You are on page 1of 13

Journal of Building Engineering 28 (2020) 101038

Contents lists available at ScienceDirect

Journal of Building Engineering


journal homepage: http://www.elsevier.com/locate/jobe

Strength and failure mechanisms of masonry prisms under compression,


flexure and shear: Components’ mechanical properties as design constraints
Gustavo Henrique Nalon a, *, Carol Ferreira Rezende Santos b, Leonardo Gonçalves Pedroti a,
Jos�e Carlos Lopes Ribeiro a, Gustavo de Souza Veríssimo a, Fl�avio Anto
^nio Ferreira a
a
Civil Engineering Department, Federal University of Viçosa, Av. Peter Henry Rolfs, Campus UFV, Dep. de Eng. Civil, Zip Code: 36.570-900, Viçosa, Minas Gerais, Brazil
b
Civil Engineering Department, Sao Carlos School of Engineering, Av. Trabalhador S~ ao-carlense, 400, Parque Arnold Schimidt, Dep. de Eng. Civil, Zip Code: 13.566-590,
S~
ao Carlos, S~
ao Paulo, Brazil

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

Keywords: This work aims to investigate the influence of mechanical properties of concrete blocks, cement-lime mortar and
Structural masonry grout on the structural behavior of prisms under compression, flexure, and shear, considering specific scenarios
Failure modes faced by designers: blocks are considerably stronger and stiffer than mortar, and vice versa. Discussions about the
Strength predictions
failure mechanisms were provided and analysis of variance was used to explain how the variation of the com­
Design constraints
ponents’ properties affected prisms’ strength and stiffness. Strength prediction models were developed in order
to account for the failure modes in different case scenarios and improve the predictive ability of formulations of
current design codes.

shape of the specimens, stiffness of bearing plates and other factors


related to characteristics of the experimental test setup [1,24,25].
1. Introduction There is a consensus that increases in the unit strength cause sig­
nificant enhancement of the ungrouted masonry strength under
In block masonry structures, vertical and horizontal loads are sup­ different kinds of loads [5–8]. Most works report small increases in the
ported and transferred to the foundation through structural elements masonry compressive strength when the mortar strength is increased [7,
constructed with blocks, mortar, and, eventually, grout and reinforce­ 9–11]. Some aspects of the structural behavior of grouted masonry are
ment. Blocks are of major importance in terms of strength of the struc­ still controversial. Some authors point out that the effect of mortar
ture, while the mortar must bond the blocks together, ensure stresses strength on the strength of grouted masonry decreases with the
transmission, and accommodate small strains. Grout is responsible for enhancement of the grout strength [12]. In contrast, others verified a
embedding reinforcing steel bars and increasing the bearing capacity of negligible influence of mortar strength on the strength of grouted ele­
the masonry [1]. ments [13]. Many studies verified substantial increases of masonry
Strength and stiffness of masonry structures are directly related to strength when grout is used [12,14–16]. Others verified that the use of
the properties of its components and the complex interaction between grouts much stronger than blocks or vice versa does not cause en­
them [2]. In fact, many of the difficulties in understanding the masonry hancements of masonry strength [8,13,17].
behavior are related to the geographical variations in materials prop­ The Brazilian code ABNT NBR 15961–1:2011 [18] covers the design
erties and the scarcity of controlled experimental tests [3]. Despite the of masonry structures made of hollow concrete blocks and states that the
large amount of studies developed over the last decades, the masonry maximum mortar compressive strength is 70% of the blocks’ net area
structural behavior is still far from being completely clarified [4]. Many compressive strength. When the strength of the mortar is close to this
different factors affect the strength, the deformability and the failure upper limit, mortar joints of masonry structures are submitted to
mode of masonry elements, including the mechanical properties of the triaxial compression, while the blocks are found in a biaxial
components, the thickness of the mortar joints, the geometry of blocks compression-tension state [4,7,19]. Many authors verified that
and the mortar bedding method [5–25]. Results of experimental eval­ ungrouted masonry prisms with these characteristics failed by vertical
uations of these characteristics are also affected by the dimensions and

* Corresponding author.
E-mail addresses: gustavo.nalon@ufv.br (G. Henrique Nalon), carolrezende@usp.br (C.F.R. Santos), lpedroti@gmail.com (L.G. Pedroti), jcarlos.ribeiro@ufv.br
(J.C.L. Ribeiro), gustavo@ufv.br (G.S. Veríssimo), flavioferreira@ufv.br (F.A. Ferreira).

https://doi.org/10.1016/j.jobe.2019.101038
Received 6 April 2019; Received in revised form 30 October 2019; Accepted 31 October 2019
Available online 4 November 2019
2352-7102/© 2019 Elsevier Ltd. All rights reserved.
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

Abbreviations and symbols t mortar joint thickness


X average of actual strengths of n number of observations
Eb elastic modulus of blocks Xa experimentally obtained strength value
Eg elastic modulus of grout Xp predicted strength value
Em elastic modulus of mortar А ratio of net area to gross area of hollow blocks
Ep elastic modulus of masonry prisms В ratio of grout area to gross area of hollow blocks
fb compressive strength of block εxx;b transversal strain of blocks
fbk characteristic compressive strength of blocks εxx;m transversal strain of mortar
fbt tensile strength of blocks υb Poisson’s ratio of blocks
fg compressive strength of grout υm Poisson’s ratio of mortar
fm compressive strength of mortar σc applied compressive stress
fmt tensile strength of mortar σp flexural bond strength of prisms
fp ultimate compressive strength of prisms σ xx and σ yy transversal stresses on the masonry component
fpa average compressive strength of prisms σ zz axial stress on the masonry component
fpk characteristic compressive strength of prisms τp initial shear strength of prisms
hb height of blocks

cracks in the blocks shells when submitted to axial compression [6,17, attention must also be given to the use of masonry units with low
19]. Bolhassani et al. [26] also state that vertical tensile splitting cracks strength and stiffness, which have been frequently used in some regions
initiated at the middle web of the blocks were observed in tests of around the world as structural components. Prakash et al. [23] mention
ungrouted prisms. In grouted prisms under compression, the authors that only bricks much softer than mortars are easily available in the
verified a similar failure mode, but characterized by diagonal cracks. Andhra Pradesh region (India). The strength and stiffness of these units
Different authors observed that the failure mode of grouted prisms re­ are very low when compared to the mechanical properties of other In­
sults from the combination of lateral tension stresses on the block shells dian regions’ bricks. Oliveira et al. [33] state that around 6000 resi­
caused by the expansion of the grout core and axial compressive stresses dential buildings in the Metropolitan Region of Recife (Brazil) were
due to the vertical load [1,12,17,27]. The failure mechanism of masonry constructed with low-strength clay blocks that support high load levels.
prisms is directly related to the deformation properties of their compo­ Pathological manifestations have been observed in these constructions
nents. In stress-strain curves of prisms under compression, Mohamad and collapses involving human deaths have been reported. Situations in
et al. [22] observed discontinuities in strain at approximately 50% of which blocks are much softer than mortars were not included in the
their ultimate stress, which is related to localized cracks. Then, the tabulated strength predictions of current guidelines. Few works have
rupture mode of these elements starts at the onset of the nonlinear dealt with these different case scenarios [23,33,34]. In these situations,
behavior. the state of stresses of the components and the failure mechanism of the
Many researchers observed that when the joint thickness is reduced, structure are completely different from those observed in conventional
increases in compressive strength and modulus of elasticity of masonry cases and considered in the prediction equations of design codes [35].
prisms are obtained, which is related to increases in the lateral The authors are unaware of any previous work that investigated the
confinement of the mortar joint [28–32]. Different researchers also compressive, flexural and shear behavior of masonry, considering these
observed that the type of mortar bedding (face shell bedding and full specific design constraints: mortar must be significantly stronger than
area bedding) also affects the structural behavior of masonry elements, units and vice versa. Despite this, many engineers around the world
since higher lateral stresses are developed in the block webs of masonry must deal with them (some practical examples are cited in Refs. [22,23,
elements whose units are face shell bedded [1,14,17,19]. 33,34]). The present research was carefully designed to investigate the
Compression, flexural and shear tests of masonry elements have influence of masonry components’ properties on the strength, stiffness,
some practical restrictions in terms of costs of manpower, materials and and failure mode of hollow concrete blocks prisms under compressive,
training, and complexities of transporting specimens to testing facilities flexural, and shear stresses, considering three distinct case scenarios: (i)
[15]. Then, current design codes propose empirical equations and the combination of properties of masonry components is in accordance
tabulated values for predicition of masonry strength. Korany and with assumptions of current guideline [18]; (ii) the cement-lime mortar
Glanville [20] studied the values estimated by the American, Canadian, is much stronger and stiffer than hollow concrete blocks; (iii) the
British, and Australian masonry codes. They concluded that those cement-lime mortar is much weaker and softer than the hollow concrete
equations were not representative of measured values and the variation blocks.
between the codes increased as the unit strength increased. Most of the
equations are based on limited data of research that was carried out in 2. Material and methods
the 1970s and 1980s [21] and are not open to the general-purpose use.
Then, it is worthwhile to re-evaluate equations and tables of masonry Two types of (14 � 19 x 39) cm hollow concrete blocks were used.
codes in order to increase their predictive ability [15]. Mohamad et al. Both of them comply with the requirements of ABNT NBR 15961-1 [18].
[22] also state that the codes must present a more appropriate way to Designated as “B1 blocks” and “B2 blocks”, they were manufactured for
account for the masonry failure modes in different situations. Some minimum gross area compressive strength of 4.5 MPa and 8.0 MPa,
design constraints related to the components’ mechanical properties can respectively. Experimental tests of “B1 blocks” and “B2 blocks” resulted
limit the use of these design formulations. in average net area compressive strength values of 10.9 MPa and
Nowadays, high-strength clay and concrete blocks have been 14.8 MPa, respectively.
developed by many manufacturers for construction of tall buildings. To produce mortars and grouts, the Portland cement CP II E 32 with
However, equations proposed by current codes to estimate masonry density of 2.95 kg/dm3, fineness of 4.0%, and 28-days compressive
compressive strength are not applicable for walls constructed with strength of 34.5 MPa was used. Hydrated lime with density of 2.32 kg/
mortars much weaker than the blocks, since the rupture starts at lower dm3 and water retention higher than 70% was also used to produce
load levels due to the early pore collapse of the mortar [22]. Special mortars. Quartzite sand with maximum size of 4.8 mm, fineness module

2
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

of 2.79, bulk specific gravity of 2.65 kg/dm3, unit weight of 1.40 kg/ lime, and water, in the appropriate proportions. The mixture was
dm3, and water absorption of 0.80% was the fine aggregate of the weighted and subjected to a maturation process for approximately 24 h.
mixtures. Grout production also required gneiss gravel with maximum In the following day, water losses resulted from evaporation were cor­
size of 12.5 mm, bulk specific gravity of 3.02 kg/dm3, unit weight of rected, and cement was added, followed by another mechanical mixing.
1.52 kg/dm3, and water absorption of 0.20% as course aggregate, and a After being prepared, mortar was submitted to the flow test and water
multifunctional plasticizer additive with 1.10 kg/dm3 density. These retention test [18]. A mortar flow in the range of (230 � 10) mm was
materials were characterized according to the test procedures cited in aimed for, as recommended by Parsekian and Soares [38]. They also
ABNT NBR 6118 [36]. assert to be desirable a high water retention index. Thus, an attempt was
The experimental program consisted of axial compression tests of made to keep this parameter in the range of (90 � 5) %. The mix-design
grouted and ungrouted two-block prisms; flexural tests of four-block method proposed by the Brazilian Association of Portland Cement
prisms; and shear tests in triplets. Most of the similar studies devel­ (ABCP) was used to produce a grout with compressive strength around
oped a multi-factor experimental design in which the factors were the 15 MPa. Grout slump was kept close to 200 mm, which is ideal for blocks
strength of blocks and the strength of the mortar. In order to analyze of moderate absorption [35].
masonry elements produced with blocks significantly softer than mor­ For compression tests, two-blocks prisms were produced, as recom­
tars and vice versa, this work proposes the variation of slightly different mended by ABNT NBR 15961–2:2011 [39]. Flexural tests followed
factors: the compressive strength of the blocks and the ratio between the recommendations of ASTM E518-15 [40], which proposes the use of
mortar compressive strength and the block’s net area compressive 460 mm minimum height specimens and at least four block courses. For
strength. A 2 � 3 factorial design was developed. For each kind of block, shear tests, triplets were used, following recommendations of the CSN
one of the mortar compositions had compressive strength close to the EN 1052–3:2002 [41]. All of the prisms were constructed by an expe­
maximum limit proposed by the ABNT NBR 15961–1:2011 [18] (70% of rienced mason. Joints were fully bedded, with thickness around (10 � 3)
the blocks’ net area compressive strength), another lower, and another mm. Concrete blocks’ surface was moistened before placing mortar and
above that limit, as shown in Table 1. The response variables of the grout. During 28 days, the prisms were kept immobile.
factorial design were the compressive strength of ungrouted and grouted Right after the production of each kind of mortar, six (4 � 4 x 16) cm
prisms, elastic modulus of ungrouted and grouted prisms, and flexural prismatic specimens and four (5 � 10) cm cylindrical specimens were
bond strength (normal to the bed joints) and initial shear bond strength cast. The curing process of mortar specimens occurred inside an envi­
(cohesion) of ungrouted prisms. ronmental room, at (23 � 2) � C and relative humidity around 95%. After
Experimental tests for determination of actual dimensions, water grout production, four (10 � 20) cm cylindrical specimens were cast and
absorption, initial water absorption, net area, tensile and compressive kept in saturated limewater at (23 � 2) � C during 28 days.
strength of the blocks were performed [18,36]. To measure longitudinal At the age of 28 days, mortar prisms were used to determine the
strains during the compression tests, two electric strain gauges were tensile strength, prism-halves (4 cm cubes) to determine the compres­
placed in each longitudinal shell of B1 and B2 blocks, following rec­ sive strength, and cylinders to determine the elastic modulus of the
ommendations of Barbosa [37]. Elastic modulus was calculated as the material (ABNT NBR 6118 [36]). Grout cylinders were used to deter­
slope of the straight line defined by the stress-strain curve’s points with mine the compressive strength and elastic modulus of the material at the
compressive stress of 0.5 MPa and 30% of the blocks strength. age of 28 days. For both, elastic modulus calculation considered the
Six different mortar compositions were produced. Mortar strengths slope of the straight line defined by the points of compressive stress of
presented in Table 1 were aimed for. One day before assembling the 0.5 MPa and 30% of the estimated strength. Strain was measured by
prisms, the mortar compositions started to be prepared by mixing sand, electric clip-gages.

Table 1
Characteristics of the specimens produced.
Test Name of the series fbk fm fg Number of specimens

Axial compression tests COM/B1/A<40/SG f B1


bk Lower than 40% of f B1
bk
Ungrouted 4
COM/B1/A<40/CG 15 MPa 4
COM/B1/A70/SG Around 70% of f B1
bk
Ungrouted 6
COM/B1/A70/CG 15 MPa 6
COM/B1/A>100/SG Higher than 100% of f B1
bk
Ungrouted 4
COM/B1/A>100/CG 15 MPa 4
COM/B2/A<40/SG f B2
bk Lower than 40% of f B2
bk
Ungrouted 6
COM/B2/A<40/CG 15 MPa 6
COM/B2/A70/SG Around 70% of f B2
bk
Ungrouted 6
COM/B2/A70/CG 15 MPa 6
COM/B2/A>100/SG Higher than 100% of f B2
bk
Ungrouted 6
COM/B2/A>100/CG 15 MPa 6
Flexural tests FLE/B1/A<40 f B1
bk Lower than 40% of f B1
bk
Ungrouted 3
FLE/B1/A70 Around 70% of f B1
bk
3
FLE/B1/A>100 Higher than 100% of f B1
bk
3
FLE/B2/A<40 f B2
bk Lower than 40% of f B2
bk
Ungrouted 3
FLE/B2/A70 Around 70% of f B2
bk
3
FLE/B2/A>100 Higher than 100% of f B2
bk
3
Shear tests SHE/B1/A<40 f B1
bk Lower than 40% of f B1
bk
Ungrouted 3
SHE/B1/A70 Around 70% of f B1
bk
3
SHE/B1/A>100 Higher than 100% of f B1
bk
3
SHE/B2/A<40 f B2
bk Lower than 40% of f B2
bk
Ungrouted 3
SHE/B2/A70 Around 70% of f B2
bk
3
SHE/B2/A>100 Higher than 100% of f B2
bk
3

3
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

The compressive strength and elastic modulus of ungrouted and Table 2


grouted two-block prisms were determined in compression tests with the Properties of the hollow concrete blocks.
load-time history proposed by ABNT NBR 15961–2:2011 [39] for me­ Property B1 Blocks (Grade B2 Blocks (Grade
chanical characterization of concrete blocks masonry prisms. Before B) A)
testing, a regularization of the bearing surfaces of the prisms was per­ Average dimensions (mm) - width x 138 � 190 x 390 140 � 190 x 390
formed with 3 mm mortar caps with compressive strength higher than height x length
the net area compressive strength of the concrete blocks. Then, capping Average thickness of longitudinal shells 30 31
provided stiff and plane surfaces perpendicular to the specimen axis. (mm)
Average thickness of transverse webs 29 31
Displacements and load levels were recorded using a QuantumX Data (mm)
Acquisition System at a 5 Hz rate, two linear displacement transducers, Net area/gross area ratio 0.54 0.55
and a 600 kN load cell. The displacement transducers were placed on Average net area compressive strength 14.1 17.1
both sides of the specimen, as shown in Fig. 1a. Elastic modulus was (MPa)
Characteristic net area compressive 10.9 14.8
calculated as the slope of the straight line defined by the points of
strength (MPa)
compressive stress of 5% and 30% of the prism strength. It is important Average tensile strength (MPa) 0.83 1.28
to clarify that small deformations of the 3 mm mortar cap slightly Water absorption (%) 6.2 5.5
affected the calculated elastic modulus, since the transducers measured Initial water absorption ((g/193.55 cm2)/ 8.27 7.09
the relative displacement between the load plates of the press. Some min)
Average net area elastic modulus (MPa) 7310 15,769
obtained stress-strain curves presented an initial concave section related
to imperfect contact between the specimen and the bearing plates of the
press. In these specific cases, the curve was normalized before calcula­ Tables 3 and 4 present the data related to the grout and mortar pro­
tion of the elastic modulus, according to procedures applied by Costigan, duction processes, respectively. Results of compression tests of prisms
Pavía, and Kinnane [42]. This correction must be qualified as a limita­ and the ratios between their components’ properties are presented in
tion of the methodology used in this research. It is important to highlight Table 5, while Table 6 shows results of flexural and shear tests, with
that the transducers could have been directly fixed to the specimens and some characteristics of the masonry components used in these series.
measured displacements fully related to them, so that similar corrections
would not be required. A digital camera was used to record compression 3.2. Structural behavior of prisms under compression
tests. Through the images and the data of load-time curves, each visible
failure warming was identified, as well as the load level in which it The performance of the structural masonry under compression can
occurred. Based on visual observations, it was possible to understand be better understood based on the tendency of differential movements
prisms’ ductility and failure modes. Strength and stiffness results were between its components, which is determined by their mechanical
subjected to analysis of variance at the significance level of 5%. properties and state of stresses. In this section, the mechanical properties
Four-block prisms were used to perform third-point beam tests and of blocks, mortar and grout used in each masonry element were
determine the flexural bond strength, normal to the bed joints. Prisms compared in order to estimate the tendency of differential lateral
were placed on steel rollers under the movable plate of the press, as movements during the compressive load application, enabling to deduce
recommended by ABNT NBR 15961–2:2011 [39] and shown in Fig. 1b. the kind of transversal stresses that act in each component and under­
Other two rollers on the central blocks of the prism were used to support stand the element’s failure mechanism. Due to the adhesion between the
a central plate where the load was applied using displacement control, components, these differential movements are restrained, and the actual
with constant speed of 0.01 mm/s. strain is an intermediate value between the amounts that each compo­
Triplets were used to perform shear tests, according to prescriptions nent tended to deform [43,44].
of CSN EN 1052–3:2002 [41]. The purpose of this work was to obtain the Ungrouted prisms produced with the stronger blocks (B2 blocks)
initial bond strength (no precompression was applied). Each specimen presented considerable compressive strength and elastic modulus gains
was placed between the movable plates of the press (Fig. 1c). Using with an enhancement in the compressive strength of the mortar. Such
displacement control and constant speed of 0.01 mm/s, the load was fact was not observed when the weaker blocks (B1 blocks) were used. An
applied on the central block of the triplet. Results of these tests were also analysis of variance proves such performance: comparing the average
submitted to an analysis of variance at the 5% significance level. compressive strength of COM/B2/A<40/SG, COM/B2/A70/SG and
COM/B2/A>100/SG series, at a 5% significance level, a P-value lower
3. Results and discussion than 0.05 was obtained, which rejects the null hypothesis that the av­
erages from the three samples are equal. When comparing the
3.1. Experimental tests results compressive strength of COM/B1/A<40/SG, COM/B1/A70/SG, and
COM/B1/A>100/SG series, the null hypothesis is confirmed, as the P-
Table 2 presents the properties of blocks. ABNT NBR 15961–1:2011 value was 0.37. Similar results were obtained from statistical compari­
[18] classifies concrete blocks in different strength categories: B1 blocks sons of the elastic modulus of the same series.
are Grade B (weaker units) and B2 blocks are Grade A (stronger units). From the results presented in Table 5, it is observed that the prisms of

Fig. 1. a) Compression test, b) Flexural test, c) Shear test.(2-column fitting image).

4
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

Table 3
Properties of grout.
Name Used in the prism series Mix proportions, in mass Slump (mm) fg (MPa) Eg (MPa)

Cement Sand Gravel Water

G15 COM/B1/A<40/CG 1.00 2.93 2.93 0.75 191 14.7 23,077


COM/B1/A70/CG 184 15.5 23,051
COM/B1/A>100/CG 191 14.7 23,077
COM/B2/A<40/CG 189 14.6 22,497
COM/B2/A70/CG 204 13.8 22,253
COM/B2/A>100/CG 190 15.7 24,350

Table 4
Mix proportion, flow and water retention of the cement-lime mortars.
Name Used in the prism series Mix proportions, in mass Flow (mm) Water Retention (%)

Cement Lime Sand Water

B1/A<40 B1/A<40/SG/CG 1.00 2.20 8.50 2.25 232 94


B1/A70 B1/A70/SG/CG 1.00 0.39 6.44 1.32 230 91
B1/A>100 B1/A>100/SG/CG 1.00 0.25 4.22 0.90 226 89
B2/A<40 B2/A<40/SG/CG 1.00 0.45 7.48 1.49 218 85
B2/A70 B2/A70/SG/CG 1.00 0.32 5.25 1.07 212 92
B2/A>100 B2/A>100/SG/CG 1.00 0.20 3.39 0.74 231 86

Table 5
Results of the compression tests of prisms and mechanical properties of their components.
Prims series fapa (MPa) fapk (MPa) fpk/fbk Eap (MPa) fm (MPa) fm/fbk (%)b fm/fbk (%)c ftm (MPa) Em (MPa)

COM/B1/A<40/SG 5.6 (8.2) 4.3 0.73 1858 (20.0) 3.7 (14.1) 33.9 62.7 1.6 (7.0) 5510 (6.3)
COM/B1/A<40/CG 11.5 (10.4) 8.7 1.47 2934 (17.7)
COM/B1/A70/SG 5.6 (9.0) 4.6 0.78 3030 (13.0) 7.0 (5.1) 64.2 118.6 2.3 (5.5) 10,403 (3.2)
COM/B1/A70/CG 12.9 (6.5) 10.4 1.76 4280 (12.1)
COM/B1/A>100/SG 6.1 (0.5) 6.1 1.03 2550 (23.2) 12.3 (12.0) 112.8 208.5 3.9 (8.3) 16,247 (4.7)
COM/B1/A>100/CG 12.1 (13.8) 9.1 1.54 3740 (23.1)
COM/B2/A<40/SG 6.4 (8.3) 5.5 0.67 2600 (17.5) 5.9 (6.9) 39.9 72.0 1.7 (9.12) 9473 (6.4)
COM/B2/A<40/CG 15.9 (4.1) 14.8 1.81 4100 (15.4)
COM/B2/A70/SG 8.6 (6.8) 7.7 0.94 3833 (16.9) 9.3 (9.8) 62.8 113.4 2.9 (10.7) 13,993 (2.7)
COM/B2/A70/CG 17.3 (1.3) 17.1 1.85 4283 (6.9)
COM/B2/A>100/SG 11.8 (7.9) 9.8 1.20 4567 (10.5) 17.1 (6.9) 115.5 208.5 4.8 (7.7) 19,388 (2.2)
COM/B2/A>100/CG 18.3 (4.4) 16.6 2.02 4717 (9.7)

Notes: The coefficients of variations results are given within the parentheses. a fpa, fpk, Ep were calculated considering the prism’s gross area. b
fbk was calculated
considering the block’s net area. c fbk was calculated considering the block’s gross area.

Table 6
Results of the flexural and shear tests and properties of the masonry components.
Series σm (MPa) τm (MPa) fm (MPa) fm/fbk (%)a fm/fbk (%)b ftm (MPa) Em (MPa)

FLE/B1/A<40 0.147 (27.6) – 3.9 (5.5) 35.8 62.7 1.8 (6.5) 5506 (6.3)
SHE/B1/A<40 – 0.067 (28.5)
FLE/B1/A70 0.210 (19.9) – 7.5 (2.0) 68.8 118.6 2.8 (7.5) 13,575 (35.7)
SHE/B1/A70 – 0.086 (1.9)
FLE/B1/A>100 0.279 (29.1) – 14.3 (7.1) 131.2 208.5 4.0 (10.2) 17,649 (8.5)
SHE/B1/A>100 – 0.124 (18.9)
FLE/B2/A<40 0.149 (23.5) – 4.3 (7.3) 29.0 52.4 1.9 (5.2) 8145 (9.1)
SHE/B2/A<40 – 0.050 (33.1)
FLE/B2/A70 0.218 (10.0) – 10.2 (3.7) 68.9 124.4 3.7 (2.3) 14,745 (5.0)
SHE/B2/A70 – 0.125 (14.1)
FLE/B2/A>100 0.408 (2.2) – 18.4 (7.4) 124.3 224.4 4.4 (7.6) 19,769 (2.0)
SHE/B2/A>100 – 0.175 (51.2)

Notes: The coefficients of variations results are given within the parentheses. a fbk was calculated considering the block’s net area. b fbk was calculated considering the
block’s gross area.

the COM/B2/A<40/SG and COM/B1/A<40/SG series were produced mortar of those series was excessively weak and prisms failed by mortar
with a very soft mortar, which suggests that the mortar tends to expand joint crushing. Close to the mortar joint, spalling of blocks occurred. As
laterally more than blocks. Due to the restriction imposed by the verified in the specimens of Fig. S1 of the Supplementary Information
adhesion between the components, there were lateral compressive file, this kind of failure may be considered ductile, since the localized
stresses in the mortar and lateral tensile stresses in the blocks. The mortar crushing starts when the compressive stress reaches around 80%

5
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

of the ultimate capacity of the prism. In some cases, mortar crushing (around 15 MPa), so the system acted more homogeneously. In fact, in
rupture was followed by cracking of the blocks, due to the concentration most of the times, grout and blocks broken practically together. Less
of stresses on the region of the block shell close to the locations where noticeable rupture warning was observed. In most of the compression
the crushed mortar was expelled. The prisms of those series presented tests, the strength of the prisms was even higher than the strength of the
lower compressive strength values. The characteristic gross area blocks or the strength of the grout. According to previous studies [1,17,
compressive strength of both series of prisms (5.5 and 4.3 MPa, 27], when the grout is not significantly stiffer than the blocks, its lateral
respectively), is lower than the characteristic gross area compressive expansion inside of the blocks causes tensile stresses on the block’s shells
strength of the blocks (8.2 and 5.9 MPa, respectively), which means that and webs. Fig. S5 of the Supplementary Information file presents the
both series presented very low efficiency. rupture mechanism of those prisms. Statistical analysis detected that
The COM/B1/A>100/SG and COM/B2/A>100/SG series were pro­ only the average compressive strength of prisms of the COM/­
duced with a stronger mortar, which suggests that the blocks tend to B2/A<40/CG series is different from the average strengths of the other
expand laterally more than the mortar, resulting in lateral compressive prism series produced with B2 blocks (COM/B2/A70/CG and COM/­
stresses in the blocks and lateral tensile stresses in the mortar. As shown B2/A>100/CG). In fact, it is around 9% below the others. It occurred
in Fig. S2 of the Supplementary Information file, the prism failed in a due to the association of an extremely flexible mortar with a consider­
brittle and explosive way by crushing of the block: the fpk of the prisms ably stiff block in the COM/B2/A<40/CG series. The mortar tended to
of those series (6.1 and 9.8 MPa, respectively) is even higher than the fbk laterally expand more than the blocks, which resulted in a significant
of the blocks they use (5.9 and 8.2 MPa, respectively). In fact, the increase of lateral tensile stresses on the block, which were added to the
compressive strength of the blocks in their confined state is higher than lateral tensile stresses resulted from the large lateral expansion of the
that in the uniaxial state. In this situation, the rupture of the structural grout. Then, the average compressive strength of the COM/­
element occurred without any significant previous warning. B2/A<40/CG series was smaller. Vertical cracks and mortar crushing
The COM/B1/A70/SG and COM/B2/A70/SG series present an configured early warning of the premature tensile splitting failure, as
intermediary situation. In these cases, the compressive strength of the presented in Fig. S6 of the Supplementary Information file.
mortar is not too low, thus preventing the joint crushing failure. The In general, the increase of the units’ strength resulted in increases of
rupture in most of the specimens occurred by transverse block splitting, the compressive strength and elastic modulus of the masonry, regardless
as shown in Fig. S3 of the Supplementary Information file. When the the ratio between the strength of mortar and blocks. Gain of compressive
compressive load reaches around 80% of the ultimate capacity of the strength of grouted and ungrouted prisms with the enhancement in the
prism, vertical cracks start to become visible in the block’s shells and blocks’ strength were larger when stronger mortars were used. For
webs, which configures a ductile rupture. The cracking increases pro­ ungrouted masonry, stiffness gain due to enhancements in the blocks’
gressively, until the moment in that the rupture happens. It is noticed strength also increased when the compressive strength of the mortar
that in these cases, the mortar is not extremely stiff, so the sudden increased. For grouted masonry, stiffness gain with enhancements in the
rupture by block crushing is prevented. blocks’ strength did not present a significant increase when changing the
Regarding the grouted prisms series produced with stronger blocks mechanical properties of the mortar.
(B2 blocks), the analysis of variance detected that the compressive For the most typical scenario, in which mortar is softer than units and
strength of the mortar affected the strength of the prism. In contrast, masonry fails by transverse blocks splitting, many authors [13,15,
when using the weaker blocks (B1 blocks), the analysis of variance 45–48] have already developed analytical and semi-empirical equations
showed that the mortar strength did not affect the prism strength. that relate the masonry compressive strength with the net area
Regardless of the type of concrete block, when comparing the elastic compressive strength of blocks and the compressive strength of mortar
modulus of the different grouted prisms series, a P-value higher than and grout, considering the functional form described by Eq. (4).
0.05 was obtained, which confirms the null hypothesis that the
fp ¼ K1 fb þ K2 fm þ K3 fg (4)
compared averages are the same. Then, according to the statistical
analysis, the mortar strength does not significantly affect the prism Based on the results of this research, multiple nonlinear regression
stiffness. analysis was carried out to develop prediction models for the masonry
In the COM/B1/A<40/CG, COM/B1/A70/CG, and COM/B1/ compressive strength in different situations. The constants K1, K2, and K3
A>100/CG series, the blocks’ net area compressive strength (10.9 MPa) were systematically varied in order to minimize the sum of square er­
is lower than the compressive strength of the grout (around 15 MPa). rors. Models are presented in Eqs. (5)–(7), which respectively consider
The elastic modulus of the grout is considerably higher than that of the the scenarios (i), (ii), and (iii) described in this work. They account for
blocks and mortar. Due to its high stiffness, the grout does not present the masonry failure modes and seem to improve the predictive ability of
very significant lateral expansion. Thus, lateral strains of the blocks is masonry strength. Eq. (6) can be used to predict the compressive
mostly restrained by the adhesion in the mortar joint. During the strength of masonry prisms when conventional block-mortar strength
application of the compressive load, cracks on the block’s shells and combinations are used, i.e., fm is between 40% and 100% of fb [1,9,17,
webs prolonged from the block/mortar interface, as presented in Fig. S4 35]. Eq. (5) can be used when concrete blocks are much stiffer than the
of the Supplementary Information file. Vertical cracks indicated that the mortar, i.e., fm is lower than 40% of fb. Eq. (7) can be used in another
rupture was about to happen. In these cases, the prism was not a ho­ specific case scenario: when the mortar is much stiffer than the concrete
mogeneous material, but a strong center column with weak shells. In blocks, i.e., fm is higher than fb.
agreement to the findings of an analysis of variance, regardless the value
fp ¼ 0:011fb þ 1:188fm þ 0:548fg (5)
of the mortar compressive strength, the ultimate stress of the prisms in
those series was around 12.2 MPa, a value between the block strength
fp ¼ 0:444fb þ 0:190fm þ 0:539fg (6)
and the grout strength. Elastic modulus also did not vary with the in­
crease of the mortar strength. fp ¼ 0:310fb þ 0:390fm þ 0:415fg (7)
A Tukey’s Test at the 5% significance level proved that the average
compressive strengths of the COM/B2/A70/CG and COM/B2/A>100/ Coefficients of determination of 0.93, 0.92, and 0.86 were obtained
CG series were equal. The same test proved that these average strengths for the regression models of scenarios (i), (ii), and (iii), respectively.
were statistically different from the average compressive strength of the They indicate that good relationships between the proposed parameters
COM/B2/A40/CG series. COM/B2/A70/CG and COM/B2/A>100/CG were obtained. It can also be observed in Fig. 2, in which experimental
series were produced with blocks that had a net area compressive results were plotted against the predicted strengths. Data points are
strength (14.8 MPa) very close to the compressive strength of the grout reasonable distributed around the equality line.

6
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

Fig. 2. Predictive ability of a) Eq. (5), b) Eq. (6), c) Eq. (7) - compressive strength.(2-column fitting image).

Two high stack bonded prisms were used in this research for deter­ adhesion. Workmanship and curing process did not change from series
mination of compressive strength and elastic modulus. However, the use to series. The initial water absorption of the B1 blocks was slightly lower
of prisms with more than two blocks, wallettes or walls in compression than that of B2 blocks. Mortar flow was around (230 � 10) mm and
tests can decrease the lateral confinement effects of the specimen pro­ mortar’s water retention was kept at a (90 � 5) % range, as recom­
vided by the loading plates of the press, leading to lower compressive mended by Parsekian and Soares [38]. Then, the main parameters that
strength values. The longitudinal deformation measured in compression changed were the mortar and the block strength. Considering these
tests of two-block prisms is also affected by the confinement effects. parameters, prediction models for flexural strength of masonry were
Another disadvantage of tests of two high stack bonded prisms is that developed using the method of least squares regression. Since all of the
this kind of specimen have only one joint, which is not able to represent prisms presented a similar rupture mode in flexural tests, scenarios (i),
the interaction between units, horizontal and vertical mortar joints [35]. (ii), and (iii) were represented together through the multiple power
Thamboo and Dhanasekar [49] state that prepend joints in the wallettes model described by Eq. (8), which presents a functional form published
also contributes to the reduction in the specimen’s strength. Based on and experimentally verified by different authors [3,15,49,52,53].
the experimental tests of 50 prisms and 40 wallettes, the authors Similarly, a regression model was created based on the results of the
concluded that wallettes compressive strength can be conservatively shear tests.
assumed as 0.67 times the prism strength. Despite these disadvantages,
σ p ¼ K4 fb α fm β (8)
most of the current design codes allow the use of compression tests of
prisms for quality control of actual masonry buildings [35], since wall The constants K4, α, and β were determined in order to minimize the
tests may be technically challenging and economically infeasible. In fact, residual sum of squares. Eq. (9) presents the obtained prediction models
two-block stack-bonded prisms are simplified models that are easy to for the flexural strength of masonry. The obtained coefficient of deter­
construct and test, considering operational and economic aspects, and mination was 0.92. Fig. 3 plots experimental results against the pre­
can still provide a reasonable estimative of the masonry structural dicted values of flexural strength. Again, data points are reasonable
behavior, which have been the concern of different recent studies [17, close to the equality line.
25,50,51]. Conclusions of these studies and results of the present work
(Eqs. (5)–(7), and figures of the Supplementary Information file) suggest σ p ¼ 0:0273fb 0:264 fm 0:661 (9)
that compression tests of two-block prisms can be used in practical case
scenarios for prediction of strength, deformability, and failure modes of 3.4. Structural behavior of prisms under shear
masonry elements.
Shear test results are also summarized in Table 6, while the rupture
mode observed in these tests is shown in Fig. S7 of the Supplementary
3.3. Structural behavior of prisms under flexure
Information file. The rupture of the block/mortar interface was also
clearly observed in the shear tests, since sudden sliding was verified
Table 6 summarizes the results of flexural tests, with some charac­
between block and mortar at the moment of the rupture.
teristics of the mortar used in each series. Fig. S7 of the Supplementary
The general behavior observed in the shear tests was the increase in
Information file shows rupture of the block/mortar interface in the
the initial shear bond strength with the increase of the compressive
flexural tests. Brittle rupture was observed in these tests due to the
weakness of the interface between the components. At the end of the
tests, blocks were practically intact.
When using B1 or B2 blocks, the flexural bond strength significantly
increased with an enhancement in the mortar strength. Therefore, a
higher amount of cement in the mortar composition provided better
mechanical interlocking between block and mortar. On the other hand,
an enhancement in the block strength did not cause an increase in the
flexural bond strength of prisms constructed with mortar whose
compressive strength is in accordance with Brazilian guidelines (FLE/
B1/A<40, FLE/B1/A70, FLE/B2/A<40 and FLE/B2/A70 series). When
the mortar compressive strength is larger than 70% of the block’s net
area compressive strength (FLE/B1/A>100 and FLE/B2/A>100 series),
the use of stronger blocks caused the increase of the flexural bond
strength of the prisms.
The observed failure mode indicates that the analysis of flexural Fig. 3. Predictive ability of Eq. (9) - flexural strength.(1.5-column
strength results must focus on the factors interfering the mortar/block fitting image).

7
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

strength of the mortar. The gain was higher when stronger blocks (B2 ratio between mortar and block’ compressive strength, and the observed
blocks) were used. When the mortar strength was increased in about rupture mode. Only data of prisms made of hollow concrete blocks, full
four times, great increases of 185% and 350% of initial shear strength mortar bedding, and joint thickness of (10 � 3) mm were considered.
were observed in series produced with B1 and B2 blocks, respectively. As However, some variations from the predicted strengths were expected
the shear bond strength is directly related to the block/mortar adhesion, due to the different geometry of hollow concrete blocks, capping ma­
a paste with higher amount of cement is better suctioned into the pores terials of specimens, loading rate of tests, press’ bearing plates thickness,
of the blocks, which improves the bond strength. For a given ratio be­ workmanship, curing and production process of specimens. Some of the
tween the mortar and block strength, it was observed that the increase in studies determined the mortar compressive strength with cylindrical
the block strength caused an increase of 41.1% in shear strength when specimens. In these situations, a factor of 1/0.85 was used to convert the
comparing the SHE/B1/A70 series to the SHE/B2/A70 series, and an compressive strength obtained from cylindrical specimens to cubic ones,
increase of 45.34% in shear strength when comparing the SHE/B1/ as suggested in Ref. [15].
A>100 series to the SHE/B2/A>100 series. Predicted values were plotted against the experimental results, as
As previously mentioned, many factors influence the mortar/block shown in Fig. 5. The accuracy of Eqs. (5)–(7) and (9) and (10) has been
bond. Then, they can directly affect the results of the prediction equa­ assessed by comparing their predictions with the experimentally ob­
tions proposed in this work. For example, the surface texture and tained strength values and quantified in terms of the error coefficient M
porosity of blocks, as well as dust and dirt on their surface decrease calculated with Eq. (11) and a coefficient P defined as the percentage of
mortar/block bond strength [35]. Poor workmanship (e.g.: mortar joints results falling within 30% of the measured values, as suggested in
not completely filled, blocks out of alignment, undesirable and unex­ Ref. [55]. M and P values calculated for each set of data are also pre­
pected movements of units over the fresh mortar) can also decrease the sented in Fig. 5. The models yield reasonably good agreement with the
brick/mortar bond strength predicted with the proposed equations. The experimental datasets compiled from the literature (M coefficients lie
absorption properties of blocks (e.g.: initial rate of absorption, sorptiv­ between 22.38% and 28.3%, with P coefficients lying between 73.77%
ity, cold and hot water absorption) and changes in mortar properties (e. and 64.74%), which suggests their appropriateness for prediction of
g.: type, flow, water retentivity, setting characteristics and air content) masonry strength.
are also reflected in mechanical strength of masonry prisms [54]. For vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uP
example, bond strength is improved if higher mortar flow is used when un
u Xp Xa

the blocks have higher water absorption, and vice-versa [35]. Different M¼
1 ti¼1
� 100 (11)
curing methods also affect masonry strength. For example, curing prisms X n 1
with plastic covering provides highest values of masonry strength, when
The predictive ability of the models developed in this work was
compared to the air curing method. On the other hand, curing masonry
compared to the predictive ability of nine different strength models
in hot weather or at lower temperatures can decrease the load capacity
available in the literature [13,17,45,47,71–74], which are presented in
of masonry elements [35].
Eq. (12) to (20) of Table 7. These models were also applied to the dataset
Since a similar rupture mode was observed in shear tests of all series,
reported by Refs. [6,25,26,56–70], and predicted and actual strength
a prediction model for shear strength of scenarios (i), (ii), and (iii) was
values were plotted together in Fig. 5. M and P values were calculated
developed through a multiple power model described by Eq. (8), using
for each prediction model and presented inFig. 6 and Table 7, in order to
the method of least squares regression. The obtained prediction model is
quantify its accuracy.
presented in Eq. (10). In Fig. 4, experimental values were plotted against
For prediction of compressive strength, the formulation of the pre­
the predicted ones. A coefficient of determination of 0.96 was obtained
sent work provided the lowest error coefficient (23.65%), followed by
and the data points are also close to the equality line.
the model proposed by Ouyang et al. [72] (38.11%). The error coeffi­
τp ¼ 0:0024fb 0:763 fm 0:825 (10) cient obtained for the flexural strength prediction model proposed in
this work (22.38%) and by Lumantarna, Biggs and Ingham [73]
(25.46%) are not much different, although the latter had a smaller
3.5. Evaluation of the predictive ability of the developed models
number of estimates falling within 30% of experimentally obtained
values (78.95%). An improvement in accuracy of prediction of cohesion
The predictive ability of Eqs. (5)–(7) and (9) and (10) was evaluated
was also provided by the model proposed in this research, since its error
in the light of a larger literature review. Values of mortar, blocks and
coefficient is 28.3%, a value lower than those obtained for the models of
grout’s compressive strength and prisms’ compressive, flexural and
Ali et al. [74] (32.98%) and Lumantarna, Biggs and Ingham [73]
initial shear strength were collected from similar published papers and
(88.64%). The models proposed by Priestley and Chai (1984) [45],
theses [6,25,26,56–70] via an extensive literature review. The appro­
Khalaf, Hendry, Fairbairn (2004) [47], Sarhat and Sherwood (2013)
priate prediction equation was applied based on the kind of test, the
[71] and Ouyang et al. [72] tend to underestimate strength, since most
of the data points are located under the equality line. On the other hand,
the model proposed by Martins et al. [17] seems to overestimate the
masonry compressive strength. The higher coefficient error and the
lower P coefficient were obtained in the cohesion model proposed by
Lumantarna, Biggs and Ingham [73].
Eqs. (5)–(7) and (9) and (10) are an improvement of previous models
because they account for the failure mechanism of the masonry elements
in different case scenarios (Table 1). Using these equations, designers
can make more accurate strength predictions and deal with the distinct
design constraints described in Refs. [22,23,33–35], which used to limit
the application of strength formulations of current masonry codes.

3.6. Final overview about the tests results


Fig. 4. Predictive ability of Eq. (10) - initial shear strength.(1.5-column
fitting image).
The experimental results of this work indicate that when low
strength units are used in ungrouted and grouted masonry, statistically

8
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

Fig. 5. Evaluation of the predictive ability of models developed in this study for determination of compressive strength, flexural strength, and initial shear strength of
masonry prisms, through comparisons with data reported in different works. (2-column fitting image).

significant changes in masonry compressive strength are not observed rupture of the block/mortar interface. When weak units were used, the
when the ratio between mortar and block strength is increased. On the flexural bond strength and the initial shear bond strength increased in
other hand, statistically significant changes in compressive strength of 89.8% and 85.1% respectively, when the ratio between mortar and
ungrouted masonry were verified when the strong blocks were used. In block strength increased from 35.8% to 131.2%. When strong units were
ungrouted and grouted prisms, compressive strength increases of 84.4% used, very high load capacity increases were also observed: the flexural
and 15.1%, respectively, were verified when the ratio between mortar bond strength and the initial shear bond strength increased in 173.8%
and block net area strength increased from 39.9% to 115.5%. High in­ and 250% respectively, when the ratio between mortar and block
creases of mortar/block bond strength were also verified in flexural and strength increased from 29.0% to 124.3%.
shear tests with the enhancement of the relative strength of mortar and B1 blocks are 23% cheaper than B2 blocks. Smaller differences are
blocks. All of these tests presented the same failure mode: a brittle verified when the costs of the different mortar compositions are

9
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038
compared. The weakest mortar (B1/A<40) is around 6% cheaper than
59.02%

62.30%
18.03%

34.43%

37.70%
70.49%
78.95%
62.50%

the strongest one (B2/A>100). Considering techno-economic aspects


6.25%

the use of weaker mortars (B1/A<40) and weaker blocks (B1 blocks)
P

seems to be interesting for the construction of single-family residential


buildings, since it reduces material expenses. The intensity of the loads
in this kind of structure is not very high. The ungrouted masonry pre­
sents a low compressive strength and stiffness, and the rupture occurs in
a ductile way by localized mortar crushing. Grout contributes for a
46.03%

48.78%
52.00%

88.60%

77.83%
38.11%
25.46%
32.98%
88.64%

significant increase in the compressive strength and stiffness of the


M

masonry. Progressive cracking of the block walls can be observed before


the rupture. The adhesion between the block and the mortar is not very
high, which compromises the flexural and shear bond strength of the
masonry. In this situation, the enhancement of the mortar strength does
not provide considerable gains of compressive strength of the masonry,
strengths. Priestley and Chai [45] also mention that
φ ¼ 0.75 would provide an effective lower bound.
represent a theoretical curve in terms of average

However, in this work comparisons are made in

but it improves the block/mortar bond strength.


​ ​
A φ factor equal to 1.0 was used, in order to

The coefficient Ch was obtained through the

�� ​ ​ if ​ ​ b < 5

For multi-storey masonry buildings, mortar with intermediate


strength (B2/A70) and blocks with higher strength (B2 blocks) seem to
t
h

be more interesting from a balanced technical and economic point of


view. Ungrouted masonry presents a considerable compressive strength
and stiffness, and the rupture occurs by transverse block splitting. Again,
terms of average strengths.

< 1 0:075 5 hb

: 1 ​ ​ ​ ​ if ​ ​ hb � 5
t

grout provides a considerable increase in the compressive strength and


stiffness of the masonry. When using grout with compressive strength
t

1
following equation:

close to the blocks’ net area compressive strength, the system acts more
homogeneously and the masonry strength increases. Furthermore, the
block/mortar bond strength is enough to provide a considerable flexural
Comments

>�
8
>
>
>
>

>
>
>
>
>

and shear strength.


Ch ¼

4. Conclusions
The results of this work provide important insights related to the
mechanisms associated with the masonry behavior, which help de­
fp ¼ 0:81Ch ​ ð0:287fb þ 0:114fm þ 0:252fg þ 0:62Þ

signers in the selection of the most adequate components to support


different kinds of loads, accommodate structure strains, and avoid un­

!

fb 1:2

desirable ruptures modes. The impact of the incompatibility of the


fg
αÞ ​ fg g ​

components’ characteristics on the performance of the structure was


αÞ
fp ¼ 1:57 ​ lnðfm Þ þ 0:75fb ​ þ 5:81 ​ ln
Strength formulations proposed in previous studies and comparisons with results of the present study.

better elucidated, considering different challenging issues sometimes


βÞfb ​ þ 0:90 ​ β ​ fg
fp ¼ 0:30fb ​ þ 0:20fm ​ þ 0:25fg ​

fp ¼ 1:335 ​ fb ​ α þ 0:616 ​ fg ​ ð ​ 1

faced by designers: the use of blocks considerably stronger and stiffer


fp ¼ φf ​ 0:59 ​ α ​ fb þ 0:90 ​ ð1

than the mortar, and vice versa. For these distinct scenarios, consistent
models for prediction of compressive strength, flexural bond strength,
and initial shear bond strength were created in order to account for the
τp ¼ 0:0326fm 0:6633 ​

masonry failure modes and improve the predictive ability of strength


σp ¼ 0:0263fm ​

τp ¼ 0:0471fm ​
fp ¼ 0:64 ​ ð ​ 1

equations of current design codes. The developed models satisfactorily


Formulation

fitted laboratory-measured data reported in the literature and proved to


be more accurate than other models reported in previous studies. The
existing database about the subject was complemented with results of
experimental tests of 100 masonry prisms constructed with hollow
concrete blocks and cement-lime mortars of different strength levels.
Recommendations of the most appropriate ways to combine masonry
components in distinct case scenarios were given, in order to guarantee
Lumantarn, Biggs, Ingham (2014) [73]

Lumantarn, Biggs, Ingham (2014) [73]


Khalaf, Hendry, Fairbairn (2004) [47]
Koksal, Karakoç, Yildirim (2005) [13]

an economic and safe construction system.


Sarhat and Sherwood (2013) [71]

This research focused on comparisons between mechanical proper­


Priestley and Chai (1984) [45]

ties and analyses of failure modes of masonry elements made of distinct


block/mortar combinations. It needs to be complemented with more

Ouyang et al. (2019) [72]


Martins et al. (2018) [17]
detailed investigations on these specific case scenarios, such as those

Ali et al. (2012) [74]


related to the nonlinear behavior of the composite, damage assessment,
stiffness plasticity degradation, variation of the Poisson’s ratio in com­
ponents under different confinement levels, and post-peak responses of
the specimens. This paper aimed to evaluate the masonry behavior using
Source

stack-bonded prisms under compression, flexure and shear. They are


simplified models that provide reasonable estimative for important pa­
rameters of the design and quality control of actual masonry structures,
since full-scale wall tests may be technically challenging and economi­
cally infeasible. However, more details about the structural behavior of
masonry can be determined with improvements of the test methods of
Eq. (12)

Eq. (13)
Eq. (14)

Eq. (15)

Eq. (16)
Eq. (17)
Eq. (18)
Eq. (19)
Eq. (20)
Table 7

# this work (e.g.: measurement of the post-peak stress-strain behavior,


transducers directly fixed to the specimens, larger number of specimens,
10
G. Henrique Nalon et al.

11
Fig. 6. Evaluation of the predictive ability of strength models previously reported in the literature.(2-column fitting image).
Journal of Building Engineering 28 (2020) 101038
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

evaluation of different pre-compression levels in shear tests, tests of [21] M. Gayed, Y. Korany, G. Sturgeon, Examination of the Prescribed Concrete Block
Masonry Compressive Strengths in the Canadian Masonry Design Standard, CSA
walls, wallettes or prisms with more than two blocks).
S304.1-2004, 15th IB2MaC, Florianopolis, Brazil, 2012.
[22] G. Mohamad, F.S. Fonseca, T. Vermeltfoort, D.R.W. Martens, P.B. Lourenço,
Funding Strength, behavior, and failure mode of hollow concrete masonry constructed with
mortars of different strengths, Constr. Build. Mater. 134 (2017) 489–496.
[23] S.S. Prakash, M. Aqhtarudin, J.S. Dhara, Behaviour of soft brick masonry small
This study was financed in part by the Coordenaça ~o de Aperfeiçoa­ assemblies with and without strengthening under compression loading, Mater.
mento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. Struct. 49 (2016) 2919–2934.
[24] J.A. Thamboo, M. Dhanasekar, C. Yan, Flexural and shear bond characteristics of
thin layer polymer cement mortared concrete masonry, Constr. Build. Mater. 46
Declaration of competing interest (2013) 104–113.
[25] F.F. Fonseca, E.S. Fortes, G.A. Parsekian, J.S. Camacho, Compressive strength of
high-strength concrete masonry grouted prisms, Constr. Build. Mater. 202 (2019)
None. 861–876.
[26] M. Bolhassani, A.A. Hamid, A.C.W. Lau, F. Moon, Simplified micro modeling of
Acknowledgements partially grouted masonry assemblages, Constr. Build. Mater. 83 (2015) 159–173.
[27] T.S. Cheema, R.E. Klinger, Compressive strength of concrete masonry prisms,
J. Am. Concr. Inst. 83 (1986) 88–97.
This study was financed in part by the Coordenaça ~o de Aperfeiçoa­ [28] J.A. Thamboo, M. Dhanasekar, C. Yan, Effects of joint thickness, adhesion and web
mento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. shells to the face shell bedded concrete masonry loaded in compression, Aust. J.
Struct. Eng. 14 (2013) 291–302.
The authors also thank the support provided by the Civil Engineering [29] F.B. Lima, A.N. Lima, W.S. Assis, Study of the influence of compressive strength
Department of the Federal University of Viçosa. The authors are also and thickness of capping-mortar on compressive strength of prisms of structural
very grateful to Dr. Rita de Ca �ssia Silva Sant’Ana Alvarenga for sug­ clay blocks, in: 15th International Brick and Block Masonry Conference,
Florianopolis, Brazil, 2012.
gesting the research topic and providing valuable guidance during the [30] M.J. Cassinello, Effect of mortar joint thickness on deformability in medieval stone
early stages of this research project. walls, Mater. Construcci� on 56 (2006) 69–80.
[31] A.J. Francis, C.B. Horman, L.E. Jerrems, The effect of joint thickness and other
factors on the compressive strength of brickwork, in: 2nd Int. Brick Masonry Conf.,
Appendix A. Supplementary data British Ceramic Association, 1972.
[32] B.V.V. Reddy, R. Lal, K.S.N. Rao, Influence of joint thickness and mortar-block
Supplementary data to this article can be found online at https://doi. elastic properties on the strength and stresses developed in soil-cement block
masonry, J. Mater. Civ. Eng. 21 (2009) 535–542.
org/10.1016/j.jobe.2019.101038. [33] R.A. Oliveira, F.A.N. Silva, C.W.A.P. Sobrinho, A.C. Azevedo, J.M.P.Q. Delgado, A.
S. Guimar~ aes, Structural performance of unreinforced masonry elements made
References with concrete and horizontally perforated ceramic blocks – laboratory tests, Constr.
Build. Mater. 182 (2018) 20–34.
[34] G. Sarangapani, B.V. Reddy, K.S. Jagadish, Brick-mortar bond and masonry
[1] M.A. Ramalho, M.R.S. Corr^ ea, Design of Structural Masonry Buildings, 2003. Pini,
compressive strength, J. Mater. Civ. Eng. 17 (2005) 229–237.
Sao Paulo.
[35] G.A. Parsekian, A. Hamid, G. Drysdale, Structural Masonry Behavior and Design,
[2] S. Sazedj, A.J. Morais, S. Jalali, Comparison of environmental benchmarks of
EdUFSCar, Sao Carlos, Brazil, 2012.
masonry and concrete structure based on a building model, Constr. Build. Mater.
[36] ABNT NBR 6118, Design of Concrete Structures — Procedure, 2014. Rio de
141 (2017) 36–43.
Janeiro.
[3] H.B. Kaushik, D.C. Rai, S.K. Jain, Stress-strain characteristics of clay brick masonry
[37] C.S. Barbosa, Strength and Deformability of Hollow Concrete Blocks, Prisms and
under uniaxial compression, J. Mater. Civ. Eng. 19 (9) (2007) 728–739, https://
Walls and Their Correlation to Mechanical Properties of Constituent Materials,
doi.org/10.1061/(ASCE)0899-1561(2007)19:9(728).
Thesis, University of Sao Paulo, 2008.
[4] F. Porto, F. Mosele, C. Modena, Compressive behavior of a new reinforced masonry
[38] G.A. Parsekian, M.M. Soares, Clay Blocks Structural Masonry - Design, Execution
system, Mater. Struct. 44 (2011) 565–581.
and Control, Telo Melo, Sao Carlos, Brazil, 2010.
[5] Q. Zhou, F. Wang, F. Zhu, X. Yang, Stress–strain model for hollow concrete block
[39] ABNT NBR 15961-2, Structural Masonry - Part 2: Execution and Site Control, 2011.
masonry under uniaxial compression, Mater. Struct. 50 (2017) 106.
Rio de Janeiro.
[6] C.S. Barbosa, P.B. Lourenço, J.B. Hanai, On the compressive strength predictions
[40] ASTM E518-15, Standard Test Methods for Flexural Bond Strength of Masonry,
for concrete masonry prisms, Mater. Struct. 43 (2010) 331–344.
2015. West Conshohocken.
[7] T. Zahra, M. Dhanasekar, Prediction of masonry compressive behaviour using a
[41] CSN EN 1052-3, Methods of Test for Masonry - Part 3: Determination of Initial
damage mechanics inspired modelling method, Constr. Build. Mater. 109 (2016)
Shear Strength, 2002. Brussels.
128–138.
[42] A. Costigan, S. Pavía, O. Kinnane, An experimental evaluation of prediction models
[8] K.V.M. Rao, B.V.V. Reddy, K.S. Jagadish, Flexural bond strength of masonry using
for the mechanical behavior of unreinforced, lime-mortar masonry under
various blocks and mortars, Mater. Struct. 29 (1996) 119–124.
compression, J. Build Eng. 4 (2015) 283–294.
[9] R.C.S.S. Alvarenga, G.H. Nalon, L.A.F. Fioresi, M.C. Pinto, L.G. Pedroti, J.C.
[43] A. Drougkas, E. Verstrynge, R. Hayen, K.V. Balen, The confinement of mortar in
L. Ribeiro, Experimental evaluation of the influence of mortar’s mechanical
masonry under compression: experimental data and micro-mechanical analysis,
properties on the behavior of clay masonry, in: Characterization of Minerals,
Int. J. Solids Struct. 162 (2019) 105–120.
Metals, and Materials, 2017, pp. 671–679.
[44] M.A. Amjad, Elasticity and Strength of Masonry, Units and Mortar, Thesis,
[10] A. Drougkas, P. Roca, C. Molins, Compressive strength and elasticity of pure lime
University of Leeds, 1990.
mortar masonry, Mater. Struct. 49 (2015) 983–999.
[45] M.J.N. Priestley, Y.H. Chai, Prediction of masonry strength from constituent
[11] V.G. Haach, G. Vasconcelos, P.B. Lourenço, Assessment of compressive behavior of
properties, New Zealand Concr. Constr. (1984).
concrete masonry prisms partially filled by general mortar, J. Mater. Civ. Eng. 26
[46] R.H. Atkinson, G.R. Kingsley, G.G. Yan, A database for compressive stress-strain
(2014).
behaviour of masonry, in: Fifth North American Masonry Conference 2, 1990,
[12] L. Huang, L. Liao, L. Yan, H. Yi, Compressive strength of double H concrete block
pp. 581–593.
masonry prisms, J. Mater. Civ. Eng. 26 (2014).
[47] F.M. Khalaf, A.W. Hendry, D.R. Fairbairn, Study of the compressive strength of
[13] H.O. Koksal, C. Karakoç, H. Yildirim, Compression behavior and failure
blockwork masonry, ACI Struct. J. 91 (1994) 367–375.
mechanisms of concrete masonry prisms, J. Mater. Civ. Eng. 17 (2005) 107–115.
[48] C. Dymiotis, B.M. Gutlederer, Allowing for uncertainties in the modelling of
[14] A.A. Hamid, A.O. Chukwunenye, Compression behavior of concrete masonry
masonry compressive strength, Constr. Build. Mater. 16 (2002) 443–452.
prisms, J. Struct. Eng. 112 (1986) 605–613.
[49] J.A. Thamboo, M. Dhanasekar, Correlation between the performance of solid
[15] S.R. Sarhat, E.G. Sherwood, The prediction of compressive strength of ungrouted
masonry prisms and wallettes under compression, J. Build Eng. 22 (2019)
hollow concrete block masonry, Constr. Build. Mater. 58 (2014) 111–121.
429–438.
[16] J.Z. Heydariha, S. Das, B. Banting, Effect of grout strength and block size on the
[50] A. Lübeck, G. Mohamad, F.S. Fonseca, L.E. Modler, R.P.B. Schmidt, Compressive
performance of masonry beam, Constr. Build. Mater. 157 (2017) 685–693.
strength and failure mode of axially loaded hollow concrete block masonry, in:
[17] R.O.G. Martins, G.H. Nalon, R.C.S.S. Alvarenga, L.G. Pedroti, J.C.L. Ribeiro,
13th Canadian Masonry Symposium, Halifax, Canada, June 4-7, 2017.
Influence of blocks and grout on compressive strength and stiffness of concrete
[51] R.P. Portella, G. Mohamad, F.S. Fonseca, A. Lübeck, R.P.B. Schmidt, L.E. Modler,
masonry prisms, Constr. Build. Mater. 182 (2018) 233–241.
Efficiency ratio of hollow clay block masonry, in: 10th Australasian Masonry
[18] ABNT NBR 15961-1, Structural Masonry – Concrete Blocks – Part 1: Design, 2011.
Conference, Sydney, Australia, February 11-14, 2018.
Rio de Janeiro.
[52] K.S. Gumaste, K.S.N. Rao, B.V.V. Reddy, K.S. Jagadish, Strength and elasticity of
[19] C.F.R. Santos, R.C.S.S. Alvarenga, J.C.L. Ribeiro, L.O. Castro, R.M. Silva, A.A.
brick masonry prisms and wallettes under compression, Mater. Struct. 40 (2007)
R. Santos, G.H. Nalon, Numerical and experimental evaluation of masonry prisms
241–253.
by finite element method, IBRACON Struct. Mater. J. 10 (2017) 493–508.
[20] Y. Korany, J. Glanville, Comparing masonry compressive strength in various codes,
Concr. Int. 27 (2005) 35–39.

12
G. Henrique Nalon et al. Journal of Building Engineering 28 (2020) 101038

[53] F.C. Christy, D. Tensing, R.M. Shanthi, Experimental study on axial compressive [64] A.A. Hamid, R.G. Drysdale, Flexural tensile strength of concrete block masonry,
strength and elastic modulus of the clay and fly ash brick masonry, J. Civ. Eng. J. Struct. Eng. 114 (1988) 50–66.
Construct. Technol. 4 (2013) 134–141. [65] P. Raposo, A. Furtado, A. Ar^ ede, H. Rodriges, H. Varum, Mechanical
[54] C.G. Yuen, S.L. Lissel, Flexural bond strength of clay brick masonry, WIT Trans. characterization of existing infill panels of concrete block on Azores islands,
Eng. Sci. 57 (2007) 253–262. Portuguese J. Struct. Eng. 3 (2017) 49–62.
[55] J.J. Brooks, Concrete and Masonry Movements, Elsevier Inc., Waltham, MA, 2015, [66] S. Zhang, D. Yang, Y. Sheng, S.W. Garrity, L. Xu, Numerical modelling of FRP-
p. 599. reinforced masonry walls under in-plane seismic loading, Constr. Build. Mater. 134
[56] L.M.L. Calçada, Evaluation of the Behavior of Grouted and Ungrouted Concrete (2017) 649–663.
Block Prisms, Thesis, Federal University of Santa Catarina, 1998. [67] M.C. Candia, L.S. Franco, Shear Adherence Strength of Unreinforced Clay Masonry,
[57] R.O. Steil, Effect of Block Geometry and Type of Mortar on the Performance of Technical Bulletin of the Polytechnic School of the University of S~ao Paulo, S~ao
Ungrouted Prisms under Compression, Thesis, Federal University of Santa Paulo, 1995.
Catarina, 2003. [68] H.F. Martins, Shear Strength of Concrete Blocks Structural Masonry, Thesis,
[58] R.P. Andolfato, J.S. Camacho, R.M. Mauricio, Study of structural masonry concrete Federal University of Santa Catarina, Florian� opolis, 2001.
block units and prisms varying the blocks cement quantity and degree of [69] I. Grilo, H. Marques, J.P. Gouveia, E. Júlio, Shear Strength of Masonry Walls of
compaction, in: Proceedings of the 7th Australasian Masonry Conference, Concrete and Clay Blocks, National Meeting of Structural Concrete, FEUP, 2012.
Newcastle, Australia, 2004, pp. 150–160. [70] G. Andreotti, F. Graziotti, G. Magenes, Detailed micro-modelling of the direct shear
[59] G. Mohamad, P.B. Lourenço, H.R. Roman, Mechanics of hollow concrete block tests of brick masonry specimens: the role of dilatancy, Eng. Struct. 168 (2018)
masonry prisms under compression: review and prospects, Cement Concr. Compos. 929–949.
29 (2007) 181–192. [71] S.R. Sarhat, E.G. Sherwood, The Prediction of Compressive Strength of Grouted
[60] J. Liu, The Effect of Height-To-Thickness Ratio on the Compressive Strength of Hollow Concrete Block Masonry Based on the Contributions of its Individual
Concrete Masonry, Thesis, University of Windsor, 2012. Components, in: 12th Canadian Masonry Symposium, Vancouver, British Columbia,
[61] O.S. Izquierdo, M.R.S. Corr^ea, I.I. Soto, The influence of mortar bedding on the June 2-5, 2013.
compressive strength of concrete block masonry structures, in: Proceedings of the [72] J. Ouyang, F. Wua, W. Lü, H. Huang, X. Zhou, Prediction of compressive stress-
15th International Brick and Block Masonry Conference, Florian� opolis, Brazil, strain curves of grouted masonry, Constr. Build. Mater. 229 (2019) 116826.
2012, pp. 1–10. [73] R. Lumantarna, D.T. Biggs, M. Dist, J.M. Ingham, Uniaxial compressive strength
[62] L.O. Castro, R.C.S.S. Alvarenga, R.M. Silva, J.C.L. Ribeiro, Experimental evaluation and stiffness of field-extracted and laboratory-constructed masonry prisms,
of the interaction between strength concrete block walls under vertical loads, J. Mater. Civ. Eng. 26 (2014) 567–575.
IBRACON Struct. Mater. J. 9 (2016) 643–681. [74] Q. Ali, Y.I. Badrashi, N. Ahmad, B. Alam, S. Rehman, F.A.S. Banori, Experimental
[63] F.E. Caldeira, Evaluation of Influence of Bedding Mortar Properties on the Stiffness investigation on the characterization of solid clay brick masonry for lateral shear
and Strength of Structural Masonry, Thesis, Federal University of Viçosa, 2018. strength evaluation, Int. J. Earth Sci. Eng. 5 (2012) 782–791.

13

You might also like