You are on page 1of 8

F2008-SC-030

NUMERICAL SIMULATION OF CRASH TEST FOR THE VEHICLE


STUDENT ROADSTER
Bial Ana*, Klemenc Jernej, Rosa Uro, Fajdiga Matija
University of Ljubljana, Faculty of mechanical engineering, Slovenia
KEYWORDS crash test, safety, Student Roadster, numerical simulation, Euro NCAP
ABSTRACT - Crash tests and numerical simulations are a vital source of information when
designing safety elements for vehicles. To obtain homologation the vehicle's crash safety
needs to be demonstrated. Based on three tests Euro NCAP program presents objective and
comparable safety performances. Student Roadster is a prototype and will remain the only one
built, therefore its adequacy has to be proven using a numerical simulation. The first stage of
the simulation was a crash of the front bumper against an obstacle. The bumper is fixed on
two dented beams that function as energy absorbers in the case of a crash. A 3D model has
been built and discretized with finite element method in Catia while the simulation of the
crash was done in Ls-Dyna. The results illustrate the importance of indentations in energy
absorbent beams for plastic deformations form in those areas. Upgrading the front bumper
model, crash test simulation for the entire vehicle SR is currently in the process.
INTRODUCTION
Student Roadster is a prototype vehicle designed and manufactured at our Faculty of
mechanical engineering. To get permission for driving the vehicle on the streets it needs to
obtain homologation. Standard homologation procedures require demonstration of vehicle
safety in case of a crash known as crash test. Since Student Roadster is the only one of its
kind, the cars safety performances can not be experimentally tested, therefore they need to be
demonstrated in a different manner. Using a 3D model of the vehicle and numerical
simulation we tend to prove that Student Roadster performs adequate in case of a crash.
Euro NCAP (New car assessment programme)
Established in 1997 the Euro NCAP programme provides reliable and comparative
assessment of the safety performance of vehicles. The program encourages the automotive
industry to exceed the minimum requirements for safety provided by legislation. The vehicles
are assessed on three tests; frontal impact, side impact and pedestrian impact. The three testes
are mandatory and can bring up to four stars. The fifth star is gained by the fourth test known
as pole impact. The ratings are presented in stars, five stars being the maximum score.
We tend to demonstrate Student Roadsters behaviour in frontal impact, the impact which
represents the most frequent type of road crash, resulting in serious injury. In frontal impact,
the car strikes the offset barrier at 64 km/h (1) as shown in Figure 1.

Figure 1: Position of the vehicle and the barrier in case of frontal impact (5)

3D MODEL AND NUMERICAL SIMULATION


Building a 3D model
The foremost important element in frontal impact is the front bumper. The bumper is fixed to
the chassis with two dented beams. Indentations in the beams behind the bumper are designed
especially to absorb the energy released at the moment of collision. The more energy is used
to deform the beams, the less of it remains to deform and damage other elements.
The bumper and its beams were built as a 3D model and meshed with finite elements (Figure
2).

Figure 2: 3D model of the front bumper and two supporting beams

In attempt to save computational time, the mesh is not consistent throughout the model.
Density of the mesh varies from areas that are affected more, where the mesh is denser to
areas that are affected less, where the mesh is less dense. Since the front bumper and the two
supporting beams used to fix it to the chassis are the first and therefore the most affected
elements in case of frontal collision, their mesh should be the densest (Figure 3).

Figure 3: Discretization of the model using finite elements

Mesh attributes for the front bumper are given in Table 1.


Number of elements

30687

Number of nodes

31682

Number of quadratic elements

30.363 (98.94%)

Number of triangular elements

324 (1.06%)

Table 1: Mesh attributes of entire model

In contrast to denser mesh in front of the vehicle, the rest of the chassis is meshed with larger
finite elements.
It should be noted that the complexity of the problem requires the model to be gradually
upgraded, therefore the bumper and two beams were analysed first and separately from the
rest of the chassis.
Simulation
As front bumper mesh was imported into Ansys the elements were prescribed element type
SHELL 163 (Explicit thin structural shell), a 4-node element with twelve degrees of freedom
at each node. The bumper mesh was connected to beam meshes using elements type BEAM
161. BEAM 161 is a simple beam element which requires three nodes where the third node
determinates initial orientation of the beam cross section. Prescribed material model for the
bumper and the beams was Bilinear kinematic. To simulate the weight of the vehicle which
was eliminated from this simulation, a square block was designed using SOLID 164 (Explicit
3D structural solid) element type, 8-node element with nine degrees of freedom. Material
model that was used was Linear isotropic. With the bumper model completed, the barrier was
simulated using a single SHELL 163 element with Rigid material model. All displacements
and all rotations of the barrier were constrained. Contact in the model was prescribed as
Automatic single surface, making all nodes contact nodes.
Keeping in mind that frontal impact test is performed with car speed of 64 km/h and taking
into consideration that the model of the bumper and its beams is significantly smaller in size

than that of the entire vehicle the initial velocity was reduced to 15 km/h and was prescribed
to all nodes in the model. The final setting of the bumper and two beams, the block that
simulates the vehicle and the SHELL 163 element simulating the barrier is shown in Figure 4.

Figure 4: Final setting of all elements

After the front bumper mesh was checked with Ls- Dyna software, the mesh was united with
those, previously made for the chassis (4). The meshes were then imported into Ansys
working environment to be completed before performing the crash simulation of entire
vehicle. Linear welds present in Student Roadsters chassis were simulated using BEAM 161
elements positioned in zigzag manner to prevent deformations of SHELL 163 elements which
were prescribed to entire model of the chassis. In attempt to simplify the problem and save
computational time the front and rear axle were also simulated using BEAM 161 elements but
of circular cross section. Twenty seven SOLID 164 elements positioned in the front part of
the chassis present the engine of the vehicle. The block elements are fixed to the chassis at
positions similar to the ones in Student Roadster using BEAM 161 elements. The offset
barrier was designed using a single SOLID 164. All degrees of freedom were taken away
from the barrier element to prevent all translations and all rotations. As required by the Euro
NCAP standard, the barrier was positioned with 40% overlap. In the model Automatic single
surface contact was applied. Initial velocity was assigned, to all nodes of the vehicle with the
value of 64 km/h, as prescribed by standard. The final model used in simulation consists of
97404 finite elements and 97090 nodes and is shown in Figure 5.

Figure 5: Final model of the vehicle and the barrier

The results acquired so far are only preliminary and act as a solid foundation for final
competition of the model.
RESULTS
To study the results of vehicles crash test simulation, results of the front bumper must be
examined first.
Solution time for the bumper model was set to 0.05sec, a value which ensures that the
structure has started to move away from the barrier. Results for von Mises stresses are shown
at instance of 0.0002sec (Figure 6), 0.01sec (Figure 7), 0.02sec (Figure 8), 0.03sec (Figure 9),
0.04sec (Figure 10) and 0.05sec (Figure 11). At t = 0.0002sec the first contact between the
bumper and the barrier occurs. The areas where von Mises stresses start to form are clearly
visible at this instance but the values taken by the stresses are not high. Stress areas are
located mostly around the indentations in the supporting beams. As the crash continues the
stresses reach their peak located around the first indentation and around the joint which fixes
the bumper to the two beams behind it. Higher values of stresses are observed also on the
front of the bumper. Stress values increase until 0.03sec, after short stabilization the values
begin to decrease. The drop of stress values is due to withdrawal of the bumper from the solid
barrier. Figures 6 to 11 show variation of von Mises stress at selected time instances.

Figure 6: t= 0.0002 sec, von Mises stress

Figure 7: t= 0.01 sec, von Mises stress

Figure 8: t= 0.02 sec, von Mises stress

Figure 9: t= 0.03 sec, von Mises stress

Figure10: t= 0.04 sec, von Mises stress

Figure 11: t= 0.05 sec, von Mises stress

Upon closely examining the results of the bumper, the results of entire vehicle can be
presented. It should be noted that due to long computational time the crash test simulation
hasnt been completed yet, so the results given below in Figure 12, 13 and 14 are only partial.
Simulation time was, once again, set to 0.05sec, however the results given are until 0.003sec
of the running time. By this time the impact is not complete yet and the vehicle is still moving
towards the barrier. Results for von Mises stresses are shown for time instances of 0.001sec
(Figure 12), 0.002sec (Figure 13) and 0.003sec (Figure 14). High stress areas are similar to
areas pointed out before in the bumper model. The highest values are located on the front of
the bumper as a direct result of the contact between the barrier and the vehicle. At t =
0.002sec von Mises stresses begin to distribute over the structure. Increased stresses occur
over the chassis but no significant displacements are present in parts other then the bumper
and supporting beams. Discussing the results at time 0.003sec we observe that stresses

continue to spread along the chassis, but still no significant displacements can be observed in
elements behind the two supporting beams. While the stresses are dispersing in the chassis the
areas with highest values remain on the front section of the bumper and around the
indentations in the two beams.

Figure 12: t= 0.001 sec, von Mises stress

Figure 13: t= 0.002 sec, von Mises stress

Figure 14: t= 0.003 sec, von Mises stress

DISCUSION
If we focus first on the results of the front bumper, we can observe that the areas of the
highest stress values are located in the front part of the bumper and around the indentations in
the supporting beams. High stress areas, which are located around the indentations in the
beams prove, that a large amount of energy dissipates in those areas, therefore the
indentations are performing suitably. Plastic deformations exceed the value of 20% very early
in the crash, which indicates structural break-down in some locations. Short stabilisation of
stress values in bumper simulation around 0.03sec coincides with the instance when the
bumper is brought to a standstill. The standstill is then followed by a rebound of the structure.
The following rebound is inevitable since it is the result of stored elastic energy and poorly
positioned indentations, however, the smaller the rebound, the better the design.
The lower part of the beams is taking a lot less released energy than the upper one, however it
seem that their function is also to prevent the upper part to tip and bend downwards, causing
less energy to be used for deformation of the beams, allowing more energy to continue to
affect other elements.
The results given from the crash simulation of the vehicle are only partial and are therefore
useless for final comments. However they indicate the stress distribution and dispersion
pattern through the chassis.
CONCLUSION
From the results of both simulations we can conclude that the stress distributions are
reasonable. The solution time of both simulations should be longer than the one currently
assigned, to ensure the crash to be completed. The beginning of the crash indicates that
indentations in the front beams are performing appropriately since large amount of energy is
used to deform them. It needs to be noted that stress values are not of vital importance to this
analysis instead final displacements of elements are. However the final displacements can
only be given once the crash is completed, so the crash test simulation for entire vehicle
continues. The model itself is still in the process of gradual upgrading until its final
completion with crash test dummies onboard.
REFERENCES
(1)
(2)
(3)
(4)
(5)

European New Car Assessment Program, Frontal impact testing protocol,


Version 4.1, 2004
LS-DYNA Keyword user's manual, Livermore Softmore Technology
Corporation, Version 970, 2003
John O. Halquist, LS-DYNA theoretical manual, Livermore Softmore
Technology Corporation, 1998
Henrik Zaletelj, The analysis of the body structure of the car Student Roadster,
Diploma thesis, University of Ljubljana, Faculty of mechanical engineering, 2006
www.euroncap.com

You might also like