Engineering Structures 177 (2018) 103–116

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

In-plane and out-of plane response of currently constructed masonry infills T

⁎
Vasiliki Palieraki , Christos Zeris, Elizabeth Vintzileou, Chrissy-Elpida Adami
Laboratory of Reinforced Concrete, Faculty of Civil Engineering, National Technical University of Athens, Greece

A R T I C LE I N FO A B S T R A C T

Keywords: In most of the Reinforced Concrete buildings in Greece, as well as in other earthquake prone countries, the
Currently constructed masonry infill current infill construction, for the exterior walls of buildings, consists in a cavity masonry wall, made of two thin
In-plane cyclic tests walls. The two walls are not transversely connected. The seismic vulnerability of those enclosures (to in-plane
Out-of-plane repeated tests and out-of-plane actions) is high, as many seismic events have shown.
Innovative infill systems
In the last decades, emphasis was given to the study of Innovative Infill Systems with improved seismic
behaviour. The in-plane and out-of-plane behaviour of the (vulnerable) currently constructed masonry infills has
not been systematically studied, experimentally and analytically. Within the present work, two full scale RC
infilled frames were tested. One was subjected to in-plane cyclic displacements; the second specimen was
subjected to repeated out-of-plane displacements, until severely damaged, and subsequently subjected to cyclic
in-plane loading. Hysteresis loops for the entire loading history, the observed damage at several drift values and
the overall behaviour of the infill are presented and discussed upon. The obtained results are compared to the
results recorded during testing of innovative infill systems. It is shown that the performance of the currently
constructed infill system is inferior in terms of both load and deformation capacity.

1. Introduction In Greece, as well as in other earthquake prone countries (e.g.

Portugal, Italy, Turkey), the following, very vulnerable, construction
Enclosures and partition walls in RC structures, currently made of type was adopted for enclosures in the ‘70s, and is still used for the
clay bricks in most European countries, are traditionally considered as construction of infill walls: cavity brick masonry walls are constructed.
non-structural elements and, thus, they are not taken into account in The typical thickness of each leaf (made, typically, of horizontally
the seismic design of buildings. perforated clay bricks) is close to 100 mm. The space between the two
Nevertheless, when structures are subjected to earthquakes, infill (unconnected) leaves is used to accommodate insulation or sliding
walls contribute to their overall seismic response. This contribution has doors and windows. In an effort to improve the behaviour of those
been proven by numerous seismic events, as well as by experimental vulnerable infills, a typical solution adopted in Greece and widely ap-
work and numerical calculations. This effect depends on several para- plied up to now is the construction of RC tie-beams at mid-height of
meters, such as the distribution of the infills (in-plane and in-height), perimeter infill walls. Although the seismic behaviour of those infills
the relative frame-infill stiffness, the infill-RC column interaction, etc. was repeatedly proven to be poor (Fig. 1), they are still in use. On the
On the other hand, a very important issue of both public safety and other hand (see Section 2), even though the behaviour of Infilled RC
economy is related to the limitation of damages in infill walls: failure of frames was experimentally investigated in numerous studies, the
infill walls may cause injuries or even casualties, whereas extensive available research data on the seismic behaviour of the Current Infill
damages of infills (caused by earthquake not necessarily as strong as the System (CIS) is rather scarce.
design one) may be significant from the economic point of view (repair This paper presents the results of two full-scale tests on RC frames
or reconstruction of infills, repair of damages to facilities, plasters, filled following the currently applied infill system (CIS). This research
painting, etc) [1]. This fact is recognized by current Codes for Earth- was motivated by the EU funded project INSYSME (www.insysme.eu,
quake Resistant Design either explicitly or implicitly [2]. Indeed, in EC8 [3]). In the framework of that project, two innovative infill walls sys-
[2], qualitative guidance is included for the protection of infills against tems, for use in new construction, were developed at the Laboratory of
premature cracking and failure. Reinforced Concrete, NTUA [4,5]. In order to document the

⁎
Corresponding author.
E-mail addresses: [email protected] (V. Palieraki), [email protected] (C. Zeris), [email protected] (E. Vintzileou),
[email protected] (C.-E. Adami).

https://doi.org/10.1016/j.engstruct.2018.09.047
Received 29 December 2017; Received in revised form 2 August 2018; Accepted 17 September 2018
Available online 27 September 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 1. (a) In-plane and (b) out-of-plane failure of infill walls during earthquakes.

preponderance of the developed infill systems, their comparison with in Portugal, presents similarities with the one in use in Greece. How-
the behaviour of the currently applied system was investigated through ever, the two leaves of the cavity wall are of unequal thickness, whereas
the experimental campaign presented herein. the infills are not provided with RC tie-beams. Nonetheless, the results
of this recent work are valuable as far as the in-plane behaviour of the
system is concerned. The experimental work presented herein aims,
2. Literature survey therefore, to contribute to the study of the enclosure system still in use
in several earthquake prone countries, taking into account the specific
The international literature is rich in results of tests of small or large features of the system in Greece.
scale RC or steel infilled frames. Tests simulate various types of infills,
constructed with a vast variety of masonry units, either in material
(clay bricks, concrete blocks, etc.) or in geometry and pattern of per- 3. Experimental programme
forations. Furthermore, various ratios of frame to infill stiffness were
examined, along with various Codes applied for the design of the frame It is reminded that the current infill construction for perimeter walls
(sub-standard or conform to current Codes), in order to investigate the (CIS) consists in a cavity masonry wall. The exterior leaves (typically,
effect of infills to the surrounding frame elements. approximately 90 mm thick), transversely unconnected, leave a space
It is to be noted that in a number of experimental campaigns, in- between them, where the insulating material is accommodated, along
filled frames were subjected to monotonically increasing load or dis- with sliding doors and windows. No special connectors or other devices
placements up to failure (see i.a., [6–8], or-more recently-[9–11]). Data are provided to connect the enclosures to the surrounding RC elements.
on the behaviour of infilled frames subjected to cyclic in-plane loads or In Greece, with the purpose of improving the behaviour of this type of
displacements are available from tests performed during the last dec- enclosures, a RC tie-beam is typically constructed at mid-height of
ades. To this purpose, scaled specimens were tested (e.g. [12–16] -scale perimeter infill walls. The tie-beams are not fixed to the RC columns.
1:2, [17–21] -scale 1:3, [22]-scale 1:5, [18]-scale 1:9). In the recent Furthermore, RC tie-beams are constructed independently to each ex-
years, emphasis was given to the development and the evaluation of terior leaf and, hence, they do not provide any transverse connection to
innovative infill systems, able to sustain low damage [3,23–27], on them. During seismic, in-plane, loading the diagonal cracks are ex-
masonry made of large width clay bricks [4,5,28,29], or on repair and pected to occur in the intersection of the two diagonals of the infill wall.
strengthening of existing infill walls, using Cement Based Composites The expected beneficial effect of the RC tie-beam is to provide extra
[22] or FRPs [19–21,30,31]. Various loading histories were applied: in resistance to the diagonal cracking of the infill wall (thanks to the ex-
a limited number of tests, load was applied until the maximum load pectedly higher tensile strength of the concrete), as well as to reduce
capacity was reached; afterwards, displacement controlled cycles were the opening of the shear cracks (thanks to the longitudinal reinforce-
applied [32]. Most of the tests were displacement controlled, with the ment of the tie-beam).
number of cycles applied per drift value varying from one This system was reproduced in the full-scale specimens tested
[15,16,20,21,30,33], to two [22], or three cycles [14,25,26,34]. The within the present work (Fig. 2). More specifically, a reinforced con-
available literature related to the out-of-plane behaviour of masonry crete frame was designed according to EC8-Part1 [2] (Fig. 3). It should
infills is not as rich as for their in-plane behaviour. In a number of be noted that, although the CIS is relevant for a significant portion of
investigations the out-of-plane behaviour of bare masonry walls is in- the existing building stock, combined with substandard RC frames, the
vestigated [30,34–38], or thick masonry infilled frames are tested [39]. decision was made to investigate the behaviour of CIS within frames
Last but not least, several full-scale tests (either cyclic or repeated) constructed according to EC8-Part1 [2]. The main reason is that the CIS
were performed on infilled frames in- or out-of-their-plane, within the is still in use in new constructions and the second reason is that its
Project INSYSME [3]. In the framework of this project, aiming at the comparison with newly developed systems for enclosures was sought. A
development of innovative infill systems, NTUA has developed two cavity infill wall was constructed within the frame (Fig. 2a). Each leaf
infill systems, in cooperation with the Greek brick manufacturing in- (90 mm thick) was made of horizontally perforated clay bricks
dustry XALKIS S.A. Data on the performance of those two systems (both (Fig. 2b). The bricks are 210 mm long, 90 mm wide and 120 mm high,
granted with a patent by the Greek Patent Office) can be found else- the average weight per unit is equal to 1.80 kg, and the voids ratio is
where [4,5,40]. equal to 35%. The strength parallel to the holes is equal to 10.00 MPa,
Although thin masonry walls (simulating masonry infills) were while the strength perpendicular to the holes is equal to 3.50 MPa. The
tested by several investigators [10,23,26,34,37,41–43], the currently diagonal compressive strength of walls constructed with the same
constructed infills are not simulated and experimentally investigated. bricks is equal to 0.35 MPa for plain walls, while it is equal to 0.47 MPa
Recently, in [41] the infill walls used in Portugal have been experi- for walls reinforced with a tie-beam at their mid-height. A general
mentally and analytically studied. The specimen tested in [41], typical purpose cement-lime mortar, classified as M1-M2 [44] was used

104
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 2. (a) Construction of the Current Infill System; preparation for the construction of the RC tie-beam, (b) the horizontally perforated clay bricks used for the
construction of the Current Infill System, (c) detail of the attachment of the infill to the RC beam, (d) the infilled frame in position for in-plane testing.

(fmc = 0.85–1.50 MPa, fm,fl = 0.40–0.80 MPa) for the construction of The mean compressive strength of concrete was equal to 15 MPa. Each
the infill. Perpendicular mortar joints were filled. The upper row of tie-beam was reinforced with one longitudinal bar (8 mm diameter,
bricks is constructed using inclined bricks. This is a common con- characteristic yield strength = 500 MPa), located at mid-width of each
struction practice, aiming to wedge the infill wall to the RC beam leaf. After completion of the in-plane test, the failed infill was removed
(Fig. 2c). Each leaf (3.0 m long and 2.3 m high) was provided with a RC and the frame was filled again with a wall that was tested out-of-its
tie-beam (90 mm thick, 120 mm high), approximately at its mid-height. plane. In this case, only one leaf (and not the entire cavity wall) was

Fig. 3. (a) The reinforcement of the RC frame, designed according to EC8 [2], (b) section of the beam and the column of the RC frame.

105
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 4. Reaction frame and test setup used for (a) in-plane and (b) out-of-plane tests.

constructed. Actually, as no connection is provided to the two leaves, (D6 and D7, Fig. 5a) were mounted at the two ends of the centerline of
each of them behaves independently from the other. The single leaf wall the top beam to measure the in-plane lateral displacements. The col-
is situated eccentrically within the frame, i.e. in its position as part of umns of the frame are not axially loaded.
the cavity wall. After the out-of-plane test was completed, the single
leaf wall has been tested in-plane, with the purpose of estimating the 3.1.2. Out-of-plane tests
effect of the out-of-plane loading on the in-plane behaviour of the infill The test set-up consists in a steel frame, designed to comply with the
wall. requirement of significant stiffness and, hence, of minimal deformations
A clarification needs to be added here related to the decision made during testing. The steel frame (Fig. 4b) was serving as a reaction
to construct the second infill within the already tested RC frame. During system to the hydraulic jack (capacity ± 500 kN) which was used for
the in-plane test of the CIS, hairline flexural cracks (hardly visible the application of deformations to the infill. A steel system with mul-
during loading and invisible after the completion of the test) occurred tiple hinges (Fig. 4b) was connected to the hydraulic jack, in order to
in the two ends of the columns of the frame. Therefore, no unfavourable reach a, as uniform as possible distribution of the applied deformations.
for the RC frame interaction with the infill was observed, whereas the During testing, the RC frame was transversely supported and repeated
damage of the RC frame was very light. Thus, the decision was made to loading was applied to the infill alone. The test was displacement-
re-use the RC frame for the second test. controlled.

3.1. Test setup 3.2. Instrumentation and test protocol

3.1.1. In-plane tests Fig. 5a shows the instrumentation of the specimen tested in its
Fig. 4a shows the test setup used for the in-plane tests: a servo- plane. The resistance of the specimen to the imposed displacements was
controlled hydraulic actuator (maximum capacity = ± 1000 kN) was measured by a load cell incorporated in the servomechanism. Twenty
used to apply the lateral displacements. The actuator was fastened to six (26) LVDTs were installed to record absolute and relative dis-
the beam of the stiff steel reaction frame. The RC frame was provided placements (e.g. the lengthening/shortening of the diagonals of the
with a strong footing, anchored to the strong floor of the laboratory, by infill, the relative displacement at infill-frame interface, the relative
means of 16 steel dowels 30 mm in diameter, with the purpose to avoid displacement between the RC tie-beam and the infill, etc.). Six (6) strain
parasitic movements of the entire specimen during testing. The actuator gauges were positioned on the reinforcing bars of the RC ties, to mea-
applies horizontal in-plane cyclic displacements at mid-height of the sure their deformations. The strain gauges were installed in the two
beam. To allow for cyclic displacements to be imposed, the actuator ends, as well as in the middle of the reinforcing bars.
was attached to a system of two steel plates (one per end-face of the RC Fig. 5b shows the instrumentation of the specimen tested in its
beam) and eight steel rods tightly connecting the two steel plates plane, after having been tested out-of-its plane. In total, nineteen (19)
(Figs. 2d, 4a). Two Linear Variable Differential Transformers (LVDTs) displacement transducers were installed to record absolute and relative

106
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 5. Instrumentation: (a) in-plane test. The LVDTs D1-D2, D9-D17 and D25-D26, are positioned on the same side of the infill wall, the LVDTs D3-D4, D18-D24 are
positioned on the opposite side, (b) in-plane test of single leaf (after its loading out-of-plane). The LVDTs D1-D2, D12-D13 and D16, are positioned on the same side of
the infill wall, the LVDTs D3-D4, D14-D15 and D17-D19 are positioned on the opposite side, (c) out-of-plane test. The LVDTs D1-D17, are positioned on the same side
of the infill wall, the LVDTs D18-D24 are positioned on the opposite side.

107
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

displacements (e.g. the lengthening/shortening of the diagonals of the and failure of bricks.
infill, the relative displacement at infill-frame interface, the relative The significant contribution of sliding between masonry wall and
displacement between the RC tie-beam and the infill, etc.). RC tie-beam, an important characteristic of the CIS, is depicted in
The CIS specimen is subject to stepwise increasing imposed hor- Fig. 8a, where the deformations measured by the LVDTs D25, D26, D11
izontal displacements (measured at the centre of the RC frame beam, and D17 (Fig. 5a), are shown. D25 and D26 are measuring the relative
D6 and D7 in Fig. 5a,b). Three full displacement cycles are imposed for displacement between the RC tie-beam and the infill, in the upper and
each selected value of maximum displacement. the lower part of the RC tie-beam respectively. D11 is measuring the
In case of out-of-plane test, twenty seven (27) measuring devices relative displacement between the RC beam and the infill wall, while
were installed at various locations (Fig. 5c) to record the deformations D17 is measuring the relative displacement between the RC footing and
of the infill, the differential out-of-plane deformation at infill/frame the infill wall. All the abovementioned LVDTs are positioned on the
interface and at infill/RC tie-beam interface, etc. In this case, repeated same leaf of the infill wall. The data plotted in Fig. 8a show that hor-
out-of-plane deformations were imposed to the infill. The mobilized izontal sliding takes place between the RC tie-beam and the infill (D25
resistance was measured by a dynamometer incorporated to the hy- and D26), as well as between the infill and the RC beam (D11) and the
draulic jack. The imposed displacement (measured at the centre of the infill at its base (D17). It is interesting to observe that the portion of the
infill, D3 in Fig. 5c) was increased stepwise. Three cycles were per- infill between the RC tie-beam and the beam of the frame seems to
formed at each displacement level. move as a solid body (compare the measurements of measuring devices
It is noted that the measuring devices shown in Fig. 5, are positioned D25 and D11). The same holds true for the lower portion of the infill,
on both sides of the infill walls. between the RC tie-beam and the base, as the comparable values of
horizontal sliding measured by D26 and D17 show. Obviously, the
4. Test results upper portion of the infill moves more than the lower one. It seems that
the sliding between the tie-beam and the infill is initiated at low drift
4.1. In-plane cyclic response, cavity wall values. This observation is in accordance with the measurements pro-
vided by the strain gauges installed on the tie-beam reinforcement
The main results of the in-plane test are shown in Figs. 6–9. The test (Fig. 8b): the strain at the reinforcement did not exceed 0.2‰, showing
was concluded after the occurrence of non-repairable damages to the the rather low degree of mobilization of the tie-beam (due to its early
infill associated with a degradation of the resistance of the infilled separation from the adjacent masonry) and, hence, its limited con-
frame larger than 20%. tribution to the lateral resistance of the infilled frame.
The hysteresis loops for the infilled frame are shown in Fig. 6a. A On the contrary, as shown in Fig. 9, the presence of the tie-beam
maximum lateral resistance equal to 500.0 kN was recorded at a drift drastically modifies the failure mode of the infill wall. Actually, hor-
value of approximately 1.10%. The test was concluded soon after the izontal cracks occur along concrete to masonry interfaces, for small
attainment of the maximum resistance, at a drift value approximately values of the imposed drift. This failure mode is typical for infill walls
equal to 1.60%. Actually, a steep falling branch (typical for brittle be- with RC tie-beam [36]. From then on, significant sliding takes place
haviour) was recorded. The force-response degradation due to cycling along those horizontal interfaces (see Fig. 8a), although this sliding
was approximately equal to 20% between the first and the second cycle does not prevent the opening of bi-diagonal cracks at larger drift values.
(Fig. 6b). Practically no further degradation occurred during the third The significant contribution of the shear-sliding mode to the behaviour
loading cycle. of the infill (due to the RC tie-beam) may adversely affect the behaviour
The deformation along the diagonals of the infill (Fig. 7) reached of the adjacent RC columns. Actually, it may impose an additional shear
values as large as 25–30 mm. While the elongation of a diagonal is due force to the columns close to their mid-height, a location where the
exclusively to the opening of cracks (as the tensile deformation of shear reinforcement may not be as closely spaced as within the critical
masonry is minimal), the shortening of the diagonals includes the regions (close to the ends) of the columns. Thanks to the design of the
closing of the cracks, the compressive deformations of masonry and, at tested frame, in which columns were provided with closely spaced
more advanced stages of loading, the deformations due to the cracking hoops over their entire height, such a negative effect was not observed.

(a) (b)
Fig. 6. In-plane test, cavity wall: (a) hysteresis loops, (b) normalized resistance vs. drift envelopes.

108
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

(a) (b)
Fig. 7. In-plane test, cavity wall: deformations along the diagonals, plotted against the imposed horizontal displacement.

Cracking and spalling of bricks (Figs. 9 and 10a) led to the completion Fig. 13c. However, early separation between the infill and the RC tie-
of the test at a rather small value of drift. beam was also recorded (see Fig. 13a: a crack along the interface be-
tween the infill and the upper face of the RC tie-beam occurred for a
value of the deflection equal to 5 mm). Those observations are con-
4.2. Out-of-plane response firmed by the profile (both horizontal and vertical) of the deflections
(Fig. 12a and b). It should be pointed out though that, as the RC frame
This section summarizes the results of the out-of-plane test of the was transversely fixed and, hence, not moving out of its plane, the
infill. It is reminded here that a single leaf infill (90 mm thick) was displacements measured close to the interfaces between the infill and
tested. Its maximum resistance was equal to 75.6 kN; it was mobilized the RC elements show that the infill was slipping out of the frame, along
for a deflection value (measured at the centre of the wall, LVDT D3, the separating cracks formed for rather small values of the imposed
Fig. 5c) equal to 30 mm. The maximum deflection imposed to the infill deflection. The deflections along the perimeter of the infill (Fig. 12)
was equal to 60 mm. The force-response degradation between the first represent the slip of the wall out of the RC frame.
and third cycle was approximately equal to 15%. At the maximum In Fig. 14, the relative displacements between the infill and the RC
imposed deflection, a force-response loss of 30% to 50% (compared to column, measured by D23a and D23k, as well as the relative dis-
the maximum resistance) was recorded (Fig. 11). placement between the RC tie-beam and the RC column are shown.
Interesting features of the out-of-plane behaviour of the infill are Those relative displacements (deflections) were measured on the two
illustrated by the distribution of deflections along the length and the ends of the tie-beam. The data plotted in Fig. 14 show that the tie-beam
height of the infill (Fig. 12), as well as by the crack pattern (Fig. 13). changes the distribution of the deflections along the height of masonry.
Actually, it seems that the infill panel behaves like a slab loaded by a The deflections measured by D23a, above the tie-beam, are larger than
uniformly distributed load, as demonstrated by the “yield lines” of

(a) (b)
Fig. 8. In-plane test, cavity wall: (a) sliding along interfaces between the infill and the frame or between the infill and the RC tie-beam, (b) horizontal force-response
vs. strain of the reinforcing bars curves for the tie-beam of the CIS wall.

109
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 9. In-plane tests. Crack pattern at various drift values: (a) d = 0.2%, (b) d = 1.10% (at maximum lateral resistance), (c) d = 1.57% (at maximum imposed drift).

Fig. 10. Cracking of the CIS specimen at the end of testing (a) in-plane, cavity wall, (b) out-of-plane, single leaf, (c) in-plane, single leaf (after out-of-plane test was
completed).

the ones measured by D23k, under the tie-beam. This is in accordance degradation larger than 20% was recorded during the third cycle
with the crack pattern of the infill (Fig. 13) showing that a horizontal (Fig. 15b), while very pronounced cracking and crushing of brick units
crack occurred along the RC tie-beam (separating the upper part of the was recorded (Fig. 10c). As shown in Fig. 15b, the force-response de-
infill from the tie-beam). Thus, the out-of-plane deflections of the two gradation recorded during the third loading cycle, was comparable with
halves of the infill are not equal (Fig. 14). the degradation recorded during the second cycle.
The deformation along the diagonals of the infill (Fig. 16) reached
values as large as 30 mm. It is to be noted that the deformation values of
4.3. In-plane cyclic response, single leaf previously tested out-of-plane the diagonals plotted in Fig. 16 represent the width of the cracks that
occurred during the in-plane loading alone. The width of the cracks that
After the completion of the out-of-plane test (see Section 4.2), the opened during the out-of-plane loading (see e.g. Figs. 10b and 13) is not
infilled frame was subject to in-plane cyclic displacements, following included.
the same loading history as for the cavity wall (Section 4.1). It should The significant contribution of sliding between masonry wall and
be noted that, after out-of-plane loading, there was residual out-of- RC tie-beam, is observed in this case too. Actually, in Fig. 17a, the re-
plane deflection of the infill. Its maximum value (recorded at the centre lative displacements between the infill and (a) the RC beam, (b) the RC
of the infill (see Figs. 11a and 12a) was equal to 30 mm. As shown in tie-beam and (c) the RC footing are shown. Those relative displace-
Fig. 15a, where the hysteresis loops are presented, the maximum lateral ments (sliding) were measured on both sides of the infill (LVDTs D12,
resistance of the infilled frame was equal to 371.2 kN. It was recorded D13, D16 and D14, D15, D19 respectively). The data plotted in Fig. 17a
at a drift value of approximately 1.80%, coinciding with the maximum show that horizontal sliding takes place between the upper part of the
value of the applied drift. Actually, the test was concluded soon the RC tie-beam and the infill (D12 and D14), while sliding between the
attainment of the maximum resistance, given that a force-response

110
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

(a) (b)
Fig. 11. Out-of-plane tests: (a) hysteresis loops, (b) normalized resistance vs. displacement curves.

lower part of the RC tie-beam and the infill (D13 and D15) is much enclosure walls to the global force displacement response of the spe-
smaller. Larger values of the horizontal sliding are recorded between cimens, has been evaluated as the difference between the envelope
the infill and the RC beam (D16 and D19). It is interesting to observe curves of the infilled frame and the envelope curves of the bare RC
that the portion of the infill between the RC tie-beam and the RC frame frame at corresponding drift levels (Fig. 18). The curve corresponding
beam does not move as a solid body (compare D12 and D16, D14 and to the behaviour of the RC frame was obtained experimentally, after the
D19), due to the fact that the in-plane loaded infill is already cracked completion of in- or out-of-plane tests on infills. Actually, at the end of
due to the out-of-plane loading that preceded. It is also noted that the the campaign, the infill was demolished and the bare frame was sub-
measurements on the two faces of the wall are almost identical. jected to in-plane displacements. The envelopes in Fig. 18 show that the
in-plane response of the infill tested in-plane after having been sub-
5. Discussion on experimental results jected to out-of-plane actions is by 20% lower than that of the infill
subject exclusively to in-plane actions. For the sake of comparison, the
5.1. The effect of out-of-plane loading to in-plane behaviour ordinates of the CIS-single wall envelope were doubled (assuming that
the in-plane resistance of two infill leaves is double the resistance of a
In order to evaluate the effect of previous out-of-plane loading on single leaf). The effect of prior out-of-plane loading is illustrated in the
the in-plane behaviour of the tested infill, the contribution of the lower in-plane stiffness of the infill, as compared to that of the infill

(a) (b)
Fig. 12. Out-of-plane tests: (a) deflections along the height, measured by D1, D2, D3, D4 and D5 (Fig. 5c), at mid-length of the infill, (b) deflections along the length,
measured by D13, D20, D3, D21 and D8(Fig. 5c), at mid-height of the infill.

111
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 13. Out-of-plane tests. Crack pattern of the CIS specimen at various deflection values: (a) defl. = 5 mm, (b) defl. = 30 mm, at maximum resistance, (c)
defl. = 60 mm, maximum imposed deflection.

exclusively loaded in its plane (Fig. 18).

It is noted that in the published test results on the interaction of in-
plane and out-of-plane loading, in most cases, in-plane loading precedes
out-of-plane tests [3,10,28,43,45]. In cases where the out-of-plane
loading is applied before the in-plane loading [45,46], the out-of-plane
loading was of rather small amplitude, and it did not appear to influ-
ence the in-plane behaviour (characterized by in-plane corner
crushing). In the current study, large out-of-plane deformations were
imposed. Furthermore, the in-plane failure was governed by diagonal
cracking of masonry, while crushing of the bricks in the corners was
limited (Fig. 10c).

5.2. Current vs. Innovative infill systems

Although a detailed comparison between the current and the in-

novative infill systems developed at NTUA is beyond the scope of this
paper, selected data are provided in Fig. 19, allowing for some major
differences in the behaviour to be detected. Thus, Fig. 19a presents the
hysteresis loops envelope obtained from in-plane testing of an RC frame
filled with a cavity infill wall (CIS), an infill constructed following
INSYSTEM1 or an infill following INSYSTEM2. In INSYSTEM1, the infill
wall was separated into three wallettes, by means of vertical joints,
Fig. 14. Out-of-plane tests: deflections measured by D23a, D23k and D24 constructed using a mortar of reduced modulus of elasticity, to allow for
(Fig. 5c). increased deformations along the predetermined planes between con-
secutive wallettes [4]. In INSYSTEM2, the infill is horizontally and
vertically reinforced. For the construction of INSYSTEM2 infills, the

(a) (b)
Fig. 15. In-plane test, single wall: (a) hysteresis loops, (b) normalized resistance vs. displacement envelopes.

112
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

(a) (b)
Fig. 16. In-plane test, single wall: deformations along the diagonals, plotted against the imposed horizontal displacement.

innovative brick units designed and produced on purpose were used As long as it regards the out-of-plane behaviour, the preponderance
[5]. It should be noted that in INSYSTEMS1&2, infills consist of single of the reinforced infill is obvious in terms of combined bearing capacity
leaf 250 mm and 280 mm thick respectively, made of vertically perfo- and deformability (Fig. 19b). Both innovative systems exhibited sig-
rated clay bricks. To compare the behaviour of the three systems, nificant out-of-plane resistance, in contrast to the CIS. The latter was,
namely CIS, INSYSTEM1 and INSYSTEM2, the contribution of the infills on the other hand, able to sustain large deformations. It should be noted
to the in-plane behaviour of the infilled frame is presented in Fig. 19a. though that, in case of CIS, significant out-of-plane movement of the
To produce the envelopes shown in that Figure, the shear force vs. infill with respect to the frame was recorded (Fig. 12). Taking into
horizontal displacement/drift curve of the respective bare frames is account that the test was quasi-static and, hence, dynamic phenomena
subtracted from the global envelopes. occurring during a seismic event are not reproduced, the quite probable
According to the data of Fig. 19a, the value of drift for which the out-of-plane collapse of the CIS enclosure (observed frequently in real
maximum lateral resistance of the infill is recorded is approximately conditions, see Fig. 1b) is not captured. This phenomenon is expected to
equal to 0.32% for the CIS, 2.00% for INSYSTEM1 and 1.10% for be aggravated by the simultaneous in- and out-of-plane actions on the
INSYSTEM2. Defining the point on the falling branch, at which a force- CIS infill.
response reduction of 20% is recorded, the respective drift values are The evaluation of the experimental results, obtained from in-plane
equal to 0.67%, 2.35% and 1.60% for CIS, INSYSTEM1 and INSYSTEM2 and out-of-plane testing RC frames infilled either with cavity masonry
respectively. walls or following the innovative systems developed at NTUA allowed

(a) (b)
Fig. 17. (a) Sliding along interfaces between the infill and the RC tie-beam, (b) sliding along interfaces between the infill and the RC beam and between the infill and
the upper part of the RC tie-beam.

113
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Fig. 18. In-plane behaviour of infill loaded in its plane (CIS cavity wall) and infill loaded in its plane after having been subjected to out-of-plane loading (CIS single
wall and CIS single wall x 2). Hysteresis loops envelopes (after subtraction of the contribution of the bare frame).

for the definition of performance levels of infills, relating the severity of constitute a damage state for the infill itself.
damage to infills with the drift imposed to the infilled frame or with the DLS is related to limited damage of the enclosures (repairable and
maximum deflection of an infill subjected to out-of-plane actions. Thus, economically sustainable). An unequivocal and objective definition of
the overall behaviour of the infill is assessed and the adequacy of each DLS being not feasible, this limit state is related for in-plane actions to
infill system for a required performance of the entire structure is the lesser of: (a) the drift level at which noticeable diagonal cracking
evaluated. Three performance levels are defined [2,29,47,48], namely, (involving masonry units as well) is initiated and; (b) the drift level at
the Operational Limit State (OLS), the Damage Limit State (DLS) and which the enclosure reaches its maximum lateral capacity. The drift
the Ultimate Limit State (ULS). The Operational Limit State is somehow values corresponding to DLS were calculated following the second cri-
arbitrarily set to a drift value of 0.20% for in-plane actions or to an out- terion, on the basis of the experimental results (see Table 1). Similar
of-plane deflection of 5.0 mm. Actually, OLS cannot be defined on the qualitative criteria are adopted for the estimation of the DLS for out-of-
basis of the available experimental data, being related mainly to slight plane actions (see Table 1). It seems that the DLS value for the CIS
damage to finishings. The early separation between the infill and the coincides with the respective values in the literature [10].
frame is not taken into account in the evaluation, as it does not The ULS (or Life Safety Limit State) is related to severe damage of

(a) (b)
Fig. 19. Comparison of the behaviour of the CIS infill, with two innovative infill systems: (a) in-plane behaviour, contribution of the enclosure walls, (b) out-of-plane
behaviour.

114
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

Table 1 loading.
Drift and deflection values corresponding to OLS, DLS, ULS calculated on the (4) The qualitative definition of performance levels of the examined
basis of the experimental in-plane and out-of-plane results, correspondingly, for infill systems prove the pronounced vulnerability of the CIS both for
the three tested infill systems. in-plane and out-of-plane actions.
Infill type OLS DLS ULS OLS DLS ULS (5) It has to be admitted that further research on the Current Infill
System is needed. Actually, as numerous earthquakes have shown,
In- plane loading Drift (%) Out-of-plane loading Deflection the main disadvantage of this system is its poor behaviour under
(mm)
combined in- and out-of-plane actions. On the contrary, the beha-
CIS 0.2% 0.32% 0.67% 5.00 25.00 55.00 viour of RC frames infilled according to INSYSTEM1 and INSYS-
INSYSTEM 1 0.2% 2.00% 2.35% 5.00 15.00 20.00 TEM2 when subjected to biaxial shaking table tests [40] was proved
INSYSTEM 2 0.2% 1.10% 1.60% 5.00 30.00 45.00 to be quite adequate even for strong input motions.

Acknowledgments
the infill. However, the infill should not be close to collapse (human
lives protection). Actually, ULS corresponds to non-repairable damage
This work was funded by the State Scholarships Foundation (IKY),
or to cases were reconstruction of the infill is preferable from the
through the program “IKY FELLOWSHIPS OF EXCELLENCE FOR
economical viewpoint. To define ULS on the basis of the experimental
POSTGRADUATE STUDIES IN GREECE- SIEMENS PROGRAM”.
results, the lesser of: (a) the drift value at which significant damage
involves masonry units, expulsion of masonry portions and/or wide-
References
spread spalling and crushing of masonry units and; (b) the drift value at
which the bearing capacity of the enclosure degrades by about 20%,
[1] Magenes G, Bracchi S, Graziotti F, Mandirola M, Manzini CF, Morandi P, et al.
compared to its maximum capacity. In this case, the second criterion Preliminary damage survey to masonry structures after the May 2012 Emilia
was adopted, as the degradation of the maximum resistance by 20% is earthquakes; 2012. v.1, < http://www.eqclearinghouse.org/2012-05-20-italy-it > .
an indicator of significant damage. Actually, as shown in Fig. 9b for CIS, [2] EN 1998-1. Eurocode 8. Design of structures for earthquake resistance – Part 1:
General rules, seismic actions and rules for buildings, Brussels; 2005.
already at the point of reaching the maximum lateral resistance of the [3] www.insysme.eu, website of the European Research Project INSYSME (INnovative
infilled frame, extensive cracking and some crushing of bricks was SYStems for earthquake resistant Masonry Enclosures in rc buidings).
observed. Simiral criteria are adopted for enclosures subjected to out- [4] Vintzileou E, Adami CE, Palieraki V. In-plane and out-of plane response of a ma-
sonry infill divided into smaller wallettes. In: Proceedings of the 16th international
of-plane actions. The respective drift and deflection values for the three brick and block masonry conference “Masonry in a World of Challenges”, 26–30
tested infill systems are included in Table 1. June, Padova, Italy; 2016.
Although this procedure is qualitative and, hence, it constitutes a [5] Vintzileou E, Palieraki V, Adami CE, Nikolopoulou V. In-plane and out-of plane
response of a reinforced masonry infill made by an innovative new brick unit. In:
rough estimation of the performance levels of enclosure systems, it al- Proceedings of the 16th Word Conference on Earthquake, 16WCEE 2017, 9-13
lows for a comparison among the tested infills, on the basis of common January, Santiago, Chile; 2017.
criteria. The data in Table 1 allow for the pronounced vulnerability of [6] Thomas FC. The strength of brickwork. Struct Eng 1953;31(2):35–46.
[7] Holmes M. Combined loading on infilled frames. Proc Inst Civil Eng 1963;25:31–8.
the CIS to be identified.
[8] Stafford-Smith. Model tests of vertical and horizontal loading of infilled frames. J
Am Concrete Inst 1968;65(8):618–24.
6. Conclusions [9] Calvi GM, Bolognini D. Seismic response of RC frames infilled with weakly re-
inforced masonry panels. J Earthquake Eng 2001;5:153–85.
[10] Vermeltfoort AT. Bed joint reinforcement and the load bearing capacity of masonry
The paper presents the results of full scale testing of the Current beams. In: Proceedings of the 16th international brick and block masonry con-
Infill System (CIS), which is used in Greece as well as in other earth- ference, “Masonry In A World Of Challenges”, 26–30 June, Padova, Italy; 2016.
quake prone countries. The current infill construction consists in a [11] Nasiri E, Liu Yi. Experimental study of the effect of interfacial gaps on the in-plane
behaviour of masonry infilled RC frames. In: Proceedings of the 16th international
cavity masonry wall, made of two thin, transversely unconnected ex- brick and block masonry conference, “Masonry In A World Of Challenges”, 26–30
terior leaves. The tests results are compared with those obtained from June, Padova, Italy; 2016.
testing innovative infill systems. [12] Zarnic R, Tomazevic M. The behaviour of masonry infilled concrete frames sub-
jected to cyclic lateral loading. In: Proceedings of the 8th World Conference on
On the basis of the results of the experimental work presented in this Earthquake Engineering, 21–28 July, San Francisco, California; 1984. p. 863–70.
paper, the following conclusions can be drawn: [13] Mehrabi AB, Shing PB. Performance of masonry-infilled R/C frames under in-plane
lateral loads. In: NCEER workshop on seismic response of masonry, San Francisco,
California; 1994. p. 45–50.
(1) The currently used system exhibits a rather poor in-plane beha- [14] Perera R. Performance evaluation of masonry-infilled RC frames under cyclic
viour. Actually, it underwent significant damages under low drift loading based on damage mechanics. Eng Struct 2005;27:1278–88.
values (1.10%) and it suffered significant force-response degrada- [15] Puglisi M, Uzcategui M, Florez-Lopez J. Modeling of masonry of infilled frames, Part
I: The plastic concentrator. Eng Struct 2008;31:113–8.
tion at a drift value of 1.60%. Due to the fact that in the RC frame
[16] Tasnimi AA, Mohebkhah A. Investigation on the behavior of brick-infilled steel
only minor cracking occurred, the limited drift values reached by frames with openings, experimental and analytical approaches. Eng Struct
the infilled frame are due to the behaviour of the infill wall. After 2011;33:968–80.
subtraction of the RC frame contribution, the lateral force vs. drift [17] Valiasis NT, Stylianidis CK. Masonry infilled R/C frames under horizontal loading:
experimental results. Eur Earthquake Eng 1989;3:10–20.
envelope representing the behaviour of the infill alone shows that [18] Manos GC, Yasin B, Valiasis T. Small scale model simulation of the cyclic behavior
significant damage of the infill is observed at a drift value of 0.32% of infill brick panels. In: Proceedings of the 6th North American Masonry
and force-response degradation equal to 20% at a drift value of Conference, 6–9 June, Philadelphia, Pennsylvania; 1993. p. 359–70.
[19] Altin S, Anil O, Kara EM, Kaya M. An experimental study on strengthening of
0.67%. masonry infilled RC frames using diagonal CFRP strips. Compos B Eng
(2) The Current Infill System (CIS) was able to sustain large out-of- 2007;39(4):680–93.
plane deflection; its resistance was, however, quite low. Moreover, [20] Yuksel E, Ozkaynak H, Buyukozturk O, Yalcin C, Dindar AA, Surmeli M, et al.
Performance of alternative CFRP retrofitting schemes used in infilled RC frames.
it presented a tendency to slide horizontally out of the RC frame. Constr Build Mater 2010;24(4):596–609.
This is in accordance with post-earthquake observations, where CIS [21] Ozkaynak H, Yuksei E, Buyukozturk O, Yalcin C, Dindar AA. Quasi-static and
infills overturn and collapse out-of-their plane. pseudo-dynamic testing of infilled RC frames retrofitted with CFRP material.
Compos B Eng 2010;42:238–63.
(3) Τhe in-plane response of the infill tested in-plane after having been [22] Kyriakides MA, Billington SL. Seismic retrofit of masonry-infilled non-ductile re-
subjected to out-of-plane actions is by 20% lower than that of the inforced concrete frames using sprayable ductile fiber-reinforced cementitious
infill subject exclusively to in-plane actions. The in-plane stiffness composites. In: Proceedings of the 14th World conference on earthquake en-
gineering, 12–17 October, Beijing, China; 2008.
of the infilled frame was also reduced due to prior out-of-plane
[23] Mohammadi M, Akrami V, Mohammadi-Ghazi R. Methods to improve infilled frame

115
V. Palieraki et al. Engineering Structures 177 (2018) 103–116

ductility. J Struct Eng 2011;137(6):646–53. or RC tie-beams. Masonry Int 2007;20(3):117–28.

[24] Tasligedik AS, Pampanin S, Palermo A. In-plane cyclic testing of non-structural [37] da Porto F, Mosele F, Modena C. Experimental testing of tall reinforced masonry
drywalls infilled within RC frames. In: Proceedings of the 15th World conference on walls under out-of-plane actions. Constr Build Mater 2010;24:2559–71.
earthquake engineering, 24–28 September, Lisbon, Portugal; 2012. [38] Papanicolaou C, Triantafillou T, Lekka M. Externally bonded grids as strengthening
[25] Preti M, Migliorati L, Giuriani E. Experimental testing of engineered masonry infill and seismic retrofitting materials of masonry panels. Constr Build Mater
walls for post-earthquake structural damage control. Bull Earthq Eng 2011;25:504–14.
2014;13(7):2029–49. [39] Hak S, Morandi P, Magenes G. Out-of-plane experimental response of strong ma-
[26] Tasligedik AS, Pampanin S. An innovative low damage seismic solution for un- sonry infills. In: Proceedings of the 2nd European conference on earthquake en-
reinforced clay brick infill walls. In: Proceedings of the 2nd European conference on gineering and seismology, 24–29 August, Istanbul, Turkey; 2014.
earthquake engineering and seismology, 25–29 August, Istanbul, Turkey; 2014. [40] Vintzileou Ε, Adami CE, Palieraki V, Mouzakis Ch, Karapitta L. Seismic behaviour of
[27] Penna A, Magenes G, Calvi GM, Costa AA. Seismic performance of Aac infill and a reinforced masonry infill made with an innovative new brick unit: results of a
bearing walls with different reinforcement solutions. In: Proceedings of the 14th shaking table test. In: Proceedings of the 6th ECCOMAS thematic conference on
international brick and block masonry conference, 13–20 February, Sydney, computational methods in structural dynamics and earthquake engineering,
Australia; 2008. COMPDYN 2017, 15–17 June, Rhodes Island, Greece; 2017.
[28] Pereira MFP, Pereira MFN, Ferreira JED, Lourenço PB. Behavior of masonry infill [41] Furtado R, Arede V. Double-leaf infill masonry walls cyclic in-plane behaviour:
panels in RC frames subjected to in plane and out of plane loads. In: Proceedings of experimental and numerical investigation. Open Construct Build Technol J
the 7th AMCM international conference, 13–15 June, Krakow, Poland; 2011. 2018;12:35–48.
[29] Morandi P, Hak S, Magenes G. Performance-based interpretation of in-plane cyclic [42] Pujol S, Fick D. The test of a full-scale three story RC structure with masonry infill
tests on RC frames with strong masonry infills. Eng Struct 2018;156(1):503–21. walls. Eng Struct 2010;32(10):3112–21.
[30] Erdem I, Akyuz U, Ersoy U, Ozcebe G. An experimental study on two different [43] da Porto F, Guidi G, Dalla Benetta M, Verlato Ν. Combined in-plane/out-of-plane
strengthening techniques for RC frames. Eng Struct 2006;28(13):1843–51. experimental behavior of reinforced and strengthened infill masonry walls. In:
[31] Derakhshan H, Lucas W, Griffith MC. Flexural behavior of FRP strengthened brick Proceedings of the 12th Canadian masonry symposium, 2–5 June, Vancouver,
cavity walls. In: Proceedings of the 16th international brick and block masonry British Columbia; 2013.
conference, “Masonry In A World Of Challenges”, 26–30 June, Padova, Italy; 2016. [44] EN 1996-1-1. Eurocode 6 -Design of masonry structures -Part 1-1: General rules for
[32] Anil O, Altin S. An experimental study on reinforced concrete partially infilled reinforced and unreinforced masonry structures, Brussels; 2005.
frames. Eng Struct 2006;29(3):449–60. [45] Flanagan R, Bennett R. Bidirectional behavior of structural clay tile infilles frames. J
[33] Pires F, Carvalho EC. The behavior of infilled reinforced concrete frames under Struct Eng 1999;125(3):236–44.
horizontal cyclic loading. In: Proceedings of the 10th World conference on earth- [46] Henderson RC, Jones WD, Burdette EG, Porter ML. The effect of prior out-of-plane
quake engineering, 19–24 July, Madrid, Spain; 1992. damage on the in-plane behavior of unreinforced masonry infilled frames. In: 4th
[34] Chiou TC, Hwang SJ. Test on cyclic behavior of reinforced concrete frames with DOE natural phenomena hazards mitigation conference, 19–22 October, Atlanta,
brick infill. Earthquake Eng Struct Dyn J Int Assoc Earthquake Eng Georgia; 1993. p. 307–16.
2015;44(12):1939–58. [47] Hak S, Morandi P, Magenes G, Sullivan T. Damage control for clay masonry infills in
[35] Griffith MC, Vaculik J. Flexural strength of unreinforced clay brick masonry walls. the design of RC frame structures. J Earthquake Eng 2012;16(1):1–35.
In: Proceedings of the 10th Canadian masonry symposium, 8–12 June, Banff, [48] Sassun K, Sullivan TJ, Morandi P, Cardone D. Characterising the in-plane seismic
Alberta; 2005. performance of infill masonry. Bull New Zeal Soc Earthquake Eng
[36] Vintzileou E, Palieraki V. Perimeter infill walls: the use of bed joint reinforcement 2016;49(1):100–17.

116