Engineering Structures 122 (2016) 108–120

Contents lists available at ScienceDirect

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

Cyclic behavior of UHPFRC flexural members reinforced with

high-strength steel rebar
Chung-Chan Hung ⇑, Chen-Yu Chueh
Department of Civil Engineering, National Cheng Kung University, No.1, University Rd, Tainan City 701, Taiwan

a r t i c l e i n f o a b s t r a c t

Article history: Ultra-high performance fiber reinforced concrete (UHPFRC) is a unique class of fiber reinforced concrete.
Received 6 August 2015 It features an ultra-high compressive strength and a ductile, tensile strain hardening behavior accompa-
Revised 3 May 2016 nied by multiple narrow cracking. The cyclic flexural performance of UHPFRC structural beams reinforced
Accepted 11 May 2016
with high-strength steel with a specified yielding strength of 680 MPa is experimentally investigated in
Available online 26 May 2016
this study. Six cantilever beams are prepared and tested under displacement reversals. The experimental
variables include the reinforcement ratio of the high-strength longitudinal rebar and the amount, loca-
Keywords:
tion, and length of steel fibers in the beams. The intermediate and ultimate behaviors of these cantilever
Ultra-high performance fiber reinforced
concrete
members are discussed using multiple performance parameters, including strength capacity, flexural
High-strength steel rebar ductility, failure pattern, hysteretic response, energy dissipation capacity, and stiffness retention. The
Cyclic flexural behavior results show that UHPFRC beams reinforced with high-strength steel are able to show satisfactory cyclic
Cantilever beams flexural performance prior to failure. The addition of steel fibers substantially enhances the damage tol-
erance ability of the high-strength beams, even when the fibers are selectively used only in the top and
bottom beam sections. The proposed composite of UHPFRC and high-strength steel rebar not only takes
advantage of the ultra-high mechanical properties of both materials, but also resolves the issue of poten-
tial premature failure patterns associated with high-strength concrete and high-strength steel rebar.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction ratio. The tailored mixing design of UHPFRC leads to an ultra-high

High performance fiber reinforced cementitious composites
(HPFRCCs) are distinguished from traditional fiber reinforced
cementitious composites (FRCC) by their unique tensile strain
hardening behavior accompanied by multiple narrow cracks [1–
5], as shown in Fig. 1. Depending on the mixing design and the
fiber type, the ultimate tensile strain of HPFRCCs ranges between
0.2% and 8%, which can be up to 100 times greater than that of con-
ventional concrete materials. When HPFRCCs are under compres-
sion, fibers can provide a confining effect similar to that of
stirrups in reinforced concrete, enhancing the compressive
strength and ductility. The appealing mechanical performance of
HPFRCCs in the material level has motivated researchers and engi-
neers to explore their applications in earthquake-resistant struc-
tures [6–12]. Ultra-high performance fiber reinforced concrete
(UHPFRC) is a special class of HPFRCCs [13–20]. It is made by an
optimized gradation of granular materials with minimal or no
coarse aggregate. Its material proportion is characterized by a high
cementitious material content with a very low water/binder (w/b)
⇑ Corresponding author. Tel.: +886 6 2757575x63130; fax: +886 6 2370804.
E-mail address: cchung@mail.ncku.edu.tw (C.-C. Hung).
compressive strength. In contrast to the brittle and explosive fail-
ure pattern of conventional high-strength concrete, UHPFRC exhi-
bits a significantly improved ductility and residual strength
under compression, and shows ductile strain hardening behavior
under tension.
Lampropoulos et al. [21] numerically investigated the feasibility
of using UHPFRC for strengthening RC beams, and suggested that
superior strengthening performance could be achieved by using
UHPFRC jackets. Xu et al. [22] experimentally studied the behavior
of UHPFRC columns subjected to blast loading. It was found that
UHPFRC columns could effectively resist the overpressures and
shock waves, restraining the maximum and residual displacements
of columns. Hosinieh et al. [23] experimentally studied the behav-
ior of UHPFRC columns under pure axial loading. It was concluded
that using UHPFRC to replace normal concrete in columns could
improve the post-peak ductility of columns more significantly
when the confinement steel was less. Yoo et al. [24] studied the
flexural behavior of UHPFRC beams reinforced with glass fiber-
reinforced polymer (GFRP) rebars and hybrid reinforcements
(steel + GFRP rebars). The results indicated that all test beams sat-
isfied the code-specified service crack width criteria and showed
very stiff load–deflection behavior after the formation of cracks.

http://dx.doi.org/10.1016/j.engstruct.2016.05.008
0141-0296/Ó 2016 Elsevier Ltd. All rights reserved.
 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120 109

37]. The objective of the study is to characterize the flexural perfor-

mance of UHPFRC structural beams longitudinally reinforced with
high-strength steel rebar under seismic action. Six structural can-
tilever beam specimens with varying designs are tested under dis-
placement reversals. The design parameters include the flexural
reinforcement ratio and the amount, type, and location of steel
fibers used in the structural beams. The test results are discussed
using multiple performance variables, including strength capacity,
flexural ductility, cracking pattern, hysteretic response, energy dis-
sipation capacity, and stiffness retention.

2. Materials

The tested structural elements are composite beams consisting

of UHPFRC and high-strength steel rebar. The details of the
employed materials are presented in this section. It should be
noted that the term UHPC (ultra high performance concrete) gen-
erally has a broader definition than UHPFRC, i.e., it can be used
to refer to the UHPFRC with or without fiber reinforcement. In
Fig. 1. Comparison between HPFRCCs and regular fiber reinforced cement-based
composites (FRCCs). order to better differentiate between fiber reinforced and non-
fiber reinforced UHPC in this study as well as to prevent exces-
sively long phrases, the term UHPC herein refers to the non-fiber
Astarlioglu and Krauthammer [25] numerically investigated the reinforced UHPC, whereas the term UHPFRC is the UHPC material
response of UHPFRC columns under impulsive loads. It was con- reinforced with fibers.
cluded that the peak displacements of UHPFRC columns were
about 30% smaller than those of regular concrete columns, and 2.1. UHPFRC
the level of axial loading had a significant influence on the behavior
of UHPFRC columns. Han and An [26] suggested the use of bundled The mixed proportions of the UHPFRC material employed in this
longitudinal reinforcing bar as a substitute for steel fibers for study are given in Table 1. The components include Type I ordinary
enhancing the flexural ductility of UHPFRC beams. Palacios [27] Portland cement, silica fume, silica sand (with particle sizes rang-
experimentally investigated the behavior of UHPFRC columns ing between 0.1 mm and 0.3 mm), quartz powder,
under displacement reversals. It was found that the use of UHPFRC polycarboxylate-based superplasticizer admixture, water, and
changed the typical failure mode of concrete columns. The results high-strength hooked steel fibers. In order to study the effect of
also indicated that UHPFRC columns exhibited higher peak fibers on the performance of beams, different amounts of steel-
strength and greater drift capacity prior to significant strength hooked fibers are used in the beam specimens, namely 0%, 1%, or
degradation compared to conventional normal strength RC col- 2%. In addition, the performance of two types of high-strength steel
umns. Ren et al. [28] studied the triaxial compressive behavior of fibers with lengths of 30 mm and 60 mm, respectively, is studied.
UHPFRC and then applied the results to analyze the behavior of Details of the fibers are listed in Table 2. The UHPFRC mixtures pre-
UHPFRC against projectile high-speed impact. Baby et al. [29] pared in this study have adequate workability, with the measured
applied the modified compression field theory to predict the diameters in the slump flow test of about 600 mm and 620 mm for
flexure-shear behavior of prestressed and reinforced UHPFRC the ones with the 30 mm-long fibers and 60 mm-long fibers,
beams. The numerical model was able to predict the reorientation respectively. It is worth mentioning that the addition of the steel
of the compressive struts due to increasing force demands. Makita fibers with a greater aspect ratio (i.e., length over diameter)
and Brühwiler [30] proposed empirical fatigue damage models for intends to decrease the workability of UHPFRC mixtures and the
UHPFRC. In their model, the tensile fatigue behavior of UHPFRC dispersion of fibers.
was analyzed based on the elementary damage mechanics theory. The compressive strength of the UHPFRC material is determined
Bastien-Masse and Brühwiler [31] suggested a composite model using compressive tests of standard cylinder specimens
for predicting the punching resistance of reinforced UHPFRC–RC (100 mm
200 mm). The tensile properties of UHPFRC are
composite slabs.
Although many studies have confirmed the superior mechanical
and damage tolerance properties of UHPFRC under monotonic or Table 1
Mix proportions of UHPFRC (proportion by weight).
blast loading, limited research has been carried out on the seismic
performance of UHPFRC structural members [27]. The study pre- Type I Silica Silica Quartz Water and Fiber (by
sented herein investigates the cyclic flexural behavior of an inno- Portland fume sand powder superplasticizer volume)
cement
vative UHPFRC structural beam reinforced with high-strength
steel rebar. The employment of high-strength steel rebar to replace 1 0.23 0.5 0.39 0.31 0%, 1%, 2%
the one with normal strength allows designers to reduce the
amount of required reinforcing steel, which eases the placement
of UHPFRC. It also takes advantage of the high strength and ductil-
Table 2
ity of UHPFRC. The use of high-strength steel rebars in conjunction Properties of hooked steel fibers.
with UHPFRC is a potential method for resolving the unfavorable
Type of Length Diameter Specified yielding Elastic
failure modes that could arise when high-strength steel is used
fibers (mm) (mm) strength (MPa) modulus (GPa)
in normal concrete members. These issues include concrete crush-
I 30 0.38 3070 210
ing before reinforcing bars yield, buckling of compressive steel
II 60 0.90 2300 210
rebar, premature crack widening, and splitting bond failure [32–
110 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120

obtained using direct tensile tests on dog-bone specimens. The

configuration, dimensions, and test setup of the dog-bone speci-
men are shown in Fig. 2. The tensile deformation of the specimen
is measured using two linear variable displacement transducers
(LVDTs) mounted on the two sides of the specimen. The gauge
length is 160 mm. Although this LVDT setup, which was also
widely adopted by other studies due to its efficiency (for example,
in [1,2]), can generally capture the tensile deformation of speci-
mens with acceptable accuracy, it should be noted that the accu-
racy of the measurement can be further enhanced by using two
additional LVDTs on the other two faces of the specimen. Because
the use of four LVDTs is able to address the deformation state of the
entire cross section, including the effects of out-of-plane bending.
The magnitude of the applied force is measured using the load cell
attached to the actuator. Both the tensile and compressive tests are Fig. 3. Stress–strain relationship of the high-strength steel rebar.
conducted using displacement control with a loading rate of
0.5 mm/min to simulate the quasi-static loading condition.
shown in Fig. 3. It can be seen that the high-strength steel exhibits
adequate hardening behavior and ductility before facture.
2.2. Steel rebar

3. Test specimens
Two types of steel rebar are used in the beams. The transverse
reinforcement in all beam specimens consists of D13 hoops with
Six structural specimens are prepared and tested. Their names,
a specified yielding strength of 420 MPa. All beams are longitudi-
design details and geometries are shown in Table 4 and Fig. 4,
nally reinforced with high-strength D25 steel bars with a specified
respectively. The specimen consists of a beam element connected
yielding strength of 680 MPa. The actual tensile properties of the
to an RC stub. All beam elements have identical dimensions,
steel reinforcements are obtained using direct tensile tests. The
namely, 2000 mm in length, 250 mm in width, and 350 mm in
results are summarized in Table 3. In addition, the measured
height. The test setup for the specimens is shown in Fig. 5. The
stress–strain relationship for the high-strength D25 steel rebar is
beam element is intended to represent a beam cantilevered from
the RC stub. It is tested under cyclic loading with the RC stub fixed
to a strong floor. In order to sufficiently restrain the RC stub in all
degrees of freedom, the RC stub is clamped by two precast RC
blocks connected by ten steel rods. Each steel rod is fixed to the
strong floor with a prestressed force of 125 kN.
The experimental variables of the six cantilever beams are the
amount of high-strength longitudinal rebar (4-D25 or 6-D25) as
well as the amount (volume fraction of 0%, 1%, or 2%), type
(30 mm-long or 60 mm-long hooked steel fibers), and location of
the fibers (in full depth or selectively in the top and bottom sec-
tions). B-2R-0SF, which is employed as the control specimen, is
entirely cast using UHPC; the used UHPC has the same mix as
UHPFRC, except that no fiber is added. For the other beams,
UHPFRC is used in the plastic hinge region, which is assumed to
be within 700 mm away from the face the RC stub. The rest of
the beam is made using UHPC. In B-2R-2SF/TB, the top and bottom
sections (80 mm depth each) in the plastic hinge region of the
beam, which have the largest demand with regard to stress and
strain reversals, are cast using UHPFRC with 2% fibers, whereas
the web section is made using UHPC. In order to compare the influ-
ence of varying fiber lengths on the flexural behavior of the
UHPFRC beam, B-1.4R-2LF uses the 60 mm-long steel fibers,
whereas other UHPFRC beams employ the 30 mm-long steel fibers.
The reinforcement details of the specimens are shown in Fig. 4.
All beams are designed to exhibit flexure-dominant behavior by
providing sufficient transverse reinforcement. D13 steel stirrups
with a close spacing of 75 mm are used in the plastic hinge region
Fig. 2. Test setup and dimensions of the tensile specimen.
to prevent shear failure. The rest of the beam has D13 stirrups
spaced at 150 mm. All specimens are longitudinally reinforced
Table 3
with high-strength D25 steel bars. In order to investigate the effect
Tensile properties of steel bars. of a larger demand of steel strain on the cyclic performance of the
beam, specimens B-1.4R-2SF and B-1.4R-2SF are designed to have a
Size Yielding strength Ultimate strength Yielding Elongation
(MPa) (MPa) strain (%)
smaller tensile reinforcement ratio (1.4%) compared to that of the
others (2.0%).
D13 477 731 0.0021 19
The material strengths of UHPC and UHPFRC used in the beam
D25 703 907 0.0022 15
specimens measured on the test days are summarized in Table 4.
 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120 111

Table 4
Details of the cantilever beam specimens.

Specimen Flexural Tensile Fiber Strengths of UHPC/ Age at testing Completed

rebar reinforcement UHPFRC drift and
ratio its cycle
Volume fraction Length Compression Tension
B-2R-0SF 6-D25 2.0% 0% 30 mm 150.8 MPa – 191 days 9%, 2
B-2R-1SF 6-D25 2.0% 1% 30 mm 148.0 MPa 7.5 MPa 173 days 8%, 3
B-2R-2SF 6-D25 2.0% 2% 30 mm 151.6 MPa 9.9 MPa 193 days 9%, 3
B-2R-2SF/TB 6-D25 2.0% 2% in top and bottom sections; 30 mm 142.6 MPa 7.5 MPa 156 days 9%, 2
0% in middle section
B-1.4R-2SF 4-D25 1.4% 2% 30 mm 146.8 MPa 8.6 MPa 150 days 7%, 2
B-1.4R-2LF 4-D25 1.4% 2% 60 mm 143.2 MPa 11.1 MPa 144 days 8%, 2

Fig. 4. Design details of test specimens (unit: mm).

Fig. 5. Test setup of the cantilever beam (unit: mm).

112 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120
In particular, the uniaxial tensile stress–strain curves of the
UHPFRC materials are shown in Fig. 6(a). It can be seen that all
UHPFRC materials exhibit an obvious hardening behavior. The
post-cracking strains, at which the cracks start to localize, are
between 0.25% and 0.6%. After crack localization occurs, the
stress–strain curve shows a softening behavior due to the further
opening of the localized crack. The ultimate failure strains of the
tensile specimens are about 2%. A representative failure pattern
of the UHPFRC dog-bone specimen is shown in Fig. 6(b), in which
a dense array of cracks can be observed.
Fig. 6. Tensile behavior of the UHPFRC materials.
Fig. 7. Applied reversed drift ratios.
Fig. 8. Locations of strain gauges in steel bars (unit: mm).
4. Test instrumentation and load protocol
Vertical displacement reversals are imposed on the end of the
beam using a hydraulic actuator with a loading capacity of
1000 kN and a stroke length of 600 mm. The center line of the
hydraulic actuator is located at 1850 mm away from the face the
RC stub. The shear span to beam depth ratio is 5.3. The applied drift
ratio history is shown in Fig. 7; the drift ratio is calculated as the
ratio of the deflection measured at the end of the beam to the shear
span. Three cycles at each target drift are applied. The magnitude
of the load that is imposed is measured using a load cell connected
to the actuator. The deflection at the end of the beam is measured
using two string potentiometers, which are attached to the two
sides the beam and aligned to the center line of the hydraulic actu-
ator, as shown in Fig. 5. The measured drift is corrected for the pos-
sible rotation of the fixed RC stub, which is measured using a
rotational gauge installed on the top of the RC block. The corrected
deflection of the beam is calculated as
D ¼ Da  Db ð1Þ
where Da is the average of the values measured by the two string
potentiometers; Db is the vertical displacement of the beam due
to the rotation of the clamped RC stub, which can be computed as
 
1
Db ¼ L þ H
h ð2Þ
2
where L is the span length of the beam, H is the depth of the RC stub,
and h is the measurement of the rotation gauge. Although the com-
puted Db is very minor compared to Da , it is taken into account in
this study to enhance the accuracy of the test. Furthermore, a dense
array of strain gauges are attached to the longitudinal and trans-
verse steel reinforcements in the plastic region of the beam, as
shown in Fig. 8, to monitor the reinforcement strain throughout
the test. The strain gauges are deployed at a spacing of 150 mm
on the longitudinal rebars. In addition, each stirrup has two strain
gauges that are 100 mm apart.
5. Evaluation of experimental results
Extensive data are obtained from the experimental tests. The
structural performance of the beams is discussed in terms of crack
and failure patterns, load–deflection relationship, energy dissipa-
tion capacity, stiffness retention, and steel reinforcement strain.
5.1. Crack patterns
Fig. 9 shows the crack patterns of the cantilever beam speci-
mens at the 5% drift response. It can be seen that the bottom cover
in the plastic hinge region of B-2R-0SF has significantly spalled due
to tension and compression reversals induced by cyclic loading. In
particular, the steel reinforcement has been exposed. In contrast,
no obvious concrete spalling is observed in UHPFRC beams,
 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120 113

Fig. 9. Crack patterns in the plastic hinge regions at 5% drift response.

114 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120

Fig. 10. Damage patterns in the plastic hinge regions at the end of the test.
 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120
signifying the confining effect provided by steel fibers. Although B-
2R-2SF/TB only has fibers in the top and bottom beam sections,
concrete spalling is still effectively prevented. In comparison with
the control beam (B-2R-0SF), narrower and more closely dis-
tributed cracks are developed over the plastic hinge regions in
the UHPFRC beams due to the bridging effect of steel fibers. The
results imply that the unique multiple cracking of UHPFRC mate-
rial observed on the material scale can also be successfully devel-
oped on the UHPFRC beams reinforced with high-strength steel
rebar on the structural scale.
At the 5% drift response, notable localized flexural cracks have
been developed in the plastic hinge regions of B-1.4R-2LF and B-
1.4R-2SF. This is because these two beams have less flexural rein-
forcement, which leads to a larger demand for reinforcement strain
compared to other beams when under displacement reversals. In
particular, it can be found that pullout of fibers occurs across the
major crack in B-1.4R-2SF. The maximum crack widths in B-1.4R-
2SF and B-1.4R-2LF are 10.0 mm and 5.1 mm, respectively, at 5%
drift. The results imply that the 60-mm long fibers perform better
than the 30-mm long fibers in terms of arresting the structural-
scale cracks.
The final failure patterns of the beams are shown in Fig. 10. The
maximum drift at which the beam attains at least a full cycle is
listed in Table 4. The failure modes of all beams are due to the frac-
ture of high-strength longitudinal steel reinforcement. It can be
observed in Fig. 10 that the number of cracks of the control speci-
men B-2R-0SF is substantially lower than that of the UHPFRC
beams. Before B-2R-0SF fails due to fracture of longitudinal rebar,
substantial concrete spalling has occurred in both the cover and
core regions in the plastic hinge zone. This damage pattern not
only limits the energy dissipation capacity of the high-strength
beam but also causes the beam to become highly structurally
unstable when subject to further loading. Unlike the substantial
spalling in B-2R-0SF, the concrete spalling and crushing is effec-
tively restrained in all UHPFRC beams, due to the confinement
effect provided by the fibers that substantially enhances the com-
pressive ductility and residual strength of high-strength concrete.
Before all UHPFRC beams fail, significantly widening of flexural
cracks extends through the depth of the beam in the plastic hinge
region. These widening cracks induce a large local stress in the
high-strength steel rebar, ultimately causing the rebar to fracture.
For B-2R-2SF/TB, the economical use of fibers in the top and bot-
tom beam sections, where the largest tension and compression
demands occur, effectively delays and restrains concrete spalling
in the beam covers. As the drift demand of B-2R-2SF/TB increases
to be more than 6%, concrete spalling in the middle beam section,
where no fibers are used, gradually becomes obvious, as can be
Fig. 11. Comparison of number of cracks in the specimens.
115
seen in Fig. 10(l). In comparison between B-1.4R-2SF and B-1.4R-
2LF, the difference between the lengths of the employed steel
fibers appears to have an insignificant effect on the final damage
pattern.
It is also worth mentioning that no premature bond failure,
which often appears as the splitting concrete crack along the flex-
ural reinforcement, is observed between the high-strength steel
rebar and UHPFRC (or UHPC) until the end of the test. The results
imply that the material characteristics of the low w/b ratio and
the dense matrix of UHPFRC (or UHPC) enable an adequate bond
strength between the matrix and the high-strength steel bars in
the flexure-dominant beam specimens under large displacement
reversals. Moreover, the combined use of UHPC and UHPFRC in
B-2R-2SF/TB does not cause visible debonding at the UHPFRC-
UHPC interface throughout the test. The presence of the interface
between UHPFRC and UHPC does not impair the performance of
the beam specimen.
5.2. Number of cracks
The relationships between the number of cracks versus the drift
demand for the various beam specimens are plotted in Fig. 11. All
UHPFRC beams have more cracks than the control beam specimen,
B-2R-0SF, in which no fiber is added. In particular, B-1.4R-2SF has
nearly twice the number of cracks as B-2R-0SF throughout the test.
The greater number of cracks in the UHPFRC beam reduces the
average crack width in comparison with the UHPC beam (B-2R-
0SF) under the same drift demand. Furthermore, Fig. 11 shows that
as the volume fraction of fibers is increased from 1% to 2%, more
cracks are developed in the beams. It can also be seen that B-
1.4R-2SF and B-1.4R-2LF have a larger number of cracks compared
to other UHPFRC beams before 6% drift demand. This can be attrib-
uted to the lower amount of longitudinal steel reinforcement in
these beams, which induces a larger demand for tensile strain in
the reinforcing steel. Overall, the test results imply that it is possi-
ble to effectively enhance the crack-width control ability of a high-
strength steel reinforced UHPC beam by means of adding high-
strength steel hooked fibers.
Fig. 11 also shows that B-1.4R-2SF has more cracks than B-1.4R-
2LF. This signifies that the 30-mm long fibers are more beneficial
than the 60-mm long ones in promoting the multiple cracking
behavior in the high-strength structural specimen. Moreover, it is
interesting to note that B-2R-2SF and B-2R-2SF/TB show similarly
good performance in terms of the number of cracks. These results
suggest that even if fibers are only added in the top and bottom
beam sections in the plastic hinge region, the damage tolerance
ability of the high-strength steel reinforced UHPC beam can still
be effectively enhanced.
5.3. Load-drift response
The load-drift hysteretic responses of the cantilever beam spec-
imens subject to cyclic loading are shown in Fig. 12. The results
show that all high-strength steel reinforced beams are able to exhi-
bit ample ductility and stable hysteretic behavior before failure,
even for the beam without fibers, i.e., B-2R-0SF. Fig. 12 also shows
that as the drift response increases, the hysteresis loops of B-2R-
0SF become slightly narrower than that of the UHPFRC beams.
The results imply that the steel fibers in the UHPFRC beams not
only help restrain concrete spalling, but also sustain the integrity
of the concrete matrix around the reinforcing rebar under cyclic
loading, thus effectively enhancing the bond-slip toughness. The
comparison between the responses of B-1.4R-2SF and B-1.4R-2LF
in Fig. 12 shows that the 30-mm long fibers lead to wider hys-
teretic loops than the 60-mm long ones, as also seen in the energy
dissipation capacity, which is reported below.
116 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120

Fig. 12. Load–displacement responses of the specimens.

5.4. Strength envelope
Fig. 13 shows the strength envelopes of the various beams sub-
ject to cyclic loading. It can be seen that as soon as B-2R-0SF
reaches its yielding strength at 3% drift, there is an obvious drop
in the strength for further drift demand, possibly due to the signif-
icant concrete spalling that reduces the effective beam depth. On
the other hand, all UHPFRC beams, including B-2R-2SF/TB in which
fibers are selectively used in its top and bottom sections, are able
to sustain their strength fairly well. It can be observed in Fig. 13
that the flexural strength capacity of the beam increases along with
the amount of steel fibers, i.e., B-2R-2SF > B-2R-1SF > B-2R-0SF.
The addition of fibers by 2% volume fraction is able to increase
the flexural strength capacity by about 20% in comparison with
B-2R-0SF. The results show that the tensile and compressive dam-
age tolerance abilities of UHPFRC in the material scale effectively
lead to a stable load carrying capacity of beams in the structural
scale.
It is also interesting to note that B-2R-0SF, B-2R-1SF, B-2R-2SF,
and B-2R-2SF/TB, which have the same steel reinforcement but dif-
ferent amounts of 30 mm long fibers, have approximately the same
drift capacity. This implies that the material ductility gained by the
addition of 30 mm long fibers is not able to translate to the struc-
tural flexural ductility of beams. In fact, it can be observed during
the test that as the UHPFRC beam experiences large displacement
reversals, crack localization occurs due to pullout failure of fibers.
The rebar strain consequently concentrates at the locations where
the localized cracks develop. Due to cyclic loading on the beam, the
 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120 117

strain localization. Therefore, although the addition of 30 mm long

Fig. 13. Strength envelopes of the beams.
concentrated rebar strain accumulates rapidly, and this facilitates
rebar fracturing. On the other hand, in the plastic hinge region of
the UHPC beam (i.e., B-2R-0SF), the flexural reinforcement is essen-
tially exposed due to significant concrete spalling. The exposed
flexural reinforcement leads to a more uniform strain distribution
in the steel reinforcement, thus delaying rebar fracturing due to
fibers is able to enhance the ductility of UHPFRC in the material
scale, it appears to have an insignificant effect on the drift ductility
of the high-strength beam under cyclic flexural loading.
The benefits of using the 60 mm long fibers to replace the
30 mm long ones in the UHPFRC beam can be observed by compar-
ing the envelopes of B-1.4R-2SF and B-1.4R-2LF in Fig. 13. In con-
trast to the results found for 30 mm long fibers, the addition of
60 mm long fibers in the beam is able to considerably enhance
the drift capacity of the beam subject to displacement reversals.
The results can be explained by the observation made during the
test that B-1.4R-2LF has substantially less fiber pullout at the local-
ized cracks than B-1.4R-2SF at 7% drift. This implies that the
60 mm long steel fibers are able to provide a more effective bridg-
ing effect across the cracks than the 30 mm long ones as the drift
response increases. In addition to the improvement of drift capac-
ity, Fig. 13 also shows that the use of 60 mm long fibers is able to
enhance the flexural strength capacity of the beam under cyclic
loading.
When the same fibers are used but the tensile reinforcement
ratio of the beam is decreased from 2% (B-2R-2SF) to 1.4% (B-
1.4R-2SF), the drift capacity of the UHPFRC beam is decreased from

Fig. 14. Energy dissipation capacities of the beams under cyclic loading.

Fig. 15. Stiffness of the beams under cyclic loading.

118 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120

9% to 7% due to the earlier rebar fracturing, as a result of the larger

demand for steel rebar strain. This trend is however opposite to
that of normal reinforced concrete flexural members. For a normal
reinforced concrete flexural member, less flexural reinforcement
results in a smaller effective compression zone, which delays con-
crete crushing and consequently leads to greater flexural ductility.

5.5. Energy dissipation capacity

The energy dissipation capacities for the various beams under

cyclic loading are compared. The energy dissipated with a specified
target drift is calculated as the enclosed area of the hysteretic loop
at the 3rd cycle to the target drift divided by the rectangular area
enclosing the loop, as illustrated in Fig. 14(a). Fig. 14(b) presents
the cumulative energy dissipation capacity versus the drift
response for the beams. The results show that the addition of fibers
enhances the energy dissipation capacity of the beam under cyclic
flexural loading, even when the fibers are only used in the top and
bottom beam sections. This is because the bridging and confining
effects provided by fibers not only prevent concrete spalling and
crushing, but also lead to more narrow cracks. Therefore, the hys-
teretic loop is fatter for UHPFRC beams. At the end of testing, B-2R-
2SF shows a cumulative energy capacity about 30% larger than that
of B-2R-0SF. In addition, the influence of the fiber length on the
energy dissipation capacity of the beam is also significant. The
short fibers appear to have a more beneficial effect in enhancing
the energy dissipation capacity. This is because they perform better
in terms of facilitating the multiple cracking pattern that sustains
the integrity of the steel–concrete composite.

5.6. Stiffness retention

Fig. 15 shows the stiffness versus drift curves for the tested
beams under cyclic loading. The stiffness at a specified target drift
is computed as the slope between the force-drift values at the
maximum and minimum target drifts, as illustrated in Fig. 15(a).
Each stiffness value is calculated at the 3rd cycle to the specified
target drift. The computed results of stiffness in Fig. 15(b) show
that when the beams have the same longitudinal reinforcement
ratio, the flexural stiffness of the beam under cyclic loading is
slightly enhanced with the increasing amount of fibers. With
regard to the influence of the longitudinal reinforcement ratio, it
can be seen that B-1.4R-2SF has a flexural stiffness that is signifi-
cantly less than that of B-2R-2SF, which contains a larger reinforce- Fig. 16. Profiles of longitudinal steel strains at various beam drifts.
ment ratio. In comparison between B-1.4R-2SF and B-1.4R-2LF, it
can be seen that the former has a larger flexural stiffness before
the drift reaches 3%, implying that 30 mm long fibers are more
effective than 60 mm long ones in retaining the flexural stiffness
of the beams before crack localization occurs. This is because under
the same volume fraction of fibers, B-1.4R-2SF, which has the
shorter fibers, has a larger number of fibers than B-1.4R-2LF in
the beam to control the crack width. As the drift response
increases, the difference between the flexural stiffnesses of B-
1.4R-2SF and B-1.4R-2LF diminishes since the bridging effect of
fibers in B-1.4R-2SF gradually becomes ineffective.

5.7. Steel reinforcement strains

Fig. 16 shows the profiles of longitudinal rebar strains in the

plastic hinge region of the beam at 1%, 3%, and 5% drifts. Due to
the fact that the data measured by the strain gauges above the
5% drift are generally invalid, they are not reported herein. It can
be seen in Fig. 16 that the high-strength longitudinal rebar in all
beams at 1% drift response just enters or is about to enter the
inelastic stage. At 3% drift, the high-strength longitudinal rebar in
all beams exhibits considerable inelastic behavior. In particular,
Fig. 17. Profiles of transverse steel strains at various beam drifts.
the rebar strains in B-1.4R-2SF and B-1.4R-2LF are larger than
0.005, which is about 2.5 times the yielding strain of steel rebar.
In general, it is found that the UHPFRC beams have a larger rebar
strain than the UHPC beam (B-2R-0SF). The results show that the
 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120
localized crack in the UHPFRC beams introduces a larger local rebar
strain. Fig. 17 shows the strain profile of the transverse reinforce-
ment in the plastic hinge region. It shows that the strains of the
transverse reinforcement are well below the yielding strain of steel
reinforcement for all beams. The measured steel reinforcement
strains imply that the cyclic behavior of all beams is dominated
by flexure.
6. Conclusions
This paper experimentally studied the cyclic flexural character-
istics of an innovative structural beam, i.e., high-strength steel
reinforced UHPFRC cantilever beam. Six cantilever beams were
tested to investigate the influence of various experimental vari-
ables on the structural performance of UHPFRC beams. The exper-
imental variables included the flexural reinforcement ratio of high-
strength steel rebar and the amount, length, and location of steel
fibers. The test results showed that all high-strength steel rein-
forced UHPC and UHPFRC structural beams under cyclic loading
were able to show ample ductility before failure, thus ensuring
appropriate moment and force redistributions among members
in statistically indeterminate structures. The material characteris-
tics of UHPFRC successfully translated into the multiple cracking
and damage tolerance in the high-strength steel reinforced struc-
tural UHPFRC beams. The failure mode of UHPFRC beams was dom-
inated by the fracture of high-strength longitudinal steel rebar, as a
result of concentrating strains at the localized concrete cracks.
Despite the high demand for bond strength induced by the use of
high-strength steel rebar, no noticeable premature bond failure
between high-strength steel rebar and UHPFRC (or UHPC) was
observed at the ultimate state. In contrast to the UHPC beam, only
minor concrete spalling was observed in the UHPFRC beams at the
ultimate stage due to the bridging and confining effects of fibers.
The addition of fibers in the high-strength steel reinforced UHPC
beam was beneficial in enhancing the damage tolerance, strength
capacity, energy dissipation capacity, and stiffness retention of
the beam under cyclic loading. In particular, the selective use of
fibers in the top and bottom beam sections of the UHPC beam
was found to effectively inhibit concrete spalling and delay local-
ized cracking, consequently improving the energy dissipation
capacity of the high-strength beams. This composite UHPFRC–
UHPC beam was shown to be an efficient and economical alterna-
tive to UHPFRC beams in terms of cyclic flexural performance. In a
comparison of the performance of the 30 mm and 60 mm long
steel fibers, the former performed better in facilitating the multiple
narrow cracking and increasing the energy dissipation capacity,
whereas the latter enhanced the strength capacity and flexural
ductility of the beam. The flexural ductility was also found to be
improved by increasing the ratio of high-strength flexural rein-
forcement from 1.4% to 2%, which is opposite to the comparable
case of normal reinforced concrete flexural members.
Acknowledgements
The research described herein was sponsored in part by the
Ministry of Science and Technology, Taiwan under Grant No.
103-2221-E-006-268. The support of the high-strength steel rein-
forcement provided by Tokyo Tekko Co Ltd, Japan is also greatly
acknowledged. The opinions, findings, and conclusions expressed
in this paper are those of the authors, and do not necessarily reflect
those of the sponsor.
References
[1] Naaman AE, Reinhardt HW. Characterization of high performance fiber
reinforced cement composites. In: Naaman AE, Reinhardt HW, editors.
119
Proceedings of the 2nd International Workshop on High Performance Fiber
Reinforced Cement Composites: HPFRCC 2, RILEM. p. 1–24.
[2] Fischer G, Li VC. Effects of matrix ductility on the deformation behavior of
steel-reinforced ECC flexural members under reversed cyclic loading
conditions. ACI Struct J 2002;99(6):781–90.
[3] Hung C-C, El-Tawil S. Hybrid rotating/fixed-crack model for high performance
fiber reinforced cementitious composites. ACI Mater J 2010;107(6):
569–77.
[4] Hung C-C, Li S-H. Three-dimensional model for analysis of high performance
fiber reinforced cement-based composites. Compos B 2013;45:1441–7.
[5] Naaman AE, Likhitruangsilp V, Parra-Montesinos GJ. Punching shear response
of high performance fiber-reinforced cementitious composite slabs. ACI Struct
J 2007;104(2):170–9.
[6] Hung C-C, Yen W-M, Yu K-H. Vulnerability and improvement of reinforced ECC
flexural members under displacement reversals: experimental investigation
and computational analysis. Constr Build Mater 2016;107:287–98.
[7] Hung C-C, Su Y-F. On modeling coupling beams incorporating strain-hardening
cement-based composites. Comput Concr 2013;12(4):565–83.
[8] Canbolat BA, Parra-Montesinos GJ, Wight JK. Experimental study on the
seismic behavior of high-performance fiber reinforced cement composite
coupling beams. ACI Struct J 2005;102(1):159–66.
[9] Hung C-C, Su Y-F, Yu K-H. Modeling the shear hysteretic response for high
performance fiber reinforced cementitious composites. Constr Build Mater
2013;41:37–48.
[10] Hung C-C, El-Tawil S. Seismic behavior of a coupled wall system with HPFRC
materials in critical regions. ASCE J Struct Eng 2011;137(2):1395–636.
[11] Lequesne RD, Setkit M, Parra-Montesinos GJ, Wight JK. Seismic detailing and
behavior of coupling beams with high-performance fiber-reinforced concrete.
ACI SP271-11 2010;272:189–204.
[12] Hung C-C, Chen Y-S. Innovative ECC jacketing for retrofitting shear-deficient
RC members. Constr Build Mater 2016;111:408–18.
[13] Wille K, Kim DJ, Naaman AE. Strain-hardening UHP-FRC with low fiber
contents. Mater Struct 2011;44(3):583–9.
[14] Wille K, Naaman AE, El-Tawil S. Optimizing ultra-high-performance fiber-
reinforced concrete. Concr Int 2011;33(9):35–41.
[15] Wille K, Naaman AE, Parra-Montesions GJ. Ultra-high-performance concrete
with compressive strength exceeding 150 Mpa (22 ksi): a simpler way. ACI
Mater J 2011;108(1):42–54.
[16] Yang SL, Millard SG, Soutsos MN, Barnett SJ, Le TT. Influence of aggregate and
curing regime on the mechanical properties of ultra-high performance fiber
reinforced concrete (UHPFRC). Constr Build Mater 2009;23(6):2291–8.
[17] Graybeal BA. Compressive behavior of an ultra-high performance fiber
reinforced concrete. ACI Mater J 2007;104(2):146–52. 2007.
[18] Hassan AMT, Jones SW, Mahumd GH. Experimental test methods to determine
the uniaxial tensile and compressive behaviour of ultra high performance fibre
reinforced concrete (UHPFRC). Constr Build Mater 2012;37:874–82.
[19] Habel K, Denarié E, Brühwiler E. Time dependent behavior of elements
combining ultra-high performance fiber reinforced concretes (UHPRFC) and
reinforced concrete. Mater Struct 2006;39:557–69.
[20] Maya LF, Zanuy C, Albajar L, Lopez C, Portabella J. Experimental assessment of
connections for precast concrete frames using ultra high performance fibre
reinforced concrete. Constr Build Mater 2013;48:173–86.
[21] Lampropoulos AP, Paschalis SA, Tsioulou OT, Dritsos SE. Strengthening of
reinforced concrete beams using ultra high performance fibre reinforced
concrete (UHPFRC). Eng Struct 2016;106:370–84.
[22] Xu J, Wu C, Xiang H, Su Y, Li ZX, Fang Q, Li J. Behaviour of ultra high
performance fibre reinforced concrete columns subjected to blast loading. Eng
Struct 2016;118:97–107.
[23] Hosinieh MM, Aoude H, Cook WD, Mitchell D. Behavior of ultra-high
performance fiber reinforced concrete columns under pure axial loading. Eng
Struct 2015;99:388–401.
[24] Yoo DY, Banthia N, Yoon YS. Flexural behavior of ultra-high-performance fiber-
reinforced concrete beams reinforced with GFRP and steel rebars. Eng Struct
2016;111:246–62.
[25] Astarlioglu S, Krauthammer T. Response of normal-strength and ultra-high-
performance fiber-reinforced concrete columns to idealized blast loads. Eng
Struct 2014;61:1–12.
[26] Han SM, An JW. The ductile behavior test of ultra high performance fiber
reinforced concrete rectangular beam by the combination of the fiber
and group of reinforcing bars. J Korea Instit Struct Maint Insp 2015;19(3):
139–48.
[27] Palacios G. Performance of full-scale ultra-high performance fiber-reinforced
concrete column subjected to extreme earthquake-type loading and effect of
surface preparation on the cohesion and friction factors of the AASHTO
interface shear equation. PhD dissertation 2015;University of Texas at
Arlington.
[28] Ren GM, Wu H, Fang Q, Liu JZ, Gong ZM. Triaxial compressive behavior of
UHPCC and applications in the projectile impact analyses. Constr Build Mater
2016;113:1–14.
[29] Baby F, Marchand P, Atrach M, Toutlemonde F. Analysis of flexure-shear
behavior of UHPFRC beams based on stress field approach. Eng Struct
2013;56:194–206.
[30] Makita T, Brüehwiler E. Damage models for UHPFRC and R-UHPFRC tensile
fatigue behaviour. Eng Struct 2015;90:61–70.
[31] Bastien-Masse M, Brühwiler E. Composite model for predicting the punching
resistance of R-UHPFRC–RC composite slabs. Eng Struct 2016;117:603–16.
120 C.-C. Hung, C.-Y. Chueh / Engineering Structures 122 (2016) 108–120
[32] Tavallali H, Lepage A, Rautenberg J, Pujol S. Cyclic response of concrete frame
members reinforced with ultrahigh strength steel. ASCE/SEI 2011 Structures
Congress, Las Vegas, Nevada; 2001.
[33] Bing L, Park R, Tanaka H. Strength and ductility of reinforced concrete
members and frames constructed using high-strength concrete. Research
Report 94–5, Department of Civil Engineering, University of Canterbury,
Christchurch, New Zealand; 1994.
[34] Watanabe F, Lee JY, Nishiyama M. Structural performance of reinforced
concrete columns with different grade longitudinal bars. ACI Struct J 1995;92
(4):412–8.
[35] Todeschini CE, Bianchini AC, Kesler CE. Behavior of concrete columns
reinforced with high strength steels. J Am Concr Instit 1964;61(6):701–15.
[36] Seliem HM, Lucier G, Rizkalla SH, Zia P. Behavior of concrete bridge decks
reinforced with high-performance steel. ACI Struct J 2008;105(1):78–86.
[37] Sato Y, Tanaka H, Park R. Reinforced concrete columns with mixed-grade
longitudinal reinforcement. Research Report No. 93-7, Department of Civil
Engineering, University of Canterbury, Christchurch, New Zealand; 1993.