Experimental Testing of Reinforced ECC Beams Subjected

to Various Cyclic Deformation Histories

Timothy E. Frank, Ph.D., P.E. 1; Michael D. Lepech, Ph.D. 2; and Sarah L. Billington, Ph.D., M.ASCE 3

Abstract: Steel-reinforced engineered cementitious composite (ECC) members have demonstrated enhanced seismic performance in struc-
tural components and systems such as coupling beams, infill panels, joints, columns, and beams. Because a large pulse in a deformation
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

history may cause fiber pullout within the ECC and alter material-level behavior, the response of reinforced ECC components subjected to
deformation histories that contain initial pulses are of particular interest. Reinforced ECC beams of various steel reinforcement ratios and
reinforcing bar sizes were experimentally subjected to one of three deformation histories. The presence and size of initial deformation pulses
affected cracking, strain development in the steel reinforcement, and hysteretic response, while the failure mode of the specimens was con-
sistently fracture of the steel reinforcing bars. Reductions in steel reinforcement strain caused by bond degradation at the steel–ECC interface
facilitated, in general, no change in ultimate structural ductility between similar specimens, regardless of deformation history, except spec-
imens containing the lowest steel reinforcement ratio, 0.73% in flexure, which had a comparatively high bond capacity relative to the bond
demand at the steel–ECC interface. DOI: 10.1061/(ASCE)ST.1943-541X.0002034. © 2018 American Society of Civil Engineers.
Author keywords: Engineered cementitious composite (ECC); Cyclic deformation history; Deformation pulse; Structural ductility;
Steel reinforcement strain.

Introduction Chompreda 2007; Yuan et al. 2012). However, the effect of defor-
mation history on reinforced ECC specimens is unclear due to lack
Engineered cementitious composite (ECC) is a class of high- of test data.
performance fiber-reinforced cement-based composite (HPFRCC) In two separate studies of reinforced HPFRCC infill panels with
materials containing a low volume fraction (typically 2% by vol- steel fibers, one was subjected to a monotonically increasing defor-
ume) of polymeric fibers in a mortar mixture. ECC has the ability to mation history (Hanson and Billington 2009) and others were in-
pseudo strain harden under uniaxial tension, and on a microscale stalled on a steel moment frame and subjected to recorded ground
ECC materials form multiple microcracks when strained (Maalej motions records (Lignos et al. 2014). Results showed the cracking
and Li 1995). On a mesoscale, ECC materials behave like an elasto- patterns within the HPFRCC materials between infill panels were
plastic solid due to fiber bridging across the microcracks, which significantly different. These differences could be attributed to de-
facilitates specimen ductility. ECC specimens of various sizes, formation history, fiber properties, mortar mixture designs, or age
shapes, and mix designs achieve specimen strains between 0.25 of specimens. In a pilot study of steel-reinforced ECC beams, also
and 5% under uniaxial tension (Douglas and Billington 2010; reported herein, Frank et al. (2015) found differences were ob-
Kanda and Li 1999; Maalej et al. 1995; Mechtcherine et al. 2011). served in the cracking pattern and strain accumulation in the steel
ECC has been proposed for a number of different structural reinforcement when specimens were subjected to different defor-
applications, including improving performance under seismic load- mation histories. The results of the pilot study warranted further
ing (Li and Kanda 1998). In seismic research, quasi-static cyclic exploration using a larger specimen sample size.
tests are conducted on structural components to gain a better The goal of the research presented herein is to understand how
understanding of response to seismic loads. It is common to use the response of reinforced ECC beams changes as reinforcement
a cyclic deformation history that begins with small, elastic cycles; ratio and bar layout vary when experimentally subjected to one
the cyclic amplitude typically increases every one to three cycles, of three different deformation histories. Varying reinforcement ra-
and cycles are imposed on the component until a predetermined fail- tios are of interest because as the steel reinforcement ratio in a
ure criterion is reached (Fischer and Li 2002; Parra-Montesinos and reinforced ECC flexural member decreases, the tensile strain at
the extreme fiber will increase at a given curvature. A beam with
1
Civil Engineer, Headquarters Air Force, 1260 Air Force Pentagon, a smaller reinforcement ratio exhibits larger tensile strain in the
Washington, DC 20330 (corresponding author). E-mail: timothy.frank@ extreme tensile fiber and a smaller area over which compression
us.af.mil acts than a beam with a larger reinforcement ratio. Larger tensile
2
Associate Professor, Dept. of Civil and Environmental Engineering, strains are expected to lead to fiber pullout or rupture at lower drifts,
Stanford Univ., Environment and Energy Bldg., 473 Via Ortega, Stanford, causing earlier strain localization in the steel reinforcement in a
CA 94305. E-mail: mlepech@stanford.edu reinforced ECC beam with a lower reinforcement ratio when com-
3
Professor, Dept. of Civil and Environmental Engineering, Stanford pared with a higher reinforcement ratio. In previous experiments,
Univ., Environment and Energy Bldg., 473 Via Ortega, Stanford, CA
flexural testing of reinforced ECC beams of various steel reinforce-
94305. E-mail: billingt@stanford.edu
Note. This manuscript was submitted on July 8, 2017; approved on ment ratios showed the number of flexural and splitting cracks re-
November 7, 2017; published online on March 30, 2018. Discussion duced as steel reinforcement ratio reduced (Bandelt and Billington
period open until August 30, 2018; separate discussions must be submitted 2016). The lack of distributed cracking in reinforced ECC beams
for individual papers. This paper is part of the Journal of Structural with comparatively low steel reinforcement ratios indicated the
Engineering, © ASCE, ISSN 0733-9445. high tensile strain at the extreme fiber was concentrated near

© ASCE 04018052-1 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 Table 1. Specimens Tested in This Study bar size decreases the ratio of total bar perimeter to total steel
Steel Number cross-sectional area increases. The reinforcing bar perimeter is pro-
Specimen reinforcement Reinforcing bar of bars, Deformation portional to bond capacity because it represents the surface area of
name ratio (%) diameter (mm) each side history the bar in physical contact with the ECC. The cross-sectional area
ECC-0.73-F 0.73 10 2 FEMA 461 of steel is proportional to bond demand because it relates to the
ECC-0.73-SP 0.73 10 2 Small pulse amount of stress the bar can develop. Thus, reinforced ECC spec-
ECC-0.73-LP 0.73 10 2 Large pulse imens reinforced with more, smaller reinforcing bars were expected
ECC-0.95-F 0.95 13 2 FEMA 461 to demonstrate the ability to maintain a strong bond with the ECC
ECC-0.95-SP 0.95 13 2 Small pulse material through larger drifts than specimens with fewer, larger
ECC-0.95-LP 0.95 13 2 Large pulse reinforcing bars.
ECC-1.0-F 1.0 10 4 FEMA 461 Not considered in Eq. (1) is that a greater number of reinforcing
ECC-1.0-SP 1.0 10 4 Small pulse bars in a section will reduce the bar spacing. A reduction in bar
ECC-1.0-LP 1.0 10 4 Large pulse
spacing may help limit crack widths through improved crack con-
ECC-1.3-F 1.3 13 2 FEMA 461
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

ECC-1.3-SP 1.3 13 2 Small pulse trol, but smaller bar spacing provides less confinement around the
ECC-1.3-LP 1.3 13 2 Large pulse bars. Due to a stronger bond between steel and the mortar matrix,
ECC-1.4-F 1.4 13 3 FEMA 461 steel-reinforced strain-hardening cementitious composite materials
ECC-1.4-SP 1.4 13 3 Small pulse require less bar spacing than steel-reinforced concrete (Kanakubo
ECC-1.4-LP 1.4 13 3 Large pulse and Hosoya 2015), but less confinement may facilitate interbar
ECC-1.5-F 1.5 16 2 FEMA 461 splitting, which would reduce bond strength and affect specimen
ECC-1.5-SP 1.5 16 2 Small pulse response. Other important seismic response parameters of struc-
ECC-1.5-LP 1.5 16 2 Large pulse tural beams, for example, shear span to depth ratio, were not ex-
perimental variables in this study.

the critical cross section. The result of high tensile strain at the criti-
cal cross section was that ECC beams reinforced to 0.70% steel Experimental Program
reinforcement ratio or less failed by reinforcement fracture at
low ultimate drifts, from 3 to 5%, relative to beams with higher Specimen Geometry and Instrumentation
reinforcement ratios (Bandelt and Billington 2016).
The second variable of interest is reinforcing bar diameter, along Flexurally dominant ECC cantilever beam specimens were selected
with steel reinforcement ratio. Both affect bond to the surrounding for this study. Eighteen specimens were monolithically cast with
ECC. The ratio of total reinforcing bar perimeter to total cross- an enlarged base to represent the half-span of a beam framing
sectional area is a function of the reinforcing bar diameter, db into a column. The naming convention of each specimen includes
[Eq. (1)] information about the material (ECC), the steel reinforcement
ratio (one of six values between 0.73 and 1.5%), and the applied
PNumber of bars deformation history (Table 1). In all specimens, the longitudinal
barperimeter πdb 4
¼ PNumber
i¼1
2
≡ ð1Þ reinforcement was mild deformed steel bars, and the transverse
bararea i¼1
of bars
πd b =4 d b reinforcement was smooth steel wire. The specimens were cast
in one of two sizes. Six specimens were cast per Geometry A
As reinforcement ratio increases due to increased reinforcing [Fig. 1(a)], and had a flexural cross section of 127 × 178 mm. Clear
bar diameter, the ratio of total bar perimeter to total steel cross- cover to the longitudinal bars was 1.9 cm. Stirrups were spaced
sectional area decreases. For a constant reinforcement ratio, as 75 mm apart throughout the beam and into the joint. Symmetric

Fig. 1. (a) Geometry A and (b) Geometry B used to cast reinforced ECC beams

© ASCE 04018052-2 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 reinforcement was provided by two reinforcing bars on each side of Table 3. Steel Reinforcement Bar Properties
the section. Three specimens reinforced with two 10-mm bars on Yield Ultimate Strain at
each side had a 0.73% steel reinforcement ratio in flexure, and three Bar diameter strength, strength, ultimate strength,
specimens reinforced with two 13-mm bars on each side had a (mm) f y (MPa) f u (MPa) εu (%)
1.3% steel reinforcement ratio in flexure. The span length to the 3.2 690 820 13
point of loading for Geometry A specimens was 760 mm. 10 445 690 18
Three specimens of each of four reinforcement ratios were cast 13 455 675 16
per Geometry B [Fig. 1(b)], and had a flexural cross section of 16 440 625 15
165 × 203 mm. Clear cover to the longitudinal bars was 3.2 cm.
Following seismic detailing per ACI 318-14 (ACI 2014), stirrups
were spaced at 37.5 mm through the joint and the plastic hinge of four 100 × 200 mm cylinders per specimen. No significant dif-
region of the beam. Beyond a distance of 342 mm from the joint, ference in compressive behavior or strength was recorded between
stirrup spacing was reduced to 75 mm. Two 13-mm and two 16-mm samples containing the 8- or 12-mm fibers. Flexural response of
80 × 80 × 305 mm ECC beams was determined through third-
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

bars on each side provided flexural steel reinforcement ratios of

0.95 and 1.5%, respectively, to these specimens. Four 10-mm point bending tests. The modulus of rupture of the ECC was
reinforcing bars on each side provided a flexural reinforcement ra- 10 MPa with a standard deviation of 1.7 MPa, based on the average
tio of 1.0%, and three 13-mm reinforcing bars on each side pro- of three beams per specimen tested at the same age as the test spec-
vided a flexural reinforcement ratio of 1.4%. The span length to imens. The deflection at which a 0.1-mm crack formed, initiating
the point of loading for Geometry B specimens was 813 mm. fiber pullout, was consistent between third-point bending test
Geometry B specimens also had a wider and taller enlarged base samples containing both 8- and 12-mm fibers; however, specimens
that was more heavily reinforced than that of Geometry A. The with 12-mm fibers sustained approximately 7.9% more deflection
increase in joint reinforcement was intended to facilitate failure before equivalent bending stress dropped to 10% of the modulus of
within the beam portion of the specimens, while limiting the impact rupture. A significant difference in structural performance of rein-
of specimen geometry on other response parameters of interest. Six forced ECC specimens was not anticipated based on the observed
strain gauges were affixed to the steel reinforcement within each difference in ECC bending response. Additional material testing
specimen: three each on a vertical reinforcing bar on either details are given in Frank (2017).
side. The locations of the strain gauges were in the joint 5 cm Yield strength (f y ), ultimate strength (f u ), and strain at ultimate
below the joint face and 5 and 15 cm above the joint face strength (εu ) of the various types of steel reinforcement used in this
[Figs. 1(a and b)]. study are shown in Table 3, based on the average of two samples
each tested quasi-statically in tension under fixed-fixed grip con-
ditions. Yield strength was as expected based on manufacturer data.
Materials
The mixture design for the ECC is shown in Table 2. The sand had Test Setup and Instrumentation
an effective size of 0.10 mm. The polyvinyl alcohol (PVA) fibers
were supplied by Kuraray (Tokyo, Japan) and were 12 mm long for During testing, the beam was oriented in the vertical direction. Each
specimens of Geometry A and, due to material availability, 8 mm end of the enlarged base was clamped to a steel wide-flange section,
long for specimens of Geometry B. All fibers were 40 μm in diam- which was bolted to a laboratory strong floor [Figs. 1(a and b)].
eter, and the fiber content was 2% by volume. Because crack widths One end of a 50-kN hydraulic actuator was bolted to a reaction
of more than 0.1 mm typically initiated fiber pullout in this ECC frame and the other was bolted to steel plates on either side of
mixture design, the different lengths of PVA fibers used between the specimen. The actuator was furnished with a load cell. The ori-
specimens were expected to facilitate comparable response at the entation of the actuator varied slightly (up to 5°) during the test
structural component-level. The high-range water reducer was Mel- because the top of the specimen moved back and forth with each
flux 1641 F manufactured by the BASF Corporation (Ludwigshafen, cycle. The recorded force from the load cell was adjusted to
Germany). The viscosity modifying admixture was Methocel manu- account for only the force acting perpendicular to the specimen.
factured by the Dow Chemical Company (Midland, Michigan). The specimens were instrumented with three string potentiom-
The ECC was made using a horizontal pan mortar mixer, cast eters. The string potentiometer at the point of loading controlled the
into wooden forms in which the longitudinal axis of the cantilever actuator’s movement, and two other string potentiometers were
beam was parallel with the casting floor, and specimens were moist equally spaced at an interval length equal to one-third of the beam’s
cured under wet burlap for 7 days. Specimens were then air cured span. All tests were conducted under displacement-controlled con-
at room temperature for 28  2 days, when testing took place. ditions at a rate of 4.23 × 10−2 mm=s. Displacement was limited to
The ECC achieved an average compressive strength of 44 MPa the plane of loading by lateral bracing positioned approximately
with a standard deviation of 4.7 MPa at 28 days based on tests 90 cm up from the base of the specimen. The tests were paused
periodically to inspect the specimen for cracks, take photographs,
and record information.
Table 2. ECC Mixture Design per Cubic Meter
Constituent Mass (kg) Deformation Histories
Type II/V portland cement 547 Three deformation histories were used in this study. The deforma-
Class F fly ash 656 tion history proposed in FEMA 461 (FEMA 2007), hereafter re-
Silica sand 438 ferred to as the FEMA deformation history, was applied to
Water 312 one specimen of each of the six steel reinforcement ratios
PVA fibers 26.0 [Fig. 2(a)]. The FEMA deformation history was composed of two
High-range water reducer 2.74 cycles per amplitude step; each step was 40% larger in amplitude
Viscosity-modifying admixture 0.613
than the previous one. The first amplitude step was 0.15% drift,

© ASCE 04018052-3 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 20 Table 4. Mean Values of Various Cracking Responses by Deformation
History
10
Deformation history
Drift (%) 0 Response FEMA SP LP
Drift at dominant flexural cracka formation (%) 2.1 2.2 2.3
-10 Drift at last new flexural crack formation (%) 3.5 2.5 —b
Drift at major splitting crackc formation (%) 5.9 5.7 6.9
-20 Number of flexural cracks per side in 12 12 11
(a) Cycles bottom third
Number of flexural cracks per side in 17 18 17
20 bottom two-thirds
Number of full-depth flexural cracks 3.0 3.5 4.0
10 Dominant crack width at 6.1% drift (mm) 5.2 5.6 6.9
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

Drift (%)

Residual splitting crack length (cm) 17 16 17

0 a
Dominant crack formed when crack width exceeded 0.1 mm.
b
Occurred during the initial pulse; drift not recorded.
-10 c
Major splitting cracks formed when crack width reached 0.5 mm.

-20
(b) Cycles
expected in structures within 20–30 km of a fault wherein a
20
forward-directivity rupture produces a velocity pulse at the begin-
ning of the time history (Baker and Cornell 2008; Sehhati et al.
10
2011). The deformation pulses were expected to initiate crack
Drift (%)

localization and induce ECC material softening. With the applica-

0
tion of two small initial pulses, it was expected that crack locali-
zation would begin and some softening would occur, but many
-10
fibers would remain bridging the dominant crack. It was expected
that two large initial pulses would sever fiber bridging across the
-20
(c) Cycles dominant crack. Crack mouth opening displacement measurements
were taken at the peak of each excursion by hand and through
Fig. 2. (a) FEMA, (b) small-pulse, and (c) large-pulse deformation image analysis of digital photos.
histories

Experimental Results and Discussion

where drift was calculated as the horizontal defection at the point of
loading divided by the span length. Cracking
The two additional deformation histories used in this study each Table 4 shows the mean values of several cracking responses for all
began with two initial pulses that preceded the FEMA deformation 18 specimens grouped by the deformation history to which they
history. One specimen of each steel reinforcement ratio was sub- were subjected. The first crack in each specimen occurred near
jected to a deformation history beginning with two initial pulses 0.21% drift, and horizontal steady-state cracks less than 0.1 mm
to either 2 or 2.5% drift, called the small-pulse (SP) deformation wide formed as loading continued. Inelastic specimen response,
history [e.g., Fig. 2(b)], and one specimen of each steel reinforce- indicated by an abrupt change in specimen stiffness, occurred
ment ratio was subjected to a deformation history beginning with simultaneously with yield of the steel reinforcing bars and the
two initial pulses to either 5.5 or 7% drift, called the large-pulse formation of a dominant crack near the base of the specimen that
(LP) deformation history [e.g., Fig. 2(c)]. Preliminary testing indi- exceeded 0.1 mm wide. Inelastic response began between 1.1 and
cated potential reinforcing bar fracture at low drifts in some spec- 2.2% drift, and occurred at lower drifts in specimens with a lower
imens (Bandelt 2015). Therefore, initial pulses of the specimens steel reinforcement ratio, e.g., 0.73 and 0.95%. During inelastic
reinforced to 0.73% steel reinforcement ratio were limited to 2 specimen response, cracks continued to form during ECC material
and 5.5% drift for the SP and LP deformation histories, respec- softening. When all the fibers had failed at the dominant crack, all
tively. Specimens of the other five steel reinforcement ratios were further specimen strain induced by the loading cycles was concen-
subjected to initial pulses of 2.5 and 7% for the SP and LP defor- trated at the dominant crack. Splitting cracks on the edge of the
mation histories, respectively. Based on monotonic tests of speci- specimens formed after yielding of the steel reinforcing bars,
mens reinforced to 0.73 and 1.3% steel reinforcement ratio, and lengthened and widened with increasing drift. The drift at
yielding of the steel reinforcing bars and crack localization would which splitting cracks reached 0.5 mm wide was believed to be
be expected at a lower drift than the SP pulse amplitude for all of significance as described in the following section, and thus is
reinforcement ratios. Each test proceeded until the strength at presented in Table 4. Similar flexural cracking patterns and splitting
the peak of an inelastic excursion was less than 75% of yield cracks have been noticed in previous experiments of steel-
strength or when at least half of the steel reinforcement on one side reinforced ECC tests (Fischer and Li 2002; Yuan et al. 2012;
of the specimen fractured. Bandelt and Billington 2016).
The inclusion of initial pulses preceding the FEMA deformation To compare cracking responses between groups, t-tests at a
history was intended to simulate a different, yet plausible, earth- 90% confidence interval were performed (Table 4). One significant
quake response from monotonically increasing cycles. Initial defor- difference was in the mean drift at which the last new flexural crack
mation pulses represent the relatively large ductility demands formed between the specimens subjected to the FEMA deformation

© ASCE 04018052-4 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 history (3.5% drift) and the specimens subjected to the SP defor- differently. Results presented herein indicate fibers in a reinforced
mation history (2.5% drift). To further evaluate the difference, the ECC structural component that are subjected to a deformation
six specimens subjected to each deformation history were split into history containing initial pulses do not have the same fiber integrity
two subgroups: three specimens with a lower steel reinforcement of a similar component subjected to a deformation history with
ratio (0.73–1.0%) and three specimens with a moderate steel monotonically increasing cycles. Fiber integrity could affect the
reinforcement ratio (1.3–1.5%). component’s ability to withstand environmental conditions because
Further analysis revealed that the difference in the drift at which cracks may be less able to remain narrow enough to prevent sig-
the last new flexural crack formed was driven by the response of nificant water and chloride ingress (Lepech and Li 2009). The dif-
specimens in the moderate steel reinforcement ratio subgroup. ferent cracking due to differences in deformation history could
Specimens with a moderate steel reinforcement ratio subjected therefore affect repair decisions following a seismic event. Sensors
to the FEMA deformation history continued to crack at larger drifts that could record story drifts and identify initial deformation pulses
than specimens with a lower reinforcement ratio, 4.5% drift com- in reinforced ECC structural components, combined with visual
pared with 2.5% drift, on average. This was expected because the inspections, may aid in determining fiber integrity during post-
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

tensile strains in the specimens with a moderate steel reinforcement earthquake inspections.
ratio would be lower at a given drift due to the neutral axis position
being closer to the extreme tensile fiber than less reinforced spec- Reinforcement Strain
imens. Lower tensile strains at a given drift delayed crack formation
until larger drifts. In specimens subjected to the SP deformation Reinforcement Strain Evolution
history, it was seldom that new cracks formed after completion Strain in the steel reinforcement was measured at six locations
of the initial pulses, to 2 or 2.5% drift. The small initial pulses in each specimen by strain gauges attached to reinforcing bars
led to the formation of a dominant crack and some softening. [Figs. 1(a and b)]. During testing of all 18 specimens, the strain
Müller et al. (2015) conjectured that polymeric fibers get damaged in four of those locations (5 cm above and 5 cm below the joint
through abrasion and crushing, and lose tensile strength during face on both sides of the specimen) exceeded yield strain, while
cyclic loading. In specimens subjected to two small initial pulses, strain in the steel located 15 cm above the joint peaked at or near
fibers began to pull out along some portions of the dominant crack yield strain through the drift at which the strain gauges failed.
during the initial pulses. It is believed that the exposed portion of Figs. 3(a–f) show the envelopes of strain in the steel reinforcement
the fibers were damaged during the loading cycles that followed the at the peak of each tension excursion recorded 5 cm above and 5 cm
initial pulses leading up to the cycles that exceeded the initial pulse below the joint face on both sides of six select specimens up until
amplitude. Damaged fibers were less able to transfer stress across strain gauge failure. Additional reinforcement strain evolution re-
the dominant crack. With a diminished ability to transfer stress sults are shown in Frank (2017). For reference, the strain envelopes
across the dominant crack, specimens subjected to the SP deforma- are plotted over the completed portion of the deformation history
tion history had softened more than specimens subjected to the up until specimen failure.
FEMA deformation history at a given drift. Because fiber integrity In specimens subjected to the FEMA deformation history,
was believed to be the root cause of the difference in mean drift at the reinforcement remained elastic through the first several cycles.
which the last new flexural crack formed between specimens sub- Upon yield, strain in the steel reinforcement increased rapidly
jected to the FEMA and SP deformation histories, this observation within the first excursion inducing yield, then typically increased
would not be expected when comparing reinforced concrete spec- gradually as drift increased during subsequent cycles as the bar
imens subjected to deformation histories with and without small hardened. When comparing strain in the steel across specimens
initial pulses. at the same level of drift and subjected to the same deformation
The second significant difference in observed cracking was history, strain was generally greater in specimens with a lower steel
between specimens subjected to the FEMA deformation history reinforcement ratio, as expected. In particular, strain in the steel
and those subjected to the LP deformation history. The width of reinforcement was greater when the reinforcement ratio was lower
the dominant crack was significantly larger (6.9 mm) in specimens by way of smaller reinforcing bars, or in other words, when the
subjected to the LP deformation history than those subjected to the ratio of bond capacity to bond demand was greater. For example,
FEMA deformation history (5.2 cm) at the drift used for compari- the peak recorded strain in ECC-0.95-F (two 13-mm bars per side)
son (6.1% drift). While this difference was observed in specimens at the first positive excursion to 6.1% drift was 6.0% [Fig. 3(a)],
with both lower and moderate steel reinforcement ratios, the sig- whereas the peak strain in ECC-1.5-F (two 16-mm bars per side)
nificant difference was primarily driven by specimens with a lower was 3.5% [Fig. 3(b)].
steel reinforcement ratio. Specimens with a lower steel reinforce- In specimens subjected to the SP deformation history, steel
ment ratio yielded at lower drifts due to the greater tensile strain in reinforcement strain increased during the initial pulses, then peak
the extreme fiber, as expected, compared to specimens with a mod- tensile strain reduced during cycles whose amplitudes were smaller
erate steel reinforcement ratio. The observation that reinforced than that of the initial pulses. As amplitude approached and
ECC specimens subjected to the LP deformation history formed increased beyond that of the initial pulses, strain in the steel
a wider dominant crack than those subjected to the FEMA defor- reinforcement increased again. One strain gauge in ECC-1.3-SP
mation history may not carry over to observations of reinforced and two strain gauges in ECC-1.4-SP did not increase once cyclic
concrete specimens subjected to deformation histories with and amplitude exceeded that of the initial pulses [Figs. 3(c and d)].
without initial pulses because several wide cracks would be ex- Rather, strain at these locations decreased. Steel strain increases
pected to form in the concrete specimens, whereas only one crack and reductions due to the cyclic nature of the loading wherein
wider than 0.1 mm occurred per ECC specimen. the steel reinforcing bars alternate from tension to compression
Final cracking patterns between reinforced ECC specimens are an expected response, and thus these intracycle fluctuations
subjected to different deformation histories can be found in are not documented in Figs. 3(a–f).
Frank (2017). While final cracking patterns appeared similar, as In this paper, steel strain reductions refer to an incident of de-
did the average number of cracks, steel reinforcement ratio creasing strain in the steel reinforcement while the reinforcing bar
and deformation history affected the fibers bridging the cracks is in tension and the drift is increasing within or between cycles.

© ASCE 04018052-5 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Strain in the steel reinforcing bars of (a) ECC-0.95-F; (b) ECC-1.5-F; (c) ECC-1.3-SP; (d) ECC-1.4-SP; (e) ECC-1.3-LP; (f) ECC-1.4-LP
recorded by four strain gauges per specimen up until strain gauge failure

Both of the aforementioned specimens that experienced strain re- these results, however, is that the largest recorded steel reinforcing
ductions in the steel reinforcement were among the moderate steel bar strains at a given drift were recorded in reinforced ECC spec-
reinforcement ratio subgroup in this study. Their relatively large imens subjected to the FEMA deformation history, and not those
cross-sectional area of steel generated a high bond demand at subjected to a deformation history containing initial pulses. Further,
the steel–ECC interface, which caused the steel to debond from specimens with a lower steel reinforcement ratio tended to have
the ECC. This debonding led to the reduction in reinforcement higher recorded strain in the steel reinforcement at a given drift.
strain that was observed in Specimens ECC-1.3-SP and ECC- When fracture of the reinforcement is the design failure mode
1.4-SP. of a reinforced ECC specimen, a deformation history containing
Most of the strain gauges in specimens subjected to the LP monotonically increasing cyclic steps may produce more severe
deformation history failed during the first pulse, with the exception loading conditions due to the higher expected accumulation of
of the three reinforced ECC specimens with steel reinforcement strain in the steel reinforcement. Without a priori information of
ratios of 1.3, 1.4, and 1.5%. While gauges in ECC-1.4-LP and what type of deformation history is expected in the field, monoton-
ECC-1.5-LP remained functional beyond the two initial pulses, ically increasing cycles in an experimental protocol should be a
strain reduced because cyclic amplitudes were smaller than that conservative approach to testing steel-reinforced ECC building
of the initial pulses [e.g., Fig. 3(f)]. As amplitude increased to levels components.
near to and above 7% drift, strain in the steel reinforcement in-
creased in ECC-1.4-LP. In contrast, strain at two locations in Reinforcement Strain Reduction
ECC-1.3-LP rose with drift, then abruptly reduced during the initial In general, during the 18 experimental tests in this study, strain in
pulses as a splitting crack opened beyond 0.5 mm wide [Fig. 3(e)]. the tensile steel reinforcement increased with each tensile excursion
Strain remained relatively low throughout the remainder of the data and decreased with each compressive excursion. Throughout the
collected and did not increase, even as drift exceeded the amplitude deformation history, strain at the peak of each tensile excursion
of the initial pulses. It is believed the splitting crack that grew more generally grew larger as the increases during tensile excursions out-
than 0.5 mm wide increased the length of reinforcement debonded paced the decreases during each compressive excursion. Given the
from the ECC material. Larger debonded lengths from splitting direct relationship between increasing strain in the steel reinforce-
cracks led to decreased strain in the steel reinforcement in previous ment and reinforcing bar fracture, reductions in steel reinforcement
experiments as well (Moreno et al. 2014). strain wherein strain at the peak of each tensile excursion decreased
Due to strain gauge failure before the conclusion of testing in as drift increased were expected to delay specimen failure.
many specimens, drawing robust conclusions about the effects of Because growth in splitting crack width occurred simultane-
deformation history, steel reinforcement ratio, or reinforcing bar ously with reductions in strain in the steel reinforcement of
size on strain in the steel reinforcement is difficult. The trend in ECC-1.3-LP, a detailed investigation into the relationship between

© ASCE 04018052-6 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 Table 5. Instances of Strain Reductions in the Steel Reinforcement and the Status of Splitting Cracks Widths at the Time of the Observations
Deformation Before any Splitting 0.1 mm ≤ splitting Splitting crack
history splitting crack width < 0.1 mm crack width < 0.5 mm width ≥ 0.5 mm
FEMA 3 1 0 0
SP 5 0 0 0
LP 0 1 1 2

observed splitting cracks in the ECC and strain gauge reductions reinforcement strain 5 cm above the joint face in ECC-1.3-LP
was performed. Within the 18 specimens in this study, 90 bars were reduced from a peak of 0.80% strain to 0.38% abruptly within
used for longitudinal reinforcement. Splitting cracks were observed the first pulse to 7% drift. Destructive posttest visual inspections
in the ECC near 78 of the 90 (87%) bars. Splitting cracks that grew of Specimens ECC-0.73-SP and ECC-1.3-F revealed abrasion on
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

to at least 0.5 mm wide, indicative of potential loss of fiber bridging the ECC, indicating interface crushing occurred, further reinforcing
strength per observations of ECC-1.3-LP, were observed in the hypothesis that the physical mechanism of crushing of the ECC
the ECC near 41 of the 90 (46%) of the bars. As described in at the steel reinforcing bar interface is associated with the FEMA
“Specimen Geometry and Instrumentation,” on 36 of the bars and SP deformation histories.
across the 18 specimens, a total of 72 strain gauges were placed
within 5 cm of the joint face. Reductions in steel strain as measured
by a strain gauge within 5 cm of the joint face were recorded by 13 Hysteretic Response
strain gauges. However, instances of strain reductions were not Fig. 4 shows the hysteretic response of six select reinforced ECC
correlated with the presence or size of splitting cracks (Table 5). specimens. Additional hysteretic responses are shown in Frank
When reductions in reinforcement strain were observed, spec- (2017). Observations of splitting cracks at least 0.5 mm wide,
imens subjected to the FEMA and SP deformation histories typi- crushing, and spalling on both sides of the specimen, as applicable,
cally experienced strain reductions prior to the formation of are indicated.
splitting cracks. In contrast, specimens subjected to the LP defor- As expected, peak specimen strength increased with steel
mation history typically experienced reductions in the steel strain as reinforcement ratio due to the additional tensile strength provided
splitting cracks first formed or as splitting crack widths were ac- by the increased longitudinal steel. Each of the six specimens sub-
tively increasing. These data suggest that drawing robust conclu- jected to the FEMA deformation history failed by fracture of the
sions on the level of strain in the steel reinforcement, including steel reinforcement. As previously noted, greater steel strain was
whether a localized strain reduction has occurred, is difficult by generally observed in specimens with a lower steel reinforcement
visual inspection of splitting cracks in the ECC near the reinforcing ratio. The increase in strain directly resulted in a reduction in
bars alone. Two possible causes for the observations presented in ultimate drift in specimens with a lower (0.73–1.0%) steel
Table 5 are discussed next. reinforcement ratio when compared to specimens with a moderate
(1.3–1.5%) steel reinforcement ratio. Ultimate drift ranged from
Reinforcement Strain Reduction Mechanisms 12% drift for the four specimens with the four lowest reinforcement
Two physical mechanisms leading to reduction in strain in the steel ratios (0.73–1.3%) to 17% drift for the two specimens with the two
reinforcement are hypothesized. With the LP deformation history, highest steel reinforcement ratios (1.4–1.5%). The trend of higher
splitting cracks were present in four of four incidents of observed ultimate drift in reinforced ECC specimens with a greater steel
strain reduction (Table 5); it is hypothesized that splitting was the reinforcement ratio is in agreement with findings in Bandelt and
dominant mechanism by which strain reduction occurred. Because Billington (2016).
three of four instances of observed reductions in steel strain that The peak strength of specimens subjected to the SP deformation
occurred during the FEMA deformation history and five of five in- history often occurred during the first positive excursion to 2 or
stances that occurred during the SP deformation history happened 2.5% drift, when the compression side of the specimen was undam-
prior to splitting cracks (Table 5), interface crushing is the pre- aged from previous cycles [Figs. 4(c and d)]. In general, hysteretic
sumed mechanism. This hypothesis is consistent with results from response was very similar between specimens subjected to the
Bandelt and Billington (2014), who also posited an increase in in- FEMA deformation history and those subjected to the SP deforma-
terface crushing in reinforced ECC beam-end specimens when sub- tion history, indicating that despite partial material softening within
jected to cyclic loading, using a protocol similar to the FEMA the first two cycles of specimens subjected to the SP deformation
deformation history, relative to specimens subjected to monotonic history, hysteretic response, as a whole, was not greatly affected.
loading. The small initial pulses of the SP deformation history did In all six specimens subjected to the LP deformation history, the
not alter the strain reduction mechanism from that of specimens peak strength occurred during the first positive excursion when the
subjected to the FEMA deformation history. compression side of the specimen was undamaged from previous
Characterization of the reinforcement strain reductions provided cycles [e.g., Figs. 4(e and f)]. Crushing was observed in four of
evidence that the mechanisms that caused strain reductions were the six specimens during the first large pulse, but did not lead
correlated to the deformation history. Strain reductions during to abrupt failure. The ability of the ECC to resist or delay spalling
the FEMA or SP deformation histories occurred somewhat gradu- despite some crushing during initial deformation pulses sustained
ally over several cycles. For example, reinforcement strain 5 cm confinement to the steel reinforcing bars and facilitated a hysteretic
below the joint face in ECC-1.3-SP reduced from a peak steel strain envelope similar to that of specimens subjected to the FEMA and
of 1.2 to 0.16% over 22 cycles from the first 2.5% pulse to the first SP deformation histories.
cycle to 6.1% drift. The rate of strain reduction during the LP de- Fig. 5 shows ultimate drift of 16 out of the 18 reinforced ECC
formation history was higher. Reduction in reinforcement strain specimens in this study that failed due to fracture of the steel
from peak steel strain to the strain at the first tensile excursion after reinforcement without significant joint damage; ECC-1.3-SP,
strain reduction initiated was, on average, 53%. For example, which experienced significant joint damage, and ECC-1.4-LP,

© ASCE 04018052-7 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

Fig. 4. Hysteresis curves of select reinforced ECC specimens: (a) ECC-0.73-F; (b) ECC-1.5-F; (c) ECC-0.73-SP; (d) ECC-1.5-SP; (e) ECC-0.73-LP;
(f) ECC-1.5-LP

which failed due to crushing, are omitted. When comparing ulti- amplitude for steel reinforcement ratios from 0.95 to 1.0% or from
mate drift between the six specimens subjected to the FEMA 1.3 to 1.5%. Despite complete fiber rupture or pullout at the loca-
deformation history with the five specimens subjected to the SP tion of the dominant crack during the first cycle in all reinforced
deformation history or the five specimens subjected to the LP ECC specimens subjected to the LP deformation history, in general,
deformation history, in general, there was no significant difference. specimen ductility was not affected relative to that of nominally
Nearly horizontal trendlines of the data grouped by steel reinforce- identical specimens subjected to the FEMA deformation history.
ment ratio indicate no trend in ultimate drift with the initial pulse Indifference of reinforced ECC flexural member ductility to various
deformation histories was unexpected, but can be explained by the
18 ability of the ECC to tolerate significant damage in compression
15 while confining the reinforcing bars and the cementitious core re-
Ultimate Drift (%)

gardless of the presence of a dominant crack. In contrast, reinforced

12 concrete specimens would not maintain the same level of confine-
9 ment when subjected to large pulses. Crushing failure may occur
after fewer cycles when a reinforced concrete specimen is subjected
6
to a deformation history containing large initial pulses than a
3 deformation history containing monotonically increasing cycles.
0
Reinforced ECC specimens with a lower reinforcement ratio,
0 1 2 3 4 5 6 7 however, trended toward a reduction in ultimate drift with the pres-
Initial Pulse Amplitude, Drift (%) ence and size of initial pulses. The three specimens with the lowest
of the six reinforcement ratios tested (0.73%) had ultimate drifts of
12, 8.5, and 6.1% when subjected to the FEMA, SP, and LP de-
formation histories, respectively. These findings suggest that initial
Fig. 5. Effect of initial pulse amplitude on ultimate drift grouped by
pulses in a deformation history may influence ductility in rein-
steel reinforcement ratio, ρ
forced ECC components, but only those with a steel reinforcement

© ASCE 04018052-8 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 ratio near the lowest steel reinforcement ratio tested in this study, component underwent than visible splitting cracks in the specimen.
perhaps 0.73% or less wherein the bond capacity to bond demand Strain in the steel reinforcement had the tendency to reduce prior to
ratio is comparatively large. the ECC material forming splitting cracks when subjected to the
The conclusion that failure mode and ductility, in general, of FEMA or small-pulse deformation histories, evidence of internal
reinforced ECC beams were unaffected by the deformation histor- damage through interface crushing. More abrupt strain reduction
ies applied in this study are important findings for the design of due to splitting cracks, in specimens subjected to the large-pulse
reinforced ECC components for an unknown future ground motion. deformation history, was believed to accommodate the strain in-
When considering design for collapse prevention under a large seis- duced by the initial pulses, and facilitated the indifference in speci-
mic event, the tendency of a ground motion to generate initial men ductility to deformation history.
deformation pulses may not be an important ground motion char- Results presented herein through experimental testing provide
acteristic to evaluate or include in numerical models or experimen- a fundamental understanding of how reinforced ECC flexural
tal validation so long as the steel reinforcement ratio is 0.95% or members with various steel reinforcement ratios and reinforcing
greater. bar sizes respond to various deformation histories. More details
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

and further analysis of these results, including energy dissipation

and stiffness degradation, are presented in Frank (2017). Numerical
Conclusions simulations validated by the results from experimental testing are
needed to complement and expand upon this work.
Eighteen steel-reinforced engineered cementitious composite
flexural members were tested quasi-statically under cyclic loading
in order to better understand their seismic response. Because the Acknowledgments
characteristics of future earthquakes are unknown and cyclic defor-
mation histories with monotonically increasing cycles dominate the The authors would like to acknowledge financial support provided
literature, the response of reinforced ECC beams subjected to de- by the Air Force Institute of Technology, the John A. Blume Earth-
formation histories with and without initial deformation pulses quake Engineering Center, and the Thomas V. Jones Engineering
were investigated. Three specimens from each of six different steel Faculty Scholarship at Stanford University. The views expressed in
reinforcement ratios were subjected to different deformation histor- this paper are those of the authors and do not reflect the official
ies; one was the deformation history recommended by FEMA 461, policy or position of the United States Air Force, Department of
and the other two were the FEMA deformation history preceded by Defense, or the U.S. Government.
two initial deformation pulses of varying amplitudes. The findings
from this study provide new insights into reinforced ECC compo-
nent response under deformation histories containing initial pulses. References
Initial deformation pulses were found to have a wide-ranging ACI (American Concrete Institute). (2014). “Building code requirements
effect on reinforced ECC flexural members when compared with for structural concrete.” ACI 318-14, Farmington Hills, MI.
similar specimens that underwent the FEMA deformation history. Baker, J. W., and Cornell, C. A. (2008). “Vector-valued intensity measures for
The specimens with the lowest steel reinforcement ratio tested pulse-like near-fault ground motions.” Eng. Struct., 30(4), 1048–1057.
(0.73%) showed reduced ductility when subjected to deformation Bandelt, M. J. (2015). “Behavior, modeling, and impact of bond in steel
histories containing initial pulses compared with no pulses, and re- reinforced high-performance fiber-reinforced cement-based compo-
duced further when the amplitude of the initial pulses was larger. In sites.” Ph.D. dissertation, Stanford Univ., Stanford, CA.
general, however, failure mode and ductility of reinforced ECC Bandelt, M. J., and Billington, S. L. (2014). “Monotonic and cyclic bond-
slip behavior of ductile high-performance fiber-reinforced cement-based
flexural members were not significantly affected by the presence
composites.” Proc., 3rd Int. RILEM Conf. on Strain Hardening Cemen-
or size of initial pulses. titious Composites (SHCC3), E. Schlangan, M. G. Sierra Beltran,
Strain in the steel reinforcement was found to accumulate at a M. Lukovic, and G. Ye, eds., RILEM, Paris, 1–8.
greater rate in specimens subjected to a deformation history con- Bandelt, M. J., and Billington, S. L. (2016). “Impact of reinforcement ratio
taining monotonically increasing cycles than those containing and loading type on the deformation capacity of high-performance
initial deformation pulses. To be conservative, future research of fiber-reinforced cementitious composites reinforced with mild steel.”
reinforced ECC components concerning the seismic performance J. Struct. Eng., 10.1061/(ASCE)ST.1943-541X.0001562, 04016084.
and failure by fracture of the steel reinforcement should use the Douglas, K. S., and Billington, S. L. (2010). “Strain rate dependence of
FEMA deformation history, or similar. HPFRCC cylinders in monotonic tension.” Mater. Struct., 44(1),
Small initial pulses in the deformation history were shown to 391–404.
FEMA. (2007). “Interim testing protocols for determining the seismic per-
soften the ECC, and subsequent cycles were believed to damage
formance characteristics of structural and nonstructural components.”
the fibers, limiting their capacity to bridge cracks and contribute FEMA 461, Washington, DC.
to the formation of new cracks. This insight is useful for postearth- Fischer, G., and Li, V. (2002). “Effect of matrix ductility on deformation
quake inspections. Partially softened ECC may perform poorly in behavior of steel-reinforced ECC flexural members under reversed
aftershocks or against postearthquake freeze-thaw cycles or water cyclic loading conditions.” ACI Struct. J., 99(6), 781–790.
ingress. Methods to identify the deformation history to which a Frank, T. E. (2017). “Response of reinforced engineered cementitious
reinforced ECC component was subjected during an earthquake, composite flexural members subjected to various cyclic deformation
e.g., from sensor data, could aid in making repair decisions based histories.” Ph.D. dissertation, Stanford Univ., Stanford, CA.
on the insights to fiber integrity it could provide in conjunction with Frank, T. E., Lepech, M. D., and Billington, S. L. (2015). “Effect of
deformation history on steel-reinforced HPFRCC flexural member
visual observations.
behavior.” Proc., 7th RILEM Workshop on High Performance Fiber
The presence of splitting cracks alone was found to be a Reinforced Cement Composites (HPFRCC7), H. W. Reinhardt,
poor indicator of the level of steel reinforcement strain accumula- G. J. Parra-Montesinos, and H. Garrecht, eds., RILEM, Paris, 367–374.
tion. Characteristics of recorded reductions in steel reinforcement Hanson, J., and Billington, S. L. (2009). “Cyclic testing of a ductile fiber-
strain, i.e., the mechanisms that cause steel–ECC bond degradation, reinforced concrete infill panel system for seismic retrofitting of steel
were more closely associated with the deformation history the frames.” Rep. No. TR 177, Stanford Univ., Stanford, CA.

© ASCE 04018052-9 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052

 Kanakubo, T., and Hosoya, H. (2015). “Bond-splitting strength of rein- Mechtcherine, V., et al. (2011). “Behaviour of strain-hardening cement-
forced strain-hardening cement composite elements with small bar based composites under high strain rates.” J. Adv. Concr. Technol.,
spacing.” ACI Struct. J., 112(2), 189–198. 9(1), 51–62.
Kanda, T., and Li, V. C. (1999). “New micromechanics design theory for Moreno, D. M., Trono, W., Jen, G., Ostertag, C., and Billington, S. L.
pseudostrain hardening cementitious composite.” J. Eng. Mech., 10 (2014). “Tension stiffening in reinforced high performance fiber
.1061/(ASCE)0733-9399(1999)125:4(373), 373–381. reinforced cement-based composites.” Cem. Concr. Compos., 50,
Lepech, M. D., and Li, V. C. (2009). “Water permeability of engineered 36–46.
cementitious composites.” Cem. Concr. Compos., 31(10), 744–753. Müller, S., Mechtcherine, V., and Zydek, M. (2015). “Behaviour of strain-
Li, V. C., and Kanda, T. (1998). “Engineered cementitious composites for hardening cement-based composites (SHCC) subject to cyclic loading.”
Proc., 7th RILEM Workshop on High Performance Fiber Reinforced
structural applications.” J. Mater. Civ. Eng., 10.1061/(ASCE)0899-
Cement Composites (HPFRCC7), H. W. Reinhardt, G. J. Parra-
1561(1998)10:2(66), 66–69.
Montesinos, and H. Garrecht, eds., RILEM, Paris, 333–340.
Lignos, D. G., Moreno, D. M., and Billington, S. L. (2014). “Seismic
Parra-Montesinos, G., and Chompreda, P. (2007). “Deformation capacity
retrofit of steel moment-resisting frames with high-performance
and shear strength of fiber-reinforced cement composite flexural mem-
fiber-reinforced concrete infill panels: Large-scale hybrid simulation bers subjected to displacement reversals.” J. Struct. Eng., 10.1061
Downloaded from ascelibrary.org by HQ USAFA/DFLIB/SER on 02/21/22. Copyright ASCE. For personal use only; all rights reserved.

experiments.” J. Struct. Eng., 10.1061/(ASCE)ST.1943-541X.0000877, /(ASCE)0733-9445(2007)133:3(421), 421–431.

04013072. Sehhati, R., Rodriguez-Marek, A., ElGawady, M., and Cofer, W. F. (2011).
Maalej, M., and Li, V. C. (1995). “Introduction of strain-hardening engi- “Effects of near-fault ground motions and equivalent pulses on multi-
neered cementitious composites in design of reinforced concrete flexural story structures.” Eng. Struct., 33(3), 767–779.
members for improved durability.” ACI Struct. J., 92(2), 167–176. Yuan, F., Pan, J., Xu, Z., and Leung, C. K. Y. (2012). “A comparison
Maalej, M., Li, V. C., and Hashida, T. (1995). “Effect of fiber rupture on of engineered cementitious composites versus normal concrete in
tensile properties of short fiber composites.” J. Eng. Mech., 10.1061 beam-column joints under reversed cyclic loading.” Mater. Struct.,
/(ASCE)0733-9399(1995)121:8(903), 903–913. 46(1–2), 145–159.

© ASCE 04018052-10 J. Struct. Eng.

J. Struct. Eng., 2018, 144(6): 04018052