Construction and Building Materials 53 (2014) 149–161

Contents lists available at ScienceDirect

Construction and Building Materials

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

Experimental assessment of in-plane behaviour of three-leaf stone

masonry walls
Bruno Silva ⇑, Massimo Dalla Benetta 1, Francesca da Porto 1, Claudio Modena 2
Department of Civil, Environmental and Architectural Engineering, University of Padova, Via Marzolo, 9, 35131 Padova, Italy

h i g h l i g h t s

 We carried out shear compression tests on 16 three-leaf stone masonry panels, before and after injecting NHL grout.
 Non-injected panels underwent external leaf separation at lower displacement levels.
 Injected walls presented enhanced behaviour and increased mechanical parameters.
 The use of scaled specimens may be considered representative of the tested masonry.

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

Article history: This paper presents an experimental campaign on multi-leaf stone masonry panels, scales 1:1 and 2:3, in
Received 4 April 2013 both original conditions and strengthened with grout injections. The panels were subjected to horizontal
Received in revised form 21 November 2013 in-plane cyclic loading combined with vertical loading for different pre-compression levels, and provided
Accepted 23 November 2013
important information on failure mechanisms, maximum displacement capacity, shear strength and
Available online 18 December 2013
other mechanical parameters, such as shear modulus and tensile strength. Further analysis provided
results on other parameters which mainly characterise the behaviour of three-leaf masonry under seis-
Keywords:
mic loads, i.e., stiffness degradation, energy dissipation, and viscous damping.
Stone masonry
Grout injection
The main aim of this study was to gather information on the static and dynamic behaviour of three-leaf
In-plane cyclic tests stone masonry structures in non-injected and injected conditions, in order to accurately characterise
Ductility their mechanical behaviour. In particular, the effectiveness of injections of hydraulic lime-based grout
Energy dissipation as a reinforcement technique was assessed and validated.
Viscous damping Ó 2013 Elsevier Ltd. All rights reserved.
Stiffness degradation
Reinforcement effectiveness

1. Introduction In order to predict the seismic resistance of masonry, study of

During an earthquake, a resistant masonry wall is subjected to
both vertical loads (due to gravity and the vertical component of
the seismic action) and horizontal loads as a consequence of
inertia-restoring forces. Multi-leaf stone masonry is particularly
susceptible to in-plane shear actions, due to its low tensile
strength. In addition, if the quality of the inner leaf with respect
to the external leaves is poor, and if there are no transversal
elements connecting the leaves, detachment and out-of-plane col-
lapse of external leaves very often occurs, as shown in Fig. 1.
⇑ Corresponding author. Tel.: +39 0498275355; fax: +39 0498275631.
E-mail addresses: bruno.silva@dicea.unipd.it (B. Silva), massimo.dallabenetta@
unipd.it (M. Dalla Benetta), francesca.daporto@unipd.it (F. da Porto), claudio.modena@
unipd.it (C. Modena).
1
Tel.: +39 0498275631; fax: +39 0498275631.
2
Tel.: +39 0498275613; fax: +39 0498275613.
its shear capacity is therefore necessary, as well as evaluation of
the effectiveness of grout injections, to prevent out-of-plane and
in-plane collapse mechanisms.
Grout injections have proved effective in improving the in- and
out-of-plane behaviour of multi-leaf stone masonry. Shaking table
tests on storey-high walls have recently also demonstrated the
effectiveness of injections in delaying leaf detachment under seis-
mic loads, significantly improving wall behaviour [1].
As regards behaviour under horizontal in-plane loads, Shear
Compression (SC) tests are typically used, to determine shear and
tensile strength, including shear modulus. Parameters such as duc-
tility, energy dissipation and stiffness degradation can also be eval-
uated by testing specimens under cyclic loading, in conditions
which buildings actually undergo during an earthquake. In this
technique, specimens are subjected to a constant vertical (axial)
load, simulating the pre-compression level acting on the building.
Cyclic lateral displacements are then applied at increasing

0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2013.11.084
150 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161

Nomenclature

A cross-sectional area of wall V vertical load

b shear stress distribution factor w width of stone masonry panels
E Young’s modulus a boundary condition parameter
Ehys dissipated energy d horizontal displacement measured at top of panels
Einp input energy du ultimate displacement
Ewc,0 elastic modulus of non-injected walls dcr displacement corresponding to cracking limit state
Ewc,s elastic modulus of injected walls df displacement corresponding to flexural cracking limit
fgr compressive strength of grout state
ft tensile strength dHmax displacement corresponding to maximum resistance
fwt,0 experimental tensile strength of non-injected walls limit state
fwt,s experimental tensile strength of injected walls D imposed displacement at top of wall
G shear modulus ci shear strain evaluated on masonry panel
Gexp shear modulus corresponding to cracking limit state qb apparent density
Gexp 30–60% experimentally obtained shear modulus, resulting qr real density
from average of values ranging from 30% to 60% of max- r00 vertical pre-compression
imum shear resistance rmax compressive strength of panels
Gk theoretical shear modulus s average nominal shear strength evaluated on panels
Gw,0 shear modulus of non-injected walls sHmax average nominal shear strength evaluated on panels at
Gw,s shear modulus of injected walls maximum resistance
h height of stone masonry panels su,0 shear strength of non-injected walls
H horizontal resistance su,s shear strength of injected walls
Hcr horizontal resistance corresponding to cracking limit l ductility
state l0 ductility of non-injected walls
Hdu horizontal resistance corresponding to ultimate dis- ls ductility of injected walls
placement limit state m Poisson ratio
Hf horizontal resistance corresponding to flexural cracking n equivalent viscous damping
limit state ncr equivalent viscous damping corresponding to cracking
Hmax horizontal resistance corresponding to maximum resis- limit state
tance limit state ndu equivalent viscous damping corresponding to ultimate
K secant stiffness displacement limit state
Kcr secant stiffness at cracking limit state nf equivalent viscous damping corresponding to flexural
l length of masonry panel cracking limit state
p0 open porosity nHmax equivalent viscous damping corresponding to maxi-
p total porosity mum resistance limit state
t thickness of stone masonry panels

amplitudes step-wise. Walls can be tested as cantilevers with cen-
tral vertical loading or as double fixed-end walls [2].
Since the early 1960s, several authors have carried out shear
compression tests on multi-leaf stone masonry, both in the
laboratory and on-site, before and after grout injection. The main
results of such experimental results are listed in Table 1. The
geometry, mechanical properties of the studied materials, and
the mechanical properties of the masonry before and after grouting
are shown.
Shear compression tests on injected and non-injected multi-leaf
stone masonry specimens show that diagonal cracks generally de-
velop in mortar joints, in some cases also passing through the
stone, particularly at higher pre-compression levels. With repeated
imposed lateral displacement, cracking becomes more extensive
and, at maximum lateral resistance, vertical cracks appear as an ef-
fect of compression [1,3]. Cracking also occurs in the transversal
sides, due to increased out-of-plane deformation of the external
leaves, reducing the resistant section of the compressed walls.
Experimental tests also show that the failure mechanisms of in-
jected masonry walls submitted to in-plane cyclic loading are
mainly governed by the slenderness ratio, pre-compression level,
and masonry bond.
As observed from the failure modes of masonry walls,
separation of external and internal leaves is due to shear failure
planes generated in the infill material, causing high horizontal
pressure on the external leaves [4]. The main cause of the im-

Fig. 1. Out-of-plane failure of stone masonry walls without transversal connections due to horizontal seismic actions, L’Aquila, Abruzzo, Italy.
Table 1
Mechanical parameters obtained in laboratory and in situ tests under shear–compression (the definition of the used symbols may be found in the symbols list).

Author Code Geometry Mechanical properties of masonry before and after grouting
Dimensions (cm3) fgr/ft (N/mm2) fwt,0 (N/mm2) fwt,s (N/mm2) r0 (N/mm2) su,0 (N/mm2) su,s (N/mm2) Ew,0 (N/mm2) Ew,s (N/mm2) K (kN/mm2) Gw,0 (N/mm2) Gw,s (N/mm2) l0 ls
Tomaževič and Sheppard [5] CAT I 100  60  265 32.5/1.9 0.02 0.12 – – – – – – 65 100 8.2 4.8
CAT II 150  100  50 32.5/1.9 0.04 – – – – – – – 85 – 11 –
150  100  50 32.5/1.9 0.04 – – – – – – – 85 – 9.0 –
CAT III 150  100  55 32.5/1.9 0.11 0.22 – – – – – – 73 125 4.7 3.1
150  100  55 32.5/1.9 0.12 0.22 – – – – – – 81 165 5.9 4.7
150  100  55 32.5/1.9 0.08 0.25 – – – – – – 94 – 5.8 –
150  105  55 32.5/1.9 0.08 0.29 – – – – – – 87 – 7.3 –
150  105  55 32.5/1.9 0.08 0.23 – – – – – – 78 173 6.0 5.4
150  105  55 32.5/1.9 0.12 0.27 – – – – – – 109 143 6.9 4.7
Tomaževič and Apih [3] A-1 100  50  100 32.5/1.9 – 0.30 – – – – – 64.69 – 160 – 2.6
A-2 100  50  100 32.5/1.9 – 0.30 – – – – – 83.85 – 200 – 3.4
B-1 100  50  100 19.7/1.6 – 0.36 – – – – – 48.85 – 117 – 2.2
B-2 100  50  100 19.7/1.6 – 0.37 – – – – – 52.31 – 122 – 2.4

B. Silva et al. / Construction and Building Materials 53 (2014) 149–161

C-1 100  50  100 6.8/0.6 – 0.20 – – – – – 53.85 – 137 – 2.6
C-2 100  50  100 6.8/0.6 – 0.42 – – – – – 57.98 – 142 – 2.5
D-1 100  50  100 12.8/1.7 – 0.33 – – – – – 57.25 – 148 – 3.1
D-2 100  50  100 12.8/1.7 – 0.39 – – – – – 60.89 – 148 – 3.1

Modena and Bettio [25] 1 – – 0.040 0.052 0.048 0.043 0.055 258.9 336.6 – 43.10 56.10 – –
– – 0.034 0.037 0.041 0.037 0.040 226.5 239.5 – 37.80 39.90 – –
2 – – 0.043 0.046 0.046 0.046 0.049 271.8 291.3 – 45.30 48.50 – –
– – 0.026 0.043 0.028 0.028 0.046 168.3 271.8 – 28.00 45.30 – –
3 – – 0.024 0.035 0.025 0.026 0.039 148.9 226.5 – 24.80 37.80 – –
– – 0.026 0.039 0.028 0.029 0.042 168.3 246 – 28.00 41.00 – –
4 – – – 0.066 – – 0.071 – 427.2 – – 71.20 – –
– – – 0.066 – – 0.071 – 427.2 – – 71.20 – –
5 – – 0.028 0.044 0.030 0.031 0.049 181.2 278.3 – 30.20 46.40 – –
– – 0.028 0.044 0.030 0.031 0.049 181.2 278.3 – 30.20 46.40 – –
6 – – – 0.052 – – 0.064 – 343 – – 57.2 – –
– – – 0.058 – – 0.059 – 394.8 – – 65.8 – –
Modena [26]a Pognana (E) – – – – – 0.114 0.237 – – – 102 268 5.7 2.3
Castelletto – – – – – 0.072 – – – – 36 – – –
Merizzo – – – – – 0.061 – – – – 74 – – –
Corradi et al. [27]a B-T-04-OR 90  180  48 – – – – 0.219 – 917 – – 546 – – –
V-T-07-IN 90  180  48 – – – – – 0.225 – 1814 – – 450 – –
P-T-15-OR 90  180  48 – – – – 0.172 – 471 – – 216 – – –
Corradi et al. [28]a TC-01-F 86  48  182 – – – 0.147 0.083 – 1298 – – 38 – – –
86  48  182 – – – 0.272 – 0.412 – 4153 – – 281 – –
TC-02-F 86.3  48  180 – – – 0.184 0.089 – 306 – – 65 – – –
86.3  48  180 – – – 0.268 – 0.196 – 1770 – – 196 – –
Galasco et al. [29] CS00 – – 0.200 – 0.2 – – – – – – – – –
CS01 – – 0.160 – 0.5 – – – – – – – – –
CS02 – – 0.100 – 0.2 – – – – – – – – –
CT01 – – 0.130 – 0.5 – – – – – – – – –
CT02 – – 0.100 – 0.2 – – – – – – – – –
Mazzon [1] R2 146  122  32 12.8/3.8 – 0.10 1.0 – 0.40 – 4057 – – 383 – –
R4 91  123  32 12.8/3.8 – 0.12 1.0 – 0.28 – 2738 – – 170 – –
S4 93  138  32 12.8/3.8 – 0.14 1.0 – 0.28 – 5513 – – 140 – –
S2 145  137  32 12.8/3.8 – 0.13 2.0 – 0.53 – 6708 – – 449 – –
S5 92  127  33 12.8/3.8 – 0.17 2.0 – 0.42 – 4640 – – 249 – –
R5 93  138  33 12.8/3.8 – 0.17 2.0 – 0.40 – 4323 – – 281 – –

151
a
In situ experimental investigation.
152 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161

proved behaviour of grouted walls is the increased shear bond
strength of the grout interface. Higher stress levels may therefore
be achieved by greater bond strength between internal and exter-
nal leaves [4].
Although few experimental studies have been carried out so far
on the tensile and shear strength of grouted multi-leaf stone ma-
sonry, available data show that, after grouting, strength is en-
hanced due to homogenisation.
The mechanical parameters shown in Table 1 show that shear
strength values between 0.026 N/mm2 and 0.219 N/mm2 for non-
injected masonry and between 0.040 N/mm2 and 0.53 N/mm2 for
injected masonry can be achieved, considering the various levels
of pre-compression. Injected specimens have shear strength on
average 1.5 times higher than that of non-injected ones.
The shear modulus (G) of non-injected masonry specimens
ranges between 24.8 N/mm2 and 546 N/mm2 and, for injected
ones; the range is 37.8–450 N/mm2, on average 2.1 times
higher. The compressive strength of the grout (fgr) clearly does
not increase the strength of the wall proportionally. Shear
compression tests with high-strength (cement-based) and
medium/low-strength (lime-based) grout shows no correlation
between the fgr/ft ratio of the grout with increased masonry
strength. The shear compression tests of [5] in specimens
injected with cement grout even showed that ductility (l)
decreases on average by 59.0% of that of non-injected
specimens.
Lastly, it should be noted that experimental results are few and
do not allow totally reliable predictions of grouting efficiency.
According to these observations, a series of experiments was set
up in order to: (i) test low-strength grout; (ii) obtain a complete set
of data on this type of material; (iii) carry out comparisons of 1:1
and 2:3 specimens, and (iv) obtain constitutive laws.
2. Experimental programme
2.1. Test programme and specimens
Before characterising injected masonry walls under in-plane horizontal loads,
behaviour in compression, under both monotonic and cyclic loads, was studied
on specimens of the same type. The main experimental mechanical parameters
resulting from simple compression tests are listed in Table 2. The complete results
of this preliminary characterisation, including experimental and numerical analy-
ses, are available in literature [6,7].
For the present experimental research, two series of stone masonry walls (F in
full scale 1:1, S in reduced scale 2:3) were constructed, see Fig. 2. According to expe-
rience with previous experimental work [1,8], the walls were constructed as a
whole and were later subdivided into individual panels for testing.
Specimens were constructed according to a technique representing one of the
most widespread structural systems applied to minor historical buildings through-
out Europe. This type of masonry is characterised by many voids in the inner leaf.
Panels were sized with a percentage of voids and a thickness ratio so that the effects
of injections were clear-cut, but were at the same time sufficiently representative of
real walls. The walls were constructed without any type of transversal connection
between the leaves, i.e., the most unfavourable situation.
The specimens were composed of two external leaves, each about 18 cm thick
(1:1 scale) or 12 cm (2:3 reduced scale). They consisted of rough-shaped limestone
blocks not more than about 25 cm long, arranged in sub-horizontal courses with
mortar joint thickness varying from 1 to 7 cm (1:1 scale) and 1 to 4 cm (2:3 scale).
The internal core, about 14 cm (1:1) or 9 cm (2:3) thick, was made with limestone
fragments poured into non-compacted layers between the two external leaves.
The materials chosen to construct the specimens combined the need to repro-
duce the conditions of historical walls (e.g., chemical and mechanical characteristics
of mortar, composition of inner cores, etc.) and to use easily available local materi-
als. Three types of calcareous stone (red stone, regular white stone, and irregular
white stone) were taken from the Cugnano quarry in Belluno, north-eastern Italy.
The lime mortar (T30V) was composed of a binder of natural hydraulic lime and
lime putty (ratio 1:3), with a lime/sand ratio of 1:3 and a water/binder ratio of 0.35.
2.2. Strengthening
Half of the three-leaf panels were strengthened with injections of the commer-
cial FEN-X/B natural hydraulic lime grout [9], see Fig. 3, a special, high-fluidity mix
with an exclusive base of Fenix NHL5, sulphate-resistant with a low water-soluble
salt content, designed for consolidating stone masonry structures. Injections were
carried out after mortar curing. First, in order to optimise the distribution of injec-
tion points and consequently of the mixture in the cores, a pre-defined triangular
mesh of equilateral triangles with 30-cm sides were drawn, allowing dense distri-
bution of holes so that the mixture could reach all points of the inner layers. Only
one of the sides was injected; on the opposite side, control holes were drilled
according to another similar mesh with 60 cm sides.
The injection procedure was quite fast and easy to perform and, including pre-
liminary preparation of walls and sealing of holes, lasted a total of about 2 h to cov-
er 8 m2. The average quantity of grout injected was 82 lt/m2 for 1:1 scale walls and
47 lt/m2 for 2:3 walls. After the walls were cut, visual inspection confirmed that al-
most all the voids and cavities had been filled, and revealed good bonding between
injection material and existing mortar.
After construction, curing and strengthening, the walls were sawn into 16 pan-
els and transported to the Building Materials Experimental Laboratory, University
of Padova, for testing.
2.3. Material properties
The whole experimental set-up was preceded by a preliminary phase, consist-
ing of: (a) characterisation of the mechanical properties of stone, mortar and in-
jected material; (b) examination of rheological and physical properties (fluidity
and stability) of grout and mechanical characteristics of composite elements (cylin-
ders) made of stone fragments (full scale 1:1, reduced scale 2:3) injected with grout,
all of which are presented in Table 3.
The results clearly showed that the grout presented greater strength than the
mortar. However, physical compatibility between the two materials was observed
by Young’s modulus, which was similar on both cases.

Table 2
Panels tested in shear compression (the definition of the used symbols may be found in the symbols list).

Specimens Condition Scale Slenderness Dimension (t  l  h) (mm) rmaxa (N/mm2) Ea (N/mm2) r00 (N/mm2) r00 /rmax (N/mm2)
SCF1.0NI Non-injected (NI) 1:1 1.2 100  50  120 2.49 2531 1.00 40%
SCF1.25NI 1.25 50%
SCF0.75NI 0.75 30%
SCF0.5NI 0.50 20%
SCF1.0I Injected (I) 4.89 5203 1.00 25%
SCF1.5I 1.50 35%
SCF1.25I 1.25 30%
SCF2.0I 2.00 50%
SCS0.5NI Non-injected (NI) 2:3 1.25 80  33  100 2.41 2392 0.50 20%
SCS1.0NI 1.00 40%
SCS0.75NI 0.75 30%
SCS1.25NI 1.25 50%
SCS1.25I Injected (I) 4.28 5125 1.25 25%
SCS1.0I 1.00 20%
SCS1.5I 1.50 30%
SCS2.0I 2.00 40%
a
This values correspond to average values resulting from simple compression tests performed on the same typology of walls and tested within the same experimental
campaign, [19].
 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161 153

Wall C - 8 panels in scale 1:1 Wall E - 8 panels in scale 2:3

Fig. 2. Stone masonry walls.

Fig. 3. Consolidation procedure. (a) Drilling holes for grout injection. (b) Cleaning drilled holes with compressed air. (c) Grout injection. (d) Sealing holes with mortar.

As regards the rheological properties of the grout, although the efflux values the wall was measured with a magnetostrictive absolute position sensor
measured according to ASTM [10] and Marsh cones [11] were very high compared (200 mm), which was also used for retro-activation of the actuators. Fig. 5 shows
with the norm, grout fluidity during injection was satisfactory. The procedure also the instrumental scheme used for the various types of specimens.
showed good stability, with little bleeding and segregation. Visual observations Pre-compression levels for all walls (0.5, 0.75, 1.0 and 1.25 N/mm2 for non-in-
showed that the grout satisfied the requirements of injectability into cylinders, with jected specimens; 1.0, 1.25, 1.5 and 2.0 N/mm2 for injected specimens, for both
good interactions between grout and stone and an effective injection procedure scales 1:1 and 2:3) were defined according to the failure fields corresponding to
with regard to void filling. the most typical failure modes (shear, compression and flexural failure mecha-
Comparison of the mechanical properties of the materials with results from cyl- nisms) for these kinds of structural elements, subjected to both compression and
inders showed that the overall behaviour of masonry depends on the weakest com- in-plane horizontal loads, see Table 2. Failure fields were computed by theoretical
ponent material – in this case, mortar – due to the high mechanical characteristics analysis based on Eqs. (1)–(3), which are normally applied for strength prediction
of the stone and the lower values obtained from the injected cylinders. However, of brick masonry walls. Further calibration was performed for three-leaf stone ma-
the resistance of composite materials is much lower than that of grout. sonry, e.g., the formulas predicting the shear failure mechanism proposed by Turnš-
ek and Cacovic [14] and refined by Turnšek and Sheppard [15]:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2.4. Test set-up, instrumentation and procedure ft r00
H ¼lt 1þ shear failure mechanism ð1Þ
b ft
Shear compression tests were performed on 16 panels in various conditions
(non-injected – NI, injected – I), scales and pre-compression levels. Specimens were  
tested in a cantilever-type boundary condition, with a fixed base. A steel track al- r00 rmax
H ¼lt 1  6h compression failure mechanism ð2Þ
lowed the free displacement of the upper part of the specimen, and two vertical rmax l
actuators were connected to the sledge, sliding on a frictionless linear guide, fixed
at the top of a horizontal steel beam for uniform vertical load distribution. Horizon-
"  2 #
r00 r00 rmax
tal displacement was applied by a third actuator, fastened to a stiffening steel struc- H ¼lt   flexural failure mechanism ð3Þ
rmax rmax 2h
ture, and connected to the panel with a forked beam [12]. The hydraulic and control l

systems were those developed by da Porto [13]. Fig. 4a shows a general view of the
where H is the wall’s horizontal resistance, rmax is the compressive strength of pan-
test rig and Fig. 4b the control and acquisition system.
els, r00 is the average compression stress due to the vertical load, t is the thickness of
The specimens were instrumented with 32 potentiometric displacement trans-
stone masonry panels and b is the shear distribution factor, defined as follows:
ducers, four to measure transversal deformation and monitor leaf separation, 24 to (
 h
measure in-plane flexural and shear deformations, and four to measure base uplift h P1 1 if l
61
and relative sliding between wall and base and also between masonry panel and b¼ b¼ ð4Þ
l 6 1:5 1:5 if h
l
P 1:5
reinforced concrete beams (top and bottom). Lateral displacement at the top of

Table 3
Average values of the mechanical characteristics of the materials and composite elements (the definition of the used symbols may be found in the symbols list).

Material rmax (N/mm2) ft (N/mm2) E (kN/mm2) m (–) qb (kg/m3) qr (kg/m3) p0 (%) p (%)
a
Stones Red Stone 93.4 17.3 54.5 – 2665 2441 1.3 8.3
Regular white 163.8 16.1a 68.0 – 2657 2405 1.4 9.1
Irregular white 189.9 29.6a 80.1 – 2642 3104 1.2 15.0
Mortar T30V 3.8 1.4a 4.7 0.27 – – – –
Grout FEN-X/B 12.5 2.8a 7.5 0.29 1871 – – –
Cylinders (stone fragments and grout) (1:1) 3.4 0.5b 4.5 – 2263 – – –
(2:3) 3.5 0.5b 3.6 – 2216 – – –
a
Tensile strength evaluated by flexural tests.
b
Tensile strength evaluated by indirect tensile (Brazilian) tests.
154 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161

Fig. 4. Shear–compression test. (a) Test rig. (b) Control and acquisition system.

Fig. 5. Instrumentation set-up for shear compression tests.

where h and l are the height and length of the masonry wall [16,17]. 20
After the preliminary phase of pre-compression, cyclic tests were carried out
under displacement control, by application of a horizontal displacement history 15
with increasing amplitudes and peaks, repeated three times for each displacement
amplitude. The amplitudes were imposed at a frequency of 0.005 Hz, see Fig. 6. All 10
tests were carried out up to failure of specimens, when a null value of lateral resis-
Displacement [mm]

tance was achieved, allowing the post-peak phase to be examined. 5

0
3. Experimental results 0 2000 4000 6000 8000 10000 12000 14000
-5
3.1. Failure modes
-10
During in-plane cyclic testing, the specimens exhibited various
types of overall behaviour, according to their condition (injected/ -15

non-injected), panel scale and applied vertical load. Four limit

states, which can be used to idealise the behaviour of masonry
panels [18], were observed, corresponding to changes in the way
specimens resist the progressive increase of applied lateral
displacement.
 The first limit state is due to the opening of the first flexural
cracks (Hf, df). Flexural cracking limit is defined by the appear-
ance of horizontal cracks on the first mortar bed joint between
specimen and lower concrete beam, or at the second mortar bed
joint. This phenomenon occurred in all specimens at
displacements of ±0.50–2.50 mm for non-injected panels and
±1.50–4.00 mm for injected ones, depending on applied pre-
compression and scale.
 After the opening of the first cracks, the overall behaviour of
each panel differed according to its condition, scale and pre-
compression level. The wall reaches the cracking limit state –
Hcr, dcr. At this point, two different main behaviours were
observed. One was the appearance of the first significant diago-
nally oriented shear crack, due to shear mechanisms; the other
was a crack pattern due to rocking, with the development of
-20
Time [s]
Fig. 6. Example of applied displacement history.
sub-vertical cracks in the compressed toes due to bending. Both
behaviours were observed in both scale panels at displacements
of ±1.0–2.5 mm for non-injected panels and ±2.5–7.0 mm for
injected ones.
 Subsequently, loads were increased gradually until maximum
load (Hmax) and the relevant displacement (dHmax) were
reached, which represents the third limit state.
 The last phase corresponds to the ultimate limit state, at which
specimens attain maximum displacement du, corresponding to
a value of residual lateral resistance Hdu.
Fig. 7 shows four specimens (SCF1.0NI, SCF1.0I, SCS1.0NI, SCS1.0I)
tested under compressive stresses of 1.0 N/mm2, and Fig. 8 shows
the corresponding lateral force vs. displacement hysteresis curves
(the whole series of crack patterns and hysteresis curves are found
in literature [19]).
 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161
In all cases, the first cracks occurred at the interface between
mortar and stone in the lower part of the specimens. However, pat-
terns differed in each case, according to the condition of the wall
(injected/non-injected), scale (1:1, 2:3) and applied pre-compres-
sion level. Cracking appeared earlier on the specimens tested with
higher vertical loads.
Non-injected panels 1:1, specimens SCF1.0NI, SCF1.25NI and
SCF0.75I, developed mixed flexural/shear behaviour. However, due
to the pre-compression levels, SCF1.25NI collapsed due to crushing
at the compressed toe. SCF0.5NI displayed dominant flexural
behaviour. All injected specimens in scale 1:1 exhibited rocking.
SCF2.0I, which was the specimen tested at the highest
pre-compression, developed diagonal cracking, followed by com-
pression failure.
Non-injected panels in scale 2:3 generally showed behaviour
governed by flexural failure. SCS1.0NI was already damaged before
it was tested. Injected 2:3 panels were influenced by rocking, ex-
cept for SCS2.0I with a higher pre-compression level, which showed
clear-cut flexural behaviour and collapse influenced by crushing, as
wide transversal cracks appeared after maximum resistance had
been reached, with cracking of the inner core.
In all cases, non-injected panels showed leaf separation at lower
displacements. Non-injected specimens failed due to buckling of
the external leaves, whereas injected specimens showed extensive
transversal cracking of cores after reaching maximum resistance,
confirming panel homogeneity due to grout injections.
3.2. Horizontal load and displacement
Fig. 9 compares limit states and idealised envelope curves of all
tested specimens, and Fig. 10 summarises the average characteris-
tic values of lateral force (H) and displacement (d) at the relative
limit states in the various masonry types.
The injected panels, in both scales, developed the first crack at
almost double displacement, twice the lateral resistance, when
compared with the non-injected panels. Non-injected 2:3 panels
Fig. 7. Crack patterns. (a) Specimen SCF1.0NI at failure. (b) Specimen SCF1.0I at maximum resistance. (c) Specimen SCS1.0NI at failure. (d) Specimen SCS1.0I after maximum
resistance.
155
reached higher displacement values at lower lateral forces with re-
spect to the 1:1 panels, and injected panels showed their first
cracks at almost the same displacement, with higher lateral forces
for 1:1.
At maximum resistance, injected panels in both scales showed
three times higher displacements at almost 2 times higher force
than non-injected ones. Comparisons of 1:1 and 2:3 panels showed
that non-injected 1:1 specimens showed maximum resistance
which was 67% higher than in the 2:3 panels, with almost the same
displacement. Instead, the injected 1:1 panels showed 20% higher
displacement values than the 2:3 panels, and also a 100% increase
in maximum lateral resistance, defining the rocking mechanism of
the 1:1 panels compared with the flexural mechanism of the 2:3
specimens with higher pre-compression loads.
During the final part of the tests, at the ultimate displacement
limit state, similar behaviour was observed in all panels, their
mechanical characteristics deteriorated, and wide cracks appeared
on transversal sides. Therefore, the different failure mechanisms
which developed during the first part of the test were always af-
fected by compression. This was most obvious in specimens sub-
jected to higher pre-compression levels.
In general, specimen comparisons, in both scales and with the
same pre-compression levels, showed that, the higher the vertical
load applied, the lower the resulting displacement, particularly
for injected elements. At maximum resistance, the difference be-
tween non-injected and injected 1:1 specimens, subjected to the
same vertical load (r0), was that the injected ones exhibited 5–6
times more displacement at an average of 40% higher maximum
resistance values. The same phenomenon was observed, with
smaller percentage differences, in the 2:3 panels, 3 times higher
displacement at an average of 27% higher maximum resistance
values. The overall behaviour of the walls was also analysed by
comparing the envelopes of hysteresis loops for each panel. As
a general rule, the non-injected panels in both scales were char-
acterised by lower horizontal forces and maximum displacement
values, 1:1 specimens exhibiting larger values in terms of initial
156 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161

200
(a) 200
(b)
150 150

100 100

50 50

H [kN]
H [kN]

0 0
-12 -8 -4 0 4 8 12 -30 -20 -10 0 10 20 30
-50 -50

-100 -100

-150 -150

-200 -200

δ [mm] δ [mm]

100 100
(c) (d)
75 75

50 50

25 25
H [kN]
H [kN]

0 0
-12 -8 -4 0 4 8 12 -30 -20 -10 0 10 20 30
-25 -25

-50 -50

-75 -75

-100 -100
δ [mm] δ [mm]

Fig. 8. Lateral force (H) vs. displacement (d) hysteretic curves. (a) SCF1.0NI. (b) SCF1.0I. (c) SCS1.0NI. (d) SCS1.0I.

Fig. 9. Comparison of limit state envelope curves for (a) full scale specimens (1:1) and (b) reduced scale specimens (2:3).

stiffness, maximum lateral resistance and maximum displace-
ment. The higher pre-compressed specimens in both scales
achieved higher maximum resistance with more brittle failure
in the final phase.
The degradation of lateral resistance and the displacement
capacity of the panels provided important information to evaluate
changes in overall behaviour. Indicators of resistance and displace-
ment capacity were calculated as the ratios of the considered mag-
nitude at two different limit states. Table 4 shows ratios Hcr/Hmax,
Hdu/Hmax and Hdu/Hcr for acting horizontal force and ratios dcr/
dHmax, du/dHmax and du/dcr for deformation capacity.
Considering first the degradation of lateral resistance of all
specimens, ratio Hcr/Hmax was almost the same for all panels, and
the cracking limit state always occurred between 87% and 95% of
the resistance force (cracking of damaged panel SCS0.5NI occurred
at 55%). This means that, when the collapse mechanism of each
specimen manifested, the maximum lateral resistance of the wall
was not achieved, although residual resistance was very low. All
specimens also exhibited a similar decrease after attaining maxi-
mum horizontal force. The second ratio, Hdu/Hmax, emphasises an
average reduction of 13% for 1:1 non-injected panels and 10% for
injected ones, and 20% (without SCS0.5NI) for 2:3 non-injected
 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161 157

Fig. 10. Average values of lateral force (H) and displacement (d) at limit states, (the definition of the used symbols may be found in the symbols list).

panels and 8% for injected ones, of the force measured when the
specimen failed. It is worth noting that SCS1.25NI showed a reduc-
tion of 36% of its strength. Lastly, the third resistance indicator, ra-
tio Hdu/Hcr, was closer to one in all specimens, indicating that the
cracking limit state occurs at approximately the same level of
resistance as the ultimate limit state.
The results from the displacement capacity indicators showed
different overall behaviour from those of the resistance indicators.
A great difference was observed with respect to the values ob-
tained from specimens subjected to different vertical loads.
Considering the first displacement capacity indicator, ratio dcr/
dHmax, the average values obtained for the various types of walls
were 0.48 and 0.38 for 1:1 non-injected and injected panels,
respectively, and 0.47 and 0.43, respectively for 2:3 panels (with-
out SCS0.5NI), non-injected and injected. These results underline
the fact that non-injected panels in both scales reached the same
average displacements in order to attain maximum resistance.
However, for the 1:1 injected elements, greater displacement
was necessary with respect to the 2:3 injected ones (11%).
Although the difference between load at cracking limit and at
maximum resistance was restricted in all types of walls, greater dis-
placements were obtained for the injected specimens (in both scales).
The average increase in displacement for the 1:1 panels was 21%, but
only 15% in 2:3 panels. Also with different vertical loads, these results
indicate the lower displacement needed for the specimens under
higher vertical loads, to reach maximum resistance.
Further significant information came from ratio du/dHmax. As ex-
pected [20], a lower pre-compression level allows greater displace-
ment before the wall collapses. However, the injected panels, in both
scales, reached greater displacements at higher vertical loads before
collapse. This phenomenon was due to the rocking and flexural
mechanisms of the specimens. In general, non-injected specimens,
in both scales, had higher ratios than injected ones (average values:
1.54 and 1.20 for 1:1 non-injected and injected, and 1.48 and 1.39 for
2:3 non-injected and injected – without SCS0.5NI). The percentage
decrease in the displacement before collapse between non-injected
and injected specimens was 20% for 1:1 scale, and 6% for 2:3.
Additional information came from ductility indicator du/dcr, re-
lated to the failure mechanism. In the non-injected specimens in
scale 1:1 (exhibiting shear failure), high values of ultimate ductility
were obtained, due to low displacements at shear cracking limit
dcr. Note, however, the lower value of SCF1.25NI compared with
the others, due to the more brittle collapse of the specimen after
higher vertical loads. In the case of injected specimens 1:1,
governed by a rocking mechanism, those subjected to lower verti-
cal stress had a greater du/dcr ratio, since the low vertical loads

Table 4
Resistance indicators and displacement capacity indicators for all the tested specimens (the definition of the used symbols may be found in the symbols list).

Specimens r00 (N/mm2) Hcr/Hmax Hdu/Hmax Hdu/Hcr dcr/dHmax du/dHmax du/dcr

SCF1.0NI 1.0 0.91 0.89 0.97 0.46 1.66 3.61

SCF1.25NI 1.25 0.89 0.92 1.04 0.60 1.16 1.93
SCF0.75NI 0.75 0.88 0.81 0.92 0.46 1.74 3.83
SCF0.5NI 0.5 0.93 0.86 0.92 0.39 1.60 4.10
Average 0.90 0.87 0.96 0.48 1.54 3.37
SCF1.0I 1.0 0.87 0.79 0.90 0.33 1.09 3.34
SCF1.5I 1.5 0.93 0.99 1.06 0.46 1.22 2.65
SCF1.25I 1.25 0.89 0.94 1.06 0.26 1.21 4.71
SCF2.0I 2.0 0.95 0.89 0.94 0.49 1.28 2.60
Average 0.91 0.90 0.99 0.38 1.20 3.32
SCS0.5NI 0.5 0.55 1.35 2.46 1.48 0.88 0.60
SCS1.0NI 1.0 0.92 0.87 0.95 0.44 1.31 3.00
SCS0.75NI 0.75 0.92 0.89 0.97 0.46 1.39 3.00
SCS1.25NI 1.25 0.93 0.64 0.69 0.53 1.74 3.32
Averagea 0.92 0.80 0.87 0.47 1.48 3.11
SCS1.25I 1.25 0.91 0.94 1.03 0.38 1.56 4.11
SCS1.0I 1.0 0.90 0.89 0.99 0.26 1.17 4.56
SCS1.5I 1.5 0.91 0.94 1.03 0.49 1.57 3.20
SCS2.0I 2.0 0.93 0.92 1.00 0.60 1.26 2.11
Average 0.91 0.92 1.01 0.43 1.39 3.50
a
Average value not including the damaged panel SCS0.5NI.
158 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161
applied allowed rocking in SCF1.0I and SCF1.25I. Instead, those sub-
jected to higher stresses – SCF1.5I and SCF2.0I – developed a smaller
du/dcr ratio.
The injected specimens in scale 2:3 showed the same phenom-
enon: SCS1.25I and SCS1.0I, subjected to lower vertical stress, had
higher ratios, allowing rocking to take place, whereas SCS1.5I and
SCS2.0I, with higher vertical stress, had lower ratios. Comparisons
of the ductility of the non-injected and injected panels, in both
scales, revealed a decrease of 1.5% for 1:1 injected panels (3.37
for non-injected and 3.32 for injected) but an increase of 12.5%
for the 1:1 panels (3.11 for non-injected and 3.50 for injected).
3.3. Stiffness degradation, energy dissipation and equivalent viscous
damping
The seismic response of buildings is related not only to the
strength and displacement capacity, i.e., ductility, of their struc-
tural elements, but also to the typical parameters of cyclic behav-
iour, such as stiffness degradation, energy dissipation capacity, and
the viscous damping coefficient.
Fig. 11 shows the relation between non-dimensional secant
stiffness degradation K/Kcr (Kcr being secant stiffness at cracking
limit state Kcr = Hcr/dcr) and measured displacement ratio d/dHmax.
As already observed for masonry in general [2] and for multi-leaf
stone walls in particular [1], the shape of stiffness degradation as
a function of lateral displacement in a non-dimensional form gen-
erally follows a power law and is quite similar in all types of ma-
sonry walls. Distribution coefficients close to one were obtained
by adjusting power law trend lines to the experimental data of
each tested panel.
Concerning stiffness degradation, a rapid decrease in all speci-
mens subjected to lower vertical stress levels was observed. In-
jected panels clearly exhibited higher starting stiffness than non-
injected ones, and a more rapid decrease due to rocking, whereas
2:3 injected specimens showed slower stiffness degradation.
Non-injected 2:3 panels, compared with the corresponding 1:1
ones, had higher starting stiffness ratios, sustaining the decrease
in stiffness and reaching a higher displacement ratio.
Analysis of energy and related viscous damping provided useful
information on overall wall behaviour. Fig. 12 and Table 5 show the
energy ratios and equivalent viscous damping values for each pa-
nel and the average values at the various limit states.
As regards energy dissipation, all non-injected specimens
exhibited initial energy ratio (Ehys/Einp) higher than injected ones.
Subsequently, the decrease in this ratio from the cracking limit
Fig. 11. Stiffness degradation vs. normalised displacement. (a) Panels in scale 1:1. (b) Panels in scale 2:3 (the definition of the used symbols may be found in the symbols list).
to the attainment of lateral resistance was almost the same for in-
jected panels in both scales. Beyond this phase, increase was lim-
ited in all specimens, apart from the non-injected 2:3 ones,
which showed a plateau up to failure. Higher values of ratio Ehys/
Einp were obtained for the non-injected specimens, but the abso-
lute value of Ehys (dissipated energy) was higher for the injected
ones, due to the greater displacement capacity reached after grout
injection.
Analysing the equivalent viscous damping (n), similar behaviour
was reported as for energy, see Fig. 12b, non-injected panels show-
ing greater values at initial damping and a substantial increase fol-
lowing the attainment of lateral resistance up to failure. The
injected panels showed a much greater increase compared with
the non-injected ones.
3.4. Mechanical parameters
Shear compression tests also revealed important mechanical
characteristics, in particular, tensile strength ft and shear charac-
teristics such as maximum shear stress and sHmax and shear mod-
ulus G. Referential tensile strength ft of the masonry was evaluated
with Eq. (1) [14], considering applied vertical stress r00 and panel
geometry, when average shear stress at maximum resistance sHmax
is known, see Table 6. sHmax is the nominal average shear stress at
maximum resistance (horizontal load divided by horizontal cross-
sectional area) [21].
The average value of ft for the panels in scale 1:1 was 0.050 N/
mm2 in non-injected conditions and 0.170 N/mm2 the injected
ones. For panels in scale 2:3, the ft values for non-injected
(0.060 N/mm2) and injected (0.150 N/mm2) conditions were very
similar to those obtained for 1:1 panels and matched the range
of values found in literature [22] for uncut stone masonry with
external layers of limited thickness and infill cores, in original con-
ditions (ft from 0.0525 N/mm2 to 0.0765 N/mm2), injected with
grout (ft ranging from 0.187 N/mm2 to 0.273 N/mm2).
Analysis of mechanical properties showed that tensile strength
ft was not significantly affected by the scale factor. However, due to
its different mechanical definition, shear stress sHmax was obvi-
ously affected by pre-compression level. In any case, for non-in-
jected masonry panels, average shear stress was found to be
0.185 N/mm2 and tensile strength 0.055 N/mm2 for injected pan-
els, the estimated ft was 0.16 N/mm2 and sHmax 0.39 N/mm2.
Shear modulus G was evaluated according to three methods.
First, it was determined according to the effective stiffness, Eq.
(5), of wall Kcr, in the boundary condition of a cantilever wall
 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161 159

50
(a) 100
(b)
90 45

80 40

70 35

60 30

50 25

40 20

30 15

20 10

10 5

0 0
0 0.5 1 1.5 2 2.5 3 3.5 0 0.5 1 1.5 2 2.5 3 3.5

Fig. 12. Energy dissipation (a) and equivalent viscous damping (b).

Table 5
Energy dissipation and equivalent viscous damping at the different limit states (the definition of the used symbols may be found in the symbols list).

Specimens r00 (N/mm2) Flexural cracking Cracking limit Maximum resistance Ultimate displacement
Ehys/Einp (%) nf (%) Ehys/Einp (%) ncr (%) Ehys/Einp (%) nHmax (%) Ehys/Einp (%) ndu (%)
SCF1.0NI 1.0 45 12 11 10 23 10 33 17
SCF1.25NI 1.25 72 16 66 16 61 16 60 16
SCF0.75NI 0.75 59 13 50 10 47 11 54 24
SCF0.5NI 0.5 55 27 44 36 40 63 46 89
Averagea 59 14 42 12 44 12 49 19
SCF1.0I 1.0 31 5 22 4 18 9 18 15
SCF1.5I 1.5 42 7 36 6 30 11 33 16
SCF1.25I 1.25 44 7 29 6 26 7 26 10
SCF2.0I 2.0 46 9 37 7 37 9 42 12
Average 41 7 31 6 28 9 30 13
SCS0.5NI 0.5 81 28 67 20 59 14 52 12
SCS1.0NI 1.0 75 21 52 9 47 13 45 14
SCS0.75NI 0.75 63 15 49 12 46 12 46 17
SCS1.25NI 1.25 75 18 61 13 51 12 51 27
Averageb 71 18 54 11 48 12 47 19
SCS1.25I 1.25 60 9 39 9 35 12 38 19
SCS1.0I 1.0 47 11 34 6 30 12 32 14
SCS1.5I 1.5 53 9 43 8 40 8 41 12
SCS2.0I 2.0 62 13 55 11 51 14 51 18
Average 56 10 43 8 39 11 41 16
a
Average value not including the damaged panel SCF0.5NI.
b
Average value not including the damaged panel SCS0.5NI.

Table 6
Tensile strength (ft) determined based on the experimental shear compression tests (the definition of the used symbols may be found in the symbols list).

Specimens sHmax (N/mm2) r00 (N/mm2) b ft (N/mm2) ft-average (N/mm2)

SCF1.0NI 0.23 1.0 1.2 0.07 0.05

SCF1.25NI 0.21 1.25 0.05
SCF0.75NI 0.15 0.75 0.04
SCF0.5NI 0.15 0.5 0.06
SCF1.0I 0.34 1.0 0.15 0.17
SCF1.5I 0.43 1.5 0.16
SCF1.25I 0.42 1.25 0.18
SCF2.0I 0.52 2.0 0.18
SCS0.5NI 0.11 0.5 1.25 0.04 0.06
SCS1.0NI 0.25 1.0 0.02
SCS0.75NI 0.20 0.75 0.08
SCS1.25NI 0.28 1.25 0.09
SCS1.25I 0.39 1.25 0.17 0.15
SCS1.0I 0.34 1.0 0.16
SCS1.5I 0.39 1.5 0.14
SCS2.0I 0.44 2.0 0.14
160 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161

Table 7
Shear modulus (G) determined according to the different approaches and for the different typologies of panels (the definition of the used symbols may be found in the symbols
list).

Specimens E (N/mm2) r00 (N/mm2) Gexp (N/mm2) Gexp 30–60% (N/mm2) Gk (N/mm2)

SCF1.0NI 2531 1.0 50.0 64.3 187.1

SCF1.25NI 1.25 212.4 332.2 216.6
SCF0.75NI 0.75 213.1 216.6 276.9
SCF0.5NI 0.5 87.2 95.9 96.6
SCF1.0I 4966 1.0 1140.0 1116.6 59.6
SCF1.5I 1.5 1386.3 1021.3 271.0
SCF1.25I 1.25 147.8 117.4 132.1
SCF2.0I 2.0 21.4 91.2 236.1
SCS0.5NI 2959 0.5 37.7 52.2 63.7
SCS1.0NI 1.0 137.0 121.9 109.0
SCS0.75NI 0.75 176.2 166.9 124.0
SCS1.25NI 1.25 573.9 429.3 267.3
SCS1.25I 5125 1.25 824.1 791.0 145.0
SCS1.0I 1.0 1496.8 1059.2 96.7
SCS1.5I 1.5 289.4 383.0 189.9
SCS2.0I 2.0 582.8 735.2 201.8

(a = 3.33) and in the hypothesis that the panel behaves as a homo-
geneous, isotropic material (Gk). The second and third methods
make use of the experimental data, estimating G with Eq. (6). In
the second method, shear modulus Gexp is considered to corre-
spond to the cracking limit state. The third method estimates
parameter Gexp 30–60%, considering the average of Gexp in the inter-
val between 30% and 60% of maximum shear resistance [23,24].
G A
kcr ¼ h k i ð5Þ
1; 2  1 þ a GEk ðhlÞ
s larger displacement values when compared with non-
Gexp ¼ P2
1
c
i¼1 i
2
where A is the cross-sectional area, h and l are the height and length
of the specimen, E is the elastic modulus evaluated by vertical com-
pression tests on the specimens, and s and ci are the average nom-
inal shear stress and the shear strain evaluated on the masonry
panel. Effective stiffness Kcr is the ratio between lateral resistance
H and the corresponding displacement d, and can be experimentally
evaluated from the envelope curves of the hysteresis loops. In this
case, it was calculated as the secant value at the cracking limit state.
Table 7 lists shear modulus values determined according to the
various approaches.
As regards evaluation of the shear modulus, the values corre-
sponding to cracking limit state Gexp were higher than those obtained
with shear modulus Gexp 30–60% (7% higher for 2:3 NI, 13% higher for
1:1 I, 17% higher for 2:3 NI, and 26% lower for 1:1 NI). The values ob-
tained with these two methods were very different from those calcu-
lated with the first method (Gk) for the specimens on which failure
was greatly affected by rocking (e.g., SCF1.0I, SCF1.5I and SCS1.0I) and,
as such, cannot be considered as perfectly representative.
4. Conclusions
Experimental shear compression testing was carried out on 16
three-leaf stone masonry panels in varying conditions (injected/
non-injected with grout), scales (full scale 1:1, reduced scale 2:3)
and pre-compression levels (from 20% to 50% of rmax), to evaluate
the effects of grout injections on cyclic in-plane horizontal loads.
The main conclusions are as follows:
 The overall behaviour of the panels was greatly influenced by
the local position of the stones, mechanical characteristics,
and pre-compression load. The more highly pre-compressed
specimens achieved greater maximum resistance with more
brittle failure at the final phase.
 Non-injected panels exhibited mainly shear mechanisms in
specimens subjected to higher vertical loads, whereas lower
pre-compressed specimens developed a flexural mechanism.
Injected panels were influenced by a rocking mechanism.
 In all cases, non-injected panels underwent leaf separation at
lower displacement levels, specimens failing due to buckling
of external leaves. Instead, injected specimens, in both
scales, showed extensive transversal cracking after reaching
maximum resistance, with cracking of the inner core, but no
significant layer separation.
 In injected panels, the first crack appeared at almost 2 times
ð6Þ
injected ones. At maximum resistance, injected panels
exhibited 3 times higher displacement at almost 2 times
higher force than the non-injected ones. A rapid decrease
in stiffness in all specimens subjected to lower vertical stress
levels was observed. Injected panels clearly had higher start-
ing stiffness compared with non-injected ones, and a more
rapid decrease due to rocking. Conversely, all non-injected
specimens had a higher initial energy ratio Ehys/Einp than
injected ones. In all types of mechanisms, the lower the
applied pre-compression, the faster the decrease, until max-
imum lateral resistance was attained. Equivalent viscous
damping showed similar behaviour to energy.
 Analysis of mechanical properties revealed that, after grout
injection, tensile strength ft increased 3 times and, in terms
of experimental values (Gexp and Gexp30–60%) the shear modu-
lus increased approximately 4.0 times; Gk remained more or
less constant.
 Changing the scale factor changed the influence of pre-com-
pression on panel behaviour. The same collapse mechanisms
for the same pre-compression values could not be repro-
duced in both scale 1:1 and 2:3 panels. However, in general,
the use of scaled specimens may be considered representa-
tive of the type of masonry tested here.
Acknowledgements
The first author would like to thank the FCT (Fundação para a
Ciência e a Tecnologia – Foundation for Science and Technology)
of Portugal. This work was carried out under EU Contract FP7-
ENV-2009-1-GA244123: ‘New integrated knowledge-based ap-
proaches to the protection of cultural heritage from earthquake-in-
duced risk – NIKER’. Research was also partially supported by the
Reluis Project and Executive Programme of Co-operation in the
Field of Science and Technology, between the governments of Italy
 B. Silva et al. / Construction and Building Materials 53 (2014) 149–161
and Japan. Thanks are due to Tassullo S.p.A. (Italy) for providing the
basic materials. Specimens were built at ESEV (masonry school,
Verona, Italy) and were tested at the Laboratory of Structural Mate-
rial Testing, University of Padova. The authors would also like to
thank engineers S. Santunione and N. Secchiero for their work as
part of their graduation theses.
References
[1] Mazzon N. Influence of grout injection on the dynamic behaviour of stone
masonry buildings. PhD Thesis. Università degli Studi di Padova. Padova, Italy;
2010.
[2] Tomaževic M. Earthquake-resistant design of masonry buildings. London
(Great Britain): Imperial College Press; 1999.
[3] Tomaževič M, Apih V. The strengthening of stone-masonry walls by injecting
the masonry-friendly grouts. Eur Earthq Eng 1993;6(1):10–20.
[4] Toumbakari EE. Lime-pozzolan-cement grouts and their structural effects on
composite masonry walls. PhD thesis. Katholieke Universiteit Leuven,
Belgium; 2002.
[5] Tomaževič M, Sheppard PF. The strengthening of stone-masonry buildings for
revitalization in seismic regions. In: Proceedings of 7th European conference
on earthquake engineering; 1982. p. 275–82.
[6] Silva B, Pigouni E, Valluzzi MR, da Porto F, Modena C. Assessment of the
effectiveness of grout injection on the consolidation of 3-leaf stone masonry
walls through the use of NDT and DT. In: Proceedings of SAHC conference,
Wroclaw, Poland; 2012.
[7] Silva B, Dalla Benetta M, da Porto F, Valluzzi MR. Compression and sonic tests
to assess effectiveness of grout injection on three-leaf stone masonry walls. Int
J Archit Heritage, 8(3), 2014, 408-435.
[8] Valluzzi MR. Comportamento meccanico di murature consolidate con
materiali e tecniche a base di calcea. PhD Thesis. University of Trieste,
Trieste, Italy (in Italian); 2000.
[9] Tassullo SpA. Tassullo Trento, Italy; 2012. Available at: <http://www.
tassullo.it>.
[10] ASTM C939, 1994. Flow cone method. Standard test method for flow of grout
for preplaced-aggregate-concrete.
[11] Miltiadou AE. Grouting as a method for the repair of masonry monuments.
MSc thesis. Institute of advanced architectural studies, University of York,
York, Great Britain; 1985.
[12] Mosele F. In-plane and out-of-plane cyclic behaviour of reinforced masonry
walls. PhD thesis. Università degli Studi di Trento; 2009.
[13] da Porto F. In-plane cyclic behaviour of thin layer joint masonry walls. PhD
thesis. Università degli Studi di Trento; 2005.
161
[14] Turnšek V, Cacovic F. Some experimental results on the strength of brick
masonry walls. In: Proceedings of 2nd international brick masonry conference,
Stoke-on-Trent, Great Britain; 1971. p. 149–56.
[15] Turnšek V, Sheppard P. The shear and flexural resistance of masonry walls. In:
Proceedings of research conference on earthquake engineering, Skopje,
Republic of Macedonia; 1980. p. 517–73.
[16] Benedetti D, Tomaževič M. Sulla verifica sismica di costruzioni in muratura.
Ingegneria Sismica 1984;I:9–16 (in Italian).
[17] Decreto Ministeriale 14/01/2008 (NTC 2008). Norme Tecniche per le
Costruzioni (Italian).
[18] Abrams DP. Performance based engineering concepts for unreinforced
masonry building structures. Prog Struct Mat Eng 2001;3:48–56. http://
dx.doi.org/10.1002/pse.70.
[19] Silva B. Diagnosis and strengthening of historical masonry structures:
numerical and experimental analyses. PhD Thesis. University of Brescia,
Italy; 2012.
[20] Vasconcelos G, Lourenço P. In-plane experimental behavior of stone masonry
walls under cyclic loading. J Struct Eng 2009;135(10):1269–77.
[21] Bernardini A, Modena C, Turnšek V, Vescovi, U. A comparison of laboratory test
methods used to determine the shear resistance of masonry. In: Proceedings of
7th world conference on earthquake engineering, vol. 7, Istanbul, Turkey;
1982. p. 181–184.
[22] Circolare. ‘‘Istruzioni per l’applicazione delle ‘nuove norme tecniche per le
costruzioni’’ di cui al decreto ministeriale 14 gennaio 2008. GU n. 47 del 26–2-
2009 - Suppl. Ordinario n.27. n. 617; 2009 (in Italian).
[23] Bosiljkov V. Experimental and numerical research on the influence of the
modified mortars on the mechanical properties of the brick masonry. PhD
thesis. University of Ljubljana, Slovenia; 2000.
[24] Bosiljkov V, Totoev Y, Nichols J. Shear modulus and stiffness of brickwork
masonry: an experimental perspective. Struct Eng Mech 2005;20(1).
[25] Modena C, Bettio C. Experimental characterization and modelling of injected
and jacketed masonry walls. Nice, France: Italian–French Symposium
Strengthening and Repair of Structures in Seismic Areas; 1994. p. 273–82.
[26] Modena C. Interpretazione dei risultati ottenuti dalle prove in sito nell’ambito
delle tre convenzioni con gli istituti di ricerca di Firenze e Milano e
modellazione del comportamento strutturale dei componenti rinforzati.
Convenzione di ricerca tra la Regione Toscana e il Dipartimento di
Costruzioni e Trasporti dell’Università degli Studi di Padova; 1999 (in Italian).
[27] Corradi M, Borri A, Vignoli A. Experimental study on the determination of
strength of masonry walls. Constr Build Mater 2003;17(5):325–37.
[28] Corradi M, Tedeschi C, Binda L, Borri A. Experimental evaluation of shear and
compression strength of masonry walls before and after reinforcement: deep
re-pointing. Constr Build Mater 2008;22(4):463–72.
[29] Galasco A, Penna A Magenes G. Caratterizzazione meccanica di muratura in
pietra. Parte seconda: prove cicliche di taglio-compressione su pannelli di
grandi dimensioni. Technical report, Università degli Studi di Pavia. Allegato
4.2-UR01-2; 2009 (in Italian).