Experimental study on strain-rate-dependent behavior and failure modes of long glass fiber-reinforced polypropylene composite

Abstract

Due to its good mechanical performances and design flexibility, long glass fiber-reinforced polypropylene (hereinafter referred to as LGFRP) composite has been increasingly used in the automotive industry, in which the LGFRP components are likely to sustain different strain rates loading during a crash event. The objectives of this study are to investigate the correlations between the LGFRP and strain rates 10⁻³s⁻¹ to 50 s⁻¹, and the corresponding failure modes of LGFRP. Therefore, tensile and compression tests are conducted at different strain rates and the corresponding microstructures of the specimens are investigated with scanning electron microscope. The experimental results show that the failure strain and ultimate strength increase as increasing strain rate. The elastic modulus is sensitive to strain rate in tensile tests, but less sensitive to strain rate in compression tests. The main failure modes of the specimens are the matrix crack and fiber pull-out. The defects such as bubbles, shrinkage cavities, or dry fibers of the specimens play important roles in the initiation and propagation of cracks during the tensile and compression tests.

Keywords

Long glass fiber scanning electron microscope failure modes thermoplastic resin

Introduction

Composites as novel materials have been increasingly used in aerospace, transportation, defense, sport, medical equipments and consumer products because of their advantages of light weight, high strength/weight, stiffness/weight, good vibration attenuation effect, design flexibility act.¹ As one class of typical composites, glass fiber-reinforced polypropylene (GFRP) has been widely adopted in automobile industry for lightweight design. A composite bus was developed using woven glass/polypropylene composites by Tillotson Pearson Inc., which led to over 30% weight reduction compared with conventional metallic bus.² Glass fibers/polypropylene is one kind of GFRP, which offers the potential for rapid manufacturing with low cycle times that allows for medium to high volume production. Moreover, glass fibers and polypropylene can be recycled.³ Therefore, glass fibers/polypropylene has been extensively adopted in vehicle components, such as roof door, body panel, frame segment, floor segments, battery access door, and seating systems with a weight reduction ranging from 40% to 60%, whilst maintaining the same or even better performance compared to conventional metallic components.^4–8 Some of these components are likely to sustain rate loading during a crash event. Moreover, accurate material model under different strain rates ( $\overset{\cdot}{ɛ}$ ) is important for numerical simulations in vehicle design and crashworthiness assessment.⁹ Therefore, it is necessary to investigate the effect of strain rate on the performance of GFRP.

The effect of strain rate on the mechanical properties of GFRP was investigated during several previous studies. The strength, fracture strain, and the impact strength of the GFRP increase with increasing strain rate, which is confirmed by literatures.^10,11 Reis et al.¹⁰ reported dynamic tensile tests for GFRP and observed that the mechanical properties of the GFRP show a strain-rate dependency. Taniguchi et al.¹¹ observed that the strength of GFRP made of T-glass increases with increasing strain rate. Papadakis et al.^12,13 investigated the effect of strain rate on the tensile and shear properties of a continuous glass fibers and polypropylene at various crosshead speeds. They observed that the Young’s modulus increases with increasing loading rate whilst the ultimate strength ( $ɛ_{b}$ ) decreases. The shear strength was reported to increase while the shear modulus decreases with increasing strain rate. Bai et al.¹⁴ studied the effect of strain rate on tensile properties of glass bead/high-density polyethylene composites. Both the Young’s modulus and tensile strength increase with increasing strain rate. Serban et al.¹⁵ investigated the influence of strain rate on the tensile properties of a PA 12-based polymer composite materials at strain rate ranging from $0.00028 s^{- 1}$ to $9.4 s^{- 1}$ . They noted that the tensile strength increases with increasing strain rate. Bonnet¹⁶ reported on the characterization of the mechanical properties of glass/polypropylene commingled woven fabric unbalanced 4:1 weave composite and a short GFRP composite at strain rate ranging from $1 s^{- 1}$ to $100 s^{- 1}$ . The results of weave composite show that the average tensile and shear modulus decrease with increasing strain rate while the tensile and shear strength increase. The polypropylene composite exhibits marginal strain rate sensitivity. There is a high level of data scatter which is partly attributed to the test setup and inertial effects.

Published papers on the mechanisms of composite failure at low and high strain rates are rarely seen. Yuan et al.¹⁷ studied impact compressive failure of GFRP unidirectional composites. The research results have shown that GFRP exhibited ductile failure for lower fiber volume fractions, but brittle failure for higher fiber volume fractions. As the temperature increased, the failure mode changed from kinking to micro-buckling. Experiments showed that the strain rate has a strong effect on the compressive strength. Suvorova¹⁸ pointed out that changing in loading rates affects the final failure modes of composite structures. Wang and Cho¹⁹ investigated the failure modes of satin weave E-glass/epoxy composite laminates. The experimental results show that the failure mode changes from splitting follows by fiber kink buckling to predominantly de-lamination and shear fracture as strain rates increases from quasi-static to high strain rates. Zhou and Zhang²⁰ researched the failure behavior of CFRP composites at high strain rates. The research shows that the main way of the dynamic compressive failures in the normal direction is shear failure due to the brittle fracture of the fiber. The results also indicate interfacial crack and de-lamination can be produced in the in-plane direction. Zhou et al.²¹ investigated the dynamic mechanical behavior of fiber-reinforced polymer composites embedded with ZnO whiskers. The investigation shows that the tensile properties of the composites are dominated by pull-out and fracture of fiber bundles. So far, according to the literatures reviewed by the authors, there is not much information that is available on the failure of long GFRP (LGFRP) under different strain rates.

Although some investigations on strain-rate-dependent behavior and failure modes of composites are reported in the above literatures, most of the above studies focus on thermosetting materials and few investigations focus on glass fibers/polypropylene composite, especially on LGFRP. LGFRP, as a new composite with advantages of favorable mechanical and shaping properties applied in vehicle industry, is important to investigate the effect of strain rate on the mechanical properties of LGFRP for vehicle design and crashworthiness assessment. Furthermore, the research of failure modes related to $\overset{\cdot}{ɛ}$ for LGFRP is also a meaningful task for the engineering application. This paper presents a systematic study of the effects of strain rate on the response of LGFRP in tensile and compression loading, and the micro- and macroscale failure modes are discussed.

Experimental setup

Material and fabrication process

The LGFRP is manufactured by air laying web technology in this paper. In the air lay process, the fibers are suspended in an air stream and then blown or forced into a continuously moving belt where the web is formed. So air laying web technology can be used to produce the three-dimensional structure of fiber mat, in which the fiber arrangement is non-direction as shown in Figure 1(a). So the strength ratio of longitudinal and transverse for the product made of the fiber mat can reach 1:1. The LGFRP used in this study is a commingled long E-glass fiber and polypropylene fiber composite, balanced 1:1 combination mat with a nominal 20% E-glass fiber volume fraction (Table 1). Figure 1 shows that the fiber mat of LGFRP, in which the specification of the fiber is 900 dtex, 40 to 50 mm length. In Figure 1(b), the white fibers are the long E-glass fibers and the blank fibers are the polypropylene fibers.

Table 1.

The materials composition of LGFRP.

Material	Length	Volume fractions	Specification	Plastication temperature
E-glass fiber	40–50 mm	20%	900 dtex	–
Polypropylene	–	80%	–	200℃

The preparation process of the experimental materials is as follows. Pre-consolidated combination mat is stacked ply-by-ply in an infrared oven, preheated at a temperature of 200℃. And the time for heat preservation is 2 min. The stack is then rapidly transferred to a preheated tool (70–80℃) which is installed in a stroking servo-hydraulic press. The tool is closed and pressure holding time is 30 s. Then, the composite plates and cylindrical specimens are obtained.

Specimens and test procedure

The dumbbell tensile specimens are cut from the flat planes, as shown in Figure 2. The thickness and density are $1.7 mm$ and $0.00178 {g / mm}^{3}$ , respectively. The compression specimen is cylindrical as shown in Figure 3. The diameter and height of compression specimen are $15 mm$ and $22.5 mm$ . Its density is the same as that of the tensile specimen.

Both quasi-static tests and dynamic tensile tests are conducted in a fatigue testing machine (MTS 647 Hydraulic Wedge Grip). The speed v of clamp is set according to the strain rate. In order to prevent the fracture of specimens in chuck position in tensile tests, the clamping parts of the specimens are reinforced by aluminum alloy plates with 1 mm thickness. The displacement response of tensile specimen is measured by an extensometer with 20 mm gauge length. To alleviate the friction between the compression specimen and fixture, butter is coated on both ends of the specimen. The tests in the same conditions are repeated five or more than five times to obtain the reliable experimental data. And then a revised stress–stain curve is obtained by taking the average value of the three similar stress–strain curves.

Phenom desktop microscope with the largest magnification 24,000 times is used to investigate failure mechanisms of the fractured composite specimens. For optical microscopic analysis, sample sections are cut from the failure region of the fractured specimens.

Results and discussion

Tensile experiments

Strain–stress curves

After processing the tensile experimental data, the stress–strain curves of LGFRP at different strain rates are obtained as shown in Figure 4. Under strain rates $0.001 s^{- 1}$ and $0.01 s^{- 1}$ , the material stress–strain responses are non-linear with premature yielding followed by maximum peak stress and strain-softening. At strain rates $0.1 s^{- 1}$ , $1 s^{- 1}$ , $10 s^{- 1}$ , and $50 s^{- 1}$ , the stress–strain responses are approximately linear elastic up to the maximum stress point followed by abrupt failure. The material stress–strain responses under different strain rates may be attributed to the time dependent. At lower strain rates, the specimens have time to distribute the load and undergo steady damage accumulation, therefore the material stress–strain responses with strain-softening under strain rates $0.001 s^{- 1}$ and $0.01 s^{- 1}$ are easily observed. However, at high strain rates, the specimens have less time to respond, thus the material stress–strain responses without strain-softening.

The elasticity modulus (E), ultimate strengths (

σ_{b}

), and failure strains (

ɛ_{b}

) of tensile test at different strain rate are presented in Table 2. The tensile ultimate strengths increases from 57.08 MPa to 121.95 MPa and failure strains increase from 0.018 to 0.045 with increasing strain rate (0.001 s⁻¹–50 s⁻¹). The ultimate strength is relatively insensitive to changes in strain rate (

0.001 s^{- 1}

–

0.01 s^{- 1}

). However, the failure strains increase significantly up to 150% from

0.01 s^{- 1}

50 s^{- 1}

. The ultimate strengths follows a similar trend with an increase of 113.65% from

0.001 s^{- 1}

50 s^{- 1}

. From

0.1 s^{- 1}

50 s^{- 1}

, the elasticity modulus increases from 2100 MPa to 3299 MPa with increasing strain rate. The effect of strain rate on elasticity modulus, ultimate strengths, and failure strains may be attributed to a combination of several factors including the visco-elastic nature of the polymeric matrix, the performance and arrangement of glass fiber, the interface between the matrix and reinforced fiber, etc.⁹ Furthermore, the strain rate sensitivity of the tensile strength can also be attributed to the strain rate dependence of the glass fibers in fiber-dominated loading mode, as suggested by Okoli.⁹ In contrast, the elastic modulus decreases from 6522.479 MPa to 2100 MPa with increasing strain rate (

0.001 s^{- 1}

–

0.1 s^{- 1}

). Therefore, the relationship of the elastic modulus and strain rate is complicated. From the queried literatures, few investigations focus on the relationship. The relationship may be related to the spread of stress wave in the material. Further detailed study is required to explain the relationship between elastic modulus and strain rate.

Table 2.

Tensile properties for LGFRP at different strain rates.

$\overset{\cdot}{ɛ}$ ( $s^{- 1}$ )	E ( $MPa$ )	$σ_{b}$ ( $MPa$ )	$ɛ_{b}$
0.001	6522.48	57.08	0.018
0.01	6504.12	66.21	0.018
0.1	2100	77.2	0.041
1	2200	82.64	0.042
10	2800	104.85	0.043
50	3299	121.95	0.045

Failure modes analysis

Under strain rates $0.001 s^{- 1}$ and $0.01 s^{- 1}$ , damage occurs around a local through-thickness fracture plane perpendicular to loading direction. However, a local through-thickness shear fracture plane at an angle of about $45^{\circ}$ to loading direction becomes more and more obvious with the increasing strain rate ( $0.1 s^{- 1}$ – $50 s^{- 1}$ ) as shown in Figure 5. This may be attributed to the nature of LGFRP. The shear failure is becoming more evident with increasing strain rate.

In the preparation of the experimental specimens, the PP fibers are plasticized to viscous state and are forced into the pores among the glass fibers by outer pressure. Then, the specimens are cooled to room temperature. In this process, some defects (as shown in Figure 6(c②)) such as bubbles, shrinkage cavities, or dry fibers may emerge. These defects may have a great influence on the properties of the material.

Based on the experimental stress–strain curves as shown in Figure 4, all the curves can be divided into three phases as shown in Figure 6(a), elastic phase indicated by I, initiation of fracture phase indicated by II, and propagation of fracture phase indicated by III. Therefore, the schematic representation of the three phases is used to analysis the failure process as shown in Figure 6(b). In the first phase, the strain is very small and the stress is linear with strain. The inner defects are no change due to the low stress. In the second phase, the elastic deformation energy increase with the increasing load. There is no evident damage in the specimens. Nevertheless, in microstructure analysis, the defects may grow and matrix cracks, even local de-bonding (Figure 6(c②)) may emerge. Due to the uneven of material property and stress distribution, the defects always firstly generate in the weakened area where stress concentration is greater. Then, the defects may generate in the other area due to “synergistic effect,” which indicates that stress concentration area transfers caused by the stress redistributes. In the third phase, when the load exceeds the critical load of crack propagation, the local defects converge, and cracks propagate at the same time. Then, the macroscopic damage can be seen, including matrix cracking, fiber or fiber bundle pull-out, and fiber breakage. As shown in Figure 6(a), the stress–strain curves decline rapidly when the load exceeds the critical load at high strain rate. Thus, the crack mechanism of specimens is brittle crack. The cause may lie in that the crack converge and propagate very rapidly.

Now, the authors discuss the failure modes (pull-out or rupture) of glass fibers. The factors influences failure modes of glass fibers include the fiber/matrix interfacial strength, the existing form of the fibers in the matrix (fiber axial and the loading direction is parallel or at an angle, fibers hook together), the fiber dispersion, and uniformity. It can be made the following analysis. The tensile breaking force F_b of glass fiber can be calculated by equation (1)

F_{b} = π r^{2} σ_{p}

(1)

where

σ_{p}

is the ultimate strength of the fiber, and r is the glass fiber radius.

The pull-out force of the fiber F_p can be calculated by equation (2)

F_{p} = 2 π rl σ_{m}

(2)

where

σ_{m}

represents the bonding strength of the interface between fiber and matrix, and l represents the effective bonding length of the fiber in the matrix.

Setting the $F_{b} = F_{p}$ , the expression of the critical length of fiber pull-out (l_c) can be obtained, as

l_{c} = \frac{r σ_{p}}{2 σ_{m}}

(3)

Therefore, the glass fiber will be broken when the embedded length of the fiber is greater than l_c, or the fiber will be pulled out. $r = 13 μ m$ , $σ_{p} = 4 GPa$ , $σ_{m} = 1.2 MPa$ of the LGFRP in this paper, the critical length of fiber pull-out is 22 mm calculated by equation (3). In Figure 7, $l'$ represents the length of glass fiber, l₁ and l₂ are the bonding lengths of the fiber in the matrix. In case of $l_{1} > l_{2}$ , the fiber is pulled out from matrix B, l₂ is called the effective bonding length of the glass fiber in the matrix. From Figure 7, it is obtained that the maximum effective bonding length of the glass fiber is $\frac{l'}{2}$ . In this study, the length of glass fibers is 40 to 50 mm; therefore, the maximum effective bonding length of the glass fiber is 20 to 25 mm. Because the value of critical length of glass fiber is between 20 mm and 25 mm, the failure modes of fibers contain two forms of fiber pull-out and fiber breakage. Actually, the length of fibers is not a unique value, but instead the length of a few fibers is greater than l_c in this study. Meanwhile, the adhesion strength of the resin and fibers is weak as shown in Figure 8(a), there is little resin on the pulled out fiber. Therefore, the main failure mode of the material is fiber pull-out as shown in Figure 8(b).

Optical micrographs of the damage region in the fractured tensile specimens are shown in Figure 8. The smooth hollow in the matrix left by the fiber pull-out can be observed shown in Figure 8(c). The end face of the pulled out fibers are flat shown in Figure 8(d), but the fibers with flat end face are not necessarily pulled out, may be fractured (when the embedded length of the fiber is greater than l_c). Figure 8(d) shows that the rugged end face of the fibers can be concluded that the failure mode of the fibers is fracture.

Compression experiments

Strain–stress curves

Typical compression stress–strain plots under different strain rates are shown in Figure 9. Under the strain rates from $0.001 s^{- 1}$ to $50 s^{- 1}$ , the stress–strain responses are characterized by an initial elastic region, then yielding followed by hardening up to the peak stress. Failure occurs in both quasi-static and dynamic tests at strain of about 0.3 to 0.4. The elastic modulus, ultimate strengths, and failure strains of compression test at different strain rate are presented in Table 3. The compression ultimate strengths increase from 70.89 MPa to 94.81 MPa and ultimate strengths increase from 0.325 to 0.405 with increasing strain rate (0.001 s⁻¹–50 s⁻¹). The elastic modulus decreases from 268 MPa to 240 MPa with increasing strain rate from $0.001 s^{- 1}$ to $1 s^{- 1}$ . In contrast, the elastic modulus increases from 240 MPa to 275 MPa with increasing strain rate from $1 s^{- 1}$ to $50 s^{- 1}$ . The elastic modulus of the specimens is little difference between quasi-static and dynamic tests. Therefore, the elastic modulus is less sensitive to strain rates in compression tests. From quasi-static to $50 s^{- 1}$ , the average compression ultimate strengths and failure strain increase by dramatic 33.7% and 24.6%, respectively. The increasing of elastic modulus, ultimate strengths, and failure strains in compression may be attributed to the strain rate dependence of the glass fibers in fiber-dominated loading mode, as suggested by Okoli,⁹ the time dependent nature of damage accumulation²² and the viscoelastic effects in the matrix material.⁹

Table 3.

Compression properties for LGFRP at different strain rates.

$\overset{\cdot}{ɛ}$ ( $s^{- 1}$ )	E ( $MPa$ )	$σ_{b}$ ( $MPa$ )	$ɛ_{b}$
0.001	268	70.89	0.325
0.01	260	72.279	0.33
0.1	250	80.3	0.381
1	240	88.84	0.397
10	255	90.93	0.399
50	275	94.81	0.405

Failure modes analysis

The different failure modes of compression specimens are shown in Figure 10. Under strain rates $0.001 s^{- 1}$ and $0.01 s^{- 1}$ , the failure zone is dispersive. On the contrary, a local through-thickness shear fracture plane at an angle of about $45^{\circ}$ to the axis of loading becomes more and more obvious with increasing strain rate ( $0.1 s^{- 1}$ – $50 s^{- 1}$ ). The reason may be that fibers in the matrix are messy, and the shear failure initiation across an axial fiber bundle becomes more evident with increasing strain rate.

Just like the tensile stress–strain curves, the compressive stress–strain curves is composite of three segments, as shown in Figure 11, which indicates the three phase in the process of compression, namely elastic phase indicated by I, densification and initiation of fracture phase indicated by II, and propagation of fracture phase indicated by III. For all the three phases, the stress is approximately linear with strain. In the first phase, the glass fiber, matrix, and defects change little, and deformation of the specimen can recover (as shown in Figure 11(b), I phase) when unloading. In the densification stage, although partial defects such as shrinkage cavity are compressed, the specimens are not completely densified. At the same time, the matrix and glass fiber around defects begin to deform. The differences of Poisson’s ratio and the elastic modulus between glass fiber and matrix cause stress concentration, which leads to the initial cracks. This is the second phase, which is shown in Figure 11(b) II phase. In the third phase, when the load exceeds the critical load of crack propagation, the disperse defects continue expanding, meanwhile the adjacent cracks converge. Then, macroscopic damage such as matrix fracture or fiber pull-out can be seen, which can lead to the final failure of the specimens. The failure mode of the compression specimens is brittle fracture, which is the same as that of the tensile specimens. In other words, the convergence and propagation of cracks are very rapidly in the third phase.

Although cracks take place in many places of composite specimens on the microlevel, there are usually only several cracks on the macro level. Through matching two sections, the crack initiation zone and extension direction can be determined. When the load is applied to composite material, the initiation and extension of the crack will be stopped by fibers, and the cracks tend to develop along the interface between the fiber and matrix, which leads to the formation of many bifurcate micro cracks. Therefore, the initiation zone of damage exists in the crack initiation zone. The information about effect of defects on crack initiation can be provided by the scanning electric microscope observation. Finally, through the analysis of the macro to micro and according to the previously discussed basic failure modes, the main failure mode with matrix cracking, fibers, or fiber bundle pull-out of compression specimens can be determined.

Comparison and analysis the results of tensile and compression experiments

The ultimate strengths and failure strains of the tensile and compression specimens are all very sensitive to strain rate as shown in earlier subsections. The elastic modulus of tensile specimens decrease from 6522.479 MPa to 2100 MPa with increasing strain rate between $0.001 s^{- 1}$ and $0.1 s^{- 1}$ . The tensile modulus increases from 2100 MPa to 3299 MPa with increasing strain rate (0.1 s⁻¹–50 s⁻¹). The elastic modulus of compression specimens decreases from 268 MPa to 240 MPa with increasing strain rate between $0.001 s^{- 1}$ and $1 s^{- 1}$ . The elastic modulus of compression specimens increases from 240 MPa to 275 MPa with increasing strain rate ( $1 s^{- 1}$ – $50 s^{- 1}$ ). In the strain rate range of 0.001 s^-1 to 50 s⁻¹, the maximum elastic modulus of tensile specimens is 2.1 times higher than the minimum elastic modulus, and the maximum elastic modulus of compression specimens is 0.146 times higher than the minimum elastic modulus. Therefore, the elastic modulus of tensile specimens is more sensitive than that of compression specimens to strain rate. Tables 2 and 3 show that elastic modulus, ultimate strengths, and failure strains of the tensile test results are up to 25, 1.29, and 0.11 times that of the compression test results at the strain rates from $0.001 s^{- 1}$ to $50 s^{- 1}$ . Consequently, the elastic modulus of the tensile specimens are much larger than that of the compression specimens and the ultimate strengths of the tensile specimens are slightly larger than that of the compression specimens under the same strain rates. This may be attributed to the compressive capacity which is worse than the tensile capacity of glass fiber. In contrast, the failure strains of the tensile specimens are smaller than that of the compression specimens under the same strain rates. It can be found from Figures 4 and 9, at strain rates $0.001 s^{- 1}$ and $0.01 s^{- 1}$ , the stress–strain responses of tensile specimens and compression specimens with strain-softening are all obvious. The local through-thickness shear fracture plane at an angle of about $90^{\circ}$ to the axis of loading under the strain rates $0.001 s^{- 1}$ and $0.01 s^{- 1}$ . Furthermore, the local through-thickness shear fracture plane at an angle of about $45^{\circ}$ to the axis of loading becomes more and more obvious with increasing strain rate $0.1 s^{- 1} - 50 s^{- 1}$ as shown in Figures 5 and 10. It is mainly because of that the failure modes of the specimens are similar to brittle failure with increasing strain rate. Figures 6 and 11 show that the breakages of the matrix with bubbles, shrinkage and other defects firstly arise after the ultimate stress. And when the carrying capacities of the specimens begin to weaken, part of the fibers are pulled out and fractured with increasing the load, until all the fibers of fracture surface are pulled out or fractured. The major macroscopic damage mechanisms of tensile and compression specimens include matrix cracking, fiber, or fiber bundle pull-out.

Conclusion

In the present work, the properties and failure modes of LGFRP composites were investigated for a range of strain rate from $0.001 s^{- 1}$ to $50 s^{- 1}$ . The experimental results showed that the tensile and compression ultimate strengths and failure strains increase with increasing strain rate. However, the tensile and compression modulus decrease in the range of strain rate from $0.001 s^{- 1}$ to $0.1 s^{- 1}$ and increase in the range of strain rate from $0.1 s^{- 1}$ to $50 s^{- 1}$ . Further detailed study is required to explain the relationship between the elastic modulus of this material and strain rate effects. Meanwhile, the failure modes of the specimens were analyzed, which indicated that the specimen failures are mainly induced by the matrix crack and the fiber pull-out. The defects such as bubbles, shrinkage cavities, or dry fibers of the specimens play important roles in the initiation and propagation of cracks during the tensile and compression tests.

Footnotes

Conflict of Interest

None declared.

Funding

This work is supported by Open Research Fund Program of the State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body (Grant No. 31115014) and the National Natural Science Foundation of China (Grant No. 51405150).

References

Duan

Tao

Han

. Investigation on structure optimization of crashworthiness of fiber reinforced polymers materials. Compos Part B 2014; 60: 471–478.

Vaidya

Samalot

Pillay

. Design and manufacture of woven reinforced glass/polypropylene composites for mass transit floor structure. J Compos Mater 2004; 38: 1949–1972.

Tornqvist R and Baser B. Structural modules with improved crash performance using thermoplastic composites. Plastics for the Automotive Industry 2002; 147–158.

Ning

Pillay

Vaidya

. Design and development of thermoplastic composite roof door for mass transit bus. Mater Des 2009; 30: 983–991.

Janowski

Vaidya

Husman

. Thermoplastic sandwich structure design and manufacturing for the body panel of mass transit vehicle. Compos Struct 2007; 80: 82–91.

Ning

Vaidya

Janowski

. Design, manufacture and analysis of a thermoplastic composite frame structure for mass transit. Compos Struct 2007; 80: 105–116.

Thattaiparthasarthy

Pillay

Ning

. Process simulation, design and manufacturing of a long fiber thermoplastic composite for mass transit application. Compos Part A: Appl Sci Manuf 2008; 39: 1512–1521.

Bartus

Vaidya

Ulven

. Design and development of a long fiber thermoplastic bus seat. J Thermoplast Compos Mater 2006; 19: 131–154.

Brown

Brooks

Warrior

. The static and high strain rate behaviour of a commingled E-glass/polypropylene woven fabric composite. Compos Sci Technol 2010; 70: 272–283.

10.

Reis

JML

Coelho

JLV

Monteiro

. Tensile behavior of glass/epoxy laminates at varying strain rates and temperatures. Compos Part B 2012; 43: 2041–2046.

11.

Gerlach R, Siviour CR, Wiegand J, et al. The strain rate dependent material behavior of S-GFRP extracted from GLARE. Mech Adv Mater Struc 2013; 7: 505–514.

12.

Papadakis N, Reynolds N, Pharaoh M, et al. Strain rate dependency of the shear properties of a highly oriented thermoplastic composite material using a contacting displacement measurement methodology - Part B: shear damage evolution. Compos Sci Technol 2004; 64: 739–748.

13.

Papadakis

Reynolds

Pharaoh

. Strain rate effects on the shear mechanical properties of a highly oriented thermoplastic composite material using a contacting displacement measurement methodology—part a: elasticity and shear strength. Compos Sci Technol 2004; 64: 729–738.

14.

Bai

S-L

Cao

Chen

J-K

. Tensile properties of rigid glass bead/hdpe composites. Polym Polym Compos 2000; 8: 413–418.

15.

Serban

Weber

Liviu

. Tensile properties of semi-crystalline thermoplastic polymers: effects of temperature and strain rates. Polym Test 2013; 32: 413–425.

16.

Bonnet B. Comportement au choc de matreriax composites pour applications automobiles. PhD Thesis, l’Ecole Nationale Supérieure des Mines de Paris, 2005.

17.

Yuan

Takeda

Waas

. Impact compressive failure of GFRP unidirectional composites. Sci Eng Compos Mater 2000; 9: 1–9.

18.

Suvorova JV, Ohlson NG, Alexeeva SI. Temperature influence in the description of time-dependent materials. Mater Design 2003; 4: 299–304.

19.

Wang

Cho

. Effect of compressive loading speed and stacking sequence on mechanical characteristics of satin weave E-glass/epoxy composites. Polym Compos 2012; 33: 1603–1612.

20.

Zhou W and Zhang X. Dynamic compressive response and failure behaviour of CFRP composites at high strain rates. Adv Mater Res 2011; (152–153): 988–991.

21.

Zhou W, Wang X, Zhang H, et al. Dynamic mechanical behavior of fiber reinforced polymer composites embedded with ZnO whiskers. Adv Mater Res 2011; (197–198): 1532–1535.

22.

Hosur

Alexander

Vaidya

. High strain rate compression response of carbon/epoxy laminate composites. Compos Struct 2001; 52: 405–417.