polymers

Article
Mechanical Characterization of the Tensile Properties
of Glass Fiber and Its Reinforced Polymer (GFRP)
Composite under Varying Strain Rates
and Temperatures
Yunfu Ou 1 , Deju Zhu 1, *, Huaian Zhang 1 , Liang Huang 1, *, Yiming Yao 2 , Gaosheng Li 1
and Barzin Mobasher 2
1 College of Civil Engineering, Hunan University, Changsha 410082, China; ouyunfu@hnu.edu.cn (Y.O.);
zhanghuaianjs@outlook.com (H.Z.); ligaosheng@hnu.edu.cn (G.L.)
2 School of Sustainable Engineering and Built Environment, Arizona State University, Tempe, AZ 85287, USA;
Yiming.Yao@asu.edu (Y.Y.); Barzin@asu.edu (B.M.)
* Correspondence: dzhu@hnu.edu.cn (D.Z.); huangliang@hnu.edu.cn (L.H.);
Tel.: +86-731-8882-3861 (D.Z.); +86-135-0731-5881 (L.H.)

Academic Editor: Alper Ilki

Received: 26 March 2016; Accepted: 10 May 2016; Published: 19 May 2016

Abstract: Unidirectional glass fiber reinforced polymer (GFRP) is tested at four initial strain rates (25,
50, 100 and 200 s1 ) and six temperatures (25, 0, 25, 50, 75 and 100 C) on a servo-hydraulic high-rate
testing system to investigate any possible effects on their mechanical properties and failure patterns.
Meanwhile, for the sake of illuminating strain rate and temperature effect mechanisms, glass yarn
samples were complementally tested at four different strain rates (40, 80, 120 and 160 s1 ) and varying
temperatures (25, 50, 75 and 100 C) utilizing an Instron drop-weight impact system. In addition,
quasi-static properties of GFRP and glass yarn are supplemented as references. The stressstrain
responses at varying strain rates and elevated temperatures are discussed. A Weibull statistics model
is used to quantify the degree of variability in tensile strength and to obtain Weibull parameters for
engineering applications.

Keywords: polymer-matrix composites (PMCs); mechanical properties; stress/strain curves;

deformation; statistics

1. Introduction
Fiber reinforced polymer (FRP) composites have many merits, such as high stiffness/weight
and strength/weight ratios, advanced fatigue and corrosion resistances, etc., providing significant
functional and economic benefits, ranging from strength enhancement and weight reduction to
durability features. With decreasing manufacturing costs, recently, they have won the attention of
engineers involved in the construction of civil structures [1]. Structural elements reinforced with FRP,
however, might be subjected to dynamic loadings, such as wind loads, earthquake loads, explosions, etc.
and vary temperature conditions during their service life. Under such conditions, the mechanical
properties of FRP involving Youngs Modulus, tensile strength, toughness, etc. may suffer great
changes [26]. Therefore, the investigation of the mechanical properties of FRP composites under
dynamic loadings and different temperatures is essential to design the structures with this kind
of materials.
A variety of testing techniques and procedures have been developed for gaining insight into
the dynamic responses of materials under different strain rates and different methods are suited for
different ranges of strain rate. The conventional screw drive load frame [7] is used for quasi-static

Polymers 2016, 8, 196; doi:10.3390/polym8050196 www.mdpi.com/journal/polymers

Polymers 2016, 8, 196 2 of 16

loading of coupons at a constant strain rate. In the medium strain-rate region for strain rates up to
approximately 200 s1 , servo-hydraulic high speed machines [8] or drop-weight impact systems [9]
can be used. The most widely used technique for obtaining direct determination of material properties
at strain rates between 200 and 104 s1 is the split Hopkinson bar, which was first introduced by
Kolsky [10]. With the help of those machines, rate dependent data of FRP composites can be obtained.
The tensile behaviors of carbon fiber-reinforced polymers (CFRP) under different strain rates were
studied by several researchers [7,1113]. It was reported that the tensile properties of CFRP are strain
rate dependent, while the average transverse modulus is independent of strain rate. Koerber et al. [14]
found that shear modulus and shear behavior of the carbon-epoxy material system IM7-8552 are
dependent of strain rate. Specifically, when the loading case varied from quasi-static to dynamic, the
in-plane shear modulus of elasticity, yield strength and pure failure strength increased by 25%, 88%
and 42%, respectively. Strain rate effects on aramid fiber-reinforced polymers (AFRP) were described
in literature [15]. It showed that both the Youngs modulus and tensile strength increase at the impact
loading rate, while the failure strain decreases slightly. As the most widely used FRP composites
in civil engineering due to lower cost of production, GFRP, especially its behaviors under dynamic
loadings has also drawn much attention. Barre et al. [16] determined the tensile behavior of GFRP
using a drop-weight dynamic testing machine. The results revealed that dynamic elastic modulus and
strength tend to increase with increasing strain rate. Shokrieh et al. [17] studied the tensile properties of
unidirectional GFRP composites at quasi-static and intermediate strain rates of 0.001100 s1 by means
of a servo-hydraulic testing apparatus. A significant increase of the tensile strength was observed with
increasing strain rate. Ochola et al. [18] investigated the strain rate sensitivity of GFRP at strain rates
of 103 and 450 s1 . The experimental results reported that the dynamic material strength of GFRP
increases with increasing strain rates.
Temperature effects on GFRP are another important issue that needs further investigation in order
to adopt the use of such composites in strengthening structural reinforced concrete (RC) members and
their connections when exposed to elevated temperatures and harsh environment. Hawileh et al. [19]
experimentally investigated the variation of mechanical properties in terms of the elastic modulus and
tensile strength of composite glass (C), composite glass (G) sheets and their hybrid combinations (CG)
when exposed to different temperatures, ranging from 25 to 300 C. Results showed that the elastic
modulus of the C, G and CG at 250 C was reduced by nearly 28%, 26% and 9%, respectively, and
their tensile strength at the same temperature was reduced by about 42%, 31% and 35%, respectively,
all as compared to room temperature values. Robert et al. [20] evaluated the variation of mechanical
properties of sand-coated GFRP reinforcing bars subjected to low temperatures (ranging from 0 to
100 C) and elevated temperatures (ranging from 23 to 315 C) and discovered that low temperature
has a positive influence on the strength of composites, while at very high temperatures, around about
the glass-transition temperatures of the polymer matrix, the mechanical properties, especially the
stiffness and the strength of the composites are decreased significantly. Reis et al. [6] conducted tensile
tests on GFRP at different strain rates and temperatures and found that strain rate greatly affects the
ultimate tensile strength, and the Youngs modulus is almost insensitive to it while temperature only
influences the modulus. GFRP samples with single yarn were tested at different strain rates from
quasi-static up to 160 s1 and temperatures from 25 to 100 C by Ou and Zhu [9] to investigate any
possible effects on their mechanical properties and failure patterns. The results showed that the tensile
strength, maximum strain and toughness increase with increasing strain rates at room temperature,
and the Youngs modulus, tensile strength and toughness decrease with increasing temperatures at the
strain rate of 40 s1 .
Although many efforts have been carried out on the strain rate and temperature effects on the
mechanical properties of GFRP, very limited information is available on intermediate strain rate ranging
from 25 to 200 s1 , and only few experimental results are accessible on the dynamic tensile behavior of
GFRP at elevated temperatures and harsh environment. Furthermore, as the main load-bearing body
of GFRP, glass fibers play a significant role in GFRP subjected to an impact load. Therefore, revealing
 Polymers 2016, 8, 196 3 of 16

gaining a deep insight into the mechanical behavior of GFRP under impact loads and then
optimizing structural design.
The primary objective of this study is to investigate the effects of intermediate strain rates and
Polymers 2016, 8, 196 on the mechanical properties of glass fiber bundle (yarn) and unidirectional GFRP3 of 16
temperatures
laminate samples. In the next section, the experimental procedure and test results are presented and
discussed. Then, comparison of the obtained values of tensile strength of GFRP specimens at
the variation of tensile properties of fibers at different strain rates is the key to gaining a deep insight
different strain rates with the values found in literature is displayed. In the fourth section, the results
into the mechanical
of Weibull behavior
statistical of based
analysis GFRPonunder impact loads
experimental resultsand then optimizing structural design.
are discussed.
The primary objective of this study is to investigate the effects of intermediate strain rates and
temperatures on theProgram
2. Experimental mechanical properties of glass fiber bundle (yarn) and unidirectional GFRP
laminate samples. In the next section, the experimental procedure and test results are presented and
2.1. Testing
discussed. Then,Materials
comparison of the obtained values of tensile strength of GFRP specimens at different
strain rates with
The theunidirectional
glass values foundfiber
in literature is displayed.
fabric made by NanjingInHitech
the fourth section,Co.,
Composites theLtd.
results of Weibull
(Nanjing,
China) was used in this study, which had been fabricated
statistical analysis based on experimental results are discussed. with individual yarns (5.3 yarns/cm), as
shown in Figure 1a. The individual yarns in the fabric consist of thousands of filaments as shown in
Figure 1b. TheProgram
2. Experimental cross-sectional area of single glass yarn were calculated as 0.473 mm2 by taking into
account the linear density of the material and dividing it by its bulk density, following the same
2.1. Testing
methodMaterials
in the reference [21]. Epoxy resin provided by Hunan Good-Bond Construction Technic
Development Co., Ltd. (Changsha, China) was utilized as the matrix. The physical and mechanical
The glass unidirectional fiber fabric made by Nanjing Hitech Composites Co., Ltd. (Nanjing,
properties of GFRP components, i.e., the glass yarn and epoxy resin, were listed in Table 1.
China) was used in this study, which had been fabricated with individual yarns (5.3 yarns/cm), as
shown in Table
Figure 1. 1a. The and
Physical individual
Mechanicalyarns in theoffabric
Properties GFRP consist
componentsof thousands of filaments
(The mechanical propertiesasofshown in
Figure 1b.glass
Theyarn
cross-sectional
are obtained under areaquasi-static
of singleloading
glass yarn were
condition andcalculated
gauge lengthasof0.473 mm
specimen is225by taking into
mm.
account theThose of epoxy
linear resinofare
density theprovided by and
material manufacturer.).
dividing it by its bulk density, following the same method
in the reference [21]. Epoxy
Tensile resin provided
strength, by HunanElongation,
Youngs modulus, Good-Bond Construction
Density, Technic
C/S Development
Area of single yarn,
Components
Co., Ltd. (Changsha, China) MPa was utilized GPaas the matrix.%The physical g/cmand
3
mechanical mm2properties of
Glass yarn 919 113 1.5 2.54 0.473
GFRPEpoxy
components,
resin i.e., 36
the glass yarn and6.1epoxy resin, were 1.8 listed in1.7
Table 1.

(a) (b)
Figure 1. (a) Woven structure and (b) microscopy image of unidirectional glass fiber fabric.
Figure 1. (a) Woven structure and (b) microscopy image of unidirectional glass fiber fabric.
2.2. Specimen Fabrication and Test Procedures
Table 1. Physical and Mechanical Properties of GFRP components (The mechanical properties of
glass Individual yarn was
yarn are obtained extracted
under carefully
quasi-static fromcondition
loading the glassand
fabric for single
gauge length yarn tests. Inisorder
of specimen to
25 mm.
reduce stress concentration and to improve the
Those of epoxy resin are provided by manufacturer.). load transfer mechanism between the specimens and
steel wedges, thin aluminum sheets (10 mm 5 mm 0.5 mm) were roughhewed by toothed steel
plate and glued at both ends of the single yarn specimens using the same epoxy. When the epoxy
Tensile Strength, Youngs Modulus, Elongation, Density, C/S Area of Single
Components
was fully cured, the final
MPaglass yarn specimen GPawas constructed%by cutting g/cm off the
3 redundantYarn,length
mm2 at
both sides using an electric scissor as shown in Figure 2a.
Glass yarn 919 113 1.5 2.54 0.473
Epoxy resin 36 6.1 1.8 1.7

2.2. Specimen Fabrication and Test Procedures

Individual yarn was extracted carefully from the glass fabric for single yarn tests. In order to
reduce stress concentration and to improve the load transfer mechanism between the specimens and
steel wedges, thin aluminum sheets (10 mm 5 mm 0.5 mm) were roughhewed by toothed steel
plate and glued at both ends of the single yarn specimens using the same epoxy. When the epoxy was
fully cured, the final glass yarn specimen was constructed by cutting off the redundant length at both
sides using an electric scissor as shown in Figure 2a.
Polymers 2016, 8, 196 4 of 16
Polymers 2016, 8, 196 4 of 16
Polymers 2016, 8, 196 4 of 16

(a) (b)
(a) (b)
Figure
Figure 2.
2. Prepared
Prepared samples
samples of
of (a)
(a) glass yarn and
glass yarn and (b)
(b) GFRP.
GFRP.
Figure 2. Prepared samples of (a) glass yarn and (b) GFRP.
GFRP were manufactured by Vacuum Assisted Resin Transfer Molding (VARTM) [22], thus
GFRP were manufactured bybyVacuum Assisted Resin Transfer Molding (VARTM) [22],[22],
thus thus
thin
thin GFRP
laminates werecomposed
manufactured of a single Vacuum
ply of Assisted
glass fabric Resin with Transfer
epoxy Moldingresin were (VARTM)
fabricated, with a
laminates
thin composed
laminates composed of a single ply of glass offabric with epoxy withresin were fabricated, with a thickness
thickness of nearly 0.52 of mm, a single
after ply curing glass
24 hfabricat room epoxy
temperature. resin were
Duefabricated, with a
to the laminate
of nearly
thickness 0.52 mm,
of nearly after curing
0.52 mm, 24 h at room
afterbecuring temperature.
24 h attoroom Due to the
temperature. laminate heterogeneity,
Due to representative
the laminate the
heterogeneity, the specimens should large enough have mechanical properties
specimens
heterogeneity, should specimens
be large enough to
behave mechanical properties representative of the material.
of the material.the Considering the should
limitation large enough
of grip, to have
specimens mechanical
were cut along properties
the fiber representative
direction from
Considering
of the limitation of grip, specimens were cut along the fiber direction from the laminate from
with
thethe material.
laminate withConsidering
a dimension theoflimitation
105 mm of 22grip,
mm specimens
(L W). Two were cut alongsheets
aluminium the fiber of 40direction
mm long, 22
a dimension
the of 105 mm 22 mm (L W). Two aluminium sheets of 40 mm long, 22 mm wide and
mmlaminate
wide and with
0.3 amm dimension
thick were of 105
glued mm on the
22 mmend (L of specimens
W). Two aluminium
to avoid stress sheets of 40 mm long,
concentration. When 22
0.3
mm mmwide thick
and were
0.3 glued
mm thickon were
the end of specimens
glued on the end toofavoid stress to
specimens concentration.
avoid stress When the epoxy
concentration. was
When
the epoxy was fully cured, the typical test specimens were shown in Figure 2b. The gauge lengths of
fully
the cured,
epoxy thefully
was typical test specimens
cured, were shown inwere Figure 2b. The gauge2b. lengths of specimen are
specimen are 25.0 mm, andthethere typical
are 8test
glass specimens
fiber yarns in theshown
width inwith
Figure a volume Thefraction
gauge lengths
of 34.3%. of
25.0 mm,
specimen and
are there
25.0 mm, are 8
andglass
there fiberareyarns
8 glass in the
fiber width
yarns with
in thea volume
width fraction
with a of
volume 34.3%.
fraction of 34.3%.
Dynamic tensile tests of glass yarn under various strain rates and temperatures were conducted
Dynamic tensiletensile tests
tests ofof glass
glass yarn
yarn under various various strain
strain rates
rates andand temperatures
temperatures were were conducted
on a Dynamic
state-of-the-art drop-weight impactunder system (Instron, CEAST9340, Shanghai, China) asconducted shown in
on
on a
a state-of-the-art
state-of-the-art drop-weight
drop-weight impact
impact system
system (Instron,
(Instron, CEAST9340,
CEAST9340, Shanghai,
Shanghai, China)
China) as
as shown
shown in
in
Figure 3a. The drop height ranges from 0.03 to 1.10 m with corresponding impact velocity ranging
Figure
Figure 3a.
3a. The drop
The drop height
height ranges from
ranges from 0.030.03 to
to 1.10
1.10 m
m with
with corresponding
corresponding impact
impact velocity
velocity ranging
ranging
from 0.77 to 4.65 m/s. The maximum load application is 90 kN with a maximum potential energy of
from
from 0.77 toto 4.65
4.65 m/s.
m/s.The The maximumload load application is
9090 kNkN with a maximum potential energy
405 J. 0.77
The required impact maximum
velocities can beapplication achieved by is releasing with theaimpactor
maximum withpotential
a weight energy
of 52 ofN
of
405405
J. J.
The The required
required impact
impact velocities
velocities can can
be be achieved
achieved by by releasing
releasing the the impactor
impactor with with
a a
weight weight
of 52 of
N
from predetermined heights. The impact force induced by the free fall weight was measured by
52
fromN from predetermined
predetermined heights.heights. The impact
The aimpact force force
inducedinducedbythe by free
the the freefall fall
weightweight was was measured
piezoelectric force transducer with capacity of 2.2 kN, and force, deformation and measured
energy versus by
by piezoelectric
piezoelectric force transducer
force transducer with
with a by a
capacitycapacity of 2.2 kN, and the force, deformation and energy
time were recorded simultaneously a highofspeed 2.2 kN, dataandacquisition
the force, deformation and energy
system at a sampling versus
rate of 1
versus
time timerecorded
were were recorded simultaneously
simultaneously by a by a speed
high high speed
data data acquisition
acquisition system system
at a at a sampling
sampling rate rate
of 1
MHz with 14-bit resolution. In addition, the CEAST Visual Impact software (INSTRON (China) Co.,
of
MHz1 MHzwithwith14-bit 14-bit resolution.
resolution. In addition,
In addition, the CEAST
themanagement
CEAST Visual Visual
Impact Impact software (INSTRON (China)
Ltd., Shanghai, China) is used for system and data software
processing (INSTRON
(including (China) Co.,
the strain
Co.,
Ltd., Ltd., Shanghai,
Shanghai, China) is used for system management and data processing (including the strain
calculation). AnChina) is used for
environmental systemcan
chamber management
be heated and data processing
by electric resistance(including
wire and cooled the strain by
calculation).
calculation). An
An environmental
environmental chamber
chamber can can be be
heated
heatedby electric
by resistance
electric wire and
resistance wire cooled
and by liquid
cooled by
liquid nitrogen, giving operating temperatures ranging from 50 to 100 C. In this work, four
nitrogen,
liquid giving giving
nitrogen, operating temperatures
operating temperaturesranging from 50 to 100 C. In
100this work, four different
different impact velocities (1, 2, 3 and 4 m/s) wereranging chosen to from
obtain 50the tostrain C.
rates Inofthis40, work,
80, 120 four
and
impact
different velocities
impact (1, 2,
velocities3 and 4
(1, m/s)
2, 3 were
and 4 chosen
m/s) wereto obtain
chosen the
to strain
obtain rates
the of 40,
strain 80,
rates 120of and
40, 160120
80, s1andfor
160 s for specimens at room temperature
1

(25 C). In addition, four different temperatures, i.e., 25,
specimens
160 s1and at room temperature
for 100
specimens (25 C). In addition, Infour different temperatures, i.e., 25, 50, i.e.,
75 and
50, 75 C wereatselectedroom temperature
for the study(25 ofC).
temperatureaddition, effect
C were selected for the study of temperature effect under an impact velocity of 1 m/s (strain
four different
under temperatures,
an impact velocity of 1 m/s 25,
100
50, 75 and
(strain rate100
= 40Cs1were
). Tenselected
samplesfor the study
were tested of fortemperature
each strain rate effectand under an impact velocity of 1 m/s
temperature.
rate = 40
(strain s =1 ).40Ten
rate s1).samples
Ten samples were tested for each
were tested forstrain rate and
each strain ratetemperature.
and temperature.

(a) (b)
(a) (b)
Figure 3. Experimental setup: (a) Instron drop-weight impact system; (b) Mechanical Testing &
Figure 3. Experimental
Simulation
Figure 3. setup:
setup: (a)
(MTS) servo-hydraulic
Experimental Instron
(a)high drop-weight
speed
Instron impact
impact system;
testing machine.
drop-weight system; (b)
(b) Mechanical
Mechanical Testing
Testing &
&
Simulation (MTS) servo-hydraulic high speed testing machine.
Simulation (MTS) servo-hydraulic high speed testing machine.
In addition, dynamic tensile tests of GFRP specimens under various strain rates and
In addition,
temperatures weredynamic tensile
performed usingtests of GFRP
an MTS specimens
high rate under various
servo-hydraulic testing strain rates
machine withand
an
temperatures were performed using an MTS high rate servo-hydraulic testing machine with an
Polymers 2016, 8, 196 5 of 16

In addition, dynamic tensile tests of GFRP specimens under various strain rates and temperatures
were performed using an MTS high rate servo-hydraulic testing machine with an environmental
chamber in Arizona State University (as shown in Figure 3b) due to the limited force measurement
capacity of the drop-weight impact system. The speed of the stroke is controlled by the opening and
closing of the servo-valve of hydraulic supply. By manually turning the servo-valve, the rate of flow of
hydraulic fluid can be controlled, thus a desired stroke speed can be obtained. In our test, the valve
was opened and the stroke accelerated until it reached a constant predetermined velocity, then the
test specimens were mounted between upper and lower grips. These grips have serrated surfaces to
effectively clamp specimens and prevent any slippage during the loading process of tests. According
to the calibration records, the stroke can reach a maximum speed of 14 m/s with a load capacity of
25 kN [23]. Operating temperatures ranging from 60 to 200 C were achieved by liquid nitrogen
and electric resistance wire of the environmental chamber, with a built-in fan to ensure the uniform
distribution of temperature in the chamber. At room temperature (25 C), dynamic tensile tests were
conducted at four loading velocities of 0.625, 1.25, 2.5 and 5 m/s. The initial strain rates are defined
as the ratio of the applied velocity to the specimens gauge length, namely 25, 50, 100 and 200 s1 ,
respectively. At the initial strain rate of 25 s1 , six different temperatures (25, 0, 25, 50, 75 and 100 C)
were selected for temperature effect testing. The automatically controlled environmental chamber was
set to the required temperature and maintained for 30 min to guarantee that the test specimens reached
a uniform temperature. Ten replicates were tested for each condition to reduce the influence of random
error. A high-speed digital acquisition card (TRIO Test & Measurement, Adelaide, Australia) was
used to collect force and displacement data at a sampling rate of 500 kHz. A Phantom v7.3 high speed
digital camera (Vision Research Phantom, Wayne, NJ, USA) was used to capture the deformation and
failure behavior of GFRP specimens with a sampling rate of 20,000 fps.
Quasi-static testing was performed on an MTS load frame (C43.304) (MTS Systems (China) Co.,
Ltd., Shanghai, China) [24] with a load capacity of 30 kN and maximum sampling rate of 1000 Hz.
In this work, two load cells with 1 and 30 kN capacities were used for force measurement of glass
yarn and GFRP, respectively, with a sampling rate of 20 Hz, and the loading velocity was set to
be 2.5 mm/min. The corresponding strain rate is 1/600 s1 by dividing the velocity with gauge
length. An extensometer was used for displacement measurement of GFRP, while, for glass yarn, the
displacement recorded by MTS load frame is very close to that measured by the extensometer with
an error less than 2%, so it is reasonable to use the crosshead measurement as the deformation of glass
yarn for strain calculation.

3. Results and Discussion

3.1. Strain Rate Effect on Tensile Properties

Figure 4a show the stress vs. strain curves of glass yarn samples at different strain rates.
The high-frequency oscillations in the stress curves obtained under dynamic loadings were caused by
the system ringing phenomenon [25,26]. In addition, the number of oscillations in the curves decreases
with increasing impact velocities, since the impact duration is much shorter at higher loading rates.
Figure 4b shows the stress vs. strain curves of GFRP samples at different strain rates. Note that the
actual strain rates during dynamic testing was remarkably different from the initial strain rates, which
is attributed to the fact that the reaction force of the test specimen slowed down the stroke, especially at
higher loading speeds when the hydraulic pressures were relative lower. In the subsequent discussion
of this paper, the initial strain rates will be used to facilitate analysis. Meanwhile, it is also worth noting
that the curves were similar with the strainstress curve of mild steel because the stress fluctuated after
a small elastic deformation until the final failure at 25 and 50 s1 . This is because the yarns of GFRP
samples do not fail simultaneously during the loading, as shown in Figure 5. However, the fluctuation
in stress reduced at a higher strain rate, and it disappeared when the strain rate reached 200 s1 .
Polymers 2016, 8, 196 6 of 16
Polymers 2016, 8, 196 6 of 16
Polymers 2016, 8, 196 6 of 16

(a) (b)
(a) (b)
Figure
Figure 4. 4. Representativestressstrain
Representative stressstrainresponses
responses ofof (a)
(a) glass
glass yarn
yarn at
atinitial
initialstrain
strainrates
ratesofof1/600, 40,40,
1/600, 80,80,
Figure
120, 1604.
s Representative
1 and (b) GFRPstressstrain
at initial responses
strain rates of of (a)
1/600, glass
25, yarn
50, at
100, initial
200 s1. strain rates of 1/600, 40, 80,
120, 160 s11and (b) GFRP at initial strain rates of 1/600, 25, 50, 100, 2001 s1 .
120, 160 s and (b) GFRP at initial strain rates of 1/600, 25, 50, 100, 200 s .

(a)
(a)

(b)
(b)

(c)
(c)

(d)
(d)
Figure 5. Failure behaviors of GFRP at different strain rates: (a) 25 s1; (b) 50 s1; (c) 100 s1 and (d) 200 s1.
Figure 5. 5.Failure
Figure Failure behaviors
behaviors of
ofGFRP
GFRPatat
different strain rates: (a) 25(a)
s1;25
(b)s50
1s; 1(b)
; (c)50
100 s1; and (d) 200
s1s and
1.
different strain rates: s1 (c) 100
(d)The
200 seffects
1 . on mechanical properties of GFRP can be summarized based on the stressstrain
curvesThe effects
under on mechanical
different properties
strain rates. Figure 6of GFRPthe
shows can be summarized
dependence based ontensile
of the dynamic the stressstrain
properties
curves under different strain rates. Figure 6 shows the dependence of the dynamic
of glass yarn and GFRP, defined in terms of tensile strength and toughness, on the varying tensile properties
initial
of The
glasseffects on mechanical
yarnrespectively.
and GFRP, The properties
defined in terms of
of GFRP
tensile can be summarized
strength and toughness,based on the
on the stressstrain
varying initial
strain rates, toughness is calculated as the area under stress-strain curves, given by:
curves
strainunder
rates,different strain
respectively. rates.
The Figure is6 calculated
toughness shows the as dependence of the
the area under dynamic tensile
stress-strain curves,properties
given by: of
f
glass yarn and GFRP, defined in terms of tensile strength
Energy and toughness, on the varying initial strain

f
U
rates, respectively. The toughness is calculated = Energy = d stress-strain curves, given by: (1)
as the area under

T
UT = Volume = 0 d (1)
Volume 0 f
where the term UT represents toughness; Energy is strain; f is the maximum strain and

is stress.
where the term UT represents toughness;
There is an apparent dependence on the dynamic
U T is strain; f is

Volume
the
d maximum strain and
tensile properties of glass yarn and GFRP on the
is stress.(1)
There
strain israte.
an apparent dependence
For the glass on the dynamic
yarn, tensile strength tensile
0 properties of glass yarn and GFRP on the
and toughness increase significantly during a
strain rate.
transition For
from the glass yarn, tensile strength and toughness increase s1significantly during a
UTquasi-static
represents loading (1/600
is sstrain;
) to dynamic loading (40strain ). and
In addition, tensile
1
where the term toughness; f is the maximum is stress. There
transition
strength from
maintainsquasi-static
growthon loading
with (1/600 s1) to dynamic loading (40 s1). In addition, tensile
is an apparent dependence the increasing strain rate,
dynamic tensile while toughness
properties abruptly
of glass yarn decreases
and GFRP on the at strain
the
strength
strain ratemaintains
of 160 s growth
1 due towith
the increasing strain
diminution of rate,Specifically,
strain. while toughness abruptly
tensile strength decreases
and at the
toughness
rate. For the glass yarn, tensile strength and toughness increase significantly during a transition from
strain rate of 160 s1 due to the diminution of strain. Specifically, tensile strength and toughness
Polymers 2016, 8, 196 7 of 16

quasi-static loading
Polymers 2016, 8, 196 (1/600 s1 ) to dynamic loading (40 s1 ). In addition, tensile strength maintains 7 of 16
growth with increasing strain rate, while toughness abruptly decreases at the strain rate of 160 s1
due to theasdiminution
increase much as 88.0% of strain. Specifically,
and 474.3% fromtensile strength
919 102 MPa and
andtoughness increase
7.0 1.1 MPa as much
to 1729 as 88.0%
67 MPa and
and
40.2 474.3% fromrespectively,
4.5 MPa, 919 102 MPa whenand the
7.0 1.1 MPa
strain rate to 1729 67
increases MPa1/600
from and 40.2
to 40s4.5 MPa, respectively,
1. After that, tensile
when thegoes
strain 1 . After that, tensile strength goes on to 1
40 sthe
strength onrate increasesnearly
to increase from 1/600
22.0% to when strain rate increases from 40 to 160 s increase , while
nearly 22.0% when the strain26.4%
rate increases from 40 160 sof
to range 1 , while toughness increases about 26.4%
toughness increases about over a strain rate 40120 s1 and then nearly drops by
over a strain rate range of 40120 s 1 and then nearly drops by 10.2% between 120 and 160 s1 . The
10.2% between 120 and 160 s . The initial significant enhancement in strength and toughness can be
1

initial
explainedsignificant enhancement
as: the friction between in adjacent
strength fibers
and toughness
always plays can an be important
explained as:rolethe
duefriction
to randombetween
and
adjacent
misaligned fibers always of
breakage plays an important
filaments in the role
yarndue to random
under tensionand andmisaligned
the contactbreakage
force amongof filaments
adjacent in
the yarn under tension and the contact force among adjacent filaments increases
filaments increases with increasing strain rate, resulting in the increase of dynamic frictional forcewith increasing strain
rate, resulting
(friction in theisincrease
coefficient assumed of dynamic frictional
to be constant). force (friction
Therefore, tensile coefficient
strengthisandassumed to be constant).
energy-consumption
Therefore,
are improved tensile strength and energy-consumption are improved noticeably.
noticeably.

(a)

(b)
Figure 6. Strain
Figure 6. Strain rate
rate effect
effect on tensile properties:
on tensile properties: (a)
(a) tensile
tensile strength and (b)
strength and (b) toughness.
toughness.

For GFRP, the tensile strength linearly increases over the strain rate range of 1/600200 s1, while
For GFRP, the tensile strength linearly increases over the strain rate range of 1/600200 s1 , while
the variation of its toughness is similar to the tensile strength of glass yarn. Concretely speaking,
the variation of its toughness is similar to the tensile strength of glass yarn. Concretely speaking, tensile
tensile strength increases nearly 49.1% when the strain rate increases from 1/600 to 200 s1. By
strength increases nearly 49.1% when the strain rate increases from 1/600 to 200 s1 .1 By contrast,
contrast, toughness increases remarkably from 15.5 3.7 MP a at a strain rate of 1/600 s to 32.5 7.8
toughness increases remarkably from 15.5 3.7 MP a at a strain rate of 1/600 s1 to 32.5 7.8 MPa at
MPa1at 25s1, enhanced by 109.7%, and then increases about 34.2% over the strain rate range of 25
25 s 1, enhanced by 109.7%, and then increases about 34.2% over the strain rate range of 25200 s1 .
200 s . Those kinds of strengthening and toughening mechanisms are different from glass yarn and
Those kinds of strengthening and toughening mechanisms are different from glass yarn and could be
could be described as: (i) there is not enough time to initiate internal defects in the material at higher
described as: (i) there is not enough time to initiate internal defects in the material at higher strain rates.
strain rates. Therefore, more energy is needed for damage initiation and propagation, which leads to
Therefore, more energy is needed for damage initiation and propagation, which leads to a higher level
a higher level of tensile strength under high strain rates, and (ii) more damage is involved with
of tensile strength under high strain rates, and (ii) more damage is involved with increasing strain rate.
increasing strain rate.
In order to further study the tensile behavior and failure pattern of GFRP, digital image correlation
In order to further study the tensile behavior and failure pattern of GFRP, digital image
(DIC) analysis was performed on the specimens tested under room temperature. The stressstrain
correlation (DIC) analysis was1performed on the specimens tested under room temperature. The
curve of a GFRP tested at 25 s together with the longitudinal strain (yy ) distributions measured
stressstrain curve of a GFRP tested at 25 s1 together with the longitudinal strain (yy) distributions
by DIC are shown in Figure 7. A color code with purple representing the lowest strain values and
measured by DIC are shown in Figure 7. A color code with purple representing the lowest strain
red at 3.0% strain is used, denoting the level of strains. The time and stage of stressstrain behavior
values and red at 3.0% strain is used, denoting the level of strains. The time and stage of stressstrain
associated with each step of the strain distributions are indicated. The damage evolution shows that,
behavior associated with each step of the strain distributions are indicated. The damage evolution
shows that, during the linear elastic stage of the test (t = 0.3 ms), a relatively uniform strain
distribution is obtained. As the load increases (t = 2.0 ms), tensile strain starts to localize near the top
and bottom edges of area of interest (AOI), which is represented by the red zone. This may imply the
Polymers 2016, 8, 196 8 of 16

during the linear elastic stage of the test (t = 0.3 ms), a relatively uniform strain distribution is obtained.
As the load increases (t = 2.0 ms), tensile strain starts to localize near the top and bottom edges of
area of interest (AOI), which is represented by the red zone. This may imply the stress concentration
Polymers 2016, 8, 196 8 of 16
due to the clamps and associate with the non-linear behavior observed in the stress-strain curve.
The inhomogeneity
stress concentrationof thedue
strain fields
to the clampsmayandbe attributed
associate with to
themultiple
non-linearaspects
behavior such as theinpresence
observed the of
stress-strain
random Polymers
defects, curve.
variability
2016, 8, 196 The inhomogeneity of the strain fields may be attributed to
in the stiffness of basalt yarns and potential eccentric loads. The multiple aspects such
8 of 16localized
as the presence
zones continue to grow of random defects, variability
with increasing load (t =in4.0 thems)
stiffness of basalt yarns
and ultimately and potential
spread out overeccentric
the entire AOI
stress The
loads. concentration due tocontinue
localized zones the clamps and associate
to grow with the
with increasing non-linear
load behavior
(t = 4.0 ms) observed in
and ultimately the
spread
prior to stress-strain
failure (t =curve.5.0 ms). Non-uniform
The prior
inhomogeneity
strain fields are alsoattributed
observed in the specimens tested at
out over the entire AOI to failure (tof=the
5.0 strain fields may bestrain
ms). Non-uniform fieldstoare
multiple aspects such
also observed in the
other strain rates,
as the presence
specimens as shown
testedofatrandom in Figure
defects,
other strain 8.
variability
rates, as shownininthe stiffness
Figure 8. of basalt yarns and potential eccentric
loads. The localized zones continue to grow with increasing load (t = 4.0 ms) and ultimately spread
out over the entire AOI prior to failure (t = 5.0 ms). Non-uniform strain fields are also observed in the
specimens tested at other strain rates, as shown in Figure 8.

Figure 7. 2D contour of yy and stressstrain response for a GFRP specimen tested at the strain rate of
Figure 7. 2D contour of yy and stressstrain response for a GFRP specimen tested at the strain rate of
25 s1 obtained by digital image correlation (DIC).
1
25 s obtained by digital image correlation (DIC).
Figure 7.
Figure 9 2D contour of
compares yyfailure
the and stressstrain response
morphologies offor
thea GFRP specimen
composite tested at the
specimens strain
after rate of under
tension
25
differents 1 obtained by digital image correlation (DIC).
strain rates,the
which clearly shows a remarkable
Figure 9 compares failure morphologies of the difference
composite in the fracture areas.
specimens after Ittension
seems under
from the fracture surface of relative low strain rate (1/600 s1), the damage area was limited to a small
different strain
Figurerates, which clearly
9 compares shows
the failure a remarkable
morphologies of thedifference in the fracture
composite specimens areas. under
after tension It seems from
region; however, under higher strain rate (200 s1), the damage path covers the entire gauge section,
different
the fracture strain
surface rates, which clearly shows a remarkable 1
difference in the fracture areas. It seems
indicating moreof relative
materials of low strain rate
the specimen (1/600 sin energy
are involved ), thedissipation.
damage area was limitedwas
This phenomenon to a small
from the fracture surface of relative low strain rate
region; also
however,
observed under higher[17]
by Shokrieh strain
in a rate s11 ),(1/600
(200experiment.
relevant
s1), the damage area was limited to a small
the damage path covers the entire gauge section,
region; however, under higher strain rate (200 s ), the damage path covers the entire gauge section,
indicating more materials
indicating of the
more materials specimen
of the specimenareare involved
involved inin energy
energy dissipation.
dissipation. This phenomenon
This phenomenon was was
also observed by Shokrieh
also observed [17][17]
by Shokrieh in ainrelevant
a relevantexperiment.
experiment.

t = 0.0 ms t = 0.0 ms t = 0.0 ms

t = 0.0 ms t = 0.0 ms t = 0.0 ms

t = 0.2 ms t = 0.3 ms t = 0.1 ms

t = 0.2 ms t = 0.3 ms t = 0.1 ms

t = 1.4 ms t = 0.8 ms t = 0.4 ms

t = 1.4 ms t = 0.8 ms t = 0.4 ms

t = 1.8 ms t = 1.4 ms t = 0.8 ms

t = 1.8 ms t = 1.4 ms t = 0.8 ms

t = 3.0 ms t = 1.7 ms t = 0.9 ms

(a) (b) (c)
t = 3.0 ms t = 1.7 ms t = 0.9 ms
Figure 8. Normal strain distribution of GFRP adherent at different loading rates: (a) 50 s1; (b) 100 s1;
(a) (b) (c)
and (c) 200 s . 1

Figure 8. Normal strain distribution of GFRP adherent at different loading rates: (a) 50 s1; (b) 100 s1;
Figure 8. Normal strain distribution of GFRP adherent at different loading rates: (a) 50 s1 ; (b) 100 s1 ;
and (c) 200 s1.
and (c) 200 s1 .
Polymers 2016, 8, 196 9 of 16
Polymers 2016, 8, 196 9 of 16
Polymers 2016, 8, 196 9 of 16

Figure 9. Fracture morphologies of GFRP under different strain rates.

Figure
Figure 9.
9. Fracture
Fracture morphologies
morphologies of
of GFRP
GFRP under
under different
different strain rates.
strain rates.
3.2. Temperature Effect on Tensile Properties
3.2. Temperature Effect on Tensile Properties
3.2. Temperature Effect on Tensile Properties
Figure 10 shows the experimental stress-strain responses of glass yarn and GFRP under
Figure 10 shows the experimental stress-strain responses of glass yarn and GFRP under
Figure
different 10 shows thewith
temperatures experimental
the samestress-strain
strain rateresponses
of 25 s1.ofThe glass yarnof
effect and GFRP underondifferent
temperature tensile
different temperatures with the same strain rate
1 of 25 s1. The effect of temperature on tensile
temperatures
strength and with the same
toughness strain of
in terms rate of 25 s values
average . The and
effectstandard
of temperature
deviationson tensile strength and
are summarized in
strength and toughness in terms of average values and standard deviations are summarized in
toughness in terms of average values and standard deviations are summarized
Figure 11. For glass yarn, tensile strength and toughness decrease nearly 25.3% and 31.1% when in Figure 11. For glass
Figure 11. For glass yarn, tensile strength and toughness decrease nearly 25.3% and 31.1% when
yarn, tensile strength
temperature increasesand fromtoughness
25 to 75 decrease
C. This nearly
is because25.3%theand 31.1%process
fracture when temperature
of fibers is increases
generally
temperature increases
C. This isfrom 25 to 75 C. This is because the fracture process of fibers is generally
from
accompanied by the damage of chemical bond, intermolecular slip and the weakening of damage
25 to 75 because the fracture process of fibers is generally accompanied by the van der
accompanied by the damage of chemical bond, intermolecular slip and the weakening of van der
of chemical
Waals force,bond, intermolecular
etc. Heating slip and the
will undoubtedly weakening
weaken of vanbond
the chemical der Waals
and van deretc.
force, Heating
Waals force will
and
Waals force, etc. Heating will undoubtedly weaken the chemical bond and van der Waals force and
undoubtedly
then lower the weaken
tensilethe chemical
strength bond and
of fibers. van der Waals
Furthermore, force and
maximum thenover
strain lower the tensile strength
temperature range of
then lower the tensile strength of fibers. Furthermore, maximum strain over temperature range of
of fibers.
2575 C Furthermore,
has no obvious maximum
change,strain over temperature
so toughness range of
also decreases. 2575 Cwhen
However, has noheated
obviousto change,
100 C,
2575 C has no obvious change, so toughness also decreases. However, when heated to 100 C,
so toughness
tensile strengthalsoofdecreases.
glass yarnHowever,
reboundswhen 19.2%heated to 100of value
at the base C, tensile
at 75strength
C, mainlyof glass
due yarn
to therebounds
fact that
tensile strength of glass yarn rebounds 19.2% at the base of value at 75 C, mainly due to the fact that
C, mainly
19.2% at the base
the frictional of value
coefficient of at
the75fiber due to thebecause
surface increases fact that of the frictional coefficient
volatilization of sizing and of the fiber
water at
the frictional coefficient of the fiber surface increases because of volatilization of sizing and water at
C, resulting
surface increases because of volatilization of sizing and water at 100
100 C, resulting in the increase of frictional force between fibers. At the same time, toughness gets a in the increase of
100 C, resulting in the increase of frictional force between fibers. At the same time, toughness gets a
frictional force between
sharp increase fibers. At
(about 44.5%), the same
which couldtime,
be toughness
attributed getsto thea sharp increase
obvious (aboutin44.5%),
increases which
both tensile
sharp increase (about 44.5%), which could be attributed to the obvious increases in both tensile
could be attributed to
strength and maximum strain.the obvious increases in both tensile strength and maximum strain.
strength and maximum strain.

(a) (b)
(a) (b)
Figure 10.
10. Representativestressstrain
stressstrain responsesofof(a)(a)glass
glass yarn at temperatures of 25, 50, 75, 100
C
Figure
Figure 10. Representative
Representative stressstrainresponses
responses of (a) glassyarn atat
yarn temperatures of of
temperatures 25, 50,50,
25, 75, 75,
100100
C and
andand (b)
(b) (b) GFRP
GFRP at temperatures
at temperatures of 25,
of 25, 0,
0, 25,25, 50,
50,50, 75,
75,75, 100
C.
100100C.C.
C GFRP at temperatures of 25, 0, 25,
Polymers 2016, 8, 196 10 of 16
Polymers 2016, 8, 196 10 of 16
Polymers 2016, 8, 196 10 of 16

(a)
(a)

(b)
Figure (b)
Figure 11.11. Temperatureeffect
Temperature effecton
ontensile
tensile properties:
properties: (a)
(a)tensile
tensilestrength
strengthand (b)(b)
and toughness.
toughness.
Figure 11. Temperature effect on tensile properties: (a) tensile strength and (b) toughness.
For GFRP, tensile strength shows almost no change (within 3%) when temperature increases
For 25
from GFRP, tensile
to 50 C. strength ashows
However, almost no change (within 3%) when temperature increases from
ForGFRP, tensile strengthsharp
shows decrease
almost(about 18.9%)
no change is observed
(within 3%) whenover the temperature
temperature range
increases
25ofto50100
50 C. However, a sharpbydecrease (about 18.9%) ismatrix
observedwhenover the temperature range of
from 25 toC, 50which is caused
C. However, the softening
a sharp of the
decrease (about resin
18.9%) is observed glass
overtransition temperature
the temperature range
50100 C,
of50100 which
epoxy resin is caused
(Tg) is by the softening of the resin matrix when glass transition temperature
of C, which is reached.
caused by This
thesoftening
softeningeffect
of thewould
resin weaken
matrix when the interfaces betweentemperature
glass transition fibers and
of of
epoxy
matrix,
epoxyresin
resin(Tg)
and (Tg)isis reached.
decrease Thissoftening
the stiffness
reached. This ofsoftening effect
epoxy effect
matrix would
during
would weaken
loading.
weaken the interfaces
the Figure
interfaces showsbetween
12 between the failure
fibers fibers
and
and matrix,
morphologiesand decrease
of GFRP the stiffness
specimens at of epoxy
different matrix during
temperatures. When loading.
the Figure
testing
matrix, and decrease the stiffness of epoxy matrix during loading. Figure 12 shows the failure 12 shows
temperature the
increases, failure
the components
morphologies
morphologies ofof of GFRP
GFRP
GFRP specimens
specimens
specimens are moretemperatures.
atatdifferent
different separated after
temperatures. failure
When
When due
the
the exactly
testing
testing to the decrease
temperature
temperature of
increases,
increases,
bonding
thethe
componentsstrength
components between
ofofGFRP fibers
GFRPspecimens and
specimensare matrix.
are more The change
more separated of toughness
after failure
separated after is
failuredue within
dueexactly 7.6% as
exactlytotothetemperature
thedecrease
decrease of of
increases
bonding
bonding from 25
strength
strength to 25 C,
between
between andand
fibers
fibers then decreases
andmatrix.
matrix. The
The aschange
much as of29.2%
of whenisthe
toughness
toughness temperature
iswithin
within 7.6%asas
7.6% is up to 100
temperature
temperature
C. Thisfrom
increases
increases noticeable
from2525totodecrease
25 C,
25 is caused
C, and
and then by the combined
thendecreases
decreases asas much
much behaviors
as 29.2%
as 29.2% (decrease)
when when of tensile
the strength
temperature
the temperature is up to isand
up to
100
maximum

100C.C.This
This strain.
noticeable decrease is caused by the combined behaviors (decrease) of tensile strength
noticeable decrease is caused by the combined behaviors (decrease) of tensile strength and
andmaximum
maximum strain.
strain.

Figure 12. Fracture morphologies of GFRP under different temperatures.

Figure
4. Comparison Figure
with 12. Fracture morphologies of GFRP under different temperatures.
Literature
12. Fracture morphologies of GFRP under different temperatures.
4.4.1.
Comparison with Literature
Strain Rate Data
4. Comparison with Literature
Figure
4.1. Strain Rate13aData
displays the comparison of the tensile strength of GFRP specimens at different strain
4.1.rates
Strain Rateliterature
with Data [9,16,17]. All of the works revealed that the tensile strength of GFRP increases
Figure 13a displays the comparison of the tensile strength of GFRP specimens at different strain
Figure 13a displays the comparison of the tensile strength of GFRP specimens at different strain
rates with literature [9,16,17]. All of the works revealed that the tensile strength of GFRP increases
rates with literature [9,16,17]. All of the works revealed that the tensile strength of GFRP increases
Polymers 2016, 8, 196 11 of 16

Polymers 2016, 8, 196 11 of 16

with increasing strain rate, especially over the intermediate strain rate range of 1200 s1 .
Assume
withtensile
that increasing strain
strength TSrate,
is anespecially over
exponential the intermediate
function of strain ratestrain
. rate
with therange of 1200
following form:s . Assume
1

that tensile strength TS is an exponential function of strain rate with the following form:
.
TS A epB(Bq) (2)
TS = A e (2)

Parameters A and B are then simply determined by a means of exponential fitting, and it proves
befitting
to be fittingwell.
well.Once
Once A and
A and B determined,
B are are determined, this model
this model can becan be extrapolated
extrapolated to the
to predict predict the
material
material behavior
behavior at any
at any strain strain
rates rates fitting
within withinrange.
fitting range. In addition,
In addition, tensiletensile strength
strength of GFRP
of GFRP singlesingle
yarn
yarn reported
reported in previous
in previous work
work [9] [9] is150%200%
is about about 150%200% higher
higher than than results
current currentusing
results
GFRPusing GFRP
samples
samples
with with
eight eight
yarns, yarns, indicating
indicating a remarkable
a remarkable size effect.
size effect.

4.2. Temperature Data

4.2. Temperature Data
Figure
Figure 13b13b shows
shows thethe comparison
comparison of of the
the tensile
tensile strength
strength ofof GFRP
GFRP specimens
specimens at at different
different
temperatures
temperatures withwith the
the literature
literature [6,9,20]. Note that
[6,9,20]. Note that the
the tensile
tensile strength
strength of of GFRP
GFRP suffers
suffers dramatic
dramatic
losses C, which is caused by the softening of the resin matrix
losses over
over the
the temperature
temperature range
range of
of 50100
50100 C, which is caused by the softening of the resin matrix
when
when its
its glass
glass transition
transition temperature
temperature (T (Tgg)) is
is reached. This would
reached. This would weaken
weaken thethe interfaces
interfaces between
between
fibers
fibers and
and matrices
matrices andand decrease
decrease the
the resistance
resistance of of matrices
matrices during
during deformation. Reference [20]
deformation. Reference [20] also
also
found
found that the test specimens experienced a significant drop in strength over a temperature range of
that the test specimens experienced a significant drop in strength over a temperature range of
300450 C. This is due to the thermal degradation of the polymer. In this case, the molecular chains of
300450 C. This is due to the thermal degradation of the polymer. In this case, the molecular chains
the polymer
of the polymer broke, leading
broke, leadingto the formation
to the formation of micro-cracks both
of micro-cracks at the
both fiber/matrix
at the interface
fiber/matrix and
interface in
and
the matrix phase [20].
in the matrix phase [20].

(a)

(b)
Figure 13.
Figure 13. Comparison
Comparisonofofthethetensile
tensilestrength
strengthofof
GFRP specimens
GFRP at different
specimens (a) (a)
at different strain rates
strain andand
rates (b)
temperatures with literatures.
(b) temperatures with literatures.

In addition, some researchers [27] proposed that the difference in contractions of fiber and
In addition, some researchers [27] proposed that the difference in contractions of fiber and matrix
matrix on cooling is also suspected to increase the residual stresses at the fiber/matrix interface, and
on cooling is also suspected to increase the residual stresses at the fiber/matrix interface, and then
then result in local micro-cracking and reduced tensile strength. Snchez-Sez et al. [7] confirmed
this when investigating the static behavior of CFRPs at low temperatures. However, reference [20]
Polymers 2016, 8, 196 12 of 16

Polymers 2016, 8, 196 12 of 16

result in local micro-cracking and reduced tensile strength. Snchez-Sez et al. [7] confirmed this when
investigating
found that the thetensile
static strength
behaviorof ofGFRP
CFRPsincreased
at low temperatures.
with decreasing However, reference
temperature [20] found
between 100 Cthatand
the
tensile strength ofanGFRP increased with decreasing temperature
between 100 C and 0 C, exhibiting
0 C, exhibiting outstanding capability of resistance to low temperature.
an outstanding capability of resistance to low temperature.
5. Weibull Analysis
5. Weibull Analysis
The determination of the statistics and probability distributions of the random variables
The determination
characterizing of the statistics
material properties plays anand probability
important role indistributions
the development of the random variables
of probabilistic-based
characterizing material properties plays an important role in the development
design specifications. The selection of the probability distribution assigned to characterize the of probabilistic-based
design
materialspecifications.
property data The selection
will have of the probability
a significant effect on the distribution assigned to
computed reliability. characterize
Assuming the
distinct
material property data will have a significant effect on the computed reliability. Assuming
distributions for the material properties may result in calculated probabilities of failure that vary by distinct
distributions
over an orderfor of the material This
magnitude. properties
can be may resultto
attributed inthe
calculated
lower tailprobabilities
behavior ofof failure that
different vary by
cumulative
over an orderfunctions,
distribution of magnitude.whichThis
hascan be attributed
become acquaintedto the aslower tail behavior of
the tail-sensitivity different
problem in cumulative
structural
distribution functions,
reliability [28]. whichdistribution
The Weibull has become acquainted
[29] is often usedas the tail-sensitivity
to describe problem
the strength in structural
of fibers [30,31]
reliability [28]. The Weibull distribution [29] is often used to describe the strength
and FRP composites [3234]. Typically, the two-parameter Weibull distribution with the basic form of fibers [30,31] and
FRP compositesprobability
for cumulative [3234]. Typically,
density isthe two-parameter Weibull distribution with the basic form for
cumulative probability density is
( )qmm
p
P () = 1 e
Ppq 1 e 00
(3)(3)

where
where is the tensile
is the strength,
tensile 0 is the
strength, is reference or scaling
the reference value
or scaling related
value to the
related mean,
to the m is
andand
mean, mthe
is
0
Weibull modulus or shape parameter. The cumulative probability density, P, is estimated
the Weibull modulus or shape parameter. The cumulative probability density, P, is estimated as: as:

ii
PP = , (4)(4)
N `+ 11
N
,
whereNNisisthe
where thetotal
totalnumber
numberof oftests and ii is
testsand is the
the current
current test
test number.
number.
Figure
Figure 14 shows the Weibull curve fitting to tensile strengthsofofglass
14 shows the Weibull curve fitting to tensile strengths glassyarn
yarnand
andGFRP
GFRPobtained
obtainedin
this work. Figure 14a,c reveal that the cumulative probability plot shifts towards
in this work. Figure 14a,c reveal that the cumulative probability plot shifts towards higher values higher values with
increasing strainstrain
with increasing rate, which is a clear
rate, which indicator
is a clear of dependence
indicator of tensile
of dependence strength
of tensile of glass
strength yarnyarn
of glass and
GFRP
and GFRP on the strain rate. Note that, as shown in Figure 14b, the change of temperatures from 25to
on the strain rate. Note that, as shown in Figure 14b, the change of temperatures from 25
75 result in obvious left-shift of the cumulative probability plot, and then the plot shifts towards
to C75 C result in obvious left-shift of the cumulative probability plot, and then the plot shifts
the higherthe
towards stress region
higher when
stress the when
region temperature increases from
the temperature increases 100 75
75 tofrom C, to
in 100
line C,
withinFigure 11a.
line with
Additionally, as displayed in Figure 14d, no distinct shift of the cumulative probability
Figure 11a. Additionally, as displayed in Figure 14d, no distinct shift of the cumulative probability plot is observed
among
plot isthe temperature
observed among range 2550 C. After
theoftemperature rangethat,
of however,
2550 C. theAfter
plot remarkably shiftsthe
that, however, towards
plot
the lower stress
remarkably region,
shifts in accordance
towards with Figure
the lower stress region,11a as well. with Figure 11a as well.
in accordance

(a) (b)

Figure 14. Cont.

Polymers 2016, 8, 196 13 of 16
Polymers 2016, 8, 196 13 of 16

(c) (d)
Figure 14.
Figure 14. Cumulative
Cumulative failure
failure probability
probabilityversus
versustensile
tensilestrength
strengthofof(a,b)
(a,b)glass
glassyarn
yarnand
and(c,d)
(c,d)GFRP
GFRP
under different strain rates and temperatures.
under different strain rates and temperatures.

The Weibull parameters extracted from tensile strength data of glass yarn and GFRP samples
The Weibull parameters extracted from tensile strength data of glass yarn and GFRP samples are
are presented in Tables 25, respectively. The scale parameter 0 generally increases with
presented in Tables 25 respectively. The scale parameter 0 generally increases with increasing strain
increasing
rates, whichstrain rates,
follows thewhich
same follows the same
change trends change strength
of tensile trends ofattensile strength
the same at conditions.
testing the same testing
Shape
conditions. Shape parameter m is related to the variation of strength among the tested specimens.
parameter m is related to the variation of strength among the tested specimens. For glass yarn, in
For glass yarn, in general, the results obtained from higher strain rate have a lower shape parameter
general, the results obtained from higher strain rate have a lower shape parameter than the counterpart.
than the counterpart. This manifests that yarn subjected to higher strain rate is usually characterized
This manifests that yarn subjected to higher strain rate is usually characterized by a more random
by a more random breakage process.
breakage process.
Table 2. Weibull parameters for tensile strength of glass yarn under different strain rates.
Table 2. Weibull parameters for tensile strength of glass yarn under different strain rates.
Strain rate (s1) 1/600 40 80 120 160
1
N (s ) 10
Strain Rate 1/600 8 40 8 80 8 120 8 160
0 N
(MPa) 93210 1,7448 1,8368 2,0278 2,1778
0 (MPa) 932 1744 1836 2027 2177
m 26.1 20.2 16.2 15.4 10.9
m 26.1 20.2 16.2 15.4 10.9

Table 3. Weibull parameters for tensile strength of glass yarn under different temperatures.
Table 3. Weibull parameters for tensile strength of glass yarn under different temperatures.
Temperature (C) 25 50 75 100
N
Temperature ( C) 825 850 875 8100
0 N
(MPa) 1,744
8 1,479
8 1,312
8 1,497
8
0 (MPa) 1744 1479 1312 1,497
m 20.2 10.3 9.3 8.1
m 20.2 10.3 9.3 8.1

Table 4. Weibull parameters for tensile strength of GFRP under different strain rates.
Table 4. Weibull parameters for tensile strength of GFRP under different strain rates.
Strain rate (s1) 1/600 25 50 100 200
N 1
Strain Rate (s ) 10
1/600 10
25 9 50 9 100 9 200
N0 (MPa) 542
10 560
10 5929 638 9 766 9
0 (MPa)
m 542
12.5 560 592 15.76389.2 766
10.1 16.2
m 12.5 10.1 16.2 15.7 9.2

Table 5. Weibull parameters for tensile strength of GFRP under different temperatures.
Table 5. Weibull parameters for tensile strength of GFRP under different temperatures.
Temperature (C) 25 0 25 50 75 100
N

Temperature ( C) 25 8 0 8 10
25 8 50 8 758 100
N 0 /MPa 8 550 8541 560
10 530 8 496 8426 8
0 /MPa m 550 9.1 541
8.7 10.1 25.4530 18.3 496
560 16.0 426
m 9.1 8.7 10.1 25.4 18.3 16.0
6. Conclusions
Polymers 2016, 8, 196 14 of 16

6. Conclusions
This experimental study focuses on the tensile characterization and failure pattern of glass yarn
and GFRP samples under different loading conditions. The strain rate and temperature effects on
the mechanical properties and fracture morphologies are investigated and discussed comparatively.
The following conclusions can be reached:

(1) There is an apparent dependence of the dynamics tensile properties of glass yarn and GFRP on
the strain rate. For the glass yarn, tensile strength and toughness increase as much as 88.0% and
474.3% during a transition from quasi-static loading (1/600 s1 ) to dynamic loading (40 s1 ),
but toughness decreases about 10.2% when the strain rate changes from 120 to 160 s1 due to
the diminution of strain. For GFRP, the tensile strength linearly increases nearly 49.1% over the
strain rate range of 1/600200 s1 , and toughness increases remarkably (about 109.7%) during
a transition from quasi-static loading (1/600 s1 ) to dynamic loading (25 s1 ).
(2) The mechanical properties of glass yarn and GFRP are also dependent on the temperature.
For glass yarn, tensile strength and toughness decrease nearly 25.3% and 31.1% when temperature
increases from 25 to 75 C. However, when heated to 100 C, tensile strength and toughness
of glass yarn rebound due to the augment of frictional force between fibers. For GFRP, tensile
strength shows almost no change (within 3%) when temperature increases from 25 to 50 C, but
decreases sharply (about 18.9%) over the temperature range of 50100 C because of the softening
of the resin matrix when Tg of epoxy resin is reached.
(3) The failure patterns of the GFRP specimens are closely related to the loading conditions. At low
strain rates, the fracture surface is limited in a small region, while, with increasing strain
rate, the damage path covers the entire gauge section where extensive debonding between
fibers and matrix was also observed, which leads to an increase in tensile strength and energy
absorption. In addition, the components of GFRP specimen are more separated after failure at
higher temperatures.
(4) The scale parameter (0 ) values are highly dependent on the strain rates and temperatures
investigated. For glass yarn, the results acquired from higher strain rate have a lower shape
parameter than the counterpart. This manifests that yarn subjected to higher strain rate is usually
characterized by a more random breakage process.

Acknowledgments: This work was supported by the funds from the National Basic Research Program of China
(973 program, Grant No. 2012CB026200), the Sci-Tech Support Plan of Hunan Province (Grant No. 2014WK2026).
Author Contributions: Deju Zhu and Liang Huang conceived and designed the experiments; Yunfu Ou,
Huaian Zhang, Yiming Yao and Gaosheng Li performed the experiments; Yunfu Ou, Huaian Zhang and
Yiming Yao analyzed the data; Deju Zhu and Barzin Mobasher contributed materials and analysis tools; Yunfu Ou
and Deju Zhu wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Bakis, C.; Bank, L.C.; Brown, V.; Cosenza, E.; Davalos, J.; Lesko, J.; Machida, A.; Rizkalla, S.; Triantafillou, T.
Fiber-reinforced polymer composites for construction-state-of-the-art review. J. Compos. Constr. 2002, 6,
7387. [CrossRef]
2. Bai, Y.; Keller, T.; Valle, T. Modeling of stiffness of FRP composites under elevated and high temperatures.
Compos. Sci. Technol. 2008, 68, 30993106. [CrossRef]
3. Bai, Y.; Valle, T.; Keller, T. Modeling of thermo-physical properties for FRP composites under elevated and
high temperature. Compos. Sci. Technol. 2007, 67, 30983109.
4. Cantwell, W.J.; Morton, J. The impact resistance of composite materialsA review. Composites 1991, 22,
347362. [CrossRef]
5. Kim, M.; Kang, S.; Kim, C.; Kong, C. Tensile response of graphite/epoxy composites at low temperatures.
Compos. Struct. 2007, 79, 8489. [CrossRef]
Polymers 2016, 8, 196 15 of 16

6. Reis, J.; Coelho, J.; Monteiro, A.; da Costa Mattos, H. Tensile behavior of glass/epoxy laminates at varying
strain rates and temperatures. Compos. Part B 2012, 43, 20412046. [CrossRef]
7. Snchez-Sez, S.; Gmez-del Ro, T.; Barbero, E.; Zaera, R.; Navarro, C. Static behavior of CFRPs at low
temperatures. Compos. Part B 2002, 33, 383390. [CrossRef]
8. Zhu, D.; Rajan, S.; Mobasher, B.; Peled, A.; Mignolet, M. Modal analysis of a servo-hydraulic high speed
machine and its application to dynamic tensile testing at an intermediate strain rate. Exp. Mech. 2011, 51,
13471363. [CrossRef]
9. Ou, Y.; Zhu, D. Tensile behavior of glass fiber reinforced composite at different strain rates and temperatures.
Constr. Build. Mater. 2015, 96, 648656. [CrossRef]
10. Kolsky, H. An investigation of the mechanical properties of materials at very high rates of loading. Proc. Phys.
Soc. Lond. Sect. B 1949, 62, 676. [CrossRef]
11. Wang, K.; Young, B.; Smith, S.T. Mechanical properties of pultruded carbon fibre-reinforced polymer (CFRP)
plates at elevated temperatures. Eng. Struct. 2011, 33, 21542161. [CrossRef]
12. Melin, L.; Asp, L. Effects of strain rate on transverse tension properties of a carbon/epoxy composite: Studied
by moir photography. Compos. Part A 1999, 30, 305316. [CrossRef]
13. Al-Zubaidy, H.; Zhao, X.; Al-Mahaidi, R. Mechanical characterisation of the dynamic tensile properties of
CFRP sheet and adhesive at medium strain rates. Compos. Struct. 2013, 96, 153164. [CrossRef]
14. Koerber, H.; Xavier, J.; Camanho, P.P. High strain rate characterisation of unidirectional carbon-epoxy
im7-8552 in transverse compression and in-plane shear using digital image correlation. Mech. Mater. 2010, 42,
10041019. [CrossRef]
15. Zhou, Y.; Wang, Y.; Mallick, P. An experimental study on the tensile behavior of kevlar fiber reinforced
aluminum laminates at high strain rates. Mater. Sci. Eng. A 2004, 381, 355362. [CrossRef]
16. Barre, S.; Chotard, T.; Benzeggagh, M. Comparative study of strain rate effects on mechanical properties of
glass fibre-reinforced thermoset matrix composite. Compos. Part A 1996, 27, 11691181. [CrossRef]
17. Shokrieh, M.M.; Omidi, M.J. Tension behavior of unidirectional glass/epoxy composites under different
strain rates. Compos. Struct. 2009, 88, 595601. [CrossRef]
18. Ochola, R.O.; Marcus, K.; Nurick, G.N.; Franz, T. Mechanical behavior of glass and carbon fibre reinforced
composites at varying strain rates. Compos. Struct. 2004, 63, 455467. [CrossRef]
19. Hawileh, R.A.; Abu-Obeidah, A.; Abdalla, J.A.; Al-Tamimi, A. Temperature effect on the mechanical
properties of carbon, glass and carbonglass FRP laminates. Constr. Build. Mater. 2015, 75, 342348.
[CrossRef]
20. Robert, M.; Benmokrane, B. Behavior of GFRP reinforcing bars subjected to extreme temperatures.
J. Compos. Constr. 2009, 14, 353360. [CrossRef]
21. Zhu, D.; Mobasher, B.; Rajan, S.D. Dynamic tensile testing of kevlar 49 fabrics. J. Mater. Civ. Eng. 2011, 23,
230239. [CrossRef]
22. Kuentzer, N.; Simacek, P.; Advani, S.G.; Walsh, S. Correlation of void distribution to vartm manufacturing
techniques. Compos. Part A 2007, 38, 802813. [CrossRef]
23. Zhu, D.; Peled, A.; Mobasher, B. Dynamic tensile testing of fabriccement composites. Constr. Build. Mater.
2011, 25, 385395. [CrossRef]
24. Ou, Y.; Zhu, D.; Huang, M.; Li, H. Gage length and strain rate effects on tensile behavior of kevlar 29 single
filament and yarn. J. Compos. Mater. 2016, 0, 115. [CrossRef]
25. Zhu, D.; Gencoglu, M.; Mobasher, B. Low velocity flexural impact behavior of ar glass fabric reinforced
cement composites. Cem. Concr. Compos. 2009, 31, 379387. [CrossRef]
26. Xiao, X. Dynamic tensile testing of plastic materials. Polym. Test. 2008, 27, 164178. [CrossRef]
27. Reed, R.; Golda, M. Cryogenic properties of unidirectional composites. Cryogenics 1994, 34, 909928.
[CrossRef]
28. Ditlevsen, O. Uncertainty Modeling with Applications to Multidimensional Civil Engineering Systems;
McGraw-Hill International Book Co.: New York, NY, USA, 1981.
29. Weibull, W. A statistical distribution function of wide applicability. J. Appl. Mech. 1951, 13, 293297.
30. Padgett, W.; Durham, S.; Mason, A. Weibull analysis of the strength of carbon fibers using linear and power
law models for the length effect. J. Compos. Mater. 1995, 29, 18731884. [CrossRef]
Polymers 2016, 8, 196 16 of 16

31. Hui, C.; Phoenix, S.; Shia, D. The single-filament-composite test: A new statistical theory for estimating the
interfacial shear strength and weibull parameters for fiber strength. Compos. Sci. Technol. 1998, 57, 17071725.
[CrossRef]
32. Alqam, M.; Bennett, R.M.; Zureick, A.-H. Three-parameter vs. Two-parameter weibull distribution for
pultruded composite material properties. Compos. Struct. 2002, 58, 497503. [CrossRef]
33. Sriramula, S.; Chryssanthopoulos, M.K. Quantification of uncertainty modelling in stochastic analysis of frp
composites. Compos. Part A 2009, 40, 16731684. [CrossRef]
34. Lekou, D.; Philippidis, T. Mechanical property variability in frp laminates and its effect on failure prediction.
Compos. Part B 2008, 39, 12471256. [CrossRef]

2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).