Experimental study of tensile behaviour of AFRP under different strain ...

11 downloads 92658 Views 6MB Size Report
field, automotive, and petrochemical industries. Applications in ... temperature conditions during their service life (Yang, Song,. & Zhang, 2015). ..... 2012CB026200); the Sci-Tech Support ... correction. Experimental Mechanics, 29, 261–267.
Journal of Structural Integrity and Maintenance, 2016 VOL. 1, NO. 1, 22–34 http://dx.doi.org/10.1080/24705314.2016.1153327

Experimental study of tensile behaviour of AFRP under different strain rates and temperatures Xiaotong Zhanga, Deju Zhua, Yiming Yaob, Huaian Zhanga, Barzin Mobasherb and Liang Huanga a

College of Civil Engineering, Hunan University, Changsha, China; bSchool of Sustainable Engineering and Built Environment, Arizona State University, Tempe, AZ, USA

KEYWORDS

ABSTRACT

This paper focuses on the tensile behaviours of Kevlar 29 and 49 fabrics reinforced polymers (AFRP-K29 and -K49) at four strain rates (25, 50, 100 and 200 s−1) and five temperatures (−25, 0, 25, 50 and 100 °C). The experimental results show that the tensile mechanical properties of two types of aramid fibrereinforced polymer (AFRP) are strain-rate dependent in terms of Young’s modulus, tensile strength and toughness at 25 °C. While the changes of mechanical properties except for the Young’s modulus of both AFRPs and the tensile strength of AFRP-K29 under different temperatures (−25, 0, 25, 50 and 100 °C) are not significant at the strain rate of 25 s−1. The failure patterns of AFRP samples are similar at the strain rates and temperatures investigated. Non-uniform strain fields on AFRP specimens were identified by digital image correlation (DIC) technique.

1. Introduction Fibre-reinforced polymers (FRP) such as carbon, glass, and aramid fibre-reinforced polymers (CFRP, GFRP, and AFRP) are valuable alternatives to conventional materials due to their high strength to weight ratio, corrosion resistance, damage tolerance, and potentially high durability characteristics. These superior properties of FRP enable its wide applications in many engineering fields including aircraft, aerospace structures, military field, automotive, and petrochemical industries. Applications in strengthening, retrofit, and rehabilitation of structural members of civil engineering infrastructure have also been extensively adopted over the past decades (Behnam & Eamon, 2013; Pessiki, Harries, Kestner, Sause, & Ricles, 2001; Zhou & Attard, 2013). FRP used in such fields might be subjected to dynamic loadings, such as wind loads, earthquake loads, etc. and varying temperature conditions during their service life (Yang, Song, & Zhang, 2015). Despite the advantages of using FRP over traditional materials, one of the major impediments to its wider utilization is the limited knowledge of the mechanical performance of such composite materials being exposed to varying temperatures and/or subjected to dynamic loads. Thus, the fundamental understandings of the strain-rate and temperature effects, or their interactions, on the tensile behaviour of FRP are particularly essential to the design of composite structures, which have attracted increasing interest. The tensile properties of CFRP materials at quasi-static and intermediate strain rates were studied by several researchers (Al-Zubaidy, Zhao, & Al-Mahaidi, 2013; Deshpande, 2006; Harding & Welsh, 1983; Melin & Asp, 1999). It was reported that the tensile properties of CFRP were strain rate-dependent, while the average transverse modulus was independent of strain rate. GFRP also showed strain rate dependence within a wide range from 1.6  ×  10−5 to 1000  s−1 (Lifshitz & Rotem, 1970; Naik, Yernamma, Thoram, Gadipatri, & Kavala, 2010; Reis,

CONTACT  Deju Zhu 

[email protected]

© 2016 Korea Institute for Structural Maintenance and Inspection

AFRP; strain rate; temperature effect; failure patterns; tensile behaviour; DIC

Coelho, Monteiro, & da Costa Mattos, 2012; Shokrieh & Omidi, 2009). Hawileh, Abu-Obeidah, Abdalla, and Al-Tamimi (2015) and Calvet, Valcuende, Benlloch, and Cánoves (2015) studied the variation in mechanical properties in terms of the elastic modulus and tensile strength of CFRP and GFRP exposed to different temperatures (5–300 °C). Brittle rupture of fibre was observed during 100–150  °C accompanied by partial loss of epoxy adhesives while sheet splitting dominated during 200–250 °C. What is more, GFRP specimens indicated a minor reduction in axial tensile strength and strain with increasing temperature. Ramroth, Asaro, Zhu, and Krysl (2006) proposed a constitutive model for FRP-laminated composites accounting for both temperature and strain rate effects, which was verified by an experiment involving a panel exposed to combined mechanical and thermal loadings. It is well known that carbon fibres are relatively brittle, while glass fibres are prone to chemical or physical attack, and also have a higher density and a poorer fibre/matrix interfacial performance (Greenhalgh, 2009). In recent years, aramid fibre is becoming popular as reinforcement in composite materials due to its superior properties, such as relatively high strength to mass ratio and modulus, low densities, and high impact resistance (Singh & Samanta, 2015). Thus, strain rate and temperature effects on the tensile behaviours of AFRP have also acquired some attention. Wang and Xia (2000) performed tensile experiments on unidirectional AFRP at strain rates of 150, 400, and 1500 s−1 and rate dependence of mechanical properties was found. By adopting the double Weibull distribution function, a modified one-dimensional constitutive equation was proposed to describe the strength distribution, and consistency between the simulated results and the experimental data was acquired. Welsh and Harding (1985) obtained the tensile stress–strain curves of Kevlar fibre woven-reinforced polyester resin composites at strain rates of 10−4, 10, and 1000 s−1, and the variation

Journal of Structural Integrity and Maintenance 

 23

Table 1. Basic material properties of Kevlar fabrics. Material Kevlar 29/49

Yarn count (yarn/cm) 6.7 × 6.7

Bulk density (g/cm3) 1.44

Linear density (g/cm) 1.656 × 10−3

Figure 1. Prepared (a) AFRP-K29 and (b) AFRP-K49 specimens for dynamic tensile tests.

of Young’s modulus, tensile strength, and fracture strain with strain rate were determined. Both the Young’s modulus and tensile strength increased at the impact loading rate, while the failure strain decreases slightly. Examination of broken AFRP specimens showed the fracture surface appeared to be rate independent. Zhou, Wang, and Mallick (2004) determined the tensile behaviour of Kevlar fibre-reinforced aluminum laminates at strain rates of 100, 500, and 1300 s−1 and found that the yield strength, tensile strength, and failure strain of the composite all increased with increasing strain rate. Tann and Delpak (2004) reported the influences of freeze and thaw actions as well as high temperatures (180  °C) on the mechanical properties of AFRP. It was found that high temperature results in reduced ultimate tensile strength, but increased Young’s modulus, while the freeze and thaw cycles exhibit relatively minor influences. Rodriguez, Chocron, Martinez, and Sánchez-Gálvez (1996) and Chocron Benloulo, Rodríguez, Martínez, and Sánchez Gálvez (1997) studied the influence of the strain rate (10−3, 1, and 103 s−1) on the mechanical properties of aramid woven fabric composites and indicated that strain rate significantly influenced the longitudinal tensile strength and failure strain. All the specimens turned out to have the same kind of rupture, and all the yarns broke without slipping. When the tension tests are conducted at high speed, a high sampling rate in the range of 10–1000 kHz (Zhu, Mobasher, & Rajan, 2010) is required to acquire sufficient data points within a few milliseconds. Furthermore, the attachment of additional transducers is not quite feasible compared to static tests due to the limitation in specimen size and the inertial effect of extra mass to the samples during dynamic testing, which may affect the test results and experimental accuracy. On the other hand, the strain measured at an isolated spot or within a gauge length by conventional transducers such as linear variable differential transformer, extensometer, and strain gauge is insufficient to study the potentially inhomogeneous behaviours. Therefore, Digital Image Correlation (DIC), as a non-contacting optical full-field deformation measurement approach, is used to better address the tensile behaviour of composites at high loading rates. DIC technique was developed by Sutton, Wolters, Peters, Ranson, and McNeill (1983)

Diameter of filament (μm) 12

Cross-sectional area per yarn (cm2) 1.15 × 10−3

and Bruck, McNeill, Sutton, and Peters (1989), and has been widely applied for composites, reinforced concrete sections (Destrebecq, Toussaint, & Ferrier, 2011; Giancane, Panella, Nobile, & Dattoma, 2010; Shah & Chandra Kishen, 2011), as well as recent applications in cement-based composites tested under dynamic loads (Gao, Huang, Xia, & Li, 2014; Koerber, Xavier, & Camanho, 2010). Compared to other FRPs (such as CFRP and GFRP), the lowest density of AFRP demonstrates far superior specific properties (Matthews & Rawlings, 1999). Also, the forte of AFRP is its enhanced impact properties (Cen, Kang, Lei, Qin, & Qiu, 2006), significantly lower fibre elongation, higher tensile strength, and modulus, as well as good high temperature properties (Singh & Samanta, 2015). It obtains applications mainly in industrial and advanced technologies like ballistic armor, military and aerospace applications, etc. However, for further enhancing its applications, proper characterization is very important. Although there are a lot of research on AFRP as introduced above, there is still very little information on the mechanical performance of AFRP at medium strain rates (10–200 s−1), and no experimental results are available on the tensile behaviours of AFRP under the coupling effect of medium strain rates and different temperatures. The objective of present study is to obtain the dynamic mechanical properties and failure characteristics of AFRP under different medium strain rates (25, 50, 100, and 200 s−1) and temperatures (−25, 0, 25, 50, and 100 °C), and to study the possible coupling effect of strain rate and temperature on the dynamic behaviour of AFRP. The results of this investigation will be valuable for theoretical research and practical application of AFRP in extreme service conditions and environments, and also can be implemented in a finite element code to develop material constitutive models of AFRP for numerical simulations.

2.  Experimental program 2.1.  Specimen preparation Thin AFRP laminates with a nominal thickness of 0.5  mm were fabricated with single ply of Kevlar fabric and epoxy resin JN-C3P by means of Vacuum Assisted Resin Infusion (Xia, Shi, & Cai, 2015). Kevlar® 29 and Kevlar® 49 fabrics (manufactured by DuPont) were chosen as reinforcement for these two types of composites, whose basic material properties are listed in Table 1. Prepared laminate plates were then cut into small size coupons with a width of 22  mm (about 15 yarns of Kevlar) in the warp direction to accommodate the size of grips. The resulting fibre content is approximately 15.7% by volume. To ensure a uniform load transfer from the grips to the specimens and prevent stress concentration, aluminum sheets with the dimension of 40 mm × 22 mm × 0.3 mm were locally bonded at the both ends of the specimens using epoxy. The total length of AFRP specimen is 105 mm with a gauge length of 25 mm, as shown in Figure 1. In present paper, the terms AFRP-K29 and AFRP-K49 are defined to represent the two types of AFRP reinforced by Kevlar® 29 and Kevlar® 49 fabrics, respectively.

24 

  X. Zhang et al.

Figure 2. Experimental set-up: (a) MTS servo-hydraulic high speed testing machine and (b) AFRP specimen held in grips.

2.2.  Loading devices and technique Tensile tests of AFRP specimens under various strain rates and temperatures were performed using a MTS high rate ­servo-hydraulic testing machine with an environmental chamber in Arizona State University (as shown in Figure 2). 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 process of tests. According to the calibration records, the stroke can reach a maximum speed of 14 m/s with a load capacity of 20 kN (Zhu et al., 2010). 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 nominal strain rates are defined as the ratio of the applied velocity to the specimen’s gauge length, namely 25, 50, 100, and 200 s−1, respectively. At the nominal strain rate of 25 s−1, five different temperatures (−25, 0, 25, 50, and 100 °C) were selected for temperature effect testing. Ten replicates were tested for each condition to reduce the influence of random error. A high-speed digital data acquisition card was used to collect force and displacement data at a sampling rate of 500 kHz. A Phantom v7.3 high-speed digital camera with a sampling rate of 20,000 fps was used to capture the deformation and failure behaviour of AFRP specimens.

3.  Results and discussion 3.1.  Strain rate effect on mechanical properties Figures 3 and 4 show the stress–strain curves of AFRP-K29 and AFRP-K49 at four different nominal strain rates (25, 50, 100, and 200  s−1). The actual strain rates, which were measured from strain measurement and test duration as indicated in

the figures, were different from the nominal strain rates. This phenomenon attributes to the fact that the reaction force of the test specimen slows down the stroke, especially at higher strain rates (and loading speeds) when the hydraulic pressures are relative lower. Note that there exists a certain degree of discrepancy in actual strain rates, and in the subsequent discussion of this paper, the nominal strain rates will be used to facilitate analysis. The stress–strain behaviours of two kinds of AFRP are similar to each other at the same strain rate, while different shape is observed as the strain rate increases from 25 to 200  s−1. The bimodal curves observed under the nominal strain rate of 25 s−1 may be caused by the partial failure of longitudinal yarns at approximately halfway of the peak stress, followed by stress redistribution and reduced tensile stiffness up to sample failure. At the other three strain rates, the behaviours of the stress–strain curves are analogous which can be characterized by three regions: initial crimp region, elastic region, and post-peak region, as shown in Figure 4(c). During initial crimp region, relatively lower stresses are needed to straighten the longitudinal undulated fibres and remove the crimp (Rodriguez et al., 1996). After the inner fibres are fully straightened, AFRP specimens exhibit a linear response, which is called elastic region, and the Young’s modulus of the fabric is defined as the slope of the curve in this region. The slope (absolute value) of the post-peak region, which indicates decrease in the loading-carrying capacity, also decreases with the strain rate increasing from 50 to 200 s−1, suggesting less dramatic failure process and more fibres breakage after AFRP reaching the maximum stress (tensile strength), due to the possible reason that there exists larger discreteness in the failure strain of fibres at higher strain rate. Figure 5 shows the dependence of the tensile properties in terms of Young’s modulus, tensile strength and toughness of AFRP on strain rates. The toughness is calculated as the area under stress–strain curves indicating unit volume deformation energy, given by: 𝜀f

energy = 𝜎d𝜀 UT = volume ∫

(1)

0

where the term UT represents toughness; ε is strain; εf is the maximum strain; and σ is stress. For both types of AFRP specimens, the Young’s modulus and tensile strength increase when the strain rate increases from 25 to 50 s−1, and then decrease as strain rate increases from 50 to 200 s−1. While with the increasing strain rate, the toughness of AFRP-K29 shows an opposite tendency, and that of AFRP-K49 reveals a fluctuant process (first decreases, then increases, and finally decreases again). Concretely speaking, the Young’s modulus of AFRP-K29 and AFRP-K49 increase about 68.6 and 87.7% with strain rate increasing from 25 to 50  s−1, then decrease from 23.1  ±  7.5 to 6.4  ±  1.9  MPa and from 19.9 ± 5.9 to 5.8 ± 0.5 MPa, respectively, with strain rate increasing to 200  s−1. Tensile strengths of these two kinds of AFRP increase 16.2 and 29.7% firstly, then decrease from 284.8  ±  41.6 to 173.1  ±  22.2  MPa and from 302.3  ±  20.7 to 173.6  ±  24.0  MPa, respectively. The toughness of AFRP-K29 displays an opposite trend, with value decreasing 9.0% firstly and then increasing 20.6%, while that of AFRP-K49 fluctuates within 27.1%. From an overall perspective, the tensile properties of AFRP-K29 and AFRP-K49 are quite comparable with each other.

Journal of Structural Integrity and Maintenance 

 25

Figure 3. Stress–strain curves of AFRP-K29 at different nominal strain rates: (a) 25 s−1, (b) 50 s−1, (c) 100 s−1, and (d) 200 s−1.

Figure 4. Stress–strain curves of AFRP-K49 at different nominal strain rates: (a) 25 s−1, (b) 50 s−1, (c) 100 s−1, and (d) 200 s−1.

3.2.  Temperature effect on mechanical properties Figures 6 and 7 show the stress–strain curves of AFRP-K29 and AFRP-K49 at the strain rate of 25 s−1 and five different temperatures, i.e. −25, 0, 25, 50, and 100 °C. The effect of temperature on Young’s modulus, tensile strength, and toughness are shown in Figure 8 in terms of average values and standard deviations. In order to perform the variation trend vividly, linear fitting is utilized. For AFRP-K29, slight ascending trend in Young’s Modulus is observed when temperature increases from −25 to 50 °C, while a decrease is found as temperature rises to 100 °C. Tensile strength increases

from −25 °C to room temperature while the influence of temperature is less significant above 25 °C. On the other hand, toughness stays nearly the same which reveals the stability in energy absorption of AFRP-K29 at varying temperatures. Specifically, the fitting line equation of Young’s modulus is { 0.06 × X + 13.83(−25 ≤ X ≤ 50) Y= (2) −0.12 × X + 23.99(50 < X ≤ 100) where the term X represents temperature, and Y is the Young’s modulus. Tensile strength increases from 228.0  ±  10.0 to 243.0 ± 21.0 MPa, with fitting bilinear equation shown below:

26 

  X. Zhang et al.

Figure 5. Strain rate effect on dynamic material properties of AFRP at the room temperature: (a) Young’s modulus, (b) tensile strength, and (c) toughness.

Figure 6. Stress–strain curves of AFRP-K29 at different temperatures and a nominal strain rate of 25 s−1: (a) −25 °C, (b) 0 °C, (c) 25 °C, (d) 50 °C, and (e) 100 °C.

{ Y=

0.49 × X + 235.08(−25 ≤ X ≤ 25) 0.04 × X + 249.21(25 < X ≤ 100).

(3)

While changes in toughness under various temperatures are below 4.3%. For AFRP-K49, the effect of temperature on tensile strength are less significant compared to AFRP-K29, while similar

Journal of Structural Integrity and Maintenance 

 27

Figure 7. Stress–strain curves of AFRP-K49 at different temperatures and a nominal strain rate of 25 s−1: (a) −25 °C, (b) 0 °C, (c) 25 °C, (d) 50 °C, and (e) 100 °C.

Figure 8. Temperature effect on dynamic material properties of AFRP at a nominal strain rate of 25 s−1: (a) Young’s modulus, (b) tensile strength, and (c) toughness.

28 

  X. Zhang et al.

Figure 9. Comparison of cumulative failure probability vs. tensile strength of (a) AFRP-K29, (b) AFRP-K49 at different nominal strain rates, (c) AFRP-K29, and (d) AFRP-K49 at different temperatures.

Table 2. Weibull parameters for tensile strength of AFRP at different strain rates and temperatures. 25 °C Experimental conditions AFRP-K29 N σ0/MPa m AFRP-K49 N σ0/MPa m

25 s−1 10 254 8.7 9 234 18.7

50 s−1 9 307 8.1 7 312 22.2

25 s−1 100 s−1 9 268 12.3 7 308 7.1

trends with respect to increasing temperature are observed in Young’s modulus, and the equation is { 0.02 × X + 10.56(−25 ≤ X ≤ 50) Y= (4) −0.07 × X + 15.16(50 < X ≤ 100). The toughness of AFRP-K49, however, slightly changes from 5.99 ± 0.79 to 6.59 ± 0.33 MPa as temperature increases from −25 to 100 °C. As far as the comparison between two AFRP systems, larger Young’s modulus values are observed in AFRP-K29 compared to AFRP-K49 at all temperatures investigated. The tensile strength of AFRP-K29 is slightly higher than that of AFRP-K49 except at −25 °C. In addition, AFRP-K49 specimens exhibit higher toughness under varying temperatures from −25 to 100  °C which can be traced back to its higher ductility as observed in the stress–strain curves. It is known that the properties of epoxy resin distinctly depend on temperature such that a transition from brittle to ductile behaviours is expected as temperature increases above the glass transition temperature (Tg). Gupta, Drzal, Lee, and Rich (1985) reported considerable decreases in the Young’s modulus of cured epoxy resin when the temperature exceeded Tg. On the other hand, Kevlar fibres have good resistance to thermal change (Wang & Xia, 1999). Due to the low fibre content (about 15.7%), the linear elastic behaviour when fibre and matrix are bonded is dominated by the matrix (temperature dependent) phase. Therefore, the decreases in Young’s modulus at 100 °C may be attributed to the fact that the Tg of the epoxy resin

200 s−1 5 212 18.0 6 173 23.1

−25 °C 8 229 25.6 8 232 12.9

0 °C 8 227 7.0 7 226 14.3

50 °C 8 251 12.8 8 237 7.2

100 °C 5 242 7.4 7 231 18.9

is exceeded. However, the ultimate strength of AFRP primarily depends on the tensile strength of Kevlar fabrics (temperature independent phase) since the mode of failure is characterized by fibre fracture. Thus the effects of temperature on the tensile strength are marginal. The difference in Young’s modulus between two types of AFRP at different temperatures may be attributed to diverse fabrics, and different coefficients of thermal expansion between Kevlar fibres and epoxy resin (Shan et al., 2015), which causes residual stress and interaction (Cen et al., 2006).

3.3.  Weibull analysis Data scatters are observed in the tensile strength of AFRP which is traced back to presence of flaws introduced during processing and handling, as well as the potential eccentric loads during the tensile tests. These factors give rise to the variability in tensile strength of AFRP, and hence the strength is of a probabilistic nature. Several statistical distributions have been used to describe this variation. Since two-parameter Weibull analysis can provide a reasonable fit line and improve computational efficiency (Alqam, Bennett, & Zureick, 2002), it becomes the most widely used approach to describe the variability in the strength of fibres (Hui, Phoenix, & Shia, 1998; Wang & Xia, 1998) and FRP composites (Alqam et al., 2002; Lekou & Philippidis, 2008). The typical form of the two-parameter Weibull distribution for cumulative probability density can be characterized as:

Journal of Structural Integrity and Maintenance 

Figure 10. Failure process of AFRP-K29 at different nominal strain rates and room temperature: (a) 25 s−1, (b) 50 s−1, (c) 100 s−1, and (d) 200 s−1.

Figure 11. Failure process of AFRP-K49 at different nominal strain rates and room temperature: (a) 25 s−1, (b) 50 s−1, (c) 100 s−1, and (d) 200 s−1.

 29

30 

  X. Zhang et al.

Figure 12. Fracture morphologies of (a) AFRP-K29 and (b) AFRP-K49 at different nominal strain rates and room temperature.

Figure 13. Fracture morphologies of (a) AFRP-K29 and (b) AFRP-K49 at different temperatures and a nominal strain rate of 25 s−1.

P=

Figure 14. Basic principle of DIC: (a) speckle pattern, AOI and subset and (b) track of subset using cross correlation.

[ ( )m ] 𝜎 P(𝜎) = 1 − exp − 𝜎0

(5)

where σ is the tensile strength, σ0 is the characteristic strength, and m is the shape parameter, characterizing the spread in the distribution of strength. P was estimated using the following equation:

i N+1

(6)

where N is the total number of tests and i is the current test number (Zhu, Mobasher, Erni, Bansal, & Rajan, 2012). Figure 9 shows the Weibull curves fitting to experimental data under different strain rates and temperatures, with Weibull distribution parameters summarized in Table 2. Figure 9(a) and (b) show that as strain rate increases from 25 to 50 s−1, the cumulative probability plot shifts towards higher stress values, while the move in opposite direction occurs when strain rate continues increasing from 50 to 200 s−1, which is in line with Figure 5. The cumulative probability plots of AFRP-K29 at different strain rates are in a wider range than AFRP-K49 except for the strain rate of 100 s−1, indicating large discreteness in tensile strength, which can also be deduced from the smaller value of m. Figures 9(c) and (d) display inconspicuous shift of cumulative probability plot under different temperatures, which is consistent with Figure 8.

3.4.  Effect of different strain rates and temperatures on failure mode Figures 10 and 11 illustrate the tensile deformation process up to sample failure under different strain rates recorded

Journal of Structural Integrity and Maintenance 

 31

Figure 15. Strain map of εyy for AFRP-K29 and AFRP-K49 specimens tested at a nominal strain rate of 100 s−1 obtained by DIC.

Figure 16. (a) Non-uniform strain distribution, (b) schematic diagram of the fabric reinforcement, and (c) 3-D surface map of longitudinal strain (εyy).

using high-speed digital camera, and Figure 12 shows the damage morphologies of tested specimens after final failure. Examination of broken AFRP specimens shows that the mode of failure appeared to be rate independent such that fibres and epoxy resin fail approximately at the same location and

a relatively flat fracture surface normal to loading direction is apparent. The fracture morphologies agree with the observations in published literature (Rodriguez et al., 1996; Welsh & Harding, 1985). The damage zone is not strictly straight, which should be explained by the randomly distributed flaws in the

32 

  X. Zhang et al.

(a)

(b)

(c)

(d)

Figure 17. (a) Time history of experimental stress and strain, and DIC strain, (b) correlation of experimental and DIC stress–strain curves; comparison of experimental stress–strain responses for (c) AFRP-K29 and (d) AFRP-K49 at different nominal strain rates.

fibres such that the fracture may initiate at slightly different locations even though a uniform stress distribution is ensured. Similar failure mode can also be observed under different temperatures (Figure 13).

4.  Image analysis 4.1.  Basic principle of DIC Conventional data acquisition and analysis techniques may not be sufficient to fully understand the material behaviour, hence further study of the strain field and damage process using image analysis is warranted. Figure 14 shows the manually specified area of interest (AOI) needed for DIC and the track of subset movement from reference to deformed images, as the basic principle of DIC. The displacements are computed at the centre point P(x,y) of each subset throughout the entire AOI to obtain full-field deformation. The use of correlation functions such as cross-correlation or normalized cross-correlation enables the tracking of subset from P(x,y) to P′(x′,y′) (Pan, Qian, Xie, & Asundi, 2009). The strain fields can be subsequently obtained by smoothing and differentiating the displacement fields. A commercial software Vic-2D 2009 developed by Correlated Solutions, Inc. was used to conduct image analysis.

4.2.  Full-field strain distributions The longitudinal strain (εyy) fields of AFRP-K29 and K49 specimens tested at 100 s−1 are shown in Figure 15 using a colour code with purple representing the lowest strain values and red at 3.0% strain. The tensile stresses associated with each step of the strain distributions are indicated below each sub-image.

The strain maps are also related to the numbers 2–4 labelled on the stress–strain curves shown at the bottom of Figure 15. Similar damage evolution can be seen in the two specimens such that at the beginning of the test (σ = 38.7/38.2 MPa), a relatively uniform strain distribution is obtained which can be correlated to the stage of uncramping. As σ increases to 197.8/118.2  MPa, a non-uniform pattern of strain field is initiated represented by distributed zones/spots in green with higher strain values (around 2%) while the tensile strain in the rest of AOI stays below 1%. Similar pattern of strain map was also observed in the specimens tested at other strain rates. The contrast becomes more evident as the tensile stress continues increasing (σ = 119.4/118.2 MPa) demonstrated by yellow/red colour which is close to a 3% strain. A cross pattern is completely formed towards the peak stress (σ = 283.4/227.5 MPa), especially in the AFRP-K49 specimen which is, to a great extent, similar to the biaxial warp-fill structure of fabrics reinforcement. Figure 16(a) shows the strain map of AFRP-K49 specimen with equally spaced lines which isolate the peak spots/zones in a cross manner and reveal the fact that these spots/zones form at approximately regular intervals. There are 15 intervals in the warp or loading direction and the distance between two adjacent vertical lines is about 1.4 mm, which correlates with the number and width of warp yarn bundles in the reinforcement. Since the warp yarns starts to take the majority of the tensile load after crimp region, larger longitudinal strains are also expected in warp yarns rather than fill yarns. Therefore, the peak zones/spots may be related to the warp junction as defined in Figure 16(b). This phenomenon is more evident in a 3-D surface plot characterized by the peaks and valleys in Figure 16(c).

Journal of Structural Integrity and Maintenance 

4.3.  Correlation with experimental stress–strain responses In order to correlate the strain obtained by DIC with experimental measurements, assumption of homogenization is used and the average strain of the entire AOI is exported and plotted as a function of time after linear interpolation and data smooth (see Figure 17(a)). In addition to the DIC strain, the experimental stress and strain vs. time histories are also plotted in the same figure, which enables the correlation of DIC strain to experimental stress. Using the correlation strategy, a comparison between the experimental and DIC stress–strain responses of an AFRP-K29 specimen is obtained, as shown in Figure 17(b), where a good agreement between the two curves can be observed. Further comparisons of two responses for representative AFRP-K29 and AFRP-K49 specimens tested at varying strain rates are illustrated in Figures 17(c) and (d). The DIC responses are generally quite comparable to the experimental stress–strain curves even though discrepancies are observed at high strain levels towards the peak. The conventional data analysis procedure is based on the assumption of uniform strain distribution such that the engineering or nominal strain is calculated as the stroke displacement over gauge length. According to the strain maps shown in Figure 15, the distribution of tensile strain is relatively uniform at lower stress level which agrees with the assumption. However, the intrinsic inhomogeneity in the FRP starts dominating the tensile behaviour at higher stress levels and the differences between strain computation methods may result in the discrepancies.

5. Conclusion This study focuses on the tensile characterization and failure pattern of AFRP composed of single ply of Kevlar 29/49 fabric and epoxy resin under different loading conditions. The strain rate and temperature effects are investigated and discussed. The following conclusions can be reached: (1) The Young’s modulus and tensile strength increase when the strain rate increases from 25 to 50 s−1, but decrease as strain rate continues increasing to 200 s−1 for both kinds of AFRP. While the toughness of AFRP-K29 shows an opposite tendency, and that of AFRP-K49 experiences decrease, increase, and decrease process. (2) The Young’s modulus and tensile strength of AFRP-K29 are the primary mechanical properties affected by the increasing temperature, while no significant changes are observed in tensile strength of AFRP-K49 and toughness of these two kinds of AFRPs. For both kinds of AFRPs, Young’s modulus decreases when temperature is elevated over Tg. The Young’s modulus of AFRP-K29 is larger than that of AFRP-K49, with toughness relatively smaller. Small difference in tensile strength exists between these two types of AFRP. (3) The failure patterns of AFRP samples are all similar at different strain rates and temperatures investigated, and the fibres and epoxy resin failed approximately at the same location, forming a relatively flat fracture surface normal to loading direction. (4)  Non-uniform strain fields in AFRP specimens are obtained by DIC method. Peak zones/spots with higher strain values are identified and associated with the biaxial structure of fabrics reinforcement. This method helps capture the non-uniform pattern of strain field

 33

which is not detectable using conventional technique, and the actual strain distribution contributes to further study of the damage process and fully understanding of the material behaviours, which make up the deficiency in experiment. The strain responses determined by means of DIC are averaged and correlated with experimental stresses. Experimental and DIC stress– strain responses confirm well with each other. (5) In the future, the mechanical properties of AFRP specimens with multi-layers of aramid fabrics subjected to different strain rates and temperatures will be tested. The results will be compared with one-layer aramid fabric-reinforced polymers mentioned in this paper to further expand the applications of AFRP.

Disclosure statement No potential conflict of interest was reported by the authors.

Funding This work was supported by the funds from National Basic Research Program of China (973 program, Project No. 2012CB026200); the Sci-Tech Support Plan of Hunan Province [grant number 2014WK2026]; the Interdisciplinary Research Project of Hunan University and the Recruitment Program of Global Youth Experts.

References Alqam, M., Bennett, R.M., & Zureick, A.-H. (2002). Three-parameter vs. two-parameter Weibull distribution for pultruded composite material properties. Composite Structures, 58, 497–503. Al-Zubaidy, H., Zhao, X.-L., & Al-Mahaidi, R. (2013). Mechanical characterisation of the dynamic tensile properties of CFRP sheet and adhesive at medium strain rates. Composite Structures, 96, 153–164. Behnam, B., & Eamon, C. (2013). Reliability-based design optimization of concrete flexural members reinforced with ductile FRP bars. Construction and Building Materials, 47, 942–950. Bruck, H., McNeill, S., Sutton, M.A., & Peters III, W. (1989). Digital image correlation using Newton-Raphson method of partial differential correction. Experimental Mechanics, 29, 261–267. Calvet, V., Valcuende, M., Benlloch, J., & Cánoves, J. (2015). Influence of moderate temperatures on the bond between carbon fibre reinforced polymer bars (CFRP) and concrete. Construction and Building Materials, 94, 589–604. Cen, H., Kang, Y., Lei, Z., Qin, Q., & Qiu, W. (2006). Micromechanics analysis of Kevlar-29 aramid fiber and epoxy resin microdroplet composite by Micro-Raman spectroscopy. Composite Structures, 75, 532–538. Chocron Benloulo, I., Rodríguez, J., Martínez, M., & Sánchez Gálvez, V. (1997). Dynamic tensile testing of aramid and polyethylene fiber composites. International Journal of Impact Engineering, 19, 135–146. Deshpande, A.B. (2006). Characterization of CFRP and GFRP composite materials at high strain rate tensile loading. (Doctoral dissertation, Department of Mechanical Engineering, Wichita State University, USA). Destrebecq, J.-F., Toussaint, E., & Ferrier, E. (2011). Analysis of cracks and deformations in a full scale reinforced concrete beam using a digital image correlation technique. Experimental Mechanics, 51, 879–890. Gao, G., Huang, S., Xia, K., & Li, Z. (2014). Application of digital image correlation (DIC) in dynamic notched semi-circular bend (NSCB) tests. Experimental Mechanics, 55, 95–104. Giancane, S., Panella, F., Nobile, R., & Dattoma, V. (2010). Fatigue damage evolution of fiber reinforced composites with digital image correlation analysis. Procedia Engineering, 2, 1307–1315. Greenhalgh, E. (2009). Failure analysis and fractography of polymer composites. Cambridge: Woodhead Publishing. Gupta, V., Drzal, L., Lee, C.C., & Rich, M. (1985). The temperature-dependence of some mechanical properties of a cured epoxy resin system. Polymer Engineering & Science., 25, 812–823. Harding, J., & Welsh, L.M. (1983). A tensile testing technique for fibrereinforced composites at impact rates of strain. Journal of Materials Science, 18, 1810–1826.

34 

  X. Zhang et al.

Hawileh, R.A., Abu-Obeidah, A., Abdalla, J.A., & Al-Tamimi, A. (2015). Temperature effect on the mechanical properties of carbon, glass and carbon–glass FRP laminates. Construction and Building Materials, 75, 342–348. Hui, C.Y., Phoenix, S.L., & Shia, D. (1998). The single-filament-composite test: a new statistical theory for estimating the interfacial shear strength and Weibull parameters for fiber strength. Composites Science and Technology, 57, 1707–1725. Koerber, H., Xavier, J., & Camanho, P. (2010). High strain rate characterisation of unidirectional carbon-epoxy IM7-8552 in transverse compression and in-plane shear using digital image correlation. Mechanics of Materials, 42, 1004–1019. Lekou, D.J., & Philippidis, T.P. (2008). Mechanical property variability in FRP laminates and its effect on failure prediction. Composites Part B: Engineering, 39, 1247–1256. Lifshitz, J.M., & Rotem, A. (1970). Longitudinal strength of unidirectional fibrous composites. DTIC Document. Haifa: IIsrael Institute of Technology. Matthews, F., & Rawlings, R.D. (1999). Composite materials: Engineering and science. Cambridge: Woodhead Publishing. Melin, L.G., & Asp, L.E. (1999). Effects of strain rate on transverse tension properties of a carbon/epoxy composite: Studied by moiré photography. Composites Part A: Applied Science and Manufacturing, 30, 305–316. Naik, N., Yernamma, P., Thoram, N., Gadipatri, R., & Kavala, V. (2010). High strain rate tensile behavior of woven fabric E-glass/epoxy composite. Polymer Testing, 29, 14–22. Pan, B., Qian, K., Xie, H., & Asundi, A. (2009). Two-dimensional digital image correlation for in-plane displacement and strain measurement: A review. Measurement Science and Technology, 20(6), 062001. Pessiki, S., Harries, K.A., Kestner, J.T., Sause, R., & Ricles, J.M. (2001). Axial behavior of reinforced concrete columns confined with FRP jackets. Journal of Composites for Construction, 5, 237–245. Ramroth, W., Asaro, R., Zhu, B., & Krysl, P. (2006). Finite element modelling of fire degraded FRP composite panels using a rate dependent constitutive model. Composites Part A: Applied Science and Manufacturing, 37, 1015– 1023. Reis, J., Coelho, J., Monteiro, A., & da Costa Mattos, H. (2012). Tensile behavior of glass/epoxy laminates at varying strain rates and temperatures. Composites Part B: Engineering., 43, 2041–2046. Rodriguez, J., Chocron, I., Martinez, M., & Sánchez-Gálvez, V. (1996). High strain rate properties of aramid and polyethylene woven fabric composites. Composites Part B: Engineering, 27, 147–154. Shah, S., & Chandra Kishen, J. (2011). Fracture properties of concrete– concrete interfaces using digital image correlation. Experimental Mechanics, 51, 303–313. Shan, X., Huang, C., Yang, H., Wu, Z., Li, J., Huang, R., & Li, L. (2015). The thermal expansion and tensile properties of nanofiber-ZrW2O8 reinforced epoxy resin nanocomposites. Physics Procedia, 67, 1056–1061.

Shokrieh, M.M., & Omidi, M.J. (2009). Tension behavior of unidirectional glass/epoxy composites under different strain rates. Composite Structures, 88, 595–601. Singh, T.J., & Samanta, S. (2015). Characterization of Kevlar fiber and its composites: A review. Materials Today: Proceedings, 2, 1381–1387. Sutton, M., Wolters, W., Peters, W., Ranson, W., & McNeill, S. (1983). Determination of displacements using an improved digital correlation method. Image and Vision Computing, 1, 133–139. Tann, D.B., & Delpak, R. (2004). Influences of freeze and thaw cycles and elevated temperature on the properties of FRP composites. In L.C. Hollaway, M.K. Chryssanthopoulos, & S.S.J. Moy (Eds.), Advanced polymer composites for structural applications in construction (pp. 611–618). Cambridge: Woodhead Publishing. Wang, Z., & Xia, Y. (1998). Experimental evaluation of the strength distribution of fibers under high strain rates by bimodal Weibull distribution. Composites Science and Technology, 57, 1599–1607. Wang, Y., & Xia, Y.M. (1999). Experimental and theoretical study on the strain rate and temperature dependence of mechanical behaviour of Kevlar fibre. Composites Part A: Applied Science and Manufacturing, 30, 1251–1257. Wang, Y., & Xia, Y. (2000). A modified constitutive equation for unidirectional composites under tensile impact and the dynamic tensile properties of KFRP. Composites Science and Technology, 60, 591–596. Welsh, L., & Harding, J. (1985). Effect of strain rate on the tensile failure of woven reinforced polyester resin composites. Le Journal de Physique Colloques, 46, C5-405–C5-414. Xia, C., Shi, S.Q., & Cai, L. (2015). Vacuum-assisted resin infusion (VARI) and hot pressing for CaCO3 nanoparticle treated kenaf fiber reinforced composites. Composites Part B: Engineering, 78, 138–143. Yang, H., Song, H., & Zhang, S. (2015). Experimental investigation of the behavior of aramid fiber reinforced polymer confined concrete subjected to high strain-rate compression. Construction and Building Materials, 95, 143–151. Zhou, H., & Attard, T.L. (2013). Rehabilitation and strength sustainability of fatigue damaged concrete–encased steel flexural members using a newly developed polymeric carbon–fiber composite. Composites Part B: Engineering, 45, 1091–1103. Zhou, Y., Wang, Y., & Mallick, P.K. (2004). An experimental study on the tensile behavior of Kevlar fiber reinforced aluminum laminates at high strain rates. Materials Science and Engineering: A, 381, 355–362. Zhu, D., Mobasher, B., & Rajan, S.D. (2010). Dynamic tensile testing of Kevlar 49 fabrics. Journal of Materials in Civil Engineering, 23, 230–239. Zhu, D., Mobasher, B., Erni, J., Bansal, S., & Rajan, S.D. (2012). Strain rate and gage length effects on tensile behavior of Kevlar 49 single yarn. Composites Part A: Applied Science and Manufacturing, 43, 2021–2029.