A Review of Experimental and Analytical

A review of experimental and analytical
studies on the out-of-plane behaviour of

masonry infilled frames
Filip Anić, Davorin Penava, Lars

Abrahamczyk & Vasilis Sarhosis
Bulletin of Earthquake Engineering

Official Publication of the European
Association for Earthquake Engineering
ISSN 1570-761X
Volume 18
Number 5
Bull Earthquake Eng (2020)

18:2191-2246
DOI 10.1007/s10518-019-00771-5
1 23
Your article is protected by copyright and
all rights are held exclusively by Springer
Nature B.V.. This e-offprint is for personal
use only and shall not be self-archived
in electronic repositories. If you wish to
self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.
1 23
Author's personal copy
Bulletin of Earthquake Engineering (2020) 18:2191–2246
https://doi.org/10.1007/s10518-019-00771-5
REVIEW ARTICLE
A review of experimental and analytical studies

on the out‑of‑plane behaviour of masonry infilled frames
Filip Anić1 · Davorin Penava1 · Lars Abrahamczyk2 · Vasilis Sarhosis3
Received: 6 June 2019 / Accepted: 11 December 2019 / Published online: 17 December 2019
© Springer Nature B.V. 2019
Abstract
This paper presents a literature review of research undertaken on the out-of-plane behav-
iour of masonry infilled frames. This paper also discusses the effects of bidirectional loads,
openings, slenderness, boundary conditions etc. As numerous researchers have reported,
these effects play a crucial role in achieving arching action cause, as they can bypass or
limit its effectiveness. Namely, arching action leads to additional compressive forces which
resist traversal ones. This is confirmed by inertial force methods of testing, while the same
cannot be claimed for inter-storey drift or dynamical methods. It is to be acknowledged that
most experimental tests were carried out using inertial force methods, mostly with the use
of airbags. In contrast, only a few were undertaken with dynamical methods and just two
with inter-storey drift methods. It was found that inertial force and inter-storey drift meth-
ods differ widely. In particular, inertial force methods damage the infill, leaving the frame
more or less intact. Conversely, drift heavily damages the frame, while infill only slightly.
Openings were investigated, albeit with contrasting results. Namely, in all cases, it was
found that openings do lower the deformational but not all load-bearing capacities. Fur-
thermore, analytical models have shown contrasting results between themselves and with
experimental data. Models’ stabilities were checked with single- and multi-variable para-
metric analysis from which governing factors, influences of frame and other parameters
were identified.
Keywords Out-of-plane behaviour · Experimental studies · Analytical studies · Masonry

infilled frames · Experimental methodologies
* Davorin Penava
davorin.penava@gfos.hr
1
Faculty of Civil Engineering and Architecture Osijek, Josip Juraj Strossmayer University of Osijek,
3 Vladimir Prelog Str., 31000 Osijek, Croatia
2
Bauhaus-Universität Weimar, Erdbebenzentrum, Marienstraße 13B, 99423 Weimar, Germany
3
Faculty of Engineering, School of Civil Engineering, University of Leeds, Leeds LS2 9JT, UK
13
Vol.:(0123456789)
2192 Bulletin of Earthquake Engineering (2020) 18:2191–2246
1 Introduction
Earthquakes are natural occurring phenomena which manifest themselves by shaking and
displacing the ground. Buildings in those areas are considered to have a certain degree of
seismic vulnerability. Above all, it is common for high-rise, i.e. multi-storey buildings, to
retain a comparably higher degree of seismic vulnerability. Thus, in order to understand
and lower the seismic vulnerability, a separate field of structural engineering, known as
earthquake or seismic engineering, has been developed. The field began as early as 1876
(Howe 1936) and it is still expanding to this day. It has contributed to various cognitions
and provisions, termed anti-seismic regulations for civic and building protection. Some
provisions are regulated nationally, and some internationally. A list of such provisions is
shown on Table 1. Meanwhile, the information in Table 1 leads to the conclusion that all
codes are more or less intertwined, with a majority of codes based upon ACI (2011) provi-
sions from the US.
The majority of high-rise, i.e. multi-storey buildings, have their structural load-bearing
framework made of either reinforced concrete (RC) or structural steel (SS) with some sort
of non-bearing infill wall/panel. In Southern Europe, the traditional structural systems are
composed of RC frames with non-bearing unreinforced masonry (URM) walls (Booth and
Key 2006), hereafter referred to as infill. Most commonly, hollow clay blocks are used as
infill units. They possess better thermal properties, greater shear strength and work effi-
ciency than solid bricks. Most often, blocks are stacked with voids which are vertically
facing. However, in the traditions of countries such as in Portugal and Turkey blocks are
laid so the voids are facing a horizontal direction (de Sousa 2014; Misir et al. 2012). When
subjected to ground motions, structures are prone to inertial and inter-storey drift forces.
With infilled frame structures, the inter-storey drift forces are transmitted as displacements
through rigid diaphragms, i.e. slabs. Meanwhile, the inertial forces are the consequence of
accelerated higher masses, predominately of the infill. Therefore, the failures from inertial
forces are expected at higher storeys and with infills that have less connections in their
boundary conditions. Greater inter-storey drift damage is expected at lower storeys, where
the shear forces are greater, along with an additional reduction in the infill’s inertial force
characteristics from gravity loads of upper storeys. Most commonly, the inter-storey drift
damage is observable in the heavy cracking of the frame and the development of plastic
hinges (Fig. 2a, b). Damage from inertial forces is observable in the heavy cracking of
the infill and, frequently, the disappearance of larger parts of the wall (Fig. 2c–e). During
ground motion excitations, infill interacts with the surrounding frame, thus rendering its
behaviour and, with it, the behaviour of the whole structure. Due to its complex problem,
this interaction is yet to be implemented in European seismic codes (BSI 2005). Hence, it
is one of the most popular researched subjects in the field. Specifically, it was known from
the early 1960s (McDowell et al. 1956b; Holmes 1961) that infill does indeed contribute to
the overall behaviour of frames. However, a way to implement the interaction is still being
thoroughly researched and, at the time of code development, largely unknown. Therefore,
with limited information, Eurocode 8 (BSI 2005) regards infill as a secondary element dur-
ing ground motions. Accordingly, this has resulted in more redundant frames.
In order to better comprehend the complex problem of the frame-infill interaction, many
studies have been conducted since the early 1960s. The field has broadened and divided
into three main components, listed by prevalence:
1. In-plane (IP), where the frames are loaded in the direction of their axial plane;
13
Table 1 Worldwide seismic building codes
# Code Remark
1 EN Eurocode Parts 8,6; BSI (2005)

2 Italian seismic codes NTC(2008)
3 American Concrete Institute ACI-318; ACI (2011)
4 Applied Technology Council (ATC)—Federal Emergency Management Agency; FEMA 273 (1997) and FEMA 356 (2000)
5 Indian Standards of Earthquake Engineering; IS 1893 (2002 )
6 Peruvian Technical Building Standard; E.030 NTS (1997)
7 Chilean; NCH 433 (1996) Based upon ACI
8 Ecuadorian INEN-5 (2001) Based upon ACI

9 Colombian Standards for Seismic Resistant Design and Construction; NSR-84 CAEE (1984) Based upon ACI
10 National Structural Code of Philippines; Vol. 1, Fourth Edition NSCP (1992) Based upon ACI
11 New Zealand: Code of Practice for the Design of Concrete Structures; Part 1 NZS-3101 (1995)
12 Israeli Design Provisions for Earthquake Resistance of Structures SI-413 (1995)
13 Nepal National Building Code; NBC-105 (1995) Largely following IS code

14 Russian Building Code on Construction in Seismic Areas; SNiP-II-7–81 (1995)
15 Venezuelan Seismic Code, 1988. Regulations for Earthquake Resistant Buildings; CNIC (1988)
16 Vietnamese Seismic Design Standard TCXDVN 375: 2006; VBS (2006) Based upon SNiP
17 Canadian Standards Association and Standards Council of Canada
2193
13
2. Out-of-plane (OoP) where the frames are loaded perpendicular to their axial plane;
3. Bidirectional (IP + OoP, OoP + IP and simultaneous).
The IP + OoP describes OoP behaviour due to previous IP damage, and vice versa in
the case of OoP + IP. Simultaneous action describes a synchronous IP and OoP loading
protocol. Furthermore, the greatest focus has been on the IP field, with less attention
on the OoP field and hardly any in the case of bidirectional behaviour (Asteris et al.
2017). In recent years, OoP research has gained momentum (Fig. 1) in the form of
experimental surveys (Tables 2, 3, 4), and micro (Reindl et al. 2011; Yuen and Kuang
2012), macro (Kadysiewski and Mosalam 2009, Al Hanoun et al. 2019; Furtado et al.
2016, Di Trapani et al. 2018) and analytical modelling (Table 6). Regarding the aspect
of seismic codes on OoP behaviour, Eurocode 8 (BSI 2005) provisions only restrict
the OoP collapse of slender infills (if h∕t > 15 ). It states that, in such cases, special
measures should be addressed, such as the application of: light wire meshes, wall ties
fixed to the columns, wind posts and concrete belts (BSI 2005). The US codes of the
Masonry Standards Joint Committee et al. (1999) propose a calculation of OoP load-
bearing capacity with the use of Dawe and Seah (1989) equations (Eqs. 3, 4). On the
other hand, the Canadian code (Canadian Standards Association and Standards Council
of Canada 1978) does not include a specific method for obtaining the OoP capacity.
However, it states that the arching action (Sect. 2.1) method should be used. US fed-
eral recommendations FEMA 356 (2000) and FEMA 273 (1997) and those based upon
FEMA, as in the case of the New Zealand codes (NZS-3101 1995), use the approaches
to assess OoP damage types and calculate OoP bearing capacity; see (Eq. 10), which is
a simplified version of the equation found in Angel et al. (1994) (Eq. 5). The first stud-
ies that researched the OoP behaviour of infill walls where from the blast, rather than
the seismic engineering field. Pioneering work was undertaken by the Armour Research
Foundation, sponsored by the US Air Force, involving authors such as McDowell et al.
(1956a, b) and Monk (1958). In blast engineering, the OoP load is introduced by the
explosion’s blast waves which pressurize the infill. Hence, the experimental work was
carried out by pressurizing the infill, most commonly with airbags. Blast engineers laid
the groundwork for the field of seismic engineering, using identical methods in order
to simulate the OoP inertial forces; hence, such methods are here referred to as iner-
tial. Uniquely, the field of seismic engineering has also developed additional methods
for testing structures, these are known as dynamical and inter-storey drift methods.
11
Number of articles (#)
10 Published articles
9
8 Trend line
7
6
5
4
3
2
1
0
Year
Fig. 1 Number of OoP—infilled frame specific articles as published by year
13
Table 2 OoP experimental loading methods and boundary conditions
# Study by OoP load type OoP load method IP loading method Gravity load Gaps considered
None @ beam @ columns
1 Dawe and Seah (1989) Monotonic Inertial—AB No – + + +

2 Henderson et al. (1993) Cyclic Inter-storey drift Cyclic – + – –
3 Angel et al. (1994) Monotonic Inertial—AB Cyclic + + – –
4 Fowler (1994) Dynamic Shaking table Dynamic + + – –
5 Klinger et al. (1996) Dynamic Shaking table Dynamic + + – –
6 Flanagan and Bennett (1999a) Monotonic & Dynamic Inertial—AB & inter-storey drift Cyclic – + – –
7 Hashemi and Mosalam (2007) Dynamic Shaking table No – + – –
8 Tu et al. (2010) Dynamic Shaking table No + + + –
9 Komaraneni et al. (2011) Dynamic Shaking table Cyclic – + – –
10 Liu et al. (2011) Dynamic Shaking table No – + + –
11 Pereira et al. (2011) Cyclic (q-static) Inertial—double leafed AB Cyclic – – + –

12 Rabinovich and Madah (2011) Dynamic Shaking table No + + – –
13 da Porto et al. (2013) Monotonic Inertial—L Cyclic + + – –
14 Hak et al. (2014) Cyclic (q-static) Inertial—P Cyclic – + – –
15 Akhoundi et al. (2015, 2016) Cyclic (q-static) Inertial—AB No – + + –
16 Furtado et al. (2015) Cyclic (q-static) Inertial—AB Cyclic + + – –

17 Petrus et al. (2015) Cyclic (q-static) Inertial—L No – + – –
18 Preti et al. (2015) Cyclic (q-static) Inertial—P Cyclic – + + –
19 Misir et al. (2016) Cyclic (q-static) Inertial—AB Cyclic + + – –
20 Mosoarca et al. (2016) Cyclic (q-static) Inertial—L No – + – –
21 Verlato et al. (2016) Monotonic Inertial—P Cyclic – + – –
22 Arede et al. (2017) Cyclic (q-static) & monotonic Inertial—AB Cyclic + + + –
23 Sepasdar (2017) Monotonic Inertial—AB No – + – –
24 Vailati et al. (2017) Cyclic (q-static) Inertial—AB Monotonic – + – –
2195
13
Table 2 (continued)
2196
# Study by OoP load type OoP load method IP loading method Gravity load Gaps considered
None @ beam @ columns
13
25 Wang (2017) Monotonic Inertial—AB No – – + +
26 Di Domenico et al. (2018) Cyclic (q-static) Inertial—P No – + + +
AB airbag, P point load, L line load

Table 3 OoP experimental tests’ material characteristics and subject studied
# Study by Frame Infill Subject studied
1 Dawe and Seah (1989) SS Concrete blocks dry and normally stacked Effects of centric window opening, dry & normally stocked block,
boundary conditions
2 Henderson et al. (1993) SS Clay block layed horizontally OoP + IP loading
3 Angel et al. (1994) RC Concrete blocks and bricks Infill slenderness variation, mortar type variation
4 Fowler (1994) SS Clay block
5 Klinger et al. (1996) RC Clay block Weak and strong frames
6 Flanagan and Bennett (1999a) SS Clay block IP + OoP & OoP + IP loading
7 Hashemi and Mosalam (2007) RC Hollow clay brick
8 Tu et al. (2010) RC Clay block Confined masonry wall and unconfined infill panels
9 Komaraneni et al. (2011) RC Clay block Weak & strong frame, confined and unconfined panels
10 Liu et al. (2011) RC Concrete blocks Special reinforced beam-infill connections observed
11 Pereira et al. (2011) RC Hollow clay brick
12 Rabinovich and Madah (2011) RC Concrete blocks Gap between infill and columns
13 da Porto et al. (2013) RC Clay block Included URM & reinforced masonry units
14 Hak et al. (2014) RC Clay block Full wall height opening and vertical stripes + IP + OoP
15 Akhoundi et al. (2015, 2016) RC Hollow clay brick Centric window opening, effects of contact surface between airbag
and infill
16 Furtado et al. (2015) RC Single & double leaf hollow clay block layed horizontally Effects of gravity load
17 Petrus et al. (2015) SS Clay block
18 Preti et al. (2015) SS Concrete block layed horizontally Considered sliding joints to limit infill-frame response (from Preti
et al. 2012)
19 Misir et al. (2016) RC Block: pumice, hollow-fired clay, autoclaved aerated Simulations IP & OoP loading, structural and design (code) rec-
concr. & Brick: insulated & horizontally layed, differ- ommendations and limitations were established
ent thickness
20 Mosoarca et al. (2016) RC Clay block Innovative infill strengthening system
21 Verlato et al. (2016) RC Clay block Innovative infill strengthening system, full wall height opening
2197
13
2198
# Study by Frame Infill Subject studied
13
22 Arede et al. (2017) RC Clay block Innovative test setup
23 Sepasdar (2017) RC Concrete block Centric window opening
24 Vailati et al. (2017) RC Clay block Innovative earthquake-safe & eco-friendly infill panel system
25 Wang (2017) RC & SS Concrete block Centric door opening, different frame material
26 Di Domenico et al. (2018) RC Clay block Boundary variation
Table 4 OoP experimental tests’ geometrical characteristics
# Study by Infill Openings Scale Specimen
h∕t l∕h Ao ∕Ain (%) Type # Storeys Bays
1 Dawe and Seah (1989) 14.74, 20.00, 31.11 1.29 19.05 Window 1:1.0 9 1 1
2 Henderson et al. (1993) 19.20 1.20 1:1.0 2 1 1
3 Angel et al. (1994) 8.73, 39.68, 16.09, 10.73 0.75 1:1.0 8 1 1
4 Fowler (1994) 10.12 1.64 1:1.0 1 1 1
5 Klinger et al. (1996) 19.52 0.74 1:2.0 2 1 1
6 Flanagan and Bennett (1999a) 22.40, 11.20, 6.79 1.00 1:1.0 3 1 1
7 Hashemi and Mosalam (2007) 23.78 1.35 1:1.0 2 1 1
8 Tu et al. (2010) 5.60 0.96 1:1.0 4 1 2
9 Komaraneni et al. (2011) 22.67, 11.33 1.78 1:2.0 3 1 1
10 Liu et al. (2011) 1.90 1.67 1:3.0 2 2 1
11 Pereira et al. (2011) 2.33 2.00 1:1.5 4 1 1

12 Rabinovich and Madah (2011) 21.00 0.58 1:1.0 2 1 1
13 da Porto et al. (2013) 22.08 1.57 1:1.0 6 1 1
14 Hak et al. (2014) 12.55 1.43 34.14 Full wall height 1:1.0 5 1 1
15 Akhoundi et al. (2015, 2016, 2018) 14.86 1.48 12.45 Window 1:2.0 6 1 1
16 Furtado et al. (2015) 2.33, 1.35 1.29 1:1.0 3 1 1

17 Petrus et al. (2015) 14.00 0.84 1:1.0 1 1 2
18 Preti et al. (2015) 11.50 1.24 7.78 Window 1:1.0 2 1 1
19 Misir et al. (2016) 8, 8.51, 10.53, 14.85, 23.53 1.80 1:1.0 6 1 2
20 Mosoarca et al. (2016) 1.40 0.76 1:1.0 3 1 1
21 Verlato et al. (2016) 9.17 1.40 33.25 Full wall height 1:1.0 4 1 1
22 Arede et al. (2017) 1.53 1.83 1:1.0 5 1 1
23 Sepasdar (2017) 10.89 1.37 17.38 Window 1:2.0 4 1 1
24 Vailati et al. 2017 7.60 1.84 1:1.0 7 1 1
2199
13
2200
# Study by Infill Openings Scale Specimen

h∕t l∕h Ao ∕Ain (%) Type # Storeys Bays
13
25 Wang (2017) 10.89 1.37 17.38 Door 1:2.0 4 1 1
26 Di Domenico et al. (2018) 29.38 1.28 1:3.0 3 1 1
Bulletin of Earthquake Engineering (2020) 18:2191–2246 2201
(a) Infill wall damaged due to inertial forces, Nepal (b) Façade wall damaged by inertial forces, Kaikoura
(c) Muisne earthquake inter-storey drift failure, Ecuador (d) Muisne earthquake inter-storey drift failure, Ecuador
(e) Muisne earthquake combined failure, Ecuador (f) Muisne earthquake combined failure, Ecuador
Fig. 2 Various failures as recorded after real earthquake scenarios (Photographs from EERI 2018)
Dynamical methods are performed on shaking tables, while inter-storey drift methods
load the frame rather than infill. Furthermore, many studies from the blast and seis-
mic engineering fields overlap. For instance, Gabrielsen et al. (1975), from the blast
engineering field, cited arching actions observed in the 1967 Caracas and the 1971 San
Fernando earthquakes. In addition, there are other branches studying the same phe-
nomena. However, these are not studied as thoroughly as the main ones, addressing, for
example, the effects of wind (Anderson and Bright 1976), load and soil pressure (Jäger
et al. 2008). This paper presents a review of the literature on the OoP and bidirectional
behaviour of infilled frames with a focus of seismic engineering. Additionally, various
OoP experiments have been conducted on load-bearing brickwork (no frame), such as in
Drysdale and Essawy (1988), Hallquist (1970), Lam et al. (2003) and Vaculik (2012).
13
Their findings show numerous similarities with inertial force tests of infilled frames.
In this paper, however, only a few such studies were included as the main focus of this
paper was on framed structures (Fig. 2).
This paper revised the literature, which has included various journal and proceedings
papers, as well as theses (MSc and PhD). The review scope included manuscripts pub-
lished until the end of 2018. Furthermore, there were three specific articles that are in
scope of this paper:—Furtado et al. (2018b) made a systematic state-of-the art paper that
gathered articles from specific electronic indexing bases; Pasca et al. (2017) made a paper
about analytical models and their reliability; Asteris et al. (2017) made a more general
overview that also included macro and micro modelling.
2 Review of experimental campaigns
This section covers a review of experimental assessments of infilled frames with regard to
OoP and bidirectional loading within the scope of seismic engineering. Data from experi-
mental campaigns are gathered and presented in Tables 2, 3, 4 and 5. The distribution of
various characteristics from Tables 2, 3 and 4 are plotted in Figs. 3, 4 and 5. From the
provided tables, it is clear that frames are made as either strong or weak (Klinger et al.
1996; Flanagan and Bennett 1999a; Wang 2017) and constructed of either RC or Struc-
tural Steel (SS). Infill panels were mostly built as single-leafed vertically laid clay or con-
crete blocks or bricks. There are exceptions, mainly in the case of Portuguese and Turkish
research, where blocks are traditionally laid horizontally. Portuguese masonry also pos-
sesses lower compressional strength (Table 5), likely due to the direction of voids. Vari-
ous conditions were tested in order to examine their influence on OoP behaviour, such as:
boundary conditions; slenderness; aspect ratio; and effects of gravity load. Furthermore,
one can point out that most experiments were conducted with inertial force methods, typi-
cally with single-leaf airbags, loaded in one direction (Table 2, Fig. 3a). The cyclic, quasi-
static airbag OoP loading method is conducted in one direction by loading and unloading
the infill, i.e. by pressurizing and depressurizing the airbag. Only a few studies are avail-
able on shaking tables, where the excitation was provided by an earthquake signal or a
sine wave (Table 2, Fig. 3b). Further, there are only two authors (Henderson et al. 1993;
Flanagan and Bennett 1999b) who have researched the inter-storey drift method, mainly as
a counterpart to bidirectional studies (Table 2, Fig. 3b). Most bidirectional tests were firstly
carried out by implementing the IP inter-storey drift force to the beam, after which, the
infill was loaded inertially with OoP loads. Conjointly, there was only one experiment car-
ried out to study the effects of previous OoP damage on IP behaviour (Flanagan 1994), and
only two with simultaneous OoP and IP loading (Flanagan 1994; Misir et al. 2016). Fur-
thermore, researchers have recognized the importance of the connection between the frame
and the infill, i.e. boundary conditions, many of which have been studied. Researchers such
as Akhoundi et al. (2016) refer to boundary conditions as effects of workmanship, as on-site
building can result in a loss of connections due to poor workmanship. Most commonly,
the effects of poor workmanship are reflected in a partial or absent connection between
the upper beam and infill (Figs. 7b, 8b, 9b), as that connection is harder to fill. Others,
such as Wilton and Gabrielsen (1973), refer to gaps as practical consideration, since infill
can move more or less independently from the frame. Such considerations were taken into
account by disconnecting the columns from the infill (Fig. 8c). Also, boundary conditions
are affected by earthquake action itself, as the first feat of infill—frame interaction is its
13
Table 5 Various geometrical and mechanical properties of specimens
Author Specimen t (mm) l (mm) h (mm) Frame type Column Beam size fm (MPa) Em (MPa) EF (MPa) Capacity w Dis- Remark
size (mm*mm) (kPa) place-
(mm*mm) ment
(mm)
Dawe and WE1 190 3600 2800 SS W250 × 58 W200 × 46 24.3 17,575 210,000* 22.30 n/a Truss type reinforc.
Seah
(1989)
WE2 190 19.20 n/a Plain masonry panel
WE3 190 7.80 n/a Dry-stack panel
WE7 190 14.70 n/a Truss type reinforc.
WE8 140 13.40 n/a Restraints against
slipping
WE9 190 17.40 n/a Window opening
Angel et al. 1 48 2740 1630 RC 305*305 203*254 11.50 8046 24,821 8.19 n/a Pure OoP, S mortar
(1994)
2 48 10.81 8046 4.02 n/a IP (0.34% dr) + OoP,
N mort.
3 48 10.14 5212 5.99 n/a IP (0.22% dr) + OoP,

L mort.
4 92 22.90 12,438 29.78 n/a IP (0.09% dr) + OoP,
N mort.
5 143 21.46 11,624 32.22 n/a IP (0.06% dr) + OoP,
N mort.
6 98 4.59 2137 12.40 n/a IP (0.25% dr) + OoP,
L mort.
7 98 11.00 2923 30.74 n/a IP (0.25% dr) + OoP,
N mort.
2203
13
2204
13
(mm*mm) ment
(mm)
8 187 3.50 2358 32.08 n/a IP (0.39% dr) + OoP,

L mort.
Flanagan 25 100 2240 2240 SS W250 × 45 W310 × 52 5.60 5300 199,948 8.10 25.30 Plain masonry panel
and
Bennett
(1999b)
18 200 W250 × 45 W310 × 52 26.60 11.50 Plain masonry panel
19 200 W250 × 45 W310 × 52 21.70 19.30 IP (0.8% dr ) +OoP
22 330 W410 × 60 W460 × 113 2.29 5040 39.50 49.50 Plain masonry panel
Hak et al. TA1 235 4220 2950 RC 350*350 350*350 4.64 5299 24,821 13.25 n/a IP (1.50% dr) + OoP
(2014)
TA2 8.11 n/a IP (2.00% dr) + OoP
TA3 13.01 n/a IP (1.00% dr) + OoP
Akhoundi SIF-A 110 2415 1635 RC 160*160 270*160 1.00 1000* 32,000* 8.86 25.00 Beam—infill gap
et al.
(2015)
SIF-B 10.13 12.00 Plain masonry panel

PIF-A 9.88 25.00 Window opening
Furtado inf_01 150 4200 2300 RC 300*300 300*500 0.53 942 24,300 7.76 22.00 OoP mono-
et al. tonic + gravity load
(2015)
inf_02 7.25 12.00 OoP cyclic
inf_03 150 + 110 1.76 1.50 OoP cyclic + IP(0.5%
dr ), 2 leafed
Sepasdar IF-W 90 1350 980 RC 180*180 180*180 9.00 7650 16,911 43.70 4.30 Window opening
(2017)
(mm*mm) ment
(mm)
IF-ND 66.30 12.50 Plain masonry panel

IF-D1 44.40 6.60 IP (0.66% dr) + OoP
IF-D2 26.40 9.90 IP (2.70% dr) + OoP
Wang IF-RC-DO 90 1350 980 RC 180*180 180*180 9.00 7650 16,911 36.20 7.90 Door opening
(2017)
IF-RC-TG 18.50 3.90 Beam—infill gap
IF-RC-SG 36.50 7.40 Columns—infill gaps
IF-RC-ID 37.60 7.70 IP (1.37% dr) + OoP
IF-S SS W150Χ30 W150Χ30 201,172 34.30 15.10 Steel frame
Di Domen- OOP_4E 80 2350 1830 RC 200*270 200*270 1.80 1517 32,000* 4.09 5.40 All bounded
ico et al.
(2018)
OOP_3E 3.39 14.60 Beam—infill gap
OOP_2E 8.42 17.00 Columns—infill gaps
*Estimated
2205
13
Monotonic,
Inter-storey Monotonic, Dynamic,
drift, 7 % 4%
27 % 8%
Dynamic,
26 % Cylic, Cylic,
46 % 46 %
None,
42 %
Inertial, Dynamic,
70 % 30 %
(a) OoP loading type (b) OoP loading method (c) Additional IP load
Fig. 3 Distribution of OoP load methods
Horizontal,
Concrete,
19 % SS, 31
31 %
%
Clay, 77 Vertical, RC, 73

% 81 % %
(a) Masonry type (b) Laying method (c) Frame type

Fig. 4 Material specifics of experimental tests
2.0
30 Max
30
Interquartile range
Third quartile
Ao / Ain (%)
20 1.5
l/h
Special measures required by EC8

h/t
20 Median
10 1.0 First quartile

Negligible openings by EC6
10 Min
0 0.5
(a) Distribution of slenderness (b) Distribution of aspect ratio (c) Distribution of opening to infill area ratio (d) Boxplot legend
Fig. 5 Distribution of geometrical characteristics
debonding, i.e. losing the connection. Correspondingly, Fowler (1994) and Tu et al. (2010),
when conducting dynamic tests, and Flanagan and Bennett (1999b), in the case of OoP
inter-storey drift cyclic tests, found that the infill and the frame move together as a single
unit. Thus, a disconnection can cause an independent action between the two, which, in
turn, magnifies the infill’s inertial forces. Therefore, due to a loss in connection, failure due
to inertial forces can occur as was the case with an A1 specimen (Fig. 24a) from Tu et al.
(2010)). Research has also been focused on openings. Namely, with IP studies, it has been
shown that openings reduce bearing capacity and affect failure mechanisms (Sigmund and
Penava 2014; Surendran and Kaushik 2012; Tasnimi and Mohebkhah 2011). However, the
same cannot be stated with certainty, as there are some opposing results in the OoP litera-
ture. Furthermore, Fig. 6 presents all the openings that have been studied in the field on
the OoP loading of infilled frames. Presumably, similar window opening as in Fig. 6a have
been tested by Dawe and Seah (1989); however, figures and detailed descriptions are miss-
ing. For the same reason, a presumably similar full-height opening, as in Fig. 6e, was tested
by Verlato et al. (2016). It should be noted that all openings were placed centrally and all
13
Wood planks
Vertical wood
elements
(a) Window without lintel Sepasdar (b) Window with lintel (Akhoundi (c) Window without lintel (Preti et
(2017) et al., 2016) al., 2012)
(d) Door without lintel (Wang, 2017) (e) Full height opening (Hak et al., 2014)
Fig. 6 Opening considered in OoP research
were tested with inertial force methods. Further, their area-to-infill ratio was Ao ∕Ai > 10 .
Hence, it is within the limit by which opening size plays a role in masonry behaviour,
as stated in the EN1996-1 provisions (BSI 2004). Figure 5c was not made as a boxplot
because there were only six data points. In the same figure, a BSI (2004) limit is also plot-
ted, which states that openings with the ratio of Ao ∕Ai ≤ 10 can be neglected.
2.1 Arching action and boundary conditions
The principal finding of the field was that, unlike the expected flexural-induced failure, a
compressive, arching action failure occurs. This was first discovered on URM wall tests
by the Armour Research Foundation in 1951 (McDowell et al. 1956b). Following this, in
1956, McDowell (McDowell et al. 1956b), under the aegis of the US Air Force, started a
discussion about the newly found arching action. He hypothesized about a solution to the
question about why masonry walls obtain a six fold greater resistance than simply sup-
ported slabs. He argued that this was due to additional compressive forces developed by
arching action. The hypothesis was approved; and, today, arching action is a standard phe-
nomenon in the field and has been translated into a theory, termed arching action theory.
McDowell et al. (1956a) was also the first to develop an equation with which to calculate
the ultimate OoP resistance of load-bearing brickwork with the effect of arching action
(Eq. 2).
The development of arching action (Fig. 7a), as described by Moghaddam and Goudarzi
(2010), Dawe and Seah (1989), starts with the bending of infill. Infill bends as a flexible
beam would, until it cracks, typically at three points: (1) interface at mid-height; (2) and
(3) boundary cracking (crushing at the frame). Those three points are clamping on one
end, and open on the other. Clamping points create compression arch or compressive mem-
brane forces (Flanagan and Bennett 1999a) (Fig. 9a) which resist transversal ones. Such
arching action was firstly reported by McDowell et al. (1956b) as a three-hinged arching
action in the case of a similar bearing mechanism of a three-hinged arch. Furthermore,
diagonal cracks start to propagate from the upper half of the infill until they meet with
horizontal cracks at the mid-points. From there, diagonal cracks start to progress on the
lower half of the infill. The yielding pattern is similar to patterns in plate theories (Quintas
13
Undeformed Beam bending Arching action Undeformed Cantilever bending Arching action
(a) Rigid arching action development (b) Gapped arching action development
Fig. 7 Arching action development phases by different boundary conditions
(a) Two-way (rigid) arching action (b) Gapped arching action (c) One-way (double-gapped) action
Hinge line Frame-infill gap
Fig. 8 Types of arching action in relation to boundary conditions
Side view
(a) Vertical strip of three (b) Vertical strip of three hinged (c) Vertical strip of two hinged (d) Vertical strip of four
hinged arching action gapped arching action arching action hinged arching action
Top view
(e) Horizontal strip of two-way, four-hinged action (f) Horizontal strip of one-way, three-hinged action
Fig. 9 Force-displacement curves of various boundary conditions
2003). Moghaddam and Goudarzi (2010) found that high slenderness, low-boundary stiff-
ness or weak infill material may cease arching action development or its benefits may be
limited. Correspondingly, Moghaddam and Goudarzi (2010) divided the problem in two
13
modes: (1) crushing at the boundary and (2) transverse instability. Crushing at the bound-
ary occurs when thick panels are crushed at the supports (frame), while transverse insta-
bility failure results from excessive transversal deflections. Similar analogy can be drawn
from the plate and shell theory, whereas the behaviour of panels with low slenderness may
be less influenced by membrane forces as is the case for thick plates and shells (Flana-
gan and Bennett 1999a). Boundary conditions affect failure modes (Fig. 8), load-bearing
and deformation capacities (Fig. 10), and arching action mechanism (Fig. 9). If the infill
is bounded by all sides, two-way action (horizontal and vertical arching) occurs (Figs. 8a,
9a) and, with it, greater load-bearing and deformations capabilities (Fig. 10). Yet, if infill is
bounded by three (infill-beam gap) or two sides (infill-columns or infill-beams gap), one-
way action occurs (Fig. 8b, c). Then, load-bearing and deformation capabilities are low-
ered in comparison to fully bounded ones (Fig. 10). Single-gapped, one-way arching action
(Fig. 8b) involves both vertical and horizontal arching. However, the horizontal arch is a
three-hinged phenomenon (Fig. 8f) and thus less effective. Meanwhile in two-way horizon-
tal arching action, four-hinged arching action occurs (Fig. 8e). Comparatively, the double-
gapped phenomenon is a truly one-way arching action (Fig. 8c), as it involves only vertical
or horizontal, three-hinged arching action and thus has the lowest OoP resistance. Fig-
ures 10 and 12b lead to the conclusion that the less of the connection there is between the
infill and the frame, the greater the loss in load-bearing and deformation capabilities. The
force–displacement curves from Dawe and Seah (1989) show greater differences in initial
stiffness and a smaller effect on deformation capabilities between various boundary condi-
tions. However, the results from Akhoundi et al. (2016) and Di Domenico et al. (2018)
show a less profound change in initial stiffness and in load-bearing outcomes (Fig. 10).
Yet, both graphs show a great reduction in deformation capabilities. These differences can
be attributed to the specimens’ slenderness. Dawe and Seah (1989) have profoundly thicker
panels when compared to those of Di Domenico et al. (2018) (Table 4). Additionally, it
should be noted that specimens containing infill-columns interface gaps demonstrated very
brittle behaviour due to crushing soon after reaching peak load (Fig. 10).
Symmetrical two-way arching action is sometimes referred to as a rigid arching action
(Fig. 10), while an unsymmetrical one-way arching action is the so-called gapped arching
55 22
50 20
45 18
40 16
Force (kN)
35 14
30 12
25 10
20 8
15 6
10 4
5 2
0 0
10 25 40 55 70 85 100 115 130 145 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
OoP deflection (mm) OoP deflection (mm)
(a) Pressure vs. deflection of Dawe and Seah (1989) (b) Force vs. deflection of Di Domenico et al. (2018)
All sides connected Infill-beam interface slip
Infill-beam interface gap Infill-columns interface gaps
Fig. 10 Force-displacement curves of various boundary conditions
13
action (Gabrielsen et al. 1975). As explained by Wilton and Gabrielsen (1973), when a
gapped panel is loaded, infill bends as a cantilever would until it reaches the frame. Then,
infill cracks at the bottom and wedges. After cracking, the divided elements develop a com-
pression arch (Fig. 7b). However, the arch has a linear (thrust) part on the upper half and a
non-linear (arch) part on the lower half of the infill.
Additional vertical arching action modes were found; however, these are not as com-
mon as three hinged action modes. In this case, they included two- and four-hinged arching
action. The two-hinged arching action (Fig. 10c) was observed in dynamical tests carried
out by Tu et al. (2010). The two-hinged arching action starts when infill cracks on the ten-
sion sides of the panel (top and base), thus a diagonal thrust forms (not arch) between the
opposite compression sides. Then, the panel segment rigidly rocks between the crushed
ends. Tu et al. (2010) concluded that such behaviour is analogous to a slender rocking pier.
The two-hinged arching action is, in essence, a compression strut; however, for the sake of
uniformity, it will be referred to as an arch. It is noteworthy that two-hinged arching action
is the only instance of specifying an arching action in dynamical experiments. Still, there
is no description of any arching action occurring with inter-storey drift method, however,
in numerical studies by Anić et al. (2019) a two-hinge arching action similar to that of Tu
et al. (2010) was found. Four-hinged vertical arching action (Fig. 9d) was reported by Var-
ela-Rivera et al. (2012) in confined masonry panels. Such arching action can be expected
with high walls and is the product of rigid, horizontal arching action (Fig. 9e).
2.2 Effects of openings
All experimental efforts to address the effects of openings were carried out using iner-
tial force methods with centrally positioned openings. The window opening in Preti et al.
(2012) (Fig. 6c) was not scoped within this paper because (of its specific infill) assem-
blage and the fact that it falls within the area of negligence Ao ∕Ai ≈ 8 < 10% accord-
ing to EN 1996-1-1 provisions (BSI 2004). Other openings had their area in the range of
Ao ∕Ai ∈ [17, 30] > 10 % when compared to the area of infill. Hence, they exceeded the
limit of negligence.
It was found that all openings cause a significant reduction in deformation capabilities
when compared to fully infilled frames (Figs. 12, 15). Yet, the same was not evident with
regard to load-bearing capacities, as different authors have reported different outcomes
(Fig. 15). For instance, Akhoundi et al. (2016) (Fig. 12b) and Dawe and Seah (1989) found
no reduction in bearing capacities, while (Wang 2017; Sepasdar 2017; Verlato et al. 2016)
observed a significant decrease (Fig. 12a, c). To relate the areas of opening to differences
in ultimate force and displacement, Fig. 15 was produced. Note that the pure OoP loading
from Verlato et al. (2016), containing an opening, was compared with the full infill which
previously has a 2.5% IP inter-storey drift, i.e. previous IP damage. By the same figure, no
logical conclusion about the impact of opening size can be drawn. Conversely, Mays et al.
(1998), in the field of blast engineering studies on RC panels with openings, proposed a
linear degradation in the ultimate resistance of the window and an increase in the ultimate
resistance of the doors in relation to the opening area. The reduction in door and window
openings, )as noted in Fig. 11, were derived from Eq. (21) and Table 9 (Mays et al. 1998) as
Fr Ao ∕Ai .
(
By observing crack patterns in infills with openings (Fig. 14), along with load-bearing
graphs (Fig. 12) and deformation map in Fig. 11b, it is evident that arching action was able
to develop. It is, above all, clear in the case of window openings (Figs. 11b, 12b, 14c) as
13
30-35
75-80 25-30
60-75 20-25
45-60 15-20
30-40 10-15
15-30 5-10
0-15 0-5
(a) Full infill specimen (SIF-O-1L-B ) (b) Window opening specimen (PIF-O-1L-B)
Fig. 11 Final OoP loading step deformation patterns in mm (Akhoundi et al. 2018)
120
110 40
100 35
90 30
80
Force (kN)
Force (kN)
70 25
60 20
50
15
40
30 10
Full infill Sepasdar (2017) Full infill
20 5
Window Sepasdar (2017) Beam-infill gap
10
Door Wang (2017) 0 Window opening
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
OoP deflection (mm) OoP deflection (mm)
(a) Window and door, no lintel (b) Window with lintel (Akhoundi et al. 2016)
205
180
155
Force (kN)
130
105
80
55
30 Full infill @ IP dr = 2.4%
5 Full height opening
0 5 10 15 20 25 30 35 40 45 50 55
OoP deflection (mm)
(c) Full height opening (Verlato et al. 2016)
Fig. 12 Force versus displacement graphs of specimens with openings
familiar rigid arching action “X” pattern emerges (Fig. 8a). Likewise, by comparing defor-
mation fields in Fig. 11b the arching on the right side of the opening is apparent. From the
same figure, it is visible that the right side is more affected; to the point that there is detach-
ment of the infill from the column. However, the loss in the deformation capabilities are
obvious when comparing it to the fully infilled frame (Fig. 11a vs. b). It is also evident that
13
3rd 4th 1st

1st 3rd
nd
2
2nd
2nd 3rd
nd
2
2nd
(a) Door and window (b) Eccentric door

Fig. 13 Results of the Mays et al. (1998) URM wall test with designated crack occurrence
(a) Centric door opening by Wang (b) Centric window opening by (c) Centric window opening with
(2017) Sepasdar (2017) lintel by Akhoundi et al. (2016)
Fig. 14 Failure patterns of OoP specimens with openings
50
35
20 Force Displacement
5 Window Door
-10
-25
-40
-55
-70
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Fig. 15 Opening to infill area ratio versus difference in force and displacement from the fully infilled frame,
window and door lines were extracted from (Mays et al. 1998)
the opening type influences the arching action. Particularly, the window has greater capa-
bilities of forming an arching action than a door or full-height openings. Openings have
a slight impact on initial stiffness (Fig. 12), whereas a door has more influence on initial
stiffness than the window opening (Fig. 12a).
The results of full-height openings by Hak et al. (2014) are not published in the cited
paper nor others. In studies by Dawe and Seah (1989), no detailed descriptions of the
ultimate displacement value, a force–displacement graph or crack patterns were found
for the specimen with a window opening. Further, in Verlato et al. (2016), no crack
13
patterns of the specimen with a full-height opening is available. In other words, there
are no data available about full-height opening crack patterns.
The copious difference in resistance reduction between the tests of Akhoundi et al.
(2016), Dawe and Seah (1989) (Fig. 12b) and Sepasdar (2017), for centric windows
(Fig. 12a, b) may be due to lintel. Window with lintel obtained resistance to almost the
same degree as a fully infilled one. Moreover, yielding lines did develop at the lower
half of the infill. The same cannot be observed in the specimen without lintel. Like-
wise, in preliminary studies by Anić et al. (2018), lintel accumulated greater amounts of
compressive stress; hence, there was an additional clamping point in the arching action
curve.
It is noticeable that crack patterns in the specimen with a door opening (Fig. 14a) do
not resemble the arching action patterns as window openings do. All cracks are more
or less vertical. Similar patterns were obtained in OoP tests of load-bearing RC walls
(without a frame) with an eccentric door opening by Mays et al. (1998) (Fig. 13b).
The door opening from Mays et al. (1998) (Fig. 13b) somewhat resembles the double-
gapped yield line (Fig. 12c). This can be explained by the fact that the first crack sepa-
rates the right and left side of the wall (Fig. 13b), thus it changed to a new boundary
condition. New boundary conditions are similar to those of the double-gapped condition
(Fig. 12c).
Mays et al. (1998) also tested a specimen with both a door and a window opening,
with similar results compared to door specimen by Wang (2017) (Fig. 13). Firstly, the
crack separated the left and right side, as an eccentric door was used (Fig. 13b). The
wall section between the window and the door acted as a double-gapped wall and cracks
with a straight line in the middle. Other parts developed diagonal cracks. From a differ-
ent point of view, the door firstly separated the left and right side, after which a single-
gapped arching action occurred on the right side (perpendicular view).
Load-bearing URM walls (without a frame) with centric and eccentric window open-
ings were tested by Griffith et al. (2007), Griffith and Vaculik (2007). In those studies,
the rigid arching action follows more or less the yielding line, as would be the case
without the openings. This is shown in Fig. 16, where, on the left side the theoretical
assumption by Mays et al. (1998) is presented and, on the right, the tested URM walls
by Griffith et al. (2007) are given. It can be concluded that the prediction correlates
Predicted Observed Predicted Observed
Boundary
Fig. 16 The Mays et al. (1998) prediction of failures versus Griffith et al. (2007) experimental outcome of
load-bearing walls
13
highly with real crack patterns. However, openings, especially their corners, did influ-
ence the position and length of the yield lines. Nonetheless, the arching action is evi-
dent. The specimen with an aspect ratio of 1:1 had more horizontal than diagonal cracks
(Figs. 14, 15, 16).
Note that the original Fig. 12a has involved pressure w (kPa) versus displacement;
however, for comparison reasons, they were recalculated with forced F by the following
Eq. (1).
(1)
( )
F = w Ainfill − Aopening .
2.3 Effects of slenderness and the aspect ratio
Intuitively, due to compressive arching action phenomena, slenderness and the aspect ratio
play a major role in OoP behaviour (Fig. 17). As stated in Sect. 2.1, slenderness may limit
or bypass the arching action (Moghaddam and Goudarzi 2010; Dawe and Seah 1989; Shap-
iro et al. 1994; Furtado et al. 2018a).
Figure 17 presents results of different authors on the effects of slenderness on OoP load-
bearing capacity. From the figure, it is clear that there is a high reduction. In Fig. 17 curves
denoted with * were derived from the results published in Angel et al. (1994), where the
change in load-bearing capacity was altered by different degrees of slenderness. Curves
were grouped by the compressive strength of similar masonry fm , IP inter-storey drift dr ,
and the same masonry unit and mortar type.
Flanagan and Bennett (1999b) found that an increase in wall thickness from 100 to
200 mm ( h∕t ∶ 22 → 11 ) resulted in an ultimate force that was three times greater. In other
words, the doubling of infill thickness tripled its OoP capacity.
By observing changes in the aspect ratio in Fig. 16, we can see an obvious change in
yield patterns. If the aspect ratio had greater values, the horizontal line would develop at
mid-points of the diagonal lines. However, when the infill is shaped like a square or has
a lower aspect ratio, no horizontal line is developed, i.e., the same as in plate theory. By
having larger height than length, a vertical four hinged, a three hinged horizontal arching
action can be expected to occur.
35
Special measures required by EC8
30
MPa, N mort.*
25
(kPa)
MPa, L mort.*
20 MPa, N mort.*
Dawe and Seah (1989)
Pressure
15 Flanagan & Bennett (1999b)
10
5 *derived from Angel et al. (1994),

also include similar IP drift
0
10 12 14 16 18 20 22 24 26 28 30 32 34
Slenderness
Fig. 17 Effects of slenderness
13
Furthermore, Komaraneni et al. (2011) found that specimens with slender walls expe-
rienced higher amplification of acceleration (inertial forces) with the maximum observed
at the mid-height. Yet, a less slender specimen along the height acquired an almost linear
profile of acceleration response, albeit with the maximum value near the top. Komaraneni
et al. (2011) had a combination of inertial and dynamical load tests, i.e. where the frame
was fixed by OoP translation and additional masses added to the URM panel.
2.4 Effect of combined in‑plane out‑of‑plane loading
Nearly half of studies included a previous IP test (Fig. 3c). Most researchers focused on
prior IP damage. The first of such studies with infilled frames were conducted by Angel
et al. (1994), where it was concluded that previous IP damage could reduce the OoP load-
bearing capacity of slender panels by as much as one half. This was included in their
equation, using the R1 factor (Eq. 5, Table 6). A force versus displacements graphs from
two different studies containing various IP drifts were plotted on Fig. 18. From there, it is
obvious that the previous damage does affect the OoP behaviour. There is a reduction in
load-bearing capacities, initial stiffness and post peak behaviour. However, the deforma-
tion capability was unaffected, this can also be observed in the studies by Akhoundi et al.
(2018), Wang (2017), Angel et al. (1994). Note that Hak et al. (2014) had nearly the same
capacity with both 1.0 and 1.5% IP drift. Nevertheless, there is an extensive loss in the ini-
tial stiffness.
Furthermore, Fig. 19 was set to in order to comprehend the influence of previous IP
damages on OoP load bearing capacities. The figure includes results from various authors.
Note that Hak et al. (2014) did not have a control (only OoP load) specimen. Rather, they
assumed a 200 kN capacity that was used in plotting the curve. Correspondingly, Akhoundi
et al. (2018) combined load specimens had a double infill wall that were compared to the
single leafed wall. Results from Hak et al. (2014) showed the lowest impact, while those of
Ricci et al. (2018) the greatest. The results from Akhoundi et al. (2018) were somewhere
in between the two. Unquestionably, there are various factors that contribute to both IP and
OoP performance, resulting in differences between the specific behaviours. However, it is
clear that previous IP loading has a considerable impact on the OoP performance and it can
reduce its capacity up to 90% (Ricci et al. 2018). In turn, we can conclude that previous
IP damage decreases initial stiffness and ultimate force. Deformation capabilities are not
greatly affected.
Henderson et al. (1993) and Flanagan and Bennett (1999b) tested previous OoP inter-
storey drift damage on IP load-bearing capacity (OoP + IP). Both concluded that prior
inter-storey drift OoP damage had minimal effect on overall IP performance, especially
with regard to its capacity. This phenomenon is certainly due to the fact that, with an inter-
storey drift method, more damage occurs to the frame than the infill. Namely, due to pre-
vious cracking of the infill, the diagonal is weaker, i.e. softer; hence, initial stiffness is
also lower when compared to the control IP groups. The corner crushing failure was not
affected as observed by both research teams. Further, in the hysteric test by Henderson
et al. (1993), the OoP response was more linear than for the IP hysteric curve. This strongly
indicates that, in the case of the drift method, there is less energy absorption compared to
IP tests.
Additionally, Flanagan and Bennett (1999b) reported previous inertial OoP load
damage followed by IP loading, and simultaneous IP and inertial OoP (airbag) load-
ing. It was concluded that, with the inertial force method, IP behaviour was profoundly
13
Table 6 Analytical models developed for the estimation of infilled frame OoP capacity
2216
# Author Load-bearing capacity equation Action Remark
13
1 Dawe and fm0.75 t2 𝛼 Gapped
w1 = 0.8 (3) 𝛼= h
Ef Ic h2 + Gf Jc th
Seah (1989) l2.5 ( w1 );𝛼 ≤ 75
1( )0.25
𝛼 𝛽 Rigid
w2 = 0.8fm0.75 t2 + (4) 𝛽= l
Ef Ib l2 + Gf Jb tl
l2.5 h2.5 w2 ;𝛼 ≤ 50
( )
1( )0.25
2fm 𝜆
( )
2 Angel et al. w = R1 R2 (5) Rigid or Gapped h∕l 𝜆 R1 by dam-

h∕t
(1994) (depends on age state
R2) (Fig. 25)
Moderate Severe
5 0.129 0.997 0.994
10 0.060 0.946 0.894
15 0.034 0.888 0.789
20 0.021 0.829 0.688
25 0.013 0.776 0.602
30 0.008 0.735 0.540
35 0.005 0.716 0.512
Alternatively, R1 can be calculated
using Eq. 17
In case of a gap:
R2 = 0.357 + 2.488e−5Ef I (use
kN/m2)
If R2 > 1 ⇒ R2 = 1
In case of no gap (fully bounded
wall) R2 = 1
3 Klinger et al. Myv Myh xyv l Rigid Use equation if h∕l ≤ 2
W = 8 h (l − h) + 8 h ln (2) x ln l−h∕2 (6)
yh
( ) ( )
(1996)
tfm
xyv =
h
� �
2 (h∕2)2 +t2
1000Em 1− √
Myv = 0.85fm ∕4 ∕ t − xyv

𝛼 𝛽
( ) ( )2
4 Flanagan and w2 = 0.73fm0.75 t2 + (7) Rigid 𝛼 = 1h Ef Ic h2

Bennett l2.5 h2.5
( ) ( )0.25
(1999a)
𝛽 = 1l Ef Ib l2
( )0.25
If h∕t < 8 → t = h∕8

5 EC 6 BSI w = fm (t∕a)2 (8) Rigid Use if t∕a < 20 ; a is l or h
(2005)
6 Drysdale 0.61fj (t∕h)2 Rigid fj = min fm , fc,mortar
and Hamid w = min 2 (9)
{ } { }
0.61fj (t∕l)
(2005)
fm 𝜆 2 h∕t
7 FEMA 356 w = 0.7 h∕t
(10) Rigid 5 10 15 25
(2000)
𝜆2 0.129 0.060 0.034 0.013
0.85fm 0.45 fm 384Ef Ib h
8 Moghaddam
Em
wcr crushing 𝛼= Em tl4
� �
and 0.18Em failure

⎧ ⎫
wmax = 4
⎪ wcr = (h∕t)2 − 0.12 + 𝛼 ⎪
Goudarzi (0.12+0.045∕𝛼)(h∕4)
w = min⎨ ⎬ (11)
⎪ ⎪
(2010)
⎩ ⎭
Rigid
wmax transverse
instability
failure
9 Morandi et al. Rigid dw ≤ dm For values
w= 0.72 ht fm + 7.2 lht2 As fy 𝛽 (12)
(2013) r dm < dw < du of r ,
( )
( )2 ⎧ (r − 1)dw ∕dm + 1
0
⎪
dw > du dm and
𝛽=⎨
⎪
du see
⎩
Table 9
As area of vertical reinforcement in
tension, fy yield strength
2217
13
2218
t1.59
13
10 Ricci et al. w = 1.95fm0.35 h2.96 (13) Rigid
(2018)
Use MPa and mm. The output of equations is in MPa
*For calculation of xyh replace h with l in xyv equation. For calculation of Myh replace xyv with xyh ; dw expected IP drift
160 110
0.00 %
100 0.16 %
140 0.37 %
90
0.58 %
120 80
OoP force (kN)
OoP force (kN)

100 70
60
80
50
60
40
40 30
20 1.0 % 20
1.5 % 10
0 2.5 %
0
5 15 25 35 45 55 65 75 85 5 15 25 35 45 55 65 75
OoP deflection (mm) OoP displacement (mm)
(a) Study by Hak et al., 2014 (b) Study by Ricci et al. (2018)
Fig. 18 Effect of previous IP damage on OoP capacity
OoP bearing capacity reduction (%)
90 Hak et al. (2014) Angel et al. (1994)

80 Furtado et al. (2016) Akhoundi et al. (2018)
70 Ricci et al. (2018) Wang, Sepasdar et al. (2018)
60
50
40
30
20
10
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
IP drift (%)
Fig. 19 OoP bearing capacity reduction in relation to IP drift
affected, unlike in the case of the drift method. With combined IP + OoP (inertial) load-
ing, Flanagan and Bennett (1999a) found a 40% reduction in IP load-bearing capac-
ity compared with the IP control specimen. Furthermore, the simultaneous IP and OoP
load Flanagan (1994) resulted in heavy damage (Fig. 21), yet remained stable. The
unperceived stability was presumably the outcome of the arching action imposed by
the airbag. Misir et al. (2016) also tested frames under simultaneous IP and OoP load
with different infills; namely, frame with: Pumice Blocks (PWF); Hollow-Fired Clay
Heat Insulation Bricks (IWF); Autoclaved Aerated Concrete Blocks (AWF); Double-
Leaf Infill Walls without Z Ties (SWF); Double-Leaf Infill Wall with Z Ties (SWZF).
It was found that usage of Z-tie’s limited the stiffness degradation in OoP direction for
SWZ under increasing IP damage, and out-of-plane stability was preserved up to 2.75%
drift (Fig. 20a). A similar improvement was also observed in IP vs. OoP stiffness curves
shown in Fig. 20b. From the figure it is observable that out-of-plane stiffness deterio-
rates faster than in-plane stiffness under combined load action.
13
110 110
PWF
100 100
90 AWF
OoP stifness (kN/mm)

OoP stifness (kN/mm)
90
80 IWF 80
70 70
SWF
60 60
50 SWZF 50
40 40
30 30
20 20
10 10
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 10 20 30 40

IP drift ratio (%) IP stifness (kN/mm)
(a) IP drift ratio vs. OoP stiffness degradation curves (b) IP vs. OoP stiffness interaction curves
Fig. 20 Stiffness degradation due to simulations IP and OoP load (Misir et al. 2016)
200
150
100
IP force (kN)
50
0
-50
-100
-150
-200
-40 -30 -20 -10 0 10 20 30 40

IP displacement (mm)
Fig. 21 IP cyclic envelope derived from Flanagan and Bennett (1999b)
2.5 Contrasting outcomes of different experimental methods
Benedetti and Benzoni (1984) stated that there are two forces acting upon a structure dur-
ing ground motions: inertial and inter-storey drift. Further, Benedetti and Benzoni (1984)
stated that inter-storey drift should cause more damage than inertial forces (Benedetti and
Benzoni 1984).
Intuitively, dynamical shaking table tests are the best experimental approaches in the
field of seismic engineering. However, the control of the experiment itself, high costs and
the need for sophisticated equipment to carry out such experiments mean that they are
rarely used. Hence, many cyclic and static experiments are conducted instead (Fig. 21).
13
Most experiments have been carried out using inertial force methods (Table 2, Fig. 3a),
which were developed (Mc Dowell et al. 1956a, b) and continue to be used in the blast
engineering field (Smith et al. 2016; Pereira et al. 2014). In those experiments, frames were
fixed from OoP translations and the infill was loaded. This can be questionable in terms of
seismic engineering, as, during the ground motion, frames are excited not fixed.
In the shaking table tests by Fowler (1994), the infill had similar or greater accelerations
as the frame (Fig. 22b). Moreover, the highest accelerations occurred on the top and over
two-thirds of the wall height (Fig. 22a), similarly to the findings of Tu et al. (2010). On the
one hand, Fowler (1994) and Tu et al. (2010) found, in their dynamic test, and Flanagan
and Bennett (1999a), in their inter-storey drift tests, that the frame and infill move together
as a single unit. This is not the case with the inertial force test.
On the other hand, tests by Fowler (1994) showed that the largest displacements of the
panel occurred at the infill’s mid-point. In other words, the infill was affected more by iner-
tial forces. Furthermore, by examining damage states from dynamical (Fowler 1994; Tu
et al. 2010) and inter-storey drift methods (Flanagan and Bennett 1999b), it is clear that
the frame sustained more damage than the infill. Further, Flanagan and Bennett (1999b)
states that, in his inter-storey drift test, the frame failed, not the infill (Fig. 23). Henderson
et al. (1993) also obtained a more linear OoP response when compared to the IP response.
Hence, with OoP inter-storey drift forces, there is less energy absorption, i.e. the infill’s
effects are not so pronounced.
Inertial force methods are certainly efficacious in the case of load-bearing brickwork,
parapet walls and framed masonry with a single gap or multiple gaps between the frame
and infill. By disconnecting the frame from infill, they can act as separate elements. This
statement goes hand in hand with the claim from Wilton and Gabrielsen (1973) that, by
disconnecting the frame from the infill, one can act separately from the other. Thus, the
greatly pronounced inertial characteristics of the infill can then lead to the inertial collapse
of itself. In other words, infill with a gap would be analogous to the parapet wall. The gap
can be caused by workmanship or by frame—infill decoupling due to ground excitation.
This was somewhat observed in one out of four specimens by Tu et al. (2010). The A1
@ 9.58 sec, slab 2.0 PMD BS CS

@ 2.56 sec, top of panel
120
1.5
100
1.0
80
Wall height (cm)
Acceleration (g)
0.5
60 0.0
40 -0.5
20 -1.0
0 -1.5
-2.0
-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 2.4 2.6 2.8 3.0 3.2 3.4
Acceleration (g) Time (sec)
(a) Instatenious peak acceleration vs. wall height (b) Acceleration vs. time (PMD: panels’ mid-point, BS: base
of specimen, CS: center of slab)
Fig. 22 Structural acceleration results from the shaking table test by Fowler (1994)
13
Restrains
Load points
Damages on the steel frame

were not mapped; however,
authors stated that the frame
failed not the infill
Cantilever load system Beam load system
Fig. 23 Crack patterns of the inter-storey drift method (Flanagan 1994)
specimen lost a connection between the beam and the infill and, as a consequence, it fell
out due to inertia (Fig. 24a). In addition, failures due to inertial forces can be expected with
very slender infills (Moghaddam and Goudarzi 2010).
All three methods should be able to address the same phenomena; yet, there are major
differences between them. However, along with a paper which computationally investi-
gated inherent differences between the approaches (Anić et al. 2019), it was clear that there
are more similarities between the dynamical and inter-storey drift methods. The greatest
argument is that the frame is more damaged than the infill (Flanagan and Bennett 1999b)
(Compare Figs. 16, 23, 24b, 25), which is in direct contrast to the claim that the frame
does not play a role if it is purely loaded in the OoP direction (Furtado et al. 2018b). Put
in another way, bare frames cannot be affected if loaded in the OoP direction. Yet, in the
research by Tu et al. (2010), bare frames also withstood great damage due to dynamical
excitations. It has to be noted that the dynamical methods referenced here were single sto-
rey, hence, more prone to inter-storey drift forces. Whereas, inertial forces will more likely
to occur in upper storeys (Fig. 2). Various instances of damage caused by inter-storey drift
forces are also observable in real earthquake scenarios (Fig. 2), especially on walls that
are orthogonal to those that have suspected high IP damage (Fig. 2d) and also in dynamic
tests (Fig. 24). Inter-storey drift can also cause the beam to twist around its axis and thrust
the upper row of blocks (Penava and Sigmund 2017). It can also lead to falling out failure
(a) Failure due to inertial forces of A1 specimen (b) Inter-storey drift force failures of PF & A1 specimens
Fig. 24 Damage states of PF and A1 specimens by Tu et al. (2010)
13
(a) No damage (b) Moderate damage (c) Severe damage

1 0.997, 0.716 0.994, 0.512
Fig. 25 Damage states for R1 factor by Angel et al. (1994)
due to inertial forces as a lost connection consequence. Regarding inertial force methods,
if the frame is discounted from the overall behaviour, the problem can be related to the
load-bearing brickwork. As in the testing of such walls, frames can be included in order to
address walls’ boundary conditions.
2.6 Additional findings
Additional findings which are not discussed separately, but are nevertheless significant, are
listed below:
1. Mortar type does influence the OoP bearing when applying inertial force methods Sha-
piro et al. (1994). Further, Shapiro et al. (1994) noted that lateral strength is directly
proportional to the compressive strength of mortar;
2. There was no significant difference between the OoP behaviour of confined masonry
walls and unconfined masonry infill panels Tu et al. (2010);
3. Angel et al. (1994) found that gravity load affected OoP stiffness up to the point where
it was overcame by OoP forces. From that point, there was no observable differences
between the specimens with and without gravity load. Conversely, gravity load affected
OoP behaviour as it modified crack patterns and post-peak behaviour, as reported by
Furtado et al. (2018a). Note that both researches had gravity load introduces to columns;
4. In the study on the effects of bounding frames, Wang (2017) compared a structural
steel versus RC frame (that was tested by Sepasdar 2017) with identical infills. The
study showed that in the case of a stronger (steel) frame there was 40% less cracking
and the ultimate load dropped 48%. Furthermore, a 83% decrease in ultimate and crack-
ing stiffness was observed with a steel frame, and only 57% in the case of RC. Author
contributed this to the fact that:—in comparison to RC, steel frame had lower torsional
stiffness;—forces pressurised the flange that basically acted as a flexural steel plate, thus
increasing twisting;—RC’s beam was clamped to the floor and steel frame was clamped
on two steel beams that might have allowed additional deflections and rotations.
3 Analytical models
In this section, a study of analytical models is presented. The study included a comparison
of analytical and experimental approaches, as well as a single- and multi-variable paramet-
ric/sensitivity analysis of the analytical models. The aim of the parametric analysis was to
gain insight into stability, reliability, model limitations and governing factors. Furthermore,
13
the study was directed at arching action theory-based models. Hence, flexural action mod-
els, such as those in Drysdale and Essawy (1988), Haseltine (1976) and Hendry (1973),
were discarded. For more information about the flexural action models, please refer to Ast-
eris et al. (2017). Furthermore, there are additional equations not covered in this paper as
Hobbs et al. (1994) where they proposed a series of equations for the evaluation of walls
OoP resistance to impact force. Also from Hrynyk and Myers (2008) equation for the cal-
culation of OoP resistance of engineered retrofitted masonry wall within the blast engineer-
ing field.
At first, the calculation of OoP load-bearing capacities was based on flexural action
theory. However, the results were inadequate, as they provided capacities that were up to
six times lower than those observed in the experimental tests (McDowell et al. 1956b). As
specified in Sect. 2.1, the first arching action-based equation (Eq. 2) was developed by the
same research team that introduced the concept of arching action McDowell et al. (1956a).
However, it was developed for the calculation of the capacity of load-bearing URM walls
(no frame). Others subsequently followed a similar pattern of equation development.
All current equations developed for the OoP load-bearing estimation of infilled frames
are displayed in Table 6. These equations (Eqs. 5 and 6) were developed to interpret the
effects of inertial forces. Correspondingly, no equations have been developed for either
inter-storey drift or dynamical methods. Further, all equations have been developed on the
basis of vertical three-hinged arching action.
fm
w=𝛾 (2)
2(h∕t)2
where 𝛾 depends on the h∕t ratio (Tables 6, 7).

Dawe and Seah (1989) were the first to develop equations specifically for infilled
frames. They developed two equations: for gapped (Eq. 3) and two-way (Eq. 4) action.
The effects of the frame’s stiffness was introduced by the factors 𝛼 and 𝛽 . The parameter
α considered the effects of columns, while 𝛽 referred to beams. Within the gapped action
equation (Eq. 3), the parameter 𝛽 was excluded, due tthe gapped arching action. In other
words, the beam was too trivial in the gapped arching action. Furthermore, Dawe and Seah
(1989) were the only equation developers to include the effects of the frame’s torsional
characteristics.
Angel et al. (1994) were the first developers of reduction factors ( R1 in Eq. 5) that resolve
the degradation in OoP load-bearing capacity due to previous IP load. These reduction fac-
tors were based on damage states which can be determined either visually, as illustrated in
Fig. 25, or by using Eq. 17. The usage of the previous IP damage factor is thus limited to
Table 7 Analytical solutions for the calculation of displacement at peak load

# Author Displacement Remark
1 Klinger et al. (1996) xy = �

tfm l
� (14)
h
1000E 1− √
2 (h∕2)2 +t2
2 FEMA 273 (1997) d= √ 0.002h2 ∕t

(15) For h∕t ≤ 22.3
1+ 1−0.002(h∕t)2
3 Flanagan and Bennett (1999b) d= √0.002h

2 ∕t
(16) For h∕t ≤ 31.6
1+ 1−0.001(h∕t)2
13
the knowledge about the damage state or displacement data, i.e. it is harder to estimate the
damage state as it is. Factor R2 takes the boundary conditions into the account, whereas, if
gapped arching action occurs, the influence of the frame is introduced via R2 ≤ 1.
Conversely, if the infill is fully bounded by the frame, R2 = 1 , the formula for capacity
then resembles that for load-bearing URM walls without the frame. In other words, the
effects of the surrounding frame are thus omitted.
𝛿
R1 = 1if < 0.5i.e.undamaged state
2𝛿cr
else (17)
[ ( )2 ( )3 ] 2𝛿𝛿cr
h h h
R1 = 1.08 − 0.015 − 0.00049 0.000013
t t t
where δ is the maximum IP displacement and 𝛿cr is the displacement at the cracking load.
Klingner et al. (1996) developed an equation based on the Bashandy (1995) model for
load-bearing brickwork. Klingner et al. (1996) combined resistance from both vertical
and horizontal arching action, resulting in the indices yv and yh . Even though Klingner
et al. (1996) tested infilled frames (Table 2), he did not include the effects of frames in his
equation. From his load-bearing equation (Eq. 6), an equation for the calculation of OoP
displacement at peak can be derived (Eq. 14). Such an equation is different from others
(Eq. 15 and 16), as it includes both geometrical and mechanical properties of the infill. In
addition, by removing horizontal or vertical arching action from the Eq. 6, the resulting
Eqs., 18 and 19, could, in essence, be the first to address the capacity of a double-gapped
infilled frame.
Myv
Wyv = (l − h) (18)
h
Myh xyv
( ) ( )
l
Wyh = 8 ln (2) ln l (19)
h xyh l − h∕2
where Wyv can be used for infill with both columns disconnected, and Wyh for infill with
both beams disconnected.
Flanagan (1994) modified the two-way equation of Dawe and Seah (1989) (Eq. 4). The
modification consisted of changing the first constants, from 4.5 to 4.1, followed by remov-
ing torsional constants from parameters 𝛼 and 𝛽 . Additionally, Flanagan (1994) modi-
fied the equation of mid-height deflection at peak load (Eq. 16) from FEMA 273 (1997)
(Eq. 15), by changing the constant 0.002 to 0.001 under the square root.
FEMA 356 (2000) and FEMA 273 (1997) used a simplified version of the Angel et al.
(1994) equations, as published in Abrams et al. (1996) paper. The simplification was
sought by removing previous IP damage while considering the infill bounded on all sides
( R1 = R2 = 1 ). FEMA 273 (1997) also developed an equation for the calculation of mid-
height deflections (Eq. 15).
Eurocode 6 (2005) purposed a simple arching action equation for load—bearing walls,
that is only dependant on slenderness and masonry’s compressive strength.
Drysdale and Hamid (2005) proposed simple arching action equation for load-bearing
walls (Eq. 9). The equation itself is somewhat similar to that of Eurocode 6 (Eq. 8). How-
ever, the equation is unique in that it calls for the use of either the compressive strength of
13
masonry unit or mortar. They have also purposed an analytical model that can be used for
rigid, gapped and axial shortening wall units. However, the equation is not as straightfor-
ward to use as it calls for hard to obtain parameters as mid-span deflection, height of an
arch, length of the compression force. Therefore, those models were not covered within
this paper.
Moghaddam and Goudarzi (2010) as explained in Sect. 2.1, proposed two modes of
failure that were based on the infill’s slenderness. Thick infills might be crushed at the
boundaries wcr while slender infills may collapse due to large transverse deflections wmax .
Hence, wmax does not include masonry’s compressive strength. The effects of the frame
are taken into the account by the parameter 𝛼 in the form of stiffness (Table 8).
Morandi et al. (2013) were the first to propose an arching action equation that includes
the implementation of panel strengthening system via rebars in the bedjoint or the mesh in
the plaster. Additionally, the researches added an parameter 𝛽 that reduces the OoP capac-
ity due to previous IP damage. The coefficient 𝛽 is determined by the relation of expected
( dw ), ultimate ( 𝛿u ) and damage limitation ( dm ) IP drift together with OoP reduction fac-
tor ra . When the strengthening part of the equation is removed and without previous IP
damage, the equations look similar to those of Eurocode 6 BSI (2005) and Drysdale and
Hamid’s (Eqs. 8 and 9).
Ricci et al. (2018) developed their equation using semi-empirical approach by observ-
ing the values of capacity with five mechanical and geometrical characteristics. Among all
produced models, the one with a minimum number of parameters and the highest p value
was accepted and proposed in (Ricci et al. 2018). From the Eq. 13 it is obvious that authors
neglected the influence of the frame. Also, Ricci et al. (2018) developed an equation for the
load calculation that would cause the cracking of the infill. However, they are not covered
here as such calculations were not in the scope of this paper.
No equations have been specifically developed for infilled frames which can include the
effects of openings. However, Mays et al. (1998) developed an equation (Eq. 20) for load-
bearing RC walls. Yet, the equation was firstly used to calculate the OoP capacity w with
an arbitrary equation, before reducing it with the Fr factor and the Ao ∕Ai ratio. Hence, it is
reasonable to use it to reduce the load-bearing capacity w of infilled frames. Furthermore,
Mays et al. (1998) developed his equations for the calculation of blast, rather than seismic
bearing capacities. Therefore, there is a reference to blast-resistant carpentry in Table 9.
( )
Ao
wo = w + wFr (20)
Ai
where Fr . is selected from Table 9, w . is an arbitrary equation used for the calculation of
OoP capacity (Table 6), Ai is the infill area and Ao is the opening area.
Notes on analytical models. During the research on the analytical models, few discrep-
ancies were observed. Therefore, readers may find equations here inconsistent with the
would-be same equations in different articles. Mostly, the constants were changed because
of differences in the system of units (US or SI), or units themselves (using MPa not kPa).
For instance, in the original paper, Dawe and Seah (1989) used a constant of 800, while
others referring to the same equation (Flanagan and Bennett 1999a; Wang 2017; Pasca
et al. 2017) used a constant of 4.5. By recalculating the units, it was shown that a constant
of 800 should be used when the fm is in MPa and 4.5 when in kPa. In this paper, for the
sake of uniformity, the constant was changed from 800 to 0.8 (Eqs. 3, 4). Hence, the out-
put takes the same unit as the input parameters (MPa). Likewise, the original equation by
FEMA 356 (2000) contains a constant of 144; however, when fm is converted from psi to
13
Table 8 Values for previous IP reduction coefficient 𝛽 in Eq. 12 by Morandi et al. (2013)
Estimated inter-storey drifts in correspondence to limit state Limit state Unreinforced Lightly reinforced
Rebars in the bed- Mesh in the plaster
joint
Damage limitation dm 0.30% 0.35% 0.50%

Ultimate du 1.00% 1.00% 2.20%
Assumed fraction of OoP resistance ra corresponding to peak infill resistance 0.20 0.30 0.40
2227
13
Table 9 Modification factor Fr Panel type Blast resist- Opening location Fr

for panels with openings (Mays ant opening
et al. 1998)
One window No Central and offset − 1.00
One door No Central + 1.36
One door No Offset − 0.13
Two windows No Evenly distributed − 0.05
One window + one door No Evenly distributed − 0.41
One window Yes Central and offset − 3.07
One door Yes Central and offset − 2.73
Two windows Yes Evenly distributed − 2.62
One window + one door Yes Evenly distributed − 2.59
Table 10 Range of parameter Parameter Range

values which are considered for
the parametric analysis
Slenderness h∕t ∈ [5, 35]
Aspect ratio l∕t ∈ [0.5, 3.0]
Frame element section size bb ∈ [50, 500] mm
Masonry’s compressive strength fm ∈ [1, 35] MPa
Table 11 Values of 𝛽 for torsional constant (Ugural and Fenster 2003)

a∕b 1.000 1.500 2.000 2.500 3.000 4.000 5.000 6.000 10.000 ∞
𝛽 0.141 0.196 0.229 0.249 0.263 0.281 0.291 0.299 0.312 0.333
MPa, the constant changes to 0.99 ≈ 1.00, as is the case in the paper by Pasca et al. (2017).
The authors here modified all equations in order for the entire input to be in MPa and mm,
resulting in an MPa output.
Further, De Risi et al. (2019) wrote and referenced equation from Ricci et al. (2018),
however, when compared, they were different between the papers. Namely, the constant
0.35 in the De Risi et al. (2019) paper was raised to an exponent and in Ricci et al. (2018) it
was left as a multiplication constant. It was found that the constant should be raised and the
equation is as written in De Risi et al. (2019). However, authors here referenced the paper
by Ricci et al. (2018), as they were their developers.
3.1 Analytical model analysis input
The analytical model analysis was carried out on the mechanical and geometrical char-
acteristics of the specimens listed in Table 5. Parametric analysis was carried out on the
Sepasdar (2017) IF-ND specimen, as it showed the greatest correlation with all analytical
models.
When calculating torsional constants of rectangular sections, Eq. (21) along with
Table 11 was used. Contrariwise, in case of non-rectangular (steel) sections, the constant
was extracted from catalogues found on Tools for Engineer website (2019).
13
Furthermore, for the purpose of parametric, i.e. sensitivity, analysis a range of values
was established as presented in Table 10. That said, when parameters such as h∕t ∈ [5, 35]
could not be incorporated as a single variable, thickness was set to the original value of
t = 90 mm, while the value for height was obtained as a range with h = t ⋅ h∕t . Analo-
gously for the aspect ratio h∕l , h was set as a constant, such that slenderness was not influ-
enced. In the case of modifying masonry’s compressive strength, elastic modulus ( Em ) was
altered with the use of Eq. (22).
J = 𝛽ab (21)
where 𝛽 is obtained through linear interpolation by using Table 10.
Em = fm,k ⋅ KE (22)
where KE is 1000 in accordance with the Annex of the EN1996-1-1 provision (BSI 2004).
3.2 Analytical model analysis results
The results of the calculated data are presented in Tables 11, 12 and 13, and Figs. 26, 27,
28, 29, 30, 31.
Table 12 shows the overall absolute difference grouped by various properties and equa-
tions. Table 13 reveals a difference between experientially and analytically obtained load-
bearing capacities. Likewise, Table 14 highlights the differences in displacements. The
term plain masonry panel in both Tables 13 and 14 mean that the panel has no special
features like gaps, previous IP damage etc.
Figure 27 displays the differences in each specimen by equations, where the limits were
set to ± 100%. Hence, some models outside the range were omitted from the plot. Like-
wise, Fig. 28 presents the absolute differences in various properties in terms of an arbitrary
analytical model and the upper limit was set to 350%. Error variabilities are presented in
Fig. 26, showing the extremes, quartiles and medians.
The findings of the single-variable parametric analysis are shown in Fig. 29 with the left
side showing the results of two-way actions and the right side showing those of one-way
Table 12 Absolute difference between experimental and calculated capacities per property

Property Difference by Eq. # (%)
3&4 5 6&19 7 8 9 10 11 12 13
Total 99 236 98 75 78 276 203 46 822 30

RC frame 159 327 131 122 89 607 402 55 1358 33
SS frame 43 66 71 38 70 46 66 40 94 27
Plain masonry panel 73 78 80 80 67 273 235 47 82 28
Beam-infill gap 303 67
Columns-infill gap 277
Openings 56 66 65 62 63 180 129 43 68 47
Concrete units 40 330 58 44 58 80 89 44 1125 32
Clay units 187 79 165 133 99 618 403 618 216 26
Previous IP load 526 2466
13
Table 13 Differences between calculated and experimental OoP capacities (Eq. 23)
2230
Author Specimen Δ by Eq. # (%) Remark

3&4 5 6&19 7 8 9 10 11 12 13
13
Dawe and Seah (1989) WE1 81 33 26 18 45 78 72 − 11 80 46 Trust type joint reinf.
WE2 84 42 36 30 53 81 76 5 83 53 Plain masonry panel
WE3 93 76 74 71 81 92 90 61 93 81 Dry-stack panel
WE4 78 38 32 − 3 37 79 74 10 82 50 Plain masonry panel
WE5 32 − 5 − 15 102 − 95 65 57 − 27 69 16 Plain masonry panel
WE6 91 68 65 61 74 89 87 47 91 74 Plain masonry panel
WE7 87 56 51 46 64 85 82 27 87 64 Trust type joint reinf.
WE8 74 26 18 − 23 25 75 69 − 8 78 40 Restraints against slipping
WE9 88 58 54 49 66 86 83 31 88 67 Window opening
Angel et al. (1994) 1 − 115 − 71 − 88 − 71 − 514 9 − 14 5 18 − 280 Pure OoP, S mortar
2 55 198 IP (0.34% dr) + OoP
3 146 599 IP (0.22% dr) + OoP
4 442 3661 IP (0.09% dr) + OoP
5 526 6413 IP (0.06% dr) + OoP
6 1026 2712 IP (0.25% dr) + OoP
7 1065 2809 IP (0.25% dr) + OoP

8 3721 7240 IP (0.39% dr) + OoP
Flanagan and Bennett (1999a) 25 − 20 13 4 − 38 − 244 29 − 17 − 8 16 − 38 Plain masonry panel
18 48 35 29 17 − 49 47 13 − 9 37 − 3 Plain masonry panel
22 69 68 52 28 12 63 39 47 56 28 Plain masonry panel
19 − 21 514 IP (0.8% dr ) +OoP
Hak et al. (2004) TA1 − 58 IP (1.50% dr) + OoP
TA2 − 74 IP (2.00% dr) + OoP
TA3 − 59 IP (1.00% dr) + OoP
Akhoundi et al. (2015) SIF-A − 90 − 523 Beam − infill gap
Author Specimen Δ by Eq. # (%) Remark
3&4 5 6&19 7 8 9 10 11 12 13
SIF-B − 117 − 188 − 216 58 − 520 − 156 − 211 26 − 124 − 700 Plain masonry panel
PIF − A − 66 − 121 − 142 68 − 375 − 96 − 138 43 − 72 − 514 Window opening
Furtado et al. (2015) inf_01 − 239 − 221 − 252 93 − 868 − 269 − 378 − 19 − 244 − 1782 OoP monotonic + gravity load
inf_02 − 217 − 200 − 229 93 − 804 − 244 − 347 − 12 − 222 − 1658 OoP cyclic
inf_03 − 13 442 OoP cyclic + IP(0.5% dr),2 leaf.
Sepasdar (2017) IF-W 42 − 3 − 13 − 21 − 66 20 3 45 30 − 117 Window opening
IF-ND 28 − 28 − 42 − 52 − 107 0 − 21 32 13 − 172 Plain masonry panel
IF-D1 − 48 306 IP (0.66% dr) + OoP
IF-D2 − 69 IP (2.70% dr) + OoP
Wang (2017) IF-RC-DO 68 43 37 33 8 56 46 70 61 − 21 Door opening
IF-RC-TG 80 − 4 64 Beam—infill gap

IF-RC-SG − 3 Columns—infill gaps
IF-RC-ID − 56 IP (1.37% dr) + OoP
IF-S 63 46 41 38 − 7 48 37 65 55 − 41 Steel frame
Di Domenico et al. (2018) OOP_4E − 58 − 119 − 141 7 − 353 − 34 − 65 43 − 19 − 221 All bounded
OOP_3E − 31 − 381 Beam—infill gap

OOP_2E − 550 Columns—infill gaps
2231
13
Eq. 13
Eq. 12
Eq. 11 Max.
Eq. 10 3rd quartile
Eq. 9
Median
Eq. 8
Eq. 7 1st quartile
Eq. 6 & 19 Min.
Eq. 5 Absolute
mean
Eq. 3 & 4
-450 -400 -350 -300 -250 -200 -150 -100 -50 0 50 100
Difference between calculated and experimental capacity (%)
Fig. 26 Variation in differences plotted as quartiles between various analytical models
Fig. 27 Differences in various analytical models for each specimen (legend as in Fig. 28)
13
D Dawe and Seah (1998) Eq. 3 & 4

A Angel et al. (1994) Eq. 5
K Klinger et al. (1996) Eq. 6
F Flanagan and Bennett (1999a) Eq. 7
E EC 6 BSI (2005) Eq. 8
Y Drysdale and Hamid (2005) Eq. 9
X Fema 356 (2000) Eq. 10
M Moghaddam and Goudarzi (2010) Eq. 11
O Morandi et al. (2013) Eq. 12
R Ricci et al. (2018) Eq. 13
Fig. 28 Absolute differences in various analytical models by specific property
actions. Note that Fig. 29c has a lower limit set to − 0.2 and an upper limit set 0.4 MPa.
The multi-variable parametric analysis is displayed in Fig. 30.
All differences were calculated with the use of Eq. (23).
Y −X
Δ= (23)
Y
where X experimental and Y calculated data.
The Klingner et al. (1996) curve in Fig. 29a reveals the results in the h∕t < 30 domain
with regard to an aspect ratio of h∕l < 2 , while the natural logarithm ln (l∕(l − h∕2)) (Eq. 6)
produced complex numbers ℂ due to ln (ℝ− ) . Likewise, as a result of the negative values
under the square root, the Eqs. 15 and 16 produce complex numbers (Fig. 31). Therefore,
they are restricted to the values of h∕t > 31.6 and > 22.3 , respectively. All the limitations
are incorporated in Tables 6 and 7 in order for readers of this paper to have insight in
advance.
4 Discussion and overview
4.1 Experimental studies
There are three main experimental methodologies: (1) inertial forces; (2) inter-storey drift;
and (3) dynamical excitation. Dynamical methodologies are certainly the best option for
understanding the behaviour of framed masonry structures under seismic load. However,
they are more expensive, harder to control and require more sophisticated equipment (in
13
0.45 0.45
0.40 0.40
0.35 0.35
0.30 0.30
0.25 0.25
w (MPa)
w (MPa)
0.20 0.20
0.15 0.15
0.10 0.10
0.05 0.05
0.00 0.00
5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
(a) Effects of slenderness on two-way action (b) Effects of slenderness on one-way action
0.45
0.20
0.40
0.35 0.15
0.30 0.10
0.25
w (MPa)
w (MPa)
0.05
0.20
0.15 0.00
0.10 -0.05
0.05
-0.10
0.00
0.5 0.8 1.1 1.4 1.7 2.0 2.3 2.6 2.9 0.5 0.8 1.1 1.4 1.7 2.0 2.3 2.6 2.9
(c) Effects of aspect ratio on two-way action (d) Effects of aspect ratio on one-way action
0.55
0.35
0.50
0.45 0.30
0.40
0.25
0.35
0.30 0.20
w (MPa)
w (MPa)
0.25
0.20 0.15
0.15 0.10
0.10
0.05 0.05
0.00 0.00
-0.05
2 5 8 11 14 17 20 23 26 29 32 35 2 5 8 11 14 17 20 23 26 29 32 35
(e) Effects of compressive strength on two-way action (f) Effects of compressive strength on one-way action
Eq. 3 & 4 Eq. 5 Eq. 6 & 20 Eq. 7 Eq. 8 Eq. 13
Eq. 9 Eq. 10 Eq. 11 Eq. 11 Eq. 12
Fig. 29 Single-variable parameter sensitivity analysis of analytical models
13
Fig. 30 Multi-variable parameter sensitivity analysis of analytical models
13
85
75 Flanagan and Bennett (1999b), Eq. 16

FEMA 273 (1997), Eq. 15
65 Klinger et al. (1996), Eq. 14
55
45
35
25
15
5 8 11 14 17 20 23 26 29 32 35
Fig. 31 Effects of slenderness ( h∕t ) on OoP displacement calculations (Table 7)
the form of a shaking table). Thus, simpler methods are used; namely, inertial and inter-
storey drift methods. Both involve the specific forces that can act upon a structure in an
earthquake scenario.
In Tables 2, 3, 4 and Figs. 3, 4 and 5, it can be observed that most experiments were
carried out using the inertial-cyclic load method, with approximately half involving a pre-
vious cyclic IP load. Most frames were made from RC, while infills were made using clay
masonry, laid out in a manner so the voids were facing in the vertical direction. The slen-
derness of the infill was in the range of h∕t ∈ [1.35, 39.68] with the majority of them below
the Eurocode 8 (BSI 2005) limit for special measures (11.20 median). The aspect ratio
was in the range of l∕h ∈ [0.58, 2.00] , with the median of 1.36. One tested the opening
in the area of negligence, while the rest were in the upper area. From Table 5, it can be
observed that masonry’s compressive strength was in the range of fm ∈ [0.53, 24.30] MPa
with a median of 9 MPa. Very low instances of fm were mainly reported in studies from
Portugal, where, traditionally, blocks have low strength which can be attributed to the fact
that they are laid out horizontally. Likewise, masonry’s elastic modulus was in the range of
Em ∈ [942, 17575] MPa with a median of 5300 MPa. The previous IP inter-storey drift was
applied in the range of dr ∈ [0.06, 2.70] % with a median of 0.39%.
On investigating different kinds of loadings, it was found that there are considerable
differences between the inertial and inter-storey drift methods. Namely, the inertial force
methods damaged the infill, while leaving the frame intact; conversely, the inter-storey drift
methods damaged the frame and slightly damaged the infill. Furthermore, there are more
matches between the inter-storey drift and dynamical methods. For instance, it was found
that the frame and the infill move as a single unit, although the frame is significantly more
damaged than the infill. It was also found that infill had a greater acceleration and/or the
same displacements when compared to the frame. Hence, infill walls are more susceptible
to inertial forces. Along with the infill walls’ proneness to inertial forces and the fact that,
with dynamical excitation, the frame and the infill act as one, it is to be expected that fail-
ure due to inertial forces occurs when there is a loss in the boundary conditions. For exam-
ple, due to IP loads, it is common for a connection between the beam and the infill wall to
be lost. Hence, when imposing OoP loads, the frame and infill will not work as a single
13
Table 14 Difference between analytical models and experimental OoP displacement (Eq. 23)
Author Specimen Δ (%) Remark
Equation (14) Equation (15) Equation (16)
Flanagan and Bennett (1999a) 25 153.67 ∈ℂ 158.81 Plain masonry panel

18 170.79 140.86 131.66 Plain masonry panel
22 − 57.32 − 19.31 − 20.29 Plain masonry panel

Akhoundi et al. (2016) SIF-B 147.86 131.83 115.14 Plain masonry panel
Furtado et al. (2016) inf_01 91.84 85.55 71.03 OoP monotonic + gravity load
inf_02 251.70 240.18 213.55 OoP cyclic
Wang (2017) IF-S − 42.46 − 24.56 − 27.10 Steel frame
Sepasdar (2017) IF-ND − 42.46 − 8.86 − 27.10 Plain masonry panel

Di Domenico et al. (2018) OOP_4E 986.99 ∈ℂ 817.16 All bounded
2237
13
unit. As a consequence, the infill wall will fall out, resulting in a failure due to inertial
forces. Likewise, the same can occur as a consequence of poor workmanship, given that it
is challenging for workers to fill the void between the beam and the infill wall. Therefore,
the role of the frame, even with the fallen infill, should be addressed. In a similar manner,
few studies had gaps between columns and infill (Dawe and Seah 1989; Wang 2017; Di
Domenico et al. 2018). Such boundary conditions are sometimes referred to as practical
considerations, since the infill and the frame behave more or less independently of each
other.
With pure OoP inertial force methods, an analogy with plate structures was determined.
In other words, it had similar yield patterns in the case of boundary condition and resisted
the transverse through membrane forces, as the infill wall went through an arching action.
As stated in Hobbs et al. (1994), the term yield line is somewhat inappropriate as during
the arching action tensile strength is neglected. Hence, a more appropriate term would be
hinge line as the parts of infill rotate about that line (Fig. 8).
The main discovery of the field was certainly the development of arching action. Arch-
ing action is what generates membrane forces according to the plate analogy. Namely,
when the infill is loaded, it bends. After reaching a critical point, the infill cracks. Thus, the
infill splits into two separate parts which clamp themselves on contacting surfaces while
being open on the opposing surfaces. These three clamping points (mid-height and sup-
ports) form a three-hinged arch, named after the same resisting mechanism of an arch with
three hinges. The arching action resists transverse forces with additional compressive, nor-
mal forces. Through arching action, infill can withstand capacities of up to six times those
predicted by flexural theory. Besides the three-hinged arching action, two- and four-hinged
arching actions were discovered. The two-hinged arching action was found in the dynami-
cal test and is a probable action of the inter-storey drift methods. Certain factors can bypass
or limit the arching action, such as: boundary conditions; slenderness; aspect ratio; gravity
loads; infill type; mortar type; previous actions; and openings.
Boundary conditions, sometimes referred to as the effects of workmanship or practical
considerations, have shown a significant impact on OoP behaviour. Also, boundary condi-
tions are changeable trough time, as during an earthquake, the first feat that occurs is infill
debonding from the frame. Further on, with every interface loss, there is a loss in bearing
and deformation capabilities. This also affects the yielding patterns (hinge lines) and fail-
ure modes. On the one hand, with all sides connected to the frame, two-way or rigid (verti-
cal and horizontal) arching action is achieved. On the other hand, with one (beam-infill)
or two (columns-infill) interfaces missing, a one-way or gapped (vertical) arching action
is developed. The authors here would recommend not using the term one-way arching as
few authors did for single-gapped systems, as this could lead the conclusion that it does not
contain either vertical or horizontal arching. However, there is a clear difference in hinge
Table 15 Terminology as used in the literature

Description Infill connected by all Infill has a gap in connection Infill has two gaps opposite to
sides to the frame to one element (usually beam) each other (usually columns)
Figure 8a Figure 8b Figure 8c
Terminology Fully connected/bounded Partly connected/bounded

Two-way arching One-way arching
Rigid arching action Single gap or Double-gap arching action
Gapped arching action
13
development and behaviour between rigid and gapped systems. Hence, a more suitable ter-
minology should be used, as derived from the literature in Table 15.
Seven studies were carried out which observed the effects of openings. These stud-
ies included: one door, four window and two full-height openings, placed centrally. All
showed a drastic reduction in deformational capabilities. However, with regard to bearing
capacities, three out of five studies showed a decrease and two showed no change in bear-
ing capacity. The reduction was observed on the door, window and full-height openings,
while two window openings showed no decrease in ultimate force. The reason for such a
discrepancy is yet to be found. However, one cause could be the lintel, as computational
studies have shown that the existence of lintel adds an additional point to the compression
arch. Furthermore, crack patterns have revealed that, even with openings, arching action
was able to develop. However, its effectiveness was limited.
Both the infill’s slenderness and the aspect ratio have shown effects on OoP behaviour.
As the arching action is a function of compressive forces, it is intuitive that slenderness
and the aspect ratio could lower or elevate the infill’s bearing and deformation capacities.
It has also been found that, depending on slenderness, the infill may fail due to transverse
instability or by crushing at boundary. In other words, with thick infills, the arching action
could be bypassed.
4.2 Analytical models
There were ten analytical models developed for the calculation of infilled frames’ load-
bearing capacities (Table 6). All of them were developed on the basis of experiments that
mimic the inertial forces. However, the Eq. 7 by Flanagan and Bennett (1999a) was a
modification of the pre-existing Eq. 4 by Dawe and Seah (1989), while FEMA 356 (2000)
uses a simplified equation from Shapiro et al. (1994). Hence, eight analytical models were
originally developed. Each model has a specific effect accounted for, such as: previous IP
damage, one- and/or two-way action, torsion of frame parts, and different modes of failure.
Of those six equations, five of them include the effects of the surrounding frame, mainly
through stiffness. Note that there are also flexural action-based analytical models. How-
ever, they were not considered in this paper.
Furthermore, there are three analytical models developed for the calculation of OoP
deflection at peak load (Table 7). The Eq. (16) by Flanagan and Bennett (1999b) is a modi-
fied Eq. (15) from FEMA 273 (1997). Additionally, Eq. (14) from Klinger et al. (1996) was
extracted from the load-bearing Eq. (6) from the same authors.
It was found that deflection and load-bearing equations differ greatly, both among
themselves and with the experimental data. When compared to other specific properties,
equations showed relatively good correlations with concrete units, and with plain masonry
systems (Table 12, Fig. 28), The greatest overall performance of the load-bearing equa-
tions was that of Ricci et al. (2018) (Eq. 13). The equation was developed using semi-
empirical approach. Hence, with the amount of data accumulated by the time of equation
development, it had the best fitting. Also, equation had the smallest amount of error vari-
ation (Fig. 26), hence it also the most stable equation. Further, Moghaddam and Goudarzi
(2010) Eq. (11) in which two modes of failure were addressed: namely, transverse instabil-
ity and crushing at the boundaries had also good correlations with the experiments. In most
cases, the transverse instability failure was the prevailing one. Furthermore, the authors of
this paper extracted a pure double gapped one-way action equation from the Klingner et al.
(1996) Eq. 6. The equation was used to calculate load-bearing capacities of specimens with
13
the columns-infill gap. The capacities had a 276.83% difference compared to the experi-
mental ones. However, in the case of Wang (2017) IF-RC-SG specimen the difference was
3.27% (Table 13). Therefore, the equation has further development potential.
The previous IP damage reduction factor R1 from Angel et al. (1994) Eq. (5) did not
perform well with other specimens. This can be addressed to the equation itself, and to
the fact that the reduction due to previous IP damage varies considerably among other
researches (Figs. 17, 19). Furthermore, similar IP damage reduction factor 𝛽 from Morandi
et al. (2013) (Eq. 12) had a unsatisfactory performance, moreover, because of the ultimate
IP inter-storey drift 𝛿u = 1% some specimens had no OoP capacities. Hence, their calcu-
lated resistance could not be evaluated. Such performance can also be addressed to the
fact that IP damage reduction varies considerably among various researches (Figs. 17, 19)
and the fact that the equations purpose was for the design work, therefore, it has additional
redundancy.
The parametric, i.e. sensitivity, analysis revealed few stability issues: in a few instances,
the load-bearing Eq. (6) and displacement Eqs. (15) and (16) had negative numbers under
the square root, resulting in complex number outputs. Hence, the authors of this paper have
incorporated the newly found limits within the equation remarks (Tables 6 and 7).
Furthermore, parametric analysis revealed that slenderness, the aspect ratio and com-
pressive strength have a significant influence on the equation outputs. However, slender-
ness and compressive strength did play a greater role in all equations when compared with
the aspect ratio. Compressive strength had a more uniformly linear response, in contrast
with the effects of slenderness and the aspect ratio. Slenderness resulted in greater differ-
ences in the case of thick infill walls, whereas there were fewer in the case of slender ones.
All equations were based on arching action; the best correlations were in the domain
where slenderness was greater. Furthermore, the effects of the frame were addressed with
reference to the column’s and beam’s cross-sectional size (Fig. 30a). The effects were mini-
mal; hence, the omission of torsional effects was justified in the modification of the Eq. (4)
by Flanagan and Bennett (1999b) in the Eq. (7). Similarly, the effects of the frame in the
case of one-way action on the Eq. (5) by Angel et al. (1994) are minimal (Fig. 30b), while
they are non-existent in the case of two-way action. Overall, the frame had a substantially
low impact on the equation’s output, as is the case with inertial force experimental meth-
ods. Hence, the absence of the frame’s influence in the Eq. (6) by Klingner et al. (1996) is
upheld. Furthermore, previous IP damage, along with slenderness, affects the outcome of
the Eq. (5) greatly. The same is true in the experimental reports.
5 Conclusions
High-rise, multi-storey buildings are frequently composed of load-bearing frames and non-
bearing infill walls. During an earthquake, inertial and inter-storey drift forces act upon
such structures, forcing an interaction between the two, thus rendering the behaviour of the
frame and, with it, the whole structure. In real earthquake scenarios, those forces act both
in the In-Plane and Out-of-Plane direction. Therefore, for a better understanding of such
behaviour, the field of research has been divided into three principal subfields: In-Plane,
Out-of-Plane and bidirectional behaviour. Bidirectional behaviour regards the effects of:
previous In-Plane damage on Out-of-Plane behaviour; previous Out-of-Plane damage on
In-Plane behaviour; and simultaneous In-Plane and Out-of-Plane action.
13
This paper has revised the literature with regard to pure Out-of-Plane and bidirec-
tional seismic loading. It has also scoped the experimental and analytical models. The
paper’s focal point was set on the field of seismic engineering and infilled frames, while
also taking the effects of openings into account.
However, due to similarities, there are only limited references in studies on blast
engineering and loadbearing walls (without the frame). There are three basic Out-of-
Plane methods for testing infilled frames: dynamic excitation, inertial forces and inter-
storey drift methods. The inertial force method is the most commonly practised, mostly
with airbags. The inter-storey drift method is poorly researched, as are dynamical
methods.
It was found that, during Out-of-Plane inertial loading, a beneficial arching action
occurs. With it, additional compressive forces are generated, which provide supplementary
resistance to the transversal ones. The only mention of arching action in dynamical meth-
ods concerns the subordinate, two-hinged action. In the case of inter-storey drift methods,
there is no mention of arching action.
It was found that both inertial force and inter-storey drift methods differ greatly. Namely,
the inertial force methods damage the infill and the frame, albeit only slightly. The exact
opposite is true with the inter-storey drift method. In dynamical methods, it was found
that the frame and the infill move as a single unit, with the frame receiving more damage.
Furthermore, infill experiences higher accelerations than the frame. Hence, as soon as the
infill loses connection with the frame (i.e. due to previous In-Plane drift), they do not act as
a single unit. Thus, a failure due to inertial forces acting on the infill may occur.
Various parameters can limit or bypass the arching action, such as: slenderness; the
aspect ratio; boundary conditions; gravity load; masonry and mortar type; openings; and
frame stiffness.
Analytical models produce widely dispersed results. The slenderness and compres-
sive strength of the infill reveal the greatest influence on ultimate force. The equation for
the calculation regarding the effects of openings by Mays et al. (1998) was shown to be
inadequate. Probably due to the fact that the equation was developed for RC not masonry
walls. The best correlations and equation stability (error variation) with analytical and
experimental outputs were found with the Eq. (13) by Ricci et al. (2018). Also, Eq. (11) by
Moghaddam and Goudarzi (2010) also provided very good correlations, pointing out the
fact that equations are becoming more accurate as the field develops. Authors here derived
Eqs. (18) and (19) for double gapped systems. In the case of Wang (2017) IF-RC-SG spec-
imen, the correlation with the experiments were promising 3.21%. However, the overall
correlation with experimental data was not adequate (Fig. 28), however, it has a potential
for further development.
More research should be focused on the inter-storey drift method and/or a combination
of inter-storey drift and inertial force methods. Further, as no studies have been conducted
into the effects of the inter-storey drift method on RC frames, an experimental campaign to
address this should be considered.
More studies should look at the effects of simultaneous Out-of-Plane and In-Plane load-
ing, while the effects of previous Out-of-Plane damage on In-Plane behaviour should be
updated.
There is a need for a systematic study on the effects of openings, especially given that
none had been identified by dynamical and inter-storey drift methods. The effects of con-
fined openings should be addressed within all methods.
It is unclear as to what type (if any) of arching action is formed during dynamical and
inter-storey drift methods. Hence, work should be undertaken to clarify this issue.
13
More research should be carried out on the Out-of-Plane behaviour of multi-storey and/
or multi-bay structures.
There is also a need to develop a universal equation for inertial force methods which
will take most of the parameters that govern Out-of-Plane behaviour into consideration.
There should also be greater efforts to develop analytical models based on dynamical and
inter-storey drift methods and/or their combinations with the inclusion of openings.
Acknowledgements The research presented in this article is a part of the research project entitled FRAmed-
Masonry composites for Modelling and Standardization (FRAMA) (HRZZ-IP-11-2013-3013), which
is supported by Croatian Science Foundation. Additionally, this research was partially supported by joint
Croatian-German research project entitled Damage Prognosis for reinforced concrete Frame structures with
masonry infill walls under 3D Earthquake action (EFRAIM) approved by Ministry of Science and Educa-
tion of Republic of Croatia (MZO) and German Academic Exchange Service (DAAD).
References
Abrams DP, Angel R, Uzarski J (1996) Out-of-plane strength of unreinforced masonry infill panels. Earthq
Spectra 12(4):825–844
ACI (2011) Building code requirements for reinforced concrete. Am Concr Inst Struct J. doi:
30CFR250.901(d)(1)
Akhoundi F, Vasconcelos G, Lourenço PB, Palha CAOF, Martins A (2015) Out-of-plane behavior of
masonry infill walls. In: 7th International conference on seismology and earthquake engineering
Akhoundi F, Vasconcelos G, Lourenco PB, Silva L (2016) Out-of-plane response of masonry infilled rc
frames: effect of workmanship and opening. In: 16th International brick and block masonry confer-
ence “masonry in a world of challenges”. Taylor & Francis, pp 1147–1154
Akhoundi F, Vasconcelos G, Lourenço P (2018) Experimental out-of-plane behavior of brick masonry
infilled frames. Int J Archit Herit 18:1–7
Al Hanoun MH, Abrahamczyk L, Schwarz J (2019) Macromodeling of in-and out-of-plane behavior of
unreinforced masonry infill walls. Bull Earthq Eng 17(1):519–535
Anderson C, Bright N (1976) Behavior of non-loadbearing block walls under windloading. Concrete
10(9):27–30
Angel R, Abrams DP, Shapiro D, Uzarski J, Webster M (1994) Behavior of reinforced concrete frames with
masonry infills. Technical report. University of Illinois Engineering Experiment Station, College of
Engineering, University of Illinois at Urbana-Champaign
Anić F, Penava D, Guljaš I, Sarhosis V, Abrahamczyk L, Butenweg C (2018) The effect of openings on out-
of-plane capacity of masonry infilled reinforced concrete frames. In: 16th European conference on
earthquake engineering
Anić F, Penava D, Abrahamczyk L, Sarhosis V (2019) Computational evaluation of experimental method-
ologies of out-of-plane behavior of framed-walls with openings. Earthq Struct 16(3):265
Arde A, Furtado A, Melo J, Rodrigues H, Varum H, Pinto N (2017) Challenges and main features on quasi-
static cyclic out of-plane tests of full-scale infill masonry walls. In: Conference: 7th international con-
ference on advances in experimental structural engineering
Asteris P, Cavaleri L, Di Trapani F, Tsaris A (2017) Numerical modelling of out-of-plane response of
infilled frames: state of the art and future challenges for the equivalent strut macromodels. Eng Struct
132:110–122. https://doi.org/10.1016/j.engstruct.2016.10.012
Bashandy TW (1995) Behavior of reinforced concrete infilled frames under dynamic loading: part 1. Mas-
ter’s thesis, The University of Texas at Austin
Benedetti D, Benzoni G (1984) A numerical model for seismic analysis of masonry buildings: experimental
correlations. Earthq Eng Struct Dyn 12(6):817–831
Booth E, Key D (2006) Earthquake design practice for buildings, Thomas Telford, London. http://ebook
s.cambridge.org/ref/id/CBO9781107415324A009
BSI (2004) Eurocode 6: design of masonry structures-Part 1-1: common rules for reinforced and unrein-
forced masonry structures. ENV 1996-1-1: 2004: E, CEN Brussels
BSI (2005) Eurocode 8: Design of structures for earthquake resistance-part 1: general rules, seismic actions
and rules for buildings. British Standards Institution
13
CAEE (1984) Colombian standards for seismic resistant design and construction nsr-84. Technical report.
Colombian Association for Earthquake Engineering, Bogota, Colombia
Canadian Standards Association and Standards Council of Canada (1978) Masonry design and construction
for buildings. 1977, The Association CNIC (1988) Venezuelan seismic code, 1988. regulations for
earthquake resistant buildings. Comision De Normas Industriales, Covenin
Committee MSJ et al (1999) Building code requirements for masonry structures. American Concrete Insti-
tute, Detroit. ISBN 1929081022
da Porto F, Guidi G, Dalla Benetta M, Verlato N (2013) Combined in-plane/out-of-plane experimental
behaviour of reinforced and strengthened infill masonry walls. In: Proceedings of 12th Canadian
masonry symposium
Dawe J, Seah C (1989) Out-of-plane resistance of concrete masonry infilled panels. Can J Civil Eng
16(6):854–864
De Risi MT, Di Domenico M, Ricci P, Verderame GM, Manfredi G (2019) Experimental investigation
on the influence of the aspect ratio on the in-plane/out-of-plane interaction for masonry infills in rc
frames. Eng Struct 189:523–540
Di Domenico M, Ricci P, Verderame GM (2018) Experimental assessment of the influence of boundary
conditions on the out-of-plane response of unreinforced masonry infill walls. J Earthq Eng. https://
doi.org/10.1080/13632469.2018.1453411
Di Trapani F, Shing PB, Cavaleri L (2018) Macroelement model for in-plane and out-of-plane responses of
masonry infills in frame structures. J Struct Eng 144:04017198
Drysdale R, Essawy A (1988) Out-of-plane bending of concrete block walls. J Struct Eng 114:121–133
Drysdale RG, Hamid AA (2005) Masonry Structures; Behaviour and Design. Canada Masonry Design Cen-
tre, Ottawa
EERI (2018) Earthquake engineering research institute. https://www.eeri.org/. Accessed 30 Oct 2018
FEMA 273 (1997) Guidelines for the seismic rehabilitation of buildings. Technical report. Federal Emer-
gency Management Agency Washington, DC
FEMA 356 (2000) Commentary for the seismic rehabilitation of buildings. FEMA-356, Federal Emergency
Management Agency, Washington, DC
Flanagan R (1994) Behavior of structural clay tile infilled frames. Technical report. Center for Natural Phe-
nomena Engineering
Flanagan RD, Bennett RM (1999a) Arching of masonry infilled frames: comparison of analytical methods.
Pract Period Struct Des Constr 4(3):105–110
Flanagan RD, Bennett RM (1999b) Bidirectional behavior of structural clay tile infilled frames. J Struct Eng
125(3):236–244. https://doi.org/10.1061/(ASCE)0733-9445(1999)125:3(236)
Fowler JJ (1994) Analysis of dynamic testing performed on structural clay tile infilled frames. Ph.D. thesis,
Oak Ridge National Lab. (ORNL), Oak Ridge
Furtado A, Rodrigues H, Arde A, Varum H (2015) Experimental characterization of the in-plane and out-
of-plane behaviour of infill masonry walls. Proc Eng 114:862–869. https://doi.org/10.1016/j.proen
g.2015.08.041
Furtado A, Rodrigues H, Arde A, Varum H (2016) Simplified macro-model for infill masonry walls consid-
ering the out-of plane behaviour. Earthq Eng Struct Dyn 45(4):507–524
Furtado A, Rodrigues H, Arde A, Varum H (2018a) Effect of the panel width support and columns axial
load on the infill masonry walls out-of-plane behavior. J Earthq Eng. https://doi.org/10.1080/13632
469.2018.1453400
Furtado A, Rodrigues H, Arde A, Varum H (2018b) Out-of-plane behavior of masonry infilled rc frames
based on the experimental tests available: a systematic review. Constr Build Mater 168:831–848
Gabrielsen B, Wilton C, Kaplan K (1975) Response of arching walls and debris from interior walls caused
by blast loading. Technical report. URS Research Co, San Mateo
Griffith MC, Vaculik J (2007) Out-of-plane flexural strength of unreinforced clay brick masonry walls. TMS
J 25:53–68
Griffith MC, Vaculik J, Lam NTK, Wilson J, Lumantarna E (2007) Cyclic testing of unreingorced masonry
wall I two way bending. Earthq Eng Struct Dyn 36(6):801–821. https://doi.org/10.1002/eqe.654
Hak S, Morandi P, Magenes G (2014) Out-of-plane experimental response of strong masonry infills. In: 2nd
European conference on earthquake engineering and seismology
Hallquist A (1970) Lateral loads on masonry walls. Norwegian building research institute Reprint from els
Symposium on bearlng walls, Warsaw 1969
Haseltine BA (1976) Design of laterally loaded wall panels. In: Proceedings of the British ceramic society,
pp 115–126
Hashemi SA, Mosalam KM (2007) Seismic evaluation of reinforced concrete buildings including effects of
masonry infill walls. Technical report. Pacific Earthquake Engineering Research Center
13
Henderson R, Jones W, Burdette E, Porter M (1993) The effect of prior out-of-plane damage on the in-plane
behavior of unreinforced masonry infilled frames. In: The fourth DOE natural phenomena hazards
mitigation conference
Hendry AW (1973) Lateral strength of unreinforced brickwork. Struct Eng 51(2):43–50
Hobbs B, Ting M, Gilbert M (1994) An analytical approach for walls subjected to static and dynamic out-
of-plane point loads. In: Proceedings of the tenth international brick/block masonry conference. Uni-
versity of Calgary, Alberta, pp 329–358
Holmes M (1961) Steel frames with brickwork and concrete infilling. In: Proceedings of institution of civil
engineers. https://doi.org/10.1680/iicep.1961.11305
Howe GE (1936) Requirements for buildings to resist earthquakes. Reprinted by American Institute of Steel
Construction, Inc., New York, USA
Hrynyk TD, Myers JJ (2008) Out-of-plane behavior of URM arching walls with modern blast retrofits:
experimental results and analytical model. J Struct Eng 134(10):1589–1597
INEN (2001) Cdigo ecuatoriano de la construccin requisitos generales de diseo (inen-5). Instituto Ecuatori-
ano de Normalizacin
IS 1893 (2002) Is 1893: Indian standard criteria for earthquake resistant design of structures: part 1 general
provisions and buildings. Technical report. Indian Standard
Jäger W, Vassilev T, Hoffmann J, Schops P (2008) Unreinforced masonry basement walls—a comparison
of theoretical design approaches and numerical simulations. In: Proceedings of the 14th international
brick & block masonry conference. The University of Newcastle, Sydney, p 68
Kadysiewski S, Mosalam KM (2009) Modelling of unreinforced masonry infill walls considering in-plane
and out-of-plane interaction. PEER 2008/102. University of California, Berkeley
Klingner R, Rubiano N, Bashandy T, Sweeney S (1996) Evaluation and analytical verification of shaking
table data from infilled frames, part 2: out-of-plane behavior. In: Proceedings of the 7th North Ameri-
can masonry conference. Elsevier Science Ltd, Colorado, pp 521–532
Komaraneni S, Rai DC, Singhal V (2011) Seismic behavior of framed masonry panels with prior damage
when subjected to out-of-plane loading. Earthq Spectra 27(4):1077–1103
Lam N, Griffith M, Wilson J, Doherty K (2003) Timehistory analysis of urm walls in out-of-plane flexure.
Eng Struct 25(6):743–754. https://doi.org/10.1016/S0141-0296(02)00218-3
Liu M, Cheng Y, Liu X (2011) Shaking table test on out-of-plane stability of infill masonry wall. Trans
Tianjin Univ 17(2):125
Mays GC, Hetherington JG, Rose TA (1998) Resistance-deflection functions for concrete wall panels with
openings. J Struct Eng 124(5):579–587. https://doi.org/10.1061/(ASCE)0733-9445(1998)124:5(579)
McDowell E, McKee K, Sevin E (1956a) Arching action theory of masonry walls. J Struct Div 82(2):1–8
McDowell E, McKee K, Sevin E (1956b) Discussion of arching action theory of masonry walls. J Struct Div
82:27–40
MHPP (1995) Nepal national building code for seismic design of buildings in nepal (nbc-105). Ministry of
Housing and Physical Planning, Department of Buildings
Misir S, Ozcelik O, Girgin SC, Kahraman S (2012) Experimental work on seismic behavior of various types
of masonry infilled RC frames. Struct Eng Mech 44(6):763–774
Misir IS, Ozcelik O, Girgin SC, Yucel U (2016) The behavior of infill walls in rc frames under combined
bidirectional loading. J Earthq Eng 20(4):559–586. https://doi.org/10.1080/13632469.2015.1104748
Moghaddam H, Goudarzi N (2010) Transverse resistance of masonry infills. ACI Struct J 107(4):461
Monk CB (1958) Resistance of structural clay masonry to dynamic forces: a design manual for blast resist-
ance, vol 7. Structural Clay Products Research Foundation, London
Morandi P, Hak S, Magenes G (2013) Simplified out-of-plane resistance verification for slender clay
masonry infills in RC frames. In: ANIDIS 2013-XV Convegno di Ingegneria Sismica
Mosoarca M, Petrus C, Stoian V, Anastasiadis A (2016) Behaviour of masonry infills subjected to out of
plane seismic actions. Part 2: Experimental testing. In: 16th International brick and block masonry
conference
NCH (1996) Earthquake-resistant design of buildings nch 433. National Institute of Normalization
NSCP (1992) National structural code of Philippines, vol. 1, 4th edn (nscp). The Board of Civil Engineering
of the Professional Regulation Commission
NTC LG (2008) Norme tecniche per le costruzioni. Italian technical norms for constructions
NTS (1997) National technical standard e-030 earthquake resistant design. Ministry of Transportation,
Communications and Housing
NZS-3101 (1995) Code of practice for the design of concrete structures, part 1 (nzs-3101). Technical report.
Standards Association of New Zealand, Wellington, New Zealand
Pasca M, Liberatore L, Masiani R (2017) Reliability of analytical models for the prediction of out-of-plane
capacity of masonry infills. Struct Eng Mech 64(6):765–781
13
Penava D, Sigmund V (2017) Out-of-plane behaviour of framed-masonry walls with opening as a result
of shaking table tests. In: 16th World conference on earthquake engineering
Pereira MFP, Pereira M, Ferreira J, Lourenco PB (2011) Behavior of masonry infill panels in rc frames
subjected to in plane and out of plane loads. In: 7th International conference on analytical models
and new concepts in concrete and masonry structures
Pereira J, Campos J, Loureno P (2014) Experimental study on masonry infill walls under blast loading.
In: The 9th international masonry conference
Petrus C, Stoian V, Marius M, Anastasiadis A (2015) Reinforced concrete frames with masonry infills.
Out of plane experimental investigation. Acta Technica Napocensis 58(3):14–23
Preti M, Bettini N, Plizzari G (2012) Infill walls with sliding joints to limit infill-frame seismic inter-
action: large-scale experimental test. J Earthq Eng 16(1):125–141. https://doi.org/10.1080/13632
469.2011.579815
Preti M, Migliorati L, Giuriani E (2015) Experimental testing of engineered masonry infill walls for
post-earthquake structural damage control. Bull Earthq Eng 13(7):2029–2049. https://doi.
org/10.1007/s10518-014-9701-2
Quintas V (2003) Two main methods for yield line analysis of slabs. J Eng Mech 129:223–231
Rabinovitch O, Madah H (2011) Finite element modeling and shake-table testing of unidirectional
infill masonry walls under out-of-plane dynamic loads. Eng Struct 33(9):2683–2696. https://doi.
org/10.1016/j.engstr uct.2011.05.019
Reindl L, Butenweg C, Kubalski T (2011) Numerical simulation of unreinforced masonry walls subject
to dynamic out-of plane loading. In: Proceedings of the 3rd international conference on computa-
tional methods in structural dynamics and earthquake engineering, pp 25–28
Ricci P, Di Domenico M, Verderame GM (2018) Empirical-based out-of-plane URM infill wall model
accounting for the interaction with in-plane demand. Earthq Eng Struct Dyn 47(3):802–827
Sepasdar R (2017) Experimental investigation on the out-of-plane behaviour of concrete masonry
infilled rc frames. Master’s thesis, Dalhousie University
Shapiro D, Uzarski J, Webster M, Angel R, Abrams DP (1994) Estimating out-of-plane strength of
cracked masonry infills. Technical report. University of Illinois Engineering Experiment Station,
College of Engineering, University of Illinois at Urbana-Champaign
SI (1995) Design provisions for earthquake resistance of structure (si-413). The Standards Institution of
Israel
Sigmund V, Penava D (2014) Influence of openings, with and without confinement, on cyclic
response of infilled rc frames an experimental study. J Earthq Eng 18(1):113–146. https://doi.
org/10.1080/13632469.2013.817362
Smith NL, Tait MJ, El-Dakhakhni WW, Mekky WF (2016) Response analysis of reinforced concrete
block infill panels under blast. J Perform Constr Facil 30(6):04016059. https://doi.org/10.1061/
(ASCE)CF.1943-5509.0000825
SNiP (1995) Building code on construction in seismic areas (snip-ii-781). The Ministry for Construction
of Russia de Sousa H (2014) Typical masonry wall enclosures in portugal. Enclosure masonry
wall systems worldwide: Typical masonry wall enclosures in Belgium, Brazil, China, France, Ger-
many, Greece, India, Italy, Nordic Countries, Poland, Portugal, the Netherlands and USA p 179
Surendran S, Kaushik HB (2012) Masonry infill RC frames with openings: review of in-plane lateral
load behaviour and modeling approaches. Open Constr Build Technol J 6(Suppl. 1-M9):126–154
Tasnimi A, Mohebkhah A (2011) Investigation on the behavior of brick-infilled steel frames with open-
ings, experimental and analytical approaches. Eng Struct 33(3):968–980
Tools for Engineer Website (2019) Tools for engineer. http://www.toolsforengineer.com/. Accessed 13
Feb 2019
Tu YH, Chuang TH, Liu PM, Yang YS (2010) Out-of-plane shaking table tests on unreinforced
masonry panels in RC frames. Eng Struct 32(12):3925–3935. https://doi.org/10.1016/j.engst
ruct.2010.08.030
Ugural AC, Fenster SK (2003) Advanced strength and applied elasticity. Pearson Education, London
Vaculik J (2012) Unreinforced masonry walls subjected to out-of-plane seismic actions. Ph.D. thesis,
The University of Adelaide; School of Civil, Environmental & Mining Engineering
Vailati M, Monti G, Di Gangi G (2017) Earthquake-safe and energy-efficient infill panels for mod-
ern buildings. In: International conference on earthquake engineering and structural dynamics.
Springer, Cham, pp 233–261
Varela-Rivera J, Polanco-May M, Fernandez-Baqueiro L, Moreno EI (2012) Confined masonry walls sub-
jected to combined axial loads and out-of-plane uniform pressures. Can J Civil Eng 39:439–447
VBS (2006) Vetnamese seismic design standard tcxdvn 375:2006. Technical report. Vietnamese Build-
ing Standard, Hanoi, Vietnam
13
Verlato N, Guidi G, da Porto F, Modena C (2016) Innovative systems for masonry infill walls based on the
use of deformable joints: combined in-plane/out-of-plane tests. In: Proceedings of the 16th interna-
tional brick and block masonry conference
Wang C (2017) Experimental investigation on the out-of-plane behaviour of concrete masonry infilled
frames. Master’s thesis, Dalhousie University
Wilton C, Gabrielsen BL (1973) Shock tunnel tests of preloaded and arched wall panels. Technical report.
URS RESEARCH CO SAN MATEO CALIF
Yuen Y, Kuang J (2012) Effect of out-of-plane loading on lateral force transfer mechanisms of infilled RC
frames. In: 14th International conference on computing in civil and building engineering
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
13

A Review of Experimental and Analytical

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Review of Experimental and Analytical

Uploaded by

Copyright:

Available Formats

A review of experimental and analytical

studies on the out-of-plane behaviour of

Filip Anić, Davorin Penava, Lars

Bulletin of Earthquake Engineering

Bull Earthquake Eng (2020)

A review of experimental and analytical studies

Filip Anić1 · Davorin Penava1 · Lars Abrahamczyk2 · Vasilis Sarhosis3

Keywords Out-of-plane behaviour · Experimental studies · Analytical studies · Masonry

1 EN Eurocode Parts 8,6; BSI (2005)

8 Ecuadorian INEN-5 (2001) Based upon ACI

13 Nepal National Building Code; NBC-105 (1995) Largely following IS code

Fig. 1 Number of OoP—infilled frame specific articles as published by year

1 Dawe and Seah (1989) Monotonic Inertial—AB No – + + +

11 Pereira et al. (2011) Cyclic (q-static) Inertial—double leafed AB Cyclic – – + –

16 Furtado et al. (2015) Cyclic (q-static) Inertial—AB Cyclic + + – –

AB airbag, P point load, L line load

# Study by Frame Infill Subject studied

11 Pereira et al. (2011) 2.33 2.00 1:1.5 4 1 1

16 Furtado et al. (2015) 2.33, 1.35 1.29 1:1.0 3 1 1

# Study by Infill Openings Scale Specimen

2 Review of experimental campaigns

3 48 10.14 5212 5.99 n/a IP (0.22% dr) + OoP,

8 187 3.50 2358 32.08 n/a IP (0.39% dr) + OoP,

SIF-B 10.13 12.00 Plain masonry panel

IF-ND 66.30 12.50 Plain masonry panel

Clay, 77 Vertical, RC, 73

(a) Masonry type (b) Laying method (c) Frame type

Special measures required by EC8

10 1.0 First quartile

Fig. 5 Distribution of geometrical characteristics

2.1 Arching action and boundary conditions

Fig. 8 Types of arching action in relation to boundary conditions

Fig. 9 Force-displacement curves of various boundary conditions

Fig. 10 Force-displacement curves of various boundary conditions

3rd 4th 1st

(a) Door and window (b) Eccentric door

Fig. 14 Failure patterns of OoP specimens with openings

Predicted Observed Predicted Observed

2.3 Effects of slenderness and the aspect ratio

15 Flanagan & Bennett (1999b)

5 *derived from Angel et al. (1994),

2.4 Effect of combined in‑plane out‑of‑plane loading

# Author Load-bearing capacity equation Action Remark

2 Angel et al. w = R1 R2 (5) Rigid or Gapped h∕l 𝜆 R1 by dam-

Myv = 0.85fm ∕4 ∕ t − xyv

4 Flanagan and w2 = 0.73fm0.75 t2 + (7) Rigid 𝛼 = 1h Ef Ic h2

If h∕t < 8 → t = h∕8

and 0.18Em failure

# Author Load-bearing capacity equation Action Remark

OoP force (kN)

90 Hak et al. (2014) Angel et al. (1994)

Fig. 19 OoP bearing capacity reduction in relation to IP drift

OoP stifness (kN/mm)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 10 20 30 40

-40 -30 -20 -10 0 10 20 30 40

Fig. 21 IP cyclic envelope derived from Flanagan and Bennett (1999b)

2.5 Contrasting outcomes of different experimental methods

@ 9.58 sec, slab 2.0 PMD BS CS

Damages on the steel frame

Cantilever load system Beam load system

Fig. 23 Crack patterns of the inter-storey drift method (Flanagan 1994)

(a) No damage (b) Moderate damage (c) Severe damage

Fig. 25 Damage states for ­R1 factor by Angel et al. (1994)

where 𝛾 depends on the h∕t ratio (Tables 6, 7).

Fig. 1 Number of OoP—infilled frame specific articles as published by year

2 Review of experimental campaigns

Fig. 5 Distribution of geometrical characteristics

2.1 Arching action and boundary conditions

Fig. 8 Types of arching action in relation to boundary conditions

Fig. 9 Force-displacement curves of various boundary conditions

Fig. 10 Force-displacement curves of various boundary conditions

Fig. 14 Failure patterns of OoP specimens with openings

2.3 Effects of slenderness and the aspect ratio

2.4 Effect of combined in‑plane out‑of‑plane loading

Fig. 19 OoP bearing capacity reduction in relation to IP drift

Fig. 21 IP cyclic envelope derived from Flanagan and Bennett (1999b)

2.5 Contrasting outcomes of different experimental methods

Fig. 23 Crack patterns of the inter-storey drift method (Flanagan 1994)

Fig. 25 Damage states for R1 factor by Angel et al. (1994)

Table 7 Analytical solutions for the calculation of displacement at peak load

Table 9 Modification factor Fr Panel type Blast resist- Opening location Fr

Table 10 Range of parameter Parameter Range

Table 11 Values of 𝛽 for torsional constant (Ugural and Fenster 2003)

3.1 Analytical model analysis input

3.2 Analytical model analysis results

Table 12 Absolute difference between experimental and calculated capacities per property

Fig. 26 Variation in differences plotted as quartiles between various analytical models

Fig. 27 Differences in various analytical models for each specimen (legend as in Fig. 28)

Fig. 28 Absolute differences in various analytical models by specific property

Fig. 29 Single-variable parameter sensitivity analysis of analytical models

Fig. 30 Multi-variable parameter sensitivity analysis of analytical models

Fig. 31 Effects of slenderness ( h∕t ) on OoP displacement calculations (Table 7)

Table 15 Terminology as used in the literature