You are on page 1of 10

Cement & Concrete Composites 56 (2015) 1524

Contents lists available at ScienceDirect

Cement & Concrete Composites


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

Strain rate dependent properties of ultra high performance ber


reinforced concrete (UHP-FRC) under tension
Sukhoon Pyo a, Kay Wille b, Sherif El-Tawil c,, Antoine E. Naaman c
a
Korea Railroad Research Institute, 176 Railroad Museum Road, Uiwang-si, Gyeonggi-do 437-757, South Korea
b
Department of Civil & Environmental Engineering, University of Connecticut, 261 Glenbrook Road Unit 2037, Storrs, CT 06269-2037, USA
c
Department of Civil & Environmental Engineering, University of Michigan, 2350 Hayward, G.G. Brown, Ann Arbor, MI 48109-2125, USA

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

Article history: The results of an experimental investigation of UHP-FRC tensile response under a range of low strain rates
Received 9 November 2013 are presented. The strain rate dependent tests are conducted on dogbone specimens using a hydraulic
Received in revised form 28 August 2014 servo-controlled testing machine. The experimental variables are strain rate, which ranges from
Accepted 22 October 2014
0.0001 1/s to 0.1 1/s, ber type, and ber volume fraction. Five different types of bers are considered
Available online 29 October 2014
including straight and twisted bers with different geometric properties. The rate sensitivity of the com-
posite material in tension is evaluated in terms of its rst cracking strength, post-cracking strength,
Keywords:
energy absorption capacity, strain capacity, elastic modulus, ber tensile stress and number of cracks.
Mechanical properties
Strain rate effect
The test results show pronounced rate effects on post-cracking strength and energy absorption capacity.
Tensile properties Further, post cracking strength varies linearly with the ber reinforcing index and energy absorption
Fiber reinforcement capacity varies linearly with the product of the ber length and the reinforcing index, as predicted from
Ultra high-performance concrete the theory for ber reinforced concrete.
Cracking 2014 Elsevier Ltd. All rights reserved.

1. Introduction A relatively recent and most promising ber reinforced cemen-


titious material for structural applications is UHP-FRC. Although
The development of high performance concrete materials hav- different researchers have dened UHPC and UHP-FRC using sev-
ing signicant strength and enhanced ductility, particularly in ten- eral criteria [1315], American Concrete Institute (ACI) Committee
sion is a subject of continuous research. Examples of concretes 239 suggests the following denition: Concrete, ultra-high
with enhanced tensile properties include Fiber Reinforced Con- performance concrete that has a minimum specied compressive
cretes (FRC) [1], Slurry Inltrated Fiber Concrete (SIFCON) [2], strength of 150 MPa (22,000 psi) with specied durability, tensile
Multi-Scale Cement Composite (MSCC) [3], Hybrid Fiber Concretes ductility and toughness requirements; bers are generally included
(HFC) [4], High Performance Fiber Reinforced Cement Composites to achieve specied requirements. There appears to be consensus
(HPFRCC) [5], Engineered Cementitious Composite (ECC) [6] and in the research literature that well designed UHP-FRC can be highly
Ultra High Performance Fiber Reinforced Concrete (UHP-FRC) [7]. durable against chemical attack, freezethaw cycles, abrasion and
As is clear from their names, all of the concretes listed above are chloride penetration [14,16,17], and therefore there is strong inter-
reinforced with bers (polymeric or/and steel). est in exploring its material properties.
Fiber reinforced concretes resist tensile stress through compos- The objective of this study is to obtain more detailed knowledge
ite action between the cementitious matrix and embedded bers. of the tensile behavior of UHP-FRC under loading with various
The transmission of forces between these two components occurs strain rates. The experimental variables are ber type, ber volume
through interfacial bond. After cracking, bers bridge the cracks, fraction and strain rate. Five different ber types are considered,
providing resistance to crack opening, enhancing structural behav- including both straight and twisted bers with a variety of geomet-
ior and durability. Many researchers have focused on characteriza- ric properties. The strain rates considered range from 0.0001 1/s,
tion of the tension behavior of these cementitious materials. which represents pseudo static loading, to 0.1 1/s, which is com-
Examples of such research efforts can be found in Refs. [6,812]. monly considered as representative of seismic loading rates
[18,19]. Dogbone shaped UHP-FRC specimens are used in the test
Corresponding author. Tel.: +1 734 764 5617. program and their performance is evaluated in terms of their rst
E-mail address: eltawil@umich.edu (S. El-Tawil).
cracking strength, post-cracking strength, energy absorption

http://dx.doi.org/10.1016/j.cemconcomp.2014.10.002
0958-9465/ 2014 Elsevier Ltd. All rights reserved.
16 S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524

capacity, strain capacity, elastic modulus, ber tensile stress and rate-dependent bridging law expressing the relation between ten-
number of cracks within the gage length. The ber tensile stress sile stress and crack opening. Wille et al. [33] reported on the strain
is the average stress resisted by the bers at the maximum post- rate dependent tensile behavior of UHP-FRC with different ber
cracking strength of the composite. Dogbone shaped specimens volume fraction at strain rates ranging from 0.0001 to 0.1 1/s. They
are selected for a tensile testing program following earlier research observed that strength and energy absorption capacity both
reported in Wille et al. [20]. Such specimens fulll AASHTO T 132- increased with an increase in ber volume fraction for a given
87 [21] requirements. strain rate.
As previously noted in the introduction, and as is clear from the
limited references listed, there is little information on the strain
2. Strain rate effect on FRC under tension rate-dependent response of UHP-FRC, especially under tension.
This study is geared towards addressing this gap by providing
Numerous tests have been carried out to investigate the information on key parameters that inuence UHP-FRC behavior
response of FRC under tension loading applied at various strain under tension.
rates. Krmeling and Reinhardt [22] carried out experimental
studies to characterize the tensile behavior of steel ber reinforced
3. Experimental program
concrete at strain rates of 1.25  10620 1/s. They noted a consid-
erable increase in tensile strength, strain at maximum stress and
A targeted experimental program was carried out to investigate
fracture energy associated with the rate effect. Zhu et al. [23] con-
the rate sensitivity of UHP-FRC under uniaxial tension. The mix
ducted a series of dynamic tensile tests (loading speed: 1000 mm/
proportions of UHP-FRC with three different volume fractions of
s) for three types of fabric (carbon, alkali resistant glass, and poly-
ber are given in Table 1. Further details about the mix constitu-
ethylene) reinforced cement composites using a high speed
ents can be found in Wille et al. [11]. Five different bers were used
servo-hydraulic testing machine, and compared their mechanical
with properties described in Table 2. The twisted bers were made
properties with the cases under quasi-static loading. They reported
in the lab out of round wire stock. They had square cross section
a signicant decrease in strength of the carbon composite with
with a pitch of 5 mm, where the ber Pitch is the length of one full
increased loading rate. They concluded that the strength decrease
(360-degree) twist around the ber axis [34]. The different bers
under high speed loading is mainly due to sliding friction of la-
are designated by their type, diameter and length as shown in
ments against their neighboring laments. Other experimental
Table 2, e.g. S-0.2-25 is a straight smooth ber with 0.2 mm diam-
research on the impact behavior of FRC can also be found in
eter and 25 mm length. Each UHP-FRC test series is designated by
[2426].
the ber used to make it, appended with a number that reects the
HPFRCC materials were investigated for their use in seismic
volume fraction of ber, e.g. S-0.2-25-2% represents a mix with 2%
applications by several researchers [27,28]. Maalej et al. [29] inves-
ber volume fraction. Nine series of UHP-FRC tensile test speci-
tigated the tensile behavior of ECC containing a combination of
mens were prepared and tested as shown in Table 3.
high-modulus steel bers and relatively low-modulus polyethyl-
Four strain rates were considered in the tests, ranging from
ene bers at strain rate ranging from 2  106 to 0.2 1/s. They con-
quasi-static (e_ 0:0001 1/s) to seismic e_ 0:1 1/s), and applied
cluded that the tensile strength doubled but there was no obvious
using a hydraulic servo-controlled testing machine (MTS-810). At
change in strain capacity with increasing strain rate. Yang and Li
least three specimens for each loading rate in each test series were
[30] investigated the tensile strength of ECC reinforced with poly-
tested, resulting in 108 specimens in total (9 series  4 loading
vinyl alcohol (PVA) bers at strain rate increasing from 105 to
rates  3 specimens) as summarized in Table 3.
101 1/s. They found that the tensile strength of ECC doubled
whereas the strain capacity decreased from 3% to 0.5%. They attrib-
uted these observations to the sensitivity of ber interfacial chem-
Table 1
ical bond strength to loading speed. Kim et al. [31] investigated the Mixtures proportions by weight (based on Wille et al. [11]).
tensile behavior of HPFRCC using two types of deformed, high-
Fiber volume fraction 1% 2% 3%
strength steel bers, namely twisted and hooked bers at strain
rates of 104 to 101 1/s. They observed that the dynamic increase Cement 1.00 1.00 1.00
Silica fume 0.25 0.25 0.25
factors (DIF) increased up to 2.0 and 1.7 for the rst-cracking
Glass powder 0.25 0.25 0.25
strength and post-cracking strength, respectively, depending on Water 0.22 0.22 0.22
ber type, ber volume fraction, and matrix strength. In contrast Superplasticizer 0.005 0.005 0.005
to the ndings of Yang and Li [30] for ECC, strain capacity remained Sand Aa 0.27 0.26 0.26
largely unaffected by strain rate. Douglas and Billington [19] inves- Sand Bb 1.07 1.05 1.02

tigated the strain rate sensitivity of two HPFRCC materials (PVA a


Maximum grain size = 0.2 mm (1/128 in.).
ber reinforced ECC and twisted high strength steel ber rein- b
Maximum grain size = 0.8 mm (1/32 in.).
forced HPFRCC) in cylindrical specimens under monotonic tension
at strain rates ranging from 2  105 to 0.2 1/s. Both HPFRCC
materials showed enhancements in strengths (25120% for ECC Table 2
and 77165% for steel ber reinforced HPFRCC) as strain rate Properties of steel bers used in this study.
increased. Both types of materials experienced a 5055% decrease
Notation Form df lf lf/df Tensile strength MPa
in strain capacity as the strain rate increased. They noted that (mm) (mm) (ksi)
unlike thin coupon specimens which tend to align bers along
T-0.3-25 Twisteda 0.3 25 83 2670 (387)b
the loading direction, cylindrical specimens allow for random, S-0.2-25 Straight 0.2 25 125 2860 (415)
three-dimensional ber alignment, leading to less efcient ber S-0.4-25 Straight 0.4 25 62.5 1850 (268)
bridging. S-0.3-18 Straight 0.3 18 60 2330 (338)
Fujikake et al. [32] investigated the tensile behavior of UHP-FRC T-0.3-18 Twisteda 0.3 18 60 2670 (387)b

using the commercial mix of Ductal Premix under various strain a


Manufactured out of round wire with df = 0.30 mm at the University of Michi-
rates ranging from 106 to 0.5 1/s. Based on the obtained stress- gan Structural Laboratories.
b
elongation relationships in uniaxial tension, they proposed a Tensile strength of ber after twisting.
S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524 17

Table 3 a dry condition at the age of 28 days, after 24 h of drying in the lab-
UHP-FRC test series investigated in this study. oratory environment.
Test series Fiber volume fraction (Vf) Vf(lf/dF) Strain rate
1% 2% 3%
3.2. Test setup and procedure
T-0.3-25-1% X 0.83 0.1
T-0.3-25-2% X 1.67 0.01
T-0.3-25-3% X 2.50 0.001 Wille et al. [20] surveyed the various techniques used for
0.0001 tensile testing of cementitious composites and selected a specic
S-0.2-25-1% X 1.25 0.1 method with the best ability to capture strain hardening behav-
0.01 ior. Fig. 1 shows the geometry of the specimen and test set up
S-0.2-25-2% X 2.50 0.001 that were also adopted in this research. The gage length of the
0.0001
dogbone type tensile test specimen is 76 mm (=3 inch). Tests
S-0.4-25-2% X 1.25 0.1 were conducted using displacement control, where the displace-
0.01
ment of the actuator in the hydraulic servo-controlled testing
S-0.4-25-3% X 1.88 0.001
0.0001 machine was used as the control variable. Two optical markers
for noncontact displacement measurement were attached to
S-0.3-18-2% X 1.20 0.1
0.01 the surface of the specimen. Three cameras were used to mea-
0.001 sure the movement of the markers, from which the measure-
0.0001 ment system calculated the deformation of the specimen. The
T-0.3-18-2% X 1.20 0.1 applicability of the measurement system for this type of
0.01 application was conrmed in a similar study done by Kim
0.001 et al. [31] for HPFRCC. The tensile load history was recorded
0.0001
via the load cell attached to the testing machine and synchro-
nized to the deformation history recorded by the measurement
system.
3.1. Materials and fabrication

A Hobart type laboratory mixer was used to prepare the UHP- 3.3. Test results
FRC mixture. The mixing procedure followed Wille et al. [11]. First,
silica fume was mixed with all the sand for approximately 5 min. Referring to Fig. 2a and b, the parameters considered in this
Then, cement and glass powder were added and mixed together investigation are: (1) the rst cracking strength (rcc), (2) the post
for at least another 5 min before the water and superplasticizer cracking strength (rpc), (3) the energy absorption capacity (g), (4)
were added. The cementitious mixture became uid after approx- the strain capacity, which is the strain value at post cracking
imately 5 min of adding the liquid ingredients. When the cementi- strength (epc), (5) the elastic modulus (Ecc) up to the rst cracking
tious mixture started to show adequate owability and viscosity, strength, (6) the ber tensile stress (rfpc) and (7) number of cracks
bers were dispersed by hand in the mix. The cementitious mix- within the gage length. It is important to note that rcc is the turn-
ture with uniformly distributed bers was poured into dogbone ing point between elastic and strain hardening parts and does not
shaped molds without any vibration. After casting, the specimens necessarily represent the stress when the rst crack developed. It
were covered with plastic sheets and stored at room temperature also should be noted that g is the area under stressstrain curve
for 24 h prior to demolding. The specimens were then placed in a up to an arbitrarily selected value: ru = 0.95. It was noted from
water tank for an additional 26 days without any special curing experience that consistent softening behavior generally occurred
such as heat or pressure treatment. All specimens were tested in beyond this point (compare the softening regimes in Fig. 2a and b).

Fig. 1. Direct tensile test setup for this experimental research according to Wille et al. [20].
18 S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524

Fig. 2. Typical interpretation and measurement of test results: (a and b) typical variables; (c) typical averaging of stressstrain curve (T-0.3-18-2% & e_ 0:01 1/s) (Note: strain
valid up to peak stress only.)

The variable rfpc represents the average tensile stress in the stressstrain curves for each series under four different strain rates
ber at maximum stress in the composite. It is calculated by Eq. are plotted in Figs. 35. All test series maintained performance
(1) as follows [20]: Level 3 as classied in Naaman and Reinhardt [35] for all loading
rates, i.e., strain hardening was observed for all loading rates.
rpc
rfpc 1 Furthermore, T-0.3-25-2%, T-0.3-25-3%, and S-0.2-25-2% series
/  Vf
maintained performance Level 4 as classied by Naaman and
where / is a ber orientation factor, taken as 0.9 for all series fol- Reinhardt [35], which is dened as high energy absorbing with
lowing Wille et al. [20]. The number is quite high because most of g P 50 kJ/m3 as suggested by Wille et al. [20] for all loading rates.
the bers are aligned in the load direction due to the casting According to Naaman [36], the post cracking strength of com-
method. Vf is the ber volume fraction. Table 4 summarizes all posites reinforced with short discontinuous bers is proportional
the test results. All the values given in Table 4 are averaged values to ber aspect ratio and their volume fraction as follows:
from at least three specimens. A dynamic increase factor (DIF),
rpc kseq V f lf =df 2
which is the ratio between dynamic response and static response,
is computed for four parameters (rcc, rpc, g, epc) to effectively where k is a factor equal to the product of several coefcients that
illustrate the effect of strain rate on strength and other material account for average pullout length, group reduction effect, and ber
parameters. In this study the values measured at the lowest strain orientation effect. seq is the equivalent bond strength. The product
rate of 0.0001 1/s were considered to be the reference, i.e. pseudo- V(lf/df) is termed the ber reinforcing index. Table 5 shows the
static loading rate. The DIF values are also listed in Table 4. effect of the ber aspect ratio on post cracking strength based on
Fig. 2c shows an example averaged stressstrain curve, plotted the results of Eq. (2) for the pseudostatic and seismic loading rates
along with raw data to give a sense of the spread in the test data. In and 2% ber content. Moreover, the surface energy of pull-out,
order to get an averaged curve, the raw data set for each individual which gives a good estimate of the fracture energy in the type of
plot is divided into two parts, one before and one after the peak tensile tests carried out here can be put in the following form [10]:
stress point. Each point on either part is assigned a fraction of peak
stress strain. The data at each fraction point is then averaged cg fseq V f l2f =df 3
across all curves for both portions of each curve. It is clear from
Fig. 2c, and this was generally observed throughout the test pro- where f is the product of k in Eq. (2) and the ratio of average bridg-
gram, that specimen-to-specimen variability within each strain ing stress to the maximum post-cracking stress over the expected
rate was relatively low for all parameters. The average tensile maximum pull-out length. It can be thus observed that the expected
S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524 19

Table 4
Rate effect on the average tensile properties of UHP-FRC.

Test series Strain rate (1/s) First cracking Post cracking Energy Strain capacity Elastic Fiber tensile Average
strength (rcc) strength (rpc) absorption (epc) modulus (Ecc) stress (rfpc) number
capacity (g) of cracks
Avg. Std. dev. DIF Avg. Std. dev. DIF kJ/m3 DIF % DIF GPa MPa EA
MPa MPa
T-0.3-25-1% 0.0001 6.89 0.84 1.00 8.31 0.28 1.00 33.5 1.00 0.19 1.00 54.5 923 7
0.001 7.54 0.29 1.09 8.91 0.65 1.07 37.9 1.13 0.23 1.20 54.6 990 7
0.01 7.04 0.83 1.02 9.42 0.71 1.13 47.6 1.42 0.26 1.38 57.1 1050 8
0.1 8.13 1.1 1.18 10.8 0.22 1.31 60.7 1.81 0.34 1.80 58.9 1200 9
T-0.3-25-2% 0.0001 9.03 0.31 1.00 11.6 1.2 1.00 63.5 1.00 0.42 1.00 60.1 644 13
0.001 9.14 0.92 1.01 12.6 0.85 1.08 83.4 1.31 0.50 1.20 60.4 698 14
0.01 8.73 1.3 0.97 13.8 0.59 1.19 84.1 1.32 0.55 1.30 54.3 764 11
0.1 8.87 1.9 0.98 14.0 0.14 1.21 103 1.61 0.70 1.66 60.8 779 11
T-0.3-25-3% 0.0001 11.8 0.28 1.00 20.9 1.8 1.00 108 1.00 0.48 1.00 71.3 775 13
0.001 11.5 0.50 0.97 21.7 0.94 1.04 143 1.32 0.62 1.28 70.0 805 13
0.01 12.0 0.61 1.02 23.5 2.1 1.12 177 1.64 0.68 1.41 70.3 869 12
0.1 12.1 0.68 1.03 24.1 1.9 1.15 186 1.72 0.79 1.64 76.1 892 12
S-0.2-25-1% 0.0001 6.22 0.79 1.00 8.11 1.4 1.00 34.5 1.00 0.33 1.00 47.7 901 9
0.001 7.86 0.19 1.26 9.49 0.60 1.17 51.9 1.50 0.49 1.47 56.1 1050 13
0.01 6.92 2.0 1.11 8.29 2.0 1.02 48.1 1.39 0.39 1.17 51.1 921 10
0.1 6.70 0.65 1.08 9.01 1.4 1.11 47.6 1.38 0.43 1.30 54.0 1000 13
S-0.2-25-2% 0.0001 9.73 0.77 1.00 14.9 0.44 1.00 84.4 1.00 0.55 1.00 66.6 829 13
0.001 9.19 1.7 0.94 14.6 2.3 0.98 83.1 0.98 0.59 1.09 63.2 813 12
0.01 11.2 0.98 1.15 15.5 2.7 1.04 87.2 1.03 0.61 1.11 64.8 858 14
0.1 9.65 1.7 0.99 15.5 0.26 1.04 89.3 1.06 0.55 1.02 67.8 862 17
S-0.4-25-2% 0.0001 6.86 0.83 1.00 8.32 1.9 1.00 35.9 1.00 0.27 1.00 54.5 462 8
0.001 6.78 0.55 0.99 8.65 0.72 1.04 36.8 1.02 0.30 1.10 51.2 481 10
0.01 8.13 0.88 1.18 9.96 0.43 1.20 48.3 1.34 0.35 1.28 57.7 553 9
0.1 8.30 1.5 1.21 9.72 1.0 1.17 56.1 1.56 0.42 1.55 55.0 540 9
S-0.4-25-3% 0.0001 7.67 0.35 1.00 9.41 0.62 1.00 41.3 1.00 0.24 1.00 61.4 349 10
0.001 8.20 1.0 1.07 9.63 0.47 1.02 48.5 1.17 0.28 1.15 61.0 357 9
0.01 8.69 1.2 1.13 12.5 1.4 1.33 74.8 1.81 0.47 1.96 62.8 463 9
0.1 8.71 1.4 1.14 12.8 0.48 1.36 78.8 1.91 0.47 1.94 59.9 474 7
S-0.3-18-2% 0.0001 7.05 0.80 1.00 8.44 0.26 1.00 20.0 1.00 0.17 1.00 56.5 469 7
0.001 8.45 1.3 1.20 9.80 0.98 1.16 28.1 1.40 0.21 1.22 61.1 544 7
0.01 7.05 0.63 1.00 9.32 1.4 1.10 34.2 1.70 0.23 1.35 53.1 518 6
0.1 7.66 0.90 1.09 9.70 0.60 1.15 38.3 1.91 0.31 1.80 53.4 539 6
T-0.3-18-2% 0.0001 7.73 0.24 1.00 9.04 0.29 1.00 27.8 1.00 0.27 1.00 56.1 502 7
0.001 8.53 0.89 1.10 10.4 0.52 1.15 34.6 1.25 0.22 0.86 55.0 578 8
0.01 7.32 0.12 0.95 9.63 0.18 1.07 47.3 1.70 0.29 1.15 50.4 535 7
0.1 7.37 1.5 0.95 10.8 1.2 1.20 54.5 1.96 0.44 1.70 56.9 600 7

fracture energy varies with the square of the ber length or with the the effect of strain rate is clear, that is, the post-cracking strength
product of the ber length by the reinforcing index. increases with an increase in strain rate.
The effect of ber aspect ratio by itself, lf/df, can be evaluated by
comparing series with 2% volume fraction from the three gures
4. Evaluation of experimental results
(Figs. 35 and Table 4). Not only does the post cracking strength
generally increases with ber aspect ratio, but also the rst crack-
The average numerical results are summarized in Table 4.
ing strength. For example, S-0.2-25-2% shows better performance
Figs. 37 and Fig. 9 illustrate the effects of various parameters on
than S-0.4-25-2% by 33% in rst cracking strength, 66% in post
the response of the composite at various strain rates.
cracking strength, 78% in strain capacity and 100% in energy
absorption averaged across all strain rates, respectively.
4.1. Separate effects of ber volume fraction, aspect ratio, length, and Although the elastic modulus was measured for all specimens
type tested, the large scatter in the data makes it difcult to draw a rm
conclusion regarding the effect of aspect ratio or volume fraction of
Fig. 3 illustrates the effects of volume fraction of ber and strain ber or strain rate on elastic modulus (see Table 4). Overall, the
rate on the tensile response of the composite reinforced with twisted ber series showed better performance than straight ber
twisted steel bers. It can be observed that the post-cracking series because of the additional anchorage effect associated with
strength and the corresponding strain generally increase with an the untwisting action that occurs during pullout. Kim et al. [37]
increase in both volume fraction of ber and strain rate. Typically discuss the untwisting mechanism. However, for a given volume
the post-cracking strength more than doubles when the volume fraction, the T-0.3-25 and the S-0.2-25 bers show similar mechan-
fraction of ber goes from 1% to 3%. Fig. 4 illustrates similar trends ical performances. For example, The T-0.3-18-2% series shows mar-
when using smooth steel bers with 1% and 2% volume content. ginally better mechanical properties than S-0.3-18-2%. The
Fig. 5 shows the response of four series of tests using either smooth increase is most prominent in energy absorption capacity, where
or twisted steel bers having about the same aspect ratio; here also it is 36% higher on average across all strain rates. This is attributed
20 S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524

(a) (a)

T-0.3-25-1% S-0.2-25-1%

(b) (b)

T-0.3-25-2% S-0.2-25-2%

Fig. 4. Observed rate effect on the tensile behavior of UHP-FRC using smooth steel
(c) bers having a high aspect ratio with increasing volume content (Note: strain valid
up to peak stress only.)

packing density of the cementitious matrix around the ber. It is


also noted that the equivalent bond is an average value estimated
over a small crack opening (related to epc) and is different from a
similar value obtained from a complete ber pull-out curve.

T-0.3-25-3%
4.3. Effect of reinforcing index Vf(lf/f)

Eq. (2) suggests that, for the same kseq , the post-cracking
strength of the composite is directly proportional to the ber rein-
Fig. 3. Observed rate effect on the tensile behavior of UHP-FRC using twisted steel
forcing index, that is, the product of the volume fraction times the
bers with increasing volume content and same ber aspect ratio (Note: strain valid
up to peak stress only.) aspect ratio of ber. Fig. 6a provides a summary of the data
observed for the post-cracking strength versus the reinforcing
index at different strain rates. The trend predicted by Eq. (2) is
to the fact that the number of S-0.2-25 bers, which are thinner clearly conrmed by the data, that is, the post-cracking strength
than the T-0.3-25 bers, is 225% more in a unit volume than increases with the ber reinforcing index. To best quantify the
T-0.3-25 bers, making up for the additional anchorage mecha- data, the least square t lines for the four loading rates ranging
nism of the twisted bers. The good performance of S-0.2-25 bers, from quasi-static (0.0001 1/s) to seismic (0.1 1/s) are listed in
however, comes at a price because the larger number of the bers Table 6 and the extreme ones shown in Fig. 6. It can be further
makes it more difcult to mix. In general, it was not possible to observed that the post-cracking strength increases with strain rate.
make mixes with more than 2% volume fraction of S-0.2-25 bers. Fig. 6b illustrates the variation of the rst cracking strength of
In contrast, mixes with up to 3% volume fraction of T-0.3-25 bers the composite versus the ber reinforcing index at different strain
were feasible. The opposite argument can be made for the T-0.3-25 rates. Here also the least square t lines of the data are plotted and
bers versus the S-0.4-25 bers. suggest that the rst cracking strength increases with both the
reinforcing index and the strain rate.

4.2. Equivalent bond


4.4. Energy absorption capacity
The results of all test series with 2% volume fraction of bers are
summarized in Table 5. Looking across all entries in the 4th and Following the format of Eq. (3), the observed energy absorption
6th columns suggests that kseq is almost independent of ber type, capacity (see Fig. 2) is plotted in Fig. 7 versus the quantity
2
diameter and length. This unexpected result may be attributed to V f lf =df for different strain rates. The trend observed conrms the-
the use of a UHPC matrix where the bond for smooth straight ber oretical predictions; that is, the energy increases with both
2
is reported to be excellent due to surface abrasion (see recent V f lf =df and the strain rate. The least square t lines provided in
paper by Wille and Naaman [34]), and is likely due to the very high the gures and Table 6 offer a good mean to quantify the data.
S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524 21

Table 5
(a) Derivation of equivalent bond strength for UHP-FRC with 2% ber content.

Test series lf/df e_ = 0.0001 1/s e_ = 0.1 1/s


rpc (MPa) kseq (MPa) rpc (MPa) kseq (MPa)

T-0.3-25-2% 83 11.6 6.98 14.0 8.44


S-0.2-25-2% 125 14.9 5.97 15.5 6.20
S-0.4-25-2% 62.5 8.32 6.65 9.72 7.78
S-0.3-18-2% 60 8.44 7.04 9.70 8.08
T-0.3-18-2% 60 9.04 7.53 10.8 9.00

S-0.4-25-2 %

(b)

S-0.4-25.3 %

(c)

S-0.3-18-2%

Fig. 6. Effects of the ber reinforcing index on mechanical properties of UHP-FRC:


(a) Post cracking strength; (b) rst cracking strength.

(d)

T-0.3-18-2%

Fig. 5. Observed rate effect on the tensile behavior of UHP-FRC using bers with
about same aspect ratio (Note: strain valid up to peak stress only.)

2
Fig. 7. Effects of V f lf =df on energy absorption capacity of UHP-FRC.
4.5. Cracking

Fig. 8 shows observed cracking patterns of UHP-FRC, which capacity increase as ber volume fraction increases. However, it is
indicates UHP-FRC materials exhibit multiple cracking after rst difcult to draw a clear conclusion about the effect of strain rate on
cracking along with hardening behavior. The number of cracks the number of cracks, due to the variability in the test data. It
within the gage length is also an important indicator of energy should be noted that the crack counting process itself is somewhat
absorption capacity and strain at peak load. It appears from the subjective because of the difculty of ascertaining the presence of a
results (see Table 4) that both the number of cracks and the energy crack after unloading.
22 S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524

Table 6
The least square t approximations of results in Figs. 7 and 8.

Index Strain rate T-ber S-ber


(1/s)
First cracking e_ 0:0001 rcc = 2.98Vf(lf/ rcc = 2.56Vf(lf/
strength df) + 4.25 df) + 3.86
e_ 0:001 rcc = 2.31Vf(lf/ rcc = 1.10Vf(lf/
df) + 5.59 df) + 6.33
e_ 0:01 rcc = 3.11Vf(lf/ rcc = 2.93Vf(lf/
df) + 3.96 df) + 3.67
e_ 0:1 rcc = 2.66Vf(lf/ rcc = 1.67Vf(lf/
df) + 5.00 df) + 5.52
Post cracking e_ 0:0001 rpc = 7.98Vf(lf/df) rpc = 5.93Vf(lf/df)
strength e_ 0:001 rpc = 8.41Vf(lf/df) rpc = 6.12Vf(lf/df)
e_ 0:01 rpc = 9.04Vf(lf/df) rpc = 6.65Vf(lf/df)
e_ 0:1 rpc = 9.40Vf(lf/df) rpc = 6.76Vf(lf/df)
Energy absorption e_ 0:0001 2
g 1:66V f lf =df
2
g 1:17V f lf =df
capacity e_ 0:001 2 2
g 2:16V f lf =df g 1:27V f lf =df
e_ 0:01 g
2
2:56V f lf =df
2
g 1:48V f lf =df
e_ 0:1 g
2
2:82V f lf =df
2
g 1:55V f lf =df

T-0.3-25-2% S-0.2-25-2%
= 0.001 1/s = 0.1 1/s

Fig. 9. Rate effect on the ber tensile stress of UHP-FRC at maximum load: (a) with
different ber volume fractions; (b) with 2% ber volume fraction.

strain capacity. Plots of DIF versus strain rate for rst cracking
strength, post cracking strength, energy absorption capacity and
strain capacity are compared in Fig. 10. The increases in DIF for
the four parameters can be reasonably simulated by a log-linear
trend with the increase in strain rate. Fig. 10a shows that the
highest rate sensitivity of rst cracking strength occurs in series
with S-0.4-25 bers, while the lowest occurs in series with T-0.3-
18 bers. Similarly, S-0.4-25 series shows the highest rate sensitiv-
ity in post cracking strength, energy absorption capacity and strain
capacity, but S-0.2-25 series shows less rate sensitivity in those
parameters as shown in Fig. 10bd, respectively.
Fitted log-linear relationships for rst cracking strength, post
cracking strength, energy absorption capacity and strain capacity
with strain rate are shown in Fig. 10. The gure indicates that
Fig. 8. Representative multiple cracking patterns in UHP-FRC specimens.
the increase in DIF is moderate and nearly linear in log-linear space
at strain rates up to 0.1 1/s. This general tendency has been
4.6. Fiber tensile stress observed in other cement-based materials such as ordinary con-
crete, high-performance concrete or UHPC [31,33,38].
Fig. 9 shows rate effects on the ber tensile stress of UHP-FRC.
The ber tensile stress (Eq. (1)), which represents the effectiveness
4.8. General trends in the test data
of ber usage, increases as strain rate increases. Even though
higher ber volume fraction led to better mechanical properties
Several general trends can be seen in the tension test results:
of UHP-FRC, it generally decreases, from a qualitative perspective,
(1) Tensile strength, energy absorption capacity and strain capacity
as ber volume fraction increases, in contrast to other mechanical
all increase as ber volume fraction increases for all ber series
parameters (see Table 4), likely due to the ber-group effect.
under all strain rates; (2) while twisted bers provided somewhat
Increases in performance attributed to volume fraction are also
better performance than equivalent straight bers, the mechanical
limited by another practical limitation, i.e. difculty of mixing with
anchorage advantage of twisted bers over smooth bers for post-
a large quantity of bers.
cracking strength could be overcome by increasing ber aspect
ratio; (3) the ber aspect ratio inuences post cracking strength
4.7. DIF and ber tensile stress, while ber shape and length play impor-
tant roles for strain capacity and energy absorption capacity; (4)
DIF was evaluated for four key parameters, rst cracking similar to other concretes, log-linear relationships of DIF of UHP-
strength, post cracking strength, energy absorption capacity and FRC are moderate and nearly linear at strain rates up to 0.1 1/s.;
S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524 23

T-0.3-25 series at the 2% ber content suggests that the mechanical


anchorage advantage can be accounted for by changing ber aspect
ratio. This has practical signicance because the price of straight
ber is generally cheaper than that of twisted ber.
Unlike results reported for HPFRCC (e.g. in Yang and Li [30] and
Douglas and Billington [19]), the strain capacity does not decrease
as strain rate increases. In fact, it almost doubles depending on ber
type and other mechanical properties such as tensile strengths and
energy absorption capacity also increase substantially as strain rate
increases. These results indicate that UHP-FRC is particularly prom-
ising for applications that involve seismic, impact or blast.

5. Conclusions

This experimental study investigated the direct tensile behavior


of UHP-FRC with ve different steel bers at strain rates ranging
from quasi-static (0.0001 1/s) to seismic (0.1 1/s). The tests were
conducted using a hydraulic servo-controlled testing machine
and results were evaluated in terms of rst cracking strength,
post-cracking strength, energy absorption capacity, strain capacity,
elastic modulus, ber tensile stress and number of cracks within
the gage length. Log-linear relationships of DIF for rst cracking
strength, post-cracking strength, energy absorption capacity and
strain capacity were presented based on the test data. The key
observations and ndings of this study can be summarized as
follows:

1. An increase in the ber volume fraction led to increases in the


composite tensile strength, energy absorption capacity, strain
capacity and elastic modulus for all ber tested under all strain
rates. In contrast, the ber tensile stress did not show a clear
trend, likely because it was inuenced by the ber group effect.
2. For the UHP-FRC tested in this study, the equivalent bond
strength for the straight steel bers seems to be of the same
order as that of the twisted bers.
3. For similar equivalent bond strength, the observed post crack-
ing strength of the composite varies linearly with the ber rein-
forcing index (Vf(lf/df)) as predicted theoretically from Eq. (2).
4. For similar equivalent bond, the observed energy absorption
capacity up to peak load varies linearly with the product
2
V f lf =df , as predicted from theory.
5. For the range of strain rates used in this study (0.0001 1/s to
0.1 1/s) both the post-cracking peak strength and the fracture
energy up to peak load increase with an increase in strain rate.
It was difcult to draw a rm conclusion regarding the rate sen-
sitivity of the elastic modulus and the number of cracks because
of observed large scatter in the test data.

It is hoped that the results from this experimental research will


provide some basic information for developing rate dependent
constitutive models for UHP-FRC. Such models are needed to sim-
ulate the response of UHP-FRC structures subjected to extreme
loading. Furthermore, the fact that UHP-FRC shows substantial
increases in energy absorption capacity as strain rate increases
implies that the material is especially promising for blast and
Fig. 10. Dynamic increase factor (DIF) of UHP-FRC: (a) First cracking strength; (b) impact applications. However, additional research is needed to
post cracking strength; (c) energy absorption capacity; (d) strain capacity.
investigate and characterize the response of UHP-FRC at strain
rates higher than those selected in this study and for a much
(5) even though there is no rate sensitivity in the number of cracks
broader range of parameters.
(see Table 4), multiple cracks developed in all UHP-FRC specimens
(see Fig. 8) promoting strain hardening behavior of UHP-FRC under
tension. Acknowledgments
Twisted bers lead to better overall mechanical performance
than equivalent straight bers, primarily because their mechanical The research described herein was sponsored by the National
anchorage mechanism is benecial. However, the fact that the Science Foundation- United States under Grant No. CMS 0928193
S-0.2-25 series shows similar mechanical performance to the and the University of Michigan, Ann Arbor. The opinions expressed
24 S. Pyo et al. / Cement & Concrete Composites 56 (2015) 1524

in this paper are those of the authors and do not necessarily reect [19] Douglas KS, Billington SL. Strain rate dependence of HPFRCC cylinders in
monotonic tension. Mater. Struct. 2011;44:391404.
the views of the sponsors.
[20] Wille K, El-Tawil S, Naaman AE. Properties of strain hardening ultra high
performance ber reinforced concrete (UHP-FRC) under direct tensile loading.
Cem Concr Compos 2014;48:5366.
References [21] AASHTO T 132-87. Standard method of test for tensile strength of hydraulic
cement mortars. American Association of State and Highway Transportation
[1] Brandt AM. Fibre reinforced cement-based (FRC) composites after over 40 Ofcials; 2009. 8 pages.
years of development in building and civil engineering. Compos Struct [22] Krmeling HA, Reinhardt HW. Strain rate effects on steel bre concrete in
2008;86:39. uniaxial tension. Int J Cem Compos Lightweight Concr 2006;36:13718.
[2] Naaman AE, Homrich JR. Tensile stress-strain properties of SIFCON. ACI Mater J [23] Zhu D, Peled A, Mobasher B. Dynamic tensile testing of fabriccement
1989;86:24451. composites. Constr Build Mater 2011;25:38595.
[3] Rossi P. High performance multimodal ber reinforced cement composites [24] Lok TS, Zhao PJ. Impact response of steel ber-reinforced concrete using a split
(HPMFRCC): the LCPC experience. ACI Mater J 1997;94:47883. Hopkinson pressure bar. J Mater Civ Eng 2004;16:549.
[4] Markovic I. High-performance hybrid-bre concrete: development and [25] Mechtcherine V, Millon O, Butler M, Thoma K. Mechanical behaviour of strain
utilisation, PhD thesis. Delft University of Technology; 2006. hardening cement-based composites under impact loading. Cem Concr
[5] Naaman AE, Reinhardt HW. Proposed classication of HPFRC composites based Compos 2011;33:111.
on their tensile response. Mater Struct 2006;39:54755. [26] Bindiganavile V, Banthia N, Aarup B. Impact response of ultra-high-strength
[6] Li VC, Wang S, Wu C. Tensile strain-hardening behavior of polyvinyl alcohol ber-reinforced cement composite. ACI Mater J 2002;99:5438.
engineered cementitious composite (PVA-ECC). ACI Mater J 2001;98:48392. [27] Li VC. On engineered cementitious composites (ECC). J Adv Concr Technol
[7] Wille K, Kim DJ, Naaman AE. Strain-hardening UHP-FRC with low ber 2003;1:21530.
contents. Mater Struct 2011;44:58398. [28] Parra-Montesinos G. HPFRCC in earthquake-resistant structures: current
[8] Dugat J, Roux N, Bernier G. Mechanical properties of reactive powder knowledge and future trends. In: Proceedings of 4th international RILEM
concretes. Mater Struct 1996;29:23340. workshop on high performance ber reinforced cement composites (HPFRCC
[9] Fischer G, Fukuyama H, Li VC. Inuence of matrix ductility on tension- 4). RILEM Publications; 2003. p. 45372.
stiffening behavior of steel reinforced engineered cementitious composites [29] Maalej M, Quek ST, Zhang J. Behavior of hybrid-ber engineered cementitious
(ECC). ACI Struct J 2002;99:10411. composites subjected to dynamic tensile loading and projectile impact. J Mater
[10] Naaman AE. Toughness, ductility, surface energy and deection-hardening FRC Civ Eng 2005;17:14352.
composites. In: Proceedings of the JCI international workshop on ductile ber [30] Yang E, Li VC. Rate dependence in engineered cementitious composites. In:
reinforced cementitious composites (DFRCC) application and evaluation Proceedings HPFRCC-2005 international workshop. Honolulu, Hawaii, USA;
(DFRCC-02), Takayama, Japan; 2002. p. 3357. 2005.
[11] Wille K, Naaman AE, Parra-Montesinos GJ. Ultra-high performance concrete [31] Kim DJ, El-Tawil S, Naaman AE. Rate-dependent tensile behavior of high
with compressive strength exceeding 150 MPa (22 ksi): a simpler way. ACI performance ber reinforced cementitious composites. Mater Struct
Mater J 2011;108:4654. 2009;42:399414.
[12] Wille K, Parra-Montesinos GJ. Effect of beam size, casting method, and support [32] Fujikake K, Senga T, Ueda N, Ohno T, Katagiri M. Effects of strain rate on
conditions on exural behavior of ultra-high-performance ber-reinforced tensile behavior of reactive powder concrete. J Adv Concr Technol
concrete. ACI Mater J 2012;109:37988. 2006;4:7984.
[13] Rossi P. Ultra high performance concretes. Concr Int 2008;30:314. [33] Wille K, El-Tawil S, Naaman AE. Strain rate dependent tensile behavior of
[14] Graybeal BA. Ultra-high performance concrete Report No. FHWA-HRT-11- ultra-high performance ber reinforced concrete. In: High Performance
038. Washington, DC: Federal Highway Administration; 2011. Fiber Reinforced Cement Composites 6. Netherlands: Springer; 2012. p.
[15] Naaman AE, Wille K. The path to ultra-high performance ber reinforced 38187.
concrete (UHP-FRC): ve decades of progress. In: 3rd International symposium [34] Wille K, Naaman AE. Pullout behavior of high-strength steel bers embedded
on UHPC and nanotechnology for high performance construction in ultra-high-performance concrete. ACI Mater J 2012;109:47987.
materials. Kassel: Kassel University Press; 2012. p. 316. [35] Naaman AE, Reinhardt HW. Setting the stage: toward performance based
[16] Graybeal BA. Material Property Characterization of Ultra-High Performance classication of FRC composites. In: Proceedings of 4th international RILEM
Concrete Report No. FHWA-HRT-06-103. Washington, DC: Federal Highway workshop on high performance ber reinforced cement composites (HPFRCC
Administration; 2006. 4); 2003. p. 14.
[17] Pfeifer CG, Moeser B, Giebson C, Stark J. Durability of ultra-high-performance [36] Naaman AE. Engineered steel bers with optimal properties for reinforcement
concrete. In: Tenth ACI international conference on recent advances in of cement composites. J Adv Concr Technol 2003;1:24152.
concrete technology and sustainability issues. No. SP-261-1; 2009. [37] Kim DJ, El-Tawil S, Naaman AE. Loading rate effect on pullout behavior of
[18] Scott BD, Park R, Priestley MJN. Stressstrain behavior of concrete conned deformed steel ber. ACI Mater J 2008;105:57684.
by overlapping hoops at low and high strain rates. ACI J Proc [38] Pyo S, El-Tawil S. Crack velocity-dependent dynamic tensile behavior of
1982;79:1327. concrete. Int J Impact Eng 2013;55:6370.

You might also like