International Journal of Impact Engineering 71 (2014) 89e96

Contents lists available at ScienceDirect

International Journal of Impact Engineering

journal homepage: www.elsevier.com/locate/ijimpeng

Fracture behaviour of steel fibre-reinforced concrete at a wide range of

loading rates
X.X. Zhang*, A.M. Abd Elazim, G. Ruiz, R.C. Yu
ETSI de Caminos, Canales y Puertos, Universidad de Castilla-La Mancha, 13071 Ciudad Real, Spain

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

Article history: Three-point bending tests on notched beams of steel fibre-reinforced concrete (SFRC) have been con-
Received 25 June 2013 ducted using both a servo-hydraulic machine and a drop-weight impact device. The shape and geometry
Received in revised form of the beams followed the RILEM recommendation, i.e., 150 mm
150 mm in cross section, 700 mm in
29 January 2014
length, notch-depth ratio was around 1/6 and the span of the tests was kept constant 500 mm. The peak
Accepted 26 April 2014
load and the fracture energy were measured over a wide range of loading rates (loading point
Available online 9 May 2014
displacement rates), spanning six orders of magnitude. Under low loading rates, from 103 mm/s to
100 mm/s, the tests were performed with the servo-hydraulic machine; from 102 mm/s to 103 mm/s, the
Keywords:
Steel fibre-reinforced concrete
drop-weight impact machine was used instead. The results show that the fracture energy and the peak
Loading rates load increase as the loading rate increases. Furthermore, such a trend is relatively mild under low rates.
Drop-weight impact machine The gain of the fracture energy and peak load is around 10% compared with its quasi-static values. It
Fracture energy could be attributed to viscous effects mainly originated by the presence of free water in voids and porous
structures in the matrix, and the weak fibre pullout resistance. However, under high rates the increases
in the fracture energy and the peak load are pronounced due to the inertia effect and the greater fibre
pullout energy. The dynamic increase factors of the peak load and the fracture energy for the fastest
loading rate are approximately 3.5 and 2.5, respectively.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
Steel fibre-reinforced concrete (SFRC) is concrete made of hy-
draulic cements containing fine or fine and coarse aggregate and
discontinuous discrete steel fibres [1]. Addition of randomly
distributed steel fibres improves concrete properties, such as static
flexural strength, ductility and flexural toughness. Some examples
of structural and nonstructural uses of SFRC are hydraulic struc-
tures, airport and highway paving and overlays, industrial floors,
refractory concrete, bridge decks, shotcrete linings and coverings,
and thin-shell structures [2].
Besides bearing quasi-static loads, many concrete structures are
subjected to short duration loads, such as the impacts from missiles
and projectiles, wind gusts, earthquakes, and machine dynamics.
Several techniques have been developed to study the dynamic
fracture behaviour of concrete and concrete structures, like modi-
fied Charpy impact test, Split Hopkinson pressure bar test, drop-
weight impact test and explosive test [3,4].
* Corresponding author. Tel.: þ34 926295300 Ext.3261; fax: þ34 926 295 391.
E-mail addresses: zhangxiaoxinhrb@gmail.com, xiaoxin.zhang@uclm.es (X.
X. Zhang).
Many researchers have shown that the impact resistance can be
increased substantially with the addition of randomly distributed
steel fibres to concretes. For instance, Namman and Gopalaratnam
[5] have used a drop-weight impact machine and an Instron uni-
versal machine to study the bending properties of steel fibre-
reinforced mortar beams at four different loading rates,
4.23
104 mm/s, 8.46 mm/s, 7.0
102 mm/s and 1.0
103 mm/s,
respectively. The dimension of the beams was
12.5 mm
75 mm
300 mm (width
depth
length) and the
span was 254 mm during the tests. Three volume fractions of fibres
(1%, 2% and 3%), three fibre aspect ratios (47, 62 and 100) were
adopted. The results showed that depending on the fibre rein-
forcing parameters the energy absorbed by the composite at static
loading rates can be one to two orders of magnitude higher than
that of the unreinforced matrix. Moreover, up to a three times in-
crease was observed in the modulus of rupture and the energy
absorbed by the composite when the loading rate increases from
4.23
104 mm/s to 1.0
103 mm/s. The similar tendency was also
obtained by Gopalaratnam and Shah [6].
Banthia [7] has adopted a drop-weight impact machine to
conduct dynamic tests on steel fibre-reinforced high-strength
concrete beams at loading rate around 4 m/s, the specimen was

http://dx.doi.org/10.1016/j.ijimpeng.2014.04.009
0734-743X/Ó 2014 Elsevier Ltd. All rights reserved.
90 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96
150 mm in width, 150 mm in depth and 1500 mm in length over a
span 960 mm. The peak bending load and fracture energy got 498%
and 640% increases, respectively, compared with their quasi-static
values. Later, a modified Charpy impact machine was designed to
test concrete and SFRC in uniaxial tension [8]. The results showed
that under impact, the higher strength of the matrix, the less
effective the fibres in improving fracture energy absorption.
ACI Committee 544 has proposed another type of drop-weight
impact test for evaluating the impact resistance of fibre-
reinforced concrete (FRC) [3], i.e., a hammer is dropped repeat-
edly to impact a disc specimen, and the number of blows required
to cause the first visible crack on the top and to cause ultimate
failure are both recorded. This method is designed to obtain relative
performance of plain concrete and FRC. Nataraja et al. [9] analysed
the statistical variation of impact resistance of SFRC under this type
of repeated drop impact condition. The disc samples were 150 mm
in diameter and 64 mm in thickness and containing 0.5% volume
fraction of round crimped steel fibres with 0.5 mm in diameter and
55 in aspect ratio. The observed coefficients of variation were 57%
and 46% for the first crack resistance and ultimate resistance,
respectively.
Most studies cited above concentrated on unnotched speci-
mens. Notched samples, though, provide results with less scatter,
since the notch cuts the bottom layer of the beam, where fibres
have a preferential orientation due to the presence of the bottom
face of the mould. Besides, the notch concentrates the stresses and
thus it usually leads to a single main crack that propagate from the
notch tip towards the loading point, nonlinear deformation being
negligible in the rest of the specimen. Thus, in 2002, the final
recommendation for the bending test for SFRC by the RILEM TC
162-TDF Committee was to use notched prisms (150 mm
150 mm
in cross section) with central point loading [10]. One of the greater
advantages of such configuration is that it guarantees the stability
during the test even for FRC with low fibre contents. In 2005, a
similar method was also proposed by EN 14651 standard and, two
years later, an updated version was available [11].
In order to get additional insights into the loading rate effect on
the fracture properties of SFRC, in this paper we present three-point
bending tests of notched specimens conducted at a wide range of
loading point displacement rates (for simplicity, it is substituted by
loading rates), from 103 mm/s to 103 mm/s, using both a servo-
hydraulic testing machine and a drop-weight impact instrument.
The shape and geometry of the beams follow RILEM recommen-
dations [10]. The results show that the fracture energy and the peak
load increase as the loading rate increases. Moreover, such a trend
is relatively mild under low rates. While it is pronounced under
high loading rates due to the inertia effect.
The rest of this paper is structured as follows: the experimental
procedure is given in Section 2, in Section 3 the results are pre-
sented and discussed. Finally, relevant conclusions are drawn in
Section 4.
2. Experimental procedure
2.1. Material characterization
A single type of SFRC was used throughout the experiments,
made with a siliceous aggregate of 12 mm in maximum size and
ASTM type II cement, 42.5R. Two types of super plasticizer (Gle-
nium ACE-325, Glenium B-255) were used in the concrete
composition. The mixing proportions by weight were 1: 0.29: 0.1:
0.048: 1.6: 1.2: 0.0066: 0.021 (cement: water: limestone filler: silica
fume: sand: coarse aggregate: nanosilica: superplasticizer).
64.5 kg/m3 of steel fibre were added as the reinforcement. The steel
Table 1a
Properties of the SFRC at an age of 65 days.
fc (MPa) E (GPa) n r (kg/m3)
Mean 92 35 0.18 2438
Standard deviation 5 1 0.01 17
Table 1b
Properties of the matrix material at an age of 65 days.
fc (MPa) E (GPa) n r (kg/m3) ft (MPa) GF (N/m)
Mean 76 32 0.18 2297 5.1 145
Standard deviation 3 1 0.01 6 0.7 18
fibre used was hooked-end with 50 mm in length, 0.75 mm in
diameter and 67 in aspect ratio, it has a tensile strength 1900 MPa.
Compressive tests were carried out at an age of 65 days,
following ASTM C39 (which is analogous to EN 12390-3), cylinders
of 150 mm
300 mm (diameter
height) were used. To perform
the compressive tests, a Servo-hydraulic testing machine was
adopted. Values of characteristic parameters: compressive strength
(fc), elastic modulus (E) and Poisson’s ratio (n) are listed in Table 1a.
Moreover, the properties of the matrix material are also presented
in Table 1b as a reference, ft and GF are tensile strength and fracture
energy, respectively.
2.2. Three-point bending tests
To study the mechanical properties of the SFRC prisms, three-
point bending tests were conducted on notched beams over a
wide loading rate range from 103 to 103 mm/s. Two testing ma-
chines were adopted to carry out the tests, one was a servo-
hydraulic testing machine and the other was a drop-weight
impact instrument.
The dimensions of the test beams were 150 mm
150 mm
(B
D) mm in cross-section, and 700 mm in total length (L). The
initial notch-depth ratio (a/D) was approximately 1/6, and the span
(S) was fixed at 500 mm during the tests, see Fig. 1.
2.2.1. Tests under low loading rates 103 to 100 mm/s
Within this low loading rate range, the tests were performed
using the servo-hydraulic testing machine under position control
coupled to a robust frame in the Laboratory of Materials and
Structures of the Civil Engineering School of the University of
Castilla-La Mancha in Ciudad Real as shown in Fig. 2. The alignment
of the supports and the loading line was checked when installing
each specimen. One of the supports was fixed, whereas the other
was free to rotate around the axis of the frame.
Three loading rates, from quasi-static level (3.33
103 mm/s)
to rate dependent levels (0.1 mm/s and 3.33 mm/s), were applied.
Six specimens were tested at each loading rate.
Force
Fig. 1. Schematic diagram of the specimen.
 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96 91

Side view Front view

Actuator and Load cell

Platens

Profiled beam

Loading bar

D = 150 mm Specimen

Span =
500 mm
L = 700 mm Rotating support
Fixed support

Steel beam

(a) (b)
Fig. 2. (a) Schematic diagram and (b) photo of the experimental set-up of the servo-hydraulic testing machine.

2.2.2. Tests under loading rates from 102 to 103 mm/s
In this high loading-rate range, all tests were conducted using
the instrumented, drop-weight impact apparatus as shown in
Fig. 3. It has the capacity to drop a 316 kg mass from heights of up to
2.6 m, and can accommodate flexural specimens with spans of up to
approximately 1.6 m. In this research, an impact hammer weighting
120.6 kg was employed and three-drop heights adopted were 40,
160 and 360 mm due to the fact that the capacity of the force
sensors installed so far on the supports of the machine was only
89 kN, thus, the maximum impact energy (the drop height
360 mm) was not enough to break the strong beam completely.
Nevertheless, it is still meaningful for studying the fracture

HOIST MACHINE
CONTROLS

MAGNETIC

SENSOR
HAMMER
DATA
ACQUISITION

TUP SYSTEM
MAGNETIC

STRIP
ACCELEROMETER

STEEL BAR
COLUMN COMPUTER
OPTICAL FIBER
PHOTOELECTRIC
SENSOR

BEAM

(a) (b)
Fig. 3. (a) Schematic diagram and (b) photo of the drop-weight impact machine.
92 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96

200 200
Reaction force
Impact force
160
160

120
Load (kN)

120

Load (kN)
80
80
start point of
40 the impact force
40
start point of
0 the reaction force
344 μs
0
0.018 0.019 0.020 0.021 0.022 0.023 0.024 0.025 0.0 0.5 1.0 1.5 2.0
Time (s) Loading-point displacement (mm)

(a) Loading rate 8.85×102 mm/s

200 200

160 160

120
120
Load (kN)

Load (kN)
80
80
start point of
40 the impact force
40
start point of
0 the reaction force
250 μs
0
0.007 0.008 0.009 0.010 0.011 0.012 0.013 0.014 0.0 0.5 1.0 1.5 2.0
Time (s) Loading-point displacement (mm)
(b) Loading rate 1.77×10 3 mm/s
200 200

160 160

120
120
Load (kN)

Load (kN)

80
80
start point of
40 the impact force
40

0 start point of
248 μs the reaction force
0
0.005 0.006 0.007 0.008 0.0 0.5 1.0 1.5 2.0
Time (s) Loading-point displacement (mm)
(c) Loading rate 2.66×10 3 mm/s

Fig. 4. Typical impact and reaction forces versus time (left), and comparison of load versus displacement (right), at loading rates (a) 8.85
102 mm/s (b) 1.77
103 mm/s and (c)
2.66
103 mm/s.

behaviour of this SFRC. The corresponding impact velocities were In order to calculate dynamic fracture energy, an “equivalent
8.85
102 mm/s, 1.77
103 mm/s and 2.66
103 mm/s, respec- static force” method was proposed due to the fact that the
tively. These force sensors used are PCB piezoelectric ones, espe- analytical solution is complex even for this relative simple impact
cially for the measurements at impact loading conditions with a problem. Namely, the reaction force from the supports is taken as
wide broadband resolution (1e10,000 Hz). Six specimens were the equivalent bending load without inertia effect on the beam
tested at each impact speed. A detailed description of the instru- when the specimen can be modelled as a single degree of
ment is given in references [4,12]. freedom system. In other words, an approximation of the beam
The impact force between the hammer tup and the specimen is response at impact loading conditions can be obtained by
measured by a piezoelectric force sensor. Moreover, the reaction assuming that all the kinetic energy is transferred to the first
force is determined by two force sensors located between the mode of vibration of the beam, the mode is similar to the static
supports and the specimen. An accelerometer bonded to the impact one but has greater flexural strength and smaller shear strength
hammer was used to measure acceleration and displacement dur- [13]. Banthia [7] verified the effectiveness of this method by
ing the impact process. comparing the simplified analytical and experimental results for
For dynamic three-point bending tests, it is well known that a the drop-weight impact three-point bending tests. When the
part of the impact force keeps the balance with the inertia force, failure pattern changes to shear or local ones instead of flexure,
while another part is used to deform and fracture the specimen. the method would not be valid any more.
 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96 93

Fig. 5. Failure modes of SFRC beams under different loading rates.

In our case, all the beams failed in flexure at a wide range of

loading rates. Then, the “equivalent static force” method was
adopted to analyse the results. As mentioned before, the system
becomes exactly the equivalent static one, structural and inertia
effects are removed, only material behaviour is taking effects.
200 -3
During the test the specimen was not broken completely, thus,
3.33×10 mm/s the area under the load versus displacement at mid-span curve up
180 0.1 mm/s to a specified deflection was used in the evaluation of fracture
3.33 mm/s energy (GF), method that was also adopted in some other references
160 2
8.85×10 mm/s [5,6,8,14,15]. Here, a cut-off point was chosen at the displacement
3
140 1.77×10 mm/s (central deflection) of 2 mm for all the tests under low and high
Load (kN)

3
2.66×10 mm/s loading rates. GF was determined by using Eq. (1) given by RILEM
120
50-FMC Technical Committee [16].
100
80 W0 þ mg SL ds
GF ¼ (1)
BðD  aÞ
60
40 where W0, B, D, a, S, L, m, ds and g are the area under the experi-
mental loadedisplacement curve, width, depth, notch, span,
20 length, mass, specified deflection of the beam (2 mm) and gravi-
0 tational acceleration, respectively. Under dynamic loading condi-
0.0 0.5 1.0 1.5 2.0 tions, W0 was obtained by the area under the reaction forcee
Displacement (mm) displacement (loadedisplacement) curves, where the reaction
force is evaluated by adding the values from both support data
Fig. 6. Loadedisplacement curves at different loading rates. points as proposed in references [12,17].
94 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96

Table 2
Experimental results at different loading rates.
$
Testing machine H (mm) d (mm/s) Pmax (kN) DIF for Pmax GF (N/m) DIF for GF Number of fibres

Servo-hydraulic testing machine e 3.33
103 49.5 (11) 1 4432 (1022) 1 267 (59)
e 0.1 55.4 (4) 1.119 4590 (566) 1.04 283 (56)
e 3.33 55.5 (6) 1.121 4676 (455) 1.06 253 (46)
Drop-weight impact machine 40 8.85
102 91.6 (6) 1.85 e e 266 (31)
160 1.77
103 134.1 (6) 2.71 8089 (782) 1.83 264 (14)
360 2.66
103 172.5 (11) 3.48 11,167 (1391) 2.52 248 (32)

Note: values in parentheses are standard deviations.

3. Results and discussion
3.1. Fracture behaviour at a wide range of loading rates
Typical impact and reaction forces versus time curves are shown
in Fig. 4 (left column). It is worth noting that the time intervals
between the start points of the impact force and the reaction force
are 344, 250, and 248 ms, corresponding to loading rates 8.85
102,
1.77
103, and 2.66
103 mm/s respectively. However, it only takes
approximately 118 ms for the shear stress wave to travel from the
impact point to the support point. Thus, there is an apparent time
delay. The reason could be the small gaps between the pieces
consisting the supports and the specimen though the contact be-
tween them is good by eyesight. It is obvious that the time delay
decreases with the increase in the loading rate, on the contrary, the
peak loads of the impact and the reaction forces increase.
Fig. 4 (right column) shows the comparison between the impact
and the reaction forces versus displacement curves. The initial time
of the reaction force has been shifted, thus, the impact and the
reaction forces have the same starting points. For the tests under
drop height 40 mm (loading rate: 8.85
102 mm/s), the impact
energy was too small to deform and fracture the beam apparently
as shown in Fig. 5(d), but the specimen had an obvious main crack
along the ligament. Hence, the beam is almost like a fixed end to the
impact hammer. According to the stress wave theory, the peak of
the reaction force (95.2 kN) is greater than that of the impact force
(70.4 kN). Furthermore, for the tests under drop heights 160 mm
(loading rate: 1.77
103 mm/s), the beam was deformed and
fractured clearly but not broken completely as shown in Fig. 5(e).
The beam undergoes first an acceleration, and then a deceleration
before it finally comes to rest. During the deceleration, the peak
load of the reaction force will be greater than that of the impact
force during this period due to the inertia effect. When the drop
height of the hammer was increased to 360 mm (loading rate:
2.66
103 mm/s), the beam was almost broken entirely as shown in
Fig. 5(f), consequently the peak of the reaction force (164.8 kN) is
less than that of the impact force (190.1 kN). The former is around
87% of the latter, i.e., majority of the impact force is used to fracture
the specimen, only a small portion is keeping balance with the
inertia force.
Failure modes of beams under low loading rates are shown in
Fig. 5(a, b, c), it can be observed that there are more branch cracks
around the main crack compared with the crack pattern of beams in
Fig. 5(d, e, f) under high loading rates. Moreover, all beams were not
broken completely, it is clear that the post-peak fracture behaviour
of the beam was greatly improved by the addition of steel fibres.
Furthermore, all the fibres in the crack surface are pulled out, no
single broken fibre is found.
Fig. 6 shows the comparison of the typical loadedisplacement
curves at different loading rates. It can be observed that the peak
load increases with increase in loading rates. However, the stiffness
of the beam does not show a similar tendency, which is due to the
sensitivity of the elastic flexibility of the beam to the boundary
conditions during the application of the concentrated load as pro-
posed in reference [18]. Nevertheless, the stiffness still gains a
sound increase when the loading rates have a big jump, i.e., from
low to high loading rates. Fig. 6 also indicates that the chosen cut-
off value for obtaining GF is adequate for this particular geometry
and material, since ds falls in the stretch of the loadedisplacement
curve following the softening part after the peak. Indeed, according
to Fig. 6, any ds greater than 1.5 mm would result in a similar rate
tendency in fracture energy due to the fact that the post-softening
behaviour of SFRC mainly depends on the pullout of fibres from the
matrix in our case.
Table 2 provides detailed information about the experimental
results. The dynamic increase factor (DIF) is defined by the ratios of
the peak load (Pmax) and fracture energy to their corresponding
quasi-static values. Here, the lowest loading rate

Fig. 7. (a) Loading rate dependence of the peak load (b) loading rate dependence of the fracture energy.
 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96 95

Table 3
Comparison of experimental results.

Suaris and Shah, Naaman and Gopalaratnam, Gopalaratnam and Shah, Banthia, Current work
1983, [14] 1983, [5] 1986, [6] 1987, [7]

Specimen size 38.1
76.2
381 12.5
75
254 25
64
203 150
150
960 150
150
500
B
D
Sa (mm)
Notch a (mm) 0 0 12.5 0 25
Fibre content 1% 1% 1% 1.5% 0.8%
(volume ratio)
Mix proportions 1:2:0:0.5 1:2:0:0.4 1:2:0:0.5 1:2:3.5:0.5 1:1.6:1.2:0.29
C:S:A:Wb
l/dc (Aspect ratio) 25.4/0.254 (100) 25/0.25 (100) 25.8/0.41 (63) 60/0.6 (100) 50/0.75 (67)
Steel fibre shape Hooked-end Smooth brass-coated Smooth brass-coated Hooked-end Hooked-end
Dynamic Drop weight impact Drop weight impact Modified Charpy Drop weight impact Drop weight
experimental set-up impact
Hammer weight (kg) 108.9 e e 42.5 120.6
Max. velocity (m/s) 1.0 1.0 2.45 3.79 2.66
fc (MPa) 57.5 64.7 30.4 50.0 79.3
E (GPa) e e 29.2 e 31.0
Pmax (kN) 7.1, (Quasi-static value 3.7) 3.9, (Quasi-static value 1.4) 3.9, (Quasi-static value 2.0) 57.3, (Quasi-static 172.5, (Quasi-static
value 11.5) value 49.5)
GF (N/m) at a central deflection of at a central deflection of at a central deflection of 12,724, (Quasi-static at a central
12.5 mm 16,471, 12.5 mm 5082, (Quasi-static 2.54 mm 3010, (Quasi-static value 1991) deflection of 2.0 mm
(Quasi-static value 8277) value 2229) value 1710) 11,167, (Quasi-static
value 4432)
DIF for Pmax 1.92 2.79 1.95 4.98 3.48
DIF for GF 1.99 2.28 1.76 6.40 2.52
a
B
D
S; Width
Depth
Span.
b
C:S:A:W; Cement:Sand:Aggregate:Water, mix proportions by weight.
c
l/d; Length/diameter of the steel fibre.
$
(d ¼ 3.33
103 mm/s) is taken as the quasi-static loading condi- the low rate range. It could be attributed to viscous effects mainly
tion. H stands for the drop height of the hammer under impact originated by the presence of free water in voids and porous
loading conditions. It is worth noting that the fracture energy under structures in the matrix [20], and also the weak pullout behaviour
drop height 40 mm was not calculated due to the fact that the between the fibre and the matrix. However, under impact loading
maximum displacement was only around 0.5 mm, it did not reach rates, the rate effect is pronounced. On the one hand, the micro-
the specified deflection 2 mm as mentioned before. inertia around the crack tip makes fracture propagation more
Fig. 7(a) shows the loading rate effect on the peak load. It is difficult [21-25]. On the other hand, hooked-end steel fibres
evident that the peak load increases with increase in loading rates. embedded in concrete matrix support a higher load under impact
It should be also noted that the tendency is minor under low and the pullout energy is also greater, the rate effect is remarkable
loading rates, while it is pronounced instead under high loading [26].
rates. A prediction equation for this rate effect is derived from the The loading rate effect on the fracture energy is also shown in
experimental results as shown in Eq. (2). Fig. 7(b). The trend is similar to that of the peak load, i.e., the ten-
dency is moderate under low loading rates, while under high
0 $ 1n loading rates it is dramatic. A similar equation is also proposed to
. d represent this behaviour, see Eq. (3).
S
DIFP ¼ Pmax Pmax ¼ Pmax =53:3 ¼ 1 þ k@ $ A
d0 0 $ 1r
0 11:01  d
 
$
$
DIFGF ¼ GF GsF ¼ GF =4565:9 ¼ 1 þ m@ $ A
¼ 1 þ 7:5
10 4 @ $d A ; for d in mm=s (2) d0
d0 0 $ 11:54
  d $
S
$ ¼ 1 þ 7:6
106 @ $ A ; for d in mm=s (3)
where
$ Pmax is the static peak load in kN, d is the loading rate in mm/
d0
s, d0 is set as 1 mm/s thus, those adjustment parameters k and n are
without units. The equation can be used to efficiently predict the where GSF is the static fracture energy, 4565.9 N/m in our case.
rate effect on the peak load and could also be helpful in performing Coefficients m$ and r are adjusting parameters without units due to
numerical simulations. Moreover, the fitting curve gives the static the fact that d0 is set as 1 mm/s as mentioned before.
value of the peak load (53.3 kN) as well, which could only be ob-
tained by a purely static test.
Since both steel and concrete are strain-rate sensitive mate- 3.2. Comparison with data already published
rials, the pullout behaviour of a fibre can be expected to be
affected by the rate at which the pullout occurs. Furthermore, in Table 3 provides a comparison of some experimental results
this study, the rate effect on the fracture behaviour of SFRC should obtained only at highest loading rates from other researches and
be mostly related with the behaviour of matrix and pullout of fi- the results presented in this paper. It is worthwhile to note that
bres due to the fact that no single broken fibre was found in the the fracture behaviour of SFRC was investigated at a wide range of
crack surface. Under low loading rates, from 0.018 mm/s to loading rates in our case, from 103 to 103 mm/s. Moreover, the
18 mm/s, high-strength hooked-end fibres showed no appreciable width of the beam adopted in references [5,6,14] was not appro-
rate sensitivity when pulled out from matrix [19]. In our case, the priate according to the RLEM TC 162-TDF, i.e., the smallest spec-
rate effect on the fracture behaviour of the SFRC is slight as well in imen dimension should be at least 2.5 times larger of the fibre
96 X.X. Zhang et al. / International Journal of Impact Engineering 71 (2014) 89e96
length. Furthermore, in reference [7], unnotched specimen was
used for the test though the cross section of the beam followed the
codes. Thus, in order to overcome these shortcomings, the stan-
dard notched prisms recommended by RILEM and EN 14651 codes
were adopted for our tests. Nevertheless, from this table, it can be
still observed that the loading rate effect on the mechanical
properties of different steel fibre-reinforced concretes is similar in
spite of the differences in the type of specimen and experimental
set-up.
The values of DIF for the peak load and the fracture energy are in
the same order of magnitude, though there is some difference
among them. Such difference is likely attributable to type, shape
and geometry of the steel fibre, fibre volume ratio, bond behaviour
between fibres and matrix, size of the specimen and the testing
conditions. Furthermore, it should be taken into account as well
that the notched specimen absorbs less dissipated energy than that
of the unnotched one due to the fact that the crack locates close to
the notch plane. Thus, it is better to use notched specimen to study
the fracture behaviour of SFRC for reducing the influence of dissi-
pated energy. Recently, Caverzan and Cadoni et al. [27] have
investigated the dynamic fracture behaviour of a high performance
fibre-reinforced cementitious composite by using a modified
Hopkinson bar, the notched cylinder specimens were used in the
test, the loading rate was 3 m/s. The results showed that the DIF
value for the fracture energy was approximately 1.5. It is worth
noting that the rate effect on the fracture energy illustrates similar
tendency with different loading methods, such as modified Charpy
impact testing machine, drop-weight impact instrument and
modified Hopkinson bar system.
4. Conclusions
The fracture behaviour of a steel fibre-reinforced concrete under
dynamic loading conditions was investigated in this study. As
opposed to previous research works, we used the standard notched
prisms that are recommended for measuring the flexural strength.
Besides, the loading rates varied considerably from a quasi-static
level to a dynamic level, the order of magnitude changed from
103 to 103 mm/s. As a result of the study, the following conclusions
can be drawn:
The peak load and the fracture energy are sensitive to the
loading rate. Under low loading rates, the rate effect is minor, while
it is pronounced under high loading rates. The dynamic increase
factor of the peak load and the fracture energy is approximately 3.5
and 2.5, respectively, for the highest impact velocity. This is
attributed primarily to the strain rate sensitivity of the matrix and
the pullout resistance of the fibres.
Two prediction-equations for the rate sensitivity of the fracture
energy and the peak load are provided. They can be helpful in
numerical simulations that evaluate the rate dependence of the
fracture behaviour.
Acknowledgements
We are grateful for the financial support received under grant
No. MAT2012-35416 from the Ministerio de Economía y
Competitividad, Spain. We also thank the INCRECYT program and
the FPI fellowship No. BES-2008-008216.
References
[1] ACI Committee 544. State-of-art report on fiber reinforced concrete, ACI
Committee 544 report 544.1R-96. Detroit: American Concrete Institute; 1996.
p. 7.
[2] ACI Committee 544. Guide for Specifying, proportioning, and production of
fiber reinforcing concrete, ACI Committee 544 report 544.3R-08. Detroit:
American Concrete Institute; 2008. p. 2.
[3] ACI Committee 544. Measurement of properties of fiber reinforced concrete
(Reapproved 2009). ACI Committee 544 report 544.2R-89. Detroit: American
Concrete Institute; 1989. pp. 6e8.
[4] Zhang XX, Ruiz G, Yu RC. A new drop-weight impact machine for studying
fracture processes in structural concrete. Strain 2010;46:252e7.
[5] Naaman AE, Gopalaratnamt VS. Impact properties of steel fiber reinforced
concrete in bending. Int J Cem Compos Lightweight Concr 1983;5:225e33.
[6] Gopalaratnam VS, Shah SP. Properties of steel fiber reinforced concrete sub-
jected to impact loading. J Am Concr Inst 1986;83:117e26.
[7] Banthia N. Impact resistance of concrete [PhD thesis]. Vancouver, Canada:
Department of Civil Engineering, the University of British Columbia; 1987.
pp. 214e28.
[8] Banthia N, Mindess S, Trottier JF. Impact resistance of steel fiber reinforced
concrete. ACI Mater J 1996;93:472e9.
[9] Nataraja MC, Dhang N, Gupta AP. Statistical variations in impact resistance of
steel fiber-reinforced concrete subjected to drop weight test. Cem Concr Res
1999;29:989e95.
[10] RILEM TC 162-TDF. Test and design methods for steel fiber reinforced
concrete-bending test. Mater Struct/Materiaux Constr 2002;35:579e82.
[11] BS EN 14651. Test method for metallic fibre concrete e measuring the flexural
tensile strength (limit of proportionality (LOP), residual). UK: BSI; 2007.
pp. 1e20.
[12] Zhang XX, Ruiz G, Yu RC, Tarifa M. Fracture behaviour of high-strength con-
crete at a wide range of loading rates. Int J Impact Eng 2009;36:1204e9.
[13] Hughes G, Beeby AW. Investigation of the effect of impact loading on concrete
beams. Struct Eng 1982;V 60B:45e52.
[14] Suaris W, Shah SP. Properties of concrete subjected to impact. J Struct Eng
1983;109:1727e41.
[15] Akcay B, Tasdemir MA. Mechanical behaviour and fibre dispersion of hybrid
steel fibre reinforced self-compacting concrete. Constr Build Mater 2012;28:
287e93.
[16] RILEM, TC50-FMC. Determination of the fracture energy of mortar and con-
crete by means of three-point bend tests on notched beams. Mater Struct
1985;18:99e101.
[17] Banthia N, Mindess S, Bentur A, Pigeon M. Impact testing of concrete using a
drop-weight impact machine. Exp Mech 1989;29:63e9.
[18] Planas J, Guinea GV, Elices M. Stiffness associated with quasi-concentrated
loads. Mater Constr 1994;27:311e8.
[19] Dong JK, El-Tawil S, Namman AE. Loading rate effect on pullout behavior of
deformed steel fibers. ACI Mater J 2008;105:576e84.
[20] Ruiz G, Zhang XX, Yu RC, Porras R, Poveda E, del Viso JR. Effect of loading rate
on fracture energy of high-strength concrete. Strain 2011;47:518e24.
[21] Brara A, Klepaczko JR. Fracture energy of concrete at high loading rates in
tension. Int J Impact Eng 2007;34:424e35.
[22] Weerheijm J, Van Doormaal JCAM. Tensile failure of concrete at high loading
rates: new test data on strength and fracture energy from instrumented
spalling tests. Int J Impact Eng 2007;34:609e26.
[23] Rossi P, Toutlemonde F. Effect of loading rate on the tensile behaviour of
concrete: description of the physical mechanisms. Mater Struct/Materiaux Et
Constr 1996;29:116e8.
[24] Bischoff PH, Perry SH. Compressive behaviour of concrete at high strain rates.
Mater Constr 1991;24:425e50.
[25] Vegt I, Breugel VK, Weerheijm J. Failure mechanisms of concrete under impact
loading. In: Carpinteri A, Gambarova P, Ferro G, Plizzari G, editors. Fracture
mechanics of concrete and concrete structures. Italy, Catania: Taylor & Francis
Group; 2007. pp. 579e87.
[26] Banthia N, Trottier JF. Deformed steel fiber-cementitious matrix bond under
impact. Cem Concr Res 1991;21:158e68.
[27] Caverzan A, Cadoni E, Prisco M. Tensile behaviour of high performance fibre-
reinforced cementitious composites at high strain rates. Int J Impact Eng
2012;45:28e38.