Static and dynamic crushing responses of CFRP

0 downloads 0 Views 6MB Size Report
Jan 9, 2017 - balls) were chosen and a series of static and dynamic tests were ... rebound (filled with aluminum honeycomb and plastic balls) and .... locity was chosen to be 2 m/s and an impactor of 19.02 kg weight fell .... 42.4/41.0 0.6.
Materials and Design 117 (2017) 396–408

Contents lists available at ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

Static and dynamic crushing responses of CFRP sandwich panels filled with different reinforced materials Yanqin Zhang a,b, Zhijian Zong a, Qiang Liu a,c,⁎, Jingbo Ma a, Yinghan Wu a, Qing Li d a

School of Engineering, Sun Yat-Sen University, Guangzhou City 510006, China Department of Mechanical and Electrical Engineering, Dongguan Polytechnic, Dongguan City 523808, China State Key Laboratory of Advanced Design and Manufacture for Vehicle Body, Hunan University, Changsha City 410082, China d School of Aerospace, Mechanical and Mechatronic Engineering, Sydney University, Sydney, NSW 2006, Australia b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Static and dynamic crash tests were conducted for sandwich panels with different cores. • Typical load-displacement curves and energy absorption behaviors were identified. • Specimens with Al honeycomb and plastic balls performed better energy absorption capabilities under static compression. • Specimens with Al honeycomb and EPP foam performed better energy absorption capabilities under dynamic impact.

a r t i c l e

i n f o

Article history: Received 9 August 2016 Received in revised form 16 November 2016 Accepted 7 January 2017 Available online 09 January 2017 Keywords: Crashworthiness CFRP Reinforced materials Dynamic response

⁎ Corresponding author. E-mail address: [email protected] (Q. Liu).

http://dx.doi.org/10.1016/j.matdes.2017.01.010 0264-1275/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t This study aims to investigate the crashworthiness of carbon fiber reinforced plastic (CFRP) sandwich panels filled with different reinforced materials under quasi-static compression and low velocity impact loading. Four lightweight filler materials (namely EPP foam, aluminum honeycomb, rubber foam balls and plastic hollow balls) were chosen and a series of static and dynamic tests were carried out to explore the damage mechanism, load bearing capacity, energy absorption and cushioning properties of these different sandwich panels. The complete core crushing contributed to the loading resistance and energy absorption under the static compression leaving the top and bottom CFRP facesheet intact. Three distinct load-displacement categories, classified as no rebound (unfilled), incomplete rebound (filled with aluminum honeycomb and plastic balls) and complete rebound (filled with EPP and rubber balls) were observed in the impact tests; the localized facesheet rupture and core crushing were dedicated to significant energy absorption during impact. The specimens filled with aluminum honeycomb and plastic hollow balls exhibited superior energy absorption capabilities (2.8 and 4.8 J/g, respectively) in the static compression testes, while the specimens filled with aluminum honeycomb and EPP foam exhibited superior capability of energy absorption (0.8 and 0.9 J/g, respectively) in the impact tests. © 2017 Elsevier Ltd. All rights reserved.

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

1. Introduction The growing effort to develop lightweight structures with better crashing performance has led to remarkable development and fairly extensive utilization of sandwich structures [1]. Typically a sandwich structure comprises a composite system featuring a lightweight core between two relatively thin high-strength facesheets or skins. Such sandwich structures have been widely used in numerous fields such as aerospace, marine, automotive, and defence industries attributable to their superior mechanical properties such as high ratio of specific stiffness to weight, excellent sound/thermal insulation, remarkable energy absorption capacity, etc. [2,3]. Nevertheless, sandwich structures are susceptible to static compression and impact loading during the service life [4], such as crushing of core filler and debonding between facesheet and core, which may cause significant property degradation and damage, severely compromising the structural integrity and functionality [2]. Structural performance of various sandwich panels has been studied extensively over the years. There has been a major concern on the effects of loading conditions (e.g. static compression, low velocity impact, medium/high velocity impact [5–7]), impactor shape and size [8,9]) as well as structural parameters (e.g. core material and thickness [10–13], facesheet type and thickness etc. [14,15]) on the crashing behaviors and damage mechanisms. As increased recognition to the critical role of core material, its selection has drawn significant attention [12] and various core materials have been considered for reinforcing sandwich panels. In literature, typically there are four groups of core materials, i.e. balsa woods, corrugated sheets [16] foam-like and honeycombs materials [17]. For example, Cesim and Cenk [18] adopted balsa wood and polyvinyl chloride (PVC) foam as core materials for studying the impact responses of sandwich panels. Pitarresi et al. [19] compared the single corrugation, double corrugation, dimpled, web and tubal cores for developing cost-effective crashworthy sandwich structures. Mamalis et al. [20] investigated the sandwich panels with four different polymeric foam cores (i.e. Polymetchacrylimide (PMI) foam, two grades of linear PVC foam and polyurethane foam) and two fiber reinforced plastics (FRP) facesheet laminates for characterizing the collapse modes and crushing characteristics through a series of edgewise compression tests. Wang et al. [12] compared the crash characteristics of sandwich panels with aluminum facesheets and five different core materials (low density balsa wood, high density balsa wood, cork, polypropylene honeycomb, and polystyrene foam) subjected to mediumvelocity impacts. Wang et al. [3] further studied the sandwich panels with polyurethane foam cores under low velocity impact, in which non-destructive inspection and destructive sectioning techniques were used to evaluate the internal and external damages in the sandwich panels. Schubel et al. [6] characterized the low-velocity crashing behaviors and damage mechanism for the sandwich panels with woven carbon/epoxy facesheets and a PVC foam core, finding that low-velocity impact was fairly similar to quasi-static counterpart except for localized damage. As a common filler material, honeycombs are very extensively used in the aerospace industry for their exceptional efficiency of high performance to lightweight and low cost. In this aspect, Meo et al. [5] investigated the low-velocity crashing responses, such as damage initiation, crack propagation, and failure mechanisms, for honeycomb sandwich panels. Hazizan et al. [21] explored the low-velocity crashing behaviors of aluminum honeycomb sandwich structures through drop-weight impact tests. Feraboli et al. [22] used a cylindrical pole to penetrate the composite sandwich panel comprised of carbon/ epoxy facesheets and a deep honeycomb core. Despite these continual attempts, challenge remains to select a proper core filler for achieving best possible overall mechanical properties and crashworthiness. As a relatively newer facesheet material, composites have exhibited substantially higher ratios of stiffness and strength to weight compared to metallic counterparts. The sandwich panels with plain weave laminated facesheets are becoming more and more prevalent attributable

397

to the fact that plain weave fabrics are of well-balanced ply properties and enhanced inter-laminar properties compared to unidirectional laminates [3]. In this regard, carbon fiber reinforced plastics (CFRP) composites have stood out to be an effective crashworthy material, offering an exceptional structural properties and weight advantage which have been extensively studied by Liu, Troiani, Jacob, Lavoie, etc. [23–27]. Much remained to be investigated to date is the crashing behaviors and failure mechanism of sandwich structures made of such CFRP composites. Despite the abovementioned studies in literature, there have been rare works concerning the characteristics of CFRP facesheets with different core materials under static and dynamic loading. This paper aims to provide an experimental study on the relative performance of different filler materials. The quasi-static compression tests were first conducted; the typical load-displacement curves and energy absorption behaviors were identified. Then, the crashing responses of CFRP sandwich panels, such as contact load-displacement curves, damage process, specific energy absorption (SEA) and acceleration-time curves, were explored in detail under a series of low-velocity impact tests. This study is expected to gain new understanding and provide some guide to the design of novel sandwich structures for enhancing crashworthiness. 2. Materials and methods 2.1. Specimens preparation CFRP has been increasingly applied in the lightweight body structures in electric vehicles (such as BMW i3) due to the restriction of energy storage in battery to travel distance. Fig. 1a exhibits the CFRP body structure in our recently developed lightweight EVs, in which most of the body components were fabricated in the CFRP materials, including the side body frame, roof, front and rear energy absorption apparatus. To improve the crashworthiness characteristics of the vehicle body, the space inside these components can be filled with lightweight reinforced materials; thus the baseline CFRP sandwich panels filled with different core materials are chosen to evaluate the reinforcement effectiveness. The preparation of specimens incorporates the facesheet panels and filled core materials (as shown in Fig. 1b). The Toray T300 plain weave carbon fiber/epoxy prepreg was used to fabricate the CFRP laminated faces by a hot press. One specific stacking sequence (thickness), i.e. [0/ 90]6 (1.2 mm in thickness), was considered for the laminated facesheets in this study. The average tensile strength and Young's modulus were 592 MPa and 57 GPa, respectively, which were obtained from the in-house mechanical tests following ASTM D3039-76. The fabricated laminates were then cut into square specimens with a side length of 100 mm, the overall thickness of 25 mm was chosen from top facesheet to bottom facesheet for accommodating the core filler (Fig. 1b). The adhesive film was used on the bonding surface of CFRP facesheets to keep free of any contaminants such as grease, oil, wax or mold release. The core material was bonded on the inner surface of facesheets using 3 M double sided adhesive tape, and then the whole sandwich panel specimen was placed in the pressing machine with the height limit block to precisely control the height of the sample and enhance the bonding effectiveness between the core and facesheets. In order to evaluate the interactive effect of different reinforcements on CFRP facesheets, four lightweight core filler materials with the same height were chosen, namely expanded polypropylene foam(EPP-F), aluminum honeycomb (Al-H), rubber foam balls (R-B) and plastic hollow balls (P-B), respectively (Fig. 1c). For the EPP-F material (Table 1), the density was 30 kg/m3; For the Al-H, the cell size of the aluminum honeycomb was 4 mm, the cell wall thickness and height was 0.1 mm and 25 mm respectively; For the rubber foam (R-B), the ball diameter was 25 mm, and the density was 1.8 × 104 kg/m3; For the plastic hollow ball (P-B), the ball diameter was 25 mm, and the ball wall thickness was 1 mm. An array with 16 balls was created between the top and

398

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

Fig. 1. Schematic of: (a) CFRP body, (b) sandwich specimen and (c) different cores.

bottom facesheets to reinforce the sandwich specimens of R-B and P-B. In addition, a set of unfilled specimens (U-F-S) were prepared for the comparison in the low velocity impact tests, which was supported by four steel columns in the corners to form the same height as other specimens. 2.2. Testing procedure According to the vehicle safety standard [28], there are two typical loading cases representing the quasi-static crushing and dynamic impact acting on vehicle body structures. Thus in this study, the quasi-static compression and low-velocity impact tests were conducted. The quasi-static compression tests were first performed in Sans 5305 hydraulically actuated testing machine, as shown in Fig. 2a, where the specimen was placed in between the support platen and loading platen without any fixture. The tests were carried out using displacement control of 2 mm/min and each specimen was crushed through a 20 mm stroke. The data (force and displacement) were recorded by an automatic data acquisition system directly. Low velocity impact tests were carried out in Instron Dynatup 8391 drop weight impact test machine, as shown in Fig. 2b. These tests were Table 1 Crashworthiness characteristics in quasi-static compressing tests. Specimen

Weight of specimen (g)

Weight of filled material (g)

Peak load (kN)

Energy absorption (J)

SEA (J/g)

EPP-F-SC-1 EPP-F-SC-2 EPP-F-SC-3 Average S.D. Al-H-SC-1 Al-H-SC-2 Al-H-SC-3 Average S.D. R-B-SC-1 R-B-SC-2 R-B-SC-3 Average S.D. P-B-SC-1 P-B-SC-2 P-B-SC-3 Average S.D.

43.1 43.1 43.7 43.3 0.3 50.3 50.9 50.7 50.6 0.3 166.9 167.1 165.3 166.4 1.0 67.3 67.7 67.3 67.4 0.2

6.9 6.9 6.9 6.9 0.0 13.8 13.8 13.8 13.8 0.0 129.9 129.9 129.9 129.9 0.0 31.3 31.3 31.3 31.3 0.0

4.4 4.4 4.5 4.4 0.1 20.4 19.4 20.5 20.1 0.6 47.0 51.0 51.8 49.9 2.6 47.3 47.9 51.5 48.9 2.3

30.0 28.2 31.1 29.8 1.5 142.8 139.3 146.9 143.0 3.8 97.7 98.2 98.7 98.2 0.5 312.7 321.9 328.1 320.9 7.7

0.7 0.7 0.7 0.7 0.0 2.8 2.7 2.9 2.8 0.1 0.6 0.6 0.6 0.6 0.0 4.6 4.8 4.9 4.8 0.1

Note that the specimen label, e.g. e.g. EPP-F-SC-1, represents the sandwich panels with EEP foam under static compression loading in Specimen #1 of this group.

conducted using an impactor with a hemispherical nose of 36 mm in diameter with rebound brake to avoid multiple impacts. The impact velocity was chosen to be 2 m/s and an impactor of 19.02 kg weight fell onto the specimens along two smooth column guides through the center hole of the pneumatic fixture of 76 mm in diameter. The responses of the impact test were collected from the data acquisition system in terms of load, time, energy, velocity and displacement. For each case, three specimens were tested and their crashing data were summarized in Tables 1 and 2, respectively. 2.3. Crashworthiness criteria Several key crashworthiness parameters were quantified to compare the performance and crushing behaviors of sandwich structures. The total energy absorbed in a crushing or impact process (following the standard in ASTM D3763-02 [29] for impact tests) is determined by integrating the area below the load-displacement curve, mathematically as, l

Ea ¼ ∫ 0 Pdl

ð1Þ

where Ea is the absorbed energy, P is the instantaneous crushing force and l is the crushing stroke or displacement. Specific energy absorption (SEA) is an important factor to measure the energy absorption capacity of different structures per unit mass, which is calculated by dividing the total energy absorption Ea by the mass of the structure. The mass m of specimen was mounted by the whole parts of sandwich panels in this study, as SEA ¼

Ea m

ð2Þ

The peak load Pmax is another key indicator to assess the crashworthiness of the structure which represents the initial peak crushing force in the quasi-static compression and low velocity impact tests. Note that Ei is the impact energy consisted of the initial kinetic energy and the potential energy of impactor, which could be determined by the highest point in energy-time curves [8]. 3. Results and discussion 3.1. Quasi-static compressing 3.1.1. Typical load-displacement curve The compression tests were conducted on four different filler CFRP sandwich specimens, and their load-displacement curves are depicted

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

399

Fig. 2. Test setups: (a) quasi-static compression, and (b) low velocity impact.

in Fig. 3. It can be seen that the load-displacement curve of sandwich panels filled with the aluminum honeycomb (Al-H-SC) differs with the other three counterparts. For the Al-H-SC specimen (the black curve in Fig. 3), the honeycomb deformed in a progressive folding fashion, where the load increased linearly to the peak value of 19.4 kN; after the peak, the load dropped dramatically to a lower but a plateau level; and followed with a steady fluctuation attributable to cell wall plastic deformation with wrinkling and fracture. The progressive crushing behavior can be typically characterized by the mean load, representing the buckling resistance under quasi-static compression. In this case, the average mean load of these three tested Al-H-SC specimens was 7.5 kN. For the sandwich panels filled with the rubber foam balls (R-B-SC, the red curve in Fig. 3), the curve represents the typical loadTable 2 Crashworthiness characteristics in low velocity impact tests. Specimen

Peak load (kN)

Maximum displacement (mm)

Maximum acceleration (m/s-2)

Ei/Ea(J)

SEA(J/g)

U-F-S-1 U-F-S-2 U-F-S-3 Average S.D. EPP-F-DI-1 EPP-F-DI-2 EPP-F-DI-3 Average S.D. Al-H-DI-1 Al-H-DI-2 Al-H-DI-3 Average S.D. R-B-DI-1 R-B-DI-2 R-B-DI-3 Average S.D. P-B-DI-1 P-B-DI-2 P-B-DI-3 Average S.D.

1.9/2.8 2.1/3.0 2.1/2.8 2.0/2.9 0.1/0.1 2.3 2.0 2.3 2.2 0.1 2.3 2.4 2.5 2.4 0.1 6.1 5.7 5.7 5.8 0.2 2.7 2.8 2.8 2.8 0.0

37.5 33.8 35.2 35.5 1.9 25.3 25.3 24.6 25.1 0.4 19.0 20.3 18.8 19.4 0.8 13.1 12.5 12.1 12.6 0.5 23.5 24.5 22.0 23.3 1.3

−130 −135 −132 −132 2.5 −127 −130 −131 −129 2.1 −123 −125 −119 −122 3.1 −301 −297 −290 −296 5.6 −156 −152 −160 −156 4.0

47.1/46.9 43.4/43.3 43.8/43.7 44.8/44.6 2.0/2.0 42.7/38.0 42.7/36.8 42.6/36.8 42.7/37.2 0.1/0.7 41.5/40.3 41.8/40.5 41.5/40.3 41.6/40.4 0.2/0.1 40.4/37.8 40.3/37.7 40.3/38.0 40.3/37.8 0.1/0.2 42.4/41.0 42.6/41.4 42.1/41.2 42.4/41.2 0.3/0.2

1.3 1.2 1.2 1.2 0.1 0.9 0.9 0.8 0.9 0.1 0.8 0.8 0.8 0.8 0.0 0.2 0.2 0.2 0.2 0.0 0.6 0.6 0.6 0.6 0.0

Note that the specimen label, e.g. EPP-F-DI-1, represents the sandwich panels with EEP foam under dynamic impact loading in Specimen #1 of this group.

displacement response of cellular structures, including an initial yield stage, a slow increase stage and a rapid increase stage. During the initial yielding stage, the load increased approximately linearly at first until reached a plateau due to the buckling of the cell structures. After the plateau, the load increased slowly due to the elastic deformation of the rubber balls and started to compact gradually; finally the load increased quickly due to the densification of the balls. The start point of the densification stage approximately corresponded to 15 mm of displacement, and to the end the load reached the peak of 51.8 kN at the crushing displacement of 20 mm. For the sandwich panels filled with plastic balls (P-B-SC, the green curve in Fig. 3), the load increased fairly quickly with the crushing displacement. After the displacement reached around 13.3 mm, the bulges (fluctuations) appeared on the increasing curve, indicating that the plastic balls lost their stability and began the plastic deformation. Interestingly, while the noticeable difference in the loading-displacement curves, both the specimens filled with plastic and rubber balls reached the similar peak loads around 49 kN at the end of the compression stage. The load-displacement curve of the sandwich panels filled with the EPP foam (EPP-F-SC, the blue curve in Fig. 3) located in the bottom of these curves, representing the typical micro-cellular material behavior (i.e. the dimension of cell is smaller than the rubber balls and plastic

Fig. 3. Load-displacement curves of CFRP sandwich specimens filled with different core materials under compression.

400

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

balls). Under compression, the cellular edges of EPP foams collapsed by elastic or plastic buckling. The faces of these closed cells bend and bead walls collapsed mainly in buckling manner [30], which accordingly densified the foams and led to a fairly smooth trend of increase in the loaddisplacement curve. 3.1.2. Damage process Fig. 4 exhibits the crushing process and damaged core for these four different CFRP sandwich specimens. It is interesting to note the top and bottom CFRP facesheets remained intact (no damaged) in the static compression events. For the specimen filled with the EPP foam (EPP-F-SC) (Fig. 3 and Fig. 4a), the contact load grew smoothly, after the crushing distance around 15 mm on average, the EPP foam stepped into the densification region, leading to slight increase in load. By comparing the initial height and final height of the specimen, the minimum reduction in height was observed (only 3.9 mm on average) in these four tested sandwich structures, attributable to the excellent elastic recoverability of the EPP foams. For the specimen filled with aluminum honeycomb (Al-H-SC) (Fig. 3 and Fig. 4b), the contact load increased almost linearly with the displacement first until reached the peak load of average 20.1 kN at displacement about 1 mm. As the platen continued moving downwards, progressive folding appeared between the contact of the upper facesheet and honeycomb core, leading to a mean load at 7.5 kN until the test finished. The final deformed honeycomb became a hexagon shape, indicating the honeycomb cells remained their self-folding form under crushing. By comparing the initial height and the final height of specimen, the maximum reduction in height was observed (19.8 mm on average) of these four tested sandwich structures, showing that aluminum honeycomb cores were of the poor recoverability and severe plastic deformation under the compression. For the specimens filled with rubber balls (R-B-SC) (Fig. 3 and Fig. 4c), the contact load increased smoothly, which was similar to the

curve of EPP-F-SC. The rubber foam balls were tightly squeezed by the loading and support platens from the both sides. Note that when the upper facesheet moved down for more than 15 mm, the rubber foam balls started strongly resisting the compression, causing a dramatic rise in the contact loading. Compared to the intact specimen, there was rare deformation on the upper facesheet and the rubber foam balls; and the reduction in height was 5.2 mm on average for these three tested specimens. For the specimens filled with plastic hollow balls (P-B-SC) (Fig. 3 and Fig. 4d), the remarkable difference can be observed compared to the RB-SC counterparts, particularly being seen in the slope of load-displacement curves and the damage appearance of the cores. As shown in Fig. 3, the load increased more quickly than that of R-B-SC under the same displacement. After the average displacement over 13 mm, the slope changed due to significant unrecoverable plastic deformation on the plastic balls, finally making the total height of P-B-SC decrease by 10.7 mm on average for these three tested specimens. 3.1.3. Energy absorption Sandwich structures have drawn increasing attention in applications for energy absorbers, thus the energy absorption and SEA were investigated herein [31]. It is interesting to note that there is no visible damage either on the top or bottom facesheets for all the specimens, therefore the test merely reflected the energy absorption behaviors of the filled core materials. As shown in Table 1 and Fig. 5, the average values of energy absorption (Ea) are 29.8, 143.0, 98.2 and 320.9 J, for EPP-F-SC, Al-H-SC, R-B-SC and P-B-SC, respectively. It is interesting to note that P-B-SC absorbed the most energy, which was mainly due to the high capacity of plastic deformation of core filler. As shown in Fig. 3, the slope of P-B-SC was greater than those of EPP-F-SC and R-B-SC; while Al-H-SC showed high slope only at the pre-compression stage. The deformation of plastic balls, especially at plastic deformation stage, greatly strengthened the CFRP facesheets and absorbed substantially more energy. On the other

Fig. 4. Failure process and damaged specimen: (a) EPP-F-SC, (b) Al-H-SC, (c) R-B-SC and (d)P-B-SC.

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

Fig. 5. Absorbed energy and SEA in the static compression tests.

hand, the progressive buckling collapse of aluminum honeycomb core provided a main source of the energy absorption. It is noted that the core materials of plastic balls and aluminum honeycomb were seriously damaged and became irreversible. However, the EPP foam and rubber foam balls showed invisible deformation on their surface, leading to a lower energy absorption. Considering the different weights of the fillers, the SEA exhibited different trends, as shown in Fig. 5. The specimen filled with the plastic hollow balls was of the highest SEA value of 4.8 J/g, followed by that

401

filled with aluminum honeycomb (2.8 J/g), and finally with rubber foam balls (0.6 J/g) and EPP foam (0.7 J/g). The low value of SEA of the EPP foam was related to the lowest crushing load (as shown in Fig. 3). Al-H-SC showed better crushing characteristics than the EPP-F-SC and R-B-SC counterparts, mainly attributable to its progressive collapse in compression, which could maintain the load at a stable level and result in good energy absorption. Similar to EPP-F-SC, R-B-SC exhibited little damage on the surface. Interestingly, it can be seen that R-B-SC had the highest peak load of these four specimens but the lowest SEA. This is mainly due to the greatest weight of 166.4 g, which was 147–284% heavier than the other specimens. Note that the introduction of plastic ball filler (P-B-SC) significantly increased the energy absorption; and the corresponding SEA was 700% higher than that of R-B-SC. By comparison of these four filled CFRP panels, it can be concluded that the Al-HSC and P-B-SC had better energy absorption capacity under the static compression.

3.2. Low velocity impact tests 3.2.1. Load-displacement curves To understand the dynamic responses of the sandwich panels filled with various reinforced materials, a series of low velocity impact tests were carried out. Fig. 6 plots the typical load-displacement curves of sandwich panels without and with different cores, which can be classified into three categories under the same initial impact kinetic energy here, namely no rebound (e.g. U-F-S in Fig. 6b), incomplete rebound

Fig. 6. Typical load-displacement curves in the low velocity impact tests.

402

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

(e.g. Al-H-DI and P-B-DI in Fig. 6c) and complete rebound (EPP-F-DI and R-B-DI in Fig. 6d). The impact load increased almost linearly with displacement until reached the initial peaks for all the specimens with different core fillers, which strengthened the resistance of whole sandwich panels against the penetration. Note that the stiffnesses of all specimens were different (as per the slope of the linear stage), indicating that in addition to the intrinsic material properties of facesheets [32], filler materials influenced the behavior of sandwich panels [12]. It can be seen clearly in Fig. 6a that the stiffness of sandwich panels with reinforced core filler was higher than that of unfilled panels except for the case with the EPP foam (EPP-F-DI). After the initial peak, the responses of loaddisplacement exhibited different characteristics. For the unfilled specimen (U-F-S), the load-displacement curve presented a double-peak shape (Fig. 6b) which could be divided into three regions with points A and B. The load increased linearly to the first peak and the slope indicated the elastic modulus of the unfilled CFRP panel. A number of oscillations occurred before the load reached the peak value, caused by invisible matrix crack and fiber breakage at the upper CFRP facesheet. After the initial peak, the load gradually declined, indicating the complete damage of top facesheet with significant transverse cracks. The load thus declined to a minimum value when the tip of spherical impactor penetrated into the top facesheet. Following this, there was a slight increase in load between point A and point B, mostly attributed to the increase in frictional resistance between the impactor and the upper facesheet with the increasing contact area due to the spherical shape of the impactor. When the impactor continued penetrating through, the load reached a higher second peak linearly and then experienced in a loading plateau due to the combined effect of friction between the impactor and top facesheet as well as progressive fracture of the bottom facesheet. After the plateau, the load dropped suddenly to nearly zero at the end of curve, indicating that the complete perforation was developed in the lower facesheet, without rebound in the loading curve.

For the sandwich panels filled with the aluminum honeycomb (AlH-DI) and plastic hollow balls (P-B-DI), the load-displacement curves could be characterized by “incomplete rebound”, which had a single sharp loading peak and a small rebound at the end of the test (Fig. 6c). The loads for these two specimens were fairly similar in the beginning of the tests with nearly the same loading peak and slope, which were higher than that of unfilled specimen (U-F-S) due to the contribution of filler materials. After the first peak, Al-H-DI exhibited a long stable plateau, which was similar to the responses of specimen under quasistatic compression, but the value of the load at this plateau was lower mainly due to sever localized destruction of the aluminum honeycomb. Along with the compaction/densification of aluminum honeycomb under crashing, the load slightly increased at the end of the plateau to form the second hump before dropped rapidly to nearly zero at the end of curve with only small rebound. While for P-B-DI, after the initial peak, the load fluctuated with the increase in the displacement, gradually dropped to below 1 kN, and then increased as shown in Fig. 6c due to the destruction of filled plastic hollow balls. Similarly to Al-H-DI, the load dropped rapidly to nearly zero at the end of curve with a small rebound. For the sandwich panels filled with the EPP foam (EPP-F-DI) and rubber foam balls (R-B-DI), the load-displacement curves had a single peak and followed with a significant rebound at the end of the test with an almost enclosed shape (Fig. 6d). The load for R-B-DI increased almost linearly with the displacement until reached the peak force, and then decreased with almost complete rebound. It is interesting to note that the stiffness of the sandwich panels filled with the EPP foam (EPP-F-DI) was even lower than that of the unfilled sandwich panels with four columns at the corners; and the response of EPP-F-DI under low velocity impact was fairly similar to bare EPP foam under quasistatic compression [33], which comprised an elastic phase and a higher plateau duration. The above results indicated that the sandwich panels filled with the EPP foam and aluminum honeycomb might be more suitable to be the

Fig. 7. Typical impacted specimen of U-F-S: (a) top view of upper facesheet (exterior), (b) back view of upper facesheet (interior), (c) top view of lower facesheet (interior), (d) back view of lower facesheet (exterior), and (e) cross-sectional view.

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

403

Fig. 8. Typical impacted specimen of EPP-F-DI: (a) top view of the upper facesheet (exterior), (b) back view of the upper facesheet (interior), (c) top view of damaged filled core, (d) back view of bottom facesheet (exterior), and (e) cross-sectional view.

Fig. 9. Typical impacted specimen of Al-H-DI: (a) top view of top facesheet (exterior), (b) back view of top facesheet (interior), (c) top view of damaged filled core, (d) back view of bottom facesheet (exterior), and (e) cross-sectional view.

404

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

reinforcement core material for their long stable plateau and relative low stiffness during impact events. 3.2.2. Damage process Damage process of the sandwich composites can be explored through examination of the corresponding load-displacement curves and the damaged specimens. The failure processes in the impact-loaded sandwich composites were first investigated here by visual inspection on the exterior (front and rear) surfaces of the damaged facesheets. Then the damage mechanisms at the interior layers and cores were ascertained through destructive analysis by sectioning the samples for more detailed analysis. The different views of deformed unfilled CFRP panels were shown in Fig. 7. From Fig. 7a and c, it can be seen that the damage zone in the woven composite laminates was in a circular shape at the top view of the upper facesheet, which differed from that of the unidirectional ply composite laminates where a peanut shape damage zone was common in the fiber directions [34]. A cross cracking pattern from the impact center was observed on the upper facesheet with significant fiber breakage and matrix-fiber delamination, showing that tear damage occurred during the impact tests. The diameter of the circular damage zone on the top side was around 42 mm on average, which was the biggest of all types of specimens and was larger than that of the impactor (36 mm), indicating that a typical complete perforation damage occurred. As illustrated in Fig. 7b, the shape of damage zone was rhombus at the back view of the upper facesheet and the damage area was larger than the top view of the facesheet, which was mostly due to significant bending failure on the interior surface of the upper facesheet. It is necessary to note that the through-thickness perforation occurred and the interior facesheet also experienced in a penetration damage.

Fig. 7c and d showed the top (interior) and back (exterior) view of the bottom facesheet; which exhibited the similar damage shape but relatively smaller damage area compared to the top facesheet. Note that only the fiber breakage was inspected in the bottom facesheet, Specifically severe delamination and fiber pull-out can be observed in the cross-sectional view of the damage specimen (Fig. 7e), where the maximum displacement was measured as 31 mm which was smaller than the maximum deflection (35.5 mm) acquired in the load-displacement curve, mainly due to elastic recovery of the bottom facesheet. Interestingly, it is observed that the typical fracture pattern of the sandwich panels filled with aluminum honeycomb (Al-H-DI), rubber foam balls (R-B-DI), and plastic hollow balls (P-B-DI) were quite similar. As shown in Figs. 9–11, only the top facesheet experienced a punchthrough damage. However, the size of damage area for the top facesheet was different for the filler materials. The damage area of specific specimen was dimensioned; and the matrix crack, fiber breakage and circumferential fracture lines were observed. All these three types of specimens (Al-H-DI, R-B-DI and P-B-DI) were non-perforated, where only the upper facesheets were penetrated and the shapes of their damage fragmentation were similar (i.e. combined damages of matrix crack and fiber breakage) with severe pyramid perforated rear damages. Tear damage were also noticed in these three types of specimens. Exception is in the test of EPP-F-DI (Fig. 8a and b), where a remarkable through-thickness crack initiated from the impact area and propagated to the edge of composite facesheet, indicating that the core material played an important role in damage mechanism and affected the shape of the damage area on sandwich panels. Fiber breakage, matrix cracks and local buckling were observed at the impacted (top) facesheet of EPP-F-DI. As shown in Fig. 8c and e, the damage modes of the EPP core included a cross-shaped indentation and tear damage on the impacted surface. A large area of elastic deformation with depth of

Fig. 10. Typical impacted specimen of R-B-DI: (a) top view of top facesheet (exterior), (b) back view of top facesheet (interior), (c) top view of damaged filled core, (d) back view of bottom facesheet (exterior), and (e) cross-sectional view.

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

25.1 mm and a local indentation with depth of 6.1 mm were also observed. The global bending and debonding between facesheet and core were also observed as shown in the cross-sectional view in Fig. 8e. Fig. 9 illustrated the typical damaged specimen of Al-H-DI. A diameter of 35.5 mm was measured in the circular damage zone on the top surface (exterior) of the upper facesheet, where the primary damage modes were similar to that in the above descriptions as shown in Fig. 9a and b. After the top facesheet was penetrated completely, the honeycomb core buckled progressively and presented the major capacity of impact resistance. The main deformation area had almost the same size as the impactor area, whereas the whole damage zone was fairly large as shown in Fig. 9c. The deformation of the honeycomb core included three modes: the core crushing, core shear and core buckling in the low-velocity impact tests, which were similar to [35,36]. It is noted that the main failure mode of specimens herein was the progressive crushing of core with depth of 19.2 mm as observed in the crosssectional view (Fig. 9e). The typical impacted specimen of sandwich panels filled with rubber foam balls (R-B-DI) was shown in Fig. 10. Clearly, this group of specimens showed minimal damage area, including matrix crack, fiber breakage and tear damage, which is even slightly smaller than the impactor area (Fig. 10a and b). Only one rubber foam ball at the impact center experienced the critical tearing damage, as shown in Fig. 10c. From the cross-sectional image (Fig. 10e), the damage of the top facesheet and rubber foam balls can be clearly seen. The penetration of the specimens was about 7 mm on average, which is smaller than the maximum displacement of impactor (12.6 mm) recorded in the load-displacement curve. This is mainly due to significant elastic recovery of the top facesheet and rubber foam balls. Similar to the unfilled sandwich panels, the P-B-DI specimen exhibited a circular hole with a diameter of 36 mm on the top facesheet as shown in Fig. 11a. The main damage modes were displayed in Fig. 11a and b. As shown in Fig. 11c, the plastic hollow ball at the impact center

405

was entirely destroyed with brittle fragments accompanied by the cracks on the adjacent four balls. The fiber pull-out and delamination of the top facesheet were also observed in the cross-sectional view in Fig. 11e. The impactor penetration in the specimens was measured approximate 11.5 mm, which was deeper than that of rubber foam balls (R-B-DI) but was still smaller than the drop displacement of impactor (23.3 mm on average), indicating significant elastic recovery. 3.2.3. Crashworthiness characteristics The crashworthiness characteristics of these five different sandwich panels are compared under the same initial impact velocity herein. The crash force-displacement curves were plotted in Fig. 6 and corresponding initial peak values and maximum displacement were compared in Fig. 12. By comparing with the unfilled sandwich specimens, it can be seen that the sandwich panels filled with different core materials presented better resistance to the impact loading (with a higher peak load and lower displacement). Interestingly, of these specimens, the R-B-DI panel (filled with rubber balls) had a highest peak load (5.8 kN) and minimum displacement (12.6 mm). On the contrary, the EPP-F-DI (filled with the EPP foam) developed a lowest peak load (2.2 kN) and greatest displacement (25.1 mm). It is worth mentioning that the lower displacement indicated better impact resistance when the specimens were subjected to the same impact kinetic energy [15]. Note that a good energy absorber should be of a low peak force to prevent severe damage and injury of passenger or goods from crashing [37]. As a result, the EPP foam and aluminum honeycomb would be a proper candidate material of choice in these different cores due to their low peak load. The rubber balls may improve the impact resistance and energy absorption of the structure in a high velocity impact loading [38] with a low displacement. Impact energy (Ei) and absorbed energy (Ea) (as defined in Section 2.3) are two important but different indicators to the evaluation of the crashing behaviors and impact resistance in the composite structures.

Fig. 11. Typical impacted specimen of P-B-DI: (a) top view of top facesheet (exterior), (b) back view of top facesheet (interior), (c) top view of damaged filled core, (d) back view of bottom facesheet (exterior), and (e) cross-sectional view.

406

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

Fig. 12. Comparison of peak loads and maximum displacements for different sandwich panels under impact tests.

The absorbed energy, calculated from associated load-displacement curves, with respect to real time is plotted in Fig. 13a. It can be seen that energy absorption increased with the time in the beginning, then decreased after the impactor reaching the maximum displacement, and finally maintained at a constant level [39]. This constant value represented the total energy (Ea) absorbed permanently by the sandwich panels at the end of an impact event. The maximum value of each curve (Fig. 13a) represented the associated impact energies (Ei) [40]. From the test results it is noted that the energy lost during the impact process, e.g. the heat due to friction, dynamic vibration, etc. was neglected in the calculation. Thus the energy absorption Ea is lower than the Ei for each specimen as shown in Fig. 13b; and their difference in the same curve is the recovered energy that rebounds the impactor from the non-perforated specimens [39] (e.g. for EPP-F-DI, Al-H-DI, RB-DI and P-B-DI). It is interesting to note that the perforated unfilled specimen (U-F-S) exhibited the highest energy absorption (44.6 J) due to large displacement and friction between the impactor and penetrated CFRP sheets after complete perforation, followed by P-B-DI (41.2 J), AlH-DI (40.4 J) and EPP-F-DI (37.2 J). The SEA is another key indicator to estimation of the crashworthiness. The average SEAs were calculated as 1.2, 0.9, 0.8, 0.2 and 0.6 J/g for the U-F-S, EPP-F-DI, Al-H-DI, R-B-DI and P-B-DI, respectively (Fig. 13b). It can be seen that the SEA of all the filled specimens are lower than that of unfilled specimen. The reason lies on the different damage modes where the unfilled specimens exhibited the penetration damage in both the top and bottom facesheets, whereas the filled specimens experienced in non-perforated damage with intact at the bottom

Fig. 14. SEA-displacement curves for the sandwich panels with different filled cores under low velocity impact.

facesheet [12]. It can be observed from Fig. 12 and Fig. 13b that the sandwich panels with EPP and aluminum honeycomb exhibited a lower peak load but a higher SEA among these filled specimens, indicating a good potential to be an effective energy absorbers. This finding was similar to the medium velocity impact response of sandwich panels with the polystyrene (PS) foam reported in literature [12]. The results may be attributed to the fact that the EPP foam and aluminum honeycomb appeared to provide much greater foundational elastic support to the sandwich skins, allowing the deformation (Figs. 8e and 9e) to spread out the entire facesheet much more evenly and allowing the skins to absorb a much more energy. In order to achieve a good crashworthiness behavior, typically a progressive crushing mode would be expected to ensure the energy absorption to be as high as possible by allowing the development of a stable high level crushing force, and avoiding sudden drop during crushing process [25]. Thus the SEA-displacement curves were plotted in Fig. 14, which showed that AL-H-DI worked most ideally with a smooth and linear increase in SEA with the displacement. Note that there was slight fluctuation in the SEA-displacement curves of U-F-S-2 and P-B-DI-1, mainly due to the progressive fracture of the bottom facesheet and destruction of the plastic hollow balls, respectively. In addition, a small increase rate of ascent can be found in the curves of EPPF-DI-2 and R-B-DI-1. The acceleration performance of vehicle in a collision is another important criterion to measure its safety. The higher and longer the

Fig. 13. (a) Typical absorbed energy vs time curves, and (b) energy absorption of different specimens under impact tests.

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

407

Fig. 15. (a) Typical acceleration-time curves and (b) maximum acceleration and duration time of sandwich panels under impact tests.

acceleration, the more harmful the passengers/goods would suffer [41]. Fig. 15a plots the relation of acceleration vs time under impact tests. It is noted that the curve of U-F-S is quite different from the others. For U-FS, there existed two peak values, −98 m/s2 at 4 ms and −132 m/s2 at 27 ms, corresponding to the breakages of the top and bottom facesheets, respectively, and the whole deformation time lasted 36 ms. For Al-H-DI, P-B-DI and EPP-F-DI, the tendency of the curves was nearly the same, showing one peak value first, then following with many oscillations and approaching to zero at the end. The average peak acceleration values were − 122 m/s2 at 21 ms, − 156 m/s2 at 4 ms, − 129 m/s2 at 20 ms, and the duration were 27 ms, 30 ms, 45 ms, for Al-H-DI, P-BDI, EPP-F-DI, respectively. Both the acceleration value and the duration time indicated that Al-H-DI was the best cushioning material. For R-BDI, although the duration was the shortest (18 ms), the acceleration peak was too high (− 296 m/s2), which could potentially cause great harm to the passengers/goods. The load bearing (peak load) and energy absorption (SEA) capabilities of the quasi-static compressing sandwich panels are much higher than those of low velocity impact loading counterparts, as shown in Fig. 16. The peak load for the specimens with EPP foam, aluminum honeycomb, rubber balls, and plastic hollow balls subjected to quasi-static compression are 4.4, 20.1, 49.9 and 48.9 kN, respectively, which are approximately 169.2%, 717.9%, 860.3% and 1630.0% of those under the low velocity impact. The SEA values for the specimens with the EPP foam, aluminum honeycomb, rubber balls, and plastic hollow balls under quasi-static compression are 0.7, 2.8, 0.6 and 4.8 J/g, respectively, which are about 77.8%, 350.0%, 300.0% and 800.0% of those subject to the low velocity impact. Based on this valuable finding, the EPP foam showed much more effective energy absorption capability in the low

Fig. 16. Comparison of crashing characteristics in the quasi-static compression and low velocity impact tests.

velocity impact than that in the quasi-static compression. It was noted that the SEAs in the low velocity impact tests were much lower than those in the quasi-static compressing tests, mainly due to the localized facesheet rupture and core crushing during impact whereas the global damage on filled cores during quasi-static compression. 4. Conclusions The crashworthiness of CFRP panels filled with different reinforced core materials were experimentally investigated under both the static and dynamic loadings in this study. Based on the test results and analyses, the following conclusions can be drawn within the limitations of the study: (1) For the quasi-static compression tests, the peak loads were 4.4, 20.1, 49.9 and 48.9 kN for EPP-F-SC, Al-H-SC, R-B-SC and P-B-SC, respectively. The specimen filled with the plastic hollow ball was of the highest SEA of 4.8 J/g, followed by the specimen filled with the aluminum honeycomb (2.8 J/g), and the rubber foam ball and EPP foam (0.6 J/g and 0.7 J/g, respectively). By comparing the initial and final displacement, the minimum reduction in height was observed in the specimens of EPP-F-SC and R-B-SC, attributable to the excellent elastic recoverability of the filler materials. (2) Three distinct load-displacement categories, classified as no rebound (U-F-S), incomplete rebound (Al-H-DI and P-B-DI) and complete rebound (EPP-F-DI and R-B-DI), were observed in the low velocity impact tests with the same initial impact kinetic energy. Both the top facesheet and bottom facesheets were completely penetrated in the unfilled specimen, whereas for the filled sandwich panels, only perforation damage was observed on the top facesheet, indicating that the filled cores was able to provide effective resistance to the impact loading. (3) Under the impact tests, the average peak loads were 2.0, 2.4, 2.8, 5.8 and 3.0 kN, the SEAs were 1.2, 0.9, 0.8, 0.2 and 0.6 J/g, for U-F-S, EPPF-DI, Al-H-DI, R-B-DI and P-B-DI, respectively. The localized facesheet rupture and core crushing dedicated major energy absorption during impact. It was noted that the SEAs in the low velocity impact tests was much lower than those in the quasi-static compressing tests. (4) Only the top facesheet experienced in penetration damage in the low velocity impact tests and the size of damage zone differed with the different filler materials. The shape of the damage area in the top woven CFRP facesheet exhibited approximately a circular shape for the sandwich panels except that filled with the EPP foam, where a remarkable through-thickness crack initiated from the impact area and propagated to the edge of CFRP facesheet. (5) It was interesting to note that these filled specimens presented different characteristics in the static compression and dynamic impact tests. The aluminum honeycomb and plastic hollow balls can be two candidate core materials to reinforce the sandwich structures under static compression loading, while aluminum honeycomb and EPP

408

Y. Zhang et al. / Materials and Design 117 (2017) 396–408

foam can be potential candidate to be an effective energy absorber under dynamic impact loading. Acknowledgements This investigation was financially supported by the National Natural Science Foundation of China (51675540), the Outstanding Young Scholars of Guangdon Province g(2015TQ01X371), the Pearl River S&T Nova Program of Guangzhou (2014J2200005), the Natural Science Foundation of Guangdong Province (2015A030313016), the Natural Science Foundation of Hunan Province (2016JJ3039) and the China Postdoctoral Science Foundation (2015M582323). References [1] J.R. Vinson, Sandwich structures, Appl. Mech. Rev. 54 (2001) 1–2. [2] G.B. Chai, S.Q. Zhu, A review of low-velocity impact on sandwich structures, Proc. Ins. Mech. Eng. Part L J. Mat. Des. Appl. 225 (2011) 207–230. [3] J. Wang, A.M. Waas, H. Wang, Experimental and numerical study on the low-velocity impact behavior of foam-core sandwich panels, J. Sandw. Struct. Mater. 17 (2015) 54–58. [4] K.R. Ramakrishnan, S. Guérard, P. Viot, K. Shankar, Effect of block copolymer nanoreinforcements on the low velocity impact response of sandwich structures, Compos. Struct. 110 (2014) 174–182. [5] M. Meo, R. Vignjevic, G. Marengo, The response of honeycomb sandwich panels under low-velocity impact loading, Int. J. Mech. Sci. 47 (2005) 1301–1325. [6] P.M. Schubel, J.J. Luo, I.M. Daniel, Low velocity impact behavior of composite sandwich panels, Comp. Part A Appl. Sci. Manufact. 36 (2005) 1389–1396. [7] T. Anderson, E. Madenci, Experimental investigation of low-velocity impact characteristics of sandwich composites, Compos. Struct. 50 (2000) 239–247. [8] A. Kurşun, M. Şenel, H.M. Enginsoy, E. Bayraktar, Effect of impactor shapes on the low velocity impact damage of sandwich composite plate: experimental study and modelling, Composites Part B Engineering. 86 (2015) 143–151. [9] E. Sevkat, B. Liaw, F. Delale, Drop-weight impact response of hybrid composites impacted by impactor of various geometries, Mater. Des. 52 (2013) 67–77. [10] O. Ozdemir, R. Karakuzu, A.K.J. Alshamary, Core-thickness effect on the impact response of sandwich composites with poly(vinyl chloride) and poly(ethylene terephthalate) foam cores, J. Compos. Mater. 49 (2015). [11] W. Hou, F. Zhu, G. Lu, D.N. Fang, Ballistic impact experiments of metallic sandwich panels with aluminium foam core, International Journal of Impact Engineering. 37 (2010) 1045–1055. [12] H. Wang, K.R. Ramakrishnan, K. Shankar, Experimental study of the medium velocity impact response of sandwich panels with different cores, Mater. Des. 99 (2016) 68–82. [13] S. Hou, S. Zhao, L. Ren, X. Han, Q. Li, Crashworthiness optimization of corrugated sandwich panels, Mater. Des. 51 (2013) 1071–1084. [14] R. Mohmmed, F. Zhang, B. Sun, B. Gu, Finite element analyses of low-velocity impact damage of foam sandwiched composites with different ply angles face sheets, Mater. Des. 47 (2013) 189–199. [15] B. Yang, Z. Wang, L. Zhou, J. Zhang, L. Tong, W. Liang, Study on the low-velocity impact response and CAI behavior of foam-filled sandwich panels with hybrid facesheet, Compos. Struct. 132 (2015) 1129–1140. [16] S. Hou, C. Shu, S. Zhao, T. Liu, X. Han, Q. Li, Experimental and numerical studies on multi-layered corrugated sandwich panels under crushing loading, Compos. Struct. 126 (2015) 371–385. [17] A. Shipsha, Failure of Sandwich Structures With Sub-interface Damage, Phd Theses Department of Aeronautics, 2001.

[18] C. Atas, C. Sevim, On the impact response of sandwich composites with cores of balsa wood and PVC foam, Compos. Struct. 93 (2010) 40–48. [19] G. Pitarresi, J.J. Carruthers, A.M. Robinson, G. Torre, J.M. Kenny, S. Ingleton, et al., A comparative evaluation of crashworthy composite sandwich structures, Compos. Struct. 78 (2007) 34–44. [20] A.G. Mamalis, D.E. Manolakos, M.B. Ioannidis, D.P. Papapostolou, On the crushing response of composite sandwich panels subjected to edgewise compression: experimental, Compos. Struct. 71 (2005) 246–257. [21] M.A. Hazizan, W.J. Cantwell, The low velocity impact response of an aluminium honeycomb sandwich structure, Composites Part B Engineering. 34 (2003) 679–687. [22] P. Feraboli, F. Deleo, B. Wade, M. Rassaian, M. Higgins, A. Byar, et al., Predictive modeling of an energy-absorbing sandwich structural concept using the building block approach, Comp. Part A Appl. Sci. Manufact. 41 (2010) 774–786. [23] Q. Liu, H. Xing, Y. Ju, Z. Ou, Q. Li, Quasi-static axial crushing and transverse bending of double hat shaped CFRP tubes, Compos. Struct. 117 (2014) 1–11. [24] Q. Liu, Z. Mo, Y. Wu, J. Ma, G.C.P. Tsui, D. Hui, Crush response of CFRP square tube filled with aluminum honeycomb, Comp. Part B Eng. 98 (2016) 406–414. [25] E. Troiani, L. Donati, G. Molinari, R.D. Sante, Influence of plying strategies and trigger type on crashworthiness properties of carbon fiber laminates cured through autoclave processing, Strojniski Vestnik. 60 (2014) 375–381. [26] G.C. Jacob, J.F. Fellers, S. Simunovic, J.M. Starbuck, Energy absorption in polymer composites for automotive crashworthiness, J. Compos. Mater. 36 (2002) 813–850. [27] J.A. Lavoie, S. Kellas, Dynamic crush tests of energy-absorbing laminated composite plates, Comp. Part A Appl. Sci. & Manufact. 27 (1996) 467–475. [28] Federal Motor Vehicle Safety Standards and Regulations (Revised December 1998), 1998. [29] D3763-02 A, Standard Test Method for High Speed Puncture Properties of Plastics Using Load and Displacement Sensors. West Conshohocken, PA, 2002. [30] R. Bouix, P. Viot, J.-L. Lataillade, Polypropylene foam behaviour under dynamic loadings: strain rate, density and microstructure effects, International Journal of Impact Engineering. 36 (2009) 329–342. [31] Q. Liu, Y.Z. Lin, Z.J. Zong, G.Y. Sun, Q. Li, Lightweight design of carbon twill weave fabric composite body structure for electric vehicle, Compos. Struct. 97 (2013) 231–238. [32] K. Mohan, T.H. Yip, S. Idapalapati, Z. Chen, Impact response of aluminum foam core sandwich structures, Mater. Sci. Eng. A 529 (2011) 94–101. [33] S. Ouellet, D. Cronin, M. Worswick, Compressive response of polymeric foams under quasi-static, medium and high strain rate conditions, Polym. Test. 25 (2006) 731–743. [34] J. Wang, A.M. Waas, H. Wang, Experimental and numerical study on the low-velocity impact behavior of foam-core sandwich panels, Compos. Struct. 96 (2013) 298–311. [35] E. Wu, W.-S. Jiang, Axial crush of metallic honeycombs, Int. J. Impact Eng. 19 (1997) 439–456. [36] M. Yamashita, M. Gotoh, Impact behavior of honeycomb structures with various cell specifications—numerical simulation and experiment, International Journal of Impact Engineering. 32 (2005) 618–630. [37] N. Qiu, Y. Gao, J. Fang, Z. Feng, G. Sun, Q. Li, Crashworthiness analysis and design of multi-cell hexagonal columns under multiple loading cases, Finite Elem. Anal. Des. 104 (2015) 89–101. [38] E. Sarlin, M. Apostol, M. Lindroos, V.T. Kuokkala, J. Vuorinen, T. Lepistö, et al., Impact properties of novel corrosion resistant hybrid structures, Compos. Struct. 108 (2014) 886–893. [39] W. Wu, Q. Liu, Z. Zong, G. Sun, Q. Li, Experimental investigation into transverse crashworthiness of CFRP adhesively bonded joints in vehicle structure, Compos. Struct. 106 (2013) 581–589. [40] C. Atas, O. Sayman, An overall view on impact response of woven fabric composite plates, Compos. Struct. 82 (2008) 336–345. [41] D. Wang, Impact behavior and energy absorption of paper honeycomb sandwich panels, International Journal of Impact Engineering. 36 (2009) 110–114.