Front. Struct. Civ. Eng.

https://doi.org/10.1007/s11709-021-0701-4

RESEARCH ARTICLE

Flexural behavior of ECC hollow beams incorporating

different synthetic fibers
Ahmmad A. ABBASa,b, Farid H. ARNA ’OTc, Sallal R. ABIDd* , Mustafa ÖZAKÇAa
a
Civil Engineering Department, Gaziantep University, Gaziantep 72310, Turkey
b
Building and Construction Materials Department, Southern Technical University-Shatrah Technical Institute, Shatrah 64007, Iraq
c
College of Technical Engineering, Islamic University in An Najaf, Najaf 54001, Iraq
d
Civil Engineering Department, Wasit University, Kut 52001, Iraq
*
Corresponding author. E-mail: [email protected]

© Higher Education Press 2021

ABSTRACT Twelve ECC beams with three different fiber types, along with four normal concrete beams, were tested
in this study to evaluate the influence of cross-sectional hollowing on their flexural performance. The fiber types used
were nylon monofilament (NM), low-cost untreated polyvinyl alcohol (PVA), and polypropylene (PP). Three different
square hole sizes of 60, 80, and 100 mm with cross-sectional hollowing ratios of 0.16, 0.28, and 0.44, respectively, were
adopted for each group of beams in addition to a solid beam. All beams were tested under four-point loading using a
displacement-controlled testing machine. The test results showed that ECC beams can mostly withstand higher cracking
and ultimate loads compared to their corresponding normal concrete versions. The results also showed that both the
ductility and toughness of the ECC beams are higher than those of the normal concrete beams and that the ductility values
of the hollow beams with a hole size of 60 mm are higher than those of the corresponding solid beams. Moreover, hollow
ECC beams with hole sizes of 60 and 80 mm exhibited a higher ductility than a solid normal concrete beam.

KEYWORDS hollow beam, hollowing ratio, ECC, flexural strength, ductility, toughness

1 Introduction polyvinyl alcohol (PVA) have good resistance to alkali and

are not affected by hydrated cement [3]. The excellent
Concrete has a low tensile strength, low ductility, and low performance of these types of fibers is often reflected by
energy absorption owing to its low toughness and their strain-hardening behavior beyond the first cracking
microscopic size defects, which are highly expected to when under tension, which is attributed to their potential
occur under local tensile stresses [1,2]. The inclusion of for crack bridging [4–7].
short randomly distributed fibers in small fractions (0.5%– During the last two decades, several types of new and
2% by volume) was found to be a candidate solution to superior fibrous concrete were introduced with attractive
improve the tensile behavior of concrete. The inclusion of tensile and energy absorption characteristics. Engineered
fibers can significantly improve the tensile strength, cementitious composites (ECCs) are modern types of
increase the energy absorption (toughness), and decrease fibrous concrete that contain no coarse particles. Instead,
the number of defects that make the material behave quasi- extremely fine fillers are used with large amounts of
brittle. Moreover, in certain cases, it is difficult to place a cementitious materials. Owing to the large amounts of fine
sufficient number of reinforcing bars in thin-walled materials, the fiber content of ECCs is lower than that
members. Therefore, fibers can be used as a partial required for other mixture types providing the required
substitution of the lost quantity of the reinforcement area. mechanical properties, such as a high tensile strength [8]
Many types of fiber are currently used. Synthetic fibers and improved ductility [9,10], with a significantly high
such as nylon-monofilaments, polypropylene (PP), and strain capacity (3%–5%) [11]. One of the most attractive
characteristics of ECCs are their ability to control the crack
Article history: Received Jul 28, 2019; Accepted Apr 25, 2020 width to less than 60 µm owing to the good stress
2 Front. Struct. Civ. Eng.

distribution [12]. A vast number of studies on the types of synthetic fibers. Using the tested beams, the load-
application of ECCs in structural members have been deflection behavior, cracking and failure patterns, ductility,
conducted. ECCs have been used to modify the flexural and toughness of the hollow beams with different wall
behavior of the beams [13,14], improve the shear thicknesses were investigated and compared to those of
resistance of structural members [15] and the behavior of solid beams. Such light weight sections can be considered
column beam connections during an earthquake [16], and successful candidates in precast building construction.
as a solution to improving the abrasion resistance of
concrete in hydraulic structures [17,18].
Although new types of concrete including ECCs have 2 Materials and methods
certain advantages, they also have some disadvantages,
including the weight of the member and high costs. As 2.1 Concrete mixture and material properties
another disadvantage, such types of materials consume
large amounts of cement, which means that their produc- In this research, ECC mixtures were created using three
tion leads to higher carbon dioxide emissions. Reducing different types of synthetic fibers: nylon monofilament
the size of the structural members will reduce the weight, (NM), low-cost untreated PVA, and PP. Table 1 lists the
cost, and of course, the carbon dioxide emitted because the characteristics of the three types of fibers. In addition to the
amount of consumed cement will be reduced. To reduce three ECC mixtures, a normal concrete mixture was
the size of the structural member, its cross-sectional prepared for comparison. Table 2 summarizes the material
dimensions should be optimized in such a way that it proportions of the four mixtures. The NM and PVA-ECC
balances the structural design requirements (strength and mixtures were mixed according to the M45, as outlined in
serviceability) and its weight. To improve the ductility, Ref. [28], in which a 2% fiber content was used, whereas
pultruded cellular sections and layered sandwich sections PP-ECC was mixed according to a mixture proposed by
using fiber-reinforced polymers have been suggested in Zhang et al. [29] with a 3% fiber content.
previous studies [19–21]. Hollow concrete sections were The experimental program was composed of two stages.
also introduced as candidates that may afford the required In the first stage, four series of mechanical property tests
balance. The reduced sectional area will mostly result in a were conducted. First, standard 100 mm
200 mm
lower strength and weaker behavior under loads. However, cylinders subjected to a uniaxial compression load were
0
fibers can be used to partially or fully substitute the loss of used to evaluate the compressive strength (fc ) of concrete
strength and performance produced by reducing the according to ASTM C39. Second, similar cylindrical
sectional area. specimens were subjected to an indirect tensile load
Several studies have been conducted on the torsional (splitting tensile test) to evaluate the concrete tensile
strength [22–24] of hollow reinforced concrete beams, strength (fct ) according to ASTM C 496. Third, similar
whereas studies on the flexural performance of hollow specimens were subjected to a compression load to
beams are too limited [25]. The available studies have evaluate the Young’s modulus according to ASTM
focused on normal concrete with either no fibers or C469. Fourth, prisms of 100 mm
100 mm
350 mm
reinforcement with steel fibers [26,27]. An ECC is were cast and tested according to ASTM C1609 to evaluate
composed completely of fine materials and is known for the flexural performance under four-point loading.
its superior ductility and flexural behavior compared to
normal concrete. Thus, the use of such materials with 2.2 Configuration of solid and hollow beams
optimized thin-walled sections may reduce the weight of
the structure, keeping the strength comparable to that of Sixteen beams were tested to investigate the flexural
solid normal concrete beams but with a much higher behavior of ECC beams with different longitudinal hole
energy absorption capacity under loading. To the best of sizes and different fiber types. The beams were divided
the authors’ knowledge, no previous studies have into four groups according to the type of fiber used. The
investigated the flexural behavior of ECC beams with beams of the first group were made of normal concrete (E
hollow sections. Therefore, in the present research, an group) for comparison purposes, whereas the beams of the
experimental program was directed for investigating the other three groups were made of an ECC but with different
flexural behavior of hollow ECC beams containing three types of fibers (NM, PVA, and PP). Three beams with
Table 1 Fiber properties
type density (kg/m3) lf* (mm) Df* (mm) lf/Df tensile strength (MPa) elastic modulus (GPa)
NM 1140 19 0.050 380 966 25.0
PVA 1260 6 0.015 400 1600 34.0
PP 910 12 0.030 400 350 1.4

*Note: lf and Df are the length and diameter of the fiber, respectively.
 Ahmmad A. ABBAS et al. Flexural behavior of ECC hollow beams 3

Table 2 Proportions of ECC and normal concrete mixtures

mix fiber (%) C (kg) FA (kg) G (kg) S (kg) S.S. (kg) S.fm (kg) HWR (kg) W (lt)
normal 0 465 – 680 1170 – 35 – 235
NM-ECC 2 551 662 – – 441 – 7.3 352
PVA-ECC 2 551 662 – – 441 – 7.3 352
PP-ECC 3 939 461 – – – – 7 371

Note: C, cement; FA, fly ash type F; G, gravel; S, river sand; S.S., silica sand; S.fm, silica fume; HWR, high water reducer; W, water.

square holes with side lengths of 60, 80, and 100 mm in 2.3 Beam test setup
addition to a solid beam were fabricated from each of the
four groups. Table 3 lists the details of the 16 beams. The All beams were tested under four-point flexure loading
number following the group symbol represents the side until failure. The beams were supported on two cylindrical
length of the hole, where 0, 6, 8, and 10 refer to hole side supports with a clear 750-mm span. The tests were
lengths of 0, 60, 80, and 100 mm, respectively. conducted using a closed-loop servo-controlled INSTRON
testing machine with a maximum capacity of 250 kN. The
Table 3 Compressive strength, splitting tensile strength, and modulus loading of the beams was conducted at a constant rate of
of elasticity of the four mixtures 0.4 mm/min. Increased monotonic loading was applied
beam Vf hole size f 'c fsp Ec through two steel cylinders attached to the crossing head.
(%) (mm) (MPa) (MPa) (GPa) The deflection of the beams was measured using linear
E0 0 0 32.3 4.47 33.4
variable differential transformers (LVDTs) that were
attached beneath the tested beam, as shown in Fig. 2. To
E6 60
determine the net center deflection, two dial gauges were
E8 80 used. One dial gage was used at the top of the free edge of
E10 100 the beam, and the other was attached to the tension side
NM0 2 0 38.4 6.46 9.3 (100 mm from the support), as shown in Fig. 2.
NM6 60
NM8 80
3 Results of control tests
NM10 100
PVA0 2 0 34.3 6.49 17.3 3.1 Compressive strength, splitting tensile strength, and
PVA6 60 modulus of elasticity
PVA8 80
The three types of ECCs and the normal concrete mixtures
PVA10 100
were designed to achieve a 28-d cylinder compressive
0
PP0 3 0 33.9 6.58 15.5 strength (fc ) of approximately 35 MPa to eliminate the
PP6 60 effect of the concrete strength on the flexural behavior of
PP8 80 the beams. Table 3 summarizes the compressive strengths
PP10 100 of the four mixtures obtained. Many previous studies have
shown that the inclusion of fibers does not significantly
influence the compressive strength of concrete [30–32].
Table 3 shows that, in general, the compressive strengths of
All tested beams are of square reinforced cross sections the four mixtures were comparable.
of 150 mm
150 mm and a length of 850 mm, as shown in In contrast to the compressive strength, the tensile
Fig. 1. The beams were reinforced with longitudinal and strengths of the ECCs for all tested specimens of different
transverse deformed bars of 8 mm and 5.3 mm diameter, fiber types were significantly improved compared to a
respectively. The average reinforcement ultimate and yield normal concrete mixture. The percentage of improvement
strength of the 8 mm diameter bars were 650 and 530 MPa, in splitting the tensile strength of the ECC mixtures
respectively, whereas for the 5.3 mm diameter bars they compared to normal concrete ranges 44.5%–47%, as
were 477 and 400 MPa, respectively. The clear concrete shown in Table 3. In spite of the differences in fiber
cover of the beams was 15 mm, which led to an average tensile strength, all ECC mixtures showed almost equal
effective depth (d) of 131 mm. Figure 1 illustrates the splitting tensile strengths. In addition, because of the effect
configuration and reinforcement details of the four of the fibers, they all failed gradually rather than directly
different beams of each of the four groups. separating into two parts as for normal concrete. By
4 Front. Struct. Civ. Eng.

Fig. 1 Configuration and reinforcement details of the tested solid and hollow beams.

were tested under three-point loading with a clear 300-mm

span. The load was applied at a rate of 0.075 mm/min up to
a mid-span deflection of 1 mm, and a 0.4 mm/min loading
rate was applied up to failure.
Table 4 shows the main parameters that were obtained
from the test results, whereas Fig. 3 shows the develop-
ment of the mid-span displacement of the beams. The
flexural strength was calculated based on the linear elastic
beam theory, as follows:
PL
f ¼ , (1)
bH 2
where f is the specified strength (MPa), P is the
characterized load (N), L is the span length (mm), and b
and H are the width and depth of the beam (mm),
respectively.
In the current study, the first peak strength (f1) and
ultimate flexural strength (fp) were almost the same for
beams of normal concrete and PVA-ECC, whereas f1 for
Fig. 2 The beam test setup under four-point loading. the NM-ECC and PP-ECC represent 68% and 65% of the
fp, respectively. It was also observed that the tensile
contrast, the experimental results obtained in this research strength obtained from a splitting test (fsp) is more than that
showed that the modulus of elasticity of the ECC is obtained from a beam bending test fp by 28%–35% for the
typically lower than that of concrete. This can be attributed three ECC mixtures and 21% for the normal concrete
to the absence of coarse aggregates. mixture.
The improvement of fiber on the flexural behavior is
3.2 Flexural test clearly reflected in the load–deflection behavior after the
first crack. In the current study, the normal concrete and
The flexural tests on prisms made from the four mixtures PVA-ECC prisms showed brittle behavior after cracking.
were conducted in accordance with ASTM C1609. The By contrast, the NM-ECC and PP-ECC prisms showed
beams were 100 mm
100 mm
350 mm in size and different post-crack behaviors. The NM-ECC showed a
 Ahmmad A. ABBAS et al. Flexural behavior of ECC hollow beams 5

Table 4 Main flexural performance indices

D D,F
mix f1 (MPa) fp (MPa) f 1 / fp T150 (J) T150 RD
T ,150 (%) RD,F
T ,150
D,0
T150 RD,0
T ,150

normal 4.97 4.97 1.00 3381.30 1.00 14 1.00

NM-ECC 3.24 4.75 0.68 15500.83 4.58 106 7.6
PVA-ECC 4.87 5.03 0.97 14463.12 4.28 46 3.3
PP-ECC 3.29 5.08 0.65 21078.60 6.23 78 5.6

4 Results of solid and hollow reinforced

beams
4.1 Flexural strength

Two major stages are important in the beam design. First,

the deflection of the beam under service loads should not
detract from the appearance. In the second state, the beam
should be safe under the worst-loading case. In this
section, the tested beams are compared and assessed under
both the ultimate and serviceability stages.

4.1.1 Serviceability stage

Fig. 3 Load-deflection curves of prisms tested according to Under low loading conditions, the beam behaves elasti-
ASTM C1609. cally, where the compression and tensile stresses in
concrete are small and proportional to the corresponding
strain softening, whereas the PP-ECC showed a hardening strains. At this stage, the section is mostly uncracked,
plateau before the softening zone. This can be attributed to where the stress at the extreme tension fiber is still less than
the higher fiber content used in the PP-ECC mixture (3%) the modulus of rupture. To observe the development of the
compared to the 2% used in the NM-ECC and PVA-ECC flexural behavior of ECC beams compared to normal
mixtures. concrete beams, the tested normal concrete beams were
The post-cracking behavior can be specified as the theoretically analyzed according to the elastic bending
residual strength at a specified deflection or toughness at theory. These beams were analyzed considering the target
the same point. In the current study, the flexural strength compressive strength, which is 30 MPa, and the experi-
ratio (RD D mental yield strength of the used steel bars, which is
T,150 ) is calculated based on the toughness (T150 ) of
up to L/150, as follows: 530 MPa. By contrast, the modulus of elasticity (Ec) was
calculated according to MC2010 [33] as follows:
150
T150
D

í 1=3
T,150 ¼
RD
100%: (2) 3 fc þ 6:4
f1
b
d 2 Ec ¼ 21:5
10 : (3)
10
The toughness results showed that all ECC prisms have
a high energy absorption compared to those made of The cracking bending strength is then elastically
normal concrete. The T150 D
of NM-ECC and PVA-ECC was calculated through the following:
D
more than 4-times the T150 of normal concrete, whereas the fr
Ig
D
T150 of the PP-ECC was approximately 6-times that of Mcr ¼ , (4)
yt
normal concrete. However, the flexural strength ratio
Ig
(RDT ,150 ) was varied based on the first crack strength, which where is the section modulus of the beam.
was small for NM-ECC and PP-ECC; hence, their RD yt
T,150
The required modulus of rupture (f1,ACI) is calculated
values were high. Table 4 shows the RD T ,150 of the three according to ACI318-14 as follows:
types of ECC mixtures compared to the normal concrete pffiffiffiffiffiffi
mixture, where the highest ratio was obtained for the NM- f1,ACI ¼ 0:62 fcí : (5)
ECC.
6 Front. Struct. Civ. Eng.

The maximum deflection at the center of the normal recorded for the PP-ECC beams and partially for the
concrete beams before cracking can be calculated accord- PVA-ECC beams, whereas those with NM fibers exhibited
ing to the elastic bending theory as a continuous decrease in the cracking load with an increase

 in the hole size, as shown in Fig. 4. Although beams with
Pa – 3L2 hole indices of 0.16 and 0.28 (except for the NM group)
δmax ¼ þa ,
2
(6)
12Ec I 4 showed encouraging results, the further increase in the hole
size to an index of 0.44 resulted in a lower cracking load;
where P is the cracking load, L is the clear span, a is the
however, it was still comparable to that of the correspond-
distance between one point load and the closer support,
ing solid beam. The maximum Pcr,test values were recorded
and Ec and I are the concrete modulus of elasticity and the
for beams made from PP-ECC.
moment of inertia of the beam section, respectively.
The analysis of the tested solid and hollow beams
showed that, in general, the theoretical cracking loads 4.1.2 Ultimate stage
(Pcr,theo) of the normal concrete beams were less than those
obtained experimentally for all beam groups. For the The ultimate moment capacity (Mu,normal ) for the solid
normal concrete beams, the Pcr,theo values were less than normal concrete beam was calculated according to Eq. (7),
the experimental cracking loads (Pcr,test) by approximately where the tensile strength of concrete in the tension zone is
15%, 27%, 36%, and 21% for E0, E6, E8, and E10, assumed to be zero, and the reinforcement bars resist the
respectively. The calculated cracking loads (Pcr,theo) were applied tensile stresses. For comparison, the safety factors
9.9, 9.6, 9.2, and 8.0 kN, respectively, whereas the for the applied moments and materials were neglected. In
cracking loads obtained from the tested beams (Pcr,test) the current study, the compression reinforcement was
are as listed in Table 5. By contrast, for most cases, Pcr,test found to avoid yielding.
of ECC beams were much higher than those obtained  a
theoretically, as shown in Fig. 4. It is clear that the values Mu,normal ¼ 0:85fcí ba d – þ Aí í
s Es εs ðd – d # Þ: (7)
2
of Pcr,theo slightly decrease with an increase in the
hollowing index, where the hollowing index (HI) refers The depth of the compression concrete block (a) can be
to the ratio of the hole sectional area to the beam total calculated as
sectional area (HI = Ahole/Atot). Although the moment of
ð0:85fcíbÞa2 þ ð0:003Aí í
s Es – As fy Þa – ð0:003As Es β1 d # Þ
inertia of the hollow beams is lower than that of the solid
beams, the values of Pcr,test for the tested beams of group E ¼ 0,
showed a continuous increase with an increase in the hole
size up to an HI of 0.28. This observation was also and

Table 5 Load and deflection results of the tested beams

beam Vf (%) HI Pcr (kN) dcr (mm) Pu (kN) du (mm) Pcr /Pu
E0 0 0.00 11.6 1.3 70.5 14.0 0.16
E6 0.16 13.2 1.2 67.5 14.0 0.20
E8 0.28 14.4 1.5 68.3 14.9 0.21
E10 0.44 10.2 1.3 47.5 3.2 0.21
NM0 2 0.00 22.4 1.2 81.3 14.3 0.28
NM6 0.16 18.8 1.2 81.1 10.7 0.23
NM8 0.28 10.3 1.2 87.3 15.6 0.12
NM10 0.44 8.20 0.8 56.4 4.6 0.15
PVA0 2 0.00 11.3 0.7 77.1 14.6 0.15
PVA6 0.16 20.1 1.1 79.6 14.9 0.47
PVA8 0.28 17.7 1.1 79.9 17.5 0.75
PVA10 0.44 13.7 0.9 60.1 2.8 0.39
PP0 3 0.00 14.9 1.4 86.8 21.5 0.17
PP6 0.16 22.5 1.9 66.2 20.0 0.34
PP8 0.28 27.1 2.1 63.7 9 0.43
PP10 0.44 17.9 1.9 49.5 4.6 0.36
 Ahmmad A. ABBAS et al. Flexural behavior of ECC hollow beams 7

Fig. 4 Effect of increasing the hole index (HI) on (a) the cracking load and (b) service load according to Euro code 2.

0:05ðfcí– 28Þ follows:

β1 ¼ 0:85 – ,
7 CECC þ TECC1 þ TECC2 þ Ts ¼ 0, (8)


í β1 d # where
εs ¼ 0:003 1 – ,
a ftí
ðh – zs – cÞ2
CECC ¼
b,
where b, d, and d # are the beam width, effective depth, and 2
nε
zs
distance from the upper fiber of the beam to the center of
the compression reinforcement, respectively. As and Aí s are ftí
nε
zs
TECC1 ¼
b,
the cross-sectional areas of the tension and compression 2
steel, respectively, whereas Es and εí s are the modulus of
elasticity and strain of the compression reinforcement, TECC2 ¼ ftí
½zs ð1 – nε Þ þ c
b,
respectively.
A comparison of the stress block depth with the Ts ¼ As
fy :
thickness of the top flange in the normal concrete hollow
beams showed that the neutral axis is less than the flange The neutral axis is the millstone for calculating the
thickness, which mathematically means that the hollow moment capacity. This can be achieved through the force
beams can be analyzed as a rectangular beam. This is equilibrium according to Eq. (8). The triangular portion of
because the concrete stress block in the tension zone is the tension stress [37] is computed as a result of the linear
neglected, as mentioned above. strain assumption and based on the yield strain ratio (nε)
Based on the residual strength obtained from the flexural between the strain of the ECC (εECC) and the yield strain of
performance test, different cases can be assumed for ECC the reinforcement (εs), where ne= εECC/εs. To find the
beams, where both fibers and reinforcement bars in the neutral axis, Eq. (8) shall be solved in terms of the
tension zone resist the tensile stress, and the linear stress unknown zs.
distribution can be used to represent the compression zone Previous researchers have adopted many values of the
[34–37]. It was found that the flexural maximum yield strain. Lepech and Li [34] found that a 0.02% yield
compressive strain is less than the strain corresponding strain of an ECC is suitable for use in the design
to the peak stress during compression, which is 3400 με considerations. This value was used in the analysis of
[37] . There are three resultant forces acting on the section. ECC beams in the current study. The yield stress
CECC acts on the compression zone resisted by the ECC, corresponding to the mentioned strain was obtained from
and three other resultant forces act on the tension zone an experimental study. For the failure case, the strains
(TECC1, TECC2, and Ts). TECC1 and TECC2 are the resistances obtained from the ECC of different fibers used in the
by ECC according to the shape of the stress block, whereas current research were 0.07 for εNM-ECC and 0.04 for both
Ts is the resistance by the reinforcement bars in the tension εPVA-ECC and εPP-ECC.
zone. The effect of the compression reinforcement is Based on the acting forces on the stress diagram [37],
ignored owing to its small distance from the neutral axis, with the location of the neutral axis obtained from Eq. (8),
which results in an extremely small moment. the beam moment capacity of the solid ECC beams was
Therefore, the equilibrium balance can be introduced as calculated as follows:
8 Front. Struct. Civ. Eng.


2ðh – zs – cÞ
Mu,ECC ¼ – CECC

3


ð1 – nε Þ
zs þ c
þ TECC1
þ nε
zs
2


2nε
zs
þ TECC2
: (9)
3
The ultimate load capacity (Pu) of the tested beams is
shown in Fig. 5. Except for the PP-ECC beams, it is clearly
shown that the Pu of other ECC beams was increased for an
HI value of 0.16. However, these load capacities decreased
with an increase in HI beyond this limit. Different load
capacities were recognized for PP-ECC beams, where Pu
decreased with an increase in the HI. This may be due to Fig. 6 Ultimate load of normal concrete beams and ECC beams
the high bond strength of the PP fibers, which causes a high of different fibers.
cracking load but early yielding of the fibers at the initial
stage. Fig. 8 shows the effect of the hole size on the behavior of
A theoretical analysis of the solid beams showed that the the beams of the same materials.
ECC beams analyzed based on Eq. (9) produced a 13% The loading processes can be divided into three main
improvement over the normal concrete beam analyzed stages based on the response of the beams (deflection).
according to Eq. (7), whereas the experimental study These are the elastic, strain hardening, and failure stages.
showed that the improvements in the Pu of the NM-ECC, In the elastic stage, it can be observed in Fig. 7(a) that the
PVA-ECC, and PP-ECC beams compared to the normal ECC-solid beams exhibited a stiffer behavior than normal
concrete beams were 15%, 9%, and 23%, respectively. concrete beams during the pre-cracking stage. This case is
These improvements were higher for the hollow beams. mostly valid for hollow beams regardless of the size of the
Figure 6 shows a comparison between the ultimate load hole. The superiority among the ECC beams at this stage
strength for the beams made from normal concrete to the was dependent on the HI. For solid beams and beams with
beams of ECCs with different fiber types. an HI of 0.16, the NM-ECC beams showed the highest
dP/dδ among the ECC beams, whereas the PP-ECC
exhibited the lowest dP/dδ. By contrast, the PVA-ECC
beams exhibited the highest dP/dδ for HI values of 0.28
and 0.44.
The elastic zone was then followed by a hardening
plateau (displacement-hardening behavior), and concen-
trated microcracking then started to form at the points
where the highest deflection occurred. For beams with a
hole size of 100 mm (HI = 0.44), there was no hardening
plateau for the normal concrete beam and an extremely
limited plateau for ECC beams. This may be attributed to
the localized shear cracks that were produced close to one
support of the beams with the largest holes (HI = 0.44).
The comparisons show that the normal, NM-ECC, and
PVA-ECC beams with 60 and 80 mm holes have almost
the same hardening plateau, as shown in Figs. 8(a)–8(c),
Fig. 5 Effect of increasing the HI on the ultimate beam strength. whereas for PP-ECC beams, the plateau decreased with an
increase in the hole size (Fig. 8(d)). The longest strain
4.2 Load-deflection behavior hardening was recorded for the PP-ECC solid beam, which
can be attributed to the higher fiber content compared to
The expected deflection for all beams was 0.05 mm, as the other mixtures.
shown in Eq. (6), which is an extremely small value
compared to those recorded for the experimental beams, as 4.3 Flexural ductility and toughness
listed in Table 5. Figure 7 shows the development of the
deflection induced from the applied load as a comparison Calculating the ductility ratios (μ) of the beams is difficult
among the beams made from different materials, whereas because of the difficulty in defining the yield point [38,39].
 Ahmmad A. ABBAS et al. Flexural behavior of ECC hollow beams 9

Fig. 7 Load-deflection curves for (a) solid beams, (b) beams with 60 mm holes, (c) beams with 80 mm holes, and (d) beams with
100 mm holes.

The ductility ratio was determined from load–deflection of solid beams, which was not the case for normal concrete
curves and is defined herein as the deflection correspond- beams. Increasing the hole size to 100 mm led to a
ing to the ultimate load (δu) normalized by the deflection significant decrease in μ for all beams. This can be
corresponding to the load at the yield point (δy), that is, attributed to the localized shear cracks, as previously
(μ = δu/δy). The yield deflection is the deflection corre- described.
sponding to the load obtained from the intersection of the Although the PP fibers have a smaller tensile strength
secant line intersecting the load–deflection curve at 75% of than the other used fibers, the PP-ECC beams showed the
the ultimate load and the horizontal line drawn from the highest ductility ratios compared to other solid beams and
ultimate load, whereas the ultimate deflection corresponds beams with 60-mm holes. On the other hand, NM-ECC
to 80% of the ultimate load at the post-peak zone [39]. and PVA-ECC beams showed similar ductility behaviors.
Figure 9 illustrates the defined deflections of the two The relation governing μ in the HI for the normal concrete
mentioned levels for one of the tested beams in this study hollow beams used in the experiments can be represented
(PP6). The ductility of all the other beams was calculated by Eq. (10), whereas Eq. (11) represents this relation for
using the same procedure. the experimental ECC beams. These regression formulas
Figure 10 shows the ductility ratios (μ) for the four were derived from the experimental results of the current
groups of beams. It is clearly shown that the μ values of study and have determination coefficients of R2 = 0.92 and
ECC beams are noticeably higher than those of normal 0.99 for normal and ECC beams, respectively.
concrete beams for both solid and hollow sections. It is
also shown in the figure that all beams with 60-mm holes N ¼ – 20ðHIÞ2 þ 5:4ðHIÞ þ 2:67, (10)
exhibited a higher μ compared to their corresponding solid
beams, whereas increasing the hole size to 80 mm led to ECC ¼ – 50ðHIÞ2 þ 17:6ðHIÞ þ 3:78: (11)
lower values of μ for all beams. However, the μ values of
the ECC beams with 80-mm holes are comparable to those Toughness is also an important reflector for the ability of
10 Front. Struct. Civ. Eng.

Fig. 8 Load-deflection curves for (a) normal concrete, (b) PVA-ECC, (c) NM-ECC, and (d) PP-ECC beams.

Fig. 9 Ductility parameters for the experimental beam PP6. Fig. 10 Effect of hole index on the beam ductility.

beams to absorb the applied energy. Mathematically, the For the ductility, the toughness values of the ECC beams
toughness can be defined as the total area under the load– were significantly higher than those of the normal concrete
deflection curve up to a certain level of loading. Figure 11 beams. This is not surprising because of the displacement
shows the toughness for the three loading stages. The hardening that characterizes the ECC behavior. Moreover,
toughness of the tested beam up to the yielding point is the Tp contributes a higher percentage to the total
labeled Ty, the toughness up to the peak load is Tp, and the toughness compared to the Ty and Tu-80 percentages for
toughness at up to 80% of the peak load within the post- the normal concrete and PVA-ECC beams, whereas the Tp
peak zone is labeled Tu-80. percentages are shown to be less than those of Tu-80 for the
 Ahmmad A. ABBAS et al. Flexural behavior of ECC hollow beams 11

Unlike that appearing on the compression faces of the

normal concrete beams (Fig. 12(a)), the fiber bridging of
the tension zone and the multi-cracking development of
ECC beams led to lower compression stresses at the
compression surface. Reaching the ultimate beam strength,
the longitudinal reinforcement bars started to yield with a
significant increase in crack opening width and deflection.

5 Conclusions
In this study, the flexural performances of solid and hollow
normal concrete and ECC beams were investigated. Three
types of fibers (NM, PVA, and PP) were used for the ECC
beams, whereas three hole sizes were studied for the three
groups of ECC beams and the normal concrete beams. The
Fig. 11 Toughness of the tested beams. following conclusions summarize the most important
results obtained from the experimental beams.
NM-ECC and PP-ECC beams. This indicates that the 1) Despite their lower moment of inertia, the cracking
contribution of the fibers beyond the peak load is more loads of normal concrete, PVA-ECC, and PP-ECC hollow
effective when NM or PP fibers are used. It should be noted beams with hollowing indices of 0.16 and 0.28 were higher
that the PVA fiber used in this study is a low-cost untreated than their corresponding solid beams, whereas those with
fiber. NM fibers exhibited a continuous decrease in cracking load
with an increase in the hole size. The maximum cracking
4.4 Failure mode load values were recorded for PP-ECC beams, which were
14.9, 22.5, 27.1, and 17.9 kN for hollowing indices of 0,
Figure 12 shows the final cracking patterns of the 16 beams 0.16, 0.28, and 0.44, respectively.
after failure. The first cracks were visually observed and 2) The improvements in the ultimate load capacity of
detected using the load-deflection curves. The location of PVA-ECC, NM-ECC, and PP-ECC solid beams compared
the first crack was found to be dependent on the size of the to the normal concrete solid beam were 9%, 15%, and
hole and the fiber presence. In general, the first crack of the 23%, respectively. These improvements were higher for
solid beams was initiated within the mid-third of the beam hollow PVA-ECC and NM-ECC beams (17%–28%)
span, showing pure flexural cracking. For ECC beams with compared to hollow normal concrete beams. Moreover,
a hole size of 60 mm, the first crack was also a pure flexural the ultimate load capacities of the hollow PVA-ECC and
crack, whereas the corresponding normal concrete beam NM-ECC beams with hollowing indices of 0.16 and 0.28
exhibited shear cracking at the far end of the outer third. By were higher than that of a normal solid concrete beam.
contrast, the first crack was initiated within the outer third 3) The ECC solid beams exhibited stiffer load-deflec-
of the span for PP-ECC and NM-ECC beams with a hole tion behavior than normal concrete beams during the pre-
size of 80 mm, whereas it was within the mid-third of the cracking stage. This case is almost valid for hollow beams
corresponding PVA-ECC beam. For the beams with regardless of the size of the hole. The hollow NM-ECC and
100-mm holes, the dominant cracking and failure pattern PVA-ECC beams with hollowing indices of 0.16 and 0.28
was pure shear cracking. showed almost the same hardening plateaus that were
ECC beams exhibited a much better cracking perfor- comparable or higher than their corresponding solid beams
mance compared to normal concrete beams. As shown in and solid normal concrete beams. By contrast, the
Fig. 12, the PP-ECC and NM-ECC beams showed multi- hardening plateau decreased with an increase in the hole
fine flexural cracking with small crack widths and close size for the PP-ECC beams but with longer hardening
crack spacings. These cracks were then propagated up to plateaus compared to other types of ECC and normal
the flexural reinforcement level with an extremely small concrete beams.
increment in crack width because of the crack bridging 4) Owing to their displacement hardening characteris-
activity of the fibers. Then, during and after the increase in tics, both the ductility and toughness were noticeably
load, the cracks further propagated upward as flexural or higher for the solid and hollow ECC beams compared to
shear–flexural cracks. Based on the opening size, some normal concrete beams. Beams with a hollowing index of
cracks were further propagated to the compression zone 0.16 exhibited a higher ductility compared to their
until failure. However, the normal concrete beams corresponding solid beams, whereas the ductility
exhibited a lower number of wider and far-spaced cracks. decreased as the size of the hole increased. However, the
12 Front. Struct. Civ. Eng.

Fig. 12 Crack pattern of the tested beams for (a) normal concrete, (b) PVA-ECC, (c) NM-ECC, and (d) PP-ECC beams.

ductility of hollow ECC beams with hollowing indices of the Doctoral Degree. Anna Arbor, MI: University of Michigan, 2005
0.16 and 0.28 mm was still noticeably higher than that of 3. Wang Y, Backer S, Li V C. An experimental study of synthetic fiber
the normal solid concrete beam. reinforced cementitious composites. Journal of Materials Science,
1987, 22(12): 4281–4291
Acknowledgements This experimental study was conducted in the
4. Li V C. Tailoring ECC for special attributes: A review. International
structural laboratory in the civil engineering department of Gaziantep
University, Turkey. The authors would like to express their gratitude to the Journal of Concrete Structures and Materials, 2012, 6(3): 135–144
Department of Civil Engineering at Gaziantep University and the staff of the 5. Kanda T, Li V C. Practical design criteria for saturated pseudo strain
structural laboratory for their support. hardening behavior in ECC. Journal of Advanced Concrete
Technology, 2006, 4(1): 59–72
6. Li V C, Wang S, Wu C. Tensile strain-hardening behavior of PVA-
References ECC. ACI Materials Journal, 2001, 98: 483–492
7. Li V C, Wu C, Wang S, Ogawa A, Saito T. Interface tailoring for
1. Michael D L, Victor C L. Large-scale processing of engineered strain-hardening PVA-ECC. ACI Materials Journal, 2002, 99: 463–
cementitious composites. ACI Materials Journal, 2008, 105: 358– 472
366 8. Said S H, Razak H A, Othman I. Flexural behavior of engineered
2. Wang S. Micromechanics based ECC matrix design. Dissertation for cementitious composite (ECC) slabs with polyvinyl alcohol fibers.
 Ahmmad A. ABBAS et al. Flexural behavior of ECC hollow beams 13

Construction & Building Materials, 2015, 75: 176–188 2014, 51: 163–178
9. Mohammed B S, Aswin M, Beatty W H, Hafiz M. Longitudinal 24. Al.Khuzaie H M A, Atea R S. Investigation of torsional behavior
shear resistance of PVA-ECC composite slabs. Structures, 2016, 5: and capacity of reactive powder concrete (RPC) of hollow
247–257 T-beam. Journal of Materials Research and Technology, 2019, 8
10. Algburi A H M, Sheikh M N, Hadi M N S. Mechanical properties of (1): 199–207
steel, glass, and hybrid fiber reinforced reactive powder concrete. 25. Murugesan A, Narayanan A. Deflection of reinforced concrete
Frontiers of Structural and Civil Engineering, 2019, 13(4): 998– beams with longitudinal circular hole. Practice Periodical on
1006 Structural Design and Construction, 2018, 23(1): 04017034
11. Zhang Q, Li V C. Development of durable spray applied fire- 26. Abbass A, Abid S, Özakça M. Experimental investigation on the
resistive Engineered Cementitious Composites (SFR-ECC). Cement effect of steel fibers on the flexural behavior and ductility of high-
and Concrete Composites, 2015, 60: 10–16 strength concrete hollow beams. Advances in Civil Engineering,
12. Hung C C, Chen Y S. Innovative ECC jacketing for retrofitting 2019, 2019: 1–13
shear-deficient RC members. Construction & Building Materials, 27. Abbass A A, Abid S R, Arna’ot F H, Al-Ameri R A, Özakça M.
2016, 111: 408–418 Flexural response of hollow high strength concrete beams
13. Ge W J, Ashour A F, Ji X, Cai C, Cao D F. Flexural behavior of considering different size reductions. Structures, 2020, 23: 69–86
ECC-concrete composite beams reinforced with steel bars. Con- 28. Lepech M D, Li V C. Large-Scale processing of engineered
struction & Building Materials, 2018, 159: 175–188 cementitious composites. ACI Materials Journal, 2008, 105: 358–
14. Qiao Z, Pan Z, Xue W, Meng S. Experimental study on flexural 366
behavior of ECC/RC composite beams with U-shaped ECC 29. Zhang R, Matsumoto K, Hirata T, Ishizeki Y, Niwa J. Application of
permanent formwork. Frontiers of Structural and Civil Engineering, PP-ECC in beam–column joint connections of rigid-framed railway
2019, 13(5): 1271–1287 bridges to reduce transverse reinforcements. Engineering Structures,
15. Kanda T, Watanabe S, Li V C. Application of pseudo strain 2015, 86: 146–156
hardening cementitious composites to shear resistant structural 30. Li V. Engineered Cementitious Composite (ECC). Concrete
elements. In: Fracture Mechanics of Concrete Structures Proceed- Construction Engineering Handbook. CRC Press, 2008, 1584
ings FRAMCOS-3, D-79104. Freiburg: IA-FraMCoS, 1998, 1477– 31. Johnston C D. Significance of tests and properties of concrete.
1490 ASTM Special Technical Publication, 1994, 169C: 547–561
16. Qudah S, Maalej M. Application of engineered cementitious 32. Zahiri F, Eskandari-Naddaf H. Optimizing the compressive strength
composites (ECC) in interior beam–column connections for of concrete containing micro-silica, nano-silica, and polypropylene
enhanced seismic resistance. Engineering Structures, 2014, 69: fibers using extreme vertices mixture design. Frontiers of Structural
235–245 and Civil Engineering, 2019, 13(4): 821–830
17. Abid S R, Hilo A N, Daek Y H. Experimental tests on the 33. International Federation for Structural Concrete (fib). MC2010.
underwater abrasion of engineered cementitious composites. Con- Model Code 2010, Final Draft. Lausanne: fib, 2012, 65–66
struction & Building Materials, 2018, 171: 779–792 34. Lepech M, Li V C. Design and field demonstration of ECC link
18. Abid S R, Shamkhi M S, Mahdi N S, Daek Y H. Hydro-abrasive slabs for jointless bridge decks. In: Proceedings of CONMAT-5.
resistance of engineered cementitious composites with PP and PVA Ann Arbor, MI: University of Michigan, 2005, 1–10
fibers. Construction & Building Materials, 2018, 187: 168–177 35. Krolicki J, Maffei J, Calvi G M. Shear strength of reinforced
19. Ferdous W, Manalo A, Aravinthan T, Fam A. Flexural and shear concrete walls subjected to cyclic loading. Journal of Earthquake
behaviour of layered sandwich beams. Construction & Building Engineering, 2011, 15(sup1): 30–71
Materials, 2018, 173: 429–442 36. Campione G, Letizia Mangiavillano M. Fibrous reinforced concrete
20. Ferdous W, Almutairi A D, Huang Y, Bai Y. Short-term flexural beams in flexure: Experimental investigation, analytical modelling
behaviour of concrete filled pultruded GFRP cellular and tubular and design considerations. Engineering Structures, 2008, 30(11):
sections with pin-eye connections for modular retaining wall 2970–2980
construction. Composite Structures, 2018, 206: 1–10 37. Qi J, Wang J, Ma Z J. Flexural response of high-strength steel-ultra-
21. Ferdous W, Bai Y, Ngo T D, Manalo A, Mendis P. New high-performance fiber reinforced concrete beams based on a
advancements, challenges and opportunities of multi-storey mod- mesoscale constitutive model: Experiment and theory. Structural
ular buildings—A state-of-the-art review. Engineering Structures, Concrete, 2018, 19(3): 719–734
2019, 183: 883–893 38. Park R. Ductility evaluation from laboratory and analytical testing.
22. Alnuaimi A S, Al-Jabri K S, Hago A. Comparison between solid and In: Proceedings of the 9th World Conference on Earthquake
hollow reinforced concrete beams. Materials and Structures, 2008, Engineering. Tokyo, Japan, IAEE, 1988, 605–616
41(2): 269–286 39. Pam H J, Kwan A K H, Islam M S. Flexural strength and ductility of
23. Lopes S M R, Bernardo L F A. Cracking and failure mode in HSC reinforced normal and high-strength concrete beams. Structures and
hollow beams under torsion. Construction & Building Materials, Buildings, 2001, 146(4): 381–389