Engineering Structures 176 (2018) 175–187

Contents lists available at ScienceDirect

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

Response of thin lightly-reinforced concrete walls under cyclic loading T

a,⁎ b c d e
Carlos A. Blandon , Carlos A. Arteta , Ricardo L. Bonett , Julian Carrillo , Katrin Beyer ,
João P. Almeidae
a
Department of Civil Engineering, Universidad EIA, km 2+200 variante aeropuerto JMC, Envigado, Colombia
b
Department of Civil Engineering, Universidad del Norte, Barranquilla, Colombia
c
Department of Civil Engineering, Universidad de Medellín, Medellín, Colombia
d
Department of Civil Engineering, Universidad Militar Nueva Granada, Bogotá, Colombia
e
Earthquake Engineering and Structural Dynamics Laboratory (EESD), École Polytechnique Fédérale de Lausanne, Switzerland

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

Keywords: During the last two decades, thin concrete walls have been frequently used to brace mid- to high-rise buildings in
Thin wall some Latin American countries. This structural system differs significantly in terms of wall geometry and re-
Reinforced concrete inforcement layout from traditional cast-in-place reinforced concrete wall buildings. Limited experimental data
Cold-drawn reinforcement on this wall system and the absence of post-earthquake field observations make it difficult to assess whether such
Welded-wire mesh
walls behave similarly to the walls designed according to the current local design code. The paper presents and
Lightly-reinforced slender walls
discusses the results of an experimental program comprising quasi-static cyclic tests of four slender, thin and
lightly-reinforced concrete walls with different geometrical configurations, steel properties and reinforcement
layouts, which correspond to a common construction practice in Colombia. The seismic response of the speci-
mens was assessed in terms of crack propagation and failure modes, hysteretic and backbone curves, con-
tribution of rocking, flexural, shear and sliding components to lateral drift, stiffness degradation, and energy
dissipation capacity. The results suggest that the response of these reinforced concrete walls does not meet the
performance specified in the Colombian regulation if they are designed to reach the maximum lateral drift
allowed by the code.

1. Introduction earthquake with a return period of 475 years, without collapsing and
limiting the structural damage.
One alternative for industrialized and low-cost housing in Latin The walls in the structural system under discussion have several
America includes concrete wall buildings using slender and thin lightly- characteristics that introduce significant differences in terms of geo-
reinforced walls which are cast conforming the architectural layout of metry and reinforcement distribution when compared to the traditional
the residential units. This construction method uses steel or aluminum cast-in-place reinforced concrete (RC) wall buildings considered by the
modular formwork that can be assembled in different configurations. ACI 318 provisions. One of the main differences is the use of walls with
The main advantage of this method is the significant reduction of the significantly reduced thickness (tw) that can be as low as 70 mm with a
construction time as nonstructural divisions or facades are considerably typical range between 100 and 150 mm [3]. Such reduced thickness can
reduced or are not required. This type of buildings has been constructed be specified by designers, as the code does not have an explicit
in low, moderate and high seismicity regions following specifications minimum value for this parameter for reinforced concrete walls.
for reinforced concrete walls defined by the Colombian Code (NSR-10) Typically, these walls only have a single curtain of web reinforce-
for Earthquake-Resistant Construction [1]. Provisions for concrete ment which is spliced to started bars of 6.3 mm (#2) or 9.5 mm (#3)
structures in all versions of NSR have been based on a previous version diameter which extend from the foundation up to the second third of
of ACI 318. The current version of the provisions for concrete struc- the first floor height, ensuring the required lap splice length according
tures, which updated a previous version issued in 1998, are based on to the code. This single curtain of reinforcement usually consists of
the 2008 version of ACI 318 [2]. Reinforced concrete buildings de- meshes made of cold-drawn electro welded wires, which provide the
signed according to the NSR-10 regulation are supposed to have the minimum steel ratio required by the local regulations. To meet the
capability of reaching a maximum lateral drift of 1.43% for the design ultimate flexural demand, additional reinforcement made of deformed

⁎
Corresponding author.
E-mail address: [email protected] (C.A. Blandon).

https://doi.org/10.1016/j.engstruct.2018.08.089
Received 22 March 2018; Received in revised form 19 July 2018; Accepted 26 August 2018
Available online 10 September 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

bars are sometimes placed at the wall edges or at connections between and low axial load ratio reached a maximum lateral drift of 1.0% and
walls. Walls with confined boundary elements are scarce or when presented a rupture of extreme tensioned vertical reinforcement, which
present, the effectively confined core area is limited or non-effective generated a severe strength degradation.
because of the small available thickness [4]. Due to the architectonic Although these references provide information about performance
and structural dual purpose of the walls, another key characteristic of and failure modes of the tested walls, the main characteristics of these
this particular system is that walls are usually connected at one or both specimens have significant differences to the walls of interest in this
edges forming I-, T-, C-, L-shaped or any other irregular shaped cross- study including axial load ratios, reinforcing steel ratios, steel me-
sections. The wall characteristics and irregularity of wall cross-sections chanical properties, transverse section geometry and steel distribution
is also typical for other countries in South America like Chile [5] and among others. The lack of experimental and numerical information for
Peru [6]. the specific type of walls of interest, hinders the possibility of verifying
Evaluation of buildings from earthquakes in Chile (2011) and New if the available design guidelines are directly applicable to system de-
Zealand (2011) indicated that structural damage of concrete walls was scribed above, as these guidelines have been defined based on in-
associated to high axial loads, low wall area per floor, irregular element formation from walls with significantly different geometrical char-
configuration and distribution and high slenderness of the walls [7–10]. acteristics and reinforcement arrangements [19–21,7].
Even if the buildings affected during these earthquakes have different This paper shows and discusses the results of an experimental pro-
configurations with respect to the Colombian case, the observed da- gram comprising quasi-static cyclic tests of four slender and lightly-
mage indicates that the transverse slenderness of the walls at the critical reinforced concrete thin walls with different geometrical configura-
section, existing in the Colombian buildings, could facilitate out-of- tions, reinforcement mechanical properties and distribution, which are
plane instability when subjected to seismic load reversals. Additionally, representative of the type of buildings described above. The seismic
other aspects of their response associated to specific features of the local response of the specimens was assessed in terms of crack propagation
design and construction methodology are worth investigating. Ac- and failure modes, hysteretic curves, contribution of rocking, flexural,
cording to complementary studies carried out by Arteta et al. [3] and shear and sliding components to lateral drift, stiffness degradation, and
Arteta [11], the Colombian thin-wall archetype has low gravity axial energy dissipation capacity.
loading (axial load ratio below 10%), non-ductile welded-wire meshes
(WWM) as longitudinal and transverse reinforcement, and pre- 2. Experimental program
dominantly low longitudinal reinforcement ratio. Additional ductile
bars at the edges can be observed in the reinforcement layout of some The experimental program comprised the tests of four reinforced
buildings but boundary elements are absent for most cases. concrete (RC) walls with characteristics similar to the construction
The laboratory experimental data and post-earthquake field ob- practice of buildings with thin and slender RC walls with single curtain
servations of walls with the above characteristics is limited, especially of web reinforcement. The specimens were tested under pseudo-static
for thin walls with a single curtain of reinforcement and M/VLw ratios reversed-cyclic loading in the Structural Mechanics Lab at the EIA
larger than two. A previous test program carried out at EPFL [12,13] University in Colombia. The test setup includes a combination of axial
addressed the seismic response of one typical wall configuration of load, shear force and flexural moment gradient that can be considered
Colombian buildings through unidirectional and bidirectional tests on as representative of the seismic force distribution in walls within a real
two walls of 80 mm and 120 mm thick. These tests showed that the building designed according to the current practice in seismic regions in
walls could be prone to out-of-plane buckling and limited displacement Colombia.
capacity, below 0.7% drift ratio. However, during this latter program,
only one wall configuration was considered and the longitudinal steel 2.1. Variables of interest and specimen definition
was significantly more ductile than typically used in Colombian con-
struction practice. A more recent program [14,15] focused on uniaxial The main characteristics of the wall specimens were defined based
tension-compression tests on a series of 12 isolated boundary elements on the statistical analysis of a database that comprised 28 RC thin-
with different thicknesses, steel reinforcement ratios, and rebar eccen- walled buildings constructed in Colombia [3]. The buildings analyzed
tricities. These tests reported the behavior of specimens representing vary between 5 and 18 stories, with wall area densities in the long-
the boundary elements, but they evidently miss the effect of the entire itudinal (Dl) and transverse (Ds) directions between 1.5 and 6%, with an
wall; in particular, as discussed by Rosso et al. [15], the influence of the average of 3.6% and a coefficient of variation of 0.27. The length of
vertical displacement profiles imposed on wall boundary elements is flanged walls carrying most of the base shear is in the range
significantly distinct from the imposed displacement on uniaxial tests. 2 ≤ Lw ≤ 8 m, with a typical length of 4.5 m. The clear height of each
McMenamin [16] carried out tests on several slender precast cantilever story is 2.4 m, with almost no variation from one structure to another.
walls; however, only two of them had an M/VLw ratio of 2.5. The height The expected gravity axial load on the walls vary between 2 and 11% of
to thickness ratio was 50 and the vertical steel ratios were 1.1% and Ag f′c, where Ag is the gross area of the cross-section, and f′c is the
0.6%. The former specimen presented reinforcement buckling and nominal concrete strength at ground floor. All walls have distributed
concrete spalling failure for a drift below 2% while the latter specimen steel in the web and the flange. Excluding the wall edges, longitudinal
presented reinforcement fracture failure for a drift below 1%. Both tests steel ratio of distributed steel in the web (ρw) varies between
did not show a significant out-of-plane response. Carrillo and Alcocer 0.2% ≤ ρw ≤ 0.7%, with a typical value of 0.25% (minimum code re-
[17] reported on results of quasi-static and dynamic tests of walls with quirement). The analysis of the database also included representative
H/Lw ratios varying between 0.5 and 2 and with web shear reinforce- values of the thickness of flanged walls, shear span ratio, steel re-
ment made of a single curtain of welded-wire meshes; however, walls inforcement ratio, number of reinforcement curtains, as well as esti-
were tested under low axial loads that are characteristic of low-rise mations of neutral axis depth from basic section analysis. Such analysis
housing. Tomazevic et al. [18] evaluated the seismic behavior of ten resulted in the definition of the specimen with the characteristics shown
(10) rectangular reinforced concrete shear-walls with H/Lw ratio of 1.4 in Table 1.
and double curtain of wire mesh. They analyzed the influence of dif- Geometry, steel layout and type of steel reinforcement were defined
ferent parameters such as the amount and distribution of the steel and as the key variables to evaluate in the experimental program. Regarding
the axial load ratio on the seismic response. The amount of horizontal the geometry, specimens W4, W5 and W6 were conceived to char-
and vertical reinforcement varied from 0.26% and 0.38% and two axial acterize full scale T-shaped walls with a thickness of 100 mm, length of
load ratio were considered (0.07 and 0.14 fc'Ag). Six unconfined spe- 2.5 m and clear inter-story height (Hw) of 2.4 m (see Fig. 1). These three
cimens were tested. The specimens with unconfined boundary elements walls were named sequentially following a previous experimental

176
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Table 1
Characteristics of test specimens.
Wall Web reinforcement Longitudinal reinforcement at wall edge Type of web reinforcement Additional wall edge reinforcement Axial load (kN)

ρs (%) db (mm) ρb (%) db (mm)

W4 0.27 6.4 – – Distributed bars None 470

W5 0.26 7.0 – – Electro-welded cold-drawn mesh None 470
W6 0.27 6.4 2.53a 12.6 Distributed bars 2#12.6 mm 470
W7 0.27 6.4 1.27b 12.6 Distributed bars 2#12.6 mm 490

a
Concrete area estimated as wall thickness (100 mm) by edge length defined as 2 times the concrete cover plus the distance between 13 mm (#4) reinforcement
bars which gives a total of 100 mm.
b
Concrete area equal to the short flange total area (200 mm × 100 mm).

program [22]. The T-shaped geometry characterizes connected walls in ρt,W5 = ρl,W5 = 0.26%) and the additional reinforcement at the edges is
the perpendicular direction at one wall edge, which is a predominant placed to provide the wall with additional strength. This additional
geometrical configurations of the construction system. The flange of the reinforcement was also added to evaluate the potential of out-of-plane
T shape walls was included to evaluate the stabilizing effect of a con- instability as this behavior is more likely to occur as the steel ratio
necting perpendicular wall on one side and to evaluate the effect of the increases. The additional steel at the edge of the web however, was
vertical reinforcement placed along the connecting wall in the per- defined based on the maximum test set up capacity. According to the
pendicular direction. However, the flange length and the steel included same construction practice, the uniformly distributed reinforcement
in such flanges was mainly defined considering construction limitations along the web is connected to the foundation using 350 mm lap splices
and the set up capacity of the laboratory and do not necessarily re- at the wall base with 6.3 mm (#2) deformed reinforcing bars placed
present a wall of any particular length. Selected web thickness fulfills inside the foundation beam. The 12.6 mm (#4) bars are also lap spliced
two requirements: (i) to be representative of construction in high at the base using dowel bars of the same diameter and extended
seismicity areas in Colombia, while (ii) still being a potentially critical 700 mm inside the wall. Dowel bars are the starter reinforcement that
case for lateral (out-of-plane) instability under cyclic loading. Such extends from the foundation into the wall. The use of low longitudinal
instability may arise from the large unsupported length of the web and steel ratio along the web has been previously identified as a critical
the large slenderness ratio (Hw/tw) [12,23–25]. condition as it could increase the crack width and lead to prompt
The web of all specimens were reinforced with uniformly dis- fracture of reinforcement in the wall section [26].
tributed vertical and horizontal steel. The flange was reinforced with Specimen W7 had a slight geometrical variation when compared to
three 12.6 mm (#4) longitudinal bars. This arrangement was of interest the other specimens. It consisted in an additional short flange con-
for this study as the additional flange reinforcement and the cross forming a pseudo I-shaped section representing the case where ortho-
section geometry may facilitate triggering flexural-compression failures gonal walls are connected at both sides of the web, which is also a
on the web edge. In addition to the distributed reinforcement described common feature for this construction system. This additional flange had
above, specimen W6 had two 12.6 mm (#4) reinforcing bars at the web two 12.6 mm (#4) reinforcing bars, similar to specimen W6. The ad-
edge to represent another typical construction practice for walls with ditional flange was included to evaluate a possible stabilization effect
larger moment demands than those represented by W4 and W5. In such on the web edge.
case, the steel reinforcement ratio at the web is close to the minimum Specimens W4, W6 and W7 had 6.3 mm (#2) reinforcing bars dis-
(0.25%) specified by ACI 318 Code [2] (e.g. ρt,W4 = ρl,W4 = 0.27% and tributed along the flange and the web. The web vertical and horizontal

Fig. 1. Geometry and reinforcement layout of the specimens.

177
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 2. Test setup: (a) experiment setup, (b) instrumentation layout and sensor numbering.

reinforcement of specimen W5 comprised a cold-drawn welded-wire

mesh (WWM). Such steel mesh is widely used as it helps reducing the
construction costs and installation time [27,28]. In fact, it is the most
common reinforcement used in RC thin walls, although its non-ductile
behavior is well known. Horizontal and transverse steel reinforcement
ratios were kept constant for all specimens.

2.2. Tests setup and loading protocol

The test setup comprised a steel reaction frame that was braced in
two orthogonal directions (Fig. 2a). Loads were applied with two ver-
tical actuators of +700 kN (−520 kN) and one horizontal actuator of
Fig. 3. Loading protocol.
+500 kN (−410 kN) of ± 250 mm maximum stroke. The vertical ac-
tuators were connected to a HEA 450 steel beam, placed on a
550 mm × 350 mm reinforced concrete beam on top of the wall. The displacement of the walls. The first yield displacement was estimated as
actuators were separated 2.2 m and applied the total the gravitational 2.0 mm according to results of a numerical model of the wall developed
force (Fvl and Fvr). This force was programmed to keep an axial load using the software DIANA [31], and calibrated based on test results
ratio during the entire lateral load protocol of 0.05P/Agf’c, based on a reported by Almeida et al. [13]. Cycles of 0.4 mm, 0.8 mm, 1.2 mm and
specified concrete strength of 35 MPa, Additionally, these actuators 1.6 mm were defined before attaining the estimated first yield dis-
applied a variable vertical force (ΔFv), that was slaved to the horizontal placement of 2.0 mm. These deformation levels were to have a better
force (Fh), programmed to ensure that the value of the shear span ratio evaluation of the stiffness degradation at low drift demands. The peak
was constant (Mb/VbLw = 2.08). Larger shear span ratios up to 3.5 displacements for the next cycles were defined to achieve drift levels of
could be observed after the statistical analysis of the building database; 0.1%, 0.13%, 0.17%, 0.33%, 0.42%, 0.5%, 0.83%, 1.16%. The walls
however, the value selected is a representative value. Additionally, this were subjected to two full cycles at each drift level to assess the cyclic
shear span fit the maximum wall length and load capacity available in degradation response of the walls. A slow loading velocity was chosen
the laboratory [3]. Experimental and analytical evidence has shown to disregard dynamic effects and to allow a continuous control of the
that lateral displacements induce significant variations on the axial load instrumentation and the hydraulic system.
[29,30]; however, this variable was not evaluated in this experimental
program. 2.3. Instrumentation
The foundation beam of the walls was bolted to a strong concrete
floor to avoid uplift and sliding during testing. The out-of-plane dis- The instrumentation scheme comprised several transducers to
placement of the specimen was restrained by two measure in-plane and out-of-plane deformations. Fig. 2b shows the
150 mm × 150 mm × 4 mm steel tubes located at both sides of the top layout of the potentiometric displacement transducers (PDT) used to
HEA 450 beam and attached to the reactions frames (only front re- measure the contribution of sliding (37), flexural (15 to 23, 32 to 36)
straining tube shown in figure). and shear deformations (24 to 31) to the total displacement of the wall.
Once the axial load had been applied to the wall under a force- In addition, wire-PDTs (1′ to 13′) were connected perpendicularly to
control protocol, a displacement-control protocol was implemented to the wall to capture the out-of-plane displacement profile along the web
apply the loading history shown in Fig. 3. The load steps inducing edge, along its length, and at 1.2 m above the wall-foundation interface.
flange compression where defined to have a positive drift. The dis- Other sensors were installed to measure sliding (39) and uplift (38, 40)
placement cycles were defined in terms of the horizontal yield of the foundation beam. To trace the cracking pattern evolution, a

178
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 4. Stress-strain curves of steel reinforcement: (a) 7 mm-diameter electro-welded wire mesh (3 samples), (b) 6.3 mm-diameter (#2) deformed bars (6 samples).

Fig. 5. Cracking pattern at the end of test of specimen W4: (a) flange compression cycles (left), (b) flange tensile cycle (right).

random speckle pattern of high contrast was painted on the web surface specimen W5 transitioned from the elastic to the plastic branch
of walls W5 to W7. Pictures were recorded every 30 s during the tests smoothly without developing a yield plateau and exhibiting a flat post-
with a full-frame digital SLR camera. Images were post-processed using yield response, characterized by a rupture strain smaller than 3% in
a digital image correlation (DIC) technique [32–34]. average (Fig. 4a). The 6.3 mm (#2) bars used to reinforce the web of
specimens W4, W6 and W7 did not show a defined yield plateau either,
however, the steel exhibited some strain hardening and the rupture
2.4. Mechanical properties of the materials
strain was approximately twice the rupture strain for the cold-drawn
wire from the mesh (Fig. 4b).
Specified nominal compressive strength for the concrete was
35 MPa. Actual concrete strength at the day of testing was 39.1 MPa for
W4, 40.1 MPa for W5, 39.2 for W6, and 47.0 MPa for W7. The tensile 3. Wall behavior
stress–strain curves of the steel reinforcement measured using coupon
tests are shown in Fig. 4. Tensile tests of reinforcement were carried out In this section, the propagation of damage of each wall specimen is
in the Materials and Structures Laboratory at the Nueva Granada evaluated in terms of its cracking patterns and failure mode. The hys-
Military University in Colombia. The yield and maximum stress for the teretic behavior is assessed in terms of lateral force and lateral drift
reinforcing steel were 563 MPa and 691 MPa, respectively, for the ratio.
6.3 mm (#2) bars; 419 MPa and 630 MPa, respectively, for the 12.6 mm
(#4) bars; and 723 MPa and 759 MPa, respectively, for the 7 mm- 3.1. Cracking propagation and failure mode
WWM.
The 12.6 mm (#4) reinforcing steel complied with ASTM-A706 Cracking for specimen W4 was recorded in the classical manner
standard [35] with the typical yield plateau and fracture elongation using markers and crack width rulers at the end of the cycles of the
larger than 10%. However, the cold-drawn wires from the mesh and the loading protocol. On the other hand, for specimens W5 to W7, a digital
6.3 mm (#2) reinforcing steel exhibited a less ductile behavior, as image correlation (DIC) technique was used to assess the crack evolu-
shown in Fig. 4. The stress-strain curve of the 7 mm-WWM used in tion pattern. Cracking is described in terms of drift ratio (ΔR). The drift

179
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 6. Cracking pattern at the end of test of specimen W5: (a) load steps inducing flange compression (left), (b) load steps inducing flange tension (right). The plot
shows the vertical strains as computed by the DIC software. The patter revealed by the DIC software does not represent the crack width but a region of large strains
around the crack.

ratio was computed as the quotient between the lateral displacements top of the wall and it was probably associated with a stress con-
measured below the top beam and the height of the wall panel (2.4 m). centration at the top beam interface. Fracture of the 6.3 mm (#2)
The cracking pattern at the end of the test of W4 is shown in Fig. 5. starter bars also occurred for a drift ratio ΔR = 0.67% during the load
The crack pattern that developed during the load steps inducing flange step that induced compression in the flange. A vertical crack appeared
compression was significantly different to the one developed during the at the web toe along the 100 mm thick wall side and continued enlar-
steps that put the flange in tension. In the first case, cracks were located ging and spreading upwards. This crack triggered a lap-splice failure of
along the web at three main locations, namely, at the wall foundation the specimen due to the slippage of the 12.6 mm (#4) reinforcing bars
interface, at 0.55 m and at 1.7 m above the panel-foundation interface. during the first load step of the 0.83% drift ratio that induced com-
When the first crack opened, the wall experienced a sudden displace- pression in the flange and tension in the lap splice at the web. This
ment that was captured by the sensors. For the load steps inducing failure mode resembled wall lap-splice failures observed in other ex-
tension cycles, cracking was evenly distributed along the wall height perimental programs [36].
(Fig. 5a). Failure, defined as a lateral force drop larger than 20% of the Cracking propagation of specimen W7 was similar to that of spe-
maximum lateral force, was reached when loading during the cycle cimen W6 for the half-cycles that compressed the edge with the short
with ΔR = 0.63% in the negative direction (flange in compression), flange. Cracks on the flange side were evenly distributed along the
when several of the 6.3 mm (#2) dowel bars connecting the wall to the height with wider cracks spreading towards the compressed toe at
foundation fractured. When reversing the load, the concrete at the web 0.7 m and 1.1 m above the foundation. There was also a wide crack at
edge started crushing followed by the fracture of the 12.6 mm steel bars the wall-foundation interface that appeared at early loading stages.
of the flange at ΔR = 0.71%. Fracture of the 6.3 mm (#2) dowel bars was observed during the second
Fig. 6 shows the cracking pattern at the end of the test of specimen load step of ΔR = 0.83% that induced compression in the flange and
W5. For the direction where the flange is in compression, a crack at the continued for ΔR = 1.16% for both directions. The fracture of these
wall-foundation interface was also observed, and it remained as the dowel bars was followed by the rupture of the 12.6 mm (#4) dowel bar
only significant crack during several cycles until a second horizontal at the wall-foundation interface during the second cycle of the same
crack appeared at the web edge. This second crack was located 1.2 m drift at the side with the shortest flange. This fracture caused the failure
above the wall base. A secondary crack located at the wall-top beam of the specimen due to the loss of the lateral strength larger than 20%.
interface also appeared at early cycles but it did not exhibit an apparent The reinforcement showed an inclined fracture that could have been
opening. Cracking propagation along the flange was similar to W4 as it cause by several tensile compressive cycles that induced buckling (see
was distributed along the entire height of the panel. The dowel bars Fig. 8).
located near the flange started fracturing at ΔR = 0.50% when com- The evolution of cracking of specimens W5 and W6 is shown in
pressing the web edge. A strength degradation was observed at Fig. 9 for three different levels of drift ratio demand. Cracking ex-
ΔR = 0.67% when more dowel bars fractured. Spalling of the com- tending more than half of the wall length, on the flange side of all the
pressed web toe also occurred at the same drift level. Failure occurred specimen, was observed for drift ratios as low as ΔR = 0.17%. The en-
due to the fracture of the 12.6 mm (#4) reinforcing bars on the flange tire crack pattern was completely developed at a drift ratio ΔR = 0.42%.
side, at the wall foundation interface, at ΔR = 0.83% (Fig. 6b). The cracks on the flange side (right hand side of the pictures) were well
Fig. 7 shows the cracking pattern at the end of test W6. Cracking on spread for all walls and remained approximately horizontal within the
the web edge side of specimen W6 was significantly influenced by the flange. The crack pattern for tensile flange cycles became then inclined
two additional 12.6 mm (#4) mm reinforcing bars. They induced a along the web with angles varying between 16 ≤ θcr ≤ 38 degrees, with
more distributed crack propagation along the entire wall height com- a typical inclination of approximately 35 degrees. On the other hand,
pared to specimens W4 and W5. Similarly to the other units, the largest crack pattern on the web-edge (left hand side of the pictures) differ
crack was located at the wall-foundation interface. One additional crack among the walls. For instance, only a few wide cracks formed on units
was observed at the web edge side, 0.70 m above the base, and was W4 and W5 (reinforced with the WWM) while a larger number of
associated to the lap splice of the 12.6 mm (#4) reinforcement bars, narrower cracks formed for specimens W6 and W7, due to the addi-
which ended at this location. Another large crack was observed at the tional reinforcement placed at the web edge (W6) or at the shorter

180
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 7. Cracking pattern at end of test of specimen W6: (a) flange compression cycles (left), (b) flange tensile cycle (right). The plot shows the vertical strains as
computed by the DIC software. The patter revealed by the DIC software does not represent the crack width but a region of large strains around the crack.

flange (W7). inducing web tension before strength started to decrease due to rupture
of the 6.3 mm (#2) dowel bars at the web side. The specimen W5
3.2. Hysteretic response reached a maximum drift of +1.10% for the flange compression load
step and −1.20% for the flange tensile load step before the fracture of
Hysteretic curves are fundamental indicators of the global strength most of the dowel bars and the crushing onset of the web edge concrete.
and deformation capacities as well as other important parameters such However, the strength degradation for those drift levels was approxi-
as stiffness degradation and energy dissipation [37]. The hysteretic mately 42% and 31%, respectively. A 20% of strength degradation,
curves of the walls are shown in Fig. 10. They are expressed in terms of which is commonly used as a criterion for failure, was observed at
lateral force and lateral drift ratio. To account for different values of ΔR = +0.80% and ΔR = −0.97% for compressive and tensile flange
concrete strength ( fc′), the ratio between the shear stress and the square load steps, respectively. Specimen W6 showed more stable loops with
root of fc′ ( fc′ ) is also shown in the figure. The shear stress was com- strength degradation lower than 13% during the second loading cycle
puted as the ratio between the lateral force and the gross area of the for a lateral drift close to ± 0.85% in both directions (Fig. 10c). A
wall web (Lwtw). sudden strength drop occurred at ΔR = +0.92% during the flange
As shown in Fig. 10a, specimen W4 showed a stepped strength de- compression load steps due to the failure of the lap splice of the
gradation after a lateral drift of +0.61% when the 6.3 mm (#2) dowel 12.6 mm (#4) reinforcing bars at the web edge. At that instant, the
bars started rupturing at the wall-foundation interface during the load strength degradation reached 16%. The hysteretic response of specimen
step that induced compression in the flange. After a sudden strength W7 was similar to that of specimen W6 up to a lateral drift of
drop at +0.75% drift, the load was reversed and a lateral drift of ΔR = ± 0.85%. As shown in Fig. 10d, the specimen W7 was able to
−0.73% was reached before the strength started to decrease due to the accommodate two additional cycles up to ΔR = 1.20% but suffered a
crushing of the concrete at the bottom web edge. As shown in Fig. 10b, strength degradation larger than 35% at this point. A 20% strength
the specimen W5 was able to reach a drift of +0.69% during a load step degradation was observed for a drift ratio close to +1.15% for the

Fig. 8. Reinforcement fracture at the shorter flange of specimen W7.

181
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 9. Evolution of crack propagation for specimens W5 and W6 at three different drift ratios. Pattern reveals regions of large vertical strains around the cracks and
not the actual crack width.

flange compressive cycle and −1.24 for the flange tensile cycle. the total response.
For special RC structural walls, ACI 318 [2] prescribes that the re-
lative contribution of concrete to nominal wall shear strength is 0.17
fc′ MPa for walls with H/Lw equal or higher 2.0. As shown in Fig. 10, 4. Analysis and discussion of results
the peak strength of wall specimens varied between 0.14 fc′ MPa for
The response of the specimens is compared next with regard to their
W5 and 0.21 fc′ MPa for W6 during the flange compression cycles.
drift capacity at certain limit states, their backbone force-displacement
These peak strength values close or lower than the contribution of
envelopes, the contributions to the drift from flexural, shear and sliding
concrete to wall shear strength indicate a limited shear contribution to
deformations, and the evolution of stiffness degradation and energy

Fig. 10. Drift ratio versus base shear relationships for walls W4, W5, W6, and W7.

182
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

fs,max = fy). The YBC limit is met when the theoretically computed sa
values reach the experimental measurements recorded with the dis-
placement transducers near the base (32 and 15). The short gage length
of these transducers (e.g. 50 mm) ensures that the recorded displace-
ments only account for crack opening at the base without much con-
tribution from strains above the crack.
The YABC limit state is set by determining the test step in which the
rebar elongation (dstr), measured with the transducers located along the
edges of the wall, exceeded the yield strain limit. The strain in the bars
at the i-th location along the height of the panel was estimated as
εStr,i = dstr,i/Ltr,i where is the Ltr,i is the gage length of the i-th trans-
ducer. This value was compared against the yield strain of each bar in at
the edge of the panel (i.e. εy = fy/Es). The PLR limit state is related to
the peak load for each loading direction and was directly obtained from
the hysteresis response. For code-based seismic design, the LLR limit
Fig. 11. Backbone curves of all specimens.
state is reached when a loss of 20% in capacity is observed after the
peak shear strength is achieved.
dissipation. Fig. 11 shows the backbone curves enveloping the previously pre-
sented hysteresis loops, marking the limit states previously described.
4.1. Limit states For the direction compressing the web edge, the response of specimens
W4, W5 and W6 was similar. This is expected because the specimens
To identify key parameters of the behavior of the walls, five per- had the same cross-section geometry, and the response in this direction
formance levels were defined on the envelope of the hysteretic response is mainly commanded by the tensile strength of the reinforcement in the
as follows: cracking moment (CM), first yield at the base crack (YBC), flange, which was the same in this set of specimens. The envelope for
first yield above the base crack (YABC), peak lateral resistance (PLR), specimen W7 exhibits a larger capacity in strength and displacement
and loss of lateral resistance (LLR). due to the enlarged element at the web edge.
The CM limit state defines the drift range over which the in-plane The response for the direction compressing the flange is dependent
stiffness of the wall panel is largest. Additionally, for lightly reinforced on the reinforcing characteristics of the specimens. Specimens W6 and
wall panels where the yielding moment may be close to the cracking W7 reached the largest capacity among all specimens thanks to the two
moment, the CM limit state definition may help detective the potential additional 13 mm bars at the web edge. Furthermore, specimen W7
of a brittle failure (e.g. if cracking moment is larger than the yielding exhibited the largest displacement capacity of the four units. Although
moment). The CM limit state was theoretically defined by computing specimen W5 was expected to reach a higher strength as compared to
the cracking moment, Mcr, of the cross section using Eq. (1) W4 because of the larger yield strength of the WWM, its capacity was
12% lower. This could be in part because the reinforcement details and
(fr + P / Ag ) Ig (0.62 fc′ + P / Ag ) Ig the type of steel at the foundation-wall interface, where most of the
Mcr = =
y y (1) plastic deformations occurred, is similar for both units.
Drift ratios at defined limit states are shown in Table 2. For the
where y is the distance from the centroid to the extreme fiber in tension,
direction compressing the web edge, the DCR limit is reached in the
Ag and Ig are the gross area and gross moment of inertia of the cross-
drift range 0.024 ≤ ΔR ≤ 0.033% for specimens W4, W5 and W6, and
section, respectively, and fr is the modulus of rupture of the concrete,
at ΔR = 0.041% for specimen W7. For the opposite direction, the DCR
which is defined in terms of the compressive strength, fc′ (MPa), at the
limit is reached in the range 0.017 ≤ ΔR ≤ 0.024% for specimens W4,
day of the test following NSR-10 provisions. The base shear force as-
W5 and W6, and at ΔR = 0.032% for W7. The YBC limit was also
sociated to the CM limit is estimated using the shear span ratio Mu/VuLw
reached early on in the loading sequence. In both loading direction, all
of the experiments, which as stated was kept constant at 2.08 for all
specimens reached the YBC limit in the range 0.08 ≤ ΔR ≤ 0.10%, ex-
tests.
cept for specimen W5 who reached at 0.03% drift when compressing
For slender reinforced concrete walls with low axial load, one of the
the flange. The limited cracking drifts values discussed above should be
main contributions to the total displacement comes from rigid body
considered in newer revisions of the Colombian Building Code, NSR-10,
rotation around the neutral axis, due to the crack at the wall-base in-
to prescribe the use of non-cracked moments of inertia when analyzing
terface [38]. Furthermore, the impact of the base-crack opening on total
the seismic action on multistory wall structural systems.
displacement is greater for lightly reinforced panels [39]. The YBC limit
Specimen W5 did not reached the YABC limit when compressing the
is defined based on the strain penetration model presented by Moehle
flange, indicating that most of the plastic deformation on the edge rebar
[40]. Such model relates slip deformation due to rigid-body rotation
with both the geometry and stress state of the reinforcing bars at the
Table 2
edge of the panels in the interface with the foundation. The model as-
Drift ratios at defined limit states.
sumes that the tension force in the bar gradually decreases from the
maximum stress demand at the wall-base interface (fs,max), to zero Limit Drift at defined limit states (%)
State
stress at the anchorage distance into the base. Bar elongation (sa) from
W4 W5 W6 W7
this point to the wall-base interface can be computed as the integral of
the strain along the anchorage length (la). If fs,max = fy (i.e. YBC limit is FCCa FTCa FCC FTC FCC FTC FCC FTC
reached), sa can be estimated using Eq. (2):
DCR 0.025 −0.024 0.024 −0.017 0.033 −0.020 0.041 −0.032
f y2 db YBC 0.08 −0.08 0.03 −0.10 0.08 −0.08 0.08 −0.10
sa = YABC N.R. −0.19 0.09 −0.26 0.24 −0.21 0.14 −0.29
8Es u (2) PLR 0.32 −0.49 0.44 −0.58 0.76 −0.79 0.85 −0.85
LLR 0.59 −1.06 0.80 −0.87 0.92 −0.86 1.15 −1.24
where db is the bar diameter, Es = 200 GPa is the elastic modulus of
steel, and u = 1.0 fc′ (MPa) is the average uniform bond stress re- N.R: limit state Not Reached.
a
commended by ACI 363-92 [37] for linear response (i.e. up to FCC: Flange compression loads step. FTC: Flange tensile load steps.

183
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

was concentrated at the base crack. Additionally, the lack of strain

hardening of the WWM on the web prevented the spread of the plas-
ticity along the web edge. The YABC limit was recorded at 0.09, 0.24
and 0.14% drift ratio for specimens W5, W6 and W7, respectively. For
the opposite direction, the YABC limit was achieved between 0.19 and
0.29% drift ratio for all specimens. When compressing the flange, the
specimen W5 reached its maximum strength at the smallest drift ratio
(i.e. ΔR = 0.32%) among all specimens. This value is 27% smaller than
that of specimen W4. Past this point, the force–displacement backbone
of specimen W5 remained flat up to the onset of strength loss at
ΔR = 0.32%. Specimen W6 and W7 reached the maximum strength
close to 0.80% drift ratio in both directions.
For the direction compressing the flange, the displacement capacity
at the LLR limit of specimens W6 was 15% and 56% larger than that of
specimens W4 and W5, respectively. This result shows the benefits of
the additional ductile reinforcement at the web edge, which provoked
an increase in both, the strength and displacement performance of the Fig. 13. Evolution of base rotation contribution to total drift.
panels. The displacement capacity of specimen W7 was 25% larger than
that of W6. For the direction compressing the thin web edge, the dis-
placement capacity of specimen W7 was at least 17% larger than that of As previously discussed, the rocking at the wall-foundation interface
any other specimen, evidencing the beneficial effect of the enlarged was the main contributor to the total displacements in view of the low
boundary. longitudinal reinforcing ratio of the specimens. Such rocking induced a
crack along the interface and subsequent large strains of the dowel bars
connecting the wall to the foundation. The rotation from the base crack
4.2. Deformation components
was largest for specimen W5 (with the WWM) at any instant of the
experiment. This is consistent with the scarcity of cracks along the
The sources of wall deformation were identified based on the re-
height of the web edge (see Fig. 6). On the other end, specimens W6 and
cordings from the different sensors located on the walls (see Fig. 12).
W7 exhibited the smallest rigid body contribution to the total dis-
Four different deformation components were identified: (i) contribution
placement greatly due to the additional reinforcement at the web edge,
of rocking or the rotation along the wall foundation, (ii) flexural de-
which prevented a wider crack opening, and facilitated the spread of
formations due to the spreading of plasticity along the wall, (iii) web
plasticity along the web edge.
shear and (iv) sliding at the base. The contribution of the four de-
formation components to the total drift at the PLR limit state is shown
in Fig. 12. The rotation along the wall-foundation interface crack in- 4.3. Out-of-plane deformation
volved the largest contribution as it reached 42.8%, 64.5%, 44.1% and
37.3% of the total displacement contribution for specimens W4, W5, The instrumental and video recordings taken at the web edge for
W6 and W7, respectively. Displacements of unit W5 when compressing specimens W4 to W6 showed that the out-of-plane displacement (OPD)
the flange are therefore mainly due to rigid body rotation, which is of the specimens was negligible, hence no signs of wall buckling were
consistent with the absence of yielding above the base crack. The observed. Maximum OPD normalized by wall thickness (OPD/tw ) was
flexural contribution to the total deformation was 42.8% for W4, 24.6% smaller than 0.085, being larger for specimen W4. The tensile strain
for W6, 27.7% for W7 and was negligible for W5 as most of the rotation applied prior to subsequent loading in compression is the parameter
was concentrated at the foundation-wall interface and only one ob- that governs the out-of-plane instability [42]. As shown in Table 3, the
servable crack formed along the wall height. Shear deformation, cal- maximum total strain along the web edge recorded from the cycles that
culated according to Hiraishi [41], had no significant contribution to put the web edge in tension was lower than 1%. Additionally, if the
total deformation of specimens W4 and W5; however, shear deforma- contribution of the base crack is not accounted for in this average strain
tion contributed 25.9% and 22.9% to the total deformation for speci- calculation, the maximum strain above the base crack did not exceed
mens W6 and W7, respectively. This increase of shear deformations is 0.4%. This low strain values above the wall-foundation interface have
explained by the increased number of cracks along the web in these likely prevented the occurrence of relevant out-of-plane deformations.
specimens, which were not observed in specimens W4 or W5. Sliding According to Eq. (4) from Parra and Moehle [25] the critical tensile
along the base was not significant for any of the walls as contributions strain (ξsm,cr) that would induce out-of-plane failure would be around
were lower than 7%. 1.6% for the walls with no additional end reinforcement (W4 and W5)
The base rotation drift is compared to the total drift when com- and around 1.1% for the wall with the additional 12.5 mm (#4) re-
pressing the flange in Fig. 13 for three limit states: DCR, YABC and PLR. inforcement bars at the wall end (W6). As one the wall had only one
reinforcement layer κ was defined as 0.5, the wall was assumed to have
perfect fix at the top and bottom beams, and therefore k was defined as
0.5. The wall thickness (bcr) used was 100 mm and the wall clear height
(hu) was 2400 mm.

Table 3
Maximum strain during load stages inducing tensile cycles at the web edge.
Specimen W4 W5 W6 W7

Average strain over wall height at web edge 0.010 0.007 0.008 0.010
(mm/mm)
Average strain if opening of base crack is 0.003 0.002 0.004 0.004
excluded (mm/mm)
Fig. 12. Contribution of deformation components to the total displacement for
Drift (%) +0.96 +0.86 +0.83 +1.14
the PLR limit state in the direction that compresses the flange.

184
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 14. Stiffness degradation of walls: (a) flange tensile cycles, (b) flange compressive cycles.

2
πb 4.5. Energy dissipation
ξsm, cr = κξcr ⎛ cr ⎞ + 0.005
⎜ ⎟

⎝ khu ⎠ (3)
The energy dissipated is a crucial seismic-performance parameter
⎛ 2m 2m 4m ⎞ 2 for structural elements subjected to several loading cycles. The varia-
ξcr = 0.5 ⎜1 + − ⎛ ⎞ +
tion of dissipated energy versus drift ratio is shown in Fig. 15a. The
0.85 ⎝ 0.85 ⎠ 0.85 ⎟
⎝ ⎠ elastic energy was obtained as the area below the lateral force vs dis-
fy placement envelope for each load step and the dissipated energy was
m=ρ obtained as the total area inside the hysteresis loop. As shown in
fc′
Fig. 15a, the specimen W5 experienced the largest amount of dissipated
energy up to a drift of 0.57%. After this drift, W4 shows the largest
4.4. Stiffness degradation dissipated energy. The reinforcement of specimen W4 included mod-
erate ductility reinforcement mesh and no additional reinforcement at
The stiffness degradation curve is commonly related with the in- the ends while W5 had the low ductility steel mesh. On the other hand,
crease of the deformation demand of the element. For practical earth- the specimen W6 and W7 experienced the lowest energy dissipation.
quake-resistant design, deformation is commonly expressed in terms of Specimen W6 was reinforced with a moderate ductility reinforcement
story drift ratio and the stiffness degradation in terms of the ratio of the mesh and additional reinforcement bars at the wall end while W7 had
stiffness of a particular cycle and the stiffness of the first cycle or initial flanges at the two ends. Wall W6 experienced a lap splice failure and
stiffness (Ke/Ki). Using the latter approach, stiffness degradation curves W7 had reinforcement fracture at both ends. At a drift ratio of 0.7%, W6
of the test units are shown in Fig. 14. Trends of stiffness degradation reached just 61% of the energy dissipated by W4 and 73% of the energy
demonstrated that walls experienced a rapid degradation as the re- dissipated by W5. W7 dissipation was lightly smaller which may in-
duction of the initial stiffness was close to 70% (Ke/Ki ≈ 0.3) for drift dicate that the even if the concentrated reinforcement at the wall ends
ratios between 0.16% and 0.25% for the flange tensile cycles, and be- improves the crack distribution; it does not necessarily improves the
tween 0.09% and 0.15% for the flange compressive tests. At 1% drift energy dissipation.
ratio, which is considered a design limit according to the local design The ratio of dissipated and elastic energy is included in Fig. 15b. As
guidelines when uncracked sections are considered in the design pro- shown in the figure, specimen W4, W6 and W7 showed a rapid decrease
cess, the effective stiffness was lower that 10% of the initial stiffness of the relative dissipated energy up to a drift ratio of approximately
(Ke/Ki = 0.1) for all the specimens. These results indicate the sig- 0.1%. After that decrease, a slight to moderate increase and a relatively
nificance of considering the cracked sections in the design process as constant region was observed up to a drift ratio of roughly 0.5%. For
drift may be significantly underestimated if gross sections are used to drift ratios larger than 0.5%, the relative energy increases with the drift
verify drift limits defined by the Colombian code [1]. Table 4 shows ratio but at a faster rate. On the other hand, specimen W5 also ex-
values of stiffness degradation (Ke/Ki) to each of the defined limit states perienced an initial decrease of ratio of dissipated energy; however, a
for all the specimens. relatively constant rate of increase was observed up to the maximum
drift recorded. The behavior for W5 may be due to the limited cracking
Table 4 during the first cycles that increased gradually during the test. The
Stiffness degradation ratios at defined limit states. limited cracking produced a smaller dissipated energy compare to the
elastic energy but this dissipation gradually increased as a couple of
Limit State Stiffness degradation at defined limit states (%)
cracks opened along the wall at the web side.
W4 W5 W6 W7
5. Comparison with plausible displacement demand at the design
FCC FTC FCC FTC FCC FTC FCC FTC
level
DCR 0.87 0.99 0.78 0.96 0.69 0.98 0.71 0.82
YBC 0.69 0.44 0.34 0.64 0.44 0.64 0.52 0.48 Results of linear response history analyses over 15 multistory thin-
YABC 0.33 0.23 N.A.a 0.35 0.23 0.34 0.33 0.24 wall buildings from Arteta [11] are adapted next to correlate demand
PLR 0.09 0.11 0.13 0.15 0.07 0.09 0.08 0.09
LLR 0.04 0.06 0.06 0.06 0.03 0.08 0.05 0.05
and capacity of thin walls buildings in Colombia. The models were
selected from a Colombian building database comprising structures
*
FCC: Flange compression loads step. FTC: Flange tensile load steps. with 4 to 18 stories in height (i.e. 10 to 43 m in height). The median
a
not available due to sensor malfunction. value of the Sozen [43] wall-area index (ratio between area of wall-

185
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

Fig. 15. Energy dissipation: (a) dissipated energy (left), (b) relative dissipated energy (right).

webs in a given direction and the first floor area) for these buildings is web. The tested specimens variables as wall length, shear span ratio and
3.7%, with minimum and maximum values of 1.4 and 9.8%, respec- additional steel at the wall edges are representative of the database.
tively. Most of the structures evaluated (approximately 70%) have wall- However, walls with different characteristics are also common within
area index larger than 3%, which are considered “robust”, and safe the database.
under seismic shaking. Further details of the building set geometry, and The results from all the specimens showed a limited deformation
their dynamic characteristics can be found in [3]. capacity well below 1.43% which is the maximum drift limit specified
Fig. 16a shows the maximum roof drift ratio (RDR) demand versus by the code for buildings designed based on cracked sections. Hence,
cracked period relationship in the longitudinal direction of the struc- these walls would not be able to perform according to the code re-
ture. Each model is represented by its structural period and was ana- quirements if included in buildings allowed to reach the maximum drift
lyzed under a set of 11 two-horizontal-component ground motions, limit. Significant lower design drift limits should be used if this type of
selected such that the median of their geomean spectra matched the walls are to be included as part of the lateral resisting system of the
design level of the code-based spectrum. A power regression for the building.
median and the 84th percentile demand is depicted along with the in- In terms of the failure mode, the walls were characterized by the
dividual response. The upper and lower bound of drift capacity at the rupture of the steel, mainly due to the concentration of plasticity in a
LLR-limit, of the panels presented herein, in the direction compressing single crack located at the wall-foundation interface. This concentration
the flange, are traced as horizontal dotted lines. It is noted that the was more evident when a low ductility cold-drawn welded mesh was
lower bound capacity is exceeded at the median and 84th percentile used as web reinforcement, instead of deformed bars with moderate
level for structural periods larger than 1.10 s and 0.85 s, respectively. ductility. Lap splice failure mode due to bar slippage was also observed
This period range is covered by structures with more than 14 stories, which induced a significant decrease of the energy dissipation com-
which represent one-third of the buildings in the database. It is worth pared to the steel rupture failure mode.
emphasizing that the elastic models do not account for flexural yielding In terms of the displacement capacity of the walls, a slight increase
of the critical cross sections, and hence, the results presented above was obtained by providing additional reinforcement with moderate
might not be conservative. ductility at the wall ends as these improved the crack distribution along
the wall height. Additional displacement capacity was obtained by
adding the short flange in W7 as it prevented the lap splice failure that
6. Conclusions
occurred in W6. There was no evidence of buckling of the wall under
load steps that induced compression on the web, possibly due to the
Four full-scale thin reinforced concrete walls were built and tested
limited elongation along the web edge, above the base crack. Stiffness
to evaluate their performance under lateral cyclic loading. These units
degradation curves demonstrated that walls experienced a rapid dete-
were aimed to represent the current local practices for one of the typical
rioration as the loss of the initial stiffness was roughly equivalent to
construction systems used in Colombia and were defined based on the
50%, 85%, 90% for drift ratios close to 0.1%, 0.5% and 0.7%, respec-
analysis of a database of buildings from 8 to 15 stories with low axial
tively. Such critical values of stiffness degradation evidenced that ty-
load ratios (below 5%) and light steel reinforcement ratios along the
pical cracking factors proposed by the local building code, which are
significantly higher, cannot be used for code-based seismic design of
walls with similar characteristics to those tested. Results of this study
are applicable only to walls with conditions similar to those reported
here. Future experimental programs should consider variations on the
axial load ratio and the impact of heavily reinforced web boundaries as
this may trigger other failure modes such as out-of-plane instability or
early concrete crushing.
A case study with a set of linear elastic response history analyses of
multistory buildings, representative of the construction practice of thin-
wall buildings in Colombia, showed that the roof drift ratio demand at
the design level of demand, might exceed the capacity of the walls
tested, specifically for structures taller than 14 stories.

Acknowledgments
Fig. 16. Displacement demand at the design level for Colombian thin-wall RC
buildings: roof-drift-ratio versus cracked period. The research was financed through an EPFL Seed Money grant

186
C.A. Blandon et al. Engineering Structures 176 (2018) 175–187

through the Ecole Polytechnique Fédérale de Lausanne (EPFL) Canterbury, Christchurch, New Zealand; 1999. 275p.
Cooperation and Development Center in Switzerland and additional [17] Carrillo J, Alcocer S. Experimental investigation on dynamic and quasi-static be-
havior of low-rise reinforced concrete walls. Earthq Eng Struct Dyn
funding was provided by Conconcreto. The authors would like to thank 2013;42:635–52. https://doi.org/10.1002/eqe.2234.
Wilson Zapata, Julian del Rio Echeverri, and Juan Fernando Garcia [18] Tomazevic M, Lutman M, Capuder F, Petkovic L. Seismic behaviour of R.C. shear-
from Conconcreto for the logistic support and technical advice. Special walls: an experimental study. In: 11th World conference on earthquake engineering,
Paper 381, Elsevier; 1996, ISBN: 0080428223.
thanks go to: Gustavo Araujo and Julio Sánchez from UNINORTE for [19] Massone L, Bonelli P, Lagos R, Lüders C, Moehle J, Wallace JW. Seismic design and
the hard work in the laboratory and reducing the experimental data and construction practices for RC structural wall buildings. Earthq Spectra
Juan Carlos Velez from Universidad de Antioquia for the laboratory 2012;28(S1):S245–56. https://doi.org/10.1193/1.4000046.
[20] Thomsen JH, Wallace JW. Displacement-based design of slender reinforced con-
assistance, to Jose Fernando Rave, Javier Penagos, Juanita Escobar, crete structural walls-experimental verification. J Struct Eng 2004;130(4):618–30.
Juan David Upegui and Jonathan Betancur at EIA University for the https://doi.org/10.1061/(Asce)0733-9445(2004) 130:4(618).
long hours preparing the specimens, and to Juan Fernando Restrepo [21] Wallace JW, Massone LM, Bonelli P, Dragovich J, Lagos R, Luders C, et al. Damage
and implications for seismic design of RC structural wall buildings. Earthq Spectra
and Julián Ossa at University of Medellín for their assistance in the
2012;28(S1):S281–99. https://doi.org/10.1193/1.4000047.
laboratory. Finally, the authors thank Mauricio Ramirez from DAQ [22] Blandón CA, Rave-Arango JF, Bonett RL. Behavior of thin reinforced concrete walls
Solutions for providing the required assistance to complete the tests. under lateral load. In: Proceedings of the VII national conference on earthquake
Authors expresses their gratitude to EIA, UNINORTE, UDEM and UMNG engineering, Bogota; 2015, Paper 25.
[23] Goodsir WJ. The design of coupled frame-wall structures for seismic actions.
universities for supporting research activities of the first four authors. Christchurch, New Zealand: University of Canterbury; 1985.
[24] Hube MA, Marihuen A, De la Llera JC, Stojadinovic B. Seismic behavior of slender
References reinforced concrete walls. Eng Struct 2014;80:377–88. https://doi.org/10.1016/j.
engstruct.2014.09.014.
[25] Parra PF, Moehle JP. Stability of slender wall boundaries subjected to earthquake
[1] NSR-10, Colombian code for earthquake-resistant construction, NSR-10, Committee loading. ACI Struct J 2017;114(6):1627–36.
100, Colombian Association of Earthquake Engineering, AIS, Colombia; 2010. [26] Dazio A, Beyer K, Bachmann H. Quasi-static cyclic tests and plastic hinge analysis of
[2] ACI 318-08. Building code requirements for structural concrete and commentary. RC structural walls. Eng Struct 2009;31:1556–71.
Committee 318. Farmington Hills, MI: American Concrete Institute; 2008. p. 467. [27] Haindl M, Hube Matías A, Arteta CA. Seismic performance assessment of a re-
[3] Arteta CA, Sanchez J, Daza R, Blandón CA, Bonett RL, Carrillo J, et al. Global and inforced concrete wall house. In: VII national conference on earthquake en-
local demand limits of thin reinforced concrete structural wall building systems. In: gineering, Bogota; 2015, Paper 23.
Proceedings of the 16th world conference on earthquake engineering, Santiago, [28] Quiroz LG, Maruyama Yoshihisa, Zavala Carlos. Cyclic behavior of thin RC Peruvian
January 9–13; 2017, Paper 2121. shear walls: Full-scale experimental investigation and numerical simulation. Eng
[4] Arteta CA, To DV, Moehle JP. Experimental response of boundary elements of code Struct 2013;52:153–67. https://doi.org/10.1016/j.engstruct.2013.02.033.
compliant reinforced concrete shear walls. In: Proceeding of the 10th US national [29] Su RKL, Wong SM. Seismic behaviour of slender reinforced concrete shear walls
conference on earthquake engineering, Earthquake Engineering Research Institute, under high axial load ratio. Eng Struct 2007;29(8):1957–65. https://doi.org/10.
Anchorage, AK, July 21–25; 2014. 1016/j.engstruct.2006.10.020.
[5] Westenenk B, de la Llera JC, Jünemann R, Hube MA, Besa JJ, Lüders C, et al. [30] Alarcon C, Hube MA, de la Llera JC. Effect of axial loads in the seismic behavior of
Analysis and interpretation of the seismic response of RC buildings in Concepción reinforced concrete walls with unconfined wall boundaries. Eng Struct
during the February 27, 2010, Chile earthquake. Bull Earthq Eng 2013;11(1):69–91. 2014;73:13–23. https://doi.org/10.1016/j.engstruct.2014.04.047.
https://doi.org/10.1007/s10518-012-9404-5. [31] TNO Building and Construction Research. DIANA – Finite Element Analysis User’s
[6] Gonzales H, López-Almansa F. Seismic performance of buildings with thin RC Manual. Release 8.1 September 2002.
bearing walls. Eng Struct 2012;34:244–58. https://doi.org/10.1016/j.engstruct. [32] Chu TC, Ranson WF, Sutton MA, Peters WH. Applications of digital-image-corre-
2011.10.007. lation techniques to experimental mechanics. Exp Mech 1985;25(3):232–44.
[7] Wallace JW, Orakcal K. ACI 318–99 provisions for seismic design of structural walls. https://doi.org/10.1007/Bf02325092.
ACI Struct J 2002;99(4):499–508. https://doi.org/10.14359/12119. [33] Pan B, Qian KM, Xie HM, Asundi A. Two-dimensional digital image correlation for
[8] Alarcon C, Hube MA, Junemann R, De la Llera JC. Characteristics and displacement in-plane displacement and strain measurement: a review. Meas Sci Technol
capacity of reinforced concrete walls in damaged buildings during 2010 Chile 2009;20(6):1–17. doi:Artn 062001.
earthquake. Bull Earthq Eng 2015;13(4):1119–39. https://doi.org/10.1007/ [34] Sutton MA, Orteu JJ, Schreier HW. Image correlation for shape, motion and de-
s10518-015-9727-0. formation measurements: basic concepts, theory and applications. New York, NY:
[9] Sritharan Sri, Beyer Katrin, Henry Richard S, Chai YH, Kowalsky Mervyn, Bull Springer; 2009.
Desmond. Understanding poor seismic performance of concrete walls and design [35] ASTM-A706. Standard specification for low-alloy steel deformed and plain bars for
implications. Earthq Spectra 2014;30(1):307–34. concrete reinforcement (pp. 6). West Conshohocken, PA: ASTM International; 2009.
[10] Kam WY, Pampanin S, Elwood K. Seismic performance of reinforced concrete [36] Almeida JP, Prodan O, Tarquini D, Beyer K. Influence of lap splices on the de-
buildings in the 22 February Christchurch (Lyttelton) earthquake. Bull NZ Soc formation capacity of RC walls. I: database assembly, new experimental tests, and
Earthq Eng 2011;44(4):239–78. findings for model development. ASCE J Struct Eng 2017;143(12).
[11] Arteta CA. Simple mechanics of reinforced concrete thin wall, design considerations [37] ACI Committee 363. Report on high-strength concrete. ACI Manual of Concrete
for Colombia. In: Proceedings of the VIII national conference on earthquake en- Practice (reapproved 1997) (p. 55). Farmington Hills, MI: American Concrete
gineering, Barranquilla, Colombia; 2017, paper 1750. Institute; 1992.
[12] Rosso A, Almeida JP, Beyer K. Stability of thin reinforced concrete walls under [38] Lowes LN, Lehman DE, Birely AC, Kuchma DA, Marley KP, Hart CR. Earthquake
cyclic loads: state-of-the-art and new experimental findings. Bull Earthq Eng response of slender planar concrete walls with modern detailing. Eng Struct
2016;14(2):455–84. https://doi.org/10.1007/s10518-015-9827-x. 2012;43:31–47. https://doi.org/10.1016/j.engstruct.2012.04.040.
[13] Almeida J, Prodan O, Rosso A, Beyer K. Tests on thin reinforced concrete walls [39] Lu Y, Henry RS, Gultom R, Ma MT. Cyclic testing of reinforced concrete walls with
subjected to in-plane and out-of-plane cyclic loading. Earthq Spectra distributed minimum vertical reinforcement. J Struct Eng 2017;143(5):04016225.
2017;33(1):323–45. https://doi.org/10.1193/101915eqs154dp. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001723.
[14] Rosso A, Jiménez-Roa LA, de Almeida JP, Guerrero AP, Blandón CA, Bonett RL, [40] Moehle JP. Seismic design of reinforced concrete buildings. New York, NY:
et al. Cyclic tensile-compressive tests on thin concrete boundary elements with a McGraw-Hill Professional; 2014.
single layer of reinforcement prone to out-of-plane instability. Bull Earthq Eng [41] Hiraishi H. Evaluation of shear and flexural deformations of flexural type shear
2018;16:859. https://doi.org/10.1007/s10518-017-0228-1. walls. Bull NZ Soc Earthq Eng 1984;17(2):135–44.
[15] Rosso A, Jiménez-Roa LA, de Almeida JP, Beyer K. Instability of thin concrete walls [42] Paulay T, Priestley MJN. Stability of ductile structural walls. ACI Struct J
with a single layer of reinforcement under cyclic loading: numerical simulation and 1993;90:385–92.
improved equivalent boundary element model for assessment. J Earthq Eng 2018. [43] Sozen MA. Earthquake response of buildings with robust walls. In: Paper presented
under review. at the fifth chilean conference on seismology and earthquake engineering, Santiago
[16] McMenamin A. The performance of slender precast reinforced concrete cantilever de Chile; 1989.
walls with roof level lateral displacement restraint under simulated in-plane seismic
loading, Research Report 99-4, Department of Civil Engineering, University of

187