Sunteți pe pagina 1din 12

Materials and Structures (2017)50:232

DOI 10.1617/s11527-017-1102-y

ORIGINAL ARTICLE

Experimental testing of reinforced concrete and reinforced


ECC flexural members subjected to various cyclic
deformation histories
Timothy E. Frank . Michael D. Lepech . Sarah L. Billington

Received: 8 July 2017 / Accepted: 5 October 2017


Ó RILEM 2017

Abstract Engineered Cementitious Composite Keywords Engineered cementitious composite


(ECC) materials have been designed to exhibit high (ECC)  Reinforced concrete  Deformation history 
tensile ductility compared to traditional concrete. Steel reinforcement strain  Energy dissipation 
ECCs have also shown improved damage tolerance Ductility
in compression. When reinforced with steel, ECC
components have been proposed for enhanced seismic
resistance in structural applications. Because of the
uncertainty associated with ground motions, deter- 1 Introduction and motivation
mining an appropriate cyclic deformation history for
seismic testing of structural components is a chal- ECC is a class of High-Performance Fiber-Reinforced
lenge. Three reinforced ECC and three reinforced Cement-based Composite (HPFRCC) that contains
concrete beams were tested under three different many similar constituents to traditional concrete. In
cyclic loading protocols. Cracking, strain in the steel contrast to concrete, however, ECC does not contain
reinforcement, and hysteretic response were moni- coarse aggregate, and includes short, randomly dis-
tored. The reinforced ECC beams exhibited an tributed fibers. ECC is engineered to form multiple,
increase in ductility and hysteretic energy dissipated steady-state cracks under uniaxial tension. On a
over the reinforced concrete beams, particularly when mesoscale, multiple cracking allows an ECC compo-
subjected to a deformation history containing large nent to gain strength beyond the formation of the first
initial deformation pulses. The presence and magni- crack, a phenomenon referred to as pseudo-strain
tude of initial pulses did not affect ductility or failure hardening [1]. Ultimate specimen strain is a function
mode of the ECC beams, and is not expected to be of specimen size, shape, mixture design, and applied
relevant in design of reinforced ECC beams for strain rate [2–4].
collapse prevention. Quasi-static cyclic tests on structural components
have been used to gain a better understanding of steel
reinforced ECC component response to seismic loads.
As ECC with embedded steel reinforcement has been
T. E. Frank (&) proposed and implemented in seismic-resistant struc-
US Air Force, Soto Cano AB, Honduras tural applications [5, 6], laboratory experimentation
e-mail: timothy.frank@us.af.mil
under cyclic loads can be a valuable tool to understand
M. D. Lepech  S. L. Billington
Stanford University, Stanford, CA, USA
232 Page 2 of 12 Materials and Structures (2017)50:232

the material’s limits and capabilities, and to make cracking and strain hardening, and thus, fundamen-
recommendations for design codes. tally change the composite material’s behavior there-
Cost and time typically limit laboratory testing to a after. Typical material characteristics of reinforced
small number of specimens. With a limited test ECCs, such as strong bond to steel reinforcement [15]
regimen, the choice of deformation history to impose and strain hardening of the steel reinforcement at low
on each specimen is important. The seismic demand levels of specimen strain [16], act as potential
on a structure depends on many variables including, vulnerabilities when subjected to large initial cyclic
for example, ground motion characteristics, soil deformations. Response of a steel reinforced ECC
properties, inertial mass, and structural period. There- component subjected to a deformation history con-
fore, selecting a single deformation history will taining initial pulses is expected to be different than
‘‘always be a compromise, but one that should be that of a similar reinforced concrete component.
conservative for most practical cases’’ [7]. Many To address some of these gaps in understanding,
researchers have used a deformation history that direct comparisons of reinforced concrete specimens
begins with small, elastic cycles. Cyclic amplitude is have been made to reinforced ECC or reinforced
then increased every one to three cycles. These cycles HPFRCC specimens in the past. Increases in energy
are imposed on the component in a deformation- dissipated, strength, and ductility, as well as a
controlled fashion until a predetermined failure crite- reduction in transverse steel required, were noted in
rion is reached. reinforced ECC beams relative to reinforced concrete
An understanding of reinforced ECC component beams [17, 18]. In another study, the authors noted a
behavior under a deformation history wherein the higher strain accumulation rate in the steel reinforcing
amplitude of each cycle does not always monotoni- bars, which resulted in failure by fracture of the steel
cally increase is of interest for several reasons. First, in reinforcement at a lower drift in a reinforced HPFRCC
previous research the cracking pattern of a reinforced beam than a reinforced concrete beam [19]. To date,
HPFRCC infill panel subjected to recorded ground however, there has been no study examining the
motions [8] was significantly different than the differences between reinforced concrete and rein-
cracking pattern of a similar component subjected to forced ECC beams considering deformation history as
a monotonically increasing cyclic deformation history an experimental variable. This study is dedicated to
[9]. Additionally, in a preliminary study preceding the exploring how the response of a reinforced ECC
work reported herein, the authors noticed differences flexural member changes when subjected to different
in cracking and strain accumulation in the steel cyclic deformation histories using reinforced concrete
reinforcement when comparing reinforced ECC spec- as a means of comparison.
imen behavior under various deformation histories
[10].
Second, structural components made from tradi- 2 Experimental program
tional reinforced concrete have been found to respond
differently to different deformation histories [e.g., 2.1 Specimen geometry
11–13]. Strength and stiffness deterioration as well as
specimen ductility varied depending on the applied The specimens reported here were approximately
deformation history. quarter-scale of what might be found in a building.
Third, initial deformation pulses from ground Differences in behavior due to differences in size
motions are not uncommon, and simulations using between these specimens and full-scale components
recorded ground motion records suggest they can was expected to be small due to the small size of the
generate story drifts of at least 7% in a generic frame particles and fibers in the ECC materials (8 mm fiber
structure [14]. The effect of initial deformation pulses length was the largest dimension of any mixture
on a reinforced ECC component may be different than constituent). The dimensions of all six specimens were
that on a reinforced concrete component because of nominally identical. Specimen dimensions and the test
the nature of the ECC material itself. Initial pulses setup are shown in Fig. 1. The vertical beam had a
imposed on an ECC component could form a domi- cross-section of 165 mm 9 203 mm. One 13 mm bar
nant crack with fiber pullout, prevent multiple embedded in each corner resulted in a 0.95% steel
Materials and Structures (2017)50:232 Page 3 of 12 232

the mixture design shown in Table 1 and delivered to


the site. The concrete for all the reinforced concrete
specimens came from the same batch.
The three reinforced ECC specimens had mixture
proportions as shown in Table 2. The silica sand had a
0.10 mm effective size. The polyvinyl alcohol (PVA)
fibers were supplied by Kuraray Co. Ltd., and
represented 2% of the composite, by volume. Fibers
were 8 mm long and had a 40 lm diameter. The high
range water reducer was Melflux 1641F, manufac-
tured by the BASF Corporation. The viscosity mod-
ifying admixture was Methocell, manufactured by the
Dow Chemical Company. The ECC was made on site,
and mixed in a horizontal pan mortar mixer in separate
batches for each reinforced ECC specimen.
The three reinforced concrete and three reinforced
ECC specimens were moist cured for 7 days and then
Fig. 1 Geometry, steel reinforcement, and location of internal allowed to cure at room temperature until being tested.
strain gages on each reinforced concrete and reinforced ECC The reinforced ECC specimens were tested on day
specimen 28 ± 2 days. To reduce the difference in estimated
compressive strength between concrete and ECC, the
reinforcement ratio in flexure. Clear cover to the reinforced concrete specimens were allowed to cure
longitudinal bars was 32 mm on all sides. longer, and were tested on day 56 ± 2 days. All six
The enlarged base had a cross-section of specimens were reinforced with deformed mild steel
229 mm 9 381 mm. A total of six, 13 mm horizontal bars with a 13 mm diameter. The reinforcing bars had
bars reinforced the enlarged base. Transverse rein- yield stress, ultimate stress, and strain at ultimate
forcement was designed as seismic detailing per ACI stress of 455, 675 MPa, and 16%, respectively, based
318-14 [20]. Stirrups were spaced at 37.5 mm on the average of two samples, each tested quasi-
throughout the enlarged base and 342 mm into the statically in tension under fixed–fixed grip conditions.
beam. The remainder of the beam stirrups were spaced Smooth, 3.2 mm diameter steel wire was used for
at 75 mm. transverse reinforcement. Two samples of the steel
Two of the four longitudinal reinforcing bars in wire were also tested quasi-statically in tension under
each specimen were instrumented with three strain fixed–fixed grip condition, and exhibited yield stress,
gages for a total of six strain gages on each specimen. ultimate stress, and strain at ultimate stress of 690,
On each instrumented reinforcing bar, one strain gage 820 MPa, and 13%, respectively.
was located within the joint, 5 cm below the joint face. Concrete from the same batch as the test specimens
Two additional strain gages were located on the was used to cast 100 mm 9 200 mm cylinders and
reinforcing bar within the beam, one 5 cm above the 80 mm 9 80 mm 9 305 mm beams. Compression
joint face, and one 15 cm above the joint face (Fig. 1). strength of the concrete was 42 MPa with a standard
Three string potentiometers placed at third points
along the length of the beam recorded displacement,
Table 1 Concrete mixture design (per cubic meter)
and resistive force was recorded by a 50 kN load cell
attached to the hydraulic actuator that displaced the Constituent Mass (kg)
specimen (not shown). Type II/V portland cement 473
1 cm pea gravel 751
2.2 Materials
Sand 828
Water 144
For the three reinforced concrete specimens, the
concrete was provided by a local ready-mix plant per
232 Page 4 of 12 Materials and Structures (2017)50:232

Table 2 ECC mixture design (per cubic meter) 12

Equivalent Bending Stress (MPa)


Constituent Mass (kg)

Type II/V portland cement 547


8
Class F fly ash 656
Silica sand 438
Water 312
PVA fibers 26.0 4
High range water reducer 2.74
Viscosity modifying admixture 0.613
0
0 1 2 3 4
Displacement (mm)
deviation of 3.3 MPa based on tests of four cylinders
per specimen, for a total of twelve samples, tested at
Fig. 2 Third point bending response of three concrete and three
56 ± 2 days. The modulus of rupture of the concrete
ECC beam samples
was 5.5 MPa with a standard deviation of 0.52 MPa
based on three beams per specimen, for a total of nine 4.23 9 10-2 mm/s, and cyclic reversals were con-
specimens, tested in third point bending. trolled by the string potentiometer at the same level as
ECC from each of the three batches was also tested the actuator. Vertical movement at the point of loading
in compression and third point bending. In compres- caused by specimen rotation during each cycle
sion, four cylinders per batch were tested, for a total of resulted in the angle between the string potentiometer
twelve samples. The mean compressive strength of the and the horizontal varying by less than ± 2°. No
batches ranged from 44 to 52 MPa, with an overall adjustments to the drift recorded by the string poten-
mean and standard deviation of 47 and 3.7 MPa, tiometers were made. Testing continued until one of
respectively. The modulus of rupture of the ECC was two failure criteria were met: one of the steel
calculated based on third point bending tests of three reinforcing bars fractured or the strength at the peak
beams per batch, for a total of nine samples. The of an inelastic excursion was less than 75% of yield
modulus of rupture of the batches ranged from 8.7 to strength (about 60–75% of ultimate strength).
10 MPa, with an overall mean and standard deviation
of 9.6 and 1.3 MPa, respectively. Equivalent bending 2.4 Applied deformation histories
stress versus vertical displacement response for three
bending tests of both concrete and ECC show repre- Three deformation histories were used in this study
sentative material response (Fig. 2). (Fig. 3); one each applied to a specimen of each
material. The first deformation history, proposed by
2.3 Methods the Federal Emergency Management Agency in
FEMA 461 [21], is similar to many used in previous
The enlarged base of each specimen was bolted to a research [e.g., 18, 22]. The FEMA deformation history
wide flange steel section, which was fixed to the consists of increasing amplitude steps with two cycles
laboratory strong floor (Fig. 1). Bending out of plane per step. The first amplitude step was 0.15% drift,
was restrained by bracing (not shown). One end of the where drift was measured as beam tip displacement
hydraulic actuator was attached to a test frame and the divided by the 813 mm beam length. Each subsequent
other to the specimen by bolts and steel plates. The amplitude step was 40% larger than the previous
orientation of the actuator varied up to ± 5° during the (Fig. 3a). The first few amplitude steps were not
test as the top of the specimen moved back and forth expected to yield the steel reinforcement.
with each cycle. The recorded force from the load cell The other two deformation histories began with two
was adjusted to account for only the force acting initial pulses followed by the FEMA deformation
perpendicular to the specimen. Axial load was not history. Initial pulses of 2.5% drift and 7% drift were
applied to any specimen. The actuator moved
Materials and Structures (2017)50:232 Page 5 of 12 232

20 2.5 Test register


(a) FEMA

10 The naming convention of each of the six specimens in


this study are given in Table 3. Each specimen name
Drift (%)

indicates the material (either RC for reinforced


0
concrete or ECC for reinforced ECC) and a descriptor
indicating the applied cyclic deformation history. The
-10 deformation history descriptors are either: F for
‘FEMA,’ SP for ‘Small Pulse,’ or LP for ‘Large
-20 Pulse.’
Cycles

20 3 Experimental results and discussion


(b) Small Pulse (SP)
3.1 Cracking
10
Drift (%)

Cracking patterns for the specimens differed more


0 between cementitious material than across deforma-
tion histories, and depictions of each can be found in
-10 Frank [23]. The cracking pattern in RC-F and RC-SP
began as horizontal flexural cracks until about 3.1%
-20 drift, when cracks inclined at approximately 45° and
Cycles formed shear cracks through the specimen’s core. The
same shear cracks occurred in RC-LP, but formed
20 during the two initial pulses. All reinforced concrete
(c) Large Pulse (LP) specimen final cracking patterns show significant
spalling and concrete material degradation (e.g.,
10
Fig. 4a). Inclined cracking was expected in the
Drift (%)

concrete due to the shear stresses developed in the


0 beam, and the spalling was expected due to the
relatively large imposed deflections.
-10 Consistent with previous tests of reinforced ECC
flexural members [17, 24], cracking in the three
reinforced ECC specimens consisted of horizontal
-20
Cycles flexural cracks with no diagonal (shear) cracks (e.g.,
Fig. 4b), owing to inherent shear strength provided by
Fig. 3 a FEMA, and b small pulse, and c large pulse the ECC composite itself. The dominant flexural crack
deformation histories

selected (Fig. 3b, c), with the intention to cause partial Table 3 Specimen naming convention
and total softening in the ECC material, respectively. Specimen name Material Deformation history
The deformation histories beginning with the 2.5 and
7% pulses will be referred to as the ‘Small Pulse’ (SP) RC-F Reinforced concrete FEMA
and ‘Large Pulse’ (LP) deformation histories in this RC-SP Reinforced concrete Small pulse
study. RC-LP Reinforced concrete Large pulse
ECC-F ECC FEMA
ECC-SP ECC Small pulse
ECC-LP ECC Large pulse
232 Page 6 of 12 Materials and Structures (2017)50:232

(a) (b) depth flexural cracks, and residual splitting crack


length varied minimally across the three reinforced
ECC specimens.
Cracking was different between the reinforced
concrete and reinforced ECC specimens. In general,
a 0.1 mm flexural crack formed at about twice the drift
in reinforced ECC (2.2% drift) than in reinforced
concrete specimens (0.81 or 1.1% drift). It was
expected that dominant crack formation in ECC would
be delayed relative to that of concrete due to the
multiple cracking nature of ECC prior to dominant
crack formation. Dominant flexural crack formation
indicated the start of material softening when fibers
began to pull out of the ECC, near the base of the beam.
At drifts beyond dominant flexural crack formation in
Fig. 4 Final cracking patterns of a RC-F and b ECC-F the reinforced ECC specimens, specimen strain con-
centrated at the location of the dominant crack.
in the reinforced ECC specimens, across which fiber There were about twice as many flexural cracks in
bridging was lost during the experimental tests due to reinforced ECC than reinforced concrete specimens,
fiber pullout, occurred at or very near to the top of the but only 25–36% extended through the full depth of
joint where the beam and enlarged base met. the reinforced ECC whereas 87–100% extended
Table 4 presents various cracking observations through the full depth of the reinforced concrete.
from the specimens. There were many similarities Flexural crack widths in the ECC remained less than or
across the three reinforced concrete specimens. equal to 0.1 mm throughout testing except for the
Almost all cracks that formed became full-depth dominant crack. Increases in cyclic amplitude at each
cracks, wherein cracks initiating from both sides step were accommodated by increases in the width of
joined. Further, the number of cracks and full-depth the dominant crack in the reinforced ECC specimens.
cracks were similar across the three reinforced con- Accordingly, the dominant crack width in the rein-
crete specimens. There were many similarities forced ECC specimens measured at 6.1% drift was
between cracking in the three reinforced ECC spec- about 2.5 times as large as the largest crack width in
imens as well. Drift at dominant flexural crack the reinforced concrete specimens. The average
formation, number of flexural cracks, number of full- residual splitting crack length after unloading in

Table 4 Cracking observations in the six specimens


Observed response RC ECC
F SP LP F SP LP

Drift at dominant flexural cracka formation (%) 0.81 0.81 1.1c 2.2 2.2 2.2c
b c
Drift at major splitting crack formation (%) 4.3 4.3 4.0 3.1 7.0 7.0c
Avg. no. flexural cracks per side in bottom 2/3 7 7 8 15 14 16
No. full-depth cracks in bottom 2/3 6 7 8 4 5 4
Residual splitting crack length (cm) 25 41 25 13 14 11
Max. flexural crack width at 6.1% drift (mm) 2.3 3.1 1.9 5.5 6.0 6.6
Max. splitting crack width at 6.1% drift (mm) 0.2 0.5 0.4 0.5 0.1 0.3
a
Dominant crack formed when crack width exceeded 0.1 mm
b
Major splitting cracks formed when crack width reached 0.5 mm
c
Occurred during the first initial pulse
Materials and Structures (2017)50:232 Page 7 of 12 232

reinforced ECC specimens (13 mm) was less than half Figure 5 shows the relationship of steel reinforce-
as long as those in reinforced concrete specimens ment strain versus specimen drift for RC-SP and ECC-
(30 mm) indicating, as expected, greater unbonded SP through the completion of the 4.3% drift cycles,
lengths of steel to the concrete formed than unbonded after which, many strain gages failed. The strain gages
lengths of steel to the ECC. Due to the small sample positioned 5 cm into the joint recorded data further
size and natural variability, effects of different defor- through the testing protocols than the strain gages
mation histories within either cementitious material 5 cm above the joint, and thus were used for compar-
could not be identified by solely observing cracking ison. The strain evolution shown in Fig. 5 were
responses. recorded on the side of the specimens that went into
tension first. The beginning of the first 0.15% drift
3.2 Reinforcement strain cycle, after the conclusion of the two initial pulses, and
the end of the second 4.3% drift cycle are indicated.
Strain in the steel reinforcement localized and accu- When comparing specimens subjected to the SP
mulated differently between specimens of different deformation history, the concrete specimen (Fig. 5a)
material and different deformation history. Steel had lower values of reinforcement strain than the
reinforcement strain was compared between materials reinforced ECC specimen (Fig. 5b) at 4.3% drift; 1.2
when specimens were subjected to the FEMA defor- vs. 1.7% steel reinforcement strain, respectively.
mation history. The reinforced concrete specimen had Between the three reinforced concrete specimens,
lower values of reinforcement strain at a given drift reinforcement strain measurements at the end of the
than the reinforced ECC specimen, e.g., 1.3 vs. 2.7% 4.3% drift step were within 7% of each other,
reinforcement strain measured 5 cm into the joint in indicating strain in the steel reinforcement was not
RC-F and ECC-F, respectively, at 4.3% drift. Less greatly affected by deformation history. Without fiber
reinforcement strain in the reinforced concrete is in confinement, cracking in the concrete led to a similar
agreement with previous researchers’ observations of level of bond degradation and thus, similar steel
reinforcement strain [16], and can be attributed to the reinforcement strain at 4.3% drift, regardless of the
wider flexural cracks and longer splitting cracks that applied deformation history. Between the reinforced
developed in the concrete than in the ECC. Wider and ECC specimens, however, reinforcement strain at the
longer cracks in concrete than ECC led to longer end of the 4.3% drift step between specimens
debonded lengths of the steel reinforcement, and thus subjected to different deformation histories were
the reinforcement strain at the any discrete strain gage significantly different. For example, ECC-SP had
location was lower. In contrast, the matrix-fiber 38% less strain in the steel reinforcement than ECC-F
composite design of ECC, which gives it enhanced at 4.3% drift. While this observation could partially be
tensile strength and ductility, improved the specimen’s due to slight differences in the location of the strain
ability to resist the cracking. gage relative to the dominant crack, it also indicates

Fig. 5 Reinforcement (a) (b)


strain versus drift measured 4 4
Reinforcement Strain (%)

Reinforcement Strain (%)

by a strain gage located


5 cm into the joint in a RC-
SP and b ECC-SP 3 3

2 2

1 1

0 0
-5.0 -2.5 0.0 2.5 5.0 -5.0 -2.5 0.0 2.5 5.0
Drift (%) Drift (%)
Start of 0.15% drift cycles
End of 4.3% drift cycles
232 Page 8 of 12 Materials and Structures (2017)50:232

that bond to the steel reinforcement in ECC may be ECC [26], as indicated by the lack of significant
more sensitive to deformation history than in concrete. crushing and spalling observed, facilitated peak
Bandelt and Billington [25] found that bond strength in the reinforced ECC specimens at compar-
strength of steel reinforcement to ECC was 39% atively lower levels of drift than the reinforced
higher than bond strength of steel reinforcement to concrete specimens (2% drift compared to 4–8.5%
concrete. Further, bond toughness of reinforced ECC drift, respectively).
beams was higher than that of reinforced concrete Figure 6 shows the hysteretic response of the RC-
beams, indicating that the steel reinforcing bars in LP and ECC-LP. Additionally, the point in the
reinforced ECC beams retained residual bond strength deformation history when splitting cracks grew to
at larger values of reinforcing bar slip than the steel 0.5 mm wide, crushing, and spalling were first
reinforcing bars in reinforced concrete beams. In the observed on both sides of each specimen, as applica-
flexural specimens tested herein, the relatively high ble, is shown. The hysteretic response of the rein-
bond toughness at the steel-ECC interface within the forced concrete specimens (e.g., Fig. 6a) showed more
reinforced ECC beams facilitated less bond degrada- pinching, or lower reloading stiffness, than those of
tion when subjected to initial deformation pulses than the reinforced ECC specimens (e.g., Fig. 6b). The
the comparatively low bond toughness at the steel– presence of the initial pulses in the SP and LP
concrete interface within the reinforced concrete deformation histories, however, generated more
beams. To illustrate this point, strain in the steel pinching than the FEMA deformation history in both
reinforcing bars of RC-SP did not increase between reinforced concrete and reinforced ECC specimens.
the first and second cycle to 4.3% drift (Fig. 5a). In Increased lateral translation facilitated by cementi-
contrast, due to the lower rate of bond degradation tious material damage and bar elongation during the
with increasing drift, strain in the steel reinforcing bars initial pulses led to the increased pinching with the
of ECC-SP increased from 3.1 to 3.3% strain (6.5% presence and size of initial pulses in the deformation
increase) between the first and second cycle to 4.3% history.
drift (Fig. 5b). Ductility is perhaps the most critical characteristic
of a structure needed to withstand a large earthquake
3.3 Hysteretic response and specimen ductility and prevent collapse. The ductility of reinforced ECC
specimens was less sensitive to 7% initial pulses than
Hysteretic response of the specimens can be found in the ductility of reinforced concrete specimens. Spec-
Frank [23]. The difference in peak specimen strength imens subjected to the FEMA and SP deformation
(28–30 kN for reinforced concrete specimens and histories made from both concrete and ECC achieved
30–32 kN for reinforced ECC specimens) was a result ductility of 12% drift, meaning the peak drift achieved
of the differences in tension and compression proper- in either direction before failure was 12% drift. When
ties between the concrete and ECC as reported in subjected to the LP deformation history, the reinforced
Sect. 2.2, e.g., mean compressive strengths of the ECC specimen also achieved 12% drift, whereas the
concrete and ECC were 42 and 47 MPa, respectively. reinforced concrete specimen reached a peak drift of
The point along the deformation history when peak 7% drift, a 42% reduction. The number of complete
strength was achieved, however, was also different excursions each specimen sustained was between 54
between reinforced concrete and reinforced ECC and 56 except RC-LP, which failed after completing
specimens. RC-F achieved peak strength during the 48 excursions.
8.5% drift cycles, whereas RC-LP achieved peak Despite sustaining damage through cracking and
strength at 4% drift during the first excursion as the some crushing and spalling, the core of ECC-LP
concrete on the compression side of the specimen was remained intact due to the relatively high shear
undamaged. Peak strength in the hysteretic envelope strength and damage tolerance of the ECC material.
of all three reinforced ECC specimens was achieved The ECC in ECC-LP was able to provide sufficient
near 2% drift regardless of the deformation history. confinement to the longitudinal steel reinforcement
The high strains that developed in the steel reinforce- and facilitate failure by fracture of the reinforcement,
ment led to strain hardening near 2% drift. This strain the same failure mode as ECC-F and ECC-SP. The
hardening, coupled with the high damage tolerance of extra plastic deformation due to the two 7% initial
Materials and Structures (2017)50:232 Page 9 of 12 232

Fig. 6 Hysteretic response (a) (b)


for a RC-LP and b ECC-LP 40 40
with icons indicating
splitting cracks at least
20 20
0.5 mm wide, crushing,

Force (kN)

Force (kN)
splitting, spalling, and
failure modes 0 0

-20 -20

-40 -40
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
Drift (%) Drift (%)

Failure by reinforcing bar fracture Crushing


Failure by cementitious material crushing Spalling
Splitting ≥ 0.5 mm wide

pulses was not a significant contribution to strain energy dissipated was determined by calculating the
accumulation in the steel reinforcement at the critical area within each hysteretic loop using the trapezoidal
section (or the strain accumulation was accompanied rule of numerical integration at steps of 0.03% drift,
by bond degradation which reduced strain), and did then normalized by the area of an equivalent elasto-
not lead to earlier bar fracture. In contrast, all three plastic hysteresis loop to the same peak force and
reinforced concrete specimens failed due to concrete displacement per [28] for each specimen. This proce-
crushing, and thus the crushing that occurred early in dure produced a hysteretic energy dissipation ratio for
the deformation history of RC-LP directly and signif- each complete hysteretic loop [29]. The hysteretic
icantly contributed to the cumulative damage that energy dissipation ratio during the first full cycle to
limited the specimen’s ultimate ductility. select drifts is shown in Fig. 7.
Because ductility of the reinforced ECC specimens Due to wider hysteretic loops, the hysteretic energy
was less sensitive to deformation history than rein- dissipation ratio was larger at every level of drift for
forced concrete specimens, reinforced ECC design for the reinforced ECC specimens than the reinforced
collapse prevention may be more consistent between concrete specimens subjected to the same deformation
situations where initial pulses are not expected and history. The hysteretic energy dissipation ratio of all
situations where initial deformation pulses may be specimens tended to increase with drift, however, it
more likely, such as near-fault locations. Due to increased at a higher rate in reinforced ECC than in
concrete crushing and subsequent loss of confinement reinforced concrete specimens. The higher rate of
to steel reinforcing bars, design for collapse preven- energy dissipation per cycle is attributed to higher
tion of reinforced concrete specimens should depend strain in the steel reinforcement, as also noted in [17].
on the expected presence and size of initial deforma- At 2.2% drift, the hysteretic energy dissipation
tion pulses. This finding supports previous research of ratios of specimens subjected to the SP and LP
reinforced concrete structures and their seismic col- deformation histories were 25–37% less than that in
lapse risk [e.g., 27]. the specimen of the same cementitious material
subjected to the FEMA deformation history. The
3.4 Energy dissipation reduced hysteretic energy dissipation ratio at 2.2%
drift occurred because cracks that formed during the
Damage from cracking and plastic deformation of the initial pulses did not dissipate energy at drifts less than
steel reinforcement contributed to the energy dissi- the amplitude of the initial pulses. At drifts larger than
pated by each specimen from the beginning of the test the initial pulse amplitude, the presence of initial
through specimen failure. To compare energy dissi- deformation pulses had less of an effect on the
pation among specimens with different strengths, hysteretic energy dissipation ratio for specimens made
232 Page 10 of 12 Materials and Structures (2017)50:232

Fig. 7 Hysteretic energy 0.6


dissipation of the first full
cycle relative to an

Hysteretic Energy Dissipation Ratio


equivalent elasto-plastic 0.5
hysteresis loop to the same
peak force and displacement
0.4

0.3

0.2

0.1

0.0
2.2 4.3 6.1 8.5
Drift (%)

RC-F ECC-F RC-SP ECC-SP RC-LP ECC-LP

of both reinforced ECC and reinforced concrete. For following initial deformation pulses were not as
example, at 4.3% drift the hysteretic energy dissipa- effective in dissipating energy as in a reinforced
tion ratio of ECC-SP was within 1.3% of ECC-F and ECC specimen. The ability of a reinforced concrete
the hysteretic energy dissipation ratio of RC-SP was component to dissipate energy during an aftershock is
within 4.5% of RC-F. expected to be more strongly dependent on the main
As drift increased beyond 4.3%, data indicate that shock’s deformation history and whether or not it
the size of the initial deformation pulse had less of an contained initial deformation pulses, than a reinforced
effect on the hysteretic energy dissipation ratio in ECC component.
reinforced ECC than in reinforced concrete speci-
mens. The hysteretic energy dissipation ratio of ECC-
LP was within 7.1% of ECC-F at 6.1% drift, whereas 4 Conclusions
the hysteretic energy dissipation ratio of RC-LP was
17% lower than RC-F. Fiber bridging across the Three reinforced concrete and three reinforced ECC
narrow crack widths in the reinforced ECC specimens specimens were tested under three deformation histo-
subjected to initial pulses facilitated greater strain ries. Differences in cracking, reinforcement strain,
accumulation in the reinforcing bars (Fig. 5), and thus hysteretic response, failure mode, specimen ductility,
steel plasticity made a greater contribution to energy and energy dissipation were observed between cemen-
dissipation than in the reinforced concrete specimens. titious materials and between deformation histories.
Results show that energy dissipated per cycle by a Cracking patterns were visibly different between
reinforced ECC component may be significantly lower materials. Cracks in the reinforced ECC specimens
during small cycles following initial deformation were narrower and more horizontal (indicating flex-
pulses when compared to cycles of the same amplitude ure) than cracks in the reinforced concrete specimens.
that are part of a monotonically increasing deforma- Spalling was more prominent in the reinforced con-
tion history. Substantial capacity exists, however, to crete specimens, whereas the fibers in the reinforced
dissipate energy if large ductility demands are ECC specimens prevented spalling. The narrower
imposed, for example from an aftershock. In a cracks in the reinforced ECC specimens facilitated
reinforced concrete specimen, however, large cycles better bond to the steel reinforcement than in the
Materials and Structures (2017)50:232 Page 11 of 12 232

reinforced concrete specimens. The increased bond conclusions from the experiments reported herein
toughness in the reinforced ECC specimens led to provide a fundamental understanding of the response
more rapid strain accumulation. Strain in the steel of reinforced ECC flexural members subjected to three
reinforcement was higher at a given drift in ECC than different deformation histories, and highlights differ-
in concrete and, for both materials, was higher at a ences in response from reinforced concrete beams.
given drift when subjected to a deformation history
without initial deformation pulses than with initial Acknowledgements This study was funded by the Air Force
Institute of Technology, the John A. Blume Earthquake
pulses.
Engineering Center, and the Thomas V. Jones Engineering
The strength of the reinforced ECC specimens was Faculty Scholarship at Stanford University.
highest near 2% drift, as the reinforcing steel strain
hardened, then decreased slightly with drift. In con- Compliance with ethical standards
trast, peak strength in the reinforced concrete speci-
Conflict of interest The authors declare that they have no
mens occurred between 4 and 8.5% drift. Pinching of conflicts of interest. The views expressed in this dissertation are
the hysteresis loops was more pronounced in rein- those of the author and do not reflect the official policy or
forced concrete than reinforced ECC, and occurred to position of the United States Air Force, Department of Defense,
or the U.S. Government.
a greater extent in specimens subjected to initial
deformation pulses. When compared to reinforced
concrete, the ductility of reinforced ECC specimens
was less sensitive to the deformation history applied. References
The high damage tolerance in compression demon-
1. Maalej M, Li VC (1995) Introduction of strain-hardening
strated by the ECC relative to concrete facilitated the engineered cementitious composites in design of reinforced
relative indifference in specimen ductility between the concrete flexural members for improved durability. ACI
FEMA deformation history, which consisted of mono- Struct J 92(2):167–176
2. Douglas KS, Billington SL (2010) Strain rate dependence of
tonically increasing cycles, and a deformation history
HPFRCC cylinders in monotonic tension. Mater Struct
consisting of two initial pulses to either 2.5 or 7% drift 44(1):391–404
followed by the FEMA deformation history. The 3. Kanda T, Li VC (1999) New micromechanics design theory
reinforced concrete specimen subjected to a deforma- for pseudostrain hardening cementitious composite. J Eng
Mech 125(4):373–381
tion history containing two initial pulses to 7% drift
4. Mechtcherine V, Silva F, Butler M, Zhu D, Mobasher B,
experienced concrete crushing during the two initial Gao S, Mäder E (2011) Behaviour of strain-hardening
pulses. The crushing of the concrete prevented the cement-based composites under high strain rates. J Adv
specimen from retaining strength during subsequent Concr Technol 9(1):51–62
5. Kanda T, Watanabe S, Li VC (1998) Application of pseudo
cycles with amplitudes exceeding 7% drift. Of the
strain hardening cementitious composites to shear resistant
three reinforced concrete specimens experimentally structural elements. In: Mihashi H, Rokugo K (eds) Proc
tested, only the specimen subjected to the LP defor- FRAMCOS-3, fracture mechanics of concrete structures,
mation history failed at a drift below 12% drift, with vol III. Aedificatio Publishers, pp 1477–1490
6. Li VC, Kanda T (1998) Engineered cementitious compos-
failure due to strength loss occurring after an ultimate
ites for structural applications. ASCE J Mater Civil Eng
drift of 7% was reached. 10(2):66–69
An increased number of cracks and increased 7. Krawinkler H (1996) Cyclic loading histories for seismic
plastic deformation of the steel reinforcement in the experimentation on structural components. Earthq Spectra
12(1):1–12
reinforced ECC specimens relative to the reinforced
8. Lignos DG, Moreno DM, Billington SL (2004) Seismic
concrete specimens led to differences in normalized retrofit of steel moment-resisting frames with high-perfor-
energy dissipated per cycle. Reinforced ECC flexural mance fiber-reinforced concrete infill panels: large-scale
components can be expected to dissipate energy hybrid simulation experiments. J Struct Eng 140:1–13
9. Hanson J, Billington SL (2009) Cyclic testing of a ductile
during an aftershock more effectively than reinforced
fiber-reinforced concrete infill panel system for seismic
concrete components, especially if the main shock retrofitting of steel frames. In: Report No. TR 177, The John
generated initial deformation pulses. Total energy A. Blume Earthquake Engineering Center, Stanford Univ.,
dissipation of the reinforced ECC specimens through Stanford, CA
10. Frank TE, Lepech, MD, Billington SL (2015) Effect of
failure was less sensitive to deformation history than
deformation history on steel-reinforced HPFRCC flexural
that of the reinforced concrete specimens. The
232 Page 12 of 12 Materials and Structures (2017)50:232

member behavior. In: Reinhardt HW, Parra-Montesinos GJ, performance Fiber reinforced cement composites,
Garrecht H (eds) 7th RILEM workshop on high perfor- pp 329–337
mance fiber reinforced cement composites, pp 367–374 20. ACI Committee, American Concrete Institute and Interna-
11. Hwang T (1982) Effects of variation in load history on tional Organization for Standardization (2014) Building
cyclic response of concrete flexural members. Dissertation, code requirements for structural concrete (ACI 318-14) and
University of Illinois commentary
12. Kawashima K, Koyama T (1988) Effect of number of 21. FEMA 461 (2007) interim testing protocols for determining
loading cycles on dynamic characteristics of reinforced the seismic performance characteristics of structural and
concrete bridge pier columns. Struct Eng Earthq Eng nonstructural components
5(1):205–213 22. Parra-Montesinos GJ, Chompreda P (2007) Deformation
13. Ohno T, Nishioka T (1984) An experimental study on capacity and shear strength of fiber-reinforced cement
energy absorption capacity of columns in reinforced con- composite flexural members subjected to displacement
crete structures. Proc JSCE Struct Eng Earthq Eng reversals. J Struct Eng 133(3):421–431
1(2):23–33 23. Frank TE (2017) Response of reinforced engineered
14. Baker JW, Cornell CA (2008) Vector-valued intensity cementitious composite flexural members subjected to
measures for pulse-like near-fault ground motions. Eng various cyclic deformation histories. Dissertation, Stanford
Struct 30(4):1048–1057 University
15. Bandelt MJ, Billington SL (2014) Monotonic and cyclic 24. Bandelt MJ (2015) behavior, modeling, and impact of bond
bond-slip behavior of ductile high-performance fiber-rein- in steel reinforced high-performance fiber-reinforced
forced cement-based composites. In Schlangan E, Sierra cement-based composites. Dissertation, Stanford University
Beltran MG, Lukovic M, Ye G (eds) Proc 3d international 25. Bandelt MJ, Billington SL (2016) Bond behavior of steel
RILEM conference on strain hardening cementitious com- reinforcement in high-performance fiber-reinforced
posites (SHCC3), pp 1–8 cementitious composite flexural members. Mater Struct
16. Moreno DM, Trono W, Jen G, Ostertag C, Billington SL 49(1):71–86
(2014) Tension stiffening in reinforced high performance 26. Li VC (1997) Damage tolerance of engineered cementitious
fiber reinforced cement-based composites. Cement Concr composites. In: Karihloo BL, Mai YW, Ripley MI, Ritchie,
Comp 50:36–46 RO (eds) Advances in fracture research, proceedings 9th
17. Fischer G, Li VC (2002) Effect of matrix ductility on ICF conference on fracture, pp 619–630
deformation behavior of steel-reinforced ecc flexural 27. Champion C, Liel A (2012) The effect of near-fault direc-
members under reversed cyclic loading conditions. ACI tivity on building seismic collapse risk. Earthq Eng Struct
Struct J 99(6):781–790 Dyn 41(10):1391–1409
18. Yuan F, Pan J, Dong L, Leung CKY (2013) mechanical 28. Lequesne RD (2011) Behavior and design of high-perfor-
behaviors of steel reinforced ECC or ECC/concrete Com- mance fiber-reinforced concrete coupling beams and cou-
posite beams under Reversed cyclic loading. J Mater Civil pled-wall systems. Dissertation, The University of
Eng 26(8):1–8 Michigan
19. Tavallali H, Lepage A, Rautenberg J, Pujol S (2012) drift 29. Olsen E, Billington SL (2011) Cyclic response of precast
limits of concrete frame members reinforced with high- high-performance fiber-reinforced concrete infill panels.
performance steel bars and fibers. In: Parra-Montesinos G, ACI Struct J 108(1):51–60
Reinhardt H, Naaman A (eds) 6th RILEM workshop on high

S-ar putea să vă placă și