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

net/publication/236677061

Damage modeling in crashworthiness of dual-phase steel

Conference Paper · May 2013

CITATIONS READS
3 994

3 authors, including:

Javad Marzbanrad Solat Norozi

Iran University of Science and Technology Iran University of Science and Technology
147 PUBLICATIONS   1,230 CITATIONS    4 PUBLICATIONS   3 CITATIONS   

SEE PROFILE SEE PROFILE

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

Safety in motorcycle View project

Topology optimization View project

All content following this page was uploaded by Solat Norozi on 15 May 2014.

The user has requested enhancement of the downloaded file.

 st
21 International Conference on Mechanical Engineering-ISME2013
7-9 May, 2013, School of Mechanical Eng., K.N.Toosi University, Tehran, Iran

ISME2013-1970

Damage modeling in crashworthiness of dual-phase steel

Javad Marzbanrad1 , Solat Noruzi2, Ali Ahmadi, S.K3
1
Iran,Tehran, Iran University of Science and Technology/School of Automotive Engineering; marzban@iust.ac.ir
2
Iran,Tehran, Iran University of Science and Technology /School of Automotive Engineering; solat.noruzi@gmail.com
3
Iran,Tehran, Iran University of Science and Technology /School of Mechanical Engineering; ali_ahmadi_sk@iust.ac.ir

Abstract microscopic variables such as voids and cracks are

Dual-Phase steel is an advanced high strength steel that examined. The second approach is the macroscopic
are used recently in auto-body for energy absorption. approach in which material is represented by its global
Fracture is an noticeable factor in crashworthiness of response. In both groups, various numerical models are
vehicle impact loading. Failure prediction of DP800 was suggested.
investigated in this studying. There are different fracture
criteria for damage modeling, but the material behavior Micro-Mechanical Gurson Model
cannot be simulated by using them, so a new approach In fact, the most known and widely used
have been used in this studying. The Gurson model in micromechanical model is Gurson-Tvergaard-
conjunction with the Johnson-Cook damage model is Needleman (GTN), which is an extended version of the
applied to simulate correctly the material behavior. By model proposed by Gurson [1]. Gurson derived a
using a new method in this research, in shear fracture pressure dependent yield function from an isolated
and ductile fracture the active mechanisms are Johnson- spherical void in a continuum. The global response of
Cook damage model and Gurson model respectively. the solid is correlated to void volume fraction. The yield
Finally, the crashworthiness of a B-Pillar in different function of Gurson model is defined as
conditions is simulated and the fracture is predicted.  q   3 p 
φ =   + 2q1 f * cosh − q2
σ   2σ 
( )2
− 1 + q3 f * = 0 (1)
Keywords: crashworthiness, dual-phase steel, failure  y  y 
prediction, advanced high strength steel where
S = pI + σ is the deviator part of the Cauchy stress
Introduction
tensor;
Crashworthiness studying of auto-body structures have
been done for many years. Recently, automotive 3
industries demand for improvement of quality of crash
q= S : S is the effective Mises stress;
2
simulations. Crashworthiness studying without taking
into account damage behavior can be leaded 1
p = − σ : I is the hydrostatic pressure;
overestimations of load-carrying capacity and absorbed 3
energy of structural component. With increasing use of
advanced high strength steels in car body structures due ( ) is the yield stress of the fully dense matrix
σ ε m pl
pl
to interest of weight reduction, car manufactures need to
material as a function of ε m , the equivalent plastic
better understanding of the behavior of these materials
strain in the matrix; and q1, q2, q3 are material
under crash loading conditions.
parameters.
In present study, dual-phase high strength steel
This model represents the commonly observed stages of
DP800 was investigated. This material shows a higher
ductile fracture, which include:
strength but significantly lower ductility than
• The formation of voids around inclusions and
conventional deep-draw steels. Due to this fact, precise
second phase particles.
failure prediction becomes more important in crash
simulation of this steel. • The growth of voids due to plastic straining and
The reliability of finite element analyses depends on hydrostatic stress.
the accuracy of constitutive and fracture models. • The coalescence of growing voids leading to
Although there are different damage models in crash fracture.
codes, most crash simulation do not take into account However, this kind of model is limited to
damage behavior. In this study two damage models are representing ductile fracture and as such ignores the
introduced and the way of using these models are fracture mechanism based on shear [2, 3].
presented. In FEM analysis, failure prediction using the Gurson
model is highly dependent on mesh refinement as this
Damage Modeling Approaches model causes strain softening prior to fracture.
Mainly two different approaches were used for damage Considering these shortcomings prevalent in today’s
modeling in engineering materials. The first one is the numerical codes, there emerges a strong need for the
microscopic approach, in which the evolution of development of a comprehensive approach for failure

1
prediction coupled with a numerically robust
implementation.

Macro-Mechanical- Johnson-cook model

Modern crash simulation models typically use element
edge lengths of 5–15 mm. The previous micro-
mechanical failure is unsuitable for such meshes and
alternative methods must be sought.
The deformation is a complex phenomenon at micro
scale. Many ductile materials at macroscale can be
assumed as a continuum. In conventional continuum
mechanics models with damage formulation,
macroscopic response is a combination of material
matrix, defects like inclusions, voids and micro cracks Figure 1: A new asymmetric 3D fracture locus.
and globally it is assumed to obey continuum mechanics
laws. The damage indicator is calculated separately and They realized that a definition of damage model
does not influence the material plasticity. Because of the depending on triaxiality alone in not adequate. The third
relative simplicity, in industrial applications the models invariant of stress, which is a measure of the Lode
are widely used. The most widely used model is angle, should be also taken into account.
proposed by Johnson and Cook [4], which incorporates Plane stress represents an important condition in
phenomenologically the strain rate and temperature sheet metal forming and crash analysis. The range of
factors in the stress-strain function, stress triaxiality that can be achieved in thin sheets
  .  usually falls between the uniaxial and equi-biaxial
[ 1 + C ln  ε p
] 
[ ] tension ( 1/3 ≤ η ≤ 2/3). This range is indicated in
 1 − (∆T )
N q
σ eq = A + Bε f  . (2)
 ε 
"Figure 1" by a shaded area, which is then enlarged in
  0  "Figure 2".
where A, B, C and N and q are five model
parameters, T, T0, Tm are temperature, room
temperature and melting temperature, respectively and ˙
ε0 is the reference strain rate. Under proportional
loadings the fracture strain εf is defined as function of
stress triaxiality η, plastic strain rate and temperature:
  . 
ε p 
(
ε f (η ) = D1 + D2e )
− D3η 

1 + D4 ln .  [1 + D5∆T ]
 ε 0 
   (3)
where D1, D2, D3, D4, D5 are five material constants
triaxiality η, defined as the ratio of mean stress to
equivalent stress: Figure 2: A 3D fracture locus for thin sheets.
σm The projections of the 3-D fracture locus on the
η= (4)
q ( ε f , ζ ) and ( ε f ,η ) planes are shown respectively in
and in the above equations we have: "Figure 3" and Figure 4". The present result points out
T − T0 difficulties in describing fracture of sheets by means of
∆T = (5)
the original Gurson model [7]. Such a model predicts a
Tm − T0
monotonic decrease of material ductility with the
This results in an exponential dependency of failure
triaxiality parameter η, while in actuality there may be a
strain on triaxiality, prescribing a monotonically falling
function of η. local minimum in the function ε f (η ) .

Stress State Dependence of Damage Modeling

In recent years, several investigations about Lode-
dependent of failure surface have been published. Based
on test results and theoretical investigations, Wierzbicki
[5] and Xue [6] have considered possible surface
shapes as shown in "Figure 1".

2 ISME2013, 7-9 May, 2013

 Figure 4: Geometry of the uniaxial tensile test specimen [10]

Gurson model have seven parameters that are not

Figure 3: A conceptual graph showing three possible cases of independent from each other. Three parameters
the dependence of material ductility (equivalent strain to
fracture) on Lode angle parameter. ( q1 , Sn , ε n ) was the same of [11] and the other four
parameters are determined by FEM simulations of
tensile test. In "Table 1" all the parameters are reported.

Table 1: Gurson model parameters for DP800

parameter q1 Sn εn ff fc fn q2
value 1.5 0.11 0.35 0.15 0.08 0.01 1

Tensile test simulation results with Gurson model

are compared with experimental test results in "Figure
5" and "Figure 6". In Abaqus model shell element with
sizes of 1.5 mm and 0.5 mm are chosen.

900

800

700
Engineering Stress (Mpa)

600

500 - Experiment800 DP
Figure 4: The projection of a 3-D plane stress fracture locus 400
into the plane of the equivalent strain to fracture and the stress Gurson-Element size=1.5mm
300
triaxiality.
200 Gurson-Element size=0.5mm
One opportunity to regard this behavior is GISSMO
100
failure criteria presented by Neukamm et al. [8] and
0
Nahshon and Hutchinson [9] proposed an extension for 0 0.05 0.1 0.15 0.2
micromechanical Gurson model, which accumulates Engineering Strain

damage in state of shear stress in dependence of Lode figure 5: Engineering stress-strain curve
parameter.
In this study, the combination of two failure
criterion, Gurson and Johnson-cook damage model is
used in plane stress condition.

Material Characterization
The Docol DP800, a dual phase steel supplied by
Swedish Steel Ltd. (SSAB) with chemical composition
(in wt % - C: 0.12, Si: 0.20, Mn: 1.50, P: 0.015, S:
0.002, Nb: 0.015, Fe: bal.) with the sheet thickness of
1.5 mm was studied. V. Tarigopula and et al. have been
reported two different types of test results [10].
figure 6: Tensile test simulation- Gurson Model
Numerical Simulation
For prediction of damage modeling in dual-phase steel
Johnson-Cook model is an accumulated failure model in
and determining of Gurson model parameters tensile
which that failure strain is more depended to stress
test is simulated by FEM explicit software. In "Figure
triaxiality than the others parameters so it is more
4" the geometry of specimen are presented.
important to determine the three parameters D 1, D2 and
D3. For determining D4 one can use the strain rate-

3 ISME2013, 7-9 May, 2013

failure strain curve. In "Figure 7" the slop of the curve is evolution law for that criterion; in the absence of a
D4. damage evolution law; however, the material stiffness is
not degraded [12]. As mentioned, no one of Gurson and
Johnson-Cook models can simulate the response of the
Engineering failure strain
24
material, independently. For covering the stress
Triaxiality and failure strain relation, a new failure
y = 0.001x + 21.49
22.5
model is needed. For answering the above need we used
the Gurson model in conjunction with the Johnson-
Failure Strain (%)

21 Cook model, such that in shear phenomenon only

Johnson-Cook model acts in the simulations and when
19.5 the stress triaxiality reaches to η=0.33 and more, only
Gurson model is an active mechanism. To see how these
18 two mechanisms happen during the simulation, we can
0.001 0.01 0.1 1 10 100
see in "Figure 9" that for example when the triaxiality is
Strain Rate (1/s)
about η=0.44, the software is calculating two active
figure 7: Determining of the D4 parameter of J-C model failure mechanisms simultaneously and independently.
In this condition failure strain for Gurson and Johnson-
For determining the D1, D2 and D3 we need at least 3 Cook models are ε f = 0.76 and ε f = 0.83 ,
experimental test results (see "Figure 8"). By using the
tensile test and shear test experiment results and by the respectively. The Gurson model initiation criterion is
help of FEM simulation with try and error method we satisfied so the material stiffness is degraded and the
calculated the parameters. Stress triaxiality is η= 0.33 element is deleted.
and η= 0.01 in tensile and shear test, respectively [10].

Figure 9: Example of interpolated curve

figure 8: failure strain-triaxiality relation in Johnson-Cook Shear test simulation by using of Johnson-Cook
failure model
model are compared with test results founded in the
literatures [10]. It can be seen from "Figure 10" that
In "Table 2" the all parameter of Johnson-cook failure
there is a good agreement between them.
model are given.

Table 2: Johnson-Cook failure model parameters for DP800

parameter D1 D2 D3 D4 D5
value 0.8 0.118 19.4 0.001 0

Also, A, B, n, m and c the plastic parameters of

Johnson-Cook model are determined from experimental
tensile test (see "Table 3").
Figure 10: Shear test simulation-Johnson-Cook odel
Table 3: Johnson-Cook plasticity model parameters for DP800
parameter A B n m c B-Pillar Crash Simulation
value 599 392.7 0.4 0 0.009 In this study, crashworthiness of B-Pillar of auto-body
is simulated by using of Gurson model in conjunction
New Damage Modeling Approach with Johnson-Coo model. Test setup is shown in
In Abaqus finite element code more than one damage "Figure 11". B-Pillar was modeled by shell element and
initiation criterion can be specified for a given material. by applying different velocity to the cylindrical
If multiple damage initiation criteria are specified for impactor.
the same material, they are treated independently. Once
a particular initiation criterion is satisfied, the material
stiffness is degraded according to the specified damage

4 ISME2013, 7-9 May, 2013

 Figure 11: Test on B-Pillar

"Figure 12" shows a fracture in B-Pillar middle

during crash simulation, so it is an important factor in
auto-body crashworthiness study to take into account Figure 14: Fracture on roof rail and B-Pillar
the damage modeling.
Conclusions
Failure prediction of dual-phase steel DP800 was
investigated in this studying. Damage modeling of
advanced high strength steel was reviewed and
concluded that, no one of the existence failure models
can simulate correctly the fracture of dual-phase steel,
because the stress triaxiality and the failure strain
relation, for plane stress conditions, in the material is in
a such way that, common damage models cannot
simulate the phenomenon. The Gurson model was used
Figure 12: Impact modeling of B-Pillar in conjunction with Johnson-Cook model for covering
all triaxiality and failure strain condition. Using the two
By changing position and velocity of the impactor, active failure mechanisms, independently and
different possible fracture in side-body of car is simultaneously, to simulate the fracture of dual-phase
investigated. In "Figure 13", three possible impact of steel can be a new way of the failure prediction. By
pole to B-Pillar of auto-body is simulated. To show the using this method, the crash response of B-Pillar of
impactor velocity effects, v=15 m/s, v=10 m/s, and v=5 auto-body was simulated. It has been shown that
m/s we applied to the rigid impactor. fracture is an important in crash studying and it may
happen during the crash loading and this means that
there will be a reduction in energy absorption of car
body.

References
[1] Gurson, A. L., “Continuum Theory of Ductile
Rupture by Void Nucleation and Growth: Part I—
Yield Criteria and Flow Rules for Porous Ductile
Materials,” Journal of Engineering Materials and
Technology, vol. 99, pp. 2–15, 1977
[2] El-Magd, E, Gese, H, Tham, R, Hooputra, H and
Werner, H. ‘Fracture Criteria for Automobile
Crashworthiness Simulation of wrought Aluminum
Figure 13: Different positions for impactor Alloy Components’, Mat.-wiss u Werkstofftech,
2001 32 712–724.
Fracture occurs, sometimes, in roof rail of auto-body [3] Schmitt, W, Sun, D Z, Blauel, J G and Christlein, J.
and in the middle of pillar as illustrated in "Figure 14". ‘Improved Description of the Material Behaviour of
This means that in crashworthiness studying whenever Aluminum Automobile Components by the Gurson
the damage occurs, the energy absorption capacity is Model’, Proceeding of the 31st International
lost in the auto-body, but if a numerical simulation Symposium on Automotive Technology and
studying does not include the damage modeling, Automation, Düsseldorf, 1998.
perhaps we will have the wrong results and over [4] G.R. Johnson and W.H. Cook. Fracture
estimations. characteristics of three metals subjected to various

5 ISME2013, 7-9 May, 2013

 strains, strain rates, temperatures and pressures. simulation. 10th International LS-DYNA Users
Eng. Fract. Mech., 21(1):31 – 48, 1985. Conference.
[5] Bao, Y., Wierzbicki, T., On fracture locus in the [9] Nahshon, K. Hutchinson, J.W. 2008. Modifications
equivalent strain and stress triaxiality space. of the Gurson model for shear failure. European
International Journal of Mechanical Sciences 46, journal of mechanics A/Solids, Vol. 27, p. 1-17.
81–98. [10] V. Tarigopula, O. S. Hopperstad, M. Langseth, A.
[6] T. Wierzbicki and L. Xue. On the effect of the third H. Clausen, F. Hild, O.-G. Lademo, and M.
invariant of the stress deviator on ductile fracture. Eriksson. 2008. " A study of large plastic
Technical report, Impact and Crashworthiness deformations in dual phase steel using digital image
Laboratory, Massachusetts Institute of Technology, correlation and FE analysis". Experimental
Cambridge, MA (USA), 2005. Mechanics, Vol. 48, p. 181-196.
[7] Feucht, M., Sun, D.Z., Erhart, T. and Frank, T. [11] Kim Lau Nielsen, 2008. "3D modelling of plug
(2006). Recent development and applications of the failure in resistance spot welded shear-lab
Gurson model. LS-DYNA Users Conference, Ulm specimens (DP600-steel)". International Journal of
2006. Fracture, Vol. 153, p. 125–139.
[8] Neukamm, F; feucht, M.; Haufe, A.; Roll, K. 2008. [12] Abaqaus Documentation, V 6.9, Analysis user's
On closing gap between forming and crsh manual, Damage and failure for ductile metals.

6 ISME2013, 7-9 May, 2013

View publication stats