06-10-2016, 10:00 AM
Experimental assessment of the seismic performance of a prefabricated concrete
structural wall system
1457961824-Experimentalassessmentoftheseismicperformanceofaprefabricatedconcrete.pdf (Size: 4.41 MB / Downloads: 7)
a b s t r a c t
In this study the authors investigated experimentally the behaviour of prefabricated reinforced concrete
sandwich panels (RCSPs) under simulated seismic loading through a large experimental campaign. Tests
were carried out on single full-scale panels with or without openings, simulating the behaviour of lateral
resisting cantilever and fixed-end walls. Tests were also carried out on a 2-storey full-scale H-shaped
structure constructed by individual panels which were properly joined together. The performance and
failure mode of all panels tested revealed strong coupling between flexure and shear due to the squattype
geometry of the panels. However due to their well-detailed reinforcement, all panels exhibited only
a relatively gradual strength and stiffness degradation and in no case did any panel suffer from sudden
shear failure. The prefabricated walls of the structural system investigated herein seem to meet all the
requirements of Eurocode 8 for walls to be designed as ‘‘large lightly reinforced walls’’; however this
assumption should be supported with further experimental and analytical studies
Introduction and background
Most current design codes of concrete buildings for earthquake
resistance give lower values of the behaviour factor q for wall
systems than for frame systems. This is attributed to a couple of
reasons; shear walls have large cross-sectional length ℓw and thus
shear plays a superior role in their seismic response and also the
seismic behaviour of the wall systems is not so well known and
understood because experimental and analytical research on walls
is practically difficult; so design codes tend to be on the safe side. To
make matters worse, when the seismic performance of structures
incorporating precast or prefabricated concrete walls as the main
lateral force resisting system is considered, this conservativeness
may become higher for designers in the seismic prone countries.
This has advocated the bad behaviour of the poorly designed and
constructed precast connections which caused the use of precast
concrete structures to be regarded with suspicion by a part of the
structural engineer’s community.
In contrast, after the excellent seismic performance of buildings
with structural walls [2,3], in very strong earthquakes (e.g. in Chile,
1985 and Kocaeli, Turkey 1999), there is presently a tendency to
acknowledge similar q-factor values for frame and wall systems.
In addition, during the 1988 Armenia earthquake poorly designed
and constructed buildings that incorporated precast concrete walls
as the main lateral force resisting system performed substantially
better than buildings built with other structural systems [4].
Although the use of precast structural walls in seismic areas of the
world has proved to be a cost-effective way for lateral resistance
of buildings, the largest part of structures is based on cast in place
reinforced concrete. This could be partly attributed to the fact that
the majority of the research on wall systems has been focused on
the seismic behaviour of cast in place walls (e.g. [5–9]).
During the past decade or so, however, the research community
has increasingly focused on the use of precast concrete walls as
the primary lateral load resisting system in seismic regions (e.g.
[10–15]). The precast concrete structural wall systems, which have
been investigated up to date, are generally arranged to provide
lateral force resistance by cantilevering from the foundation
structure, through coupling with beams or other special devices
and by rocking about their foundation. Moreover structural wall
systems showing strong nonlinear response can be grouped into
either equivalent monolithic or jointed systems. An analytic review
of the precast structural wall systems can be found in a relatively
recent bulletin of fib [16].
In this paper the seismic performance of a prefabricated equivalent
monolithic structural system comprising large reinforced
concrete sandwich panels (RCSPs) is investigated. Despite the
fact that the use of prefabricated RCSPs (described in detail below)
has been introduced in the construction industry for more than 40 years [17], they have been used in practice primarily as
gravity load bearing structural elements (e.g. [18,19]). More recently,
in the last decade, many companies from the international
precast construction industry have started manufacturing RCSPs
commercially with the aim of developing a quick, permanent and
cost-effective building system which is supplemented with a satisfactory
earthquake resistance. According to the major companies
involved in panel construction, it is estimated that the construction
cost of a (low-rise) residential building with prefabricated RCSPs is
about 2/3 of an equivalent (framed, walled or mix-type) RC structure.
The seismic performance of this structural system, which can
be possibly qualified as ‘‘Large Lightly Reinforced Walls’’ according
to Eurocode 8 (as explained later), has not been addressed up
to date. The investigation of the seismic performance of prefabricated
RCSPs, which still remains a challenging task, is addressed in
this study for the first time in a systematic way through full-scale
seismic testing.
In the present study the authors investigate experimentally the
behaviour of prefabricated RCSPs under simulated seismic loading
through a large experimental campaign. Tests are carried out both
on single full-scale walls with or without openings, reproducing
the behaviour of lateral load resisting cantilever and fixed-end
walls, and on a 2-storey full-scale H-shaped structure constructed
by individual panels which were properly joined together. In the
latter case the seismic performance of a complete prefabricated
structure including the connections between RCSPs was addressed
Another innovative aspect is the experimental investigation of
a full-scale H-shaped structure consisting of more than one
rectangular part. Such types of structures have high stiffness and
strength in both horizontal directions, and despite the fact that
they appear to be more cost effective than the combination of their
constituent parts as individual rectangular walls, our knowledge of
their behaviour under cyclic bending and shear is very limited.
2. Description of the structural system
2.1. Reinforced concrete sandwich panel (RCSP)
A reinforced concrete sandwich panel (RCSP) is composed
of an Expanded Polystyrene (EPS) foam core with prefabricated
galvanised steel wire mesh reinforcement encased in two layers
of sprayed concrete on both sides, as shown in Fig. 1(a). The
steel wire mesh of reinforcement mounted on each face of the
polystyrene foam is drawn with hot galvanisation and consists
of 2.5 mm and 3.5 mm diameter horizontal and longitudinal
reinforcement, respectively, spaced at 65 mm; this gives a
longitudinal reinforcement ratio of 0.42%, which is more than
the minimum longitudinal reinforcement of 0.2% of the Eurocode.
The connection between the two concrete layers through the
core of the wall panel is secured with 3 mm diameter steel
connectors welded to the front and back wire meshes through
the polystyrene. These connectors (∼80/m2
) could be straight
or inclined depending on the manufacturing plan. The uniform
connection between the parts of the sandwich panel is also
favoured by the surfaces of the polystyrene which have been
initially corrugated. The panels considered in this study have depth
and length of corrugation equal to 10 mm and 70 mm, respectively
(Fig. 1(a)). In this way the assembly develops nearly full composite
behaviour in stiffness and shear transfer.
Each type of steel wire mesh used (horizontal, longitudinal,
connector ties) has a nominal yield stress of 600 MPa. The shotcrete
has typically a thickness of 35 mm (also greater values could
be considered) and a characteristic 28 days cube compressive
strength higher than 25 MPa. It should be noted that in order
to control shrinkage a fiber reinforced concrete could be used.
Connection between the panels and the foundation or floor is made
by means of starter steel bars projecting from the foundation (or
floor) as shown in Fig. 1(b). In the present study, the transfer of
the tensile forces from the panels to the foundation was made by
8 mm diameter deformed bars which were placed at distances of
300 mm. These bars had a yield stress of 550 MPa.
2.2. Lateral load resisting system
In a large panel system composed of prefabricated RCSPs, the
wall and slab panels are connected in the vertical and horizontal
directions so that the walls enclose appropriate spaces for the
rooms within a building. The height of the panels is equal to the
storey height while horizontal floor and roof panels span either as
one-way or two-way slabs. When properly joined together, these
horizontal elements act as diaphragms that transfer the lateral
loads to the walls. All the walls are continuous throughout the
building height. Joint system is developed such that all structural
elements work together as a box-type system.
Panel connections represent the key structural components
in this system. Based on their location within a building, these
connections can be classified into vertical and horizontal joints.
Vertical joints connect the vertical faces of adjoining wall panels
and primarily resist vertical seismic shear forces. Horizontal joints
connect the horizontal faces of the adjoining wall and floor
panels and resist both gravity and seismic loads. Vertical and
horizontal connections are accomplished by means of dowels (starter longitudinal bars), projected from the foundation or the
(panel) slab of the previous storey, and of hairpin hooks site welded
to the dowels as shown in Fig. 1(b). Both the horizontal and vertical
joints are grouted in situ using concrete (same mix as used in the
panel construction). Gravity loads are received both by vertical and
horizontal panels.
Typical characteristics of low-rise buildings constructed with
this construction system comprise 1–5 stories with 3 m height
and typical span of the walls 3–5 m. The wall thickness ranges
from 150 mm to 200 mm depending on the thickness of the EPS
foam, while the density of the sandwich panel may be varying
between 0.9 and 1.1 t/m3
. The fundamental period of vibration for
a characteristic 3-storey building of this type is low, and in general
does not exceed 0.2 s, if full fixity at the base is assumed. However it
should be kept in mind that the wall structure, which generally has
small natural periods (in the elastic range), may often lengthened
substantially if soil–structure interaction is properly accounted for.
3. Experimental program
3.1. Test specimens’ layout
The experimental program aimed to provide a fundamental
understanding on the seismic behaviour of prefabricated lightly
RC panels and provide data valuable for modelling such types of
sandwich panels. The test specimens were divided into two groups.
The first phase of the experimental campaign includes a total
of ten full-scale RC panel specimens with and without openings,
which were constructed and tested under cyclic uniaxial flexure
with constant axial load in single or double bending. The specimens
were constructed with stiff top and bottom beams. The top
beam served to distribute the horizontal and axial loads to the
wall while the bottom beam, clamped to the laboratory strong
floor, simulated a rigid foundation. The geometries of the ten
single storey structural panels are shown in Fig. 2 whereas
their reinforcing details were identical to the prefabricated RCSPs
described previously. These specimens were designed such that
the effect of a series of parameters on their seismic behaviour could
be investigated. These parameters comprised: panel length; testing
configuration; level of the axial load; and the presence or not of
openings in the panel. A summary of the experimental parameters
is presented in Table 1. In brief, the notation of panels is PLX_A_O,
where the letter P stands for panel (or wall), L defines the panel
length (3 or 4 m), X defines the testing configuration (S for single
and D for double bending), A denotes the level of the applied axial
load during test (150 or 300 kN) and O denotes the presence of
opening on the panel (W for window, D for door). Note here that
the need of assessing the performance of panels tested with both
single and double bending configurations was illustrated by some
preliminary FE analyses (simulating RCSPs), performed on several
residential buildings. These analyses have pointed out that the
overall behaviour of these buildings is in some cases dominated
by a box-type response, which resulted in very squat walls.
The second part of the experimental campaign comprises
the test of a 2-storey full-scale H-shaped structure which was
subjected to horizontal cyclic loading applied in the plane of the
web under constant vertical load. The H-shaped structure, which
was constructed by six horizontal wall panels (three in each floor)
and two 0.2 m thick RC slabs, is 3.50-by-2.75 m in plan; it has
two 2.75 m tall stories while it is fixed into a heavily reinforced foundation 4.6 × 3.25 m in plan with a height of 0.6 m, as shown
in Fig. 3. Additional masses of 5.62 t and 6.16 t were mounted on
the second and first floors, respectively, to reproduce the effect
of loads beyond self-weights. The axial load ratio N/Ac
fc resulting
for the panels of the ground floor is 2% while the corresponding
value for the top floor panels is 1%; these values of axial loads
are comparable to those applied in practice. The self-weight of the
structure is 30 t. Note here that the steel wire mesh in the panels
of the H-shaped structure (Fig. 1(b)) consists of 2.5 mm diameter
horizontal and longitudinal reinforcement, spaced at 50 mm; this
gives a longitudinal reinforcement ratio of 0.29%.