Construction and Building Materials 153 (2017) 846–856

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: [Link]/locate/conbuildmat

Review

Review of strain rate effects for UHPC in tension

R.J. Thomas ⇑, Andrew D. Sorensen
Dept. of Civil and Environmental Engineering, Utah State University, Logan, UT, United States

h i g h l i g h t s

Dynamic increase factors for UHPC in tension are reviewed.

Existing DIF models for conventional concrete underestimate DIF for UHPC.

Improved DIF models for tensile properties of UHPC are recommended.

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

Article history: The superb mechanical performance of ultra-high performance concrete (UHPC) has led to its rapid adop-
Received 25 April 2017 tion within the industry. In addition to mechanical strength exceeding that of conventional concrete by
Received in revised form 21 July 2017 an order of magnitude, UHPC exhibits improved energy absorption and ductility, making it an ideal mate-
Accepted 22 July 2017
rial for resisting high-energy dynamic loads like impact and blast. A number of recent studies investigate
Available online 28 July 2017
the effect of strain rate on the mechanical response of UHPC in tension. These studies report dynamic
increase factors for first cracking strength, ultimate tensile strength, modulus of elasticity, energy absorp-
Keywords:
tion, and a variety of other mechanical properties. This paper presents a review of these studies and an
Ultra-high performance concrete
Strain rate
analysis of the data presented therein. Data from existing literature suggest that existing models under-
Dynamic increase factor estimate the strain rate sensitivity of the tensile properties of UHPC, especially at high strain rates.
Compression Improved models based on existing data are proposed.
Tension Published by Elsevier Ltd.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847
2. Ultra-high performance concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847
3. Strain rate sensitivity of concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847
4. Relevant studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
5. Ultimate tensile strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
5.1. Effect of fiber orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
6. First cracking strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 850
7. Flexural strength (modulus of rupture) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 851
8. Modulus of elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 852
9. Strain capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 853
10. Energy absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 853
11. Other tensile properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
11.1. Crack spacing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
11.2. Crack opening displacement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
11.3. Maximum fiber stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
12. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
13. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 855
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 855
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 855

⇑ Corresponding author.
E-mail addresses: thmsrj@[Link] (R.J. Thomas), [Link]@[Link] (A.D. Sorensen).

[Link]
0950-0618/Published by Elsevier Ltd.
 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856 847

Nomenclature

0
fc compressive strength q fiber aspect ratio (q ¼ Lf =Df )
0 0
f c;0 reference compressive strength (f c;0 ¼ 10 MPa) Lf fiber length
ft ultimate tensile strength Df fiber diameter
fc first cracking strength DIF dynamic increase factor
CODt crack opening displacement corresponding to f t DIFf 0c dynamic increase factor for compressive strength
MOR modulus of rupture DIFf t dynamic increase factor for ultimate tensile strength
et strain corresponding to f t e_ strain rate
rfiber maximum fiber stress e_ s quasi-static strain rate (Unless otherwise specified,
F fiber factor, F ¼ V f bq e_ s ¼ 106 s1 )
Vf fiber volume fraction Tp peak toughness
b fiber bond factor (0.5 for straight fibers, 1.0 for hooked g energy absorption prior to softening
or twisted fibers [1–4]) g soft energy absorption during softening

1. Introduction The review is organized in the following manner. First, a brief

introduction to UHPC describes the history, development, and
Ultra-high performance concrete (UHPC) is a newly emerged applications of the material. Next, the state of the art with respect
advanced cementitious composite that has rapidly gained favor to strain rate sensitivity of concrete is discussed. This is followed
due to its improved mechanical strength and energy absorption by an overview of salient literature wherein the materials, meth-
capacity compared to conventional concrete. An emerging trend ods, and strain rates employed by relevant studies are discussed.
in the literature is the use of UHPC to resist high strain rate, The next several sections present aggregated DIF reported for the
dynamic loads like projectile impact or explosive blast. Recent lit- various tensile properties of UHPC, including ultimate tensile
erature suggests that the performance of UHPC structures under strength, first cracking strength, modulus of elasticity, and others.
such dynamic loads is admirable [5–7]. In order to design struc- The final section provides a very brief overview of DIF for UHPC in
tures to resist high strain rate loads, it is necessary to understand compression and shear, but does not purport to provide an inclu-
the effect of strain rate on the mechanical properties of UHPC. To sive review of these topics.
this end, many recent studies focus on describing the strain rate
sensitivity of UHPC. This paper presents a review and analysis of 2. Ultra-high performance concrete
such literature.
Concrete and other cementitious composites mechanical prop- Ultra-high performance concrete (UHPC) is a recently emerged
erties are known to be rate-sensitive, where the strength and other advanced cementitious composite that exhibits superb mechanical
properties improve as the strain rate increases. It is conventional to strength, ductility, and durability. First introduced in the mid-
describe the degree of strength improvement under dynamic load- 1990s by de Larrard and Sedran [8] as well as Richard and Cheyrezy
ing using the dynamic increase factor (DIF). The DIF is a simple [9,10], UHPC has seen widespread implementation throughout the
multiplier that describes the strength of a material at a dynamic world. UHPC differs from conventional Portland cement concrete
strain rate relative to the strength at a quasi-static strain rate. in several ways. A low water-to-cementitious materials ratio
Many recent studies report DIF for UHPC; since many of the perfor- (w=cm) and the inclusion of silica fume, a fine pozzolan, promote
mance benefits associated with UHPC relate to its tensile perfor- high mechanical strength. Furthermore, UHPC contains no coarse
mance, the bulk of these studies focus on DIF for UHPC in aggregates. Instead, optimized proportions of Portland cement,
tension. Some recent literature describing dynamic strength silica fume, and fine sand provide dense granular packing;
increase for UHPC in tension suggests that DIF for UHPC may differ improving mechanical properties and durability. Finally, the
from those for conventional concrete, hence the need for this inclusion of fiber reinforcement results in post-cracking strength
review. retention and a level of ductility that is uncharacteristic of
The scope of this review is defined thusly: This paper reviews conventional concrete.
existing peer-reviewed literature that presents dynamic increase
factors for material properties of UHPC in tension at quasi-
static, intermediate, and high strain rates. In addition to UHPC, 3. Strain rate sensitivity of concrete
this includes studies that report DIF for reactive powder concrete
(RPC) and engineered cementitious composites (ECC) with quasi- Concrete structures are subject to a wide array of strain rates
static compressive strength of at least 150 MPa; which is gener- in service. These strain rates range from the order of 108 s1
ally accepted as the minimum strength for classification as UHPC. for creep, to the order 103 s1 for projectile impact or blast.
Because only the material properties are of interest, studies that Typical strain rate ranges for a variety of loading types are shown
discuss dynamic performance of UHPC structures or structural in Fig. 1.
elements are not included. Furthermore, since the intent of the Concrete and other cementitious composites mechanical prop-
review is to examine DIF as a function of strain rate, only studies erties are known to be rate-sensitive. Abrams first demonstrated
that directly report the strain rate or provide sufficient informa- the strain rate sensitivity of concrete in 1917 when he noted
tion to compute the strain rate (e.g., stress rate and modulus of increased compressive strength at higher loading rates [12]. No
elasticity) are included in this review. Similarly, only studies that universally accepted explanation for the strain rate sensitivity of
provide sufficient information to compute DIF are included; this concrete exists, but three hypotheses prevail:
means that either the DIF must be directly reported or the
mechanical properties must be reported at both quasi-static and 1. The Stefan effect, which attributes rate sensitivity to the pres-
dynamic rates. ence of viscous pore fluid [13–15];
848 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856

Fig. 1. Typical strain rate ranges for concrete structures [11].

8
1:016d
2. Rate sensitivity of crack propagation, which can alter the >
< e_e_s ; e_ < 30 s1
fracture mechanism, attenuate crack velocity, and cause DIFf t ¼ ð4Þ
>
13
preferential cracking through regions of higher strength : b e_ ; e_ P 30 s1
[16–18]; and e_ s

3. Inertial effects [11,19,20].

1
d¼ 0 ; log b ¼ 7:11d  2:33 ð5Þ
The strain rate sensitivity of conventional concrete depends 10 þ 6 ff0 c
c;0
heavily on concrete quality, where higher-strength concretes expe-
rience a lesser degree of strength improvement under dynamic Based on their analysis, Malvar and Ross [28] suggested an
loading [21]. Meanwhile, the strain rate sensitivity of fiber- alternative definition for DIFf t , which is given by Eqs. (6) and (7).
reinforced concrete relies heavily on the fiber volume fraction, 8
1:016d
geometry, and bond [22–26]. >
< e_e_s ; e_ < 1 s1
The dynamic increase factor (DIF) describes the degree of DIFf t ¼
 ð6Þ
> e_ 3
:
1
improvement in strength or any other mechanical property (e.g., b e_ s ; e_ P 1 s1
modulus of elasticity, energy absorption). DIF is the simple ratio
of some material property C at some dynamic strain rate e_ to that
1
at the quasi-static strain rate e_ s , as shown in Eq. (1). d¼ 0 ; log b ¼ 6d  2 ð7Þ
1 þ 8 ff0 c
Cðe_ Þ c;0
DIFðe_ Þ ¼ ð1Þ
Cðe_ s Þ Fig. 2 shows DIFf t for conventional concrete according to FIB
Bischoff and Perry [21] summarize DIF reported for the com- (Eqs. (4) and (5) [1] and Malvar and Ross (Eqs. (6) and (7)) [28].
pressive strength of conventional concrete at strain rates in the Dynamic increase factors are shown for quasi-static compressive
range 108 to 102 s1. Normalized at the quasi-static strain rate strengths of 30 MPa—which is typical of conventional concrete—
of 105 s1, dynamic increase factors range from 0.75 at creep and 150 MPa—which is typical of UHPC.
strain rates of the order 108 s1 to 2.25 at dynamic strain rates
of the order 102 s1. The International Federation for Structural
Concrete (Fédération international du béton—FIB) defines the
dynamic increase factor for compressive strength (DIFf 0c ) of con-
ventional concrete as given in Eq. (2), where e_ s ¼ 106 s1 [27].
0
The coefficients a and c are defined by Eq. (3), where f c is the
0
quasi-static compressive strength and f c;0 ¼ 10 MPa.
8
1:026a
>
< e_e_s ; e_ < 30 s1
DIFf c ¼
0

 ð2Þ
>
: c e_ 3 ;
1

e_ s e_ P 30 s1

1
a¼ 0 ; log c ¼ 6:156a  2 ð3Þ
5 þ 9 ff0 c
c;0

Citing the abundance of papers discussing the dynamic strength

increase for concrete in compression and the distinct lack of those
discussing the same in tension, Malvar and Ross [28] reviewed fif-
teen studies of the strain rate sensitivity of conventional concrete
in tension at strain rates in the range 106 to 102 s1. Normalized
at the quasi-static strain rate of 106 s1, dynamic increase factors
exceed 8 at dynamic strain rates of the order 101 s1. FIB defines
the dynamic increase factor for tensile strength (DIFf t ) of conven-
tional concrete according to Eq. (4), where e_ s ¼ 3  106 s1. The 0
Fig. 2. DIFf t for concrete with f c ¼ 30 and 150 MPa according to FIB [27] and Malvar
coefficients d and b are defined by Eq. (5). and Ross [28].
 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856 849

Existing models for DIF for concrete suggest reduced strain rate fiber factor (Fig. 3(d)). The aggregated data are plotted with
sensitivity in high strength concrete. Furthermore, prevailing theo- estimates for the dynamic increase factor for the ultimate tensile
ries explaining strain rate effects in concrete suggest that the pore strength (DIFf t ) according to FIB [27] and Malvar and Ross [28]
structure affects the strain rate sensitivity. Finally, fiber volume 0
with quasi-static compressive strength f c ¼ 150 MPa. The figure
fraction, geometry, and bond significantly affect the strain rate sen- does not present data sorted by quasi-static compressive strength
sitivity of concrete with fiber reinforcement. UHPC exhibits higher for two reasons: (1) The effect of quasi-static compressive strength
compressive strength and lower porosity than conventional con- is minimal in the range of compressive strengths typical of UHPC,
crete. Furthermore, UHPC includes fiber reinforcement, which is and (2) many studies do not report quasi-static compressive
known to affect strain rate sensitivity [29]. It is therefore of great strength.
interest to study the strain rate sensitivity of UHPC. Several recent Neither the FIB model [27] nor the model proposed by Malvar
papers have done so. A few studies have investigated the rate sen- and Ross [28] adequately describe the dynamic increase factor
sitivity of UHPC in compression, reporting less rate sensitivity than for the ultimate tensile strength of UHPC at high strain rates, as
conventional concrete [30–34]. One study investigated the rate shown in Fig. 3. Both models fit the data reasonably well at low
sensitivity of UHPC in shear, finding that the shear strength of and intermediate strain rates, but neither inflection point (30 s1
UHPC is insensitive to strain rate [35]. However, most existing for FIB [27] and 1 s1 for Malvar and Ross [28]) agrees well with
studies focus on the rate sensitivity of UHPC in tension. This paper the data. Park et al. [44] proposed the relationship given in Eq.
focuses on the results and implications of those studies. (8) for DIFf t , where d ¼ 0:01465 and b ¼ 0:002352. This relation-
ship is shown on Fig. 4. Eq. (8) fits the aggregated data slightly bet-
ter than Eqs. (4) or (6), but still leaves much room for
4. Relevant studies
improvement.
Thirteen studies discuss the effects of strain rate on the tensile
(
ðe_ =e_ s Þ ; e_ 6 25 s1
d
properties of UHPC using a variety of techniques [30,31,36–48]. DIFf t ¼ ð8Þ
bðe_ =e_ s Þ ; e_ > 25 s1
0:3735
Most of these studies tested UHPC specimens in direct tension
using servo-hydraulic testing machines for quasi-static and inter- An improved model was developed by nonlinear piecewise
mediate strain rates and strain energy frames or Hopkinson bars orthogonal regression modeling. A simple model of the same form
for high strain rates. A few reported on indirect tensile tests using given by Eqs. (2)–(8) was desired. An inflection point at an integer
Hokinson bars in spallation [38] or split tension [30,31] configura- power of ten was preferred. Parameters were rounded to nearest
tions, and two reported on dynamic bending experiments using a simple fractions to give simple equations. The improved model
block-bar [37] or drop-weight [35] device. Strain rates ranged from developed for DIF f t is presented in Eq. (9). The parameter
quasi-static to the order 102 s1. All studies included steel fiber  0 0 
d ¼ 1= 1 þ 8f c =f c0 given by Malvar and Ross [28] was a good fit
reinforcement, including straight, twisted, and hooked fibers. A
and was therefore adopted in Eq. (9), while the inflection point
few studies included hybrid combinations of fiber types. Volume
shifted to 10 s1. log b ¼ 7d  5:25 in order to satisfy continuity
fractions were typically in the range of 0.5–5%, but one study
at the inflection. The power on e_ =e_ s above the inflection was very
included a volume fraction of 11% [37]. Several studies also
close to 3/4. The proposed model, shown in Fig. 4, correlates well
reported results for UHPC matrix without fiber reinforcement.
Table 1 includes a brief summary of each of these studies in with the data. The power of 3=4 on e_ =e_ s for e_ > 10 s1 is substan-
chronological order of publication. tially higher than those in Eqs. (4), (6), and (8). This suggests that
Several caveats with respect to the interpretation of these data UHPC is much more strain rate sensitive than conventional con-
merit mention. First, while FIB formulations for DIF consider a crete at high strain rates.
quasi-static strain rate of 106 s1, several of the studies listed in (
ðe_ =e_ s Þ ; e_ 6 10 s1
d
Table 1 consider quasi-static strain rates of the order 104 s1. This DIFf t ¼ ð9Þ
bðe_ =e_ s Þ ; e_ > 10 s1
3=4
is not ideal for comparative and modeling purposes, but the
dynamic strength increase is minimal in this range of strain rates. In the absence of complete information regarding the quasi-
Additionally, a few of these studies report stress rates instead of static compressive strength of the mixtures used in several of the
strain rates. In these cases, very close approximations for strain studies listed in Table 1, the derivation of Eq. (8) assumes a com-
rate are made assuming a modulus of elasticity near 50 GPa. This pressive strength of 150 MPa. The definition for d in Eq. (8) is then
is consistent with values reported throughout the literature, and simplified to d ¼ 1=121.
it will be shown later in this paper that this value is mostly insen- All of the studies listed in Table 1 showed fiber parameters (i.e.,
sitive to strain rate. Finally, specimen sizes and geometries differ volume fraction, geometry, and bond) to affect the mechanical
between studies. Geometric differences, as well as differences properties under both quasi-static and dynamic loading. As such,
between test methods and analyses, are expected to affect test it was initially hypothesized that an improved estimate for DIFf t
results. The authors are unaware of any evidence that these differ-
for UHPC would consider the fiber type, volume fraction, or fiber
ences will affect rate sensitivity, although such effects would not
factor. Despite this, no trend is apparent from the data presented
be unexpected. Nevertheless, the analysis in this paper neglects
in Fig. 3(b)–(d). Furthermore, modeling efforts showed these
these potential effects. Given that specimen size variations are
parameters to be insignificant.
small relative to the maximum aggregate size, this effect is also
expected to be minimal.
5.1. Effect of fiber orientation

5. Ultimate tensile strength The distribution of fibers in fiber-reinforced concrete is pre-

dominately random. This is beneficial in many cases, such as rein-
Eleven of the studies summarized in Table 1 report ultimate forcement of randomly oriented shrinkage cracks. If the viscosity of
tensile strengths for UHPC at strain rates in the range the fresh concrete is sufficiently low, the fibers may preferentially
106–102 s1. Fig. 3 presents the aggregated data sorted by author orient in the direction of casting. This may result in optimal
(Fig. 3(a)), fiber type (Fig. 3(b)), volume fraction (Fig. 3(c)), and mechanical performance in cases where the plane of maximum
850 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856

Table 1
Summary of studies evaluating strain rate effects on tensile properties of UHPC.

Study [Ref] Test method Strain rates Materials Properties

(s1) reported
Fujikake et al. [36] Direct tension 106–101 RPC with 2% straight steel fibers rc
Servo-hydraulic rt
CODt
Parant et al. [37] Bending 106–101 UHPC matrix and UHPC with 11% straight steel fibers rf
Block-bar device
Nöldgen et al. [38] Spallation 102 UHPC with 0–2% straight steel fibers rt
Hopkinson bar
Millard et al. [35] Bending 106 UHPC with 0–6% straight steel fibers rf
Drop weight 101–100
Tran et al. [39] Direct tension 104 UHPC with 1.5% straight and twisted steel fibers rt
Strain energy 100–101 et
frame g
Nc
Pyo et al. [40] Direct tension 104–101 UHPC with 1–3% straight and twisted steel fibers rc
Servo-hydraulic rt
et
g
Nc
E
rfiber
Ranade et al. [41] Direct tension 104–101 ‘‘High-strength-high-ductility concrete” with 1.7% polyethylene fibers rc
Servo-hydraulic rt
et
Xu & Wille [42] Direct tension 104–101 UHPC with 1.5–3% straight, hooked, and twisted steel fibers g
Servo-hydraulic scr
Cadoni & Forni [43] Direct tension 101 Self-consolidating UHPC with 0–5% straight steel fibers (specified fiber rt
Hopkinson bar orientation)
Park et al. [44] Direct tension 104 UHPC with 1.5–2% straight steel fibers rt
Strain energy 101–102 et
frame Eabs
Nc
Pyo et al. [45] Direct tension 101–102 UHPC with 1–3% straight and twisted steel fibers rc
Strain energy rt
frame et
g
Su et al. [30] Splitting tension 100 UHPC with 2.5% straight steel fibers and 1–5% nanoparticles rt
Hopkinson bar
0 1
Su et al. [31] Splitting tension 10 –10 UHPC with 0.5–2.5% straight and twisted steel fibers rt
Hopkinson bar
Wille et al. [46] Direct tension 104–101 UHPC with 1.5–3% straight, twisted, and hooked steel fibers rc
Servo-hydraulic rt
E
Ehard
esoft
g
rfiber
sc
Tran & Kim [47] Direct tension 104 UHPC with 1.5% blended straight, twisted, and hooked steel fibers rt
Strain energy 101 et
frame g
Park et al. [48] Direct tension 104–102 UHPC with 2% blended short and long smooth steel fibers rt
Strain energy et
frame g
Nc
dc

stress is known with certainty. Cadoni and Forni [43] studied the 6. First cracking strength
effect of fiber orientation on the dynamic tensile properties of
UHPC by testing UHPC specimens oriented parallel and orthogonal First cracking strength f cr corresponds to the transition from
to the direction of casting. Fig. 5 shows reported values for DIFf t for elastic to strain-hardening behavior. Five of the studies listed in
fiber volume fractions in the range 0–5%. Results indicated signifi- Table 1 reported values for the first cracking strength f cr of UHPC
cant improvements in tensile strength, strain capacity, and energy at various strain rates [36,40,41,45,46]. Fig. 6 shows dynamic
absorption for specimens tested parallel to the direction of pre- increase factors for first cracking tensile strength (DIFf cr ). The figure
dominate fiber orientation. As shown in Fig. 5, dynamic increase also shows models for DIFf t according to FIB (Eq. (4) [27] and Mal-
factors were, in general, slightly higher for specimens tested paral- 0
var and Ross (Eq. (6) [28] for the case f c ¼ 150 MPa. The first crack-
lel to the direction of casting, but the improvement is small relative ing strength roughly corresponds to the tensile strength of the
to the variability in the reported DIF. As discussed above, no trend matrix, although some authors have demonstrated that first crack-
was observed between DIFf t and fiber volume fraction for either ing strength is dependent on fiber geometry or volume fraction
orientation. [40,49]. It is therefore reasonable to assume that these models
 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856 851

(a) by author (b) by fiber type

(c) by fiber volume fraction (d) by fiber factor

Fig. 3. DIFf t for UHPC from existing literature.

could give a good estimate of DIFf cr for UHPC. Unfortunately, the ported geometries. Millard et al. [35] reported MOR for UHPC with
limited quantity of data at high strain rates means that it is difficult 0–6% straight steel fibers, including some specimens with hybrid
to draw sound conclusions with respect to the goodness of fit of combinations of fiber lengths. Fig. 7 shows dynamic increase fac-
these existing models or to develop any improved models. In fact, tors for modulus of rupture (DIFMOR ) reported by these studies.
since there is a gap in available data between strain rates of the Since MOR is an estimate of tensile strength, the estimates for
order 100 s1 and 102 s1, it is impossible to provide an accurate DIFf t according to FIB [27] and Malvar and Ross [28] with quasi-
0
estimate of location of the inflection point. Nevertheless, both Eq. static compressive strength f c ¼ 150 MPa are shown in Fig. 7, as
(4) [27] and Eq. (6) [28] show reasonably good agreement with is the proposed model for DIFf t from Eq. (9). The model for DIFf t
existing data at quasi-static and intermediate strain rates. Eq. (6) proposed by Malvar and Ross [28] and given by Eq. (6) appears
[28] tends to overestimate DIFf cr at high strain rates, while Eq. to be the best fit of these three equations; the inflection point at
(4) [27] provides a conservative estimate. 1 s1 fits the data very well, but the equation tends to underesti-
mate DIFMOR at strain rates above this inflection point. An
7. Flexural strength (modulus of rupture) improved model was again determined by modified nonlinear
piecewise orthogonal regression modeling, where the parameter
Flexural strength or modulus of rupture (MOR) is often taken as d was again very close to that in Eq. (9) (i.e.,
0
an estimate of the direct tensile strength. Two studies discuss d ¼ 1=ð1 þ 8f c Þ ¼ 1=121, is given by Eq. (10). Similarly, the power
strain rate effects on the MOR of UHPC [35,37]. Parant et al. [37] on e_ =e_ s above the inflection point as determined by orthogonal
reported MOR for UHPC with 11% volume fraction of multi-scale regression was very close to 3/4; and log b ¼ 6d  4:5 in order to
fiber reinforcement, which included steel fibers of three unre- satisfy continuity at the inflection. Eq. (10) is also shown on Fig. 7.
852 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856

Fig. 4. Proposed model (Eq. (8) for DIFf t for UHPC. Fig. 6. DIFf cr for UHPC from existing literature.

Fig. 5. Effect of fiber orientation on DIFf t [43]. Fig. 7. DIFMOR for UHPC from existing literature [35,37].
(
ðe_ =e_ s Þ e_ 6 1 s1
1=121
; point—but the strain rate sensitivity becomes important at an
DIFMOR ¼ ð10Þ order of magnitude lower strain rate. This is evidenced by the
bðe_ =e_ s Þ ; e_ > 1 s1
3=4
inflection in DIFMOR occurring at an order of magnitude lower strain
Fig. 8 compares dynamic increase factors for modulus of rup- rate. As a result, flexural testing may be an inadvisable means of
ture and for ultimate tensile strength. Although MOR is frequently estimating the dynamic tensile strength of UHPC. Further investi-
used to estimate tensile strength in cementitious composites— gation may prove or disprove this recommendation.
which is mainly due to logistic difficulties with direct tensile
tests—the dynamic increase factors reported for MOR are quite dif- 8. Modulus of elasticity
ferent than those reported for ultimate tensile strength based on
direct tensile tests. The available evidence clearly suggests that Only two of the studies listed in Table 1 reported modulus of
the strain rate sensitivity of flexural strength is different from that elasticity for UHPC [40,46]. Both were based on direct tension tests
of tensile strength. Flexural strength is not more strain-rate sensi- of UHPC with up to 3% steel fiber reinforcement at strain rates
tive than direct tensile strength—in fact both dynamic increase fac- between 104 and 101 s1. Moduli of elasticity from both studies
tors increase according to the same power law above the inflection were typically in the range of 50–60 GPa, with some very high val-
 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856 853

9. Strain capacity

Seven of the studies listed in Table 1 discussed the effect of

strain rate on strain capacity as measured by the strain et corre-
sponding to the ultimate tensile strength f t [39–41,44,45,47,48].
All of these studies reported results based on direct tensile tests
using hydraulic testing machines or strain energy frames. Fig. 10
presents the aggregated results from these studies. Results from
the aforementioned studies were in good agreement with one
another, with the notable exception of the data presented by
Ranade et al. [41], which suggested that strain capacity was
reduced at increased strain rate. Park et al. [44] proposed the rela-
tionship given by Eq. (11) for DIFet , which is shown in Fig. 10. Eq.
(11) fits the data from Park et al. [44] very well, but underestimates
the strain rate sensitivity reported by the remaining studies. Solved
in the same manner as before, an improved model for DIFet is given
by Eq. (12), where b ¼ 1:2847  1011 , the inflection point reduced
to e_ ¼ 20 s1 , and the power on e_ =e_ s increased to 1.5. The proposed
model is also shown on Fig. 10. The power of 1.5 in the high strain
rate region indicates that the strain capacity is much more rate
sensitive than the other properties at strain rates above about
20 s1.
(
ðe_ =e_ s Þ e_ 6 25 s1
0:01465
;
Fig. 8. Comparison of DIFf t and DIFMOR for UHPC. DIFet ¼ ð11Þ
0:003998ðe_ =e_ s Þ ; e_ > 25 s1
0:3286

ues near 70 GPa. Dynamic increase factors for modulus of elasticity (

ðe_ =e_ s Þ e_ 6 20 s1
1=121
(DIFE ) reported by Pyo et al. [40] and Wille et al. [46] are shown in ;
DIFet ¼ ð12Þ
bðe_ =e_ s Þ ; e_ > 20 s1
1:5
Fig. 9. With the exception of two errant values, the dynamic
increase factors were very near one throughout the range of strain
rates represented by the data. The absence of data above strain
rates of the order 101 s1, which likely results from difficulties 10. Energy absorption
accurately measuring strain at high rates of loading, prevents
drawing any conclusions with respect to the strain rate sensitivity Eight studies reported fracture energy or energy absorption
of the modulus of elasticity of UHPC at high strain rates. However, during tensile loading for UHPC [39,40,42,44–48]. All of the afore-
the data do suggest that the modulus of elasticity of UHPC is not mentioned studies evaluated energy absorption under direct ten-
sensitive to strain rate at quasi-static and intermediate strain rates. sile testing of UH DIFg PC using servo-hydraulic testing machines
or strain energy frames. Fig. 11 shows dynamic increase factors
for peak toughness (DIFg ) reported within the literature. Park
et al. [44] proposed Eq. (13) for DIFg , which is also shown on

Fig. 9. DIFE for UHPC from existing literature [40,46]. Fig. 10. DIFet for UHPC from existing literature [39–41,44,45,47,48].
854 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856

when hooked fibers were used [42]. Pyo et al. [40], Tran et al.
[30], and Park et al. [44,48] reported the average number of cracks
within the gauge length of UHPC direct tension specimens; the
number of cracks is approximately the inverse of average crack
spacing. The former two studies reported that the average number
of cracks was, on average, unchanged over the range of strain rates
104–101 s1. Park et al. [44,48] reported nearly twice the number
of cracks as the strain rate increased from 104 to 102 s1. This
result directly contrasts that of Wille et al. [46]. The lack of consen-
sus suggests that more research is necessary in this area.

11.2. Crack opening displacement

Fujikake et al. [36] reported critical crack opening displace-

ments (CODu ) for notched UHPC prisms tested in direct tension
at strain rates in the range 106–101 s1. Typical crack opening
displacements were near 100 mm with 2% volume fraction of
13  0.2 mm steel fibers at the quasi-static strain rate of
106 s1. Average DIF were 1.12, 1.22, and 1.84, respectively, at
strain rates of 2  103, 2  102, and 5  101 s1. Park et al.
[48] also reported crack opening displacements in UHPC with 1%
of both short straight and long hooked steel fibers. Crack opening
displacements at quasi-static rate of the order 104 s1 were about
Fig. 11. DIFg for UHPC from existing literature [39,40,42,44–48]. half that reported by Fujikake et al. [36]. Average DIF were 1.0, 0.9,
1.4, and 1.8, respectively, at strain rates of the order 103, 102, 101
and 102 s1. DIF reported by Park et al. [48] were generally lower
Fig. 11. Eq. (13) severely underestimates the dynamic increase fac- than those reported at similar strain rates by Fujikake et al. [36].
tors reported by other authors. Determined in the same manner as
before, an improved model for is given by Eq. (14), where
11.3. Maximum fiber stress
¼ 8:2887  1014 , the inflection point reduced to e_ ¼ 20 s1 , and
the power on e_ =e_ s increased to 1.8. The improved model is also
Pyo et al. [45] and Wille et al. [46] reported the maximum fiber
shown on Fig. 11. As before, the increased power on e_ =e_ s indicates
stress experienced during direct tensile loading. Both studies
the UHPC is more strain rate sensitive than conventional concrete
demonstrated minor increases in maximum fiber stress as the
at high strain rates.
( strain rate increased through the range 104–101 s1, with typical
ðe_ =e_ s Þ e_ 6 25 s1
0:01634
; DIF near 1.2 at the highest strain rate.
DIFg ¼ ð13Þ
0:00105ðe_ =e_ s Þ ; e_ > 25 s1
0:4192

( 12. Discussion
ðe_ =e_ s Þ e_ 6 25 s1
1=121
;
DIFg ¼ ð14Þ
bðe_ =e_ s Þ ; e_ > 25 s1 The proposed empirical relationships for DIF for the tensile
1:8

properties of UHPC are based on relevant data gathered from pre-

vious publications. In general, these relationships suggest that the
11. Other tensile properties rate UHPC is more rate-sensitive than conventional concrete.
Before suggesting possible explanations for this phenomenon, it
The previous sections review and analyze dynamic increase fac- is necessary to recall the prevailing theories with respect to the
tors reported for ultimate tensile strength, first cracking tensile mechanism behind rate sensitivity in cementitious composites,
strength, modulus of elasticity, strain capacity, and energy absorp- viz. the Stefan effect [13–15], rate sensitivity of crack propagation
tion. Several papers report dynamic increase factors for lesser- [16–18], and inertial effects [11,19,20], as discussed in Section 3 of
discussed properties, including crack spacing, critical crack open- this manuscript. Although the mechanistic differences between the
ing displacement, and maximum fiber stress. We summarize these rate sensitivity of conventional concrete and UHPC have not been
results in this section but do not attempt to model dynamic discussed in any great detail in the literature, some educated spec-
increase factors for these properties because of the scarcity of ulation is possible. First, it is known that fiber reinforcement
available data. increases rate sensitivity even in non-UHPC mixtures [29]. Fiber
reinforcement arrests crack growth and contributes post-cracking
11.1. Crack spacing strength retention, and will therefore alter the rate sensitivity of
the composite. Furthermore, the Stefan effect is dependent on
Wille et al. [46] reported average crack spacing sc based on the pore structure within the matrix. In short, the resistance due
direct tension tests of UHPC at strain rates in the range 104– to the Stefan effect increases with rate and increases with pore
101 s1. Average crack spacing increased steadily with strain rate, radius [13–15]. The low water-to-cementitious materials ratio
with typical dynamic increase factors near 1.2 and the highest and the inclusion of pozzolanic silica fume in UHPC promotes the
strain rate. Xu and Wille [42] also reported average crack spacing formation of a very fine pore structure, which will necessarily
for UHPC in the same strain rate range, but results were mixed. increase the rate sensitivity. It is unlikely that inertial effects,
Average crack spacing tended to increase with strain rate when which are more a function of the applied load and the specimen
twisted fibers were used, but decreased with strain rate than of the material, differ between conventional concrete and
when straight fibers were used [42]. Mixed effects were noted UHPC.
 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856 855

13. Conclusions [9] P. Richard, M. Cheyrezy, Reactive powder concretes with high ductility and
200–800 MPa compressive strength, ACI Special Publication 144 (1994) 507–
518.
The results of thirteen previously published studies of the strain [10] P. Richard, M. Cheyrezy, Composition of reactive powder concretes, Cem.
rate sensitivity of UHPC in tension with the intent of describing Concr. Res. 25 (7) (1995) 1501–1511.
[11] S. Hentz, F.V. Donzé, L. Daudeville, Discrete element modeling of concrete
dynamic increase factors for the various tensile properties of UHPC
submitted to dynamic loading at high strain rates, Comput. Struct., 82 (29)
are reviewed and analyzed. A comparison of the reported DIF for (2004) 2509–2524.
the ultimate tensile strength, first cracking tensile strength, modu- [12] D.A. Abrams, Effect of rate of application of load on the compressive strength
of concrete, ASTM J. 17 (2) (1917) 70–78.
lus of rupture, strain capacity, energy absorption, and modulus of
[13] P. Rossi, A physical phenomenon which can explain the mechanical behavior
elasticity of UHPC with existing estimates for DIF for conventional of concrete under high strain rates, Mater. Struct. 24 (1991) 422–424.
concrete and with those previously recommended for UHPC is [14] P. Rossi, J.G.M. van Mier, C. Boulay, F. le Maou, The dynamic behavior of
made. Where necessary, proposed improved relationships for the concrete: influence of free water, Mater. Struct. 25 (1992) 509–514.
[15] P. Rossi, F. Toutlemonde, Effect of loading rate on the tensile behavior of
various DIFs for UHPC are given. concrete: description of the physical mechanisms, Mater. Struct. 29 (1996)
The following conclusions are based on the results of this study: 116–118.
[16] E. Cadoni, K. Labibes, C. Albertini, M. Berra, M. Giangrasso, Strain-rate effect on
the tensile behaviour of concrete at different relative humidity levels, Mater.

The strain rate sensitivity of the tensile properties of UHPC is Struct. 34 (1) (2001) 21–26.
similar to that of conventional concrete at quasi-static and [17] W. Suaris, S.P. Shah, Rate-sensitive damage theory for brittle solids, J. Eng.
intermediate strain rates. Mech. 110 (6) (1984) 985–997.
[18] R. John, S.P. Shah, Y.S. Jeng, A fracture mechanics model to predict the rate

UHPC is much more strain rate-sensitive than conventional con- sensitivity of mode I fracture of concrete, Cem. Concr. Res. 17 (2) (1987) 249–
cretes at high strain rates. 262.

The strain rate at which the rate sensitivity becomes important [19] H.W. Reinhardt, J. Weerheijm, Tensile fracture of concrete at high loading rates
taking into account of inertia and crack velocity effects, Int. J. Fract. 51 (1991)
(i.e., the inflection point in the piecewise definition for DIF) is
31–42.
typically lower for UHPC than for conventional concrete. [20] G. Cusatis, Strain-rate effects on concrete behavior, Int. J. Impact Eng. 38

The strain rate sensitivity of UHPC does not depend on fiber vol- (2011) 162–170.
[21] P.H. Bischoff, S.H. Perry, Compressive behaviour of concrete at high strain
ume fraction, fiber geometry, fiber factor, or compressive
rates, Mater. Struct. 24 (6) (1991) 425–450.
strength. [22] W. Suaris, S.P. Shah, Strain-rate effects in fibre-reinforced concrete subjected

An improved model for the DIF for ultimate tensile strength of to impact and impulsive loading, Composites 13 (2) (1982) 153–159.
UHPC is given by Eq. (9). [23] V.S. Gopalaratnam, S.P. Shah, Properties of steel fiber reinforced concrete
subjected to impact loading, J. Am. Concr. Inst. 83 (1) (1986) 117–126.

DIF for first cracking tensile strength are well described by [24] Z.L. Wang, Y.S. Liu, R.F. Shen, Stress-strain relationship of steel fiber-reinforced
existing relationships for DIF for tensile strength of conven- concrete under dynamic compression, Constr. Build. Mater. 22 (5) (2008) 811–
tional concrete. 819.
[25] V. Bindiganavile, N. Banthia, Impact response of the fiber-matrix bond in

An improved model for the DIF for modulus of rupture of UHPC concrete, Can. J. Civ. Eng. 32 (5) (2005) 924–933.
is given by Eq. (10). [26] J.K. Dong, S. El-Tawil, A.E. Naaman, Rate-dependent tensile behavior of high

An improved model for the DIF for strain capacity of UHPC is performance fiber reinforced cementitious composites, Mater. Struct. 42 (3)
(2009) 399–414.
given by Eq. (12). [27] Comité Euro-International du Béton, Concrete structures under impact and

An improved model for the DIF for energy absorption of UHPC is impulsive loading, CEB Bulletin 187 (1988).
given by Eq. (14). [28] L.J. Malver, C.A. Ross, Review of strain rate effects for concrete in tension, ACI
Mater. J. 95 (6) (1998) 735–739.

The modulus of elasticity of UHPC in tension is not strain rate
[29] E. Cadoni, A. Meda, G.A. Plizzari, Tensile behaviour of FRC under high strain-
sensitive up to strain rates of the order 101 s1. rate, Mater. Struct. 42 (2009) 1283–1294.
[30] Y. Su, J. Li, C. Wu, P. Wu, Z.X. Li, Influences of nano-particles on dynamic
strength of ultra-high performance concrete, Compos. B 91 (2016) 708–718.
[31] Y. Su, J. Li, C. Wu, P. Wu, Z.X. Li, Effects of steel fibres on dynamic strength of
Acknowledgement UHPC, Constr. Build. Mater. 114 (2016) 708–718.
[32] J. Clark, Preliminary Investigation of Ultra-High Performance Concrete
Behavior at High Strain Rates Using the Split-Hopkinson Pressure Bar,
The authors gratefully acknowledge the financial support of the Thesis, Michigan Technological University, 2013.
Intelligence Community Postdoctoral Fellowship (IC Postdocs) pro- [33] D. Van Slembrouck, Compression Behavior at High Strain Rates for an
Ultra High Performance Concrete, Thesis, Michigan Technological University,
gram. The opinions, findings, and conclusions presented here are
2015.
those of the authors and do not reflect the positions of the IC Post- [34] A.B. Groenveld, Effect of Fiber Orientation on Dynamic Compressive Properties
docs program. of an Ultra-High Performance Concrete, Dissertation, Michigan Technological
University (2016).
[35] S.G. Millard, T.C.K. Molyneaux, S.J. Barnett, X. Gao, Dynamic enhancement of
References blast-resistant ultra high performance fibre-reinforced concrete under flexural
and shear loading, Int. J. Impact Eng. 37 (2010) 405–412.
[36] K. Fujikake, T. Senga, N. Ueda, T. Ohno, M. Katagiri, Effect of strain rate on
[1] B.P. Hughes, N.I. Fattuhi, The workability of steel-fibre-reinforced concrete,
tensile behavior of reactive powder concrete, J. Adv. Concr. Technol. 4 (1)
Mag. Concr. Res. 28 (96) (1976) 157–161.
(2006) 79–84.
[2] F.W. Faisal, S.A. Ashour, Mechanical properties of high-strength fiber
[37] E. Paran, P. Rossi, E. Jacquelin, C. Boulay, Strain rate effect on bending behavior
reinforced concrete, ACI Mater. J. 89 (5) (1992) 449–455.
of new ultra-high performance cement-based composite, ACI Mater. J. 104 (5)
[3] M. Khuntia, B. Stojadinovic, S.C. Goel, Shear strength of normal and high-
(2007) 458–463.
strength fiber reinforced concrete beams without stirrups, ACI Struct. J. 96 (2)
[38] M. Nöldgen, W. Riedel, K. Thoma, E. Fehling, Properties of ultra high
(1999) 282–289.
performance concrete (UHPC) in tension at high strain rates, in: VIII
[4] Y.K. Kwak, M.O. Eberhard, W.S. Kim, J. Kim, Shear strength of steel fiber-
International Conference on Fracture Mechanics of Concrete and Concrete
reinforced concrete beams without stirrups, ACI Struct. J. 99 (4) (2002) 530–
Structures (FraMCoS-8) (2013).
538.
[39] N.T. Tran, T.K. Tran, D.J. Kim, High rate response of ultra-high-performance
[5] S. Astarlioglu, T. Krauthammer, Response of normal-strength and ultra-high-
fiber-reinforced concretes under direct tension, Cem. Concr. Res. 69 (2015)
performance fiber-reinforced concrete columns to idealized blast loads, Eng.
72–87.
Struct. 61 (2014) 1–12.
[40] S. Pyo, K. Wille, S. El-Tawil, A.E. Namman, Strain rate dependent properties of
[6] P. Máca, S. Radoslav, P. Konvalinka, Mix design of UHPFRC and its response to
ultra high performance fiber reinforced concrete (UHP-FRC) under tension,
projectile impact, Int. J. Impact Eng. 63 (2014) 158–163.
Cem. Concr. Compos. 56 (2015) 15–24.
[7] J. Li, C. Wu, H. Hao, An experimental and numerical study of reinforced ultra-
[41] R. Ranade, V.C. Li, W.F. Heard, Tensile rate effects in high strength-high
high performance concrete slabs under blast loads, Mater. Des. 82 (2015) 64–
ductility concrete, Cem. Concr. Res. 68 (2015) 94–104.
76.
[42] M. Xu, K. Wille, Fracture energy of UHP-FRC under direct tensile loading
[8] F. de Larrard, T. Sedran, Optimization of ultra-high-performance concrete by
applied at low strain rates, Compos. B 80 (2015) 16–125.
the use of a packing model, Cem. Concr. Res., 24 (6) (2994) 997–1009.
856 R.J. Thomas, A.D. Sorensen / Construction and Building Materials 153 (2017) 846–856

[43] E. Cadoni, D. Forni, Experimental analysis of the UHPFRCs behavior under [47] N.T. Tran, D.J. Kim, Synergistic response of blending fibers in ultra-high-
tension at high stress rates, Eur. Phys. J. Special Topics 225 (2016) 253–264. performance concrete under high rate tensile loads, Cem. Concr. Compos. 78
[44] S.H. Park, D.J. Kim, S.W. Kim, Investigating the impact resistance of ultra-high- (2017) 132–145.
performance fiber-reinforced concrete using an improved strain energy [48] J.K. Park, S.W. Kim, D.J. Kim, Matrix-strength-dependent strain-rate sensitivity
impact test machine, Constr. Build. Mater. 125 (2016) 145–159. of strain-hardening fiber-reinforced cementitious composites under tensile
[45] S. Pyo, S. El-Tawil, A.E. Naaman, Direct tensile behavior of ultra high impact, Compos. Struct. 162 (2017) 313–324.
performance fiber reinforced concrete (UHP-FRC) at high strain rates, Cem. [49] S.T. Kang, J.I. Choi, K.T. Koh, K.S. Lee, B.Y. Lee, Hybrid effects of steel fiber and
Concr. Res. 88 (2016) 144–156. microfiber on the tensile behavior of ultra-high performance concrete,
[46] K. Wille, M. Xu, S. El-Tawil, A.E. Naaman, Dynamic increase factors of strain Compos. Struct. 145 (2016) 37–42.
hardening UHP-FRC under direct tensile loading at low strain rates, Mater.
Struct. 49 (2016) 1351–1365.