See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/291553290

Crushing analysis of thin-walled beams with various section geometries

under lateral impact

Article  in  Thin-Walled Structures · May 2016

DOI: 10.1016/j.tws.2016.01.017

CITATIONS READS

12 522

4 authors, including:

Hanfeng Yin
Hunan University
49 PUBLICATIONS   684 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

surrogate model, optimization, composite structure, crashworthiness View project

All content following this page was uploaded by Hanfeng Yin on 04 June 2019.

The user has requested enhancement of the downloaded file.

 Thin-Walled Structures 102 (2016) 43–57

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: www.elsevier.com/locate/tws

Full length article

Crushing analysis of thin-walled beams with various section

geometries under lateral impact
Tao Tang, Weigang Zhang, Hanfeng Yin n, Han Wang
State Key Laboratory of Advanced Design and Manufacture for Vehicle Body, Hunan University, Changsha, Hunan 410082, PR China

art ic l e i nf o a b s t r a c t

Article history: Due to the advantages of light weight and excellent energy absorption capacity, thin-walled beams are
Received 27 August 2015 widely used as strengthen parts or energy absorbers in vehicle body. Thus, the collapse behaviors and
Received in revised form mechanical properties of thin-walled beams under the static and dynamic loadings have drawn great
11 January 2016
attentions of the researchers. In vehicle side crash accident, the contact parts of the vehicle usually
Accepted 16 January 2016
deformed in bending mode. Thus, it is significantly important to investigate the bending collapse be-
haviors of these parts. In this study, the bending behaviors of several thin-walled beams with simple
Keywords: cross section subjected to lateral impact were investigated using analytical and numerical methods. The
Thin-walled beam crashworthiness parameters such as energy absorption (EA), average crash force (Favg), peak crash force
Lateral impact
(Fmax) and crash force efficiency (CFE) were employed to evaluate the bending resistant property. In order
Bending collapse
to study how geometry shapes affect the bending performance, the main geometry parameters such as
Energy absorption
radius/radius ratio of circular/ elliptical section, side length ratio of rectangular section, height and base
angle of hat section were chosen as the design parameters. After investigating the beams with simple
cross sections, some new beams with complex cross sections were constructed by combing these simple
beams or adding reinforce ribs to these simple beams. Then, the bending performance of these new
beams was studied by numerical simulation and compared with original section beams. Result shows
that the section type b performs better in SEA and CFE than original simple section beams. It also can be
concluded that adding ribs can greatly improve the bending resistant performance for circular, rectan-
gular and hat section beams, and the vertical ribs can achieve the best reinforce effect for the circular and
rectangular section beams.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction rectangular and square prismatic beams was made by Kecman

Due to the advantages of low price, light weight, high strength
and stiffness, high reliability and excellent energy absorption ca-
pacity, thin-walled beams are widely used in crashworthiness
applications such as automobile, train and aeronautical industries
to protect passengers from severe injury [1–3]. To better under-
stand the crashing mechanism and find more effective absorbers, a
lot of research works on the axial crushing behavior of thin-walled
beams have been carried out [4–6]. However, for side impact, the
thin-walled beams will collapse in bending mode [7–8]. A study
on the real world vehicle crashes presented by Kallina showed that
up to 90% involved structural members failed in bending collapse
mode [9]. Therefore, the bending behaviors of thin-walled mem-
bers should be considered when vehicle body is designed.
The first comprehensive study on deep bending collapse of
n
Corresponding author.
E-mail address: yinhanfeng@hnu.edu.cn (H. Yin).
[10]. In his study, simple failure mechanisms involving stationary
and moving plastic hinge lines were proposed, and the moment-
rotation characteristic was calculated in the post-failure range.
However, the theory is semi-empirical because the rolling radius
of moving hinge lines needs to be acquired from tests. A similar
approach was developed by Abramowicz [11]. Koteko investigated
the yield line mechanism (YLM) of rectangular and trapezoidal box
section beams with a high width to depth ratio compared to
Kecman’s sections [12]. Park developed a solution for the bending
collapse of trapezoidal section beams with different folding
lengths on the top and bottom flanges and minimization of the
total plastic energy rate included to determine the mode para-
meters [13]. This method however did not consider the mean
crumpling moment. Kim and Reid gave a kinematically admissible
new model and proposed a more accurate theoretical solution to
predict the bending collapse of thin-walled beams based on some
existing analytical relationship [14]. Huang and Lu proposed an
empirical formula based on dimensional analysis [15]. The

http://dx.doi.org/10.1016/j.tws.2016.01.017
0263-8231/& 2016 Elsevier Ltd. All rights reserved.
44 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57

crushing behavior of circle tube under bending has also been

studied analytically and experimentally. Elchalakani conducted a
plastic mechanism analysis for thin-walled hollow section tubes
deforming in a diamond or multi-lobe collapse mode under large
deformation pure bending [16]. The predicted post-bulking mo-
ments and slopes of the collapse curves showed very good
agreement with those obtained from experiments carried out by
the researchers. Based on Elchalakani’s research, Soheila applied
the YLM technique using the energy method to investigate the
collapse behavior and energy absorption capability of circle tubes
under large deformation due to lateral impact load [17]. In his
study, analytical solutions for the collapse curve and in-plane ro-
tation capacity were also developed and used to model the de-
formation behavior and energy absorption. Apart from closed form
section thin-walled beams, the bending behavior and crushing
mechanism of open section beams was also investigated by re-
searchers. Liu and Day studied the bending collapse of the thin-
walled channel section beams and derived their approximation
moment-rotation characteristics [18].
Fig. 1. YLM for circular tubes.
In order to achieve better bending performance and weight
efficiency in energy absorption, ultra-light metal fillers such as
aluminum foam, aluminum honeycomb introduced into the thin-
walled structures has attracted increasing interest. Chen con-
ducted the study on the crush performance of sheet aluminum
foam-filled sections in compression and bending modes [19–20]. A
study on the bending collapse of thin-walled beam filled with
aluminum foam or honeycomb was also carried out by Chen [21].
In his study, the strengthening effects of ultra-light metal fillers
were quantified through numerical and experiment method. The Fig. 2. YLM for rectangular tubes.
results showed the potential of significant weight saving and vo-
lume reduction by utilizing metal filler. It is found that thin-walled
foam-filled beams as weight-efficient crash energy absorber have
promising future. Santose and Wierzbicki carried out quasi-static
three point bending simulations and experiments to study the
effect of foam-filling on the bending resistance of a thin-walled
beam [22–24]. The results showed that the presence of foam filler
retards the inward sectional collapse at the compressive flange
and changes the crushing mode from a single stationary fold to a Fig. 3. Scheme of dynamic three-point bending test for circular and square beam.
multiple propagating fold which prevents the drop in the loading
capacity due to local section collapse. It was shown that filling of
Table 1
metal foam could improve the load-carrying capacity by offering
Mechanical properties of material.
additional support from inside and increase the energy absorption.
McGregor conducted experiments for aluminum hat section Material Density/kg/m3 Yield Young's Poisson's Tangent
beams undergoing cantilever bending [25]. stress, modulus, E/ ratio modulus/
sy/MPa GPa MPa
In this paper, the bending collapse behaviors of different cross
sections thin-walled beams under lateral impact were investigated AISI1080 7860 869 205 0.28 5669
through theoretical analysis and numerical method. The force- AlMg0.5F22 2700 227 68.6 0.29 321
displacement curves were obtained and the crashworthiness in-
dexes such as EA, Favg, Fmax, CFE were adopted to analysis the
bending resistance performance of thin-walled beams. The effect time. Thus, the theoretical analysis is an efficient approach for the
of geometry parameters of each section on the bending behavior early step of design.
of beam was studied. After that, several cross sections and When circular tube is subjected to bending load, the plastic
strengthen schemes were proposed and studied to better under- deformations occur over some folding lines which are called
stand the effect of topology pattern on the bending behavior of “hinge lines”. When hinge lines are completed around the crush-
thin-walled beams. ing area of the structure, global or local collapse will progress. The
internal energy of structure is determined by the summation of
plastic energy dissipated in each hinge line. In S. Poonaya’s study
2. Bending theory of thin-walled beams [26], the collapse mechanism of circular tube subjected to bending
is divided into three phases: elastic behavior, ovalisation plateau
2.1. Circular section beam and structural collapse, and different deformation mode occurs in
each phase. Under each phase, the internal energy dissipation
Generally, there are several approaches to determine the me- rates were calculated for each of the hinge lines which were de-
chanical properties of structural members such as finite element fined in terms of velocity field. The experiment was conducted
analysis, experiments and theoretical analysis, among which ex- with a number of tubes having various diameter to thickness ra-
perimental approach costs much and finite element method pro- tios, and the theoretical prediction and experimental results
vides relatively accurate results but usually requires extensive showed great agreement.
 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57 45

Table 2
Details of three-point bending test.

Section Material D/mm Thickness/mm L/mm S/mm R1/mm R2/mm m/kg v/(m/s)

Circular AISI1080 28 2 960 760 100 50 148 3.4

Square AlMg0.5F22 55 1.5 550 430 50 50 128 4.4

Fig. 4. Deformation pattern comparison of circular beam.

Fig. 5. Deformation pattern comparison of square beam.

Fig. 6. Experimental and numerical impact force curves for the circular beam.
Fig. 7. Experimental and numerical impact force curves for the square beam.

Based on S. Poonaya's theory, Soheila [17] developed the yield

line mechanism (YLM) model for hollow circular section tube Table 3
Unified simulation details for thin-walled beams.
using experimental observations. The established YLM model is
shown in Fig. 1. Through establishment of YLM model and calcu- Material L/mm S/mm R1/mm R2/mm m/kg v/(m/s)
lation of the energy absorption of each hinge line, the total ab-
sorbed energy can be estimated and a failure curve can be plotted. AlMg0.5F22 960 745 100 50 200 5.8
The total energy absorption of the YLM model is the sum of each
individual hinge line work expressed as bellows [17]:
46 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57

Fig. 8. Finite element model of circular beam.

Fig. 9. Profile and parameters of simple sections.

Fig. 10. Deformation patterns of simple profile section thin-walled beams: (a) circular, (b) elliptical, (c) rectangular, (d) trapezoidal, (e) hat section.

n
dW
W (θ ) = ΣWi (θ ), M (θ ) = ,
i (1) dθ (2)

where Wi is the absorbed energy of hinge line, n is the number of where W is the total absorbed energy, θ is the rotation angle of
hinge line in the model. beam. While the ending moment of different rotation angle is
The associated bending moment can be expressed as: determined, the moment-rotation graph can be plotted. The work
components for each plastic hinge are also derived by Soheila [17].
 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57
Fig. 11. Comparison of hinge mechanism of simple profile section thin-walled beams at various stages: (a) circular, (b) elliptical, (c) rectangular, (d) trapezoidal, (e) hat
section.
2.2. Rectangle section beam
Fig. 2 shows the YLM model of rectangle tubes under bending
collapse developed by Kecman [10], this model is based on the
observation in actual accidents involving vehicles and in static and
dynamic laboratory tests of “bus rings”. The simplified hinge me-
chanism was verified by fifty six quasi-static bending tests for
twenty seven different sections. In Kecman’s model, the bound-
aries between plastic deformation segments and undeform por-
tion are clearly defined by stationary hinge line. The collapse
process is identified four distinct phases. In the first phase, a bulge
starts and continues up to a hinge rotation of 5–10° below the
compression flange. In the second phase, the rolling deformation
appears and continues up to 25–35°. In the third phase, new
bending lines are produced under the summit point of the bulge
on the web. In the final phase, contact occurs between the two
buckled halves of compression flange. The theoretical model of
rectangular beam under bending is based on the second phase of
collapse. The calculation of energy absorption of structure can is
based on the assumed collapse mechanism with specific yield lines
length, relative rotation angles and rolling radius. The total energy
absorption can be expressed as:
W (θ ) = W1(θ ) + W2(θ ),
where W1 and W2 present the absorbed energy of stationary and
travelling hinge lines respectively, θ is the rotation angle.
3. Numerical simulations
3.1. Finite element modeling
In this study, the numerical model is developed with explicit
non-linear finite element code LS-DYNA. Fig. 3 shows the thin-
walled beam subjected to lateral impact. The thin-walled beam
lays on two cylindrical supports and the diameters and span are R2
and S, respectively. A cylindrical punch with a diameter of R1 and a
47
mass of m impacts onto the beam with an initial velocity of v in
the mid-span. Belytschko-Tsay reduced integration shell elements
with five integration points through the shell thickness were
employed to model the beam, support and punch. Because the
supports and punch would be almost not deformable compared to
the thin-walled beam, the supports and punch were treated as
rigid bodies. To avoid spurious zero energy deformation modes
and volumetric locking, stiffness-based hourglass control and re-
duced integration were used. To account for contact between the
punch and beam, and the contact between the supports and beam,
“Automatic surface to surface” contacts were defined with coeffi-
cient friction of 0.3. The “Automatic single surface” contact was
also used to prescribe the beam itself to avoid interpenetration of
beam folding during bending collapse.
The materials of thin-walled beams considered herein were
high-strength steel AISI1080 and aluminum alloy AlMg0.5F22. To
simulate the material properties, Mat 24, Piecewise-Linear-Plasti-
city was used to model elastic-plastic materials. The mechanical
properties of materials are shown in Table 1. The strain rate effect
was neglected as the impact velocity is relatively low.
3.2. Validation of finite element model
(3)
To validate the simulation models, the circular and square
beams were first investigated under three-point bending test. The
experiment details are shown in Table 2. Figs. 4 and 5 display the
deformation pattern comparison of circular and square beams
between the experiment and simulation, respectively. It can be
seen that the deformation patterns of circular and square in si-
mulation showed good agreement with those in test. Figs. 6 and 7
display the comparison of impact force versus punch displacement
curves between the experiment and the simulation. It also can be
seen that the force versus displacement curve of simulation was
able to capture the true bending behavior. Thus, the validation of
the developed finite element models of circular and square beams
can be verified against the experimental data, and the simulation
results agreed fairly well with the corresponding experimental
48 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57

results. As a result, the FE models developed herein were accurate

and effective for the subsequent analysis.

4. Bending analysis of simple profile section beam

In order to further explore the bending behaviors and energy

absorption characteristics of thin-walled structures, more finite
element simulation models of beams with different cross sections
were developed based on the three point bending test as shown in
Fig. 3. To get comparative results, the experiment conditions were
set the same and specified in Table 3. The masses of beams were
set equivalent by adjusting wall thickness to highlight the effect of
cross section shapes. As an example, the finite element model of
circular beam is shown in Fig. 8.

4.1. Crashworthiness indicators

In order to evaluate the mechanical property of the thin-walled

beam under bending collapse, crashworthiness indices such as
energy absorption (EA), average crash force (Favg), specific energy
absorption (SEA), peak crashing force (Fmax) and crash force effi-
ciency (CFE) were used in the study.
The energy absorption of a structure under bending condition
can be expressed as:
δ
EA = ∫0 F (x)dx,
(4)

where δ is the deformation distance and F(x) is the crash force.

The average crash force is the mean crash force expressed as
the total absorbed energy divided by the corresponding crushing
displacement and can be written as:
EA
Favg = ,
δ (5)

The specific absorbed energy is defined as the ratio of the ab-

sorbed energy to the mass of the structure, as:
EA
SEA = ,
M (6)

where M is the total mass of structure.

The CFE is defined as the ratio of the average crash force to the
maximum crash force, which can be given as:
Favg
CFE = ,
Fmax (7)

where Favg is the average crash force, Fmax is the maximum crash
force during the whole collapse progress.
In the paper, the crash distance of punch was set as 200 mm
and the mass values of beams were the same. Thus, the EA, SEA
and Favg were in proportion to each other. The energy absorption
(EA) was chosen as the main indicator to evaluate the bending
resistance of thin-walled beams. When the beam has a bigger EA
value, it means the beam can absorb more energy in the same
crashing distance, and the bending resistance performance of the
beam is better. The peak crash force can have bad influence to the
passenger protection, thus the Fmax should have a relative small
value. The bending stability can be evaluated with CFE. If the value
of CFE is higher, it means the crash force is more uniformly dis-
tributed in the crashing distance thus the bending stability of the
beam is better.
Fig. 12. Crashworthiness indicators of simple profile section thin-walled beams: (a)
Fmax & Favg, (b) EA, (c) SEA, (d) CFE.
4.2. Sections and parameters

The simple cross sections chosen to study in this paper were

 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57 49

Table 4
Section dimensions and thicknesses for simple profile section beams.

No. Section Geometric parameters Thickness/mm Mass/kg

R/mm a/mm b/mm A/deg H/mm

1 Circle 10 – – – – 3.3 1.55

2 Circle 15 – – – – 2.3 1.63
3 Circle 20 – – – – 1.7 1.61
4 Ellipse – 20 10 – – 2.3 1.68
5 Ellipse – 10 20 – – 2.3 1.68
6 Rectangle – 50 25 – – 1.4 1.60
7 Rectangle – 40 40 – – 1.4 1.70
8 Rectangle – 25 50 – – 1.4 1.60
9 Rectangle – 30 30 – – 1.76 1.60
10 Trapezoid – – – 70 30 1.6 1.75
11 Trapezoid – – – 75 30 1.6 1.72
12 Trapezoid – – – 100 30 1.6 1.70
13 Trapezoid – – – 110 30 1.6 1.75
14 Hat – – – 70 25 2.3 1.57
15 Hat – – – 70 30 2.1 1.57
16 Hat – – – 70 35 1.96 1.6
17 Hat – – – 70 40 1.85 1.63
18 Hat – – – 60 35 1.96 1.58
19 Hat – – – 80 35 1.96 1.65

Fig. 13. Force-Displacement of circular beam with different radius. Fig. 15. Force-Displacement of rectangle beam with different side length ratio.

Fig. 14. Force-Displacement of ellipse beam with different radius ratio. Fig. 16. Force-Displacement of trapezoid beam with different base angle.

circle/ellipse section, rectangle/ trapezoid section and hat section, different geometry features would be studied in the following
the topology profiles and main geometry parameters of the sec- chapters.
tions are presented in Fig. 9. The bending behaviors of beams with
50 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57
Fig. 17. Force-Displacement of hat section beam with different height.
Fig. 18. Force-Displacement of hat section beam with different base angle.
Fig. 19. Force-Displacement of circular, rectangular and hat section beams.
4.3. Plastic hinges analysis
The plastic deformation of thin-walled beams under bending
mode is localized in stationary and moving hinges [27]. The wall of
beams between every two adjacent hinges either moves rigidly or
experiences small deformation. Therefore, the energy absorption
capacities of thin-walled beams under bending collapse depend on
the number of hinge lines created and the amount of deformation
in the beam wall. The deformation patterns and hinge patterns of
circular, elliptical, square, trapezoidal and hat section beams are
shown in Figs. 10 and 11 respectively. For closed profile sections,
the inward penetration of walls will greatly influence the load
carrying capacity and bending stability of beam in the post col-
lapse stage. The decrease of cross section area in plastic hinge zone
will lead to poor load carrying capacity and drop of energy ab-
sorption of thin-walled beams. As illustrated, the circular, square
and trapezoidal sections were easier to collapse into flat shape
when compared to elliptical section. Thus, the bending behavior of
elliptical section beams is better than others, which was also
shown in the load-displacement curves. It also can be seen that
the hinge mechanism of the hat section beam is a lot different
from closed profile sections, which was caused by insufficient
support to the side walls. Among the five sections, hat section was
mostly seriously compressed after bending, thus the hat section
beams had the poorest bending stability and energy absorption
capacity, which could be proved by further parameter analysis.
4.4. Parameter analysis
4.4.1. Effect of radius on circular section
In order to explore the effect of radius on the crashworthiness
of circular section beam under lateral impact, the value of R was
varied with 10, 15 and 20 mm. The crashworthiness parameters of
thin-walled beams are shown in Fig. 12 and Table 4.The force-
displacement curves of circle section beams are shown in Fig. 13. It
could be seen that radius has noticeable effect on the bending
behavior of circular beams. Specifically, larger values of radius
could increase the values of Fmax, and it also caused the decreasing
of CFE. It meant that the bending stability of circular beams aw-
fully deteriorated with the increasing of radius. In other words,
when radius was increased and beam mass was kept unchanged,
the wall of beam would get thinner and plastic hinge would be
easy to be developed, thus the load carrying capacity dropped. The
drop of crash force was greatly influenced by the development of
plastic hinge. The smaller the radius, the later the plastic hinge
was developed, and the better the bending stability of beam is.
Among the three studied circular beams, the beam with radius of
15 mm absorbed the most energy and had the biggest SEA value,
which meant it had the best composite bending resistance
performance.
4.4.2. Effect of radius ratio on elliptical section
Elliptical section could be treated as the extension of circular
section. For circular section beams, when the section radius in-
creased, the bending stability of beam got better, but the crash
force decreased at the same time, which made it hard to have the
best bending resistance performance. Here, the bending resistance
of elliptical section beams was investigated. The geometry shape
of elliptical section is mainly determined by radius along the major
axis and minor axis, so the effect of radius ratio (η ¼b/a) on the
bending resistance and energy absorption capacity elliptical sec-
tion beams was investigated.
Fig. 14 displays the force-displacement curves of elliptical sec-
tion beams with η values of 0.5, 1 and 2. It was observed that the
radius ratio had noticeable effect on crashworthiness performance
of elliptical section beam. Generally, the larger the radius ratio
was, the larger the crash force was and the later the crash force
started to drop. As for the point of crashworthiness parameters,
the EA, SEA and CFE values all increased monotonically with the
increasing of radius ratio. Most importantly, the SEA of beam with
radius ratio 2 was triple more than that with radius ratio 0.5. In
other words, the improvement of elliptical section beams in both
bending resistance performance and energy absorption capacity
 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57 51

Fig. 20. Cross sections of composited complex profile.

Fig. 21. Deformation patterns of composite profile section thin-walled beams: (a) section a, (b) section b, (c) section c, (d) section d, (e) section e.

could be achieved through increasing of radius ratio. When it came
to deformation analysis, it was found that the elliptical cross sec-
tion with bigger radius ratio experienced more deformation before
it got flat and formed plastic hinges. As a result, the beam became
harder to be bent. However, it should be noted that while circular
section beam had the same bending resistance characteristic in
any loading direction, the bending characteristic of elliptical sec-
tion beam was variable with the change of loading direction. And
when lateral impact happened in the major axis direction, the
ellipse section beam achieved the best bending resistance
performance.
4.4.3. Effect of side length ratio on rectangular section
Considering that the radius ratio had noticeable effect on the
bending performance of elliptical section beams, there is a ques-
tion: whether similar conclusion could be drawn to the rectan-
gular section beams. Herein, the effect of side length ratio (η ¼b/a)
on the bending performance of rectangular section beams was
investigated. Fig. 15 depicts the force-displacement curves of rec-
tangular beams with side length ratio 0.5, 1 and 2. It was found
that the side length ratio also affected the bending performance of
rectangular beams, but not in the same degree as elliptical ones. It
was observed that the peak crash force increased up to a high
value with the increasing of η, which was similar to elliptical
section beams. Also, the EA and SEA of rectangular beams values
got bigger as η was increasing. The SEA of rectangular section
beam with side length ratio 2 was twice more than that with side
length ratio 0.5. It meant that the beam with a higher η could
absorb more energy in bending condition with the same mass.
Comparatively, the effect of η on CFE was more complex, and the
beam with η value 1 had larger CFE than its counterparts. It can be
noticed that a second peak crash force appeared for the beam with
side length ratio 2, and this was considered to be a result of self-
contact in the beam at the post crashing stage.
52 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57
Fig. 22. Hinge mechanism of composite profile section thin-walled beams at various stages: (a) section a, (b) section b, (c) section c, (d) section d, (e) section e.
4.4.4. Effect of base angle on trapezoidal section
The trapezoidal section could be treated as rectangular section
with rotated side walls, thus the effect of base angle A on the
bending performance of trapezoidal section beam was in-
vestigated. The trapezoidal section beams with 70°, 75°, 100° and
110° base angles were chosen to study, the force-displacement
curves are shown in Fig. 16. It could be found that the effect of base
angle on beam bending performance was relatively slight. The SEA
and CFE of these beams were almost the same. But the peak crash
force decreased monotonically when the base angle increased.
4.4.5. Effect of height on hat section
In addition to closed profile sections beams, the open profile
section types such as hat section were also widely used as struc-
ture member or energy absorber. Here, the bending behavior and
crashworthiness of hat section beam were studied through nu-
merical simulation.
Firstly, the effect of section height H on the crashworthiness of
hat section beams was explored. The value of height H was varied
with 25, 30, 35 and 40 mm while the base angle A was fixed to 70°.
Fig. 17 depicted the force-displacement curves of these beams. It
could be observed that the curves had clear tendency to shift from
right to left and from bottom to up when section height H in-
creased from 25 to 40 mm. As a result, the peak crash force
increased and CFE decreased simultaneously. When H increased
from 25 to 40 mm, the CFE decreased from 0.58 to 0.4. In other
words, the bending resistance performance of hat section beams
with higher height was better than that with lower height at early
stage of bending procedure, but dropped rapidly with further
bending of structure. Consequently, it was a complicated re-
lationship between SEA and H, and the beam with H ¼30 mm had
the largest SEA value.
4.4.6. Effect of base angle on hat section
In addition to height, the base angle A is also a main geometry
parameter of hat section and could affect the bending behavior of
beams. We varied the base angle with A ¼60°, 70° and 80° while
fixing section height. The force-displacement curves are shown in
Fig. 18. It could be seen that the base angle had greater influence
on hat section beams than on trapezoidal section beams. Specifi-
cally, the Fmax increased from 13.84 to 16.71 kN, the SEA increased
from 0.73 to 0.96 J/g and CFE increased from 0.42 to 0.47 when A
increased from 60° to 70°. Clearly, the bending resistance and
energy absorption capacity both improved with the increasing of
base angle. It could be explained through deformation analysis,
since the plastic hinges development of hat section beams in-
volved the outward fold of side walls, a bigger base angle could
delay the development of plastic hinge thus improved the bending
 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57 53

Table 5
Thicknesses and masses of complex profile section beams.

Section Thickness/mm Mass/kg

a 1.1 1.78
b 1.27 1.72
c 1.27 1.72
d 1.4 1.73
e 1.7 1.70
Ⅰ 1.52 1.73
Ⅱ 1.52 1.73
Ⅲ 1.82 1.71
Ⅳ 1.82 1.71
Ⅴ 1.9 1.61
Circular 2.3 1.63
Square 1.76 1.60
Hat 1.96 1.6

Fig. 24. Force-Displacement of composited complex profile.

resistance of beams. It should be noted that the hat section beams

had the lowest SEA and CFE compared to other section type beams,
which was a result of their special bending collapse mode.

4.5. Comparative analysis of simple section beams

Considering that the bending performance of thin-walled beam

Fig. 23. Crashworthiness indicators of composite profile section thin-walled
beams: (a) Fmax & Favg, (b) EA, (c) SEA, (d) CFE.
had preference to the geometric parameters of cross section, the
beam sections with the same geometry size were chosen to con-
duct comparative analysis. The beam No. 2, No. 9 and No. 15 which
had circular, rectangular and hat cross section respectively were
chosen herein. In Fig. 19 the loading force verse displacement
curves of these beams were presented. Here, it could be seen that
the loading conditions of these beams shared the same tendency
which was comprised by three different stages. At the first stage,
the walls of beam were going to be hinged and the contact area
between beam and impactor was going to increase, which caused
the result of maximum load. At the second stage, the loading force
remained in a relatively stable level with the further development
of plastic hinge. At the last stage, the plastic hinge was completely
developed and load bearing capacity of beam suffered a sufficient
drop which led to the decrease of contact force.
After comparison of force-displacement curves, it was found
that the recorded peak force of rectangular section beam was
somewhat higher than that of circular and hat section beams.
Among these beams, the loading force of hat section beam ex-
perienced the harshest drop at the third stage. It meant the beam
could hardly stand the lateral impact load at the post plastic
bending stage. Compared with the other two beams, the loading
force of circular beam was better distributed in the crashing
54 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57

Fig. 25. Rib arrangement of simple profile sections.

Fig. 26. Deformation patterns of ribbed thin-walled beams: (a) section I, (b) section II, (c) section III, (d) section IV, (e) section V.

distance which would result in good bending resistance perfor-
mance. As for the point of energy absorption capacity, the SEA of
circular beam was 10.8% and 36.46% greater than that of rectan-
gular and hat section beams respectively. The CFE of circular, rec-
tangular and hat section beams were 0.74, 0.60 and 0.47 respec-
tively, which meant that the bending stability of circular beam was
better than its counterparts.
circular, square and hat sections; the second method was to add
reinforce ribs to simple profile sections. To distinguish from ori-
ginal sections, we classified these new constructed sections as
“complex profile sections”. The crashworthiness of these complex
profile sections thin-walled structures was investigated using the
same analysis method as used before, and their bending resistance
property were compared with the original simple profile
structures.

5. Topological modification 5.1. Profile combination

After the bending behaviors of different simple profile section
beams under lateral impact were investigated, there was a ques-
tion: if we could generate some new cross sections which had
better bending resistance capacity than these sections studied
above or how could we reinforce these commonly seen section
beams to achieve better load carrying capacity. To answer this
question, two kinds of methods were adopted to modify the cross
section and generate new thin-walled structures. The first method
was to combine the existing simple profile sections such as
The complex sections generated by combining simple sections
are shown in Fig. 20. Similarly, the masses of these complex profile
beams were also uniformed by adjusting the thickness of beam to
better evaluate the effect of section change on the crashworthiness
of beam. The bending conditions and material parameters were
exactly the same as those of simple profile section beams.
The plastic deformation pattern and hinge mechanism of these
complex section beams are shown in Figs. 21 and 22 respectively.
Compared with the simple profile section beams, the collapse
 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57 55

Fig. 27. Hinge mechanism of ribbed thin-walled beams: (a) section I, (b) section II, (c) section III, (d) section IV, (e) section V.

modes of these complex section beams were a lot more complex.

It also could be seen that the plastic hinge of beams with section b
was not completely developed in the crashing distance and more
material near the impact zone was involved in deformation. As a
result, the bending resistance capacity of beam was greatly
improved.
The crashworthiness parameters are shown in Fig. 23 and Ta-
ble 5. And the force-displacement curves of these beams are
shown in Fig. 24. It could be found that section type b beam had
the greatest SEA and CFE of 1.49 J/g and 0.87 respectively which
meant it had better crashworthiness performance than its coun-
terparts. Obviously, section type a beam had the greatest
Fmax ¼16.99 kN. Compared with square section beam, the SEA and
CFE of section type b beam had been improved by 25% and 25%
respectively. As for section type c and d, the SEA was less than that
of square tube, but the Fmax had been decreased by 18% and 8% and
CFE increased simultaneously. When compared with circular sec-
Fig. 28. Force-Displacement of ribbed simple profile beam.
tion beam, only section type b showed improvement in SEA and
CFE with range of 14% and 18% respectively. As hat section beam
56 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57
was comparatively a weak type in bending resistance behavior,
both section type d, and e showed great improvement in SEA and
CFE compared with hat section. For type d, the SEA and CFE values
increased by 29% and 70% respectively. For type e, the data were
54% and 79%.
5.2. Adding ribs
To strength the rigidity of structures, ribs are prevalently used
in thin-walled beams. To explore how the ribs’ arrangement could
modify the energy absorption capacity and structure rigidity of
thin-walled beams, various ribs were employed to the simple
profile sections beams whereas all of them had the same material
and thickness as the beams had. As shown in Fig. 25, beam I and III
were employed with a horizontal reinforce rib while beam II and
IV were employed with a vertical reinforce rib. Considering the
geometry feature of hat section, two vertical reinforce ribs were
employed adjacent to the impact area.
Fig. 26 shows the bending deformation patterns of ribbed
simple profile section beams. It could be seen that the ribbed
beams commonly underwent less deformation and had larger
deformation zone compared with the unribbed ones. The hinge
patterns of these ribbed beams are shown in Fig. 27. Usually, the
development of plastic hinge in the impact area would tre-
mendously influence the bending resistance performance of thin-
walled tubes, and the adding of ribs could effectively change the
hinge mode or hinder the development of plastic hinge. As a re-
sult, the bending performance would be greatly improved.
Fig. 28 depicts the force-displacement curves of these ribbed
beams. It could be clearly seen that the crashing force of section
type II, III and IV did not experience dramatic drop at the post
bending stage. From Fig. 24, it was found that the plastic hinge of
beam was not completely formed. For both rectangular and cir-
cular section beam, the arrangement of vertical ribs could have
better strengthen effect than horizontal ribs. It was observed that
the SEA and CFE of section type II were, respectively, 55% and 48%
greater than that of unribbed square section, and the data for
section type I were 3% and 28%. Moreover, the SEA and CFE of
section type IV were, respectively, 18% and 19% greater than that of
unribbed circular section, and the data for section type III were
5% and 19%. For the hat section beam, the SEA and CFE were 22%
and 58% improved after adding ribs.
6. Conclusions
(1) In bending collapse mode, the thin-walled beam usually de-
forms in the form of plastic hinge which greatly influences the
bending resistance performance of beam. While the plastic
hinge modes of thin-walled beams with different sections
have significant difference, the force-displacement curves of
these beams have the same tendency. In the bending collapse
mode, the crash force of thin-walled beam first reaches a peak
value, followed by preserving and then drops. Since different
section beams have diverse characters in the bending proce-
dure, the actual bending resistance performance is different.
Comparatively, the elliptical beam has better bending re-
sistance performance and energy absorption capacity than its
counterparts, and hat section beam has the worst bending
performance.
(2) The cross section geometry parameters affect the bending
resistance performance of thin-walled beams under lateral
impact in different range. Basically, the parameters related to
geometry size along impact direction, such as: radius of cir-
cular section, radius ratio of elliptical section, side length ratio
of rectangular section and height for hat section, have sig-
nificant influence on crashworthiness of beams under lateral
impact. Usually, the section with a larger size along impact
direction will result in a better crashworthiness performance.
(3) After compositing the simple profile sections or adding re-
inforcing ribs, some new thin-walled beams with complex
profiles were constructed. Among the composite structures,
the beam with section type b showed the best improvement in
bending resistance performance and crashworthiness prop-
erty. Adding ribs is an effective way to improve the bending
resistance performance of thin-walled beams. For the closed
profile section types, the vertical ribs have better performance
than horizontal ones. For the open profile such as hat section,
the adding of ribs will greatly improve the energy absorption
and bending stability thus leads to an enhanced bending
performance.
Acknowledgments
The support from Key Project of National Natural Science
Foundation of China (51275164).
References
[1] A.A.A. Alghamdi, Collapsible impact energy absorbers: an overview, Thin-
walled Struct. 39 (2) (2001) 189–213.
[2] W. Abramowicz, Thin-walled structures as impact energy absorbers, Thin-
Walled Struct. 41 (2) (2003) 91–107.
[3] Guangyong Sun, et al., Dynamical bending analysis and optimization design
for functionally graded thickness (FGT) tube, Int. J. Impact Eng. 78 (2015)
128–137.
[4] Tomasz Wierzbicki, Abramowicz Wlodzimierz, On the crushing mechanics of
thin-walled structures, J. Appl. Mech. 50 (4a) (1983) 727–734.
[5] W. Abramowicz, T. Wierzbicki, Axial crushing of foam-filled columns, Int. J.
Mech. Sci. 30 (3) (1988) 263–271.
[6] Markus Seitzberger, et al., Experimental studies on the quasi-static axial
crushing of steel columns filled with aluminium foam, Int. J. Solids Struct. 37
(30) (2000) 4125–4147.
[7] Weigang Chen, Experimental and numerical study on bending collapse of
aluminum foam-filled hat profiles, Int. J. Solids Struct. 38 (44) (2001)
7919–7944.
[8] Zonghua Zhang, Shutian Liu, Zhiliang Tang, Design optimization of cross-
sectional configuration of rib-reinforced thin-walled beam, Thin-walled Struct.
47 (8) (2009) 868–878.
[9] Kallina Ingo, Zeidle Falk, Baumann Karl-Heinz and Scheunert Dieter. The offset
crash against a deformable barrier, a more realistic frontal impact, in: Pro-
ceedings: International Technical Conference on the Enhanced Safety of Ve-
hicles, National Highway Traffic Safety Administration, Vol. 1995, 1995.
[10] Dušan Kecman, Bending collapse of rectangular and square section tubes, Int.
J. Mech. Sci. 25 (9) (1983) 623–636.
[11] W. Abramowicz, Simplified crushing analysis of thin-walled columns and
beams, Rozpr. Inzynierskie 29 (1) (1981) 5–26.
[12] M. Kotelko, T.H. Lim, J. Rhodes, Post-failure behaviour of box section beams
under pure bending (an experimental study), Thin-walled Struct. 38 (2)
(2000) 179–194.
[13] M.S. Park, B.C. Lee, Prediction of mode parameters and moment-rotation
curves for crushed thin-walled trapezoidal tubes in bending, J. Appl. Mech. 63
(2) (1996) 453–459.
[14] T.H. Kim, S.R. Reid., Bending collapse of thin-walled rectangular section col-
umns, Comput. Struct. 79 (20) (2001) 1897–1911.
[15] X. Huang, G. Lu, Bending hinge characteristic of thin-walled square tubes, Int.
J. Crashworth. 10 (3) (2005) 275–285.
[16] Mohamed Elchalakani, Raphael Grzebieta, Xiao-Ling Zhao, Plastic collapse
analysis of slender circular tubes subjected to large deformation pure bending,
Adv. Struct. Eng. 5 (4) (2002) 241–257.
[17] Soheila Maduliat, et al., Performance of hollow steel tube bollards under quasi-
static and lateral impact load, Thin-Walled Struct. 88 (2015) 41–47.
[18] Y.C. Liu, M.L. Day, Bending collapse of thin-walled beams with channel cross-
section, Int. J. Crashworth. 1 (3) (2006) 251–262.
[19] W. Chen, T. Wierzbicki, Experimental study on the crushing behavior of alu-
minum closed-hat foam-filled sections, Impact & Crashworthiness Laboratory.
No. 26. Report, 1999.
[20] Weigang Chen, Dubravko Nardhi, Experimental study of crush behaviour of
sheet aluminium foam-filled sections, Int. J. Crashworth. 5 (4) (2000)
447–468.
 T. Tang et al. / Thin-Walled Structures 102 (2016) 43–57
[21] W. Chen, T. Wierzbicki, S. Santosa, Bending collapse of thin-walled beams with
ultralight filler: numerical simulation and weight optimization, Acta Mech.
153 (3–4) (2002) 183–206.
[22] Sigit Santosa, John Banhart, Tomasz Wierzbicki, Bending crush behavior of
foam-filled sections. Metal Foams and Porous Metal Structures, Verlag-MIT,
Bremen, 1999, pp. 337–345.
[23] S. Santosa, J. Banhart, T. Wierzbicki, Experimental and numerical analyses of
bending of foam-filled sections, Acta Mech. 148 (1–4) (2001) 199–213.
[24] Sigit Santosa, Tomasz Wierzbicki, Effect of an ultralight metal filler on the
bending collapse behavior of thin-walled prismatic columns., Int. J. Mech. Sci.
View publication stats
57
41 (8) (1999) 995–1019.
[25] I.J. McGregor, D.J. Meadows, C.E. Scott, A.D. Seeds, Impact performance of
aluminum structures, in: N. Jones, T. Wierzbicki (Eds.), Structural Crash-
worthiness and Failure, Elsevier, Amsterdam, 1993.
[26] S. Poonaya, U. Teeboonma, C. Thinvongpituk, Plastic collapse analysis of thin-
walled circular tubes subjected to bending, Thin-walled Struct. 47 (6) (2009)
637–645.
[27] M.H. Shojaeifard, H.R. Zarei, R. Talebitooti, M. Mehdikhanlo, Bending behavior
of empty and foam-filled aluminum tubes with different cross-sections, Acta
Mech. Solida Sin. 25 (6) (2012) 616–626.