actuator was just connected to the threaded steel rods embedded
in the RC loading beam and it was found that this system could
not ensure sufficient control of displacements in the pull loading
direction. For the other walls, the connection between the hydrau-
lic actuator and the RC loading beam was supplemented by two
steel plates attached on each side of the RC beam and maintained
with four steel rods to ensure full bearing of these plates to the
76 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83
loading. For the GFRP-reinforced wall, only the loading in the push
direction, considered not impacted by this default, will be analysed
later. The damage mechanisms result in cracks in the first layer of
concrete blocks in the unreinforced zone and in a macrocrack in
the compressed block at the wall toe (Fig. 6(c)).
3.3. Behaviour of the TRC-reinforced walls
The CTRC-reinforced wall displays shear failure with two cracks
formed along the diagonals, in the unreinforced zone at the bottom
half of the wall (Fig. 7(a)). These cracks are tilted at an angle of
approximately 30 degrees in the push direction and approximately
40 degrees in the pull direction. During cycles, short cracks have
arisen and have multiplied along the TRC sheets to eventually lead
to local delamination of the TRC.
Damage of the GTRC-reinforced wall (Fig. 7(c)) is evidenced by
many short cracks in the unreinforced zones between composite
strips. There is also, in the thickness of the wall, a significant hor-
izontal crack, and the location can be correlated with the multi-
cracks in the TRC-sheets. A local delamination of the composite
has occurred at the basis of the TRC strip on the loading side. This
delamination concerns only one of the two faces of the wall.
In short, damage mechanisms are concentrated in the lower half
of the walls with cracks with the size and configuration depending
Fig. 6. FRP-strengthened walls. Failure modes and load-displacement curves (a), (b) for the CFRP-UW wall and (c), (d) for the GFRP-UW wall.
on strengthening patterns and strengthening materials. In the case
of FRP, dissipative mechanisms occur in the unreinforced zone
with macrocracks that widen over time whereas in the case of
TRC, a process of additional dissipation takes place in the reinforce-
ment with a damage observed in the cementitious matrix, which
sometimes leads to a local delamination of these composites.
4. Comparative analysis of the walls
4.1. Analysis of the envelope curves
The comparative analysis is based on performance parameters
that are determined from the envelopes of the hysteretic curves
(Fig. 8(a)) and that are considered relevant to assess and quantify
the behaviour of the masonry walls under a lateral loading. At
low load levels, the walls exhibit a linear behaviour, characterized
by a stiffness K el .
To obtain the lateral displacement de associated with the end of
this linear phase, an approach similar to that of Pozza et al. [15] in
the context of seismic behaviour of wood-concrete walls is
adopted here. This approach can be transposed into our study
because the reinforced walls exhibit envelope curves with a second
linear phase with a positive hardening. An exponential approxima-
 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83
Fig. 7. TRC-strengthened walls. Failure modes and load-displacement curves (a), (b) for the CTRC-UW wall and (c), (d) for the GTRC-UW wall.
tion of each envelope curve (Fig. 8(b)) is then accomplished,
according to Eq. (1). Such an approximation is also adopted in [16].
  
K0d
H ¼ ðbd þ H0 Þ 1  exp 
H0
where H is the applied lateral load associated with the displace-
ment d; b is the hardening stiffness (the slope of the second linear
phase) and H0 is the intercept of this second phase. K 0 is the tangent
stiffness of the best fit exponential regression of the experimental
data. a is a secant stiffness, obtained by imposing the equality
between the strain energies under the best exponential curve with
the strain energy under the idealized curve formed by the two lines
with slopes a and b, respectively. The meeting point of these two
lines defines the displacementde .
The second linear phase extends to the maximum load Hmax
associated with the displacement dHmax . The ultimate lateral load
is conventionally assumed to be the load corresponding to 80% of
the maximum load after the peak, and this load is associated with
a displacement du . Table 2 lists the values of these various param-
eters for all the walls tested. For illustrative purposes, the vertical
reinforcement ratio for the strengthened walls is also given in this
table.
The initial stiffness (K el ) is obtained by a linear regression at the
first cycle of loading. The differences in the initial stiffnesses calcu-
lated for the different walls cannot be attributed to the reinforce-
ments, even if the initial stiffnesses tend to increase with the
reinforcement thickness and with the number of layers of textile
fabric. Strengthening materials start to contribute to the load car-
rying when the masonry experiences cracking. The average initial
77
stiffness (calculated from all walls and for both loading directions)
is about 2239 daN/mm and the coefficient of variation is 32%.
This dispersion may be related to natural variability of the hol-
low concrete block masonry. Vanin et al. [17] have studied a data-
ð1Þ
base of 123 stone masonry walls to estimate the variability of
different parameters (stiffness, strength, drift) and it appears that
such a coefficient of variation (32%) is commonly found in litera-
ture with masonry walls.
From the elastic theory, a relationship (Eq. (2)) can be estab-
lished between the lateral stiffness of the walls and the masonry
shear modulus. By using the experimentally determined stiff-
nesses, the shear modulus values, G, are derived for all the tested
walls and are given in Table 3.
GLx t
K el ¼   2  ð2Þ
0G
1; 2Lz 1 þ k E
Lz
Lx
where Lx ; Lz et t are, respectively, the width, the height and the
0
thickness of the masonry wall. k is a coefficient related to the
0
boundary conditions (k = 0.83 for a fixed-end wall and 3.33 for a
cantilever wall [18]). A fixed-end condition is adopted here because
of friction mechanisms between the metal plate in contact with the
hydraulic actuator and the concrete loading beam which can pre-
vent the wall from rotating during loading. E denotes the elastic
modulus of the masonry, measured from uniaxial compression tests
on three masonry panels whose dimensions have been chosen in
accordance with EN 1052-1 [19]. The small-scale masonry prisms
(four blocks high and two blocks wide), laid in running bond, were
tested under compression. The tests were performed under mono-
78 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83

the vertical axial load has a relatively low value (which allows
one to better assess the effectiveness of the strengthening materi-
als). This property may contribute to the justification of the dis-
crepancy with the Eurocode value. Bosiljkov et al. also specify
that the use of G = 0.4 E leads to a correct stiffness if the compres-
sive loading is predominant but is not appropriate in the prediction
of the stiffness of laterally loaded masonry structural elements.
They mention that the shear modulus obtained from the diagonal
tests can be used to estimate the lateral stiffness for masonry
structures assembled with a stiff mortar. For brickwork made with
weak (lime) mortars, the determination of the lateral stiffness from
a shear modulus remains an open problem because the shear mod-
ulus values obtained from both compressive and diagonal tests are
not suitable for prediction of the lateral stiffness.
(a) Envelope curves for all the walls tested Lateral strength. The reinforcement patterns that have been
adopted lead to an increase in the maximum lateral load of the
strengthened walls compared to the unreinforced walls, for all
strengthening materials. The gain is at least 87% for the CTRC-
UW wall and grows to 267% for the GFRP-UW wall.
b-coefficient, slope of the second linear phase. This coefficient
seems to be related to the nature of the matrix that surrounds
the reinforcing fabric sheets. Indeed, the b-coefficients for CTRC-
UW and GTRC-UW walls have close values (71; 77 and 113), clearly
inferior to those of CFRP-UW and GFRP-UW walls (249; 317; 395).
For a same type of matrix (cementitious or polymeric), b-
coefficients values remain relatively close to each other.

(b). The process for identifying the parameters
walls.
Fig. 8. Results and analysis.
tonic displacement control, until panel failure, at a loading rate of
0.05 mm/min. Two LVDTs were placed on both faces to record the
masonry strains. For each prism, the elastic modulus is defined as
the secant modulus, calculated between 5% and 30% of the compres-
sive strength. The mean value of elastic modulus obtained from
these tests is 5560 ± 700 MPa. Details of these experimental tests
are given in [20].
On average, the shear modulus thus determined is approxi-
mately 9% of the elastic modulus. This percentage is much lower
than the 40% recommended by Eurocode 6 [21]. This result con-
firms the Tomazevic recommendation [18] that prompts us to con-
sider that the shear modulus is equal to 10% of the elastic modulus.
Bosiljkov et al. [22], however, underline that it is essential to
remain cautious in the use of Eq. (2) to determine the masonry
shear modulus because this method does not consider the influ-
ence of the compression level. Moreover, in the present study,
4.1.1. Ductility
The ductility coefficient, l (Table 2), defined as the ratio of the
ultimate displacement to the displacement at the end of the elastic
phase, only increases with the TRC-strengthening patterns How-
ever, few values are available in this study for FRP materials and
the ductility coefficient for the reference wall highly depends on
the loading direction and on the damage mechanisms involved in
each direction. A more consistent experimental program would
be essential to support this assumption. This parameter depends
on the applied vertical compression load, which is rather weak in
the present study.
4.2. Analysis of the hysteresis loops
The measurement of structural seismic performance cannot sim-
ply rest on the ductility coefficient or the strength. The degree of
damage to the structures must also be considered, but such damage
indices are not easy to define because of the diversity and the com-
plexity of potential damage mechanisms. A first approach consists of
considering the stiffness decay during cycles as a damage parameter.
Another method assumes a correlation between damage (or plastic-
ity) and the amount of dissipated energy. These two approaches are

Table 2
Parameters describing the performance levels of all the walls tested.

Wall qv Loading Direction K el K0 a b de dHmax H max Load gain du l ¼ ddue

[%] [daN/mm] [mm] [mm] [daN] [%] [mm]
URW – Push 2210 2210 - - 0.45 0.83 1200 - 2.3 5.1
Pull 1800 1800 - - 0.49 0.64 1081 - 0.64 1.3
CFRP-UW 0.17 Push 1291 2183 1229 249 1.76 3.88 2731 +128 5.6 3.2
Pull 1801 3309 1873 395 1.1 3.71 3119 +189 (*) -
GFRP-UW 7.04 Push 3081 4166 2252 317 1.33 5.67 4406 +267 (*) –
CTRC-UW 1.55 Push 2427 4393 2232 71 0.94 6.16 2456 +105 8 8.5
Pull 1409 2463 1291 113 1.14 5.82 2019 +87 7.3 6.4
GTRC-UW 9.32 Push 2516 4370 2224 77 1.47 10.7 3950 +229 12.2 8.3
Pull 3616 6175 3144 113 0.91 7.8 3913 +262 (*) -

(*) Test was stopped before the load dropped 20%.

 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83 79

Table 3
Shear modulus of the masonry, calculated from the elastic modulus.

Loading direction Unreinforced Wall CFRP-UW GFRP-UW CTRC-UW GTRC-UW

Push Pull Push Pull Push Push Pull Push Pull
G [MPa] 479 382 268 383 697 531 294 553 841
G/E [%] 8.6 6.9 4.8 6.9 12.5 9.6 5.3 10.0 15.1

Table 4
Adjustment parameters to describe the stiffness degradation during cycles.

Wall ɣ – Push Loading Direction ɣ – Pull Loading Direction

Unreinforced wall 1.045 1.067
CFRP-UW 0.739 0.723
GFRP-UW 0.74 –
CTRC-UW 0.856 0.83
GTRC-UW 0.838 0.843

4.2.1. Stiffness decay during cycles

(a) Assignment of the parameters to define the stiffness degradation
(b) Normalized stiffness decay
The stiffness of masonry walls is analysed in terms of the secant
stiffness, K s;i , defined as the ratio of the maximum load of a cycle to
the associated displacement (Fig. 9(a)). The hysteresis loops show
that the secant stiffness decreases with an increase in the applied
displacement. In the literature, some authors [23,24] are more
interested in the secant stiffness at each cycle, normalized with
respect to the secant stiffness at maximum load (K s;i =K s ). They fol-
low its evolution in terms of the ratio of the maximum displace-
ment at each cycle to the displacement associated with the
maximum load (dmax;i =dHmax Þ. In their example, only one parameter,
c (hereinafter called the adjustment parameter), is needed to fit the
results to quantify the stiffness degradation (Eq. (3)).
 c
K s;i dmax;i
¼ ð3Þ
Ks dHmax
Fig. 9(b) and (c) highlight the relevance of this equation to
assess the stiffness degradation for masonry walls reinforced with
composite materials.
Table 4 provides information on the adjustment parameter
obtained for each of all the walls tested. Even if these adjustment
parameters must be confirmed by a larger experimental campaign,
these first results show a correlation between the adjustment
parameter and the strengthening material. Indeed, whatever the
, versus , adopted strengthening configuration and whatever the loading
direction, FRP-reinforced walls have similar adjustment parame-
ters (between 0.74 and 0.723), also true for TRC-reinforced
walls for which the adjustment parameter varies from 0.856 to
0.83. Damages in the cementitious matrix of TRC materials lead
to a greater stiffness decay than with FRP, agreeing with the beha-
viour of TRC on the material scale. Indeed, the matrix damages
(cracking) imply a greater fibre contribution, not necessarily uni-
form, but inevitably at the expense of the global stiffness of the
composite, which is significantly reduced.

4.2.2. Energy dissipation during cycles

The dissipated energy during a cycle is defined as the area
formed by the loop in the load–displacement curve (hatched area
in Fig. 10(a)), and it is calculated by using the trapezoidal rule.
The total energy transferred to the specimen during a cycle corre-
sponds to the area under the load–displacement curve (shaded
area in Fig. 10(a)). Fig. 10(b) shows the cumulative dissipated ener-
Fig. 9. Stiffness degradation.
gies for all the walls tested. In Fig. 10(c), the cumulative dissipated
energies of the reinforced walls are analysed with respect to the
employed below to support the comparison between the different cumulative dissipated energy of the unreinforced wall at the same
strengthening configurations. maximum displacement, measured at the top during the cycle.
80 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83

Fig. 10(b) shows that the cumulative dissipated energies of 4.2.3. Equivalent viscous damping ratio
masonry walls, regardless of whether the wall is reinforced, are To quantify the energy dissipated by a structure under cyclic
governed by second degree equations as the maximum displace- solicitations, the equivalent viscous damping ratio, neq , is com-
ment increases. Fig. 10(c) indicates that for all the tested configu- monly selected. To account for the asymmetrical behaviour of
rations, a displacement exists from which the cumulative the walls depending on the loading direction (push or pull), the
dissipated energy of a reinforced wall becomes higher than the equivalent viscous damping ratio is calculated here on only a
cumulative dissipated energy of the unreinforced wall. This dis- half-cycle of the hysteretic curve (Fig. 11(a)) by using Eq. (4) [25],
placement is larger for FRP-reinforced walls than for TRC-
1 Ed
reinforced walls: this difference can be related to the linear elastic neq ¼ ð4Þ
behaviour of the FRP reinforcements. These materials succeed in
p Hmax;i dmax;i
limiting masonry damages at low displacements, thus generating
a lesser energy dissipation. Fig. 10(c) also shows that, on the range
of applied displacements, the cumulative dissipated energy of the
GTRC-UW wall is five times higher than the cumulative dissipated
energy of the unreinforced wall. The multiplication factor reaches
two or three for the CTRC- and CFRP-reinforced walls.

(a) Definition of the parameters for the calculation

of the equivalent viscous damping ratio.
(a) Definition of input and dissipated energies from the load-
displacement curve associated with one cycle.

(b) Cumulative dissipated energy for each wall tested. (b) Equivalent viscous damping ratio with respect to
the maximum displacements measured during cycles.

(c) Evolution of the ratio of the cumulative dissipated energies of the

reinforced walls to the cumulative dissipated energy of the unreinforced wall at
the same displacement, with respect to the maximum
displacements measured during cycles. (c) Residual displacements (return to zero force) during cycles.

Fig. 10. Energy dissipation. Fig. 11. Main parameters obtained from experimental data.
 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83 81

Fig. 12. Measured longitudinal strains in the FRP-strengthening materials versus the maximum measured displacements during cycles.

where Ed measures energy dissipation capacity of the wall and the
product Hmax;i dmax;i depicts double the recoverable elastic strain
energy stored in an equivalent linear elastic system.
Fig. 11(b) shows the variation of the equivalent viscous
damping ratio with the maximum measured displacement dur-
ing cycles for all the tested walls. Except for the unreinforced
wall in the push loading direction for which the equivalent vis-
cous damping ratio is approximately equal to 15% (such a value
can be explained by friction during slippage in this direction),
the other walls reach rather similar damping coefficients –
approximately 6% - in the range of the applied displacements.
A correlation can be established between the evolution of the
equivalent viscous damping ratio and the residual displacements
(displacement for which the specimen returns to zero force),
given in Fig. 11(c). The equivalent viscous damping ratio
increases with increasing residual displacements. However, one
wall seems to constitute an exception: the GTRC-reinforced wall
in the push loading direction displays relatively low residual dis-
placements, but its damping coefficient increases. It is likely that
multi-cracking mechanisms in TRC, especially involved in this
loading direction, contribute to the evolution of the equivalent
viscous damping ratio.
82 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83

Fig. 13. Measured longitudinal strains in the TRC-strengthening materials versus the maximum measured displacements during cycles.

4.3. Analysis of the local parameters
In Fig. 12 and in Fig. 13, the evolution of longitudinal strains in
reinforcements with maximum measured displacements during
cycles is provided. Strain levels between FRP and TRC reinforce-
ments can hardly be compared because strain gauges are bonded
on the outer surfaces of the strengthening materials. However,
strains vary with the reinforcement thickness [26] and changes
can be noticeable, in particular with relatively thick TRCs. There-
fore, strains will deliberately be analysed qualitatively here, not
quantitatively.
For all the reinforced walls, whatever the strengthening pattern
and whatever the strengthening material, the reinforcements only
have a significant contribution to the tension: strains recorded dur-
ing compressive loading remain low when compared to strains
recorded during tensile loading.
Quite logically, the reinforcements have little role during the
first linear phase. The displacement associated with the first signif-
icant evolution in the strain values of one of the gauges corre-
sponds approximately to the beginning of the second linear zone
of the load-displacement curve (the line portion with slope b).
Indeed, whatever the loading direction, significant changes in
 N. Reboul et al. / Construction and Building Materials 164 (2018) 70–83
strains occur at displacement de . or just slightly after, depending on
the size of the transition zone between the two linear phases of the
load-displacement curve.
For CFRP-UW (Fig. 12(a)), in both loading directions, strains
increase linearly between de et dHmax and remain nearly constant
beyond dHmax . Some gauges do not appear to show compressive
strains. In the push loading direction, the ultimate displacement
is associated with a sudden change in the strain of gauge 3, located
on the block in tension at the first layer, where the crack has been
detected. Even if the recorded strains are close to the measured
ultimate tensile strain of the composite, no CFRP failure has been
observed. For GFRP-UW (Fig. 12(b)), the most solicited strain gauge
is near the anchorage - which is consistent with the failure mode of
this wall (cracking at the first layer of concrete blocks) – then, the
gauges glued on composites, over mortar joints, and eventually,
the gauges glued over concrete blocks.
For TRC-reinforced walls (Fig. 13(a) and (b)), strains change
more sharply than with FRP-reinforced walls. There is a concern
about clarification if this evolution is characteristic of the TRC-
composite, or if the evolution is a bias linked to the thickness of
TRC. Nevertheless, with TRC, strain increase is not gradual, and
the sudden changes in strain appear close to the characteristic dis-
placements (de , dHmax or du ).
5. Conclusions
The experimental study presented in this paper underlines the
benefits of using externally bonded composite materials (ade-
quately anchored) to strengthen masonry structures. Indeed, all of
the strengthened walls that were tested exhibited an improved
behaviour, both in terms of lateral load and ductility. Depending
on the adopted reinforcement ratios, the increase in lateral capacity
brought by FRPs ranges from 128 to 267% and the gain provided by
TRCs ranges from 87 to 262%. The load transfer towards the rein-
forced concrete foundation is pertinent here because the foundation
has sufficient strength to sustain the load increments. This point is
important to bear in mind when considering the strengthening of
real structures. Under the compression load adopted here, TRC-
strengthening materials contribute more effectively to an increase
in ductility than FRP-composites, but FRP-reinforcements have
proved to be effective in delaying the onset of damage. Cracking
mechanisms differ depending on the strengthening configurations
and on the strengthening materials, but these cracking mechanisms
generally involve macrocracks, concentrated in unreinforced zones
where the opening grows during cycles. The GTRC-UW wall, charac-
terized by a high reinforcement ratio and a significant reinforced
area, is an exception because there is an increasing number of cracks
in the reinforcement itself during cycles. The dissipation capacity of
this wall is improved, regardless of the dissipative mechanisms of
the masonry itself.
Acknowledgements
The supply of the glass fabrics for the TRC-composites by Texi-
nov, a French society specializing in warp-knitted fabrics, is grate-
fully acknowledged. The authors wish also to thank Emmanuel
Janin and Norbert Cottet, the laboratory technicians, for providing
a significant help in this experimental campaign.
83
References
[1] T. Triantafillou, Strengthening of masonry structures using epoxy-bonded FRP
laminates, J. Compos. Constr. 2 (1998) 96–104.
[2] G. Marcari, D. Oliveira, G. Fabbrocino, P. Lourenco, Shear capacity assessment
of tuff panels strengthened with FRP diagonal layout, Composites Part B 42
(2011) 1956–1965.
[3] F.G. Carozzi, G. Milani, C. Poggi, Mechanical properties and numerical modeling
of Fabric Reinforced Cementitious Matrix (FRCM) systems for strengthening of
masonry structures, Compos. Struct. 107 (2014) 711–725.
[4] J. Hegger, S. Voss, Investigations on the bearing behaviour and application
potential of textile reinforced concrete, Eng. Struct. 30 (2008) 2050–2056.
[5] W. Brameshuber, Textile Reinforced Concrete – State-of-the-art Report of
RILEM TC 201-TRC, 2006.
[6] A. Prota, G. Marcari, G. Fabbrocino, G. Manfredi, C. Aldea, Experimental in-
plane behavior of tuff masonry strengthened with cementitious matrix-grid
composites, J. Compos. Constr. (2006) 223–233.
[7] S. Babaeidarabad, D. Arboleda, G. Loreto, A. Nanni, Shear strengthening of un-
reinforced concrete masonry walls with fabric-reinforced-cementitious-
matrix, Constr. Build. Mater. 65 (2014) 243–253.
[8] C. Faella, E. Martinelli, E. Nigro, S. Paciello, Shear capacity of masonry walls
externally strengthened by a cement-based composite material: An
experimental campaign, Constr. Build. Mater. 24 (2010) 84–93.
[9] C.G. Papanicolaou, T.C. Triantafillou, K. Karlos, M. Papathanasiou, Textile-
reinforced mortar (TRM) versus FRP as strengthening material of URM walls:
in-plane cyclic loading, Mater. Struct. 40 (10) (2007) 1081–1097.
[10] L. Koutas, S. Bousias, T. Triantafillou, Seismic strengthening of masonry infilled
RC frames with textile-reinforced mortar: experimental study, J. Compos.
Constr. 19 (2) (2015).
[11] T.-L. Bui, A. Si Larbi, N. Reboul, E. Ferrier, Shear behaviour of masonry walls
strengthened by external bonded FRP and TRC, Compos. Struct. 132 (2015)
923–932.
[12] S. De Santis, G. de Felice, Tensile behaviour of mortar-based composites for
externally bonded reinforcement systems, Composites Part B 68 (2015) 401–
413.
[13] R. Contamine, Contribution à l’étude du comportement mécanique de
composites textile-mortier. Application à la réparation et/ou renforcement
de poutres en béton armé vis-a-vis de l’effort tranchant (Doctoral thesis),
Université de Lyon, 2011.
[14] H. Takemura, K. Kawashima, Effect of loading hysteresis on ductility capacity
of reinforced concrete bridge piers, J. Struct. Eng. 43 (1997) 849–858.
[15] L. Pozza, R. Scotta, A. Polastri, A. Ceccotti, Seismic behaviour of wood-concrete
frame shear-wall system and comparison with code provisions, in:
Proceedings of the 45th meeting, International council for research and
innovation in building construction, working commission W18, Timber
structures, CIB-W18, Vaxjo, Sweden, 2012.
[16] A. Jorissen, M. Fragiacomo, General notes on ductility in timber structures,
Eng. Struct. 33 (2011) 2987–2997.
[17] F. Vanin, D. Zaganelli, A. Penna, K. Beyer, Estimates for the stiffness, strength
and drift capacity of stone masonry walls based on 123 quasi-static cyclic tests
reported in the literature, Bull. Earthquake Eng. (2017).
[18] M. Tomazevic, Shear resistance of masonry walls and Eurocode 6: shear versus
tensile strength of masonry, Mater. Struct. 42 (2009) 889–907.
[19] EN 1052-1, Method of Test for Masonry. Part 1: Determination of Compressive
Strength, Brussels, 1999.
[20] L. Bui, N. Reboul, A. Si Larbi, E. Ferrier, Mechanical in-plane behaviour of
masonry walls reinforced by composite materials: Experimental and analytical
approaches, J. Compos. Mater. (2017).
[21] Eurocode 6, Design of Masonry Structures-part 1-1: Common Rules for
Reinforced and Unreinforced Masonry Structures. EN 1996-1-1+A1, Brussels,
2013.
[22] V. Bosiljkov, Y.Z. Totoev, J.M. Nichols, Shear modulus and stiffness of brickwork
masonry: An experimental perspective, Struct. Eng. Mech. 20 (2005) 21–43.
[23] S. Churilov, ‘‘Experimental and analytical research of strengthened masonry,”
University ‘‘Ss. Cyril and Methodius”, Skopje, 2012.
[24] V. Haach, Development of a Design Method for Reinforced Masonry Subjected
to In-plane Loading Based on Experimental and Numerical Analysis,
Universidade do Minho, 2009.
[25] R. Crambuer, Contribution à l’identification de l’amortissement. Approches
expérimentale et numérique, Ecole Normale Supérieure de Cachan, 2013.
[26] L. Bui, Contribution à l’étude de murs maçonnés renforcés par matériaux
composites (FRP et TRC). Application aux sollicitations dans le plan (Doctoral
thesis), Université Lyon 1, 2013.