Experimental and Numerical Assessment of Crash Behavior of Welded Thin Wall Rectangular Steel Sections

52© 2010 AMAE
DOI:1.IJPIE.01.01.53
AMAE Int. J. on Production and Industrial Engineering, Vol. 01, No. 01, Dec 2010
Experimental and Numerical Assessment of Crash
Behavior of Welded Thin Wall Rectangular Steel
Sections
Veerasamy M and Raghu V. Prakash
Department of Mechanical Engineering, Indian Institute ofTechnology Madras, Chennai, India
Email: raghuprakash@iitm.ac.in
Abstract—The crash behavior of Cold Rolled Mild Steel
(CRMS) closed form thin section produced by stitch welding
at periodic intervals of length was studied by conducting axial
compressive tests at loading velocities of 5 mm/min and 6000
mm/min. The deformation shape, peak forces and energy
absorption capacity of the sections estimated numerically
showed a good correlation with the experimental data.
Index Terms — axial crushing, energy absorption, peak force,
welded joints, cut slot.
I. INTRODUCTION
Welding is the primary method of joining employed in
the car industryfor structural construction work. Quite often,
due to limitations in manufacturing of closed form sections,
thin sheets/plates are bent to form a closed form section and
are seam welded (continuous welding) or stitch welded
(welding at periodic intervals of section). When such welded
structural members are subjected to high velocity impact,
such as in the case of a crash, the deformation behavior is
very different from conventional closed form sections. The
purpose of this study is to evaluate the crash behavior of
stitch welded closed from sections. The crash performance
at twodifferent velocities ofimpact are evaluated numerically
and validated through experiments. Crash behavior differs
from conventional quasi-static compression behavior of
material in view of dynamic buckling. The speed at which
the plastic waves propagate in a material becomes one of
the major parameters in determining the type of buckling.
Progressive buckling, typical of crash, can occur only when
the applied axial compressive load is higher than the yield
stress of the material. During crash, the main energy
dissipating mechanisms are plastic deformation and fracture
or tearing. A vehicle’s crash performance depends on the
weld structural integrity, peak force, mean effective crushing
force and energy absorbing capacity of the structure. As
passenger safety is closely related to peak force, keeping
the minimum peak force and maximum energy absorption
capacity of the compartment frame are desirable. The
displacement of the space frame should be such that there
will not be anyintrusion into passenger compartment. Strain
rate, also expressed as deformation rate plays an important
role in the determination of load carrying capacityand energy
absorption capacity of structural members.
II. EXPERIMENTAL EVALUATION OF MATERIAL
PROPERTIES
Numerical simulation of crash behavior requires input
on material stress-strain properties at high strain rates. This
was obtained by conducting tensile tests on base material at
different rates of loading.
A. Tensile Test Specimen Specification
Four specimens made of cold rolled mild steel of thin
section were prepared as per ASTM E-08 standards (Fig. 1).
Tensile tests were carried out on a 100 kN- MTS servo
hydraulic material test system at the loading velocities of 5,
100, 500, and 1000 mm/min. at room temperature.
Figure 1 Monotonic test specimen drawing (ASTM E 8/E 8M – 08)
From the tensile test, tensile load and the corresponding
displacement data were extracted to plot the engineering
stress- strain curves as shown in Fig. 2. The required material
properties such as, tangent modulus, young’s modulus and
yield stress were obtained from the stress strain curve and
fed into the numerical simulation.
B. Strain rate hardening
The uniaxial stress-strain behavior of the material was
fitted to the Cowper-Symonds equation (1) which gives the
strain rate hardening effect, Cowper Symonds coefficients
and Cowper Symonds exponent.
Cowper-Symonds equation:
q
d
D
/1
0
1













 
0
; d
(1)
Where
σd
o
Dynamic flow stress at uniaxial plastic strain rate.
εì Strain rate
σ Associative static flow stress
D Cowper-Symonds coefficient
q Cowper-Symonds exponent

53
© 2010 AMAE
Figure 2. Engineering stress - strain curve at various strain rate.
The parameter “q” is the slope of straight line and the
intercept on ordinate is “log D”. The Cowper – Symonds
parameters were determined from the graph shown in Fig. 3
and the values are tabulated in Table 1
Figure 3 Cowper - Symonds parameters
TABLE I
COWPER – SYMONDS PARAMETERS
III. EXPERIMENTAL CRASH TESTS
A. Specimen preparation
Commercially available thin walled rectangular cross
section of CR mild steel having thickness of 0.8 mm was cut
to the length of 140 mm. A slot of 2 mm width was cut
longitudinally on one side of the specimen for the entire
length to simulate the open section, which is typically
obtained when CR sheets are bent multiple times to obtain a
closed form section. The cut slot has been joined for every
20 mm using TIG weld and 20 mm has been left open. Figure
4 and 5 shows the details of thin walled sections.
Figure 4. Detail of the drawing specification
Figure 5. Specification of welded joint with cut slot
B. Compression test
The rectangular thin walled specimen was kept in
between the compression platens of the test system. During
the test, the top platen was kept rigid and the bottom platen
which has mounted onto the actuator was moved upwards
with the aforementioned velocities to a pre-determined
distance of 50 mm of compression. The upward movement
of the actuator makes the specimen to buckle within the
compression length of 50 mm. The buckled samples are
shown in Fig. 6 and Fig. 7.
Figure 6. Deformed shape of the specimen at 5 mm/min
Figure 7. Deformed shape of the specimen at 6000 mm/min
IV. NUMERICAL SIMULATION OF CRASH TEST
The buckling characteristics of thin walled rectangular
welded section have been determined numericallyusing LS
DYNA Explicit Solver. The geometry of the specimen was
modeled using shell 163 element with 5 integration points.
The contact Belystschko shell element and rigid block were
defined using “Automatic Single Surface Contact” (ASSC),
and a coefficient of friction of 0.25 was assumed to avoid
lateral movements. The normal force at failure of 157.8 MPa
and shear force at failure of 37.6 MPa were assumed as an
input data. The meshed model with welding nodes is shown
in Fig. 8. Cowper Symonds equation parameters for CR Mild
Steel thin walled specimens shown in Table 1 were fed as
input parameters for the numerical analysis.
Figure 8. Meshed weld model with boundary conditions.

© 2010 AMAE
54
The numerical simulations corresponding to the loading
velocity of 5 mm/min and 6000 mm/min were obtained for
the time period of 600 s and 0.5 s respectively. The
dimensions of the shell element are length = 140 mm, width
= 50 mm, height = 25 mm, thickness = 0.8 mm, Young’s
modulus = 192 GPa and density = 7800 kg/m3
. The
compressive deformed shapes of the model are shown in
Fig. 9 and Fig. 10.
Figure 9. Deformed shape of the model at 5 mm/min
Figure 10. Deformed shape of the model at 6000 mm/min
V. COMPARISON OF AXIAL COMPRESSION TEST
RESULTS
From the compression test readings of each samples and
numerical results, peak force and maximum energyabsorbed
by the samples were calculated. The corresponding load vs.
displacement graphs are plotted as shown in Fig. 11 and Fig.
12 and their energy absorption capacity were plotted in Fig.
13 and Fig. 14.
Figure 11. Force vs. Displacement plots at the velocity of 5 mm/min
Figure 12. Force vs. Displacement plots at the velocity of 6000 mm/min
Figure 13. Energy vs. Displacement plot at the velocity of 5 mm/min
Figure 14. Energy vs. Displacement plots at the velocity of 6000 mm/
min
The entire load – displacement curves show higher
compressive repulsive force when the first fold is formed
and subsequent folding is formed under smaller repulsive
force. The symmetric mode behavior of the thin walled
section is associated with the formation of lobes.
TABLE II
DEFORMED COMPRESSION TEST RESULTS AT 5 MM/MIN
TABLE III
DEFORMED COMPRESSION TEST RESULTS AT 6000 MM/MIN
DISCUSSION
The crash characteristics of the thin walled rectangular
cross section has been determined experimentally and
validated numerically. Under the static and dynamic loading,
two side walls were protruded in one folding and other two
walls recessed on the next folding. It clearly shows that the
lobes deform in the same manner having symmetric mode
as shown in Fig. 9 and Fig. 10. During buckling, no crack
and no fracture was observed. It was observed that due to
increase in strain rate, the energy absorption by the welded
section was calculated as 517.13 J and 624.6 J at the velocity
rate of 5 mm/min and 6000 mm/min respectively (Table 2
and 3). Figures 11 and 12 show the peak loads and shape of
the curves gives good correlation with the experimental
results. Figures 13 and 14 show the comparison between the
simulation and experimental results of energy vs.
displacement curve.

© 2010 AMAE
55
CONCLUSIONS
The experimental crash behavior of CRmild steel welded
joint shows good correlation with numerical results in terms
of energy absorption capacity and peak force. Thus the
numerical method can effectively be used to estimate the
crash characteristics prior to conducting the actual
compressive tests. The strain rate affects the hardening
behavior of material which in turn affects the peak force as
well as energy absorbing characteristics. Based on the
experimental results, it is established that size of the lobe
and distance between the lobes varies if the strain rate
increases. Further studies are required to study the energy
absorption capacity in the actual crash events at the strain
rate of 103
s-1
to 105
s-1
.
REFERENCES
[1] F.Schneider and N. Jones, “Influence of spot-weld failure on
crushing of thin-walled structural section”, Elsevier, Int. J.
of Mechanical Sciences,Vol. 45, pp.2061-2081, Dec. 2003.
[2] W. Abramowicz and N. Jones, “Dynamic axial crushing of
square tubes”, Int. J. lmpact Engng. Vol.2, No.2, pp.179-208,
1984.
[3] B. Chinn, “Crashworthiness and its influence on vehicle
design”, in An introduction of modern vehicle design, chp. 6,
pp. 157-184, 2004.
[4] D. Karagiozovaa and N. Jones, “Dynamic buckling of elastic–
plastic square tubes under axial impact—II: structural
response”, Int. J. of Impact Engineering. Vol.30, pp. 167 –
192, April 2004.
[5] K. Mizuno, Y.Arai and C.A. Newland, “Compartment strength
and its evaluation in car crashes”, Int. J. Crash, Vol. 9, No.5,
pp. 547-557, 2004.
[6] J. S. Qiao, J. H. Chen and H. Y. Che, “Crashworthiness
assessment of square aluminum extrusions considering the
damage evolution, Thin walled structures”, vol. 44, pp 692-
700, 2006.

Experimental and Numerical Assessment of Crash Behavior of Welded Thin Wall Rectangular Steel Sections

More Related Content

What's hot (17)

Viewers also liked (8)

Similar to Experimental and Numerical Assessment of Crash Behavior of Welded Thin Wall Rectangular Steel Sections (20)

More from IDES Editor (20)

Recently uploaded (20)

Experimental and Numerical Assessment of Crash Behavior of Welded Thin Wall Rectangular Steel Sections