MATFEM Publications and Speeches

Transcrição

MATFEM
Partnerschaft
Dr. Gese & Oberhofer
Maschinenbauingenieure
Publications and
Speeches
Veröffentlichungen und Vorträge
February 2016
MATFEM
Nederlingerstrasse 1
D-80638 München
Tel.: (089) 1 59 89 79-0
Fax: (089) 1 59 89 79-9
http://www.matfem.de
F. Brenner
H. Gese
H. Dell
G. Oberhofer
Use of Specific Models for Creep,
Plasticity and Fracture in Forming
and Crashworthiness Simulation via
UCREEP and VUMAT
G. Metzmacher
Deutsche SIMULIA-Konferenz,
12.–13. November 2015, Aachen
MATFEM Partnerschaft
Dr. Gese & Oberhofer,
München, Germany
For the layout of various cold and non-isothermal forming processes and for the predictive crashworthiness simulations of metals it is essential to describe all relevant aspects of material behavior. Depending on the alloy and the simulated loading various aspects like creep, plasticity, hardening and fracture as well
as strain-rate and temperature dependency must be accounted
for. Three examples are presented in which forming behavior and
failure prediction is significantly improved by coupling specific
user subroutines to Abaqus/Standard or Abaqus/Explicit.
For the simulation of a hot forming process of fuselage panels
with inherent stress relaxation a temperature dependent material model with orthotropic plasticity and orthotropic creep is
used. Due to the large temperature range in the process, numerous basic material tests are required to derive the numerical
material model. A UCREEP-routine that describes the creep behavior of the AlMg sheets is used as none of the implemented
creep models is able to represent the relaxation behavior which
is found in the test satisfactorily.
A multiple bending operation on a dental implant is simulated
using Abaqus/Explicit and the user subroutine VUMAT with the
material model MF GenYld + CrachFEM. The material model is
capable of describing the pronounced asymmetry of the yield
stress in tension and compression for the titanium alloy under
consideration on the one hand and of rating the load reversals
correctly when accumulating damage with a tensorial model on
the other hand. Both effects must be taken into account to allow
for an accurate prediction of the fracture risk.
Due to the process design light metal extrusion exhibit a distinct
orthotropy in the plastic regime and in their fracture limits with
typically a minimum ductility transverse to the extrusion direction. As currently no models for orthotropic fracture are available
in commercial FEA codes, the user material model MF GenYld +
CrachFEM is coupled to Abaqus/Explicit for the crash simulation
of a double chambered profile.
M. Buckley1
M. Reissner2
G. Oberhofer2
H. Gese2
10th European LS-DYNA
Conference, Würzburg,
Germany
(1) Jaguar Land Rover Limited,
Gaydon, United Kingdom
(2) MATFEM Partnerschaft
München, Germany
Simulation of the manufacturing
process of self-piercing rivets with
LS-DYNA with focus on failure
prediction for sheets and rivet
Some activities on the use of 2d-rotional symmetric elements
with r-adaptivity for the simulation of the manufacturing process
of self-piercing rivets have been already published in LS-DYNA
user meetings in the last years. The focus of the research work in
this paper was an improved prediction of material failure in this
simulation environment. The functionality of r-adaptivity in LSDYNA includes a geometric criterion for the splitting of the
upper sheet during the riveting process. The geometric criterion
is defined by a residual thickness of the upper sheet for which
elements are deleted to finally split this sheet. This approach
might give acceptable results for very ductile sheet qualities (e.g.
mild and high strength steels). However for aluminium sheets
and AHSS grades the upper sheet may be split earlier due to limited ductility. Also there is a risk for fracture in the lower sheet
which has to be estimated with an appropriate fracture model. In
case of riveting AHSS grades there is also a chance of a rivet
fracture.
In a first attempt the numerical parameters for r-adaptivity of
2d-rotational symmetric elements have been optimized to ensure
correct results (e.g. fulfill volume constancy of the sheets). In a
second step the user material model MF GenYld + CrachFEM
with failure criteria for ductile normal fracture (DNF) and ductile shear fracture (DSF) has been introduced to allow for a damage accumulation in the sheet materials and the rivet material.
Besides an improved prediction of the sheet failure of the upper
sheet, the failure criteria also allow to estimate the margin of
safety for a fracture of the rivet and the lower sheet. The accumulated damage in sheets and rivet can be used later on for simulations of virtual mechanical tests on the riveted structure
(e.g. to derive strength values for crashworthiness simulation).
G. Oberhofer
M. Oehm
crashMAT 2015, 7. Freiburger
Workshop zum Werkstoff- und
Strukturverhalten bei Crashvorgängen 21–22 April 2015,
Fraunhofer EMI, Freiburg
München, Germany
Modelling Failure of Short Fiber
Reinforced Thermoplastics with
Respect to Local Degree of Fiber
Orientation
This speech was a slides-only presentation, for which there is no
abstract available.
H. Gese
F. Brenner
G. Oberhofer
Automotive CAE Grand
Challenge 2015 31 March – 1
April, 2015 Hanau, Germany
München, Germany
Crashworthiness simulation of high
pressure die cast aluminum alloys
and aluminum extrusions with MF
GenYld + CrachFEM
abstract available.
M. Oehm
A. Heath
H. Dell
New features in material model MF
GenYld + CrachFEM for sheet metal
forming simulation
H. Gese
Challenge 2015 31 March – 1
April, 2015 Hanau, Germany
München, Germany
abstract available.
H. Gese
Technisch- wissenschaftliche
Seminarreihe
Hochleistungsstrukturen im
Leichtbau, Hochschule
München, 12. November 2014
München, Germany
Verbesserte Abbildung von
Leichtbauwerkstoffen mit einem
modularen Materialmodell
Zu desem Vortrag gibt es keine Zusammenfassung.
H. Gese
International CAE Conference
2014, Pacengo del Garda,
Verona, 27–28 October 2014
München, Germany
Failure Prediction of Quenched Boron
Steels in Crashworthiness Simulation
The criteria for passenger safety have been tightened step by
step in the last years. This has caused a wide use of quenched
boron steels for the passenger cab design, especially for A-pillar
and B-pillar. Meanwhile also components with tailored tempering are used to adapt the material properties according to local
needs. The introduction of those boron steels has brought some
new challenges for the numerical simulation. In the field of
crashworthiness simulation a predictive simulation of quenched
boron steel components without a comprehensive assessment of
material failure ? including failure initiated from spot welds – is
not possible.
The material model MF GenYld+CrachFEM – can be used together with all major explicit finite element codes for crashworthiness simulation via the user material interface – provides comprehensive failure criteria for quenched boron steels. This includes the prediction of localized necking, ductile normal fracture
and ductile shear fracture. It is possible to adapt the material
properties locally by a mapping of the hardness provided by the
quenching simulation. The material model also provides the possibility to directly model the material failure in the HAZ around
spotweld with solids or with a macro model in combination with
a shell mesh. The lecture will introduce the material model, discuss the necessary experiments for the parameter identification
and will provide simulation examples based on FEA code LSDYNA.
H. Gese
H. Dell
FLC Conference, 6–7 November
2014, IVP, ETH Zurich,
Switzerland
München, Germany
CrachFEM – A Comprehensive
Approach for the Numerical
Prediction of Instability And Fracture
in Sheet Metal Forming Operations
The classical forming limit diagram is the most common method
in industrial sheet metal forming simulation for predicting possible sheet failure. This criterion is based on the onset of localized
necking. Its validity is limited to linear strain paths. For AHSS,
UHSS and aluminium sheets fracture can also appear without
prior necking. This kind of sheet failure is not covered by the
concept of the classical forming limit diagram. Ductile normal
fracture and ductile shear fracture – both in-plane and out-ofplane – are two possi-ble fracture modes in sheet materials. Depending on the hardening behavior and ductility of the sheet,
shear band instability can precede shear fracture.
The module CrachFEM of material model MF GenYld + CrachFEM offers a comprehensive approach for the prediction of material failure in sheet metal forming operations. MF GenYld +
CrachFEM can be used for a discretization with shell and solid
elements. For a shell discretization it is not possible to resolve
the onset of necking directly as the width of the necked region is
typically smaller than the edge length of a shell ele-ment. Therefore CrachFEM uses the submodule Crach to model the onset of
necking with a detailed discretization of the neck. The module
Crach uses a plastic material model which accounts for orthotropic plasticity, isotropic-kinematic hardening and strain rate sensitivity. The discretization is based on an initial imperfection to
allow for a realistic prediction of necking. The strain hardening,
strain rate sensitivity and the hardening due to change of stress
state inside the necked region (change to plane stress condition
in neck; increase of stress component normal to sheet, kinematic
hardening) contribute to the limit strain for unstable necking. As
the main physical effects are accounted for, Crach can also be
used in the case of nonlinear strain paths. Necking is detected by
the non-convergence of the mechanical model. The strain distribution across the neck at the onset of necking can be analyzed
with this numerical model as well. In analogy to the cross-section
method, the internal limit strain of algorithm Crach (maximum
strain inside the neck at onset of instability) is used as a criterion for sheet metal forming simulations. Besides tensile tests no
extra tests are needed to predict the forming limit diagram.
For the prediction of fracture (without prior necking) CrachFEM
uses phenomenological models for the failure modes ductile normal fracture (DNF) and ductile shear fracture (DSF). DNF accounts for frac-ture which is caused by void growth and void coalescence (in the case of ductile sheets, voids only appear immediately before fracture) with a fracture surface normal to the direction of the first principal strain. DSF accounts for material
failure due to shear band localization in the material. This phenomenon can appear for sheets in-plane and out-of-plane. The
equivalent plastic strain at fracture as a function of the relevant
stress state parameter is used as a failure criterion in the case of
linear strain paths. For nonlinear strain paths the initial fracture limit curves are used as master curves for integral damage
accumulation laws. In CrachFEM a tensorial description of damage is used. The limit strains for DNF and DSF cannot be predicted from elastoplastic properties as the ductility of metals depends on the microstructure of the material. A set of specimens is
needed, which allows the identification of equivalent plastic
strains at fracture for defined stress states. In the ideal case the
stress state should be constant for the whole deformation of each
kind of specimen.
A model for the post-instability strain (PIS) completes the current CrachFEM approach in the case of shell discretization. The
PIS model accounts for the additional strain in a shell element
from the onset of necking until final fracture – by DNF or DSF –
inside the neck. The PIS model provides a better assessment of
sheet failure – a tensile instability of a single shell element may
not develop into a wider neck in a zone with high strain gradients.
One focus of new developments in CrachFEM is an improved approach for ductile shear fracture (DSF). For most of the deep drawing sheet materials, shear band localization appears prior to
shear fracture. This phe-nomenon has been experimentally monitored for a number of sheet qualities in cooperation with the
Insti-tute for Manufacturing Technology at the University
Erlangen-Nürnberg. For the prediction of shear band localization
in sheet metal forming CrachFEM has been extended by a new
algorithm shearIS (shear instability). Potential shear instability
is checked for the in-plane and for the out-of-plane orientation.
With the new approach it is possible to differentiate between the
onset of shear band instability and final fracture in general. As
the strain gradient has an influence on the growth of the shear
band there are still open issues to be solved.
H. Dell1
V. Yelisseyev2
Failure prediction for non-reinforced
and short fiber reinforced polymers
G. Oberhofer1
13th LS-Dyna Forum 2014,
Bamberg, 6–8 October 2014
München, Germany
(2) MATTEST, Voronezh,
Russian Federation
The specific behavior of non-reinforced and short fiber reinforced
polymers in case of crash or drop test load cases imposes special
demands on the material model as well as the experimental determination of material parameters. If the visco-plastic material
characterization accounts for plastic compressible behavior the
characterization of material failure must also account for plastic
compressible behavior. The relevant material parameters must
be adjustment in an appropriate way. In case of short fiber reinforced polymers a significant anisotropy of failure can be observed and this anisotropy has to be considered in addition to the
anisotropy of the elasto-visco-plastic behavior. The subsequently
described simulation method is based on a phenomenological approach which shows high performance with respect to calculation
time and stability. It is therefore well suited for industrial applications.
M. Reissner
H. Dell
H. Gese
CrachFEM – A comprehensive
approach for the prediction of failure
in metallic materials
G. Oberhofer
11th World Congress on
Computational Mechanics
(WCCM XI) and 5th European
Conference on Computational
Mechanics (ECCM V), July
20–25, 2014, Barcelona, Spain
München, Germany
With new and increasing applications of advanced metallic materials, for instance Advanced High Strength Steel (AHSS) and
light-metal alloys, there are new and different challenges for numerical procedures. Well known is the classical forming limit
diagram, which is mainly used in industrial sheet metal forming
simulation to predict a possible sheet failure. This criterion is
based on the onset of localized necking. Its validity is limited to
linear strain paths. Advanced models should be used to achieve
also valid predictions for nonlinear cases. For AHSS, aluminium
and magnesium components fracture can also appear without
prior necking. Depending on the stress state and on the forming
history ductile normal fracture or ductile shear fracture can occur. Also the damage accumulation for fracture should be valid in
case of a nonlinear strain path.
The module CrachFEM of the material model MF GenYld +
CrachFEM offers a comprehensive approach to meet all of these
challenges. MF GenYld + CrachFEM can be used as a user material model for different commercial explicit-dynamic FE codes.
For industrial use the material models offers reliable phenomenological models for the failure modes given above. For academic
research and a better understanding of the macroscopic failure
phenomena the fracture models can be combined with a damage
plasticity model to achieve a more detailed prediction.
For a shell discretization it is not possible to resolve the onset of
necking directly as the width of the necked region is typically
smaller than the edge length of a shell element. Therefore CrachFEM uses the submodule Crach to model the onset of necking
with a detailed discretization of the neck. The module Crach uses
a plastic material model which accounts for orthotropic plasticity, isotropic-kinematic hardening and strain rate sensitivity. The
discretization is based on an initial imperfection to allow for a
realistic prediction of necking. The strain hardening, strain rate
sensitivity and the hardening due to change of stress state inside
the necked region (change to plane stress condition in neck; increase of stress component normal to sheet) contribute to the
limit strain for instable necking. As the main physical effects are
accounted for, Crach can also be used in case of nonlinear strain
paths. Besides tensile tests no extra tests are needed to predict
the forming limit diagram. For the prediction of fracture (without prior necking) CrachFEM uses phenomenological models for
the phenomena of ductile normal fracture (DNF) and ductile
shear fracture (DSF). DNF accounts for a fracture which is
caused by void growth and void coalescence (in case of ductile
sheets voids do only appear just before fracture) with a fracture
surface normal to the direction of the first principal strain. DSF
accounts for a shear band localization in the material which is
followed by fracture. The equivalent plastic strain at fracture as
a function of a relevant stress state parameter is used as a failure criterion in case of linear strain paths. In case of nonlinear
strain paths the initial fracture limit curves are used as a master
curve for an integral damage accumulation law. In CrachFEM a
tensorial description of damage is used. The limit strains for
DNF and DSF cannot be predicted based on elastoplastic properties as the ductility of metals depends on the microstructure of
the material and must be measured with different kind of specimens. A model for the post-instability strain (PIS) completes the
CrachFEM approach in case of shell discretization. The PIS model accounts for the additional strain in a shell element from the
onset of necking until final fracture by DNF or DSF inside the
neck.
For the forming simulation of thick sheets or bulk forming problems CrachFEM can also be used in combination with solid elements. The algorithm Crach for the prediction of instability is
not used in this case. Necking must be directly resolved by the
FE mesh. The fracture criteria for DNF and DSF in CrachFEM
are based on two different stress state parameters beta and
theta. Both parameters can be used for a general 3d-stress state
as they depend on 2 stress invariants each.
As an example for the advantage of the tensorial damage accumulation in fracture models of CrachFEM a cutting process of a
sheet edge followed by a forming operation of the sheet is discussed.
On a next and even more detailed level the introduction of a damage plasticity model in MF GenYld + CrachFEM (i.e. accumulated damage of CrachFEM influences elastoplastic model in module GenYld) allows also for a mesoscopic modeling of fracture
initiation (i.e. development of shear bands) and fracture propagation. The model for damage plasticity has been introduced in line
with the model suggested by [3]. However a damage plasticity
model causes a mesh dependent solution. Corrective measures
are suggested for this problem. Monotonic crack propagation
problems in aluminium sheets have been modelled as an example.
W. Entelmann1
2
R. Stelling
3
F. Brenner3
G. Metzmacher3
3
H. Gese
7th Forming Technology Forum
Warm and Hot Forming,
September 15–16, 2014
Enschede, the Netherlands
(1) Airbus Operations GmbH,
Bremen, Germany
(2) Premium Aerotec GmbH,
Nordenham, Germany
München, Germany
Relaxation Forming of Double Curved
Fuselage Panels – Assessment of
Formability And Optimization of
Process Parameters With Numerical
Analysis
Room temperature stretch forming is the state-of-the-art technology for the manufacturing of double curved fuselage panels for
aeroplanes. A new generation of AlMgSc alloys opens up a new
manufacturing scenario for those fuselage panels. As the AlMgSc
alloys are stable up to temperatures beyond 300 °C a hot forming
process with subsequent stress relaxation is possible. Within the
research projects NEXXT and INPRO the process of relaxation
forming has been investigated both in hardware tests and with
numerical simulation. The challenge is the identification of process parameters which avoid an early buckling of the panel and
which minimizes the elastic springback finally. The numerical simulation of the forming process allows the ?a priori? optimization of the process parameters. A temperature-dependent material model with orthotropic plasticity and orthotropic creep has
been derived. The FEA code ABAQUS/Standard with a user-defined creep law has been used. To identify the two-dimensional
geometry that is related to the curved target geometry, an inverse Finite Element simulation is performed in a first step. The
actual forming process simulation is based on the derived, flat
panel geometry. The stress relaxation and creeping of the metal
at elevated temperature must be captured precisely to gain a
dependable prediction of buckling as well as springback after
successful forming. Real hardware tests are presented and compared to the findings of the simulation approach.
A.Heath
H. Gese
Modeling Failure for Nonlinear Strain
Paths With CrachFEM
G. Oberhofer
H. Dell
Numisheet 2014, Melbourne,
Australia
München, Germany
This paper describes a general technique for simulating sheet
failure under complex deformation. Separate failure risks are
calculated for unstable necking and ductile fracture modes associated with void growth and shear banding. Necking is detected
by a multi-scale method which considers sheet inhomogeneity,
strain hardening, Bauschinger effects, strain rate sensitivity and
the multi-axial stress-state in the neck. The model is easily calibrated from uniaxial test data and can be combined with a meshindependent treatment of post-necking deformation. Ductile fracture risks are based on damage tensors calculated as integrals of
the effective plastic strain weighted by the stress state. The tensorial formulation accounts for fracture strain recovery following
load reversal. Model parameters can be evaluated with standard
testing machines using tailored specimens. The failure methodology is realized in the software module MF GenYld + CrachFEM,
which can be coupled to all major explicit FEM codes.
In this study the explicit-dynamic solver LS-DYNA® has been
used together with the material model MF GenYld + CrachFEM®. A full characterisation of the relevant aluminium cast
alloy had been available for this study. Besides the viscoplastic
behaviour of the material the material model includes a prediction of material failure due to necking (only in case of shell discretisation), ductile normal fracture and ductile shear fracture.
For both fracture models the equivalent plastic strain at fracture
is a function of a stress state parameter which gives a unique
fracture strain in case of a general three-dimensional stress
state.
A shell discretisation is the typical and effective approach in industrial crashworthiness simulation for sheets, extrusions and
thin-walled castings. However, a shell discretisation of a geometrically complex cast component is not a straightforward approach. The definition of the middle plane is not always clear at
junctions of walls and ribs. The stiffness of T-joints or K-joints
between shells does not represent the stiffness correctly. In
addition, the stress state (e.g. stress triaxiality) is not predicted
quantitatively correct at junctions. An alternative meshing strategy is a full three-dimensional discretisation of the cast component with quadratic tetrahedrons. Meshing of the three-dimensional geometry is pretty straightforward with this kind of elements. However, there is an upper limit of mesh size due to high
CPU effort.
In this study generic components are first used to quantify the
predictive quality of different meshing strategies. Finally, a real
shocktower component test (fracture has been achieved by a high
overload in this test) is modelled with shells and quadratic tetrahedrons (2 mesh sizes each). The results are compared against
the experimental findings. The capability of fracture prediction
in geometries with complex stress states is found to be insufficient and strongly mesh dependent using thin shell elements. In
contrast to that, the discretisation with quadratic tetrahedrons
predicts the location of crack initiation with excellent accuracy.
However, due to the limitations in mesh refinement (model size
of the fine model is already around 2 million elements), the time
of fracture is not predicted with the same accuracy as the location is.
In this study quasi-homogeneous properties are assumed for the
cast material. In future steps the variation of material properties
through the component due to different filling and solidification
history will be accounted for. The initiation of local material properties based on a casting simulation has been already
established in material model MF GenYld + CrachFEM.
H. Gese
G. Oberhofer
VDI-Konferenz, Simvec Spezial,
Simulation des Werkstoffverhaltens für automobile
Anwendungen, Baden-Baden,
10.–11. Dezember 2013
München, Germany
Durchgängiges Materialmodell für
Fertigungs- und Crashsimulation von
Metallen und Kunststoffen
H. Gese
G. Oberhofer
M. Oehm
A. Heath
Forming Technology Forum
2013, 19–20 September, 2013,
Herrsching, Germany
München, Germany
Approach for the Prediction of Sheet
Failure in Multi-Step Forming and
Subsequent Forming and Crash
Simulations
The classical forming limit diagram is mainly used in industrial
sheet metal forming simulation to predict a possible sheet failure. This criterion is based on the onset of localized necking. Its
validity is limited to linear strain paths. Therefore it is not applicable to multistep forming simulations and for subsequent forming and crash simulations. For AHSS, UHSS and aluminium
sheets fracture can also appear without prior necking. This kind
of sheet failure is not covered by the concept of the forming limit
diagram.
The module CrachFEM of material model MF GenYld + CrachFEM offers a comprehensive approach for the prediction of material failure in sheet metals. MF GenYld + CrachFEM can be used
for a discretization with shell and solid elements. For a shell discretization it is not possible to resolve the onset of necking directly as the width of the necked region is typically smaller than
the edge length of a shell element. Therefore CrachFEM uses the
submodule Crach to model the onset of necking with a detailed
discretiza-tion of the neck. The module Crach uses a plastic material model which accounts for orthotropic plasticity, isotropickinematic hardening and strain rate sensitivity. The discretization is based on an initial imperfec-tion to allow for a realistic
prediction of necking. The strain hardening, strain rate sensitivity and the harden-ing due to change of stress state inside the
necked region (change to plane stress condition in neck; increase
of stress component normal to sheet) contribute to the limit
strain for instable necking. As the main physical effects are accounted for, Crach can be used also in case of nonlinear strain
paths. Besides tensile tests no extra tests are needed to predict
the forming limit diagram. For the prediction of fracture (without prior necking) CrachFEM uses phenomenological models for
the phenomena of ductile normal fracture (DNF) and ductile
shear fracture (DSF). DNF accounts for a fracture which is
caused by void growth and void coalescence (in case of ductile
sheets voids do only appear just before fracture) with a fracture
surface normal to the direction of the first principal strain. DSF
accounts for a shear band localization in the material which is
followed by fracture. The equivalent plastic strain at fracture as
a function of a relevant stress state parameter is used as a failure criterion in case of linear strain paths. In case of nonlinear
strain paths the initial fracture limit curves are used as a master
curve for an integral damage accumulation law. In CrachFEM a
tensorial description of damage is used. The limit strains for
DNF and DSF cannot be predicted based on elastoplastic properties as the ductility of metals depends on the microstructure of
the material. A set of specimens is needed, which allow to identify the equivalent plastic strain at fracture for a defined stress
state.
For resume runs in multi-step forming simulations and for a
mapping between forming and crash simulation with different
mesh discretization the full membrane deformation history of an
element must be transferred from one step to the next to ensure
a correct prediction of necking. The deformation history must be
ex-pressed relative to a reference orientation. Therefore also the
orientation of the rolling direction has be mapped. For the fracture models the cumulated values of the damage tensor has to be
mapped.
In the presentation an example for a multi-step forming simulation with LS-DYNA and a process chain of deep drawing and
crash simulation (for cases with AUTOFORM to LS-DYNA and
LS-DYNA to LS-DYNA) will be shown.
F. Brenner1
M. Buckley2
H. Gese1
G. Oberhofer1
1
9th European LS-DYNA Users’
Conference, Manchester, 2–4
June 2013
München, Germany
Influence of Discretisation on
Stiffness and Failure Prediction in
Crashworthiness Simulation of
Automotive High Pressure Die Cast
Components
In recent years, castings are widely used as part of the car
chassis in automobile manufacture because of their light weight
and the flexibility of the design process. Due to the comparable
low ductility of castings, it is essential for crash simulations to
gain dependable analyses. However, modelling casting parts correctly for finite element analyses is an issue for several reasons.
In order to represent the elastoplastic stiffness correctly and
thus to obtain reliable failure predictions, an accurate prediction
of plastic strains and the corresponding stress states is required.
To meet these conditions an adequate material model as well as
an appropriate discretisation are needed. The geometric discretisation is a trade-off between computational costs, meshing effort
and the quality of the results that can be achieved in simulations.
In this study the explicit-dynamic solver LS-DYNA® has been
used together with the material model MF GenYld + CrachFEM.
A full characterisation of the relevant aluminium cast alloy had
been available for this study. Besides the viscoplastic behaviour
of the material the material model includes a prediction of material failure due to necking (only in case of shell discretisation),
ductile normal fracture and ductile shear fracture. For both fracture models the equivalent plastic strain at fracture is a function
of a stress state parameter which gives a unique fracture strain
in case of a general three-dimensional stress state.
A shell discretisation is the typical and effective approach in industrial crashworthiness simulation for sheets, extrusions and
thin-walled castings. However, a shell discretisation of a geometrically complex cast component is not a straightforward approach. The definition of the middle plane is not always clear at
junctions of walls and ribs. The stiffness of T-joints or K-joints
between shells does not represent the stiffness correctly. In
addition, the stress state (e.g. stress triaxiality) is not predicted
quantitatively correct at junctions. An alternative meshing strategy is a full three-dimensional discretisation of the cast component with quadratic tetrahedrons. Meshing of the three-dimensional geometry is pretty straightforward with this kind of elements. However, there is an upper limit of mesh size due to high
CPU effort.
In this study generic components are first used to quantify the
predictive quality of different meshing strategies. Finally, a real
shocktower component test (fracture has been achieved by a high
overload in this test) is modelled with shells and quadratic tetrahedrons (2 mesh sizes each). The results are compared against
the experimental findings. The capability of fracture prediction
in geometries with complex stress states is found to be insufficient and strongly mesh dependent using thin shell elements. In
contrast to that, the discretisation with quadratic tetrahedrons
predicts the location of crack initiation with excellent accuracy.
However, due to the limitations in mesh refinement (model size
of the fine model is already around 2 million elements), the time
of fracture is not predicted with the same accuracy as the location is.
In this study quasi-homogeneous properties are assumed for the
cast material. In future steps the variation of material properties
through the component due to different filling and solidification
history will be accounted for. The initiation of local material properties based on a casting simulation has been already
established in material model MF GenYld + CrachFEM.
H. Gese
G. Metzmacher
M. Oehm
Challenge 2013, 10–11 April
2013, Hanau, Germany
München, Germany
From sheet metal to material input for
FEA – a systematic approach
M. Franzen1
M. Thornagel2
G. Oberhofer3
Integrative Simulation von
faserverstärkten Kunststoffen unter
Crash-Beanspruchung
3
VDI-Konferenz, Kunststoffe im
Automobilbau – Plastics in
Automotive Engineering,
Mannheim, 13.–14. März 2013
(1) Ford Research & Advanced
Engineering Europe
(2) SIGMA Engineering GmbH
München, Germany
Injection-moulded fibre-reinforced thermoplastics show a high
potential for cost effective weight reduction in the automotive
area. However these materials feature a complex mechanical behaviour which is dependent on the injection moulding process.
Local micro-structure and fibre orientation / distribution, which
are highly influencing the elasto-plastic and fracture characteristics, are defined by the manufacturing process.
To improve the crash simulation quality of injection moulded
fibre reinforced parts an integrative simulation approach seems
to be inevitable. Ford Research and Advanced Engineering in
cooperation with the companies MATFEM Partnerschaft
Dr. Gese & Oberhofer and SIGMA Engineering GmbH have developed a methodology to couple injection simulation with crash
simulation which can be used in the development process at Ford
Motor Company. The material model MF GenYld + CrachFEM is
used within RADIOSS® and LS-Dyna® by taking into account
the injection moulding manufacturing process results delivered
by SIGMASOFT®.
Automotive prototype parts have been manufactured and subjected to various load cases to validate the advanced methodology
by correlating the CAE models with the test data in terms of
measured force-deflection curves and observed material failure.
The integrative simulation approach leads to a much better correlation in terms of deflection, deformation and fracture behaviour compared to conventional approaches.
M. Oehm
G. Oberhofer
H. Gese
Enginsoft International
Conference 2012, 22–23
October, Lazise, Italy
München, Germany
Advanced modelling of metals and
thermoplastics for crash applications
under consideration of the process
history
Effective CAE assessment of crashworthiness requires not only
comprehensive material models; local effects introduced by the
manufacturing process should also be considered.
This presentation shows how the predictive material model MF
GenYld + CrachFEM can incorporate the process history from
cold forming of sheets, from Mg high-pressure die casting and
from mould-injection of polymers.
H. Gese
G. Oberhofer
“Vehicle Property Validation
2012”, Automotive Circle
International, Bad Nauheim,
19–20 June 2012
München, Germany
Improved material models for metals
and plastics in crash simulation – a
prerequisite for the virtual validation
of novel lightweight design concepts
Effective CAE assessment of crashworthiness requires not only
comprehensive material models; local effects introduced by the
manufacturing process should also be considered.
This presentation shows how the predictive material model MF
GenYld + CrachFEM can incorporate the process history from
cold forming of sheets, from Mg high-pressure die casting and
from mould-injection of polymers.
M. Reißner1
S. van der Veen2
M. Groß1
Prediction of plane stress fracture
toughness for aluminium sheet
materials with numerical methods
H. Gese1
ESIS-Workshop on
Computational and Experimental Failure Mechanics, 14–15
June 2012, BAM Berlin,
Germany
München, Germany
(2) AIRBUS France, Materials
& Processes, Toulouse, France
The evaluation of the residual strength of aircraft fuselage panels with fatigue cracks is a core topic for a damage-tolerant design. For thinner sheets the plane-stress fracture toughness is
the relevant criterion. It is well known that fracture toughness
decreases with increasing thickness because thick sheets undergo the fracture-prone plane-strain condition. Whereas the planestrain fracture toughness is a real material parameter the planestress fracture toughness depend also on the geometry (thickness, crack length). For some of the assessed aluminium materials, however, the plane-stress fracture toughness decreases also
slightly towards very low sheet thicknesses.
The aim of the presented work is to determine the thickness
dependence of plane-stress fracture toughness (PSFT) by means
of finite-element simulations. C(T) 127 specimens of three different aluminium alloys (well known quality AA2024-T3 and two
other alloys Al–01 and Al–02) and with five different sheet thicknesses from 0.8mm to 3.2mm each were simulated with the FE
code ABAQUS/Explicit in combination with the advanced user
material and failure model MF GenYld + CrachFEM. The material description for the simulations includes an in-plane plastic
orthotropy of the sheet materials and stress state-dependent
fracture models that account for ductile normal and ductile shear
fracture. A shear fracture mode can be predicted for a single
finite element, but an overall shear band across the sheet thickness cannot be modeled with this approach. As the available C(T)
127-experiments show a crack propagation with slanted cracks
(can be single slant or V-shaped), the material model had to be
extended to a damage plasticity approach. Matfem implemented
damage plasticity into MF GenYld + CrachFEM along the models published by Xue et al.. In the current implementation the
flow curve of individual elements is weakened with increasing
fracture risks for ductile normal fracture and ductile shear fracture.
The crack propagation area was modeled with a very fine solid
mesh. The remaining C(T) specimen was modeled with shells ?
using a shell-solid-coupling at the interface. The existing mesh
dependency in damage plasticity was considered by scaling the
fracture curves using a concept by Wilkins et al., which requires
that plastic deformation occurs in a critical radius around the
crack tip before the crack propagates. Fracture curves and the
damage plasticity parameters were calibrated with experimental
data from only one combination (Al–02, 3.2mm). The resulting
numerical parameters were then used for all combinations of alloys and thicknesses.
A parameter study showed that the simulation of a slanted crack
is possible. The variable parameters (damage plasticity parameters, scaling of fracture curve, material description and mesh
size) affect each other, however. For the main simulations one set
of parameters for the damage plasticity model and the scaling of
the fracture curves was determined and used throughout the
project. The results for the different sheet thicknesses and different Al-alloys show differences in the onset of fracture, i.e. the
length of straight crack before slanted crack, and in the mode of
the slanted crack. The force in the simulation, normalized to the
sheet thickness, also shows differences between the different alloys and the different sheet thicknesses. The drop of the normalized force for very low sheet thickness is pronounced for Al–01,
moderate for AA2024-T3 and negligible for Al–02.
The performed simulations show that the used material model is
able to model a slanted crack and a drop of normalized force
response for decreasing thicknesses. Because the same material
model was used for all simulations of a given alloy, the predicted
thickness effects are owed to differences in the stress state. The
material model itself does not account for texture or
microstructure. Hence, the causes of thickness effect are plastic
orthotropy and a different characteristic of fracture strains vs.
stress state, at least to some extent.
H. Gese
H. Dell
2011, 17–18 May 2011, IVP,
ETH Zurich, Switzerland
München, Germany
Contribution of improved material
models to the virtual assessment of
the process robustness of sheet
forming processes
Today, most of the industrial sheet metal forming simulations
are still based on simple orthotropic yield loci, isotropic strainrate independent hardening laws and failure prediction with the
classical forming limit curve. This approach does not exploit the
full potential of the virtual simulation with finite elements. Some
of the pre-optimized parts still fail in the physical try-out runs or
are unstable in the real production process with different material batches. An improvement can be achieved by the identification of materials and processes that are not covered adequately
by the standard simulation approach and introducing improved
and more robust models. Candidates for improved material models are aluminium sheets, austenitic steels and advanced high
strength steels. Typically the classical Hill–1948 yield locus does
not fit for those materials. In addition to the forming limit diagram criteria for ductile normal fracture and ductile shear fracture are needed. This paper focuses on extended criteria for the
prediction of sheet failure. Besides the deterministic prediction of
a failure risk a failure probability can be derived from the scatter
of one material batch. An implementation of a model for failure
probability is discussed in detail. If the behavior of one material
batch can be predicted with sufficient accuracy a stochastic approach can be used on top to account for the influence of the
scatter between consecutive material batches.
M. Franzen1
G. Oberhofer2
M. Thornagel3
3
CrashMat Freiburg, Fraunhofer
EMI, 24.–25. April 2012,
Freiburg, Germany
Engineering Europe
München, Germany
(3) SIGMA Engineering GmbH
Advanced Modeling of Fiber
Reinforced Thermoplastics for Crash
Simulation under Consideration of the
Injection Molding Process
M. Groß1
H. Dell1
S. van der Veen2
T. Billaudeau2
H. Gese1
S. Abdel-Malek3
L.W. Meyer3
7th Nordmetall Colloquium,
18–19 April 2012,
Wasserschloss Klaffenbach
München, Germany
(2) Airbus SAS, Toulouse,
France
(3) Nordmetall GmbH, Adorf,
Germany
Beitrag verbesserter Materialmodelle
zur numerischen Auslegung von
Leichtbaukomponenten – A
Contribution of Advanced Material
Models for Numerical Simulation of
Lightweight Components
U. Weiss1
A. Bach1
La Metallurgia Italiana – n.
11–12/2011
Ford Research & Advanced
Engineering Europe
Magnesium castings offer significant weight saving potential for
many crash-relevant structures in the vehicle. Until now, proper
cast magnesium design was difficult and time consuming, as reliable CAE tools were not available. In the European funded research project NADIA, a new set of CAE tools have been developed for AM60 and AM50 alloys that combine local casting process simulation results with crash failure CAE modelling to reliably predict component level crash behaviour. These CAE tools
have been made commercially available and integrated into existing CAD / CAE codes.
H. Gese
G. Oberhofer
VDI-Fachkonferenz “Simulation
im automobilen Leichtbau”,
Baden-Baden, 22. und 23.
November 2011
München, Germany
M. Buckley1
M. Selig2
M. Oehm3
8th European LS-DYNA Users
Conference, 23 – 24 May 2011,
Strasbourg
(2) Autoform Development
GmbH, Zurich, Switzerland
München, Germany
A new method for CrachFEM damage
parameter calculation and transfer
from Autoform to LS-Dyna
H. Gese
G. Oberhofer
M. Oehm
automotive CAE Grand
Challenge 2011, 19 – 20 March
2011, Hanau, Germany
München, Germany
Contribution of Material Model MF
GenYld + CrachFEM to the Virtual
Assessment of Sheet Forming
Processes
G. Oberhofer
4A Engineering
Technologietage, 3. + 4. März
2011, Schladming, Austria
München, Germany
Modelling Short-Fiber Reinforced
Polymers with Material Model MF
GenYld + CrachFEM
The lecture summarizes the functionality of material model MF
GenYld + CrachFEM for the simulation of short-fiber reinforced
polymers.
H. Gese
Enginsoft International
Conference 2010, 21 October,
Fiera Montichiari, Italy
München, Germany
Advanced prediction of material
failure in metal forming and crash
simulation with material model MF
GenYld + CrachFEM
A wide range of new materials has been introduced by the automotive industry to fullfill the increased needs for lightweight design and passenger safety. Examples are advanced high strength
steels, ultra high strength steels and light-metal structures
(sheets, extrusions and castings from aluminium and magnesium alloys). A predictive simulation of these materials must include a comprehensive failure prediction.
The material model MF GenYld + CrachFEM has been developed
for an advanced failure prediction. This material model can be
coupled to LS-DYNA and other explicit FEA codes via the user
material interface. The module MF GenYld allows for a modular
combination of yield loci and hardening laws to cover the elastoviscoplastic material behaviour. The module CrachFEM can be
used for a failure prediction including criteria for localized necking (in case of shell elements), ductile normal fracture and ductile shear fracture.
Failure prediction for different kind of materials will be shown in
the lecture:
failure prediction in the process chain of deep drawing and crash
of steel sheets failure prediction for aluminium and magnesium
extrusions in crash failure prediction for magnesium high pressure die castings in crash
M. Groß
S. van der Veen
J. Jendrny
Deutsche SIMULIA-Konferenz
2010, Heidelberg, 20.–21.
September 2010
München, Germany
(2) AIRBUS France, Materials
& Processes, Toulouse
(3) AIRBUS Germany,
Materials & Processes, Bremen
Semi-automated bearing strength
prediction of aerospace materials
using Abaqus and Python scripting
C. Lange1
F. Bron2
P. Hänggi1
T. Möller3
H. Gese4
D. Daniel2
C. Leppin1
IDDRG 2010, Graz, 31 May 31 –
2 June
(1) Alcan Technology &
Management, Neuhausen,
Switzerland,
(2) Alcan CRV, Voreppe, France
(3) GOM mbH, Braunschweig,
Germany
München, Germany
Forming simulation of aluminum car
body sheet with different yield models
and comparison with experiments
In recent years, significant effort has been invested in improving
the accuracy of the formability assessment of aluminum car body
panels by numerical deep drawing simulations. Adequate modeling of the plastic deformation of the sheet and the friction between sheet and tools are both crucial for correct predictions of
the final thickness and fracture. In this paper, three different
yield constitutive models are investigated for the AA6016 T4 aluminum alloy car body sheet DR100, using the commercial software package MF GenYld + CrachFEM in combination with the
explicit finite element code LS-Dyna: Barlat 2000, Bron-Besson
2004 and the simple Hill 48 model. For the prediction of necking
failure, the software package also includes a perturbation method along the lines of the original idea by Marciniak and Kuczynski. Comprehensive characterization tests such as tensile,
shear, equi-biaxial tests in different directions are considered to
identify the model parameters. The performance of the three different yield models is compared in deep-drawing tests, namely
stretching with hemispherical punch (LDH test) and deep drawing with a cross-shaped punch (cross-die test). A sensitivity
analysis to friction coefficient is performed to determine the most
appropriate friction values. For an in depth validation, the numerical simulations are compared with tests that are performed
with the Argus optical strain measurement system. Using the
latest software feature developed by GOM, the distribution of the
error between numerical results and tests is visualized and evaluated. As expected, the Barlat 2000 model shows superior performance compared to the simple Hill 48 model. However the BronBesson 2004 yield model shows a significant improvement of the
thickness prediction, especially in the critical area where the
thickness is minimal, and is also recommended for fracture prediction in sheet forming.
H. Gese
H. Dell
G. Metzmacher
13. Workshop Simulation in der
Umformtechnik, Universität
Stuttgart, 19. März 2010
Dr. Gese & Oberhofer, München
Beschreibung der isotropen, isotropkinematischen und anisotropen
Verfestigung von Blechwerkstoffen –
experimentelle Aufnahme und
numerische Modellierung
(Zusammenfassung nicht verfügbar.)
G. Oberhofer
H. Gese
M. Oehm
Challenge 2010" Hanau, 30–31
March 2010
Modelling Non-Reinforced and FiberReinforced Polymers with Material
Model MF GenYld + CrachFEM
The material model MF GenYld + CrachFEM can model non-reinforced and short fibre-reinforced polymers in crash simulations. MF GenYld + CrachFEM is a modular approach: The module GenYld (Generalized Yield Model) can be used to describe
elastoviscoplastic material response of metals and polymers. For
polymers, GenYld allows to describe an anisotropic evolution of
the yield locus, i.e. the hardening can be modelled as a function
of the stress state. This anisotropic hardening can be combined
with an elastic and plastic orthotropy for fibre reinforced polymers. The plastic orthotropy can be different for tension and
compression. This is relevant to describing the real behaviour of
short fibre reinforced polymers. The module CrachFEM predicts
the onset of material failure due to ductile normal fracture and
ductile shear fracture. The fracture strain is a function of the
stress state and can be orthotropic for short-fibre reinforced polymers. The lecture also covers the modelling of structural polymeric foams.
H. Gese
G. Oberhofer
M. Oehm
Challenge 2010" Hanau, 30–31
March 2010
Crash Simulation with Material Model
MF GenYld + CrachFEM based on
Results of Process Simulation
The lecture describes the use of material model MF GenYld +
CrachFEM in multi-trade simulations to evaluate the crashworthiness of components. Multi-trade simulations are analyses of
linked physical phenomena where it seems to be sufficient to
perform one analysis and use the results to modify or establish
the inputs of a second analysis.
The following multi-trade examples are covered in this lecture:
Sheet metal: Deep drawing – crash; Sheet metal: Initialization of
local properties from processing and joining; Cast components:
Casting – crash
H. Gese1
H. Klamser2
P. Stolfig3
M. Groß1
L. W. Meyer4
S. Abdel-Malek4
Nordmetall-Kolloquium 2009
am 2.12. und 3.12.2009 in Adorf
bei Chemnitz (Organisation:
Prof. Dr.-Ing. L.W. Meyer)
(2) Dr.Ing. h.c. F. Porsche AG,
Weissach
(3) Fa. Stolfig GmbH, Geisenfeld
(4) Nordmetall GmbH, Aadorf
Gemeinde Neukirchen
Weiterentwickeltes Materialmodell
zur Bewertung des Crashverhaltens
von Mg-Extrusionsprofilen
Magnesium-Extrusionsprofile bieten prinzipiell ein hohes
Leichtbaupotential in Fahrzeugstrukturen. Beim Entwurf und
der Auslegung von Komponenten aus Mg-Extrusionsprofilen
müssen aber die besonderen Eigenschaften dieser Werkstoffgruppe Berücksichtigung finden. Hier sind das nichtlinear-elastische Verhalten, die ausgeprägte plastische Orthotropie – bedingt durch die geringe Zahl von Gleitsystemen der hdp-Kristallstruktur von Mg – und die anisotrope Verfestigung der Werkstoffgruppe zu nennen.
Im Vortrag wird die Auslegung von Mg-Extrusionsprofilen bei
Crashbelastung behandelt. Im Detail wird dabei ein BumperDoppelkammerprofil aus der Mg-Legierung MnE21 (Hersteller
Fa. Stolfig GmbH) betrachtet. Für die Simulation wird der explizite FEM-Code LS-DYNA zusammen mit dem User-Materialmodell MF GenYld + CrachFEM der Fa. MATFEM herangezogen.
Im Materialmodell MF GenYld + CrachFEM kann eine anisotrope – vom Spannungszustand abhängige Verfestigung – berücksichtigt werden. Bei Mg-Profilen zeigt sich diese durch eine ausgeprägte Asymmetrie der Verfestigung in Zug und Druck.
Gleichzeitig wird die plastische Orthotropie und der
Dehnrateneinfluß berücksichtigt. Ein mögliches Werkstoffversagen wird mit Modellen für lokale Einschnürung, duktilen Trennbruch und duktilen Scherbruch bewertet.
Bei Nordmetall wurden Versuche mit unterschiedlichen statischen und dynamischen Lastfällen an dem Mg-Extrusionsprofil
aus der Legierung MnE21 durchgeführt. Neben den KraftDeformations-Verläufen wurde dabei auch die lokale Dehnungsverteilung auf der Bauteil-oberfläche gemessen. Alle Lastfälle
wurden parallel mit dem beschriebenen Materialmodell simuliert und die Qualität der Prognose bewertet. Vergleichend wurden auch Simulationen mit einfacheren Materialmodellen auf
Basis isotroper Plastizität und isotroper Verfestigung durchgeführt. Die Ergebnisse zeigen, dass nur die Verwendung von erweiterten Materialmodellen eine ausreichend genaue Bewertung
des Crash-Verhaltens von Mg-Profilen zulässt.
M. Groß
G. Oberhofer
H. Gese
ANSYS Conference & 27.
CADFEM Users’ Meeting
Leipzig, 18–20 November 2009
Advanced Modeling of Metallic
Materials in Metal Forming and Crash
with LS-Dyna and Modular User
Material MF GenYld + CrachFEM
(To be submitted.)
H. Gese1
G. Oberhofer1
A. Bach2
U. Weiss2
N. Nowack3
P. Bernbeck4
8th International Conference on
Magnesium Alloys and their
Applications, Weimar, 26–29
October 2009
Engineering Europe
(3) IMPERIA GmbH
(4) MAGMA
Giessereitechnologie GmbH
Crashworthiness Simulation of Mg
Die Casting Parts Including Local
Properties from Process Simulation
For the production of crash resistant Mg die castings, it is necessary to understand all influencing process parameters during the
whole process chain. Up to now the results of the process simulation are often not included into the crash simulation.
This paper will show a new methodology to improve the predictive quality of crashworthiness simulations for Mg die castings
by including local properties from the casting and solidification
simulation into the crash simulation. The work performed within
the European Research Project “NADIA”. The casting simulations and casting trials are presented in a parallel paper by Hepp
et al.
A user-material MF GenYld + CrachFEM which can be coupled
with explicit FEA codes has been used as a basis for this work.
The module CrachFEM is able to account for the different failure
mechanisms in metallic materials (i.e. ductile shear fracture,
ductile normal fracture, localized necking). This material model
has already been successfully used for the prediction of fracture
in quasi-homogeneous metallic materials (e.g. sheets and extrusions). For the prediction of failure in magnesium die cast parts a
yield strength factor, a ductility impairment factor and a porosity
factor have been introduced. These factors are derived from characteristic process parameters coming from solidification simulation (e.g. flow length, local solidification time, physical porosity)
and are used to modify the flow stress curves and failure curves
locally. An extended version of software MAGMAlink has been
used to map the local process parameters from the casting simulation mesh ? based on MAGMAsoft – onto the crash simulation
mesh.
A Mg cast component is used in the NADIA project to validate
the new simulation approach. The results of crash simulations
are compared with crash tests performed in parallel. Components with different ingate systems are used to show the sensitivity of the new approach with local material properties.
H. Gese
M. Groß
EUCOMAS 2009, 1–2 July
2009, Augsburg, Germany
Current developments in modelling
material failure of metallic structures
The introduction of a new generation of alloys into the fuselage
structure of commercial aircrafts is a cost intensive and time
intensive approach. A wide range of properties for the manufacturing process and the in-service behaviour have to be checked
with already large structures (e.g. stretch forming of fuselage panels, wide M(T) specimens for evaluation of residual strength). In
the early stages of the validation process there is even a lack of
sheet material in the necessary dimensions. It would be favourable to have a physically based simulation tool which allows to
predict the behaviour of large structures based on a fundamental
material characterization on specimen level.
The paper will present the application of a comprehensive material model for FEA codes with explicit-dynamic time integration
scheme. This material model allows for a modular combination of
yield loci and hardening models and contains relevant criteria for
the prediction of failure of metallic materials under monotonic
loading:
* failure due to localized necking (this failure is caused by a tensile instability of a sheet material); as localized necking cannot
be resolved by shell elements a special criterion has to be used to
predict the onset of necking based on the macroscopic strains of
the shell element; an algorithm CRACH is used which is based
on a mechanical model with an initial imperfection;
* Failure due to ductile normal fracture (this failure is caused by
void nucleation, void growth and void coalescence)
* Failure due to ductile shear fracture (this failure is caused by
shear band localization).
Whereas the models for ductile normal fracture and ductile shear
fracture can be applied in case of shell and solid discretization,
the criterion for localized necking is used only for shell discretization. The failure models for ductile normal fracture and ductile
shear fracture have been already validated for the correct prediction of fracture limit curves of the alloy AA2024-T3.
However the prediction of material failure depends on a correct
modeling of the local strain distribution in a loaded structure. A
comprehensive description of plastic orthotropy and plastic hardening is a necessary prerequisite. The presentation will give a
number of examples which highlight the necessity of a combination of advanced plasticity and failure models to achieve a good
prediction of failure in loaded structures.
The first example is the prediction of failure in a tensile specimen (AA2024-T3 sheet) with a circular hole. This example is modeled with shell and solid discretization. The failure is initiated
by localized necking at the rim of the hole. Final fracture occurs
in the necked region. The crack propagates until final failure of
the specimen. The use of solid elements in combination with failure criteria for ductile normal fracture and ductile shear fracture allow to predict the specimen behaviour in a correct way.
However a very fine mesh with high computational effort has to
be used. In a second step a shell model is used in combination
with the failure models given above. The onset of necking is predicted correctly by algorithm CRACH. However the use of the
onset of necking as a failure criterion is too conservative. Therefore a model for the post-instability strain is used in addition to
account for the extra strain between onset of localized necking
and final fracture in the neck. With this extended approach a
very effective prediction based on shell discretization is possible.
The second example is a high speed impact of a spherical part on
a metallic sheet structure (e.g. runway debris during landing or
take-off). During this type of impact scenario local strain rates
beyond 1000 1/s can occur locally. A special hardening model has
been implemented for this application. The hardening model allows for a transition from a pure isothermal deformation up to a
pure adiabatic deformation. It can be combined with orthotropic
yield loci to fully cover the behaviour of the relevant aluminium
quality. The experimental programme for the derivation of the
fracture limit curves is discussed. A shell discretization is used
for the FEA simulation. The study shows that the type of failure
initiation depends on the diameter of the impacting sphere and
on the friction condition. For a small impactor diameter the ductile normal fracture is the dominant failure mode. For an increased impactor diameter the dominant failure mode is localized necking. The sensitivity of the modelling approach to rank
different aluminium alloys is discussed.
The third example is the use of the models for ductile normal
fracture and ductile shear fracture to predict the residual
strength of a sheet structure with an initial crack. A detailed model of a double cantilever beam (DCB) with solid discretization is
used. An orthotropic yield locus for the general 3D stress state is
included as the plastic orthotropy can have influence on the direction of the crack propagation. In principal the fracture criteria
can be also applied to crack tip problems. However the results
can be mesh-dependent. Non-local approaches have to be used to
avoid this problem. The results of the detailed simulation of the
crack tip can be used to evaluate relevant criteria (e.g. CTOD) for
the simulation of larger structures with shell discretization.
In the outlook of the presentation the potential to apply the simulation on large structures is discussed.
G. Oberhofer1
A. Bach2
M. Franzen2
H. Gese1
H. Lanzerath2
7th European LS-DYNA
Conference, 14–15 May 2009,
Salzburg
(2) Ford Research &
Advanced Engineering Europe,
Aachen, Germany
A Systematic Approach to Model
Metals, Compact Polymers and
Structural Foams in Crash
Simulations with a Modular User
Material
Today the automotive industry is faced with the demand to build
light fuel-efficient vehicles while optimizing its crashworthiness
and stiffness. A wide variety of new metallic and polymeric materials have been introduced to account for these increased requirements. Numerical analysis can significantly support this process
if the analysis is really predictive. Within the numerical model a
correct characterization of the material behaviour – including
elasto-viscoplastic behaviour and failure – is substantial. The
particular behaviour of each material group must be covered by
the material model.
The user material model MF GenYld + CrachFEM allows for a
modular combination of phenomenological models (yield locus,
strain hardening, damage evolution, criteria for fracture initiation) to give an adequate representation of technical materials.
This material model can be linked to LS-DYNA when using the
explicit-dynamic time integration scheme.
This paper gives an overview on the material characterization of
ultra high strength steels (with focus on failure prediction), nonreinforced polymers (with focus on anisotropic hardening of polymers), and structural foams (with focus on compressibility and
stress dependent damage evolution) with respect to crash simulation. It will be shown that a comprehensive material model –
including damage and failure behaviour – enables a predictive simulation without iterative calibration of material parameters.
A testing programme has been done for each material group in
order to allow a fitting of the parameters of the material model
first. In a second step different component tests have been carried out, which were part of a systematic procedure to validate
the appropriate predictions of the crash behaviour with LS-Dyna
and user material MF GenYld + CrachFEM for each material
group.
Gese, H.
Dell, H.
Obehofer, G.
2009, 5–6 May 2009, IVP, ETH
Zurich, Switzerland
Models for Isotropic-Kinematic
Hardening and Anisotropic Hardening
in Material Model MF Genyld +
Crachfem
The authors have developed the user material MF GenYld +
CrachFEM, which may be coupled to commercial finite element
codes for industrial deep-drawing and crash simulations.
MF GenYld is a modular plasticity model and CrachFEM is a
module for comprehensive failure prediction. MF GenYld offers
various yield loci, all of which may be combined with different
hardening models in order to reflect the observed behaviour of a
wide range of metallic materials. This paper highlights the available models for anisotropic hardening and isotropic-kinematic
hardening. Typically, isotropic hardening is assumed in industrial deep-drawing simulations. Most technical sheet qualities show
a more or less pronounced anisotropic hardening, however,
where hardening is a function of the stress state. MF GenYld allows to scale the yield locus in a defined stress regime. Each
scaling factor can be a function of equivalent plastic strain. The
model for anisotropic hardening can be combined with isotropic
hardening or with isotropic-kinematic hardening. A modified
Chaboche model is available in MF GenYld to describe kinematic
hardening. The module CrachFEM includes the algorithm Crach
for predicting localized necking. For non-linear strain paths the
influence of kinematic hardening on the onset of localized necking must be included. Crach uses a Backhaus model.
Gese, H.
Oberhofer, G.
Oehm, M.
Challenge 2009, 2–3 March
2009, Hanau
Vision and Current Development of
an FEA-Code Independent Material
Model for the Process Chain of
Manufacturing and Crash Simulation
A step forward towards a predictive crash simulation is only possible if the local properties of the involved parts of an automotive
structure are included from a simulation of the manufacturing
process. Examples for such processes are: deep drawing of
sheets, hot forming and quenching of sheets, casting of structural
metallic components and injection moulding of polymer components with or without short fibres
In some cases it is sufficient to map the final properties after the
production process and use them for the initialization of the
crash material model on the level of the gauss points of the FE
model (e.g. local preferred orientation of fibres in injection moulded polymers, hardness and sheet thickness after hot forming
and quenching of sheets, microstructure and porosity in cast metallic structures). In other cases even the deformation history in
the production process is necessary to characterize the material
behaviour for the crash simulation (e.g. deformation history of
sheet in deep drawing process).
Different FEM codes are used typically throughout process simulation and crash simulation. Even the crash simulation might be
be performed with different FEM codes in one development process (code A for components at the supplier, code B for full bodyin-white at the OEM). The material characterization should be
consistent for the full range of the CAE development.
A universal FEM user material model MF GenYld + CrachFEM
for metals and polymers will be presented which can be combined with the most relevant explicit-dynamic FEM crash codes
(Abaqus/ Explicit, LS-Dyna, PAM-Crash, Radioss). The material
model has a modular structure. MF GenYld allows for a flexible
combination of yield loci (isotropic and orthotropic) and hardening laws (isotropic, isotropic-kinematic and general anisotropic)
to model the elasto-viscoplastic behaviour of metallic and polymeric materials. CrachFEM is a modular failure model including
failure criteria for localized necking (for shells) and for ductile
normal fracture and ductile shear fracture.
A lock-and-key principle has been introduced in MF GenYld +
CrachFEM which identifies the type of mapping file. Via flags
the initialization of the Gauss points in the crash model is activated individually for different process simulations (e.g., a mould
injection simulation of a polymer is identified by flag –1 and a
deep drawing simulation by flag –2). In this way the mapping of
all parts of an automotive crash can be done even for a heterogeneous mapping from different types of process simulation. Some
of these couplings are already standard (mapping of deep drawing results including strain history). Some other couplings are
developed in current research projects (mapping of local properties from casting simulation, mapping of fiber orientation of
mould injection simulation).
Simulation examples will be presented to show the current status of the universal material model. The final outlook will identify missing links towards the vision of a general material model in
CAE environment.
Groß, M.1
Gese, H.1
Reese, E.D.2
20. Deutschsprachige AbaqusBenutzerkonferenz,
22.–23. September 2008, Bad
Homburg
(2) EADS Innovation Works,
München
Modellierung von orthotroper
Plastizität und Versagen bei Dellen in
integralen Flugzeughautfeldern mit
dem User-Materialmodell MF GenYld
+ CrachFEM
Das User-Materialmodell MF GenYld + CrachFEM wurde in den
letzten Jahren bei MATFEM entwickelt. Das Materialmodell basiert auf einem modularen Konzept. Das Modul MF GenYld erlaubt die Kombination von weiterentwickelten orthotropen
Fließortbeschreibungen mit verschiedenen Verfestigungsmodellen (isotrop, isotrop-kinematisch und anisotrop). Das Modul
CrachFEM erlaubt die Vorhersage eines Materialversagens
durch lokale Einschnürung (bei Schalenvernetzung) und durch
Scher- und Trennbruch (bei Schalen- und Solidvernetzung).
MF GenYld + CrachFEM kann an FEM-Codes mit explizit-dynamischer Zeitintegration angekoppelt werden. Im Falle von
ABAQUS/ Explicit erfolgte die Kopplung über die Schnittstelle
VUMAT. In diesem Vortrag wird eine Anwendung des Materialmodells MF GenYld + CrachFEM für die orthotrope Modellierung der Plastizität sowie für die Bewertung der Schädigung und
des Versagens von Dellen in integralen Flugzeughautfeldern vorgestellt.
Dellen in einem Flugzeugrumpf können einerseits durch Zusammenstöße mit Transportfahrzeugen während des Rangierens
oder Be- und Entladens statisch – bei vergleichsweise kleinen
Dehnraten – und andererseits in der Start- und Landephase
durch aufgewirbelte Gegenstände dynamisch – bei extrem hohen
Dehnraten – entstehen. Diese Dellen müssen hinsichtlich Schädigung untersucht und bewertet werden. In ernsten Fällen ist eine sofortige Reparatur unumgänglich. Zusätzlicher Inspektionsaufwand ist für Rümpfe in integraler Bauweise notwendig, da
neben Stringer und Hautfeld auch die Schweißnaht untersucht
werden muss.
Um die orthotrope Plastizität des Hautfelds und der Stringer zu
modellieren, wurde der Fließort Barlat–2000 verwendet. Wegen
der hauptsächlich monotonen Belastung wurde für die betrachteten Aluminiumlegierungen eine isotrope Verfestigung angenommen. Im Falle der hochdynamischen Belastung steht ein adiabates Verfestigungsmodell zur Verfügung.
Im Rahmen dieser Arbeit wurde die Rumpfstruktur mit Schalenelementen vernetzt. Die Schweißnaht ist dabei durch ein Makromodell abgebildet. Zusätzlich wurden mit Solids vernetzte Detailmodelle verwendet, um Materialinhomogenitäten in der
Laserschweißnaht berücksichtigen zu können. In diesem Vortag
werden unterschiedliche Aufprallszenarios vorgestellt und mit
den experimentellen Daten verglichen.
Dell, H.
Gese, H.
Oberhofer, G.
Numisheet 2008, September
1–5, Interlaken, Switzerland
Dr. Gese & Oberhofer, Munich
Advanced Yield Loci and Anisotropic
Hardening in the Material Model
MF GenYld + CrachFEM
The authors have developed the user material MF GenYld +
CrachFEM, which may be coupled to commercial finite element
codes for industrial deep-drawing and crash simulations.
MF GenYld is a modular plasticity model and CrachFEM is a
module for a comprehensive failure prediction. MF GenYld offers
various yield loci, all of which may be combined with different
hardening models in order to reflect the observed behaviour of a
wide range of metallic materials. This paper highlights some extended features of the plasticity model MF GenYld. One extended feature is the possibility to superimpose two base yield loci to
better adapt the material behaviour of some textured sheet materials. A new orthotropic yield locus Dell–2006 has been implemented; it allows to model materials with a low yield strength in
shear but elliptical locus in the tension-tension regime for both
plane-stress and general three-dimensional stress conditions.
Another feature in MF GenYld is the so-called yield locus modification, which allows to reach a better representation of the material behaviour. This modification uses factors to scale the locus in
a defined stress regime. This technique allows for example to model an asymmetry in the yield strength in tension and compression. This feature of yield locus modification can be extended to
anisotropic hardening where each scaling factor can be a function of equivalent plastic strain. With the anisotropic hardening
model the behaviour of materials like stable austenitic steel
sheets or magnesium sheets can be covered.
Daniel, D.1
Leppin, C.2
Lange, C.2
Gehrig, M.3
Gese, H.4
Dell, H.4
International Deep Drawing Research Group, IDDRG 2008
International Conference, 16–18
June 2008, Olofström, Sweden
(1) Alcan Centre de Recherches
de Voreppe, Centr’Alp, France
Switzerland
(3) Alcan – Neuf-Brisach,
France
München, Germany
Formability Prediction of Aluminium
Sheet for Car Body Panels
For a proper CAE-based feasibility assessment of the production
processes of such automotive components involving forming, appropriate constitutive models for the deformation and fracture
behavior are crucial. In the following paper, a full mechanical
characterization of the AA6016 T4 aluminum alloy car body
sheet DR100 is presented. A comprehensive experimental program was performed to identify and model the orthotopic elastoplastic deformation behavior of the material and its fracture characteristics including criteria for localized necking, ductile normal fracture and ductile shear fracture. The commercial material
model MF GenYld + CrachFEM has been used in combination
with the explicit finite element code LS-Dyna. Technological experiments, namely, prediction of the Forming Limit Curve (FLC),
stretching with hemispherical punch (LDH test) and deep drawing with a cross-shaped punch have been used to validate the
material model. The focus is on the correct prediction of the limits of the material in such processes.
Kessler, L.1
Gese, H.2
Metzmacher, G.2
Werner, H.3
2008 SAE World Congress. April
14–17, 2008, Cobo Center,
Detroit, Michigan, USA
(1) ThyssenKrupp Steel AG,
Duisburg
(3) BMW Group, Munich
An approach to model sheet failure
after onset of localized necking in
industrial high-strength steel
stamping and crash simulations
In large-scale industrial simulations the numerical prediction of
fracture in sheet metal forming operations as well as in crash
events is still a challenging task of high social and economic relevance. Among several approaches presented in literature, the authors and their colleagues developed a model which accounts
each for three different mechanisms leading finally to fracture in
thin sheet metals: the local instability (necking), ductile normal
fracture and ductile shear fracture. The focus of this paper is to
develop and validate a new approach to improve the predictive
capabilities for fracture triggered by localized necking for a wide
variety of steel grades. It is well known that after the onset of a
local instability additional strain is still necessary to induce fracture. In a numerical simulation using shell elements this post instability strain becomes of increasing importance when the ratio
of the characteristic shell element edge length to its thickness
decreases. Today’s shell element lengths in industrial applications can be of the same order as the width of the necking zone.
Therefore, the post instability strain may contribute to a significant percentage of the shell elements total elongation up to fracture and cannot be neglected any longer.
The enhanced necking model termed Post Instability Strain
Model for Shells (PIS Model) combines the model for localized
necking with the model for ductile shear fracture. Guided by
careful Nakajima type tests of three representative steel grades,
the development of the PIS Model also focuses on minimizing the
influence of varying shell element edge lengths. The model is implemented in such a way that the elongation of a shell element
after onset of necking is highly independent of the element edge
length.
The improvements achieved with this model are demonstrated
by validation examples which include small specimens (tensile
specimen with circular hole), technological sheet metal forming
experiments (Nakajima tests) and finally the 3-point bending
test of an automotive component.
Gese, H.
Oberhofer, G.
Dell, H.
6th German LS-Dyna Forum
2007 October 11 – 12, 2007,
Frankenthal, Germany
Consistent Modelling of Plasticity and
Failure in the Process Chain of Deep
Drawing and Crash with User
Material Model MF-GenYld +
CrachFEM for LS-Dyna
In automotive industry the need to build light weight structures
with increased demands for passenger safety can be significantly
supported by numerical simulation. However a correct material
description is essential. For a correct prediction of failure behaviour of metallic materials in the crash simulation it can be necessary to consider the complete process chain of forming and crash.
Today crash analysis cannot be performed with the same degree
of discretization like in forming simulation. In order to transfer
results between different meshes a mapping process of the relevant element variables has to be introduced.
The user material model MF GenYld + CrachFEM has been developed at MATFEM as an universal material model which can
be combined with finite element code LS-Dyna and other finite
element codes with explicit-dynamic time integration scheme.
The module MF GenYld describes the elastoplastic material behaviour. It has a modular structure which allows to combine different yield loci with a variety of hardening models. Besides different models for isotropic hardening, it includes advanced models for isotropic-kinematic hardening and anisotropic hardening. The comprehensive failure model CrachFEM allows to predict material failure due to localized necking (for shell discretization), ductile fracture and shear fracture. The models for ductile
and shear fracture can be used consistently for shell and solid
discretization. As the algorithm Crach for the prediction of localized necking requires information about the complete deformation history of an elementan orthotropic mapping is inevitable.
Within this investigation the application of material model MFGenYld + CrachFEM for forming and crash simulations is discussed. The technical background for the mapping of the deformation history is introduced.
Gese, H.
8th European Conference for
Forming Automotive
Aluminium and Steel materials,
27–28 September 2007, Bad
Nauheim, Germany
Current status on determining
material characteristics for the design
of sheet metal forming operations
Major motivations for an improved representation of the sheet
material in the virtual development process of forming operations is to reduce the number of critical processes in the press
shop, to predict properties of the formed part (local hardening,
thickness etc.) and to be able to assess new materials without
having empirical knowledge a priori.
The comprehensive test programmes for the characterization of
plasticity and failure for standard and advanced sheet materials
are outlined. These test programmes are more expensive than
standard test suites, but allow for a better representation of the
investigated material in simulation, which also benefits subsequent simulations such as crash or fatigue analyses.
Lanzerath, H.1
Bach, A.1
Failure prediction of Boron steels in
crash
Oberhofer, G.2
Gese, H.2
2007 SAE World Congress. April
16–19, 2007, Cobo Center,
Detroit, Michigan, USA
(1) Ford Research and Advanced
Engineering Europe
Hot-formed steels (here called “Boron steels”) offer a great
weight saving potential versus conventional coldformed steels
used for crash relevant structural parts.
Boron steels allow complex shaped parts due to the hotforming
process, which can be a direct or indirect process. In the direct
hot forming process first the sheet metal with an initial yield
strength of around 400 MPa is blanked and then heated in an
oven up to some 950 °C. In the next step the “hot” sheet metal is
stamped and at the same time rapidly cooled down (quench hardening process) in the stamping die.
During this process the yield strength increases up to approximately 1100 MPa in the final stamped part. Due to the enormous
strength and the very good dimensional control (nearly no
springback), more and more hot-formed parts are used in vehicle
design. Especially in the body structure hot-formed steels are
used for crash relevant parts.
However the quench hardening process causes a significant reduction of the material ductility. The designer has to ensure via
crash simulation that the crash impact can be handled without
fracture of the Boron steel component. The prediction of the deformation and fracture behavior therefore is one of the major
interests during the vehicle development process.
The paper will give an overview on material properties of Boron
steels, a failure modeling approach for crash simulation, the validation of the failure modeling approach and application examples that demonstrate the lightweight potential of Boron steels
for structural applications and the benefit of the chosen simulation approach.
Dell, H.
Gese, H.
Oberhofer, G.
NUMIFORM ’07, Materials
Processing and Design:
Modeling, Simulation and
Applications, 18 – 21 June 2007
München, Germany
Approach for the Prediction of Sheet
Metal Failure
A correct prediction of a possible sheet metal failure is essential
to sheet metal forming simulations. The use of the conventional
forming limit curve (FLC) is the standard approach on industrial
level for this problem. The FLC concept is limited to the case of
linear strain paths, however. The initial FLC is no longer valid in
the case of nonlinear strain paths. The algorithm Crach allows
for a transient prediction of the forming limit for localized necking in the case of arbitrary strain paths. For high strength steels
and aluminium sheets there is also a risk of fracture without
prior localized necking. This paper presents a fracture model
that accounts for ductile fracture (caused by void nucleation, void
growth and void coalescence) and shear fracture (caused by shear
band localization). For both types of fracture, stress state parameters are introduced which can be applied for the plane stress
state and for the general 3D stress state. The fracture limits are
defined by the equivalent plastic strain at fracture as a function
of the stress state parameter based on different experiments
with nearly constant stress state parameter. These fracture limit
curves are a basis for an integral damage accumulation in the
sheet metal forming simulation for arbitrary nonlinear strain
paths. The model Crach for the prediction of localized necking
and the two fracture models for ductile and shear fracture are
combined in the comprehensive failure model CrachFEM. CrachFEM can be linked via a user material model MF GenYld to different explicit FEM codes.
Leppin, C.1
Daniel, D.2
Formability Prediction of Aluminum
Sheet in Automotive Applications
Shahani, R.3
Gese, H.4
Dell, H.4
NUMIFORM ’07, Materials
Processing and Design:
Modeling, Simulation and
Applications, 18–21 June 2007
Switzerland
(2) Alcan – Centre de
Recherches de Voreppe,
Centr’Alp – Voreppe, France
(3) Alcan – Neuf-Brisach, ZIP
Rhénane Nord – Biesheim,
France
(4) Matfem Partnerschaft
München, Germany
In the following paper, a full mechanical characterization of the
AA6016 T4 aluminum alloy car body sheet DR100 is presented. A
comprehensive experimental program was performed to identify
and model the orthotopic elastoplastic deformation behavior of
the material and its fracture characteristics including criteria for
localized necking, ductile fracture and shear fracture. The commercial software package MF GenYld + CrachFEM in combination with the explicit finite element code Ls-Dyna is used to validate the quality of the material model with experiments, namely,
prediction of the FLD, deep drawing with a cross-shaped punch
and finally, analysis of a simplified hemming process using a
solid discretization of the problem. The focus is on the correct
prediction of the limits of the material in such processes.
Oberhofer, G.1
Franzen, M.2
Dell, H.1
Kunststoffe+Simulation 13. –
14. Juni 2007,
Fellbach/Stuttgart
(2) Ford Forschungszentrum
Aachen GmbH
Modelling of Non-Reinforced
Polymeric Material Behaviour in the
Explicit FEM Method for Crash
Simulation
The material behaviour of polymer materials in automotive
crash simulation has hitherto been described with limited accuracy due to the fact that advanced material models, which take
into account the complex yield and fracture behaviour of polymeric materials, are commercially not available. To significantly
improve crash simulations, which means the improved prediction of elasto-plastic behaviour as well as failure behaviour, advanced polymer material models, material data analysis and material tests are inevitable.
In this presentation a material model, which is able to describe
the complex hardening and failure behaviour will be presented.
This includes suitable methods for material testing, material
data evaluation as well as material data preparation for the
input in Finite Element Codes. MATFEM has developed the material model MF-GenYld which, as well as taking into account
the varying hardening behaviour in tension, compression, shear
and biaxial loading, is also capable of accounting for the high
degree to which the hardening behaviour depends on the strain
rate. This material model has been coupled with CrachFEM for
taking into account the influence of loading condition and strain
rate onto the fracture prediction.
On behalf of the Ford Forschungszentrum Aachen GmbH extensive experimental tests have been carried out for non-reinforced
polymeric materials. In a research project Ford Forschungszentrum Aachen GmbH and MATFEM validated the material model
for automotive applications. For this case the process of material
data processing for the use in material models for the explicit
FEM simulation is pointed out and first validation results are
shown.
Gese, H.
Oberhofer, G.
Dell, H.
NAFEMS Conference “Material
Modeling”, December 5 – 6,
2006, Niedernhausen near
Wiesbaden, Germany
MF GenYld + CrachFEM – A Modular
Material and Failure Model for
Structural Materials to be Used in
Metal Forming and Crash Simulations
The material and fracture model MF GenYld + CrachFEM has
been developed as a universal material model to be combined
with finite element codes based on the explicit time integration
scheme. In this way, it allows a consistent description of a structural material – metal or polymer – in a development process
using different finite element codes at various suppliers and
OEMs. The material model MF GenYld has a modular structure
which allows to combine different yield loci with a variety of hardening models. Besides different models for isotropic hardening,
it includes advanced models for isotropic-kinematic and anisotropic hardening. The comprehensive failure model CrachFEM allows to predict material failure due to localized necking (for shell
discretization), ductile fracture and shear fracture. The models
for ductile and shear fracture can be used consistently for shell
and solid discretization.
A universal material model approach enables the use of standardized experimental and theoretical methods for the evaluation of
the required material parameters. This might improve the process of quality assurance in the field of material data processing.
Oberhofer, G.1
Gese, H.1
Groß, M.1
Kühling, M.2
Numerical Analysis of the Balloon
Dilatation Process Using the Explicit
Finite Element Method for the
Optimization of a Stent Geometry
Seidel, D.2
LS-Dyna Forum 2006, 12.–13.
Oktober 2005, Ulm
(2) Boston Scientific
Technologiezentrum GmbH,
München
Endovascular stent surgery is a minimally invasive surgical procedure to treat disorders of the circulatory system as blockage of
blood vessels caused by the build up of plaque (fatty deposits, calcium deposits, and scar tissue) in the arteries, a condition called
atherosclerosis. Nearly all of the medium-sized and large blood
vessels in the body’s vascular system can be accessed by a catheter system. This fact has contributed to a rapid increase in the
performance of endovascular stent surgery.
Before implantation the stent is crimped onto a balloon which results in a diameter reduction. The balloon catheter with the collapsed stent is placed in the narrowed artery. In the blood vessel
the stent is dilated through inflation of the balloon. Then, the
balloon catheter is deflated, leaving the stent in place to hold the
artery open. The catheter and the guide wire are removed. Deflation of the balloon leads to a certain amount of recoil, reinforced
by the outer pressure of the blood vessel. The remaining mean
stresses at this state are the initial condition for a possible fatigue calculation. Due to heartbeat induced blood pressure oscillation the stent is exposed to high cycle fatigue loading on one
side and to low cycle fatigue loading due to daily body movement
on the other side.
The introduction of new stent materials and the optimization of
the stent design can be supported by Finite Element Simulation.
Within this study the load steps crimping, balloon dilatation and
recoil will be investigated. The correct modelling of the described
load steps with the finite element method demands the use of an
isotropic-kinematic hardening model as changes in the loading
direction appear. Therefore the Chaboche model for a combined
isotropic-kinematic hardening has been used in the virtual development process with the FEM-code. The Chaboche model is an
optional hardening model in the general user material MF GenYld developed by MATFEM. Additionally it has to be taken into
account that the maximum fracture strain is a function of the
stress state. The algorithm CrachFEM can be coupled with MF
GenYld for a prediction of fracture initiation during dilatation.
As a prerequisite for the simulation of the balloon dilatation process a folded structure of the balloon has to be generated. For the
given balloon geometry – with conical shapes at both ends – analytical folding tools for airbags exhibit problems in mapping the
unfolded structure to the folded structure. Therefore the balloon
folding process has been simulated directly with FEA.
The explicit FEA code LS-DYNA in MPP version has been used
for all analyses in this project. The explicit-dynamic integration
method has significant advantages for FEA model with a great
number of DOFs (CPU times increases only linearly with number of DOFs) and complex contact conditions (i.e. contact of tools
and stent, possible self contact in the final phase of crimping and
contact between stent and balloon).
Kühling, M.1
Scheuermann, T.1
Seidel, D.1
Oberhofer, G.2
Gese, H.2
Metzmacher, G.2
Gross, M.2
Novel Learning Tool to predict
Mechanical Stent Performance for
Next Generation Materials Beyond
Stainless Steel
World Congress of
Biomechanics WCBM, Jul 29 –
Aug 4, 2006, Munich, Germany
Background. Expertise on stent design development with stainless steel (SST) materials and close-to-SST alloys is available at
Boston Scientific. But new-generation (NG) metal alloys beyond
SST with other material deformation characteristics during
crimping and deployment cannot leverage these experiences.
(1) Boston Scientific
Technologizentrum GmbH,
München
Motivation. During the prototype development of a balloon-expandable iliac stent, BCTZ has developed new concepts with
MATFEM which assisted us in a highly valuable way during design development and freeze. With the mutually developed simulation tool, variations of designs and processes (e.g. for heattreatments) can easily be evaluated and the resulting stent performance can be assessed. This modular concept contains standalone stent design tools, balloon tools and material cards.
Method. Easy specimen test data generated at test are transferred to MATFEM. MATFEM generates individual material cards
per process run. These material data and the specific design data
are put together in the FEM software LS-Dyna. This numeric
data model then simulates stent performance from crimping, deployment and compression.
Two diagrams record mechanical stent data (crimp springback,
stent length change, recoil, compression resistance) during expansion. A third diagram records acute fracture resistance during expansion. With these diagrams, stent criteria can easily be
balanced. The impact of design – or process changes – on stent
performance can be predicted.
Summary. This method’s great value is to learn precisely about
NG materials’ mechanical behaviour and to optimise stent designs with respect to that specific characteristic. Since all the
material data come from simple specimen testing (e.g. on rings),
there is no need to have stent fabrication processes for these NG
materials available.
The method’s value and the predicted results for stent performance were benchmarked against stent testing. The results are
very promising.
Gese, H.
Dell, H.
FLC Zürich, March 15 – 16,
2006, IVP, Eidgenössische
Technische Hochschule Zürich,
Switzerland
Numerical Prediction of FLC with the
Program Crach
This paper describes the numerical algorithm Crach for the prediction of localized necking in sheet metal. Crach is based on an
instability model with an initial imperfection. In contrary to the
Marciniak model the imperfection has a finite dimension in the
Crach algorithm which improves the accuracy of the model for
cases with biaxial tension. Crach uses a comprehensive material
model which includes all relevant effects as plastic orthotropy,
combined isotropic-kinematic hardening and strain rate sensitivity. Therefore Crach can be applied to a wide variety of materials and boundary conditions (i.e. necking during hot forming or
necking at high strain rates). The special strength of Crach is the
prediction of necking for nonlinear strain paths. This ability is
based on an accurate model for isotropic-kinematic hardening.
Two different software products have been developed on the
basis of the algorithm Crach at MATFEM. The PC software
CrachLab is a tool for the prediction of the initial FLC or the FLC
after a defined pre-straining. All necessary input can be derived
from standard tensile tests. CrachLab has been validated for a
wide range of sheet qualities. The most attractive application of
the algorithm Crach is the software CrachFEM. CrachFEM can
be directly coupled with finite element codes for sheet metal
forming and crash simulation for a transient prediction of localized necking for shell elements with arbitrary strain paths. Criteria for ductile and shear fracture have been included in CrachFEM to cover the whole variety of fracture modes for sheet materials.
Gese, H.
Dell, H.
Oberhofer, G.
LS-Dyna Forum 2005, 20.–21.
Oktober 2005, Bamberg
Systematische Aufbereitung von
Werkstoffdaten für die
Blechumformsimulation
Motivation. Die geometrische Diskretisierung der Werkzeuge
und der Platine in der industriellen Blechumformsimulation mit
der Finite-Elemente-Methode hat heute bereits einen sehr hohen
Stand erreicht. Das größte Potential zur Verbesserung der
Vorhersagegenauigkeit steckt im Moment in der Entwicklung
und im Einsatz von höher entwickelten Materialmodellen. Das
Materialmodell umfasst dabei die Beschreibung der elastoplastischen Eigenschaften des Bleches und dessen Versagensgrenzen.
Fließortmodelle. Neben dem bisher überwiegend in der Blechumformsimulation verwendeten Fließort nach Hill 1948, welcher
nur für eine begrenzte Gruppe von Stahlblechen ausreichend
genau ist, wurden zwischenzeitlich höher entwickelte Fließorte
in LS-DYNA implementiert (Barlat-Lian 1989, Barlat 1996, Barlat 2000). Diese erlauben eine Beschreibung von Al-Blechwerkstoffen und eine differenzierte Beschreibung von unterschiedlichen Stahlblechgüten. Während für die Definition des Fließortes nach Hill–1948 die Angabe von R-Werten in 3 Orientierungen zur Walzrichtung (R-Wert beschreibt die plastische Anisotropie des Bleches) ausreichend ist, erfordert die Anpassung
der höher entwickelten Fließorte ein erweitertes Prüfprogramm.
Neben dem uniaxialen Fließwiderstand müssen auch Fließgrenzen bei anderen Spannungszuständen herangezogen werden. Im
Vortrag wird ein Versuchsprogramm vorgestellt, welches die
Messung des äquibiaxialen Fließwiderstandes und des Schubfließwiderstandes beinhaltet. Mit einem Optimierungsprogramm
können aus allen Versuchsdaten die Parameter des Fließortmodells abgeleitet werden. Beispielhaft wird dieses Vorgehen im
Vortrag
anhand
des
Fließortes
nach
Barlat
2000
(*Mat_Barlat_Yld2000 in LS-DYNA) beschrieben. Die Unterschiede zum Fließort nach Hill–1948 werden diskutiert.
Verfestigungsmodelle. Wie sich der Fließwiderstand mit zunehmender Formänderung und in Abhängigkeit der Dehnrate entwickelt, wird durch das Verfestigungsmodell beschrieben.
Üblicherweise wird bei Tiefziehsimulationen ein isotropes Verfestigungsmodell ohne Dehnrateneinfluss verwendet ? was eine
proportionale Aufweitung des anfänglichen Fließortes mit zunehmender Vergleichsformänderung bedeutet. Für einige Blechwerkstoffe (z.B. weiche Tiefziehstähle) ist die Berücksichtigung
des Dehnrateneinflusses für ein brauchbares Simulationsergebnis jedoch notwendig.
Bei Blechumformoperationen (Tiefziehen, Innenhochdruckumformen u.a.) können lokal sehr große plastische Vergleichsdehnungen auftreten. Um diese Prozesse mittels FEM-Simulation
genau abbilden zu können, müssen Fließwiderstandskurven bis
zu den maximal auftretenden Vergleichsformänderungen vorgegeben werden. Typischerweise wird hierzu die Fließwiderstandskurve aus dem uniaxialen Zugversuch in Walzrichtung mit analytischen Verfestigungsmodellen approximiert und für hohe
Formänderungen extrapoliert. Diese Vorgehensweise ist mit entsprechenden Unsicherheiten behaftet. Im Vortrag werden experimentelle Methoden behandelt, welche es erlauben die Verfestigung eines Blechwerkstoffes auch bei hohen Formänderungen zu beschreiben.
Versagensmodelle. Ein wesentliches Ziel der Umformsimulation
ist die Prognose eines Werkstoffversagens und dessen Vermeidung durch iterative Optimierung der Prozessparameter. Deshalb ist es notwendig, dass die verwendeten Materialmodelle
auch geeignete Versagenskriterien beinhalten. Insbesondere bei
Al-Blechwerkstoffen und höherfesten Stählen muss ein potentielles Versagen in Betracht gezogen werden. Bei der monotonen
plastischen Verformung von Blechen sind grundsätzlich zwei
unterschiedliche Versagensarten festzustellen ? der duktile
Trennbruch (ausgelöst durch Porenbildung, -wachstum und -vereinigung) und der Scherbruch (ausgelöst durch Scherbandlokalisationen).
Darüber hinaus kann in dünnwandigen Strukturen ? beispielsweise Blechen – eine lokale Einschnürung (Zuginstabilität) dem
Bruchversagen unmittelbar vorausgehen. Da die Breite der lokalen Einschnürung aber in der Größenordnung der Blechdicke
liegt, kann die Einschnürung bei Verwendung von Schalenelementen im Simulationsmodell nicht direkt abgebildet werden.
Deswegen muss die Instabilität als eigenständiges Versagenskriterium behandelt werden.
Die bisher in kommerziellen FEM-Programmen verfügbaren Materialmodelle sind hinsichtlich der Versagensbeschreibung noch
sehr eingeschränkt. Meist wird bei der Blechumformsimulationen nur das konventionelle Grenzformänderungsschaubild herangezogen. Dieses gilt eingeschränkt nur für den Fall linearer
Dehnungspfade. Für die Bewertung eines Versagens durch duktilen Trennbruch oder Scherbruch liegen keine umfassenden Modelle vor. Ein für alle Versagensarten der Blechumformung maßgeschneidertes Versagensmodell wurde von MATFEM mit der
Software CrachFEM realisiert, welche als User-Materialmodell
an LS-DYNA gekoppelt wurde. Im Vortrag werden die notwendigen Versuche zur Kalibrierung des Versagensmodells von
CrachFEM diskutiert.
Anwendungsbeispiele. Anhand von einigen Umformbauteilen
wird im Vortrag der Einfluss der Qualität des Materialmodells
auf die Wiedergabegenauigkeit der Dehnungsverteilung und der
Prognose des Versagens aufgezeigt.
Werner, H.1
Hooputra, H.1
Weyer, S.1
Applications of Phenomenological
Failure Models in Automotive Crash
Simulations
Gese, H.2
VIII International Conference
on Computational Plasticity
COMPLAS, Oñate E.; Owen
D.R.J. (Eds.) – CIMNE,
Barcelona, 2005
(1) BMW AG, Forschungs- und
Innovationszentrum, München,
Germany
München, Germany
The numerical prediction of the structural behaviour in extreme
loading cases like crash events continues to be a challenging task
in the foreseeable future. The need to reduce weight which in
turn may lead to reductions in fuel consumption, CO2 emissions
and an increase in driving dynamics can only be accomplished by
materials combining high-strength with sufficient ductility.
Therefore, it is of increasing importance to simulate the failure
modes of a body in white as well as energy absorbing trim components within the passenger compartment or the exterior to mitigate pedestrian impact situations.
In the presentation the essential features of a comprehensive failure model based on macroscopic stresses and strains will be presented. This model takes three different mechanisms into consideration which lead to failure in structures:
Instability which manifests itself in the formation of a neck in
thin sheet (metals) loaded in the membrane plane. Necking in
itself is not equivalent to failure, but it is immediately followed
by material failure and can therefore be utilized under certain
conditions in numerical modelling. Typical fields of application
are structures, discretized with coarse shell elements.
Ductile fracture which is conceptually based on the growth, nucleation and finally coalescence of microvoids up to fracture.
Shear fracture induced by a localization of shear bands.
Although the conceptual splitting into three failure modes mentioned above is well justified for metallic materials, it is used for
plastics as well. The basic methods of material characterisation
to establish the parameters of the material model are similar.
Differences arise from fabrication effects like sheet forming, die
casting of components or the fibre orientation in reinforced plastics. By mapping the dominant parameters of the fabrication process as initial conditions to a subsequent crash simulation, some
of these effects are incorporated in the failure analysis. The capabilities of the failure model are demonstrated by comparing component tests from structural components to experimental tests.
Axially crashed aluminium profiles, joined with self-piercing rivets whose failure characteristics are also modelled by a phenomenological model will be shown. Further examples include impact tests on ribbed plastic components where failure plays an
essential role. A structural reinforcement component subjected
to bending will serve as an example for the necessity of using a
comprehensive model to simulate the failure process.
Kessler, L.1
Beier, Th.1
Werner, H.2
Horstkott, D.3
Dell, H.4
Gese, H.4
6th Numisheet Conference
2005, “On the Cutting Edge of
Technology” August 15–19,
2005, Detrit, Michigan, U.S.A.
(1) ThyssenKrupp Stahl AG,
Dortmund, Germany
(2) BMW Group, Forschungsund Innovationszentrum,
Munich, Germany
(3) ThyssenKrupp
Umformtechnik GmbH,
Bielefeld, Germany
Dr. Gese & Oberhofer, Munich,
Germany
Material Selection for an Ultra High
Strength Steel Component Based on
the Failure Criteria of CrachFEM
An increasing use of combining more than one process step is noticed for coupling crash simulations with the results of forming
operations – mostly by inheriting the forming history like plastic
strain and material hardening. Introducing a continuous failure
model allows a further benefit of these coupling processes; it
sometimes can even be the most attractive result of such a work.
In this paper the algorithm CrachFEM for fracture prediction
has been used to generate more benefit of the successive forming
and crash simulations – especially for ultra high strength steels.
The choice and selection of the material grade in combination
with the component design can therefore be done far before the
prototyping might show an unsuccessful crash result; and in an
industrial applicable manner.
Gese, H.
Dell, H.
Oberhofer, G.
DVM-Tag 2005: Dünnwandige
Strukturbauteile, 27.–29. April
2005, Berlin
Verbesserte Materialmodelle für die
virtuelle Auslegung von
Leichtbaustrukturen im Automobilbau
bei Crashbelastung
Um die steigenden Anforderungen an die Crashsicherheit einer
Fahrzeugkarosserie bei gleichzeitigem Leichtbau zu erfüllen,
wurden in den letzten Jahren eine Vielzahl neuer Werkstoffe im
Automobilbau eingeführt. Hierbei sind höher- und höchstfeste
Stähle, Aluminium- und Magnesiumwerkstoffe sowie unverstärkte und kurzfaserverstärkte Kunststoffe zu nennen. Für jedes
Bauteil ist abzuwägen, mit welchem Werkstoff die gestellten Anforderungen erfüllt werden können. Im Hinblick auf die virtuelle
Fahrzeugentwicklung stellt dabei die prozessübergreifende Auslegung von Komponenten unter Berücksichtigung ihres Fertigungsprozesses eine besondere Herausforderung dar. Neben einer adäquaten Beschreibung der elasto-viskoplastischen Werkstoffverhaltens ist die korrekte Prognose des lokalen Werkstoffversagens von zentraler Bedeutung zur Vorhersage der Bauteildeformation und seiner Energieaufnahme.
Im Vortrag wird eine Übersicht über verbesserte Plastizitätsund Versagensmodelle gegeben. Zudem werden Simulationstechniken beschrieben, welche es ermöglichen, den Fertigungsprozess der Komponenten mit zu berücksichtigen. Neben bereits realisierten Entwicklungen werden auch noch existierende Defizite
aufgezeigt. Beispiele in der Übersicht sind ein durchgängiges
Plastizitäts- und Versagensmodell für Blechwerkstoffe zur Beschreibung der Prozesskette Tiefziehen-Crash und ein neues Materialmodell für Kunststoffe, welches die anisotrope Verfestigung
dieser Werkstoffgruppe abbildet. Am Beispiel von dünnwandigen
Gussbauteilen werden noch existierende Defizite bei der Berücksichtigung der Prozesshistorie (Einfluss der Gieß- und Erstarrungsparameter auf die mechanischen Eigenschaften) aufgezeigt.
Gese, H.1
Keller, S.2
Dell, H.1
NeMa-Tagung “Numerische Simulation –
Verarbeitungsprozesse und
prozessgerechte
Bauteilgestaltung, Fokus
Metalle”, Bayreuth 2. – 3.
November 2004
Nederlingerstrasse 1, D–80638
München
(2) Hydro Aluminium
Deutschland GmbH, Forschung
und Entwicklung, Bonn
Verbesserte Plastizitäts- und
Versagensmodelle in der
Umformsimulation
Mit der heute verfügbaren Rechnerleistung sind Blechumformsimulationen auf Basis der Methode der finiten Elemente mit einem hohen geometrischen Diskretisierungsgrad möglich. Das
größte Potential für eine Erhöhung der Prognosegenauigkeit
liegt in der Verwendung von höher entwickelten Plastizitätsund Versagensmodellen. In diesem Beitrag wird ein mögliches
Vorgehen am Beispiel des Tiefziehens einer AlMgMn-Blechlegierung aufgezeigt.
In einem ersten Schritt werden unterschiedliche Fließorte diskutiert, die eine Möglichkeit bieten, die Plastizität von Al-Blechen
in guter Näherung abzubilden. Das Versuchsprogramm zur
Messung der Fließwiderstände bei unterschiedlichen Spannungszuständen wird vorgestellt. Im Weiteren wird das Verfestigungsverhalten behandelt und es werden experimentelle Methoden zur Messung des Fließwiderstandes bei hohen Vergleichsformänderungen diskutiert. Eine genaue Beschreibung der lokalen plastischen Verformung ist eine grundlegende Voraussetzung für eine gute Versagensprognose. Zur Versagensprognose
wird bei der Blechumformsimulation konventionell das Grenzformänderungsdiagramm (GFD) eingesetzt. Dieses dient zur
Vorhersage einer lokalen Einschnürung mit nachfolgendem
Bruch. Es ist aber nur bei linearen Dehnungspfaden gültig. Andere Versagensarten, wie der Scherbruch oder der duktile Trennbruch, werden heute bei der Versagensprognose meist nicht herangezogen. In diesem Beitrag wird das neu entwickelte Versagensmodell CrachFEM vorgestellt, welches gleichzeitig die
Versagensgefahr hinsichtlich lokaler Einschnürung, duktilem
Trennbruch und Scherversagen für nichtlineare Dehnungspfade
bewerten kann. Das Potential dieses Versagensmodell wird im
Beitrag am Beispiel der Umformgrenzen beim Tiefziehen eines
rechteckigen Napfes (“Butterdose”) aufgezeigt. Die Simulationen
wurden mit dem kommerziellen FEM-Code PAM-Stamp durchgeführt.
Hooputra, H.1
Gese, H.2
Dell, H.2
A Comprehensive Failure Model for
Aluminium Extrusions
Werner, H.1
International Journal of
Crashworthiness, Vol. 9, No.5,
Woodhead Publishing (2004),
pp. 449–463.
(1) BMW Group, Forschungsund Innovationszentrum,
Knorrstrasse 147, D–80788
München
Nederlingerstrasse 1, D–80638
München
A correct representation of the plastic deformation and failure of
individual component parts is essential to obtaining accurate
crashworthiness simulation results. The aim of this paper is to
present a comprehensive approach for predicting failure in a
component based on macroscopic strains and stresses. This approach requires the use of a number of different failure mechanism representations, such as necking (due to local instabilities),
as well as ductile and shear fracture. All failure criteria have
been developed in a way to include the influence of non-linear
strain paths. The effectiveness of this approach in predicting failure is then discussed by comparing numerical results with test
data by three point bending and axial compression tests of double chamber extrusion components. All studies presented in this
paper were carried out on extrusions made from aluminium alloy
EN AW–7108 T6.
Gese, H.1
Werner, H.2
Hooputra, H.2
Dell, H.2
Failure Model for Metallic Structures
in Sheet Metal Forming and Crash
Simulation
Heath, A.3
Europam 2004, October 11th to
13th, Paris, France
Metallic materials generally fail due to one or a combination of
the following mechanisms after plastic deformation:
ductile fracture (based on initiation, growth and coalescence of
voids), shear fracture (based on shear band localisation), instability with localised necking (followed by ductile or shear fracture
inside the neck area).
(2) BMW Group, München
(3) ESI GmbH, Eschborn
The first two fracture modes have been implemented in PamCrash as material model 52 for solid elements in a common project of BMW AG, ESI and MATFEM. The two fracture modes are
treated independently from each other. An integral fracture criterion according to V.L. Kolmogorov is used for the damage accumulation in both fracture modes in material 52. The equivalent
plastic strain is a function of the stress triaxiality for ductile fracture and the newly introduced stress parameter for shear fracture (by Professor H. Dell). The potential of the Kolmogorov-Dell
model (material 52) is highlighted in the presentation by a geometrically detailed analysis of a spot weld failure using solid elements. Material 52 covers both failure modes occurring in experiments with different sheet qualities – shear failure of spot weld
and peeling of spot weld.
Instability is necessary as a third failure criterion, if thin walled
structures are discretized with shell elements. The effect of localized necking (well known from the forming limit curve in deep
drawing) cannot be modelled directly with shell elements as the
width of the localized neck is in the dimension of the sheet thickness and the typical edge length of the shells are 5–10 times
sheet thickness in crash simulation. An algorithm CRACH is
used to predict localized necking based on the macroscopic
strains of the shell elements. The CRACH algorithm uses an isotropic-kinematic hardening model to account for nonlinear strain
paths. The algorithm CRACH for the prediction of localized necking has been combined with the criteria for ductile and shear
fracture to form the comprehensive failure model CrachFEM. A
tensorial damage accumulation for ductile and shear fracture is
used in this case to account for nonlinear strain paths in extension to material 52. CrachFEM has been linked to material model 128 in PamCrash V2004. CrachFEM transiently predicts the
failure risks for the 3 fracture modes. Element elimination can
be activated based on a critical value of the overall failure risk of
an element (maximum of all 3 failure modes). The effectiveness
of this approach in predicting failure is then discussed by comparing numerical results with test data in axial drop tests of double
chamber extrusion components. The results presented were carried out on extrusions made of aluminum alloy AA7108.50-T6.
The outlook of this presentation discusses the industrial use of
material 128 in the crash analysis at BMW.
Schmid, M.1
Gese, H.2
Plege, B.3
Claus, J.1
Ein neuer Prüfkörper für das
experimentelle Ranking von AlZylinderkopflegierungen bei
thermomechanischer Beanspruchung
Langer, S.4
Werkstoffwoche 2004, 21.–23.
September 2004, München
(1) DaimlerChrysler AG,
Stuttgart
(3) DaimlerChrysler AG,
Friedrichshafen
(4) Rautenbach Aluminium
Technologie GmbH,
Wernigerode
Aufgrund der steigenden Spitzentemperaturen in Hochleistungsdieselmotoren nimmt der Anteil der thermomechanischen Belastung (TMF) in Zylinderköpfen an der Gesamtschädigung zu. Deshalb muss die thermomechanische Ermüdung bei der Werkstoffauswahl neuer Bauteile berücksichtigt werden. Für ein Werkstoffranking bietet sich als Zwischenschritt zwischen der uniaxialen Prüfung von Rundproben (einachsiger, homogener Belastungszustand) und einem Bauteil-/ Motortest (sehr komplexe Belastungsverhältnisse) die Prüfung einer technologischen Probe
mit inhomogenem, mehrachsigen Spannungszustand an. Die so
genannte “Stegprobe” soll hier näher vorgestellt werden.
Die Stegprobe ist eine Gussplatte mit zwei kreisförmigen Durchbrüchen. Der Mittelsteg zwischen den Löchern wird von einer
Seite lokal induktiv geheizt. Die beiden äußeren Stege bleiben
kalt und behindern die Ausdehnung des Mittelsteges. Die Wärmezufuhr im Mittelsteg wird über periphere Kühlkanäle wieder
abgeführt. Durch die transiente Erwärmung wird die Dehnungsbehinderung an der Stegoberseite verschärft und es werden
Druckspannungen induziert. Durch die Geometrie der Probe
wird im kritischen Querschnitt ein mehrachsiger Spannungszustand erreicht. Bereiche mit unterschiedlichen Temperatur-ZeitVerläufen in der Stegprobe altern zudem unterschiedlich schnell,
was in einem konventionellen TMF-Versuch nicht berücksichtigt
wird. Durch die Wiederholung des Temperaturwechsels erfährt
die Probe eine thermische Ermüdung. Mit einem verschärften
Temperaturzyklus kann ein Anriss innerhalb von einem Tag
Prüfzeit erzielt werden.
Anhand von elastoplastischen FEM-Simulationen der Stegprobe
werden die Belastungsverhältnisse in der Stegprobe bei der thermomechanischen Belastung qualitativ beschrieben und die Belastungsgrößen für einen Werkstoff beispielhaft quantifiziert.
Für die Herstellung der Stegprobe wurde eine Kokille konstruiert, welche die Einstellung eines porenfreien Gefüges mit einem
kleinen Dendritenarmabstand im Mittelsteg ermöglicht. Zudem
lässt sich die Erstarrungsgeschwindigkeit in diesem Bereich
über eine Kühlung im Werkzeug gezielt beeinflussen.
Im Vortrag werden mit der Stegprobe ermittelte Lebensdauern
von verschiedenen Legierungssystemen und von Proben mit
unterschiedlichen Gießparametern gegenübergestellt.
Die vorgestellten Arbeiten wurden im Rahmen eines von BMBF
geförderten Verbundprojektes durchgeführt. Die Autoren danken dem BMBF für die Förderung der Arbeiten.
Gese, Helmut
Werner, Heinrich
Aufsatz in Konstruktion, Heft
7/8–2004, Springer Verlag,
S.4–6
Virtuelle Fahrzeugentwicklung
verbessern mit neuen
Materialmodellen: Crashsimulation im
Automobilbau – Modellierung
metallischer Werkstoffe
Zentrales Thema der Crashsimulation ist die Bewertung und
Verbesserung konstruktiver Maßnahmen zum Schutz der Verkehrsteilnehmer innerhalb und außerhalb des Fahrzeuges. Die
raum-, gewichts- und kosteneffiziente Umsetzung kinetischer
Energie in plastische Arbeit bei der weitgehend irreversiblen Deformation von Strukturkomponenten stellt den Kern dieser Aufgabe dar. Neben der konstruktiven Gestaltung einer Strukturkomponente hat der verwendete Werkstoff einen wesentlichen
Einfluss auf deren Verhalten bei einer Crashbelastung. Den mathematischen Modellen für die Beschreibung des spezifischen
Werkstoffverhaltens kommt daher eine besondere Bedeutung zu.
Werner, H.1
Hooputra, H.1
Gese, H.2
Dell, H.2
Heath, A.3
Pyttel, T.3
Keßler, L.4
Yelisseyev, V.5
CrashMAT2004, 27./28. April
2004, EMI, Freiburg
(1) BMW Group, München,
(4) Thyssen Krupp Stahl AG,
Dortmund
(4) TEST, Woronesch (Russland)
Ein umfassendes Materialmodell für
die Versagensbewertung in der
Prozesskette Tiefziehen-Crash
Die Zahl der in der Automobilindustrie eingesetzten Werkstoffe
ist in den letzten Jahren deutlich angestiegen. Als Beispiel ist
der Einsatz von Aluminium- und Magnesiumlegierungen zu nennen. Zur Gewichtsoptimierung werden zudem weiche Tiefziehstähle durch höher- und höchstfeste Stahlgüten substituiert. Im
Hinblick auf die virtuelle Fahrzeugentwicklung stellt dabei die
prozessübergreifende Auslegung von Komponenten unter Berücksichtigung ihres Fertigungsprozesses eine besondere Herausforderung dar.
Bei den oben genannten Werkstoffen ist die korrekte Prognose
des lokalen Werkstoffversagens neben einer adäquaten Beschreibung der Werkstoffplastizität von zentraler Bedeutung. Zwei
unterschiedliche Versagensarten können bei der statischen und
dynamischen Verformung von Metallen beobachtet werden. Der
duktile Trennbruch entsteht infolge von Porenbildung, -wachstum und -koaleszenz; der Scherbruch resultiert aus einer Scherbandlokalisation. Bei Verwendung von Schalenelementen zur
Abbildung von dünnwandigen Strukturen (z.B. Bleche und
Strangpressprofile) ist darüber hinaus die lokale Einschnürung,
welche sich aufgrund einer plastischen Instabilität unter Zugbelastung ausbildet, zu berücksichtigen. Mit Hilfe des Algorithmus
CrachFEM, der in dem kommerziellen FE-Programmsystem
PAM-Solid implementiert ist, erfolgt eine kontinuierliche Bewertung der genannten Versagensmoden während der zeitlichen Integration der Bewegungsgleichungen. Für Volumenelemente findet dabei nur das Trenn- und Scherbruchkriterium Berücksichtigung; für Schalenelemente alle drei der oben genannten Kriterien.
Zur Validierung des numerischen Modells wurde durch die
quasi-statische Intrusion eines Impaktors in einer überwiegend
aus höherfesten Stahlgüten gefertigten Rohkarosseriestruktur
ein Versagen provoziert. In der Simulation wurde dieser Vorgang
unter Berücksichtigung der Umformhistorie aus dem Fertigungsprozess modelliert. Als Beispiel für eine Versagensbewertung von Strukturen aus der Werkstoffgruppe Aluminium werden Komponententests von Extrusionsprofilen vorgestellt und
mit Simulationsergebnissen korreliert.
Gese, Helmut
Werkstofftag 2003 — Neue Materialien für die
Automobilindustrie, 4.
November 2003, Gürzenich,
Köln
Verbesserte Material- und
Versagensmodelle für die Umformund Crashsimulation
Die Zahl der in der Automobilindustrie eingesetzten Werkstoffe
ist in den letzten Jahren extrem angestiegen. Als Beispiel ist der
Einsatz von Aluminium- und Magnesiumlegierungen zu nennen.
Zur Gewichtsoptimierung werden zudem weiche Tiefziehstähle
durch höher- und höchstfeste Stahlgüten substituiert. Diese
Werkstoffe müssen in der Umform- und Crashsimulation korrekt
abgebildet werden. Neben der Beschreibung der Werkstoffplastizität spielt auch die korrekte Prognose des lokalen Werkstoffversagens bei diesen Werkstoffen eine zentrale Bedeutung. Im Vortrag werden verbesserte Plastizitätsmodelle und die notwendigen Schritte für eine umfassende Versagensbewertung behandelt.
Werner, H.1
Dell, H.2
Metzmacher, G.2
Kessler, L.3
Heath, A.4
17th, Mainz, Germany
Dr. Gese & Oberhofer
(3) Thyssen-Krupp Stahl AG,
Dortmund
Methodology, Validation and
Application of a Failure Model Based
on Transient Forming Limit Curves for
Coupled Stamping and Crash
Processes as Part of the IMPACT
Project
Predicting the failure of structural components in a crash event
is one of the major challenges of numerical simulations in the
automotive industry. Within the IMPACT project, supported by
the European Communities Framework Program V, BMW Group
and their partners focused on the development of a numerical
model to enhance the failure prediction. Selected high strength
sheet steel materials, typically used in the body in white, have
been investigated.
The presentation will give an overview of a failure model which
is based on the transient prediction of localized necking in sheets
taking into account the forming history of the body in white´s
components. To validate the model, two- step forming processes
have been carried out on a laboratory scale experimentally as
well as numerically. These allowed for nonlinear loading path
histories to assess the differences in forming limits as predicted
by conventional forming limit diagrams (derived from experiments with linear strain paths). Finally the numerical model was
applied to an industrially relevant test case. A body in white
from a series sedan was subjected to a quasistatic B-pillar intrusion up to failure. A comparison of experimental and numerical
results will show the capabilities of the transient failure model
and indicate future steps of development, i.e. enhancing crack
propagation simulation.
Hooputra, H.1
Gese, H.2
Dell, H.2
A New Comprehensive Failure Model
for Crashworthiness Simulation –
Validation for Aluminum Extrusions
Werner, H.1
Heath, A.3
17th, Mainz, Germany
(1) BMW Group, München,
A correct representation of the plastic deformation and failure of
individual component parts is essential to obtaining accurate
crashworthiness simulation results. The aim of this paper is to
present a comprehensive approach for predicting failure in a
component based on macroscopic strains and stresses. This approach requires the use of of three different fracture mechanism
representations, such as necking (due to local instabilities), as
well as ductile and shear fracture. The fracture criteria have
been developed in a way to include the influence of non-linear
strain paths. The fracture criteria have been included into the
software CrachFEM. CrachFEM has been coupled with PamCrash via a new material model 128. CrachFEM transiently predicts a possible failure of elements and causes PamCrash to eliminate these elements.
The effectiveness of this approach in predicting failure is then
discussed by comparing numerical results with test data in three
point bending and axial drop tests of double chamber extrusion
components. All studies presented in this paper were carried out
on extrusions made of aluminum alloy AA7108.50-T6.
Schmid, M.1
Gese, H.2
Metzmacher, G.2
Claus, J.3
3
VDI-Tagung “Zylinderlaufbahn,
Hochleistungskolben, Pleuel”,
München, 16./17. September
2003
(1) EADS CRC, Ottobrunn
(3) DaimlerChrysler Forschung,
Stuttgart
Bewertung der thermomechanischen
Ermüdung von Kolben aus AlGusslegierungen im
außermotorischen Experiment und
durch FEM-Simulation
In diesem Beitrag wird das Ermüdungsverhalten von Kolben bei
thermomechanischer Beanspruchung untersucht. Eine außermotorische Versuchsmethode mit realistischen Aufheiz- und Abkühlbedingungen zur schnellen Überprüfung der Lebensdauer
von Kolben bei isolierter thermomechanischer Belastung wird
vorgestellt. Um Informationen zur Höhe der mechanischen Dehnungen am Kolbenmuldenrand zu bekommen, wird eine faseroptische Dehnungs- und Temperaturmessmethode eingesetzt. Die
außermotorische Kolbenprüfung wird parallel mit FEM simuliert und die prognostizierten Lebensdauern mit dem Experiment verglichen. Längerfristig kann das numerische Modell zur
schnellen Vorauslegung von Kolben verwendet werden.
Schmid, M.1
Claus, J.2
Thermal fatigue component test of
engine parts
Gese, H.3
Löhe, D.4
4
International Congress Thermal
Stresses ICTS 2003, 8. – 11.
Juni 2003, Blacksburg Virgina
USA
(1) EADS Corporate Research
Center, München
(2) DaimlerChrysler AG Research & Technology, Stuttgart
(4) Universität Karlsruhe (TH),
Institut für Werkstoffkunde I
In this study, the effect of thermal fatigue on pistons and cylinder heads is investigated. A test method with realistic heating
and cooling conditions is presented. To get information about the
constraints at critical regions of the investigated engine components, fiber optic strain and temperature gauges are used. The
ratio of measured total strain and thermal strain gives information on proportion of thermal strain being obstructed. The thermal fatigue piston test is simulated with FE methods. The results of experimental and simulated life times are compared.
Oberhofer, G.1
Dell, H.1
Dell, D.1
Gese, H.1
Lanzerath, H.2
Wesemann, J.2
Hombergsmeier, E.3
Improved Plasticity and Failure
models for Extruded Mg-Profiles in
Crash Simulations
3
4th European LS-DYNA Users
Conference, May 22nd and 23rd,
2003, Ulm, Germany
Aachen
Center, München
The Crash Simulation of Magnesium Structures with Finite Element Methods demands the use of suitable material and failure
models. An associated plasticity model describing the complex
asymmetric yield behaviour in tension and compression of Mg extrusions was developed during the InMaK-project (Innovative
Magnesium Compound Structures for Automobile Frames) supported by the German Federal Ministry for Education and Research (BMBF). Differences to the material model 124 in LSDYNA are exposed. In order to describe the failure behaviour of
Mg extrusions under multi-axial loading in FEM crash simulation this constitutive model has been combined with a fracture
model for ductile and shear fracture. The fracture model has
been added to the user defined constitutive magnesium model in
LS-DYNA. The experimental investigations carried out on model
components are compared with numerical derived results. Experimental methods for fracture parameter evaluation are shown
and general aspects of metal failure due to fracture as well as different modelling techniques are discussed.
Lanzerath, H.1
Wesemann, J.1
Gese, H.2
Oberhofer, G.2
Dell, H.2
Hombergsmeier, E.3
Crash Simulation on Body Structural
Components Made out of Extruded
Magnesium
3
SAE Technical Paper Series
2003–01–259, March 3rd to 6th,
2003, Detroit
Aachen
Center, München
On the base of the Finite Element Method (FEM) a new constitutive model for extruded magnesium profiles, useful for crash relevant body components will be developed. This constitutive model
will be implemented into commercial software as a user material
law and can be used in the development of vehicle structures.
First step is the material characterization of the magnesium extruded materials. On this base the constitutive model will be developed and validated. In the last step simulations will be performed with the new simulation tool and the results will be compared to test results.
Gese, H.1
Dell, H.1
Keller, S.2
Yeliseyev, V.3
DVM-Tagung
“Werkstoffprüfung 2002 Kennwertermittlung für die
Praxis”, 5. und 6. Dezember
2002, Bad Nauheim
(2) VAW aluminium AG – A
Hydro Aluminium Company,
Forschung und Entwicklung,
Bonn
(3) Software Center TEST,
Woronesch (Russland)
Ermittlung von
Fließwiderstandskurven bei großen
Formänderungen für die
Blechumformsimulation
Bei Blechumformoperationen (Tiefziehen, Innenhochdruckumformen u.a.) können lokal sehr große plastische Vergleichsdehnungen auftreten. Um diese Prozesse mittels FEM-Simulation
genau abbilden zu können, müssen Fließwiderstandskurven bis
zu den maximal auftretenden Vergleichsformänderungen vorgegeben werden. Typischerweise wird hierzu die Fließwiderstandskurve aus dem uniaxialen Zugversuch in Walzrichtung mit geeigneten Verfestigungsmodellen approximiert und für hohe
Formänderungen extrapoliert. Diese Vorgehensweise kann aber
mit großen Fehlern behaftet sein. Am Beispiel einer Al-Legierung vom Typ AA5xxx wird gezeigt, dass bekannte Verfestigungsmodelle (Swift, Voce, Hockett-Sherby) zu stark abweichenden Extrapolationen der Fließwiderstandskurve führen. Es ist
aber a priori nicht bekannt, welches der Verfestigungsmodelle
für den jeweiligen Werkstoff gültig ist. Weiterhin können einige
Verfestigungsmodelle im Bereich der Extrapolation im Sinne einer physikalischen Verfestigungsbeschreibung instabil werden.
Ein Ausweg ist hier die direkte Messung der Verfestigung bei hohen Formänderungen. Hierzu werden drei Prüfmethoden beschrieben und diskutiert:
Torsionsversuch in der Blechebene; hydraulischer Tiefungsversuch; Schichtstauchversuch.
Für den Torsionsversuch in der Blechebene wurde der in [1] beschriebene Versuch weiterentwickelt. Durch Einbringen einer
konzentrischen Nut zwischen innerem und äußerem Spannring
wurde die plastische Verformung auf eine Zone mit definierter
Breite reduziert. Der Versuch liefert eine über alle Orientierungen gemittelte Schubfließwiderstandskurve. Die Ableitung
einer Fließwiderstandskurve für den uniaxialen Zugversuch
unter Verwendung eines orthotropen Fließortes wird diskutiert.
Die Versuchstechnik bei der Messung der Fließwiderstandskurve bei äquibiaxialem Zug mit der hydraulischen Tiefung wird
nachfolgend beschrieben. Die Ableitung einer Fließwiderstandskurve für den uniaxialen Zugversuch erfordert ebenfalls die Verwendung eines orthotropen Fließortes.
Mit dem Schichtstauchversuch (mehrere Lagen von Blech werden in Blechdickenrichtung uniaxial gestaucht) kann ebenfalls
der Fließwiderstand bei äquibiaxialem Zug aufgenommen werden. In den Plastizitätsmodellen von Metallen wird unterstellt,
dass die Überlagerung eines hydrostatischen Druckes die Fließspannung nicht beeinflusst wird. Deswegen ist der uniaxiale
Druckversuch in Blechdickenrichtung im plastischen Bereich
gleichwertig einem äquibiaxialen Zug in der Blechebene. Die Ableitung einer Fließwiderstandskurve für den uniaxialen Zugversuch aus der Fließwiderstandskurve des Stauchversuchs erfordert ebenfalls die Verwendung eines orthotropen Fließortes.
Die Fließwiderstandskurven aller 3 Versuchsarten werden hinsichtlich ihrem Verfestigungsverhalten und der Reproduzierbarkeit verglichen. Die experimentell bei hohen Formänderungen
gemessenen Fließwiderstandskurven werden mit extrapolierten
Fließwiderstandskurven aus dem uniaxialen Zugversuch verglichen.
Werner, H.1
Gese, H.2
2
DVM-Tagung
“Werkstoffprüfung 2002 —
Kennwertermittlung für die
Praxis”, 5. und 6. Dezember
2002, Bad Nauheim
Zur Bedeutung dehnratenabhängiger
Werkstoffkennwerte in der
Crashsimulation
Zentrales Thema der Crashsimulation ist die Bewertung und
Verbesserung konstruktiver Maßnahmen zum Schutz der Verkehrsteilnehmer innerhalb und außerhalb des Fahrzeuges. Die
raum-, gewichts- und kosteneffiziente Umsetzung kinetischer
Energie in plastische Arbeit bei der weitgehend irreversiblen Deformation von Strukturkomponenten stellt den Kern dieser Aufgabe dar. Im Grenzfall ist darin auch die bei Bruchvorgängen für
den Rissfortschritt aufgezehrte Energie enthalten, deren Vorhersage zur Zeit noch nicht mit ausreichender Genauigkeit möglich
ist.
Da die charakteristischen Zeiten für Strukturdeformationen bei
Crashvorgängen im Bereich von Millisekunden liegen, ist die
Kenntnis des Werkstoffverhaltens bei höheren Deformationsgeschwindigkeiten für die Simulation bedeutsam. Der Einflussfaktor Deformationsgeschwindigkeit steht im Mittelpunkt des Vortrages.
Die Genauigkeit einer Crashsimulation wird durch eine Vielzahl
weiterer Parameter beeinflusst, welche aber im Zeitrahmen des
Vortrages nicht diskutiert werden können.
Ausgehend von Crash-Lastfällen für die Strukturauslegung werden typische Dehngeschwindigkeiten in Fahrzeugkomponenten
aus metallischen Werkstoffen ermittelt, welche die Anforderungen an die experimentelle Werkstoffprüfung definieren. Die
Dehnratenempfindlichkeit von unterschiedlichen Stahlsorten
und Al-Legierungen wird im relevanten Dehnratenintervall diskutiert. Die Aufbereitung von Fließwiderstandskurven aus den
Hochgeschwindigkeitsprüfungen muss unter Berücksichtigung
der physikalischen Vorgänge im Werkstoff und des verwendeten
Simulationsmodells erfolgen. Für industrielle Crashsimulationen mit explizit-dynamischen FEM-Codes werden adiabate
Fließwiderstandskurven benötigt. Neben der Abbildung der
Werkstoffplastizität ist die Prognose des Werkstoffversagens eine weitere Aufgabe der Crashsimulation. Der wesentliche Versagensmechanismus bei duktilen Blechwerkstoffen ist die Zuginstabilität. Bei linearen Dehnungspfaden entspricht dieses Versagen dem Grenzformänderungsschaubild. Ist das Verfestigungsverhalten eines Werkstoffes signifikant von der Dehngeschwindigkeit beeinflusst, so sind gegenüber dem quasistatischen Fall
im Grenzformänderungsdiagramm verschobene Instabilitätsgrenzen zu erwarten. Dies wird anhand eines Beispiels erläutert.
Im letzten Abschnitt des Vortrages werden anhand von Simulationsergebnissen unter Berücksichtigung bzw. Vernachlässigung
der Dehnratenabhängigkeit die Einflüsse quantitativ diskutiert.
Oberhofer, G.
Dell, H.
Dell, D.
Gese, H.
Proceeding of 7th International
LS-Dyna Users Conference
2002, May 19th to 21st, 2002,
Dearborn, Michigan, U.S.A.
LSCT ETA, 2002 LS-Dyna
Users Conference, September
19th and 20th, 2002, Bad
Mergentheim, Germany
Enhanced Failure Prediction in Sheet
Metal Forming Simulations through
Coupling of LS-Dyna and Algorithm
CRACH
In sheet-metal-forming the forming limit curve (FLC) is used for
ductile sheets to predict fracture in deep drawing. However, the
use of the FLC is limited to linear strain paths. The initial FLC
cannot be used in a complex nonlinear strain history of stamping
and crash including a significant change in strain rate. The
CRACH software has been developed to predict the forming limit
of sheets for nonlinear strain paths. It has been validated to predict instability for bilinear strain paths with static loading in the
first path and dynamic loading in the second path for mild steels.
As the postprocessing of single strain paths from LS-Dyna in
CRACH is not economic for industrial applications MATFEM initiated a project to couple CRACH directly with LS-Dyna using a
user-defined material model. This allows a prediction of possible
failure during the simulation for all elements in respect of their
complete strain history. A special strategy has been developed to
include CRACH without extensive increase in total CPU time.
The developed interface to LS-Dyna allows also the implementation of other failure criteria demanding the history of deformation like for example a tensorial fracture criterion. In this case
the problem is similar but the calculation time steps for failure
are different.
In order to test the reliability of the calculated safety factor experimental tests for bilinear strain paths as described in [2] have
been simulated. In this case the experimental and numerical investigations have been made on two-stage forming processes. In
reality the first of these two stages reflects a stamping process,
followed by a second stage stamping process. This second stage
could also be a crash process. Further industrial applications are
discussed.
The coupling of LS-Dyna and CRACH has the potential to predict
possible fracture in deep drawing and crash loading at an early
design stage and allows to optimise geometry and material quality to significantly reduce later problems in real components.
Gese, H.1
1
Dell, H.
2
Heath, A.2
Payen, F.3
Continuous Failure Prediction for
Steel Sheets in Successive Stamping
and Crash Simulations through
Coupling of PAM-Solid and CRACH
3
Werner, H.
EuroPAM ’01 (CD-ROM),
October 18th and 19th, 2001,
Heidelberg, Germany
(2) ESI Group, Eschborn
The mapping procedure has been introduced into PAM-SOLID
recently to account for the strain hardening from the deep drawing process in the crash simulation. This technique allows to increase the accuracy in predicting energy absorption during crash.
But there is still a lack for a common failure model in successive
stamping and crash simulations. The forming limit curve (FLC)
is used for ductile sheets to predict fracture in deep drawing.
However the use of the FLC is limited to linear strain paths. The
initial FLC cannot be used in a complex nonlinear strain history
of stamping and crash including a significant change in strain
rate. The CRACH software has been developed to predict the
forming limit of sheets for nonlinear strain paths. It has been validated to predict instability for bi-linear strain paths with static
loading in the first path and dynamic loading in the second path
for mild steels. As the postprocessing of single strain paths from
FEM in CRACH is not economic for industrial applications BMW
initiated a project to couple CRACH directly with PAM-Solid.
This allows a prediction of possible failure during the crash simulation for all elements. A special strategy has been developed
to include CRACH without extensive increase in total CPU time.
The safety factor against fracture (defined by onset of instability)
is calculated with a two-level technique. On the first level a
rough and fast calculation is used. If the safety factors is lower
than a given value a precise calculation with CRACH is started.
Real components of a body in white have been simulated to validate the PAM-Solid/CRACH coupling. The deep drawing simulation of the different components of a longitudinal (engine mount)
have been performed. The results have been mapped to the corresponding component crash model. The component testing has
been done parallel to validate the results. An accurate prediction
of cracks in deep drawing and crash have been found.
The coupling of PAM-SOLID and CRACH will be an effective tool
in the design of car components and can support the choice of appropriate steel qualities. Additionally there is a high potential
for application in forming operations with highly nonlinear
strain paths and multi-stage operations. The link between PAMSolid and CRACH has been developed by ESI GmbH and
MATFEM up to an “experimental” level on behalf of BMW. For
an industrial application the interface still has to be “streamlined”.
Hooputra, H.1
Werner, H.1
Metzmacher, G.2
Fracture Criteria For Crash
Simulation of Wrought Aluminium
Alloy Components
2
EuroPAM ’01, October 18th and
19th, 2001, Heidelberg,
Germany
The prediction of plastic deformation and fracture of components
is critical to correctly represent the transient energy absorption
through the car structure in crashworthiness simulation. Thus, a
validated material fracture model which adequately represent
this phenomenon is indispensable. The aim of this paper is to
compare and validate two existing numerical approaches to predict failure with test data by using the aluminium wrought alloy
AlMgSi1-F31.
(2) ESI Group, Eschborn
The first approach is based on failure curves expressed by instantaneous macroscopic stresses and strains (i.e. maximum
equivalent plastic strain vs. stress triaxiality). Notched tensile
specimens with varying notched radii and axisymmetric shear
specimens were used to produce ductile fractures and shear band
fractures at different stress states.
The second approach is based on the modified Gurson model and
uses state variables at the mesoscopic scale (i.e. critical void volume fraction). The critical macroscopic and mesoscopic damage
values at the fracture initiation locations were evaluated using
FEM simulations of the different specimens.
The derived macroscopic and mesoscopic fracture criteria were
then subsequently applied to crashworthiness experiments with
real components. The quality of the prediction on component
level is discussed for both types of criteria.
El-Magd, E.1
2
Gese, H.
3
Tham, R.4
Hooputra, H.4
Fracture Criteria for Automobile
Wrought Aluminium Alloy
Components
4
Werner, H.
Materialwissenschaften u.
Werkstofftechnik 32, 712–714
(2001), WILEY-VCH Verlag
GmbH, Weilheim
(1) RWTH Aachen
(3) Frauenhofer Institut für
Kurzzeitdynamik, Freiburg
In automobile crashworthiness simulation, the prediction of plastic deformation and fracture of each significant, single component is critical to correctly represent the transient energy absorption through the car structure. There is currently a need, in
the commercial FEM community, for validated material fracture
models which adequately represent this phenomenon. The aim of
this paper is to compare and to validate existing numerical approaches to predict failure with test data.
All studies presented in this paper were carried out on aluminium wrought alloys: AlMgSi1.F31 and AlMgSiCu-T6. A viscoplastic material law, whose parameters are derived from uniaxial
tensile and compression tests at various strain rates, is developed and presented herein. Fundamental ductile fracture mechanisms such as void nucleation, void growth, and void coalescence as well as shear band fracture are present in the tested
samples and taken into consideration in the development of the
fracture model.
Two approaches to the prediction of fracture initiation are compared. The first is based on failure curves expressed by instantaneous macroscopic stresses and strains (i.e. maximum equivalent
plastic strain vs. stress triaxiality). The second approach is based
on the modified Gurson model and uses state variables at the
mesoscopic scale (i.e. critical void volume fraction). Notched tensile specimens with varying notch radii and axisymmetric shear
specimens were used to produce ductile fractures and shear band
fractures at different stress states. The critical macroscopic and
mesoscopic damage values at the fracture initiation locations
were evaluated using FEM simulations of the different specimens. The derived fracture criteria (macroscopic and mesoscopic)
were applied to crashworthiness experiments with real components. The quality of the prediction on component level is discussed for both types of criteria.
Dell, H.1
Gese, H.1
Keßler, L.2
Werner, H.3
Continuous failure Prediction Model
for Nonlinear Load Paths in
Successive Stamping and Crash
Processes
Hooputra, H.3
SAE Technical Paper Series
2001–01–1131
(2) Thyssen Krupp Stahl AG,
Dortmund
The validity of numerical simulations is still limited by the unknown failure of materials when nonlinear load paths in successive stamping and crash processes occur. Localized necking is the
main mechanism for fractures in ductile sheet metal. The classical forming limit curve (FLC) is limited to linear strain paths. To
include the effects of nonlinear strain paths a theoretical model
for instability (algorithm CRACH) has been used. The algorithm
has been developed on the basis of the Marciniak model [8]. The
calibration and validation of this approach is done by a set of
multistage experiments under static and dynamic strain rates
for a mild steel.
Gese, H.
Dell, H.
Y. Bréchet (Ed.),
Microstructures, Mechanical
Properties and Processes –
Computer Simulation and
Modelling, EUROMAT 99 –
Volume 3, Wiley-VCH,
Weinheim, 2000
Macroscopic and Mesoscopic Models
for Fracture in Sheet Metal Forming
Numerical simulation of sheet metal forming is already a standard development tool in industry. However there are still limitations in adequate criteria for sheet fracture. A complete description of forming failures must include local necking of sheet
(followed by fracture in neck), sudden fracture (i.e. at small radii)
and shear fracture. In practice only the conventional forming
limit curve (FLC) is used for interpretation of simulation results.
The FLC is a criterion for local necking and its correct application is limited to linear strain paths (this limitation is mostly
violated in real industrial drawing operations). The experimental
effort to build up a FLD is very high. From this view one can
derive that there is a need for more refined criteria. The presentation gives an overview on refined macroscopic and mesoscopic
criteria for failure in sheet metal forming. Mesoscopic criteria
are based on a phenomenological model for void nucleation, void
growth and void coalescence. They are based on the plasticity
model defined by Gurson. The practical problem is the experimental evaluation of the material parameters. This type of criterion is limited to the prediction of ductile material fracture. Another approach is the use of refined macroscopic criteria. The authors present a newly developed algorithm which checks the possibility of local necking or fracture based on the local strain path
from the forming simulation. The algorithm uses a refined plasticity model which includes an anisotropic yield locus and an anisotropic hardening rule including the Bauschinger effect. The
algorithm checks whether there is a stable elastoplastic solution
for the given strain path. Including the Bauschinger effect it is
possible to study strain paths with sudden changes in drawing
direction. Practical examples are given to support the theoretical
information on mesoscopic and macroscopic failure criteria.
Gese, H.1
Juhl, K.2
Plastizitätsmodelle für die Simulation
der Blechumformung
Lang, R.3
Reese, E.D4
4
Kopp, B.; Beiss, P.; Herfurth, K.;
Böhme, D.; Bormann, R.; Arzt,
E. und Riedel, H. (Hrsg.):
Tagungsband der
Werkstoffwoche ’98, Band VI,
WerkstoffInformationsgesellschaft,
Frankfurt, 1999
(1) MATFEM Ingenieurbüro Dr.
Gese, München
(2) Daimler-Benz Aerospace
Airbus GmbH, Bremen
(3) Daimler-Benz Aerospace AG,
Ottobrunn
(4) Daimler-Benz AG,
Ottobrunn
In der Luftfahrtindustrie werden Strukturbauteile aus Al- und
Ti- Blechhalbzeugen durch Kaltumformung hergestellt. Zur
schnelleren Überprüfung der Umformbarkeit neuer Geometrien,
zur Bewertung der Eignung neuer Legierungsqualitäten und zur
Ermittlung von Prozessfenstern für Umformvorgänge wird zunehmend die Simulation der Blechumformung auf der Basis der
Methode der finiten Elemente eingesetzt. Die korrekte Beschreibung der Werkstoffplastizität ist neben einer geeigneten Diskretisierung der Geometrie die wesentliche Voraussetzung für eine
genaue Ermittlung des Spannungs- und Deformationszustandes
während des Umformprozesses und somit wesentlich für die Bewertung der Umformsicherheit und für die Quantifizierung der
elastischen Rückfederung.
Die Erfahrungen zeigen, dass isotrope, transversal-isotrope und
konventionelle orthotrope Fließortbeschreibungen nicht für die
Beschreibung von Al- und Ti-Werkstoffen ausreichend sind. Für
verbesserte Aussagen mussten weiterentwickelte Materialmodelle eingeführt und verifiziert werden. Das plastische Fließverhalten von Al-Blechhalbzeugen kann nur durch eine nicht- quadratische Fließortformulierung angenähert werden. Bei den TiWerkstoffen müssen Fließorte Verwendung finden, welche eine
Asymmetrie im Zug- und Druckbereich zulassen. Die experimentellen Arbeiten zur Ermittlung der notwendigen Plastizitätskennwerte und die mathematische Beschreibung höher entwickelter Fließorte werden diskutiert. Die Werkstoffplastizität bestimmt zudem die Umformgrenzen eines Werkstoffes hinsichtlich Instabilität und Bruch. Die angewendeten Kriterien werden
diskutiert.
Im Ausblick werden Simulationsergebnisse des Streckziehens
von Al-Blechen und des Tiefziehens von Ti-Blechen mit weiterentwickelten Materialmodellen vorgestellt und deren Genauigkeit in Relation zu Simulationen mit konventionellen Materialbeschreibungen und zum realen Fertigungsprozess diskutiert.
Gese, H.
“Impulse für kleine und
mittelständische Unternehmen
durch Neue Materialien”,
Vortragsveranstaltung der
Interessengemeinschaft Neue
Materialen in NRW e.V. am 23.
September ’98 im
Technologiepark Bergisch
Gladbach
MATFEM Ingenieurbüro Dr.
Gese, München
Schnelle Bewertung der industriellen
Anwendbarkeit Neuer Materialien
durch Simulation
Neue Materialien bieten parallel zur konstruktiven Optimierung
ein großes Potential zur Verbesserung technischer Bauteile. Die
Einführung neuer Werkstoffe ist aber mit einem großen Entwicklungs- und Testaufwand verbunden. Unerwartete Schwierigkeiten in der Verarbeitung des Werkstoffes und Probleme mit
der Betriebsfestigkeit erster Prototypen können für kleine Firmen zu einem unkalkulierbaren finanziellen Risiko werden.
Eine Methode, um das Entwicklungsrisiko bei Neuen Materialien frühzeitig bewerten zu können, ist die virtuelle Entwicklung
von Bauteilen. Bisher wurde diese Methodik der virtuellen Entwicklung nur im konstruktiven Bereich mittels 3D-CAD umgesetzt (Visualisierung, Rapid Prototyping, Digital Mock-Up). Die
virtuelle Entwicklungsschiene ist aber nur dann für die Bewertung neuer Werkstoffe sinnvoll, wenn auch der Fertigungsprozess des Bauteils und die Betriebsbelastungen des Bauteils im
Simulationsmodell vorab überprüft werden können. Hierzu müssen Berechnungen mit der Methode der Finiten Elemente (FEM)
und Verwendung von Kennwerten des neuen Materials durchgeführt werden. Die verfügbare Rechenleistung hat sich in den
letzten Jahren extrem gesteigert und stellt heute keine wesentliche Beschränkung für eine Anwendung der FEM im industriellen Maßstab mehr dar. Die korrekte Werkstoffbeschreibung in
den verwendeten Simulationsmodellen ist die zweite Voraussetzung für gute Prognosen. Im Bereich der metallischen Werkstoffe wurde zwischenzeitlich ein sehr guter Stand erreicht. Materialmodelle für Kunststoffe sind allerdings noch unzureichend.
An einigen Beispielen aus der Simulation von Fertigungsverfahren (Schmieden, Blechumformen) und der Simulation von Betriebsbelastungen (Crash, Ermüdung) werden die Möglichkeiten
der Simulation aufgezeigt. Die dazu eingesetzten Materialmodelle und Berechnungsmethoden werden hinsichtlich ihrer Genauigkeit diskutiert. Die für die Beschreibung der Werkstoffe notwendigen Materialparameter werden beschrieben.
Die Simulationsrechnung erlaubt heute somit, die an Laborproben gemessenen Werkstoffeigenschaften in die geplante Bauteilgeometrie zu “projizieren”. Somit können neue Materialien
schnell bewertet und die Zahl der notwendigen Prototypen und
Fertigungsversuche reduziert werden.
Gese, H.
Werkstoff-Forum auf der
Hannover-Messe 1998,
Vortragsveranstaltungen im
Innovationszentrum “IngenieurWerkstoffe”
Gese, München
Überblick zur
Werkstoffcharakterisierung in der
Simulationsrechnung
Zur Verkürzung von Entwicklungszeiten wird heute die virtuelle
Entwicklung von technischen Bauteilen gefordert. Bisher wurde
dies nur im Bereich der Geometrieverarbeitung mittels 3D-CAD
zufriedenstellend erreicht. Die virtuelle Entwicklungsschiene ist
aber nur dann sinnvoll, wenn auch der Fertigungsprozess des
Bauteils und die Betriebsbelastungen des Bauteils im Simulationsmodell vorab überprüft werden können. Hierzu müssen Berechnungen mit der Methode der finiten Elemente (FEM) durchgeführt werden. Die verfügbare Rechenleistung hat sich in den
letzten Jahren extrem gesteigert und stellt heute keine wesentliche Beschränkung mehr dar. Die Werkstoffbeschreibung in den
verwendeten Simulationsmodellen ist aber noch unzureichend.
An einigen Beispielen aus der Simulation von Fertigungsverfahren (Schmieden, Blechumformen) und der Simulation von Betriebsbelastungen (Crash, Ermüdung) werden die Schwächen
von bisher verwendeten Materialmodellen beschrieben und
Wege zu einer verbesserten Materialbeschreibung aufgezeigt.
Gese, H.
Workshop
Endoprothesenwechsel
Kniegelenk, 12.–14. April 1998,
Klinik für Orthopädie und
Sportorthopädie an der
Technischen Universität
München
Gese, München
Metalle als Werkstoff der
Revisionsendoprothetik
Es wird ein Überblick über die in der Endoprothetik eingesetzten
Metall-Legierungen (Fe-Basis, Ti-Basis und Co-Basis-Legierungen) gegeben. Die Mikrostruktur der Legierungen und deren
Verarbeitung (Gießen, Schmieden, Kaltumformen) entscheiden
über die Gebrauchseigenschaften der Endoprothesen. Die elastischen Eigenschaften haben dabei eine wesentlichen Einfluss auf
die Belastung des umliegenden Knochens und dessen lastadaptive Umbaureaktionen. Dies wird am Beispiel von Simulationsrechnungen aufgezeigt. Steife Implantate mit einem hohen Elastizitätsmodul (z.B. aus CoCrMo-Legierungen) reduzieren die
Dehnungen im umliegenden Knochen und können zu einer Inaktivitätsosteoporose führen. Steife Implantate reduzieren aber auf
der anderen Seite bei zementfreien Implantaten die post op auftretenden Relativverschiebungen zwischen Endoprothese und
Knochen, welche zur Ausbildung einer mechanisch instabilen
Bindegewebeschicht zum Implantat führen können. Neben der
mechanischen Reaktion auf den Knochen ist für den dauerhaften
Erfolg des Implantates die Ermüdungsfestigkeit der Legierungen von großer Bedeutung. Die Dauerschwingfestigkeit der Implantatlegierungen in korrosivem Körpermilieu unterscheidet
sich zum Teil deutlich und muss in der konstruktiven Auslegung
berücksichtigt werden. CoCrMo-Gusslegierungen bieten beispielsweise für intrameduläre Verankerungsstifte mit kleinem
Durchmesser oft eine zu geringe Dauerschwingfestigkeit. Insbesondere bei zementfreien Implantaten mit einer Oberflächenstrukturierung ist der Kerbeinfluss auf die Dauerschwingfestigkeit zu beachten. Ti-Basis-Schmiedelegierungen, welche grundsätzlich eine sehr hohe Dauerschwingfestigkeit besitzen, zeigen
eine starken Abfall der Dauerschwingfestigkeit bei Kerbwirkung. Am Beispiel der kondylären Komponente einer Knieendoprothese wird der Einsatz der Simulationsrechnung (FEM) für
die dauerfeste Auslegung von Endoprothesen beschrieben.
Gese, H.1
Stretch Forming of Aluminium Sheets
Beck, W.2
Reese, E.D.3
PAM ’97, October 16th to 17th
1997, Prague, Czech Republic
(1) MATFEM Ingenieurbüro
Gese, München
(2) Daimler-Benz Aerospace
Airbus GmbH, Dept. EVM,
Bremen
(3) Daimler-Benz AG,
Ottobrunn, Germany
Daimler-Benz Aerospace Airbus GmbH produces large aircraft
skin panels with double curved geometry from Al-alloys on a
stretch forming press. Up to now, manufacturing engineers have
used empirical methods in order to determine optimum stretching conditions. The aim is to avoid failure of the sheet material
during forming and to reduce springback to an acceptable level.
Numerical methods are introduced now to speed up the development of the manufacturing process for new geometries and alloys. The Daimler-Benz Forschungsinstitut has developed the
basis for a “virtual” stretch forming press on behalf of DaimlerBenz Aerospace Airbus GmbH. MATFEM has been subcontracted to support the numerical techniques.
The finite element code PamStamp has been introduced as a core
tool. The explicit solver is used for the simulation of the stretch
forming process and the implicit PCG-solver for the calculation of
the elastic springback. Successful manufacturing of a given skin
panel can be achieved by variation of the alloy heat treatment
and variation of the press kinematics. Therefore a refined plasticity model for aluminium alloys had to be introduced which is
sensible to small changes in material plasticity. The new material model 109 of PamStamp has been successfully tested in the
project. Additionally, an individual software tool has been developed which can generate the optimium stretch kinematics for a
given tool geometry and is able to convert press kinematics into
velocity boundary conditions for the clamped nodes of the sheet
mesh.
The presentation includes benchmark problems which show the
accuracy of the new material model 109 and the new implicit
solver for elastic springback. Simulation results will be given for
the stretch forming of different skin panels. The examples cover
parts with 1- and 2-step forming, trimmimg and springback. The
sensibility of the virtual process to changes in material properties is shown.
Finally, some still unsolved software features are addressed and
ideas will be given for future developments of PAM-Stamp.
Oberhofer, G
Gese, H.
“Die Methoden der Finiten Elemente in der Biomedizin und
angrenzenden Wissenschaften”,
Workshop 1996, 22. und 23. Juli
1996 an der Universität Ulm,
Universitätsverlag Ulm GmbH,
ISBN 3–89559–232–3
MATFEM Ingenieurbüro Gese,
München
Optimierung des Designs von
Knieendoprothesen zur Reduzierung
der Materialbelastung
Die Polyethylenkomponenten von Knieendoprothesen werden
bei der Gelenkartikulation hohen Kontaktspannungen ausgesetzt. Alternative Materialien mit entsprechender in-vivo-Testerfahrung existieren bislang nicht. Mit Hilfe der FE-Methode
sollte zunächst eine Strukturanalyse des tibiaseitigen HDPEKörpers mit nachfolgender Designoptimierung zur Reduzierung
der auftretenden Spannungen durchgeführt werden. Die Optimierungsmöglichkeiten beschränkten sich dabei auf die Designparameter »Krümmungsradien der kontaktierenden Oberflächen« und »Polyethylendicke«. Weitere geometrische Veränderungen waren aufgrund der vorgegebenen Kinematik des Kniegelenks nicht möglich. Die Analyse des femoro-tibialen Kontaktes ergab für eine vorgegebene Gesamtbelastung von 2400 N bereits Spannungen in der Höhe der Fließgrenze von Polyethylen.
Eine Optimierung des medialen Krümmungsradius der Tibiakomponente in medial-lateraler Richtung erbrachte eine Reduzierung des Spannungsniveaus. Nachfolgende Analysen mit optimierter Geometrie verdeutlichten eine Spannungszunahme bei
zunehmenden Flexions-und Rotationswinkeln. Mit zunehmender
Flexionsstellung kamen dabei kleinere Krümmungsradien der
Femurkomponente in den Kontaktbereich und es ergaben sich
bei gleichen äußeren Belastungen ansteigende Kontaktspannungen. Auch der Einfluss der Polyethylendicke auf die Höhe der
auftretenden Spannungen im Tibiaplateau konnte nachgewiesen
werden. Das Maximum der von-Mises-Vergleichsspannung lag
dabei erwartungsgemäß im Inneren des Polyethylens.
Die Patellakomponenten erfuhren ihre maximale Belastung bei
einer Flexionsstellung von 90°. Die deutlich günstigeren Belastungs-verhältnisse ergaben sich bei Variante WALLABY I aufgrund der kongruenten Krümmungs-radienpaarung in der Ebene senkrecht zur Flexionsachse. Dem natürlichen Verhalten entsprechend verschieben sich bei diesem Konstruktionsprinzip die
kontaktierenden Bereiche mit zunehmender Flexionsstellung
von der Mittelebene in die medial-lateralen Randbereiche der
Komponente. Analog zu den Ergebnissen beim femoro-tibialen
Kontakt muß festgestellt werden, dass auch bei guter Wahl der
Kontaktradienpaarung , wie sie bei beiden Implantattypen in
großen Flexionsbereichen vorliegt, ein Überschreiten der Fließgrenze im HDPE für das Patellaimplantat nicht vermieden werden kann. Die Wahl des Lastniveaus von 2400 N, einem experimentell ermittelten Maximalwert, erscheint jedoch als Grundlage für alle Flexionsstellungen zu hoch. Bei Flexionsstellungen
zwischen 30° und 90° kann dieser Wert nur bei dynamischer Belastung erreicht werden.
In bestimmten Belastungsfällen z. B. dem linienförmigen Kontakt zwischen Patella und Femurkomponente bei WALLABY II
erlaubt die Auflösung des FE-Netzes keine genaue Aussage über
den Absolutwert des Spannungsmaximums. Eine Netzverfeinerung in diesem Bereich wäre für detailliertere Aussagen notwendig. Weitere Modelltechnische Verbesserungen könnten in einer
dynamischen FEM-Analyse unter Berücksichtigung des viskoelastischen Materialverhaltens von Polyethylen liegen. Aufgrund
des stabileren Kontaktverhaltens expliziter FE-Codes ist eine
feinere Diskretisierung bei akzeptablen Rechenzeiten möglich.
In diesem Zusammenhang wäre dann auch eine Analyse der dynamischen Gleiteffekte wie sie bei einer Beugung bzw. Streckung des Kniegelenks neben der Normal-Belastung auftreten
möglich.
Gese, H.
Oberhofer, G.
13. CAD-FEM Users’ Meeting,
25.–27. Oktober 1995, Bad
Wildungen, Vortrag I–6, 6
Seiten, CADFEM GmbH, 1995
MATFEM Ingenieurbüro Gese,
München
Nichtlineare Modellierung von
Kontaktflächen bei Implantaten
Der Vortrag zeigt anhand von drei Anwendungsbeispielen aus
der Entwicklung von neuen Implantaten in den Bereichen Dentaltechnik und Endoprothetik die Notwendigkeit der nichtlinearen Modellierung von Grenzflächen zwischen benachbarten Materialien auf. Die mit ANSYS bearbeiteten Problemstellungen beinhalten den point-to-point-Kontakt mit dem Element Contac52
bei 2- und 3-dimensionalen Modellen und den allgemeinen 3-dimensionalen point-to-surface-Kontakt mit dem Element
Contac49.

MATFEM Publications and Speeches

Transcrição

Documentos relacionados

Dena Yago You and You`re People When Franz Biberkopf

In Englisch verhandeln, ohne dass es zum Fiasko wird…

Wuensche der Mediatoren an das Christkind

Erfahrungsbericht Auslandsstudium

pdf

User Guide

Heft 03/04 2009

Cost Structure Surveys for Germany

přestavby budov zedník jazykové verze

2009 - 12. - 10. >> Floor1