Construction and Building Materials 299 (2021) 124276

Contents lists available at ScienceDirect

Construction and Building Materials

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

Full size testing and detailed micro-modeling of the in-plane behavior of

FRCM–reinforced masonry
Giulio Castori a, Marco Corradi b, c, *, Emanuela Sperazini a
a
University of Perugia, Department of Engineering, Via Duranti 93, Perugia 06125, Italy
b
Northumbria University, Department of Mechanical and Construction Engineering Newcastle upon Tyne NE1 8ST, United Kingdom
c
University of Perugia, Department of Engineering, Perugia 06125, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: Fiber Reinforced Cementitiuos Mortars (FRCM) are a relatively new class of reinforcement used in conservation
Brickwork masonry engineering. However, the structural response of shear walls reinforced with non-isotropic composite material
Earthquake engineering nets, bonded with relatively low-strength and non-elastic mortar matrices is difficult to study with numerical
Mechanical testing
models. The reliability of numerical models could be improved if experimental data was available from full size
In-plane loading
Retrofitting methods
testing. However, such data is expensive to obtain as large specimens are needed. In this paper, an experimental
campaign was carried out at the laboratory on 12 full size wall panels. These were tested in diagonal tension to
assess their shear behavior in unreinforced and FRCM-reinforced configurations. Some manipulation of the data
is required in order to interpret the experimental results, but diagonal tension testing has many advantages over
other testing methods to study the shear behavior of masonry walls. In general the results have demonstrated that
FRCM strengthening of shear walls is an effective retrofitting method. Different reinforcement schemes (single-
and double-sided retrofits) and bonding inorganic matrices have been studied. Lastly, a two-step non-linear
numerical procedure, calibrated with characterization test data and benchmarked on the laboratory outcomes (i.
e. full scale testing), has been also implemented to define an effective tool for predicting the shear behavior of
reinforced walls.

1. Introduction retrofit methods have been used for decades in Italy to reinforce shear
walls: ferro-cement is a well-known example of such incorrect method.
The 2016 Central Italy earthquake was another strong reminder of The problem of corrosion of the steel reinforcement and the difficulties
the seismic risk that exists in the Eastern Mediterranean area [1–2]. The associated with the detection of this type of deterioration have been
quake caused 92% of the historic buildings of the hill town of Amatrice underestimated for a long period of time [19–20].
to collapse, 299 deaths and approximately 400 injuries [3]. The most When engineers and the construction industry started to seek alter­
part of these buildings were unreinforced either irregular stone or natives to steel reinforcement in the late 1990s, a potential solution was
brickwork masonry (URM) [4–6]. The typical collapse mode of URM the use of Fiber Reinforced Polymers (FRPs). The function of the fibers
historic buildings is due to out-of-plane rocking or shear cracking of was to carry tensile stresses, and prevent masonry cracking [21–24].
load-bearing walls. This manly depends on the direction of the seismic This can only achieved by using strong adhesives to connect the masonry
action (out-of-plane or in-plane), but, while out-of-plane rocking can be with the fibre reinforcing, and epoxy resins have been typically
studied using rotational equilibrium equations, taking into account employed. Initial results were promising: by adding a small quantity of
slenderness and conditions of constraint of the loaded walls [7–9], in- new mass, it was possible to increase the tensile strength of masonry
plane capacity highly depends on the mechanical properties of the members, and this has positive effects on the shear capacity of the wall
masonry material [10–12]. panels and the overall seismic capacity of the buildings. However, if the
Shear reinforcement is often necessary to increase the seismic ca­ “structural” problem was partially solved, other issues remained with no
pacity of wall panels, and a large amount of research had been devoted immediate answer, such as the heritage, community and aesthetic
to this problem in the last 3 decades [13–18]. Incorrect or ineffective values, the architectural interest, the authenticity, the long run

* Corresponding author at: Northumbria University, Department of Mechanical and Construction Engineering Newcastle upon Tyne NE1 8ST, United Kingdom.
E-mail address: [email protected] (M. Corradi).

https://doi.org/10.1016/j.conbuildmat.2021.124276
Received 26 March 2021; Received in revised form 18 June 2021; Accepted 15 July 2021
Available online 24 July 2021
0950-0618/© 2021 Elsevier Ltd. All rights reserved.
G. Castori et al. Construction and Building Materials 299 (2021) 124276

performance of the reinforcement, the compatibility between materials. embedment of the mesh itself into the mortar coating, because of a
In fact, several aspects need to be considered in the seismic upgrade higher matrix-reinforcement interface area. Fig. 1 shows different types
of heritage masonry buildings: not only the social and financial costs of fibrous meshes used to form FRCMs.
associated with this task are high and typically supported with public Substantial research has been dedicated to the analysis of the
funding by governmental agencies, but also several other problems bonding performance and behavior between fibrous nets and stone/
should be analyzed. The need to guarantee compatibility between new brickwork masonry, while a limited amount of research has been dedi­
(reinforcement and matrix) and old materials (masonry, brick, stone, cated to full-size testing of shear walls, especially if made of irregular
mortar), and reversibility (i.e. “any measures adopted should be stone masonry. This paper is aimed at providing experimental evidence
reversible so that they can be removed and replaced with more suitable to this aspect and proposing a numerical method for analysis. The
measures when new knowledge is acquired”, according to the ICOMOS bonding was experimentally studied using single-lap shear tests
2003 charter), has additionally complicated the already difficult task of [34–35]: results demonstrated the non-linear shear load vs. slip rela­
engineers and architects in this area [25–27]. tionship: the different rigidities (i.e. Young’s modulus) of the three
The use of epoxy-bonded FRPs demonstrated to be problematic not bonded materials (typically, high Young’s modulus and low axial
only to meet the requirement of reversibility (the resin during applica­ deformation at failure for brick, stone and mortar coating, and higher
tion penetrates irreversibly into the stones or bricks), but also for the deformation and lower modulus for the fibrous net) reduced the extent
unsatisfactory long term behavior of the epoxies when exposed to of a compatible deformation under loading. The consequence of this is
environmental actions (freeze-thaw cycles, UV radiation, humidity, etc.) an initial cracking of the mortar coating, followed by deboning or fiber
[28–29]. Another critical limitation of FRPs is their poor performance cracking phenomena. However, test results have also demonstrated that
when exposed to high temperatures. This can compromise the efficiency it is possible to reach an exploitation rate (given by the ratio between the
of FRP systems. net’s tensile strength σf and the average shear stress (at maximum load)
Against this background, Fiber Reinforced Cementitiuos Mortars σu) ranging from 35% to 45%.
(FRCM), also known as Textile Reinforced Concrete (TRC), or Textile
Reinforced Mortar (TRM), have been proposed as an alternative solution 2. Experimental work
[30–33]. The fibrous reinforcement (glass, carbon, aramid or basalt fi­
bres) is bonded to masonry using an inorganic mortar (typically made 2.1. Geometry
with cement, but also lime mortars can be used for this purpose).
Because the bond between the reinforcement and the masonry substrate To investigate the reinforcement effect, twelve full size
is weaker compared to when epoxy resin is used, larger surfaces of (1200x1200x240 mm) masonry panels were constructed and tested at
masonry are typically reinforced, using low weight-density fibrous net the Structures Laboratory of the University of Perugia (Fig. 2). Wall
or mesh. The thickness of mortar coating normally varies between 10 specimens were assembled and reinforced by an experienced mason.
and 40 mm, depending on the type of masonry to retrofit, and shape of Two masonry types were used for construction: stone and brickwork
the fibrous net. The use of a fibre mesh also guarantees a better masonry. Overall, the selected masonry typologies reflect the typical

Fig. 1. Different types of fibrous meshes, used in FRCM.

2
G. Castori et al. Construction and Building Materials 299 (2021) 124276

load. Finally, a digital acquisition system was used to measure the

magnitude of the diagonal load and the displacements as a function of
the time, at a frequency of 2 Hz.
Using the mechanical analogy of an in-plane loaded membrane, the
plane stress state (Fig. 4) at the panel’s centre is given by:
F
σxx = σ yy = − 0.56
An
(1)
F
σxy = σ yx = 1.05
An

where σxx and σxy are the normal and shear stresses, respectively; F is the
shear (diagonal) load, and An is the wall cross section (thickness ×
width).
The maximum and minimum normal stresses (i.e. principal stresses)
σI and σII can be easily calculated with the method of the Mohr circle:
F
σI = 0.50
An
Fig. 2. The wall panels tested in the experimental study.
(2)
F
σII = − 1.62
An
ones used historically in Italian masonry construction.
Different interpretations exist to estimate the masonry shear strength
For both stone and brickwork specimens, wall thickness was 240
using the diagonal tension test. According to the RILEM guidelines [37],
mm. Table 1 shows the test matrix: brickwork specimens were made of
failure occurs when the tensile principal stress σI is equal to the masonry
solid bricks (MAT- series), while the stone walls (PIE- series) were
tensile strength ft:
constituted by irregular (rubble) calcareous stones. The letter designa­
tions U, D and S were used to identify unreinforced, double-sided and Fmax
ft = σ I = 0.50 (3)
single-sided strengthenings, respectively. An

where Fmax is the maximum shear load. Finally, the masonry shear
2.2. Test setup
strength, τ0, can be calculated with:
Wall panels have been tested in shear according to the test setup ft
τ0 = (4)
shown in Fig. 3. This arrangement is known as ‘Diagonal Tension Test’, 1.5
and it has been standardized by the ASTM International (ASTM 519, With regard to the deformation characteristics of the wall panels,
[36]). The shear load was applied diagonally using a single-effect 50 these were studied by using a contact instrumentation, applied to
tons hydraulic cylinder: the wall panel typically fails by a diagonal shear measure the movements of the panels under the shear loading. 4 Linear
crack, originating from the panel’s centroid where a highly strained Variable Differential Transducers (LVDT) were used to measure the
region is assumed to form. This test method allows the determination of contractions and elongations of the four diagonals (on both sides,
the shear strength (diagonal tensile strength) of masonry assemblages by labelled with letters A and B, of the wall panels). The shortenings of the
loading them in compression along one diagonal. Two steel I-beams and wall’s loaded diagonal are ΔlCSA and ΔlCSB for side A and B, respectively,
two steel shoes are placed at the panel’s corners to apply the diagonal while the corresponding elongations of the unloaded diagonal are ΔlTSA
and ΔlTSB.
Table 1 The normal strains are given by:
The test matrix.
ΔlCSA Δl
Masonry Single- or Double sided Mesh Coating εCSA = εTSA = TSA
Reinforcement Mortar
lCSA lTSA
(5)
ΔlCSB Δl
MAT- – – – εCSB = εTSA = TSA
01-U lCSB lTSA
MAT- Double Type Type A
02-D 1 where lCSA and lCSB are the gage lengths of the LVDTs applied along the
MAT- Brickwork Single Type Type A compressed diagonals on side A and B, respectively. Similarly, lTSA and
03-S 1
lTSB are the gage lengths for the stretched wall diagonals. Hence, the
MAT- Double Type Type A
04-D 2
compressive and tensile axial strains, εc and εt, and the shear strain, γ,
MAT- Single Type Type A are given by
05-S 2
MAT- Single Type Type B
εCSA + εCSB
εC =
06-S 1 2
(6)
MAT- Single Type Type B εTSA + εTSB
07-S 2
εT =
2
PIE-01- – – –
U γ = |εC | + εT (7)
PIE-02- Stone Double Type Type A
D 1 The shear modulus Gbil of the masonry material has been calculated
PIE-03-S Single Type Type A
assuming a bilinear Stress Block response (Parabola-rectangular stress
1
PIE-04- Double Type Type A
block). The shear modulus is given by the slope of the line representing
D 2 the elastic phase.
PIE-05-S Single Type Type A In Fig. 5, the shear stress τ is given:
2

3
G. Castori et al. Construction and Building Materials 299 (2021) 124276

50-ton
Hydraulic Cylinder

Steel Shoe
Contact Instrumentation
LVDT transducer

Steel Rod Tie

Wall Panel
Brickwork
1200x1200x250 mm

H-shaped
Steel Profile

Fig. 3. Shear test setup.

and
τ (
A
)
γy = 2 γu − (9)
Shear Stress τu
1.05 F/A n
where τu and γu are the maximum shear stress and the corresponding
Principal Stress σI angular strain, respectively, calculated in correspondence of Fmax. It can
be noted thatγ y is calculated by imposing a condition of energy equality,
-1.62 F/An 0.5 F/A n
σ i.e. the equality between the area (A) underlying the diagram of the
experimental curve (solid black line in Fig. 5) and that underlying the
Principal Stress σII bilinear diagram (dashed lines). This underlying area has been calcu­
lated up to an angular strain of γu.

Fig. 4. Representation of stress state at the panel’s centroid, using the

2.3. Materials properties
Mohr circle.

2.3.1. Masonry material

Solid un-sanded standardized (in accordance with EN771 [38])
bricks have been used for the construction of the shear walls. These were
τ ≅ σ xy (8) clay, durable, high strength, solid engineering bricks (240x120x55 mm
in dimensions), available on the construction market in Italy and

Fig. 5. The method used for the calculation of the shear modulus of elasticity Gbil.

4
G. Castori et al. Construction and Building Materials 299 (2021) 124276

manufactured by Toppetti Company (located in Masserano, Italy) to reinforce the masonry wall panels. The mesh spacing (i.e. the aperture
cope with exposure to aggressive conditions (product denomination size) is 50 and 30 mm for Type 1 and Type 2, respectively. In both cases
“Mattone Pieno”). The mechanical properties of the bricks are summa­ these are made of a square grid pre-impregnated mesh. Fig. 7 shows a
rized in Table 2: to note a high compressive (38.75 MPa) and flexural detail of the mesh: it can be noted the warp and weft wire components
(3.61 MPa) strengths. used in weaving to turn the fiberglass fiber into the meshed fabric and
It is important to discuss the brickwork bonding pattern, as this may the joint at their intersection. In the weft direction (horizontal in Fig. 7),
have a critical effect on the lateral strength and deformation capacity of a single strand made of un-twisted fiberglass filament was used, while
the wall panels. Shear cracks are inclined at 45◦ and typically develop in for warp direction 6 or 4 twisted strands were employed.
the bed and head mortar joints (as the mortar is normally weaker than To prevent buckling and debonding of the FRCM from the masonry
the bricks in historic walls) following a zig-zag pattern. The used substrate, through L-shaped connectors (unequal leg angles, 100x200
bonding pattern is shown in Fig. 6: it can be noted an alternate use of mm) have been applied (Fig. 8a). These were also made of fiberglass (the
stretchers and headers on the brick courses. This a common bonding producer declares in the data sheet a tensile load capacity of 18.9 kN,
pattern historically used in Italy, and it is similar to the well-known diameter 8 mm, Young’s modulus 80 GPa). The longest connector leg
Flemish bond. has been inserted into a 12 mm-diameter hole drilled in the masonry and
Stone walls were assembled using barely-cut calcareous, high density filled with the same mortar used for the coating (Fig. 8b). A total number
stones. The objective was to simulate typical traditional stone masonry of 4 connectors were used for reinforced panels (2.8 connectors/m2).
constructions, which are common not only in Italy but also in many Two types of mortars were used for the coating: Both are ready-to-
other European and Middle East countries. Usually river-bed pebbles use mortars, having a special fibre-reinforced composition. Mortar
and other local stones found on the fields were collected and re-used in Type A is a lime (hydraulic) mortar and the producer (Kimia ltd.) reports
construction, following multi-centennial old building practices. Stone in the data sheet a compressive strength >15 MPa. Mortar Type B is a
samples have not been tested at the laboratory, but available literature cement mortar with a compressive strength >25 MPa. Table 2 also
data from testing on this type of stone demonstrate high mechanical shows the test results from mechanical characterization for the two
properties: compressive strength 45–60 MPa, flexural strength 5–15 mortars used for coating. The thickness of each mortar coating was
MPa, weight density 2200–2400 kg/m3 [39]. 20–25 mm and 30–40 mm for brickwork and stone masonry walls,
A low-cement bedding mortar was used for the construction of stone respectively (Fig. 9).
and brickwork panels (mix design by weight: 80% sand, 15% lime, 5%
Portland cement). This was mechanically characterized in compression
and bending. In total 12 bending and 24 compression tests were con­ 2.4. Retrofitting method
ducted. 40x40x160 mm prisms were tested at the laboratory and the
following strength values were found: mean compressive strength 2.69 To reinforce the wall specimens, two types of fibrous meshes have
MPa (standard deviation 0.278 MPa), flexural strength 0.76 MPa been used. The mechanical properties of the reinforcement, the mesh
(standard deviation 0.184 MPa) [40–41]. The failure tests were initially geometry and the cement matrices used to bond them to the masonry
carried out in three-point bending (central point load) after 28 days from substrate are shown in Table 3, respectively. Because the main function
the date of casting, and subsequently the mortar prisms, broken in of the cement matrix is not only to encapsulate and protect the fibers,
flexure, were tested in compression. The choice of using a low-strength but also to facilitate the transfer, by shear, of the tensile stresses acting in
bedding mortar was motivated by the desire or reproducing a historic the masonry, the choice of the cementitious mortar is critical: this should
mortar, typically weak and deformable. be workable (easy to apply, able to penetrate and embed the fibres), non-
shrinkable, viscous enough to avoid drips due to its application on
2.3.2. Meshes vertical surfaces (i.e. walls), with a pot-life (working time) sufficiently
Two types of Alkali Resistant fiberglass textiles have been used to long to allow for multiple layers of reinforcement, if needed (Table 4).

3. Test results
Table 2
Bricks and mortars properties. Standard deviation in (). 3.1. Unreinforced specimens
Brick Dimensions (mm) 240x120x55

Compressive strength (MPa) 38.75 (3.57) Table 6 shows the values of the maximum shear (diagonal) force,
Flexural strength (MPa) 3.61 (0.45) Fmax, recorded during the testing for both unreinforced and reinforced
Bedding mortar Compressive strength (MPa) 2.69 (0.21) specimens. Load was applied in cycles of increasing magnitude up to
Flexural strength (MPa) 0.76 (0.184) failure (Fig. 10). The total number of the load cycles as well as the shear
Mortar coating Compressive strength (MPa) 9.83 (1.02)
Type A Flexural strength (MPa) 2.19 (0.39)
modulus are also given in this table (Table 7).
Mortar coating Compressive strength (MPa) 25.6 (2.72) The masonry shear strength τ 0 for each wall panel together with a
Type B Flexural strength (MPa) 5.71 (0.89) description of the masonry material, and type of reinforcement are also
shown. Unreinforced wall panels of both Series (brick (MAT) and stone

Fig. 6. (a) Used materials for wall construction: Solid bricks (240x120x55 mm); (b) Bonding pattern; (c) Uncut or barely-cut calcareous irregular stone.

5
G. Castori et al. Construction and Building Materials 299 (2021) 124276

30 mm
50 mm

(a) (b)
50 mm 35 mm

Fig. 7. The two types of fiberglass mesh used for reinforcement: (a) Type 1, (b) Type 2.

Fig. 8. Reinforcement materials and application: (a) 200x100 mm fiberglass L-shaped connectors, (b) the mesh reinforcement before the application of the second
layer of mortar coating.

(PIE) work) failed through shear cracking and accordingly developed a

diagonal crack along the bed and head joints (zig-zag pattern). This
failure denotes shear cracking at the mortar joints of the wallette which
Brickwork
afterwards begins to behave as a rigid body rotating and translating
wall,
about the diagonal shear crack, i.e. without further demolition of the thickness
specimen. Fig. 11 shows the shear stress vs. angular strain plots for 240 mm
unreinforced specimens: it should be noted that the maximum value of Ready-to-use
angular strain (0.77 and 1.12 ‰ for MAT-01-U and PIE-01-U, respec­ mortar coating,
thickness 25-30 mm
tively) is not a significant parameter for the analysis, given the method
of application of the diagonal load (manually, using a pump). This in­
duces a pseudo-ductility (i.e. a horizontal plateau of the stress–strain Low cement 100x200 mm
curve) with limited mechanical significance. The main parameters to bedding 8 mm diameter
mortar L-shaped
consider are the capacity load, which was similar for both panels (67.03
Fiberglass rod
and 73.8 kN), the shear modulus (1080 and 2092 MPa for MAT-01-U and
PIE-01-U, respectively) and the failure mode. The diagonal shear crack
Fiberglass
of the brickwork panel was made of horizontal and vertical alternating
mesh
segments along the mortar joints, while, for stone masonry panel, this
was more inclined given the irregular shape of the stone blocks.
Fig. 9. Detail of single-sided reinforcement (brickwork wall panel).
The strength and stiffness values of unreinforced specimens are
consistent with the typical mechanical characteristics found in the

6
G. Castori et al. Construction and Building Materials 299 (2021) 124276

Table 3 scientific literature for traditional historic masonry in Italy [42–44].

Two types of fiberglass mesh used for reinforcement: physical and mechanical
properties.
3.2. Reinforced specimens
Type 1 Type 2

Mesh Material Fiberglass, AR Fiberglass, AR Reinforced wall panels consisted of ten wall specimens. Four were
Weight density (dry fiber) (g/m2) 235 465
made of rubble stones and the remaining six of solid bricks. FRCM
Weight density (pre-preg) (g/m2) 335 581
Mesh size (mm) 50x50 35x30 reinforcement was applied to virgin, i.e. undamaged walls.
Mesh unit tensile strength (kN/ 63 110
m) 3.2.1. Single-sided FRCM reinforcement
Single Young’s modulus (GPa) 72 72 Single-sided reinforcement should be avoided, as it produces unde­
Cord
Elongation at failure (%) 3.5 1.5
sirable second order effects to the structural responses of the shear walls
Tensile strength (MPa) 727 1270 (asymmetry in the stress distribution on the wall cross sections, likely
out-of-plane deformations in shear walls). The double-sided reinforce­
AR = Alkali Resistant
ment also activates the so-called “sandwich-panel” behavior, with a

Table 4
Results of shear tests.
Designation Wall Thickness Single- or Double sided No. of loading Maximum diagonal Masonry Tensile Masonry Shear Shear Modulus
(mm) Reinforcement cycles Load Fmax (kN) Strength ft (MPa) Strength τ 0 (MPa) Gbil (MPa)

MAT-01-U 240 – 7 67.03 0.116 0.077 1080

MAT-02-D 291 Double 14 204.9 0.353 0.236 1490
MAT-03-S 260 Single 9 100.2 0.175 0.117 1052
MAT-04-D 297 Double 14 199.8 0.350 0.233 2512
MAT-05-S 270 Single 10 113.4 0.200 0.133 2703
MAT-06-S 267 Single 10 120.9 0.211 0.141 1624
MAT-07-S 265 Single 10 120.9 0.208 0.138 1834
PIE-01-U 245 – 7 73.80 0.126 0.084 2092
PIE-02-D 303 Double 13 182.3 0.309 0.206 2034
PIE-03-S 285 Single 11 136.1 0.233 0.155 954
PIE-04-D 307 Double 14 209.6 0.363 0.242 876
PIE-05-S 293 Single 11 138.2 0.240 0.160 1902

Table 5
Summary of the test results, in terms of δτ = Fmax,reinf/Fmax,URM and δG = Gbil,
reinf/Gbil,URM ratios.

Designation Material & Reinforcement Experimental NTC 2018

δτ δG δτ δG

PIE-U Stone, URM 1.0 1.0 1.0

PIE-D Stone, double-sided 2.675 0.695 2.0
PIE-S Stone, single-sided 1.88 0.685 –
MAT-U Brickwork, URM 1.0 1.0 1.0
MAT-D Brickwork, double-sided 3.035 1.94 1.5
MAT-S Brickwork, single-sided 1.712 1.667 –

Fig. 10. Example of the cycling loading protocol.

Table 6
Table 7
Material properties of masonry components and reinforcement matrix deter­
Predicted versus experimental versus load-carryng capacities.
mined via FEM simulation .
Designation Single- or Double Predicted Experimental Error of
Young’s Modulus Poisson ratio Shear
sided load load capacity the model
(MPa) (–) transfer
Reinforcement capacity Fmax,exp (kN) Fmax,FE/
coefficient
Fmax,FE (kN) Fmax,exp
(βt) (βc)
MAT-01-U – 65.69 67.03 0.98
Solid brick properties 9000 0.20 0.5 1.0 MAT-02-D Double (Type 1) 176.25 204.9 0.86
Bedding mortar properties 320 0.15 0.2 0.5 MAT-03-S Single (Type 1) 112.5 100.2 1.12
Reinforcement matrix (mortar) 9600 (Type A) 0.17 (Type 0.2 0.5 MAT-04-D Double (Type 2) 172.5 199.8 0.87
properties 19800 (Type B) A) MAT-05-S Single (Type 2) 112.5 113.4 0.99
0.21 (Type MAT-06-S Single (Type 1) 135.0 120.9 1.12
B) MAT-07-S Single (Type 2) 135.0 120.9 1.12

7
G. Castori et al. Construction and Building Materials 299 (2021) 124276

positive structural response against out-of-plane bending loads. How­ more sustained by the GFRP reinforcement. As a consequence, while
ever, in real applications, it is common that double-sided reinforcement unreinforced specimens exhibit an abrupt loss of shear stiffness after
cannot be applied, for example for the requirement to keep the stones or cracking, for single-sided reinforced ones this is smoother. This implies
the bricks visible (fair-face aspect), or for frequently encountered con­ that the bi-linear approximation method used for shear modulus
straints inside the buildings (partition walls, the desire of avoid use calculation is able to better “capture” the structural response of unre­
disruption, high costs of restoring i.e. re-plastering, re-painting of the inforced wall panels compared to reinforced ones. Fig. 15 shows this
walls, modifications of the electrical and water systems). As a conse­ limitation for PIE-03-S test.
quence, structural engineers are often forced to design single-sided Table 5 shows the increment in terms of lateral load capacity and
reinforcement. shear modulus for all reinforcement layout. This table also compares test
Figs. 12 and 13 show the shear stress-angular strain curves for stone results from the campaign with the suggested multiplication factors
and brickwork specimens. It can be noted that single-sided reinforce­ suggested by the Italian Seismic code (NTC 2018). This is actually the
ment (Fig. 14) produced an increase in the lateral load capacity of 88 only national building code where FRCM–reinforcement of pre-existing
and 71%, for stone and brickwork specimens. On opposite, for the shear masonry buildings is considered, using a simplified approach consisting
modulus a decrement in magnitude was recorded: hoverer this may in the application of a multiplication factor to be applied to the me­
depend on the method used to calculate this mechanical parameter chanical properties (compressive and shear strengths, Young’s and shear
(Fig. 4) and on the resisting mechanism of reinforced walls panels. moduli) of the unreinforced existing masonry. The Italian code only
During the initial loading phase, the masonry material, which has suggests double-sided reinforcement, which is clearly to be preferred to
bigger resisting sectional area and is stiffer that the reinforcement, is prevent second order effects resulting from a single-sided application (i.
able to resist to the diagonal load. However, by increasing the diagonal e. out-of-plane wall deflections, asymmetric stress distribution, etc.). In
load, the masonry starts cracking and the tensile stresses are more and real practice, it is common to apply a single-sided reinforcement.

Fig. 11. Unreinforced wall panels: (a) MAT-01-U, (b) PIE-01-U.

Fig. 12. Envelope curves for the brickwork panels: shear stress (τ) versus angular strain (γ).

8
G. Castori et al. Construction and Building Materials 299 (2021) 124276

Fig. 13. Envelope curves for the stone panels: shear stress (τ) versus angular strain (γ).

3.2.2. Double-sided FRCM reinforcement

The wall panels with high GFRP-to-masonry ratio (double-sided
reinforcement) exhibited the highest strength and stiffness increment,
compared to URM specimens, as shown in Figs. 12 and 13. After the
development of the diagonal shear cracks, the loading continued up to
an angular strain of about 0.002, and then unloaded for safety consid­
erations. Typically, double-sided reinforced specimens exhibited an ul­
timate strength 203 and 167 percent higher than that of the
unstrengthened brickwork and stone walls, respectively.
Furthermore, it can be noted that the data collected during the
experimental testing showed very good repeatability between identical
masonry specimens. All wall panels experienced shear failure, but this
was different from the one recorded for unreinforced specimens. Typical
diagonal shear cracks can be seen in Fig. 16: a large number of parallel,
45◦ inclined, diagonal cracks formed along the compressed wall diago­
nal, while for unreinforced specimens only one large diagonal crack
normally developed.
The Italian Seismic Code underpredicts the shear strength by a factor
Fig. 14. Single-sided retrofit.
of 1.3 to 2, with the largest error occurring for the brickwork wall
panels. However, this code adopts a very simplified design approach, not
considering, for example, the reinforcement ratio.

4. Shear behavior of FRCM–reinforced masonry: Numerical

approach

To validate the modeling method, an ad hoc numerically-based two-

step procedure has been used. A preliminary step (Step 1: Calibration
process) involved a series of FE analyses carried out at the scale of the
constituents (brick units, mortar, reinforcement matrix and FRCM
strengthening). This was necessary to define the material parameters
given the discontinuous nature of masonry. The results of Step 1 and the
experimental data (reported in Section 3) were used as input data for a
refined micromodeling of both unreinforced and reinforced brickwork
walls (Step 2: FE model). It is worth noting how these two conceptually
related steps must not be intended as a one-way process, but feedback to
the results of the FE analysis have to come from an iterative check of the
evidences emerged during the calibration process.

Fig. 15. PIE-03-S test: shear stress vs. angular strain plot.

9
G. Castori et al. Construction and Building Materials 299 (2021) 124276

Fig. 16. Failure modes: (a) Unreinforced brickwork panel, (b) cracks on the double-sided reinforced wall.

Fig. 17. Willam-Warnke yield surface: (a) meridian plane (let ξ the hydrostatic stress invariant); (b) deviatoric plane.

10
G. Castori et al. Construction and Building Materials 299 (2021) 124276

4.1. Calibration process (Step 1) ⎧

⎪
⎪
⎪
⎪
⎪
To achieve a reliable numerical reference, the first step of the pro­ ⎨
Eβt [
posed procedure was aimed at defining the resistance parameters of the σ ij = Dck
ijhk εhk ⇒ Dck
ijhk = δih δjk
⎪
⎪ 2(1 + ν)
⎪
constituent materials not derived directly from the experimental char­ ⎪
⎪
⎩
acterization. Several FE models were carried out, using the ANSYS FEM [ ]
package, to reproduce the main mechanical properties of both masonry ]
+δik δjh cracking direction
Eν
δij δhk
components (i.e. solid bricks and bedding mortar) and the reinforce­ (1 + ν)(1 − ν)
ment, made of glass fibers embedded into a cementitious matrix.
Brick units, mortar and reinforcement matrix were modelled as a E [ ]
meshed FE continuum of eight-noded 3D isoparametric elements of + δih δjk + δik δjh other directions (12)
2(1 + ν)
hexahedral shape (Solid 65), allowing the treatment of non-linear
behavior through the assumption of isotropic multilinear stress–strain
curves, modified according to the quality of the different materials [45]. where Dck
ijhk represents the constitutive matrix of the active cracks, i.e.,
The tensile and compression responses were studied by assuming a
the ones which were not completely closed.
linear softening in tension and a tri-linear curve, characterized by a bi-
Conversely, if the cracks are re-closed and consequently all
linear strain hardening followed by a residual plateau in the post-peak
compressive stresses perpendicular to the crack faces can be transmitted
range, in compression.
across the cracks, then the constitutive relationship becomes:
Prior to failure, the response is elastic and is governed by the
⎧
following equation: ⎪
⎪
⎪
[ ] ⎪
⎪
Eν E [ ] ⎨[ Eν
]
Eβc [
σij = Dijhk εhk ⇒ Dijhk = δij δhk + δih δjk + δik δjh ck ck
σ ij = Dijhk εhk ⇒ Dijhk = δij δhk + δih δjk
(1 + ν)(1 − 2ν) 2(1 + ν) ⎪
⎪ (1 + ν)(1 − 2 ν) 2(1 + ν)
⎪
⎪
(10) ⎪
⎩
[ ]
where the constitutive matrix (Dijhk) is characterized by two constants, ] Eν E [
+δik δjh cracking direction δij δhk + δih δjk
the Poisson ratio (υ) and Young’s modulus (E), whose values, due to the (1 + ν)(1 − 2ν) 2(1 + ν)
lack of experimental data (Section 2.3.1), were estimated via FEM
simulation (i.e. by extensive numerical trials and matching FE and ]
experimental results). Since it was not possible to use strain gauges to + δik δjh other directions
measure the strain, the calibration of the two constant values has
required sensitivity analyses.
The material response over the elastic range of both masonry com­ (13)
ponents (i.e. solid bricks and bedding mortar) and cementitious matrix In case of crushing, the stiffness of the corresponding element was
was then accomplished by accounting for both cracking and crushing assumed to be negligible along all principal directions, implicating that
failure modes through the use of the Rankine criterion combined with the element can be computationally omitted.
the Willam and Warnke (W-W) model [46–47]. The adopted yielding Numerical models were developed to simulate the set-up of both
surface (having a parabolic trace in the meridian plane (Fig. 17a) and an three-point bending (Fig. 18a) and uniaxial compressive (Fig. 18b) tests.
elliptical curve for interpolating the intermediate failure meridian be­ The predictive performances of the proposed FE models were subse­
tween the two extremes (tension and compression meridians) on the quently compared with the experimental results to obtain all those
deviatoric plane (Fig. 17b)), can be defined by two (rather than five) material properties not deriving directly from the experimental
material parameters: the uniaxial compressive (fc) and tensile (ft) characterization.
strength (experimentally measured). When the hydrostatic stress is The values of the elastic parameters (E and υ) and the shear reduction
limited by (√3)fc, the other three parameters (theoretically required) factors (βt and βc) were varied until the failure strength and the vertical
can be assumed as follows: and lateral displacements of the specimens, after each characterization
fcb = 1.2fc f1 = 1.45fc f2 = 1.725fc (11) test, matched the results obtained from the FE models. The calibrated
material properties used for the computational models (Step 2) are
where fcb is the biaxial compressive strength, and f1 and f2 represent the illustrated in Table 6.
compressive strength of a biaxial and uniaxial compression state Regarding the FRCM strengthening, among the various modeling
superimposed on the hydrostatic state of stress. This model is able to strategies adopted in literature (smeared or discrete representation), it
simulate the shear behavior using two additional shear reduction factors was decided to use a discrete element approach [48–49]. A rigorous
(βt and βc) for those subsequent loads that cause sliding across the crack method should involve the full modeling of the fiberglass textiles
face for open (βt) or re-closed (βc) cracks. Since it was not possible to embedded in the mortar coating, also considering their fragile behavior
experimentally measure these parameters, their values were evaluated and the possible slippage of the grid inside the matrix. As the compu­
via FEM simulation by means of sensitivity analyses1. tational difficulties would be more than obvious even for small scale
The yielding surface was completed using cut-off failure conditions. structures, a simplification of the model appeared to be necessary. Not
The stress-strain relations were modified using a plane of weakness taking into account any possible slippage phenomenon between mortar
normal to the crack face. For the cracking direction, both normal and matrix and yarns, the glass grid was discretized as a continuous bi-
shear stiffness were reduced, the tensile strength suddenly dropped to directional grid of non-linear truss elements (using two-noded tension-
zero and the constitutive law can be re-written as: only bar elements, LINK180) perfectly bonded to the masonry support.
This because, as confirmed experimentally (see section 4), FRCM com­
posite delamination on entire walls was unlike and failure was instead
driven by tensile rupture of the wires. As observed during full scale test,
in fact, once the tensile strength of the mortar matrix was reached,
1
several cracks started to open determining a progressive decrease of the
In [47], the recommended range was: 0 (complete loss of shear transfer, i.e.,
elastic stiffness, until the ultimate stress was reached. Being the matrix
smooth crack) ≤ βt ≤ βc ≤ 1 (no loss of shear transfer, i.e., rough crack)

11
G. Castori et al. Construction and Building Materials 299 (2021) 124276

Fig. 18. FE modeling of characterization test set-ups: a) three point bending test; b) compressive test.

Fig. 19. Finite element modeling of single lap shear test set-ups.

Fig. 20. Geometric configuration and discretization of brickwork masonry specimens.

12
G. Castori et al. Construction and Building Materials 299 (2021) 124276

Fig. 21. Comparison between experimental and numerical damage maps: (a) unstrengthened specimens; (b) strengthened specimens.

completely cracked, tensile actions transfer to the grid, which started to yarn),a FE model (Fig. 19) was also developed with the aim at repro­
be subjected to not negligible levels of stresses up to failure. ducing the double lap shear test set-up of a previous laboratory inves­
Adopting a shared node approach, the connection between the glass tigation2 to obtain the value of the missing material properties. The
grid and the mortar coating meshes was achieved treating the glass value adopted for fdeb,yarn in the FE analysis (fdeb,yarn = 622 MPa) was
yarns as a slave material merged to the surrounding master material found through an iterative trial-and-error process aimed at equating the
(mortar coating). It is worth noting that, as a consequence, the step of experimental and numerical values of the failure load and the rein­
the truss elements in the FEM was lower than the real grid spacing of the forcement strain at maximum load, respectively.
reinforcing system (ranging from 30 to 50 mm for Type 1 and Type 2,
respectively) and thus, to account for the equivalent amount of rein­
4.2. FE model (Step 2)
forcement, the cross-section area of the glass yarns was decreased
accordingly.
The second step was performed at a structural level (based on the
The glass yarns were assumed unable to withstand compressive
material relationships previously determined (Step 1)) into a detailed
stresses, in order to disregard unrealistic instability phenomena, and
micromodeling of both unreinforced and FRCM-reinforced brickwork
perfectly elastic in tension. Accordingly, prior to failure, the initial
specimens.
linear-elastic phase is governed by:
Instead of adopting a macro-modeling strategy based on the use of a
masonry continuum with homogenized properties, masonry was regar­
• the tangential stiffness: k = EglassAyarns
ded as a heterogeneous material, where blocks, joints and FRCM
strengthening were meshed separately. To reproduce the experimental
where Eglass is the Young’s modulus of the fiberglass fibers, whereas
cracking pattern, an approach based on the use of a 3D mesh was
Ayarn is the cross-section area of the yarns, whose values were deter­
preferred in order to take into account the out-of-plane deformations
mined experimentally
induced by the strengthening. The FE mesh of the panels periodically
repeating arrangement was adjusted so as to have at least three brick
• the failure strength: fu = min{fu,yarn; fdeb,yarn}
elements along the panel height, two elements along the thickness of
each bed joint (it is worth noting that a coarser mesh – a single element
where fu,yarn and fdeb,yarn are the tensile rupture and debonding
along the thickness – was used for the head joint) and again two ele­
strength of the reinforcing grid, respectively (whose values were again
ments across the mortar coating. This allowed to capture the more
determined experimentally
critical details avoiding distorted meshes and identification of shear lock
As for the post-peak behavior, it was assumed that, after the first yarn
reaches its ultimate strength (fu), the failure is instantaneous (brittle
mechanism) and the stress immediately drops off to zero. 2
Experimental tests were performed on single lap joints realized with a glass
Without characterization test results for the debonding strength (fdeb,
grid bonded along the center line of the front of a single clay brick.

13
G. Castori et al. Construction and Building Materials 299 (2021) 124276

effects. Fig. 20 shows the FE model: it consists of 165,619 elements and reinforcement on high-thickness (>500 mm) shear walls and on
107,064 nodes, with 320,044 degrees of freedom (DOF). different types of masonry typologies and stone/brick arrangements.
Force-controlled analyses were carried out on both unreinforced and Experimental study should be completed to evaluate the structural
strengthened panels. The load initially included the panel own weight, response of FRCM-reinforced triple-leaf shear walls. It would be inter­
and subsequently a monotonic load at the top corner, with the opposite esting to apply the Italian design provisions to a larger set of reinforced
corner fully constrained. masonry walls, also considering the corrective factors suggested by the
To investigate the reliability of the proposed modeling approach, the above mentioned standard.
predicted collapse mechanisms have been compared with those As for the numerical results, after calibrating the FE model with
observed experimentally, by checking the evolution of the damage characterization test data and then using the experimental data obtained
pattern at different load stages. In agreement with experimental evi­ from diagonal tension tests to validate the proposed two-step non-linear
dence, the numerical failure mode of unreinforced panels was charac­ numerical procedure, the authors were suggested and generally felt to
terized by a brittle failure concentrated close to the panel’s diagonal line have sufficient evidence base for the use of such a procedure in the
of symmetry, after the peak shear stress was reached. As shown in simulation of brick masonry reinforced with FRCM strengthening.
Fig. 21a, this failure mode was governed by the low strength of the Although it constitutes only a first step toward the complete under­
bedding mortar, which caused a high level of damage of both bed and standing of the macroscale structural behavior of the reinforced panels
head joints (zig-zag pattern) anticipating the brick failure as found and can clearly be optimized (e.g. model is not used in other studies (e.g.
experimentally. Also for reinforced walls, the FE models adeguately parametric) or in comparison with data obtained by other researchers),
simulated the strengthening effect of FRCM. Because of the partially when appropriately applied, the presented procedure (Steps 1 and 2)
continuous nature of the FRCM composite system (rather than discon­ represents a sound basis for selecting this kind of intervention strategy as
tinuous like in FRP strip reinforcement), the tensile stresses were uni­ well as controlling their efficiency.
formly redistributed on the entire surface allowing to preserve the global
integrity of the reinforced walls at relatively high level of the shear load Declaration of Competing Interest
(Fig. 21b). The effect of the GF (Glass Fibres) grid was to improve the
structural behavior of the panel in withstanding both compressions and The authors declare that they have no known competing financial
tensions, avoiding damage localization and thus inhibiting the brittle interests or personal relationships that could have appeared to influence
behavior observed for unreinforced specimens. the work reported in this paper.
A further validation of the proposed modeling strategy was obtained
by comparing the results of the numerical load-carrying capacity with
Acknowledgements
those obtained experimentally. A reasonable fitting was observed be­
tween numerical and experimental results (Tab. 7). The lower capacity
The authors are grateful for the support of Kimia Co., Perugia, Italy
results of the FE model (the deviations between the experimental and
for providing the reinforcement materials used in the experimental
predicted values was found to be no >14%.) for the double-sided
program. The authors would like to thank Alessio Molinari and Giordano
strengthened specimens can be attributed to the model strategy adop­
Bisciotti and the technical staff of the Structures Laboratory (Lastru) of
ted for the strengthening system (discrete element approach), which was
the University of Perugia, Italy for their help with this project.
unable to simulate properly the interface tensile stresses from masonry
to GF grid. On the contrary, the model did not fully capture the out-of-
plane deformations of the single-sided reinforcement, which can explain Funding
the higher load-carrying capacity of FE models (+12%) in comparison
with the experimental results (with the exception of MAT-05-S test, The authors acknowledge the support of Kimia Inc. (Perugia, Italy).
where a lower ultimate load was numerically observed).
References
5. Conclusions [1] G. Falcone, A. Mendicelli, F. Mori, S. Fabozzi, M. Moscatelli, G. Occhipinti,
E. Peronace, A simplified analysis of the total seismic hazard in Italy, Engineering
From the results of this experimental campaign on the shear response Geology 267 (2020) 105511, https://doi.org/10.1016/j.enggeo.2020.105511.
[2] P. Zimmaro, G. Scasserra, J.P. Stewart, T. Kishida, G. Tropeano, M. Castiglia,
of masonry wall panels reinforced with FRCM jacketing subjected to
P. Pelekis, Strong ground motion characteristics from 2016 Central Italy
static-cyclic lateral (in-plane) loading the following conclusions can be earthquake sequence, Earthquake Spectra 34 (4) (2018) 1611–1637.
drawn: [3] Italian Senate, Report No. 01077470 (2017) Earthquakes. Central Italy 2016,
Emilia 2012, L’Aquila 2009: reconstruction resources and laws, Rome, Italy.
[4] Dina F. D’Ayala, Sara Paganoni, Assessment and analysis of damage in L’Aquila
• A series of 12 masonry specimens were tested in this investigation. historic city centre after 6th April 2009, Bulletin of Earthquake Engineering 9 (1)
These specimens were reinforced to represent a potential rehabili­ (2011) 81–104.
tation method to enhance their ability to withstand an earthquake. [5] Kazuhiko Kawashima, Ömer Aydan, Takayoshi Aoki, Ichizo Kishimoto,
Kazuo Konagai, Tomoya Matsui, Joji Sakuta, Noriyuki Takahashi, Sven-
Single- and double sided FRCM reinforcement were used and Peter Teodori, Atsushi Yashima, Reconnaissance investigation on the damage of
experimented. the 2009 L’Aquila, Central Italy earthquake, Journal of Earthquake Engineering 14
• Test results have demonstrated that FRCM-reinforcement for both (6) (2010) 817–841.
[6] Luigi Sorrentino, Serena Cattari, Francesca da Porto, Guido Magenes,
stone and brickwork wallettes can cause a significant increment of Andrea Penna, Seismic behaviour of ordinary masonry buildings during the 2016
the lateral-load capacity. For single-sided reinforcement the lateral central Italy earthquakes, Bulletin of Earthquake Engineering 17 (10) (2019)
load capacity typically increased by a factor of 2, while up to 3 was 5583–5607.
[7] J. Gustavo, N. Galati, A. Nanni, Fiber-reinforced polymer strengthening of
noted for double-sided reinforcement. unreinforced masonry walls subject to out-of-plane loads, ACI Structural Journal
• Actually, the only design code providing instructions for the appli­ 100 (3) (2003) 321–329.
cation of FRCM reinforcement of historic rubble stone and brickwork [8] E. Hamed, O. Rabinovitch, Out-of-plane behavior of unreinforced masonry walls
strengthened with FRP strips, Composites Science and Technology 67 (3-4) (2007)
masonry is the recent Italian Seismic Code. However, test results
489–500.
have shown that the Italian design provisions are conservative for [9] J.I. Velazquez-Dimas, M.R. Ehsani, Modeling out-of-plane behavior of URM walls
FRCM-reinforced shear walls by a factor of 1.3 and 2 for stone and retrofitted with fiber composites, Journal of Composites for Construction 4 (4)
brickwork, respectively. (2000) 172–181.
[10] T.T. Bui, A. Limam, V. Sarhosis, M. Hjiaj, Discrete element modelling of the in-
plane and out-of-plane behaviour of dry-joint masonry wall constructions,
Future research should be conducted to evaluate the effect of FRCM Engineering Structures 136 (2017) 277–294.

14
G. Castori et al. Construction and Building Materials 299 (2021) 124276

[11] Marco Corradi, Antonio Borri, Giulio Castori, Romina Sisti, The Reticulatus method [29] S. Babaeidarabad, G. Loreto, D. Arboleda, A. Nanni, FRCM-strengthened CMU
for shear strengthening of fair-faced masonry, Bulletin of Earthquake Engineering masonry walls subjected to out-of-plane load, Mason Soc J 32 (1) (2014) 69–84.
14 (12) (2016) 3547–3571. [30] Susanna Casacci, Cristina Gentilini, Angelo Di Tommaso, Daniel V. Oliveira, Shear
[12] Claudio D’Ambra, Gian Piero Lignola, Andrea Prota, Elio Sacco, strengthening of masonry wallettes resorting to structural repointing and FRCM
Francesco Fabbrocino, Experimental performance of FRCM retrofit on out-of-plane composites, Construction and Building Materials 206 (2019) 19–34.
behaviour of clay brick walls, Composites Part B: Engineering 148 (2018) 198–206. [31] Catherine G. Papanicolaou, Thanasis C. Triantafillou, Kyriakos Karlos,
[13] T. D’Antino, F.G. Carozzi, P. Colombi, C. Poggi, Out-of-plane maximum resisting Myrto Papathanasiou, Textile-reinforced mortar (TRM) versus FRP as
bending moment of masonry walls strengthened with FRCM composites, strengthening material of URM walls: in-plane cyclic loading, Materials and
Composite Structures 202 (2018) 881–896. structures 40 (10) (2007) 1081–1097.
[14] P.B. Shing, M. Schuller, V.S. Hoskere, In-plane resistance of reinforced masonry [32] C. Buratti, E. Belloni, L. Lunghi, A. Borri, G. Castori, M. Corradi, Mechanical
shear walls, Journal of structural Engineering 116 (3) (1990) 619–640. characterization and thermal conductivity measurements using of a new’small hot-
[15] Anastasios Drougkas, Pere Roca, Climent Molins, Experimental analysis and box’apparatus: innovative insulating reinforced coatings analysis, Journal of
detailed micro-modeling of masonry walls subjected to in-plane shear, Engineering Building Engineering 7 (2016) 63–70.
Failure Analysis 95 (2019) 82–95. [33] Zuhair Al-Jaberi, John J. Myers, Mohamed A. ElGawady, Pseudo-static cyclic
[16] Garbin, E., Valluzzi, M.R., Modena, C., & Oliveira, D.V., & Lourenço, P.B. (2006). loading comparison of reinforced masonry walls strengthened with FRCM or NSM
Experimental investigation on the structural behaviour and strengthening of three- FRP, Construction and Building Materials 167 (2018) 482–495.
leaf stone masonry walls, In Proceedings of Structural Analysis of Historical [34] Francesca Giulia Carozzi, Alessandro Bellini, Tommaso D’Antino, Gianmarco de
Constructions – SAHC, New Delhi 2006. Felice, Francesco Focacci, Łukasz Hojdys, Luca Laghi, Emma Lanoye,
[17] M. Corradi, A. Borri, A. Vignoli, Experimental evaluation of in-plane shear Francesco Micelli, Matteo Panizza, Carlo Poggi, Experimental investigation of
behaviour of masonry walls retrofitted using conventional and innovative tensile and bond properties of Carbon-FRCM composites for strengthening masonry
methods, Masonry International 21 (1) (2008) 29. elements, Composites Part B: Engineering 128 (2017) 100–119.
[18] A. Borri, G. Castori, M. Corradi, E. Speranzini, Shear behavior of unreinforced and [35] Alessandro Bellini, Seyedmohammad Kahangi Shahreza, Claudio Mazzotti, Cyclic
reinforced masonry panels subjected to in situ diagonal compression tests, bond behavior of FRCM composites applied on masonry substrate, Composites Part
Construction and Building Materials 25 (12) (2011) 4403–4414. B: Engineering 169 (2019) 189–199.
[19] Sachin B. Kadam, Yogendra Singh, Bing Li, Strengthening of unreinforced masonry [36] ASTM, E. (2007). 519-07. Standard Test Methods for Diagonal Tension (Shear) in
using welded wire mesh and micro-concrete–Behaviour under in-plane action, Masonry,” American Society of Testing and Materials, West Conshohocken, PA.
Construction and Building Materials 54 (2014) 247–257. [37] RILEM. LUMB6 – Diagonal tensile strength tests of small wall specimens. Technical
[20] Hernan Santa-Maria, Pablo Alcaino, Repair of in-plane shear damaged masonry Report, RILEM; 1994.
walls with external FRP, Construction and Building Materials 25 (3) (2011) [38] BS EN 771-1 (2011) Specification for masonry units. Clay masonry units.
1172–1180. [39] M. Corradi, A. Borri, A. Vignoli, Experimental study on the determination of
[21] A. Gabor, A. Bennani, E. Jacquelin, F. Lebon, Modelling approaches of the in-plane strength of masonry walls, Construction and building materials 17 (5) (2003)
shear behaviour of unreinforced and FRP strengthened masonry panels, Composite 325–337.
structures 74 (3) (2006) 277–288. [40] ASTM C348 – 21, Standard Test Method for Flexural Strength of Hydraulic-Cement
[22] K.M.C. Konthesingha, M.J. Masia, R.B. Petersen, N. Mojsilovic, G. Simundic, A. Mortars.
W. Page, Static cyclic in-plane shear response of damaged masonry walls retrofitted [41] ASTM C349 – 18 Standard Test Method for Compressive Strength of Hydraulic-
with NSM FRP strips–An experimental evaluation, Engineering Structures 50 Cement Mortars (Using Portions of Prisms Broken in Flexure).
(2013) 126–136. [42] Italian Building Code (2018) Norme tecniche per le Costruzioni – NTC 2018, Rome,
[23] S. Saileysh Sivaraja, T.S. Thandavamoorthy, S. Vijayakumar, S. Italy (in Italian).
Moses Aranganathan, A.K. Dasarathy, Preservation of historical monumental [43] M. Corradi, A. Borri, A database of the structural behavior of masonry in shear,
structures using fibre reinforced polymer (FRP)-case studies, Procedia Engineering Bulletin of Earthquake Engineering 16 (9) (2018) 3905–3930.
54 (2013) 472–479. [44] S. Boschi, L. Galano, A. Vignoli, Mechanical characterisation of Tuscany masonry
[24] Venice Charter (1964) International Charter for the Conservation and Restoration typologies by in situ tests, Bulletin of Earthquake Engineering 17 (1) (2019)
of Monuments and sites, (ICOMOS). In Proceedings of the 2nd International 413–438.
Congress of Architects and Technicians oh Historical Monuments, Venice, Italy. [45] M. Corradi, G. Castori, A. Borri, Repairing brickwork panels using titanium rods
[25] ICOMOS–ISCARSAH Committee (2003) ICOMOS Charter—Principles for the embedded in the mortar joints, Constr Build Mater 206 (2020) 19–34.
analysis, conservation and structural restoration of architectural heritage. In [46] Willam KJ, & Warnke ED. Constitutive model for the triaxial behaviour of concrete.
Proceedings of the ICOMOS 14th General Assembly and Scientific Symposium, In Proceeding of the International Association for Bridge and Structural
Victoria Falls, Zimbabwe (Vol. 2731). Engineering, Bergamo, Italy; 1975.
[26] Borri, A., & Corradi, M. (2019). Architectural heritage: A discussion on [47] C.V. Uday Vyas, B.V. Venkatarama Reddy, Prediction of solid block masonry prism
conservation and safety. Heritage, 2(1), 631-647. compressive strength using FE model, Mat Struct 43 (2010) 719–735.
[27] F. Stazi, M. Giampaoli, M. Rossi, P. Munafò, Environmental ageing on GFRP [48] E. Bertolesi, G. Milani, C. Poggi, Simple holonomic homogenization model for the
pultruded joints: comparison between different adhesives, Composite Structures non-linear static analysis of in-plane loaded masonry walls strengthened with
133 (2015) 404–414. FRCM composites, Compos Struct 158 (2016) 291–307.
[28] S. Marouani, L. Curtil, P. Hamelin, Ageing of carbon/epoxy and carbon/vinylester [49] J. Scacco, B. Ghiassi, G. Milani, P.B. Lourenço, A fast modeling approach for
composites used in the reinforcement and/or the repair of civil engineering numerical analysis of unreinforced and FRCM reinforced masonry walls under out-
structures, Composites Part B: Engineering 43 (4) (2012) 2020–2030. of-plane loading, Compos B Eng 180 (2020), 107553.

15