Contents lists available at ScienceDirect

Engineering Structures
journal homepage: [Link]/locate/engstruct

Effects of vertical stresses and flanges on seismic behavior of unreinforced

brick masonry
⁎
Hizb Ullah Sajid , Mohammad Ashraf, Qaisar Ali, Sikandar Hayat Sajid
Department of Civil Engineering, University of Engineering and Technology, Peshawar, Pakistan

A R T IC L E IN FO A BST R A CT

Keywords: Lateral in-plane response plays crucial role in seismic behavior of masonry structures. The aim of this article is to
Brick masonry experimentally investigate the effects of vertical stresses and flanges (transverse walls) on the lateral in-plane
Vertical stresses response of the unreinforced brick masonry (URBM) walls. The experimental work included lateral in-plane
Flange effects quasi-static cyclic tests on full-scale walls (both with & without flanges). The vertical stresses resulting from
In-plane response
typical single and two story unreinforced masonry buildings were simulated on full scale URBM walls. Flanges
Seismic behavior
were introduced at both ends of the in-plane wall in single direction. In essence, the lateral in-plane stiffness &
strength, deformability and energy dissipation of the two classes of walls are compared and the differences are
quantified to help understand the effects of flanges on the in-plane response of masonry walls. The resulting
damage mechanism and failure modes for each case are critically analyzed. The experimental results indicate
that both vertical stresses and flanges incorporation significantly improved seismic response of URBM walls. In
addition, the participation of flanges is critical in both vertical stress conditions.

1. Introduction (flanges) on the in-plane behavior of the URBM structures.

Vertical stresses are inherently present in all load-bearing URBM
Brick masonry is one of the most widely used building material in structures. Vertical stresses resulting from the diaphragm, adjacent
construction industry throughout the world. For centuries, masonry parts of the buildings and occupancy loads, constitute one of the im-
construction has been preferred primarily due to plentiful and in- portant parameter which influence the behavior of URBM under the
expensive availability of constituent materials and ease of construction. action of lateral loads. A number of researchers have studied vertical
Historically, such type of construction is located in areas that are prone stresses for their effects on masonry behavior and response. Research
to seismic activity [1]. URBM offers a weak response when subjected to work carried out by (Magenes, G., Calvi [15], Irimies & Bia [16]; Yi
lateral loads such as earthquake. As witnessed in 2005 Kashmir, Paki- et al. [17]; Haach et al. [18]; Tomaževič & Weiss [19]; Salmanpour
stan earthquake, and November 2015 earthquake in Afghanistan and et al. [20]; Salmanpour et al. [21], Salmanpour et al. [22]; Javed et al.
Pakistan, URBM suffered heavy damages [2–8]. Masonry structures [23]) utilized both solid and hollow brick and block masonry to study
(brick and stone) suffered yet again in the more recent earthquakes of the effects of vertical stresses. Quasi-static cyclic testing has been per-
2009 and 2016 in Italy [9–12] and 2015 in Nepal [13,14]. These events formed in most of these studies due to its advantages over other types of
have further strengthened the need of continuing research to under- testing techniques [24].
stand and improve unreinforced masonry construction. It is evident, Most of these researchers tested isolated masonry piers or walls
that such type of construction, although more economical and popular, either full scale or half scaled. However, Irimies & Bia [16] also in-
pose high risk to the safety of masses during earthquakes. Extensive corporated flanges in the specimens. Axial loads, masonry unit types,
research has been carried out and more work is in progress to under- geometrical characteristics and material strengths were varied to study
stand the behavior of masonry structures and to develop better design their effects on response parameters. All the researchers concluded that
guidelines for the same. vertical stress is a principal parameter that affects the strength and
Seismic capacity of masonry structure is affected by a variety of deformability characteristics of the masonry.
parameters that include aspect ratio, level of vertical stress, mechanical Apart from vertical stresses, boundary conditions, particularly
properties of masonry and its constituents and flanges This paper in- flanges, also play important role in the load in the behavior of URBM.
vestigates the effects of vertical stresses and boundary conditions Boundary conditions though inherently present in the brick masonry

⁎
Corresponding author.
E-mail addresses: hezb_uetian@[Link] (H.U. Sajid), engineerashraf@[Link] (M. Ashraf), drqaisarali@[Link] (Q. Ali), shsajid1@[Link] (S.H. Sajid).

[Link]
Received 21 April 2017; Received in revised form 3 October 2017; Accepted 8 November 2017
H.U. Sajid et al.

Load application Load application

LVDT LVDT

Masonry panel Masonry panel

Fig. 1. Compressive strength testing of brick masonry panels.

walls, are frequently ignored in the research owing to their limited 2. Experimental program
contribution. Limited research is available in literature on the con-
tribution of flanges on the deformability and strength characteristics of To investigate the effects of vertical stresses and flanges on strength
URBM walls. Earliest studies on masonry carried out by Tomaževič and deformability characteristics of URBM, a detailed experimental
et al. [25], Costley and Abrams [26], and Tomaževič [1] suggested the scheme was designed. The experimental program included selection of
need to evaluate flanges effects on masonry response. More recent work representative constituent materials for construction of test specimens,
on flanges participation was reported by Paquette and Bruneau [27], determination of constituent material properties, construction of full-
Moon [28] and Yi et al. [29]. scale specimens and quasi-static cyclic testing of specimens.
Moon [28] carried out one of the earliest and most comprehensive
work to investigate the effects of flanges on capacity and response of 2.1. Material properties
URBM. Moon [28] conducted lateral load tests on a two storey full scale
unreinforced masonry (URM) structure before and after retrofitting. Prior to fabrication of specimens, a preliminary study was carried
The variables of interest included overturning effects, flanges partici- out to select representative materials to be used in the specimens. All
pation and global rocking. Based on the results, a number of mod- the specimens were constructed using baked clay solid bricks having
ifications has been suggested in the current FEMA 356 model [30] for nominal size of 229 × 114 × 76 mm that are commonly used in the
in-plane analysis of perforated walls. For the purpose of modification in local construction industry in Pakistan. For brick units, initial rate of
the FEMA 356 [30] model, flanges have been divided into three cate- absorption and compressive strength were determined according to
gories which include compressions flange, global tension flange and ASTM C67-16 [36] and the cement mortar compressive strength was
component tension flange. The research observed considerable parti- determined following ASTM C109 [37]. The water-cement ratio was
cipation of each category of flange in influencing the capacity and decided based on the consistency of the mortar in order to achieve a
failure modes in in-plane URBM walls. workable mix. To characterize the compressive strength and diagonal
Yi et al. [31] developed a new pier model to investigate the effects tension behavior of masonry, masonry assemblages were prepared and
of flanges on the behavior and response of URM piers. The proposed tested in compliance with test protocols of ASTM C1314 [38] and ASTM
model is modification of the effective pier model developed by Yi et al. E59 [39], respectively, as shown in Fig. 1. Table 1 represents me-
[29]. The modified model is analyzed for a non-rectangular URM pier to chanical properties of constituent materials.
study flange effects. Russell and Ingham [32], Russell and Ingham [33]
and Russell et al. [34] studied flanges contribution for material and
loading characteristics experienced in brick masonry structures in New 2.2. Construciont of full scale walls
Zealand. Their results confirmed the accuracy of the model proposed by
Yi et al. [31] for estimating strength of flanged in-plane wall. Fur- In order to investigate the effects of variables under consideration, a
thermore, the results emphasized the effects of flanges on failure mode total of four (04) full-scale URBM specimens were constructed. The
and lateral strength of unreinforced masonry. specimens were named as S1, SF1, S2 and SF2. Specimens S1 and S2 are
Most recent work on flanges contribution in URBM has been re- full scale perforated walls without incorporation of flanges. Specimens
ported by Khanmohammadi et al. [35] who tested four half scaled SF1 and SF2 are full scale perforated walls with incorporation of
masonry specimens. The research work mainly focuses on block ma- flanges. The numeric values, 1 and 2, in the specimen IDs (S1, S2, SF1,
sonry, with either solid or perforated geometric nature and made
Table 1
comparisons between experimental work and proposed models that
Mechanical properties of constituent materials.
yielded acceptable prediction under controlled conditions.
Due to limited work on flanged unreinforced brick masonry Material Property Results Testing
(FURBM), further investigation is required to yield data for developing (average) Standard
acceptable analytical models for capacity and behavior prediction of
Brick units Compressive strength 16.20 ASTM C67-3a
masonry. Furthermore, due to the variability in material characteristics (MPa)
and different geometric as well as loading configurations that were Brick units Initial rate of absorption 0.07 ASTM C67-14
applied by various researchers, the results obtained do not have (kg/min)
straightforward application to the masonry structures that constitute Cement mortar Compressive strength 7.95 ASTM C109
(1:4) (MPa)
building stock of developing countries. The research work presented in
Brick masonry Compressive strength 4.34 ASTM C1314
this article encompasses the effects of target variables on baked clay (MPa)
solid brick units that are typically employed in construction industry in Brick masonry Diagonal tensile strength 0.17 ASTM E519
developing countries. (MPa)
H.U. Sajid et al.

Fig. 2. Geometric detailing of specimen S1, SF1 (isolated masonry wall) and S2, SF2 (flanged masonry wall).

masonry walls using dead load of steel beam and vertical load cell.
Fixed-cantilever condition was simulated in both specimens. Geometric
as well as loading details of test specimens are given in Table 2.

2.3. Test setup

Instrumentation plan for typical test specimens is shown in Fig. 5.

Linear variable displacement transducers (LVDTs) were installed for
recording displacements at various target locations to capture wall
mechanism, as shown in Fig. 5. LVDT-1 was installed on the front face
at the top of the wall, at horizontal level, to record the horizontal dis-
placement of the wall. This transducer also served as control gauge.
LVDT-2 and LVDT-3 were installed at top of north and south pier re-
spectively, to record top displacement of individual piers. LVDT-4 was
installed at mid point of lintel beam. The data from this transducer will
Fig. 3. Typical test specimen. be used to establish sliding of piers, if any. LVDT-5 was installed at the
bottom of the window opening. LVDT-6 is provided at vertical level, at
SF2) represent the vertical stress levels that are typically experienced in the bottom of the wall to study global rocking of the specimen. LVDT-7
single and double storey residential room (366 cm × 457 cm), respec- is provided at horizontal level, at the bottom of the wall, to capture
tively. Each specimen was 296 cm long and 304 cm tall and have a global sliding of the specimen. LVDT-8 and LVDT-9 were installed at the
thickness of 225 mm. English bond pattern was utilized in the speci- top corners of the wall, in the out-of-plane direction. The purpose of
mens. Average thickness of mortar joints was around 1.25 cm. Typical these transducers were to capture any possible out-of-plane rotation of
test specimens are shown in Figs. 2 and 3. Specimen S1 and S2 are the specimen so that test can be stopped in the event of excessive out-of-
isolated walls with no flanges at the ends. Specimen SF1 and SF2 are plane rotation of the specimen. LVDTs installed on the specimens were
provided with flanges along one direction at both ends. Each specimen capable of recording a linear displacement of 50 mm in a single direc-
contains two piers with identical geometric characteristics. A 17-cm tion. For recording displacement in both negative and positive direction
thick lintel beam is provided in both specimens. of loading, the LVDTs were set at 25 mm at the start of the test. Hence,
Fabrication process of full-scale URBM specimens was divided in they were able to measure a displacement up to 25 mm in either posi-
two phases. In the first phase, reinforced concrete pad was constructed tive or negative direction, which was within the anticipated displace-
to serve as platform for the specimen. In the second phase, full-scale ment range of test specimens. LVDTs provide a convenient mean to
URBM specimens were constructed. The platform for the brick masonry record linear displacement in masonry structures and have been used
specimens was constructed using 1:2:4 concrete mix. Upon this plat- successfully in previous studies.
form the test walls were then fabricated. In order to make sure that All the test specimens were tested inside the loading frame of
platform will be stable and remain intact during testing, the footing was Structural Engineering Laboratory at Department of Civil Engineering,
anchored to the floor using anchor-bolt system, as shown in Fig. 4. University of Engineering and Technology, Peshawar - Pakistan. On
Epoxy grout was used for installing anchoring system in the floor. global level, the specimens were fixed at bottom and were free to
Lateral loading is imposed on the specimens with a reinforced translate at the top, thereby simulating cantilever boundary condition.
concrete beam on each specimen. Vertical stresses were applied on Piers were in double fixed condition. Vertical and horizontal loads were
applied on the specimens using hydraulic jacks that were operated
H.U. Sajid et al.

Fig. 4. Geometric detailing and finished platform for test specimens.

Table 2 same loading conditions are simulated in both push and pull (positive
Geometric and loading details of test specimens. and negative displacement) cycles. Two steel plates were placed on the
top R.C. beam. Lubricated steel rollers were placed on each steel plate.
Specimen ID S1, SF1 S2, SF2
A steel beam was placed on this assembly and vertical load was then
Wall Dimensions L (cm) 296 296 applied to the steel beam. This mechanism was used to simulate the
H (cm) 304 304 actual lateral loading conditions experienced by a full-scale masonry
Pier Dimensions Lp (cm) 110 110
wall. In case of specimen SF1, the vertical load was small, hence it was
Hp (cm) 122 122
tp (cm) 22.5 22.5
applied on flanges using dead load of steel beams and sand bags.
Aspect Ratio 1.11 1.11 However, in case of specimen SF2, the vertical load was not manage-
Vertical stress at top of wall (MPa) 0.05 0.16 able using dead loads, hence hydraulic jack load was used for both in-
Vertical stress at top of pier (MPa) 0.08 0.23 plane and out-of-plane walls. For this purpose, two short steel beams
were separately designed to carry the target loads. The load from ver-
tical load cell was transferred to the main steel beam first. The load
using a hydraulic pump. Load cells, having capacity of 500 kN, were
from the steel beam was transferred to short steel beams, from where it
used to apply horizontal and vertical loads. Loading shoes were at-
was distributed to in-plane and out-of-plane walls. The location of short
tached to both ends of the top concrete beam. This was two ensure that

Fig. 5. Instrumentation plan of typical test specimen.

8 1 9

2 3

6
7
H.U. Sajid et al.

Fig. 6. Instrumentation and loading assembly of

Steel Beam Loading specimen SF1.
Shoe
Steel Roller Steel Plates
Loading
Shoe
Horizontal
Load Cell

North South
Pier Pier

Test Specimen SF1

Fig. 7. Instrumentation and loading assembly of

Steel Rollers specimen SF2.

Steel Beam

Short Steel
Beams

Loading Loading Horizontal

Shoe Shoe Load Cell

North South
Pier Pier

Test Specimen SF2

beams and long steel beam was set such that desired load levels are 2.4. Test procedure
produced at both in-plane and out-of-plan walls. The horizontal loads
were applied to in-plane walls. The testing setup and loading assembly Quasi-static cyclic test was carried out on each specimen to evaluate
for all the specimens is shown in Figs. 6 and 7. Fig. 8 depicts hydraulic effects of target variables on response parameters. Quasi-static cyclic
pump system and data acquisition system used during testing of spe- testing was selected herein due to its advantages over other type of
cimens. dynamic testing procedures. These include the ease of applying a higher
magnitude force to the test specimens and relatively inexpensive
H.U. Sajid et al.

0.5 (a)
0.4
0.3

Top Drift (%)

0.2
0.1 Displacement
0
-0.1 Cycles
-0.2
-0.3
-0.4

0.5 (b)
0.4
Hydraulic pump 0.3
system 0.2

Top Drift (%)

Data acquisition
system 0.1
Displacement
0
Cycles
-0.1
-0.2
Fig. 8. Data acquisition system and hydraulic pump system. -0.3
-0.4
-0.5
equipment [24]. Furthermore, the progression of damage mechanism
and crack patterns can be closely captured in such type of test. Owing to Fig. 10. Displacement time history: (a) S1 and SF1 walls, (b) S2 and SF2 walls.
these advantages, researchers have frequently adopted quasi-static
cyclic testing for investigating seismic behavior of unreinforced brick
masonry [5,6,16,17,20,22,23,34,35,41,46]. Typical testing setup

Fig. 9. Typical test setup of specimen SF1 and SF2.

H.U. Sajid et al.

shown in Fig. 9. Both force and displacement control cycles were ap- observed in these specimens and they continued to increase in size at
plied with stepwise linearly increasing pattern to simulate lateral forces higher drift ratios. Other cracks were minor in nature.
on the specimen. Each force or displacement cycle was repeated three In case of high vertical stress specimens (S2 and SF2), cracks in-
times. Force and displacement as well as cracks initiation and propa- itiated at vertical face of lintel beam and propagated diagonally in
gation was monitored during and after completion of the specified cy- bottom direction at lower drift ratios and in top direction during higher
cles. Typical displacement histories imposed on specimens are shown in drift ratios. Cracks also initiated in pier location of these walls during
Fig. 10. Displacement loading cycles were continued until either 20 initial drift ratios. Diagonal cracks passed both through joints and
percent drop occurred in the lateral strength capacity of specimen or masonry units during high displacement cycles. Petry and Beyer [41]
severe physical damage occurred in the specimen. The rates of dis- reported similar steepness of diagonal cracks and passing of diagonal
placement varied from an average of 50 seconds per cycle for low cracks through masonry units under high vertical stresses in un-
displacement cycles to an average of 150 seconds per cycle during high reinforced masonry. Furthermore, breakage of masonry units was ob-
displacement served at higher drift ratios. Due to high vertical stress in these speci-
mens, shear failure dominated and hence more abrupt failure was
3. Test results and discussions observed which is consistent with results reported by Petry and Beyer
[42].
In this section, results of experiments carried out on each test spe- Flanged specimens (SF1 and SF2) sustained relatively less damage,
cimen are presented. Data obtained from various gauges is processed to as evident from Figs. 14 and 15. Flanged portion of SF1 sustained more
determine target parameters and behavior under lateral loading. damaged as compared to flanged portions of SF2 wall. At 0.20% drift
Damage mechanism and failure modes, force-displacement behavior of ratio, diagonal crack was observed in upper part of flanged portion of
wall and pier, and spandrel behavior for each specimen are discussed. SF1 wall. At higher drift ratios, more diagonal cracks as well as stepped
Lateral strength, stiffness, energy dissipation, ductility and damping are vertical crack continuing throughout height of flanged portion was
calculated for each specimen. Furthermore, effects of vertical stresses observed in SF1 wall. Flanged portion of SF2 wall remained undamaged
and flanges on the above parameters of interest are also discussed. till 0.40% drift ratio. Diagonal cracks and stepped vertical cracks, both
of relatively lesser extent was compared to SF1 wall, were observed at a
drift ratio of 0.46%. Spandrel portions of all wall specimens remain
3.1. Damage mechanism
undamaged throughout the test.
Figs. 11 and 12 depict the final crack patterns of each test specimen
while Figs. 13–15 represent propagations of cracks at different limit 3.2. Force deformation behavior
states (IO, LS and CP) in the walls as per definition of ASCE 41-06 [40]
performance levels. It is worth mentioning that three cycles of each Force-deformation behavior of each test specimen, as shown in
drift ratio were applied, therefore all crack patterns shown in Figs. 11 Figs. 16 and 17, is discussed below:
and 12 are corresponding to third cycle. All walls, except S1, exhibited
similar elastic behavior up to drift of 0.045% and a load of about 16 kN. 3.2.1. Wall S1
In S1 wall, a minor bed joint crack was observed at a drift of 0.035%. Tight loops were observed till drift ratio of 0.08% which indicates
First cracks appeared in specimens at the corners of the window low energy dissipation. Beyond 0.08% drift ratio, loops widened at high
opening in the walls. In case of low vertical stress specimens (S1 and rate indicating high energy dissipation and damage in specimen.
SF1), cracks initiated at window opening corners propagated diagonally Stiffness degradation started at 0.04% drift ratio. Stiffness degradation
towards respective top corners of the specimens. Furthermore, stepped occurred more rapidly in positive loading as compared to negative
diagonal cracks passing through mortar joints were observed at higher loading due to more damages in negative loading cycles. At higher drift
drift ratios. No masonry units breakage was observed except at corners. ratios, rocking behavior dominated the specimen response as evident
The initial diagonal cracks from window opening corner to the re- from end loops and damage pattern. Low vertical stresses favored
spective bottom and top corners of the wall were the major cracks flexural response which is consistent with the results reported in Petry

(a) (b)
Fig. 11. Crack patterns: (a) S1 wall, (b) S2 wall.