Mechanics of Advanced Materials and Structures

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/umcm20

Crashworthiness of bionic tree-shaped hexagonal

hierarchical gradient structures under oblique
crushing conditions

Yuwen Chen, Xiaolin Deng, Huilan Huang, Hailong Ran & Chengming Wang

To cite this article: Yuwen Chen, Xiaolin Deng, Huilan Huang, Hailong Ran & Chengming Wang
(27 Jul 2023): Crashworthiness of bionic tree-shaped hexagonal hierarchical gradient structures
under oblique crushing conditions, Mechanics of Advanced Materials and Structures, DOI:
10.1080/15376494.2023.2240328

To link to this article: https://doi.org/10.1080/15376494.2023.2240328

Published online: 27 Jul 2023.

Submit your article to this journal

Article views: 328

View related articles

View Crossmark data

Citing articles: 19 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=umcm20
MECHANICS OF ADVANCED MATERIALS AND STRUCTURES
https://doi.org/10.1080/15376494.2023.2240328

ORIGINAL ARTICLE

Crashworthiness of bionic tree-shaped hexagonal hierarchical gradient

structures under oblique crushing conditions
Yuwen Chena, Xiaolin Dengb, Huilan Huanga, Hailong Rana, and Chengming Wanga
a
School of Mechanical Engineering, Guangxi University, Nanning, China; bSchool of Electronics and Information Engineering, Wuzhou
University, Wuzhou, China

ABSTRACT ARTICLE HISTORY

To improve the energy absorption characteristics of thin-walled structures under oblique loads, a Received 3 May 2023
new hexagon hierarchical gradient structure (HHGS) was designed in this work, which was inspired Accepted 20 July 2023
by the fractal structure of tree gradient in nature. The crashworthiness of the structure under
KEYWORDS
oblique loads was systematically studied by numerical simulation method. The force-displacement
Crashworthiness; oblique
curve of this structure under oblique loads is different from that of conventional thin-walled tubes impact; thin-walled
under axial loads. Crashworthiness comparisons show that gradient structure can significantly structures; hexagonal;
reduce initial peak forces and maintain good deformation patterns and energy absorption advan- hierarchical; gradient
tages compared to other structures under oblique loads, at 15
load, the specific energy absorp-
tion of the third-order structure was increased by 59.20% at the highest level, at the same time,
the initial peak force was reduced by 94.23%. At 30
load, the specific energy absorption of the
third-order structure was increased by 32.30% and the initial peak force was reduced by 84.45%.

1. Introduction self-similar cross section, used different strategies to obtain a

variety of new multi-cell tubes, which provided an efficient
With the increase in car ownership, the number of car colli-
way to design self-similarity-inspired multi-cell structures In
sion events is also increasing. People are pay increasing
addition, biomimetic multi-cell structures attracted wide
attention to the safety of cars, and thus the crashworthiness
attention due to their excellent energy absorption properties.
requirements of cars are becoming increasingly high. Thin-
Du et al. [12] proposed a multicellular thin-walled structure
walled structures are widely used as energy absorbers in
inspired by horsetail, and the specific energy absorption and
vehicles, trains and other vehicles due to their advantages in
strength, energy absorption and light weight. The thin- crushing force efficiency of the structure were significantly
walled structures undergo plastic deformation to absorb the higher than those of traditional structures. Gao et al. [13]
kinetic energy resulting from the collision to reduce passen- inspired by the stem of wheat and the cross-sectional struc-
ger injury. ture of bamboo, and developed a hierarchical bionic multi-
With the continuous improvement of crashworthiness cell hexagonal tube. The specific energy absorption of the
requirements, the traditional thin-walled structures have been third-order tube was approximately 2.5 times that of the
unable to meet the requirements due to their cross-section ordinary six-cell hexagonal tube. Shi et al. [14] designed four
characteristics and material distribution. Therefore, to further new types of multi-cell columns, which were composed of an
improve the energy absorption performance of thin-walled outer tube, five square subunits and connecting ribs, and the
structures, many new energy-absorbing structures were pro- translation of four subunits and the rotation of the central
posed and studied, such as multi-cell structures [1, 2]. subunit formed four new types of multi-cell columns.
Research [3] proved that multi-cell structures have better Compared with other typical multi-cell columns, it has
energy absorption characteristics than single-cell structures of remarkable crushing property. Wang et al. [15] inspired by
the same weight, and the impact force fluctuation of the plat- the sponge, and designed a novel thin-walled multi-cell tube
form stage of the multi-cell structure is relatively more uni- with a modified face-centered cubic (MFCS) cross section,
form and stable [4]. Much research has been done on the which had a more stable deformation pattern and superior
mechanical behavior of multi-cell structures under axial energy absorption characteristics compared to traditional
impact [5–9]. Gong et al. [10] proposed a new type of multi- thin-walled tubes. In addition, bionic structures based
cell square tube and its hierarchical structure by introducing on bamboo [16, 17], beetle forewing [18, 19], spider webs
square subunits at each intersection of the multi-cell tube. [20], tree rings [21], sponges [22], conches [23] and other
The new structure has better crashworthiness than the natural structures also showed excellent energy absorption
traditional multi-cell square tube. He et al. [11] inspired by characteristics.

CONTACT Xiaolin Deng [email protected] School of Electronics and Information Engineering, Wuzhou University, Wuzhou, China.
ß 2023 Taylor & Francis Group, LLC
2 Y. CHEN ET AL.

In recent years, geometric fractal structures have been It can effectively prevent the overall bending of the structure
widely studied because of their ability to significantly under oblique impact.
improve crushing performance compared with conventional The above studies show that although the fractal structure
structures. Wu et al. [24] designed triangular, square and has better energy absorption characteristics than traditional
pentagonal bionic tree fractal structures and verified the multicellular tubes, it has no advantage in reducing the ini-
accuracy of the models through simplified super folded tial peak force and resisting oblique impact, while the func-
element theory and experiments. Compared with the single- tional gradient structure has a natural advantage in reducing
cell structure, the tree fractal structure has great potential the initial peak force. In view of the above characteristics,
for improving specific energy absorption and impact force this study creatively combined gradients and fractals and
efficiency. Gong et al. [25] proposed a lotus inspired bionic proposed and studied the bionic tree hexagonal hierarchical
multicellular tube with better energy absorption performance gradient structure. The structure with a gradient change in
than traditional multicellular tubes, based on the characteris- wall thickness could not only effectively reduce the initial
tics of the lotus leaf veins branching from the center. He peak force but also have a good deformation mode under
et al. [26] proposed an excellent energy absorption structure oblique load. Therefore, this study combined two design
methods of bionic design and gradient design, inspired by
by imitating the fractal structure of the veins of the lotus
tree fractal and tree rings, proposed a hexagonal hierarchical
leaf. Ngoc San Ha et al. [27] proposed a bionic fractal
gradient structure (HHGS) with variable axial thickness, and
multi-cell bi-tubular structure with a circular cross section
systematically studied the crashworthiness of HHGS under
by imitating the fractal structure of trees and the water lotus
oblique loads.
plant. Peng et al. [28] designed bionic tree-shaped structures
with circular and regular polygonal edges according to neu-
ron branches and tree fractal structures, and adopted a 2. Materials and methods
multi-objective optimization method to conduct a compre-
2.1. Structural design
hensive evaluation and analysis of the crashworthiness of
the structures. They found that the hexagonal structure with As seen from Figure 1a, the branches of trees in nature pre-
a third-order fractal structure had the best overall crash- sent a gradient structure from thick to thin from the bottom
worthiness. Structures with excellent energy absorption often of the tree to the branches. The branches develop branches,
require a smaller initial peak load and greater specific energy showing a fractal structure. This unique gradient fractal
absorption. Although bionic fractal structures have excellent structure enables the tree to generate more branches and
energy absorption characteristics, however, there is no obvi- leaves for photosynthesis to meet its own nutritional needs.
ous advantage in reducing the initial peak force. At the same time, a stronger trunk is better able to support
The functional gradient is an effective method to reduce the side trunk and branches of the tree, making it more
the initial peak force. The functional gradient structure has resistant to wind erosion in different directions. Similarly,
been widely considered because of its natural advantages in the rings of a tree spread from the inside out, showing a
reducing the initial peak load [29]. The functional gradient hierarchical structure. Inspired by the gradient fractals and
includes the axial gradient change of the diameter of the cross tree rings of trees in nature, this study proposed hexagon
section [30–32], width [33], wall thickness [34–36], gradient hierarchical gradient structures (HHGS) with variable axial
change of filling density [37, 38] or structure [39, 40]. The thickness. Figure 1b shows the evolution of the structure
functional gradient with axial or transverse gradient changes and the structure diagram of wall thickness variation along
the axial gradient. For rotating structures with variable wall
in wall thickness is the most widely studied [34–36, 41].
thickness, CNC machining [45] can be used, for simple
While most of these studies are concerned with axial shocks,
structures with variable wall thickness, electronic wire cut-
in fact, structures can be crushed by shocks from different
ting [46] or variable gauge rolling [47] can be used, and for
directions. Therefore, many scholars have studied the crash-
complex structures, it is more convenient to manufacture by
worthiness of structures under oblique impacts. The deform-
3D printing. Structural dimensions are as follows: L0 ¼
ation mode of the structure under an oblique load is no 50 mm, H ¼ 150 mm, L1 ¼ 35 mm, L2 ¼ 20 mm, and we
longer stable [42, 43], on the other hand, under oblique load, define d1 ¼ L0  L1 , d2 ¼ L1  L2 : Previous studies by
the energy absorption capacity of the tube decreases [12, 44]. experts and scholars on the crashworthiness of axial thin-
Generally, the structure is prone to overall bending under wall energy-absorbing structures with variable thickness fully
oblique loading. For this reason, many effective methods have confirmed that the distribution of thinner walls at the
been proposed to prevent the overall bending of the structure. impact end is more conducive to reducing the initial peak
Studies [36] have shown that the deformation mode of a load of the structure [29]. The axial thickness change can be
structure with a wall thickness gradient change under an realized by power-law function [48, 49], linear function [35,
oblique impact is more stable than that of a structure with 46] and other methods. Considering the complexity of the
uniform wall thickness. The crashworthiness of the thin-walled structure in this paper, linear function is adopted. Therefore,
tube under the ascending gradient mode is generally better in this study, a thinner wall tmin was configured at the
than that of the tube under the decreasing gradient mode. impact end of the axial variable thickness tube, and a thicker
That is, a smaller wall thickness is distributed at the impact wall tmax was configured at the fixed end. The wall thickness
end, and a larger wall thickness is distributed at the fixed end. between the impact end and the fixed end increased linearly
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 3

Figure 1. Structure design: (a) inspired by branch fractals and tree rings, (b) hierarchical fractal structure and axial gradient thickness diagram [50].

along the axial direction according to Eq. (1), forming a

structure with gradient changes in axial wall thickness [34].

tx ¼ tmin þ kx
(1)
k ¼ ðtmax  tmin Þ=H

where tx is the wall thickness at position x where the dis-

tance from the impact end is. In Eq. (1), k is the coefficient
of axial thickness change. Equation (1) shows that when
k ¼ 0, the structure becomes a tube with uniform wall thick-
ness. By changing tmin and tmax, different axial thickness
variation coefficients k can be obtained. It is easy to show
that the wall thickness of the middle section is shown in
Eq. (2), where tm is the wall thickness of the middle section.
tmax þ tmin
tm ¼ (2) Figure 2. Engineering stress-strain curve of AA6061-O [51].
2
For simplicity, “HHGSn” is used where n indicates the
order of the HHGS. Adding -k to the structure name indi- 2.3. Finite element model
cates the value of k. For example, HHGS3-k6 indicates the 2.3.1. Establishing the finite element model
third-order HHGS, and k is 0.06. In this study, Abaqus/Explicit was used for numerical simu-
lation, and the finite element model is shown in Figure 3.
2.2. Material properties The tube was placed between the upper and lower rigid
body plates, and the lower rigid plate was fixed to the tube.
The material used in this paper is aluminum alloy AA6061- At the same time, all degrees of freedom of the lower rigid
O, and the engineering stress-strain curve of the material is plate were restricted. Four-node reduced integral shell elem-
shown in Figure 2. Material density q ¼ 2:7  103 kg=m3 , ent S4R was used to simulate the thin-walled structure and
and the mechanical properties of materials were: Young’s 5 integration points are used along the thickness direction.
modulus E ¼ 68GPa, Poisson’s ratio  ¼ 0:33, initial yield The inertia effect has less effect on the crashworthiness of
stress ry ¼ 71MPa, and ultimate stress ru ¼ 130:7MPa: the structure at low speed impact, coupled with the low
4 Y. CHEN ET AL.

Figure 3. Finite element model.

Figure 4. Mesh size analysis results: (a) force-displacement curves, (b) IPCF and energy absorption.

strain rate sensitivity of aluminum alloy AA6061-O [52, 53]. Table 1. Detailed data from the mesh size analysis.
Therefore, considering the analytical efficiency of the struc- Mesh size (mm) EA (J) Diff (%) IPCF (kN) Diff (%)
ture, the upper rigid body had a mass of 600 kg [54], the 1.3 8336.23 – 115.09 –
upper rigid body plate had a certain angle a with the hori- 1.4 8455.78 1.43 115.04 0.04
zontal line, and the upper rigid body plate had a vertical 1.5 8587.13 3.01 116.45 1.18
1.7 8822.26 5.83 116.74 1.43
downward impact at a fixed speed of 10 m/s [28]. The 2.0 9239.76 10.84 114.87 0.19
impact distance was set to 110 mm. The four-node reduced 2.5 9919.21 18.99 114.78 0.27
integral shell element was used to simulate the tube. The
model included a self-imposed universal contact algorithm,
and the friction coefficient was 0.2 [55].

2.3.2. Mesh test

The mesh size of the finite element model had a direct influ-
ence on the results. The larger the mesh size was, the larger
the error of the results. Therefore, convergence analysis with
six different mesh sizes was performed, the mesh sizes were
1.3 mm, 1.4 mm, 1.5 mm, 1.7 mm, 2 mm and 2.5 mm.
The model used was HHGS2, and the results are shown in
Figure 4. See Table 1 for detailed data. The energy absorp-
tion (EA) [56], mean crushing force (MCF), specific energy
absorption (SEA) and initial peak crushing force (IPCF) were
used to evaluate the energy absorption characteristics of
thin-walled structures. As shown in Figure 5, Where F(x) is
the instantaneous impact force and d is the impact distance,
m is the mass. In Figure 4a, the force-displacement curves of
various mesh sizes had the same trend except for the mesh Figure 5. Force-displacement curve and crashworthiness index diagram.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 5

Figure 6. Layer test results: (a) three different stratification methods, (b) force-displacement curves and energy absorption.

Table 2. Detailed data from the layer test.

only 3.63%. From the final deformation mode diagram of the
Layers EA (J) Diff (%) IPCF (kN) Diff (%)
structure in Figure 6b, the final deformation state of the
45 11355.29 – 63.96 –
structure with 15 layers and 45 layers had good consistency.
15 11274.06 0.72 61.64 3.63 Therefore, 15 layers could fully meet the needs of simulation
9 11077.07 2.45 60.85 4.87 analysis.

2.3.4. Model validation

size of 2.5 mm. Before the structure became dense, the force-
To verify the accuracy of the numerical model,
displacement curves of the mesh sizes of 1.3 mm, 1.4 mm
Abaqus/Explicit was used to construct the same finite elem-
and 1.5 mm were nearly identical. As seen from Figure 4b
ent model as the tube with the hexagonal cross section in
and Table 1, the difference between the energy absorbed
the experiment of Zhang [51]. The size of the structure was
when the mesh size was 1.5 mm and that when the mesh
consistent with that in Ref. [51], that is, the section side
size was 1.3 mm was only 3.01%, and the energy absorbed
length was 36 mm, the height was 120 mm, the wall thick-
when the mesh size was 2.5 mm was 18.99% larger than that
ness was 1.2 mm, and the material used was AA6061O.
when the mesh size was 1.3 mm. When the mesh size Zhang et al. [51] adopted a quasi-static compression test
changed, the initial peak force changed little, all at approxi- with a compression speed of 0.5 mm/min. Considering the
mately 115 kN, 1.5 mm mesh size and 1.3 mm mesh size, and low variability and sensitivity of AA6061O material [52, 53],
the IPCF difference was only 1.18%. The calculation time we adopted a mesh size of 1.5 mm and an impact velocity of
increased as the mesh size decreased. Therefore, considering 1 m/s in the simulation [57, 58]. The simulation results
the error and time cost overall, we finally chose a mesh size obtained are shown in Figure 7. The experiment was densi-
of 1.5 mm for subsequent analysis. fied at 71.4 mm, and thus the energy absorption before the
densification stage was used for comparison, as shown in
2.3.3. Layered test Table 3. The difference between the simulation results and
Next, the HHGS was divided into several layers of equal the experimental results was only 4.97%. The bottom of the
height along the axial direction, and the wall thickness was experimental specimen is welded and reinforced, and the
increased with fixed increments along the axial direction by simulation model is ideal, so the force-displacement curve of
assigning different thicknesses to different layers. Obviously, the experimental specimen first enters the densification
the more layers were split, the more evenly the wall thickness stage, but Figure 7c shows that the deformation modes of
the two sets of results were very similar. Considering that
of the structure changed along the axis. However, considering
the simulation results were very close to those of the experi-
the consumption of computing resources, the 2nd-order
ment conducted by Zhang et al. [51], we believed that the
HHGS (k ¼ 0.08, tm ¼ 1.2 mm) was divided into 9, 15 and 45
finite element model had sufficient accuracy and could be
layers, as shown in Figure 6a. The simulation test results of
used in subsequent simulation research.
structures with different layers under axial impact are shown
in Figure 6b. The force-displacement curve in Figure 6b
shows that the force-displacement curves of 15 layers and 45 3. Simulation results and discussion
layers were almost identical. According to the crashworthi-
3.1. Force-displacement curves and energy absorption
ness data given in Table 2, the values of energy absorption of
15 layers and 45 layers when compressed to 120 mm were Next, the finite element simulation results of HHGS1,
11274.06 J and 11355.29 J, respectively. This meant that the HHGS2 and HHGS3 were analyzed, and tm and k of the
energy absorption of 15 layers was only 0.72% different from model used were 1.3 mm and 0.1, respectively. The force-
that of 45 layers. The IPCF values of 15 layers and 45 layers displacement curve of the traditional thin-walled structure
were 61.64 kN and 63.96 kN, respectively, with a difference of under axial load is shown in Figure 8a. First, the increasing
6 Y. CHEN ET AL.

Figure 7. Experimental [51] and numerical simulation validation: (a) the force-displacement curves and EA, (b) experimental setup [51], and (c) deformation model.

Table 3. Finite element model verification results.

EA (J) Diff (%) IPCF (kN) Diff (%)
Experiment [51] 1735.62 – 42.94 –
Simulation 1649.37 4.97 44.29 3.14

Figure 8. Force-displacement curves and energy absorption: (a) conventional thin-walled tubes, (b) HHGS1, (c) HHGS2, and (d) HHGS3.

elastic stage appeared, and when the force-displacement densification stage. The force-displacement curves of the
curve reached a local maximum peak, it began to decline. HHGS under various oblique loads of 10
, 20
and 30
are
Then, the force-displacement curve presented regular fluctu- shown in Figure 8b–d. The force-displacement curves of the
ations, and this was the plateau stage. Finally, the force- HHGS under oblique loads were different from that of trad-
displacement curve continued to rise and entered the itional thin-walled tubes under axial load. Because the wall
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 7

Table 4. Energy absorption of structures at different impact angles. thickness of the structure varied along the axial gradient, the
HHGS1 HHGS2 HHGS3 general trend of the force-displacement curve constantly
10
3527.45 J 9013.13 J 18590.35 J increased. Here, the force-displacement curves of HHGS
20
2812.52 J 6571.02 J 13179.06 J under various oblique loads of 10
, 20
and 30
could be
30
2180.46 J 4818.29 J 9128.23 J

Figure 9. The deformation process of HHGS1 under multiple loading conditions.

Figure 10. The deformation process of HHGS2 under multiple loading conditions.
8 Y. CHEN ET AL.

Figure 11. The deformation process of HHGS3 under multiple loading conditions.

Figure 12. Analysis results for different wall thicknesses: (a) different wall thicknesses in the middle section, (b) force-displacement curves, (c) MCF, and (d) SEA.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 9

roughly divided into two types. One was a nearly linearly impact experiment process of the finite element analysis
increasing force-displacement curve without obvious plateau simulation was 300 frames. As seen from Figures 9–11,
stage, as shown in Figure 8c and d and the force-displace- some structures (inside the rectangular frame) did not par-
ment curves of HHGS2 and HHGS3 under 10
oblique ticipate in deformation at the impact end point, that is, all
loads. The other was the steeply increasing force-displace- structures did not reach the densification stage. Under 10

ment curve. There were several short plateau stages, as oblique load, the deformation mode of all structures was
shown in Figure 8c for the HHGS2 force-displacement curve progressive folding deformation. Under inclined loads of 20

under 20
and 30
oblique loads. None of the force-displace- and 30
, local deformation occurred on the side that first
ment curves entered the densification stage. Table 4 shows contacted the impact plate. At frame 60, with increasing
the EA of each structure at different impact angles. It can be angle, the local deformation (the part in the red circle)
seen that the EA decreases with the increase of angle and became more obvious. As the impact plate moved down
increases with the increase of level. continuously, the structure was in complete contact with the
impact plate, and the structure deformed along the normal
direction of the impact plate. The structure inclined more
3.2. Deformation modes of structures under multiple obviously along the normal direction of the impact plate.
loads
The deformation patterns of HHGS1, HHGS2 and HHGS3 4. Parameterization study
were analyzed next, and the values of tm and k in the model
used were 1.3 mm and 0.1, respectively. There were 300 data 4.1. Influence of wall thickness t
output points, each of which was 1 frame. Therefore, the In this section, HHGS3-k6 with tm¼ 1.2 mm, 1.4 mm,
1.6 mm and 1.8 mm were selected to study the impact of dif-
Table 5. Crashworthiness data of HHGS3 with different wall thicknesses. ferent wall thicknesses on the crashworthiness of the struc-
tm (mm) m (g) EA (J) SEA (J/g) MCF (kN) ture under 10
oblique impact. Figure 12a shows a diagram
1.2 615 16960.42 27.58 154.19 of the mid-section with different wall thicknesses. Figure 12b
1.4 718 21526.04 29.98 195.69
1.6 820 26440.68 32.24 240.37 shows the force-displacement curves of HHGS3 with differ-
1.8 923 31450.86 34.07 285.92 ent wall thicknesses under 10
oblique impact, and the force-
displacement curves of the structure generally showed an

Figure 13. Impact deformation states of HHGS3 with different wall thicknesses.
10 Y. CHEN ET AL.

Figure 14. Crashworthiness of different levels: (a) force-displacement curves, (b) MCF, (c) IPCF, and (d) SEA.

Table 6. Crashworthiness data of HHGSs with different hierarchical levels. near B were analyzed. In the early stage of impact, due to
Tube tm (mm) m (g) EA (J) SEA (J/g) MCF (kN) IPCF (kN) the inclined angle of the oblique impact, only part of the
HHGS1 1.7518 426 5522.24 12.96 50.20 13.29 structure’s section deformed in contact with the impact plate.
HHGS2 1.2000 426 7294.48 17.12 66.31 13.64 At the same time, due to the gradient change in wall thick-
HHGS3 0.8303 426 8320.08 19.53 75.64 3.32
ness along the axis, the changes in the two main factors of
contact area and wall thickness led to a rapid increase in the
upward trend. Each force-displacement curve could be div- force on the structure, which was represented by the force-
ided into two sections, the first half including AB and the displacement curve of AB. As the impact plate continued to
second half including BC. In addition, the slopes of the move down, the structure was in complete contact with the
force-displacement curves in AB and BC were inconsistent. impact plate. In stage B, as shown in Figure 13, the force-dis-
The slope of the force-displacement curve in AB is greater placement curve had a turning point near point B, and then
than that in BC and there is an obvious turning point near the impact plate continued to move down. The force on the
B. The force and energy absorbed by the structure increased structure was mainly affected by the gradient change in wall
with increasing wall thickness, and thus the MCF and SEA thickness along the axis, and thus the slope of the force-dis-
of the structure also increased with increasing wall thickness, placement curve in section BC was less than that in sec-
as shown in Figure 12c and d. Table 5 shows the crash- tion AB.
worthiness data of structures with different wall thicknesses.
The MCF and SEA of the structures increased with increas-
4.2. Hierarchical influence
ing wall thickness. When tm ¼ 1.8 mm, the SEA and MCF of
the structures were the largest. The SEA of HHGS3 with tm In this section, HHGS1, HHGS2 and HHGS3 (tm¼ 1.2 mm,
¼ 1.8 mm was 23.53% larger than that of HHGS3 with tm ¼ k ¼ 0.05) were selected to study the impact of different levels
1.2 mm. Figure 13 shows the deformation modes of struc- on the crashworthiness of the structure under 15
oblique
tures with different wall thicknesses under oblique impact. It impact. HHGS2 (tm ¼ 1.2 mm) was used as a benchmark to
can be seen that the deformation modes at the end of impact ensure that each structure was of the same quality.
were all progressive folding deformation. Combined with the Figure 14 shows the force-displacement curves, energy
structural deformation mode given in Figure 13, the reasons absorption and crashworthiness indices of each structure.
for the force-displacement curve to have a turning point Figure 14a shows that the overall trend of the force-
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 11

displacement curve of the structure with axial variable thick- HHGS2 and 32.09% larger than that of HHGS1. Obviously,
ness under oblique impact was constantly rising. Although energy absorption increased with increasing hierarchy. As a
the force-displacement curve of HHGS1 flattened as the result, the MCF and SEA of HHGS3 were larger than those
impact progressed, the overall trend was still upward. In of HHGS2 and HHGS1, as shown in Figure 14b and d.
addition, the force of HHGS1 fluctuated considerably, but In addition, under the condition of the same mass, the
the force-displacement curve of HHGS3 increased almost IPCF decreased with increasing hierarchy. As shown in
linearly as the level increased. Figure 14a shows that the Figure 14c, the IPCF of HHGS3 was significantly less than
energy absorption of HHGS3 was 14.06% larger than that of that of HHGS1 and HHGS2. As seen from the

Figure 15. Impact deformation state of the HHGSs at different levels.

Figure 16. Force-displacement curves and energy absorption of HHGSs at different impact angles: (a) oblique impact and model diagram, (b) HHGS1, (c) HHGS2,
and (d) HHGS3.
12 Y. CHEN ET AL.

crashworthiness data given in Table 6, the SEA of HHGS3 that of HHGS3. HHGS3 could effectively reduce IPCF while
was 14.06% larger than that of HHGS2, and the SEA of improving SEA. This phenomenon was caused by the fact
HHGS3 was 50.69% larger than that of HHGS1. The IPCF that the cross section of HHGS3 was more complex than
of HHGS1 and HHGS2 was more than 4 times larger than those of HHGS1 and HHGS2, and thus the wall thickness of
HHGS3 was the smallest among the three under the condi-
Table 7. Crashworthiness data of HHGSs with different impact angles. tion of the same mass. Meanwhile, the wall thickness varied
Tube m (g) Angle (
) EA (J) SEA (J/g) MCF (kN) along the axial gradient. As a result, the wall thickness of
HHGS1 292 10 3070.46 10.52 27.91 HHGS3 at the impact end was smaller, approximately
15 2733.42 9.36 24.85 0.4553 mm, which resulted in a significant reduction in the
20 2582.41 8.84 23.48
25 2203.11 7.54 20.03 IPCF of HHGS3. The corresponding deformation mode of
30 1905.50 6.53 17.32 the structure obtained by finite element numerical simula-
HHGS2 426 10 8087.83 18.99 73.53 tion under the 15
oblique impact is shown in Figure 15.
15 6889.25 16.17 62.63
20 5963.93 14.00 54.22 With increasing level, the wavelength of the folds decreased,
25 5199.50 12.21 47.27 and the number of folds increased, which was conducive to
30 4408.11 10.35 40.07 absorbing more energy. As seen from Figure 15, HHGS3
HHGS3 615 10 16926.02 27.52 153.87
15 14420.14 23.45 131.09 had significantly more folds (in the blue rectangle) than
20 11910.69 19.37 108.28 HHGS1 and HHGS2 at Frame 300.
25 10034.57 16.32 91.22
30 8413.81 13.68 76.49

Figure 17. Crashworthiness of HHGSs at different impact angles: (a) MCF, (b) SEA.

Figure 18. Deformation modes of the HHGSs at different impact angles.

 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 13

Figure 19. Response curves for different values of k: (a) force-displacement curves and energy absorption at a ¼ 10
, (b) force-displacement curves and energy
absorption at a ¼ 30
.

Figure 20. Crashworthiness of HHGS3 with different k values: (a) MCF and (b) SEA.

Table 8. Crashworthiness data of HHGS3 with different k values. from Figure 16b–d, the general trend of the force-displace-
Angle (
) m (g) k EA (J) SEA (J/g) MCF (kN) ment curve of HHGS was constantly rising with the process
10 615 0.05 17160.18 27.90 156.00 of impact, the force-displacement curve of HHGS1 presented
0.06 17010.42 27.66 154.64
0.07 17025.36 27.68 154.78 a gradual upward trend with a large degree of fluctuation,
0.08 16925.91 27.52 153.87 while the force-displacement curve of HHGS2 and HHGS3
0.09 16623.75 27.03 151.12 increased almost linearly. In addition, Figure 16b–d show
0.10 16474.13 26.79 149.76
30 615 0.05 8803.16 14.31 80.03 that the force exerted on the HHGSs decreased with increas-
0.06 8705.12 14.15 79.14 ing impact angle. When a ¼ 10
, the force-displacement curve
0.07 8506.27 13.83 77.33 of each structure was the largest; when a ¼ 30
, the force
0.08 8413.81 13.68 76.49
0.09 8250.05 13.41 75.00 exerted on each structure was the smallest, which led to an
0.10 8026.26 13.05 72.97 increase in the angle and a decrease in the energy absorption
of the structure. When a ¼ 10
, the EA of the structure was
the highest, and the EA of the structure was the lowest when
a ¼ 30
. It can be intuitively seen from Figure 17a and b that
as the impact angle a increased, the MCF of the HHGS gen-
4.3. Influence of impact angle a
erally showed a downward trend, and the SEA of the HHGS
Next, the crashworthiness of the structure was studied under also decreased with increasing impact angle. The SEA of
different impact angles. Here, the HHGS1, HHGS2 and HHGS2 and HHGS3 decreased more than that of HHGS1.
HHGS3 models with tm¼ 1.2 mm and k ¼ 0.08 were selected According to the crashworthiness data in Table 7, when
for analysis under five oblique impact conditions of 10
, 15
, a ¼ 20
, the SEA of HHGS1 decreased by 6.67% compared
20
, 25
and 30
. Figure 16a is a diagram of the oblique with a ¼ 15
. Similarly, the SEA of HHGS2 decreased by
impact angle and the model HHGSs. The crashworthiness 18.64% when t a ¼ 20
. The HHGS3 SEA was 23.30% less
data of different HHGSs were compared in Table 7. As seen than for a ¼ 15
. As seen from the deformation mode shown
14 Y. CHEN ET AL.

Figure 21. Deformation modes of HHGS3 with k values from 0.05 to 0.07 under 10
oblique load.

Figure 22. Deformation modes of HHGS3 with k values from 0.08 to 0.1 under 10
oblique load.

in Figure 18, when a ¼ 10
, the deformation modes of conducive to energy absorption, resulting in a significant
HHGS1, HHGS2 and HHGS3 were progressive folding reduction in the energy absorption of the structure.
deformation. As the angle increased, bending deformation
occurred on one side of the structure (red circle), and some
4.4. Influence of the axial variation coefficient k
materials failed to participate in the deformation, resulting in
reduced energy absorption by the structure. When the impact Next, the influence of the axial variation coefficient k on the
angle changed from 15
to 20
, the final modes of HHGS2 crashworthiness of the structure was studied. The HHGS3
and HHGS3 were that the side that first contacted the impact model with tm¼ 1.2 mm was used for analysis under oblique
plate was compressed, and the other side tended to undergo impacts of a ¼ 10
and a ¼ 30
, with k values ranging from
bending deformation. The deformation mode was not 0.05 to 0.1. Figure 19 shows the response curve of HHGS3
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 15

with different k values under oblique impact. As shown by impact, the energy absorbed by HHGS3 with a larger k value
the force-displacement curves in Figure 19a and b, the over- was almost the same as that absorbed by HHGS3 with a
all curves presented a rising trend. At the initial impact smaller k value. However, under 30
oblique impact, the
stage, with increasing k value, the force on the structure energy absorbed by HHGS3 with a larger k value was sig-
decreased, and the increase in the k value meant that a nificantly less than that absorbed by HHGS3 with a smaller
smaller wall thickness was allocated at the impact end, k value.
which was conducive to the reduction in the IPCF. Figure 20 and Table 8 show the crashworthiness data of
According to the energy absorption in Figure 19, at the end HHGS3 with different k values under oblique impact. Figure
of the impact, the energy absorbed by HHGS3 with a larger 20a and b show that at both a larger angle (a ¼ 30
) and a
k value (k ¼ 0.1) was less than that absorbed by HHGS3 smaller angle (a ¼ 10
), the change in the axial coefficient k
with a smaller k value (k ¼ 0.05). Under the 10
oblique had little impact on the two crashworthiness indices SEA

Figure 23. Deformation modes of HHGS3 with k values from 0.05 to 0.07 under 30
oblique load.

Figure 24. Deformation modes of HHGS3 with k values from 0.08 to 0.1 under 30
oblique load.
16 Y. CHEN ET AL.

Figure 25. Comparison of structures with same mass: (a) force-displacement curve under 15
oblique load, (b) EA under 15
oblique load, (c) force-displacement
curve under 30
oblique load, (b) EA under 30
oblique load.

Figure 26. Impact deformation final states: (a) under 15
oblique load, (b) under 30
oblique load.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 17

Figure 27. Comparison of crashworthiness under 15
oblique load: (a) MCF, (b) PCF, and (c) SEA.

Figure 28. Comparison of crashworthiness under 30
oblique load: (a) MCF, (b) PCF, and (c) SEA.
18 Y. CHEN ET AL.

and MCF. It is worth noting that under a larger impact axial gradient, the deformation mode changed from progres-
angle (a ¼ 30
), the change in axial coefficient k had a sive folding deformation at the initial impact stage to local
greater effect on SEA and MCF than under a smaller impact bending deformation at the later impact stage, while at a
angle (a ¼ 10
). Under a larger impact angle (a ¼ 30
), the smaller impact angle (a ¼ 10
), the structural deformation
MCF at k ¼ 0.1 was 8.82% smaller than that at k ¼ 0.05, and modes of k from 0.05 to 0.1 were progressive folding with
the MCF at k ¼ 0.1 was 4% less than that at k ¼ 0.05 at a little difference. Therefore, from the aspect of energy absorp-
smaller impact angle (a ¼ 10
). Combined with the deform- tion, for a larger angle of impact, the application of a
ation modes of different axial coefficients k shown in smaller k value was conducive to improving SEA and MCF,
Figures 21–24, it was found that at a larger impact angle but a smaller k value caused a relatively large IPCF in the
(a ¼ 30
), due to the increase in wall thickness along the early stage of the impact. The value of k has little influence
on SEA and MCF. For a smaller angle of impact, the larger
Table 9. Comparison of crashworthiness under 15
oblique load. the value of k was, the smaller the wall thickness allocated
tm m SEA at the impact end. It was beneficial to reduce the value of
Cross section Structure (mm) (g) (J/g) MCF (kN) IPCF (kN)
the IPCF.
HHGS3-k8 1.2000 615 23.45 131.09 5.25

5. Comparative analysis of crashworthiness

H-KF1 [59] 2.3500 615 18.79 105.07 43.81
The crashworthiness of HHGS3-k8 and H-KF1 [59], HM1-
T-T1 [60], HS2 [20], HV1-T-H1 [61], HV1-T-T1 [61], and
HM1-T-T1 [60] 2.3828 615 18.85 105.37 74.07 PSP06 [62] were compared under oblique impact angles
a ¼ 15
and a ¼ 30
. The cross-sectional dimensions and
height of all structures were consistent with the HHGS3
HS2 [20] 2.5318 615 20.28 113.39 71.10 based on the mass of the HHGS3-k8 with tm¼ 1.2 mm, and
the quality of all structures was ensured by changing the
wall thickness of other structures. Figure 25 shows the
HV1-T-H1 [61] 2.3500 615 17.75 99.25 75.05
force-displacement curves and energy absorption curves of
each structure under oblique loads of a ¼ 15
and a ¼ 30
.
HV1-T-T1 [61] 2.2025 615 14.73 82.37 91.10 Figure 25a and c show the force-displacement curves of
each structure under oblique impacts of a ¼ 15
and a ¼ 30
,
respectively. The force-displacement curves of HHGS3 gen-
PSP06 [62] 1.8298 615 16.10 90.00 79.64 erally showed a rising trend due to the axial gradient change
in wall thickness. The IPCF of HHGS3 was much smaller
than that of other structures. Figure 25b and d show the
energy absorption of each structure under oblique impacts

Table 10. Comparison of crashworthiness under 30
oblique load.

tm m SEA
Cross section Structure (mm) (g) (J/g) MCF (kN) IPCF (kN)
HHGS3-k8 1.2000 615 13.68 76.49 3.09

H-KF1 [59] 2.3500 615 10.62 59.36 35.85

HM1-T-T1 [60] 2.3828 615 10.93 61.09 11.84

HS2 [20] 2.5318 615 12.30 68.79 43.32

HV1-T-H1 [61] 2.3500 615 11.43 63.91 21.43

HV1-T-T1 [61] 2.2025 615 10.34 57.81 19.87

PSP06 [62] 1.8298 615 10.47 58.56 21.32

 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 19

of a ¼ 15
and a ¼ 30
, respectively. HHGS3 had obvious 3. Under 15
oblique impact, the force-displacement curve
advantages in energy absorption compared with the other of HHGS fluctuated less, and the force-displacement
structures. Figure 26 shows the final deformation mode of curve of HHGS3 rises almost linearly with increasing
each structure. Figure 26a showed that all structures exhib- hierarchy. Energy absorption increased with increasing
ited progressive folding deformation under an oblique hierarchy, the energy absorption of HHGS3 was 14.06%
impact of a ¼ 15
, and HHGS3 had more folds than other larger than that of HHGS2 and 32.09% larger than that
structures. As shown in Figure 26b, under 30
oblique of HHGS1. HHGS3 could effectively reduce the IPCF
impact, all structures except HHGS3 bent at the bottom of while improving the SEA and crashworthiness.
the structure. Obviously, the deformation mode of HHGS3 4. As the loading angle a increased, the MCF and SEA of
was more stable than the other structures under oblique the structure gradually decreased. When a  15
, the
impact, which effectively prevented the overall bending of deformation mode of the structure was stable. When
the structure. 15
< a  30
, the final deformation mode of the struc-
Figures 27 and 28 show the comparison of crashworthi- ture was compression on one side and bending on the
ness data of various structures under oblique impacts at other side, but the overall deformation was relatively
angles of a ¼ 15
and a ¼ 30
, respectively. Detailed crash- stable, and no overall bending occurred.
worthiness data are shown in Tables 9 and 10. As shown in 5. Under an impact at a small angle (a ¼ 10
), the change
Figure 27, the MCF and SEA of HHGS3 were larger than in k value had little effect on the SEA and MCF, but
those of other structures at a ¼ 15
oblique load. Compared the larger the k value was, the more beneficial it was to
with other structures, the SEA of HHGS3 increased by the reduction of the IPCF. Under an impact at a large
59.20% at the highest level and 15.63% at the lowest level. angle (a ¼ 30
), energy absorption and the initial peak
In Figure 27b, in terms of IPCF, the introduction of the force should be considered comprehensively to select an
HHGS3 gradient design was significantly improved com- appropriate value of k.
pared with other structures, greatly reducing the IPCF, com- 6. Compared with other structures, the HHGS3 could sig-
pared with HV1-T-T1, the IPCF of HHGS3 decreased by nificantly reduce the IPCF while maintaining the advan-
94.23%. As seen from Figure 28, the MCF and SEA of tage of energy absorption under 15
and 30
oblique
HHGS3 were also larger than those of other structures at impacts. At a ¼ 15
oblique load, the SEA of HHGS3
a ¼ 30
oblique load. Compared with other structures, the increased by 59.20% at the highest level and 15.63% at
SEA of HHGS3 increased by 32.30% at the highest level and the lowest level and the maximum reduction of IPCF
11.22% at the lowest level. According to Figure 28b, the was 94.23%. At a ¼ 30
oblique load, the SEA of
IPCF of HHGS3 was significantly improved compared to HHGS3 increased by 32.30% at the highest level and
other structures, compared with HV1-T-T1, the IPCF of 11.22% at the lowest level. The maximum reduction of
HHGS3 decreased by 84.45%. In addition, the IPCF of each IPCF was 92.87%. The HHGS deformation pattern was
structure was significantly reduced under 30
oblique impact stable under oblique impact due to its variable axial
compared to the IPCF under 15
oblique impact. thickness characteristics.

6. Conclusions Disclosure statement

Inspired by trees with fractal structure and tree rings in No potential conflict of interest was reported by the authors.
nature, this study proposed a new type of hexagonal hier-
archical gradient structure (HHGS) with variable axial thick- Funding
ness. The crashworthiness of the structure under oblique
This work is supported by the National Natural Science Foundation of
load was studied by using the verified finite element numer-
China (No. 52065059).
ical model system. The conclusions were as follows:

1. The force-displacement curves of the HHGS with axial References

variable thickness under oblique load were different [1] S. Qin, X. Deng, and X. Liu, Crashworthiness analysis of bioins-
from those of traditional thin-walled structures under pired hierarchical gradient multicell tubes under axial impact,
axial loads, and the curves constantly rising. Under Thin. Walled Struct., vol. 179, p. 109591, 2022. DOI: 10.1016/j.
oblique loads, the HHGS deformed along the normal tws.2022.109591.
direction of the impact plate. The larger the impact [2] C. Yang, Y. Tian, P. Xu, S. Yao, Z. Li, and M.S. Alqahtani,
Crashworthiness optimization of a multicellular thin-walled
angle was, the more obvious the inclination trend of the tube with triangular cells, Mech. Adv. Mater. Struct., vol. 29,
structure along the direction normal to the impact no. 28, pp. 7277–7293, 2022. DOI: 10.1080/15376494.2021.
plate. 1995548.
2. Under 10
oblique impact, the energy absorption of the [3] X. Zhang, and H. Zhang, Energy absorption of multi-cell stub
structure increased, and the MCF and SEA of the struc- columns under axial compression, Thin. Walled Struct., vol. 68,
pp. 156–163, 2013. DOI: 10.1016/j.tws.2013.03.014.
ture also increased with increasing wall thickness. The [4] X. Zhang, and H. Zhang, Axial crushing of circular multi-cell
SEA of HHGS3 with tm ¼ 1.8 mm was 23.53% larger columns, Int. J. Impact Eng., vol. 65, pp. 110–125, 2014. DOI:
than that of HHGS3 with tm ¼ 1.2 mm. 10.1016/j.ijimpeng.2013.12.002.
20 Y. CHEN ET AL.

[5] A. Alavi Nia, and M. Parsapour, Comparative analysis of energy [22] Y. Li, D. Hu, and Z. Yang, Crashworthiness design of a sponge-
absorption capacity of simple and multi-cell thin-walled tubes inspired multicell tube under axial crushing, Int. J. Mech. Sci.,
with triangular, square, hexagonal and octagonal sections, Thin. vol. 244, p. 108070, 2023. DOI: 10.1016/j.ijmecsci.2022.108070.
Walled Struct., vol. 74, pp. 155–165, 2014. DOI: 10.1016/j.tws. [23] X. Song, L. Lai, S. Xiao, Y. Tang, M. Song, J. Zhang, and Y.
2013.10.005. Zhang, Bionic design and multi-objective optimization of thin-
[6] H. Fan, Y. Luo, F. Yang, and W. Li, Approaching perfect walled structures inspired by conchs, Electron. Res. Arch., vol.
energy absorption through structural hierarchy, Int. J. Eng. Sci., 31, no. 2, pp. 575–598, 2023. DOI: 10.3934/era.2023028.
vol. 130, pp. 12–32, 2018. DOI: 10.1016/j.ijengsci.2018.05.005. [24] J. Wu, Y. Zhang, F. Zhang, Y. Hou, and X. Yan, A bionic tree-
[7] H.S. Abdullahi, and S. Gao, A novel multi-cell square tubal liked fractal structure as energy absorber under axial loading,
structure based on Voronoi tessellation for enhanced crash- Eng. Struct., vol. 245, pp. 112914, 2021. DOI: 10.1016/j.eng-
worthiness, Thin. Walled Struct., vol. 150, p. 106690, 2020. struct.2021.112914.
DOI: 10.1016/j.tws.2020.106690. [25] C. Gong, Y. Hu, and Z. Bai, Crashworthiness analysis and opti-
[8] X. Deng, S. Qin, and J. Huang, Crashworthiness analysis of gra- mization of lotus-inspired bionic multi-cell circular tubes,
dient hierarchical multicellular columns evolved from the spa- Mech. Adv. Mater. Struct., p. 2111622, 2022. DOI: 10.1080/
tial folding, Mater. Des., vol. 215, p. 110435, 2022. DOI: 10. 15376494.2022.2111622.
1016/j.matdes.2022.110435. [26] Q. He, Y. Wang, X. Shi, X. Jing, and Y. Jiang, Crushing behav-
[9] K. Wang, Y. Liu, J. Wang, J. Xiang, S. Yao, and Y. Peng, On ior on the cylindrical tube based on lotus leaf vein branched
crashworthiness behaviors of 3D printed multi-cell filled thin- structure, Mech. Mater., vol. 165, p. 104205, 2022. DOI: 10.
walled structures, Eng. Struct., vol. 254, p. 113907, 2022. DOI: 1016/j.mechmat.2021.104205.
10.1016/j.engstruct.2022.113907. [27] N.S. Ha, T.M. Pham, W. Chen, and H. Hao, Energy absorption
[10] C. Gong, Z. Bai, and Y. Hu, Crushing behaviors of novel multi- characteristics of bio-inspired hierarchical multi-cell bi-tubular
cell square tubes and its hierarchical multi-cell structures under tubes, Int. J. Mech. Sci., vol. 251, p. 108260, 2023. DOI: 10.
axial loading, Mech. Adv. Mater. Struct., vol. 30, no. 15, pp. 1016/j.ijmecsci.2023.108260.
3156–3171, 2023. DOI: 10.1080/15376494.2022.2070309. [28] Y. Peng, T. Li, C. Bao, J. Zhang, G. Xie, and H. Zhang,
[11] Y. He, X. Li, T. Jin, Q. Ji, and X. Shu, The crashworthiness Performance analysis and multi-objective optimization of bionic
design of multi-cell structures using the tessellations of self- dendritic furcal energy-absorbing structures for trains, Int. J.
similar inspired tubes, Thin. Walled Struct., vol. 180, p. 109810, Mech. Sci., vol. 246, p. 108145, 2023. DOI: 10.1016/j.ijmecsci.
2022. DOI: 10.1016/j.tws.2022.109810. 2023.108145.
[12] J. Du, P. Hao, and L. a Li, Finite element analysis of energy [29] F. Xu, X. Zhang, and H. Zhang, A review on functionally
absorption characteristics for biomimetic thin-walled multi-cel- graded structures and materials for energy absorption, Eng.
lular structure inspired by horsetails, Mech. Adv. Mater. Struct., Struct., vol. 171, pp. 309–325, 2018. DOI: 10.1016/j.engstruct.
vol. 29, no. 27, pp. 6982–6993, 2022. DOI: 10.1080/15376494. 2018.05.094.
2021.1991059. [30] A. Shiravand, and M. Asgari, Hybrid metal-composite conical
[13] Z. Gao, H. Zhang, J. Zhao, and D. Ruan, The axial crushing tubes for energy absorption; theoretical development and
performance of bio-inspired hierarchical multi-cell hexagonal numerical simulation, Thin. Walled Struct., vol. 145, p. 106442,
tubes, Int. J. Mech. Sci., vol. 239, p. 107880, 2023. DOI: 10. 2019. DOI: 10.1016/j.tws.2019.106442.
1016/j.ijmecsci.2022.107880. [31] X. Xiang, C. Xiao, N.S. Ha, G. Lu, S. Zhang, and Y. Liu, The
[14] Y.-Q. Shi, R.-X. Qin, and B.-Z. Chen, Crashworthiness design quasi-static compressive mechanical properties of barnacle bio-
and analysis for novel multi-corner square columns under axial inspired structures, Eng. Struct., vol. 275, p. 115307, 2023. DOI:
loading, Mech. Adv. Mater. Struct., vol. 29, no. 27, pp. 5843– 10.1016/j.engstruct.2022.115307.
5859, 2022. DOI: 10.1080/15376494.2021.1967529. [32] J. Zhou, S. Liu, Z. Guo, S. Xu, J. Song, and M. Zou, Study on
[15] P. Wang, F. Yang, H. Fan, and G. Lu, Bio-inspired multi-cell the energy absorption performance of bionic tube inspired by
tubular structures approaching ideal energy absorption per- yak horn, Mech. Adv. Mater. Struct., vol. 29, no. 28, pp. 7246–
formance, Mater. Des., vol. 225, p. 111495, 2023. DOI: 10.1016/ 7258, 2022. DOI: 10.1080/15376494.2021.1995088.
j.matdes.2022.111495. [33] X. Zhao, G. Zhu, C. Zhou, and Q. Yu, Crashworthiness analysis
[16] J. Fu, Q. Liu, K. Liufu, Y. Deng, J. Fang, and Q. Li, Design of and design of composite tapered tubes under multiple load
bionic-bamboo thin-walled structures for energy absorption, cases, Compos. Struct., vol. 222, p. 110920, 2019. DOI: 10.1016/
Thin. Walled Struct., vol. 135, pp. 400–413, 2019. DOI: 10. j.compstruct.2019.110920.
1016/j.tws.2018.10.003. [34] X. Deng, S. Qin, and J. Huang, Multiobjective optimization of
[17] H. Liang, W. Hao, H. Sun, Y. Pu, Y. Zhao, and F. Ma, On axially varying thickness lateral corrugated tubes for energy
design of novel bionic bamboo tubes for multiple compression absorption, Mech. Adv. Mater. Struct., vol. 29, no. 25, pp.
load cases, Int. J. Mech. Sci., vol. 218, p. 107067, 2022. DOI: 10. 4259–4272, 2022. DOI: 10.1080/15376494.2021.1924901.
1016/j.ijmecsci.2022.107067. [35] D. Han, Y. Zhang, X.Y. Zhang, Y.M. Xie, and X. Ren,
[18] L. Zhang, Z. Bai, and F. Bai, Crashworthiness design for bio- Mechanical characterization of a novel thickness gradient aux-
inspired multi-cell tubes with quadrilateral, hexagonal and etic tubular structure under inclined load, Eng. Struct., vol. 273,
octagonal sections, Thin. Walled Struct., vol. 122, pp. 42–51, p. 115079, 2022. DOI: 10.1016/j.engstruct.2022.115079.
2018. DOI: 10.1016/j.tws.2017.10.010. [36] X. Qin, Q. Ma, X. Gan, M. Cai, and W. Cai, Failure analysis
[19] L. Zhang, Y. Zhong, W. Tan, C. Gong, Y. Hu, and Z. Bai, and multi-objective optimization of crashworthiness of variable
Crushing characteristics of bionic thin-walled tubes inspired by thickness Al-CFRP hybrid tubes under multiple loading condi-
bamboo and beetle forewing, Mech. Adv. Mater. Struct., vol. tions, Thin. Walled Struct., vol. 184, p. 110452, 2023. DOI: 10.
29, no. 14, pp. 2024–2039, 2022. DOI: 10.1080/15376494.2020. 1016/j.tws.2022.110452.
1849880. [37] H. Zhang, B. Chang, K. Peng, J. Yu, and Z. Zheng, Anti-blast
[20] Y. Zhang, J. Wang, C. Wang, Y. Zeng, and T. Chen, analysis and design of a sacrificial cladding with graded foam-
Crashworthiness of bionic fractal hierarchical structures, Mater. filled tubes, Thin. Walled Struct., vol. 182, p. 110313, 2023.
Des., vol. 158, pp. 147–159, 2018. DOI: 10.1016/j.matdes.2018. DOI: 10.1016/j.tws.2022.110313.
08.028. [38] D. Wang, P. Xu, C. Yang, X. Xiao, and Q. Che, Crashing per-
[21] D. Le, N. Novak, A. Arjunan, A. Baroutaji, Q. Estrada, T. Tran, formance and multi-objective optimization of honeycomb-filled
and H. Le, Crashworthiness of bio-inspired multi-stage nested thin-walled energy absorber with axisymmetric thickness, Mech.
multi-cell structures with foam core, Thin. Walled Struct., vol. Adv. Mater. Struct., vol. 30, no. 11, pp. 2203–2220, 2023. DOI:
182, p. 110245, 2023. DOI: 10.1016/j.tws.2022.110245. 10.1080/15376494.2022.2053765.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 21

[39] Q. Gao, and W.-H. Liao, Energy absorption of thin walled tube [50] X. Deng, Y. Chen, and J. Huang, Crashworthiness analysis of
filled with gradient auxetic structures-theory and simulation, hexagonal hierarchical gradient tubes with axial variable thick-
Int. J. Mech. Sci., vol. 201, p. 106475, 2021. DOI: 10.1016/j. ness inspired by tree fractal structure, Mech. Adv. Mater.
ijmecsci.2021.106475. Struct., p. 2215778, 2023. DOI: 10.1080/15376494.2023.2215778.
[40] H. Liu, Z.X.C. Chng, G. Wang, and B.F. Ng, Crashworthiness [51] X. Zhang, and H. Zhang, Energy absorption limit of plates in
improvements of multi-cell thin-walled tubes through lattice thin-walled structures under compression, Int. J. Impact Eng.,
structure enhancements, Int. J. Mech. Sci., vol. 210, p. 106731, vol. 57, pp. 81–98, 2013. DOI: 10.1016/j.ijimpeng.2013.02.001.
2021. DOI: 10.1016/j.ijmecsci.2021.106731. [52] J. Fang, Y. Gao, G. Sun, Y. Zhang, and Q. Li, Parametric ana-
[41] J. Zhou, C. Dong, Z. Wang, B. Chen, R. Qin, and X. Niu, lysis and multiobjective optimization for functionally graded
Approaching ideal energy absorption through the multicellular foam-filled thin-wall tube under lateral impact, Comput. Mater.
structure with gradient material distribution, Int. J. Mech. Sci., Sci., vol. 90, pp. 265–275, 2014. DOI: 10.1016/j.commatsci.2014.
vol. 225, p. 107355, 2022. DOI: 10.1016/j.ijmecsci.2022.107355. 03.044.
[42] M. Yang, B. Han, P. Su, F. Li, Z. Zhao, Q. Zhang, Q. Zhang, Z. [53] X. Zhang, and H. Zhang, Some problems on the axial crushing
Hong, and T.J. Lu, Oblique crushing of truncated conical sand- of multi-cells, Int. J. Mech. Sci., vol. 103, pp. 30–39, 2015. DOI:
wich shell with corrugated core, Mech. Adv. Mater. Struct., vol. 10.1016/j.ijmecsci.2015.08.026.
28, no. 23, pp. 2458–2471, 2021. DOI: 10.1080/15376494.2020. [54] S. Pirmohammad, S. Ahmadi-Saravani, and J. Zakavi S,
1743396. Crashworthiness optimization design of foam-filled tapered dec-
[43] S. Pirmohammad, and Z. Ghahremanzadeh, Crushing behavior agonal structures subjected to axial and oblique impacts, J.
of multi-cell tubes with a novel pattern of design for their Cent. South Univ., vol. 26, no. 10, pp. 2729–2745, 2019. DOI:
cross-section under multiple crushing angles, Mech. Adv. 10.1007/s11771-019-4209-1.
Mater. Struct., vol. 29, no. 28, pp. 7441–7458, 2022. DOI: 10. [55] N.S. Ha, T.M. Pham, H. Hao, and G. Lu, Energy absorption
characteristics of bio-inspired hierarchical multi-cell square
1080/15376494.2021.2000079.
tubes under axial crushing, Int. J. Mech. Sci., vol. 201, p.
[44] A. Asanjarani, A. Mahdian, and S.H. Dibajian, Comparative
106464, 2021. DOI: 10.1016/j.ijmecsci.2021.106464.
analysis of energy absorption behavior of tapered and grooved
[56] N.S. Ha, and G. Lu, A review of recent research on bio-inspired
thin-walled tubes with the various geometry of the cross sec-
structures and materials for energy absorption applications,
tion, Mech. Adv. Mater. Struct., vol. 27, no. 8, pp. 633–644,
Compos. Part B: Eng., vol. 181, p. 107496, 2020. DOI: 10.1016/
2020. DOI: 10.1080/15376494.2018.1488311.
j.compositesb.2019.107496.
[45] N. Jafarian, and M.J. Rezvani, Crushing behavior of multi-com-
[57] Y. Zhang, M. Lu, G. Sun, G. Li, and Q. Li, On functionally
ponent conical tubes as energy absorber: a comparative analysis
graded composite structures for crashworthiness, Compos.
between end-capped and non-capped conical tubes, Eng. Struct., vol. 132, pp. 393–405, 2015. DOI: 10.1016/j.compstruct.
Struct., vol. 178, pp. 128–135, 2019. DOI: 10.1016/j.engstruct. 2015.05.034.
2018.09.092. [58] Y. Zhang, N. He, X. Song, T. Chen, and H. Chen, On impact-
[46] T. Pang, G. Zheng, J. Fang, D. Ruan, and G. Sun, Energy ing mechanical behaviors of side fractal structures, Thin.
absorption mechanism of axially-varying thickness (AVT) mul- Walled Struct., vol. 146, p. 106490, 2020. DOI: 10.1016/j.tws.
ticell thin-walled structures under out-of-plane loading, Eng. 2019.106490.
Struct., vol. 196, p. 109130, 2019. DOI: 10.1016/j.engstruct.2019. [59] J. Wang, Y. Zhang, N. He, and C.H. Wang, Crashworthiness
04.074. behavior of Koch fractal structures, Mater. Des., vol. 144, pp.
[47] R. Lu, X. Liu, S. Chen, X. Hu, and L. Liu, Axial crashing ana- 229–244, 2018. DOI: 10.1016/j.matdes.2018.02.035.
lysis for tailor rolled square tubes with axially graded both wall [60] Y. He, X. Zhang, J. Sun, X. Shu, X. Li, B. Su, and G. Xiao,
thickness and material strength, Thin. Walled Struct., vol. 117, Energy absorption of self-similar inspired multi-cell hexagonal
pp. 10–24, 2017. DOI: 10.1016/j.tws.2017.04.001. tubes, Eng. Struct., vol. 277, p. 115455, 2023. DOI: 10.1016/j.
[48] G. Li, Z. Zhang, G. Sun, F. Xu, and X. Huang, Crushing ana- engstruct.2022.115455.
lysis and multiobjective optimization for functionally graded [61] Y. He, T. Jin, X. Li, J. Qiu, and X. Shu, Crashworthiness design
foam-filled tubes under multiple load cases, Int. J. Mech. Sci., of hexagonal tubes using self-similar inspired structures, Mater.
vol. 89, pp. 439–452, 2014. DOI: 10.1016/j.ijmecsci.2014.10.001. Today Commun., vol. 33, p. 104934, 2022. DOI: 10.1016/j.
[49] G. Li, F. Xu, G. Sun, and Q. Li, A comparative study on thin- mtcomm.2022.104934.
walled structures with functionally graded thickness (FGT) and [62] W. Liu, Z. Lin, N. Wang, and X. Deng, Dynamic performances
tapered tubes withstanding oblique impact loading, Int. J. of thin-walled tubes with star-shaped cross section under axial
Impact Eng., vol. 77, pp. 68–83, 2015. DOI: 10.1016/j.ijimpeng. impact, Thin. Walled Struct., vol. 100, pp. 25–37, 2016. DOI:
2014.11.003. 10.1016/j.tws.2015.11.016.