Haufe GISSMO

GISSMO Material Modeling with
a sophisticated Failure Criteria

Andr Haufe, Paul DuBois, Frieder Neukamm, Markus Feucht
LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe
Dynamore GmbH
Industriestrae 2
70565 Stuttgart
http://www.dynamore.de
Motivation
Technological challenges in the automotive industry

Weight
Safety requirements
New materials
Composites
High strength steel
Light alloys
Polymers
New power train
technology
Cost effectiveness
Design to the point
Technological challenges in the automotive industry

Weight
Safety requirements
New materials
Composites
High strength steel
Light alloys
Polymers
Damage
New power train
technology
Cost effectiveness
max
Anisotropy
c
Design to the point
b
a
Fracture growth
Failure
E
Plasticity
fail true
Debonding
E ( e )
Closing the process chain: Standard materials / state of the art

Forming simulation
Crash simulation
Hill based models
v. Mises or Gurson model
Anisotropiy of yield surface
Strain rate dependency
Kinematic/Isotropic hardening
Isotropic hardening
State of the art: Failure by FLD

(post-processing)
Damage evolution
NEW: Computation of damage

(GISSMO)
Failure models
(mapping of damage variable)
Mapping
II
II
II
I
I
III
III
III
Preliminary considerations
for plane stress
Some thoughts on damage formulations

Is path dependence critical for the target application (crashworthiness)?
Is a phenomenological model good enough?
Should we take orthotropic effects into account (i.e. tensorial damage)
or is a scalar formulation fine?
Shall we accumulate the damage value based on the stress state
or on the strain state?
Shall the damage accumulation be isotropic, orthotropic or at least
pressure dependent?
Isotropic (scalar) damage

Effective stress concept (similar to MAT_81/120/224 etc.)
J. Lemaitre, A Continuous Damage

Mechanics Model for Ductile Fracture
Overall Section Area

containing micro-defects
Reduced (effective)
Section Area
S S
Reduction of effective cross-section leads to
reduction of tangential stiffness
Phenomenological description
Measure of
Damage
S S
D
S
* 1 D
Plane stress condition

xx
Typical discretization with shell elements:

yy
Principle axis
1

2

3
0
0

0
Plane stress
Parameterised
1 (,)
2 (,)
3 0
Definition of stress triaxiality:
xy
yx
1
2 k 1
1 0 0
0 k 1 0
0 0 0
vm
vm
1 (k 1)
3 1 (k 1) k 12
1 (k 1)k 12
(k 1)
sign( 1 )
3 1 (k 1) k
Haigh-Westergaard coordinates in principle stress space
Lode angle
1
I
tr( ) 1
3
3
2 J2 s : s
Deviatoric
plane
Definition of stress triaxiality:
3 3 J3
1
arccos
1.5
3
2
J
2
vm
10
A toy to visualize stress invariants

(downloadable from the www.dynamore.se)
Crafting instructions
page 1:
Download the PDF-file

Print on thick piece of paper
Cut out where indicated
Add four wooden sticks (15cm)
Add some glue where necessary
(engineers should find out the locations without
further instructions all others contact their
local distributor)
Have fun!
11
A toy to visualize stress invariants

(downloadable from the www.dynamore.se)
Crafting instructions
page 2:
Page 2 of the set may be added for further

clarification of the triaxiality variable.
Final shape of toy
12
Plane stress parameterised for shells

Triaxiality
vm
1 (k 1)
3 1 (k 1) k 12
(k 1)
sign( 1 )
3 1 (k 1) k
Bounds:
Compression
lim lim
k
(k 1)
1
sign( 1 ) sign( 1 )
3
3 1 (k 1)k
Tension
lim lim
(k 1)
1
sign( 1 ) sign( 1 )
3
3 1 (k 1)k
tension
vm
compression
Biaxial tension
lim lim
k 1
k 1
(k 1)
2
sign( 1 ) sign( 1 )
3
3 1 (k 1)k
13
How to define the accumulation of damage ?

A comparison of model approaches
Investigation of failure criteria for the following case:
Plane stress:
Small elastic deformations: 1
p1 and 2 p 2
Damage or failure criteria
Isochoric plasticity:
Proportional loading:
p3
2 a 1
p 2 b p1
p1
1 2b
2b
p2
vm
p
vm
4
3
2
p1
2
1
b2
a2
1 a
3 1 a2
14

A comparison of classical model approaches
Some typical loading paths
15

A comparison of classical model approaches
Some typical loading paths
Four criteria
Principal strain:
1
p1
max
max
Equivalent plastic strain:

p
4
3
2
p1
max
3
4 1
p1
2
max
2
Thinning:
max
p3
Diffuse necking:
p1
max
p3
p1
max
2 1
2 1 b2 b
1 b 2 b 2b 2
16
Failure models in the plane of principal strain
Failure strain under

uniaxial tension is
set the same in all
4 criteria.
Thinning and FLD predict
no failure under pure
shear loading.
17
Failure models in the plane of major strain vs. b
18
Failure models in the plane equivalent plastic strain vs. b
Calibrating different criteria

to a uniaxial tension test
can lead to considerably
different response in other
load cases.
19
Failure models: equivalent plastic strain vs. triaxiality
For uniaxial and biaxial tension

different criteria lead to a factor
of 2:
2, p
1, p
0.5
2, p
1, p
p
p
1, p
1, p
20
Johnson-Cook criterion (Hancock-McKenzie )
pf
d1
d1
d3
3
2
d2
1f
d2e
d3
p
vm
1
2
Johnson-Cook and
FLC are very close in
the neighborhood of
uniaxial tension.
21
Parametrized for 3D stress space
22
Lode-angle: Extension- and Compression test

III
III
Possible value for first
principle stress
II III
and I 0
II III
and I 0
30
Compression
Compression
30
Extension
II
II
I
View not parallel to hydrostatic axis
View parallel and on hydrostatic axis

(perpendicular to deviator plane)
23
3D-Stress state parameterised for volume elements

1
0
extension
1
4
0
0
1
4
0
1
compression
1
2
0
1
0
0
1
2
vm
24
Invariants in 3D stress space

Failure criterion extd. for 3D solids
Parameter definition
Stress domain in
sheet metal forming
[Source: Wierzbicki et al.]
m
I
1
vM 3 vM
27 J 3
2 vM 3
mit
J 3 s1 s2 s3
1 or 30
0 or 0
1 or 30
25
Deriving GISSMO from that
26
Closing the process chain: UHS materials for press hardening

Forming simulation
Crash simulation
Hill based models
v. Mises or Gurson model
Ortho- or isotropic yield surface
Strain rate dependency
Isotropic hardening
Isotropic hardening (locally)
Feasibility by thinning or
damage (calibration?!)
Damage evolution (locally)

Failure models
(mapping damage variable)
Dependency on temperature
Microstructural evolution (#244)
Mapping
II
II
II
I
III
III
III
27
Different ways to realize a consistent modeling
One Material Model for Forming and

Crash Simulation
Requirements for Forming Simulations:

Anisotropy, Exact Description of Yield
Locus, Kinematic Hardening, etc.
Requirements for Crash Simulation:
Dynamic Material Behavior, Failure
Prediction, Energy Absorption, Robust
Formulation
Leads to very complex model
Modular Concept for the Description

of Plasticity & Failure
Plasticity and Failure Model are treated

separately
Existing Material Models are kept
unaltered
Consistent modeling through the use of
one damage model for forming and crash
simulation
*MAT_ADD.(damage)
28
Produceability to Serviceability: Modular Concept
Mapping
Forming simulation
Material model
pl , t
Crash simulation
pl ,0 , t0
Material model
, pl
D
Damage model
, pl
Damage model
Modular Concept:
Proven material models for both disciplines are retained
Use of one continuous damage model for both

29
Produceability to Serviceability: Modular Concept

Current status in 971R5
Mapping
Barlat
, pl , t
, pl
GISSMO
Fortran-Program
Forming simulation
Crash simulation
0 , pl ,0 , t0
Mises
, pl
D
GISSMO
Ebelsheiser, Feucht & Neukamm [2008]

Neukamm, Feucht, DuBois & Haufe [2008-2010]
30
GISSMO
Failure criterion extd. for 3D solids
Parameter definition
m
I
1
vM 3 vM
Stress domain in
sheet metal forming
27 J 3
2 vM 3
mit
[Source: Wierzbicki et al.]
J 3 s1 s2 s3
Xue
Hutchinson
Gurson std.
Xue
Hutchinson
Gurson std.
[Experimental data
by Wierzbicki et al.]
31
GISSMO - a short description

Ductile damage and failure
Damage evolution
Failure curve
Mises
Gurson
Crash
Damage regularly
overestimated for linear
damage accumulation!!
GISSMO
Forming
triaxiality
32
GISSMO a short description
t
1
Engineering approach for instability failure

Instability evolution
v ,loc
Instability curve definition
1 1
F n v
Mises
Gurson
Flachzugprobe
DIN ENDIN
12001
Tensile
test specimen
EN 12001
0,50
0,45
0,40
0,35
0,30
0,25
0,20
0,15
0,10
0,05
0,00
0,00
Crash
Forming
Material Instability
Simulation
Versuch
0,05
0,10
0,15
0,20
0,25
0,30
triaxiality
33
GISSMO a short description

Mesh size dependency: Simple regularisation
Test program and calibration
Inherent mesh-size dependency of results

in the post-critical region
Simulation (and calibration) of tensile test
specimen with different mesh sizes
34
GISSMO a short description for the uniaxial case

Generalized Incremental Stress State dependent damage MOdel
DMGTYP: Flag for coupling (Lemaitre)
DCRIT, FADEXP: Post-critical behavior
D DCRIT
*
1
1 DCRIT
1 D
True Stress
True Stress
FADEXP
m=2
m=5
m=8
GISSMO dmgtyp2
MAT_024
True Strain
True Strain
35
GISSMO
Non-proportional loading influence on failure behavior
Failure strain depending on
Loading state
fail ( )
Load path and pre-damage

Accumulation of damage
Failure
fail
m / eq
36
Non-Proportional Loading
Compare 4 load paths:
A
uniaxial tension
B
uniaxial tension with lateral confinement
C
uniaxial tension with/without confinement
D
uniaxial tension without/with confinement
A
D
f
B
0.577
C
D
p
vm
0.333
Triaxiality is no longer constant

over the stress-strain path if
lateral constraint is imposed or
lifted.
37
Non-proportional loading criterion without triaxiality
d1
0.4
p
dt
p
vm
0.577
0.333
38
Non-proportional loading criterion with triaxiality
d1
d2e
d3
dt
0.66e
vm
1.5
p
vm
cte
vm
p
vm
0.577
0.333
39
Comparison: Uniaxial tension with/without confinement
p
vm
0.577
0.333
40
Input of GISSMO
Input definition in LS-DYNA (1/4)
*MAT_PIECEWISE_LINEAR_PLASTICITY
$
mid
ro
e
pr
100 7.8000E-9 2.1000E+5 0.300000
$
c
p
lcss
lcsr
0.000
0.000
99
0
...
*MAT_ADD_EROSION
$
MID
EXCL
100
$
MNPRES
SIGP1
$
$
IDAM
1
SIZFLG
0
DMGTYP
0
REFSZ
5
sigy
etan
fail
tdel
Standard material
Input (i.e. MAT_24)
vp
1.000
MXPRES
MNEPS
EFFEPS
VOLEPS
NUMFIP
NCS
SIGVM
MXEPS
EPSSH
SIGTH
IMPULSE
FAILTM
LCSDG
11
NAHSV
4
ECRIT
0.3
DMGEXP
2
DCRIT
0.2
FADEXP
2
LCREGD
12
If IDAMG=1 GISSMO is invoked.

Further parameters are expected.
Standard failure
parameters (optional)
GISSMO failure
parameters
DMGTYP=0: Damage is accumulated

But not coupled to stresses, no failure.
DMGTYP=1: Damage is accumulated
and coupled to flow stress if D>Dcrit.
Failure occurs if D=1.
43
Input of GISSMO
$
mid
ro
e
pr
100 7.8000E-9 2.1000E+5 0.300000
$
c
p
lcss
lcsr
0.000
0.000
99
0
...
*MAT_ADD_EROSION
$
MID
EXCL
100
$
MNPRES
SIGP1
$
$
IDAM
1
SIZFLG
0
DMGTYP
0
REFSZ
5
sigy
etan
fail
tdel
Standard material
Input (i.e. MAT_24)
vp
1.000
MXPRES
MNEPS
EFFEPS
VOLEPS
NUMFIP
NCS
SIGVM
MXEPS
EPSSH
SIGTH
IMPULSE
FAILTM
LCSDG
11
NAHSV
4
ECRIT
0.3
DMGEXP
2
DCRIT
0.2
FADEXP
2
LCREGD
12
ID of curve defining equivalent plastic

strain at failure vs. triaxiality.
Damage will have values from 0 up to 1.0.
Exponent for nonlinear damage

accumulation:
Standard failure
GISSMO failure
parameters
ECRIT: Equiv. plastic strain at instability onset.

If negative a curve is expected that defines
ECRIT vs. traiaxiality.
If ECRIT=0 the ordinate value Dcrit is used.
44
Input of GISSMO
$
mid
ro
e
pr
100 7.8000E-9 2.1000E+5 0.300000
$
c
p
lcss
lcsr
0.000
0.000
99
0
...
*MAT_ADD_EROSION
$
MID
EXCL
100
$
MNPRES
SIGP1
$
$
IDAM
1
SIZFLG
0
DMGTYP
0
REFSZ
5
sigy
etan
fail
tdel
Standard material
Input (i.e. MAT_24)
vp
1.000
MXPRES
MNEPS
EFFEPS
VOLEPS
NUMFIP
NCS
SIGVM
MXEPS
EPSSH
SIGTH
IMPULSE
FAILTM
LCSDG
11
NAHSV
4
ECRIT
0.3
DMGEXP
2
DCRIT
0.2
FADEXP
2
LCREGD
12
Fading exponent as defined in:
D DCRIT
* 1
1 DCRIT
FADEXP
Standard failure
GISSMO failure
parameters
Load curve defining element size dependent

regularization factors vs. equiv. plastic strain to
failure.
SIZFLG=0 (default) Characteristic element is

computed with respect to initial configuration.
SIZFLG=1 Characteristic element updated in each
increment (actual configuration, not recommended)
45
Input of GISSMO
$
mid
ro
e
pr
100 7.8000E-9 2.1000E+5 0.300000
$
c
p
lcss
lcsr
0.000
0.000
99
0
...
*MAT_ADD_EROSION
$
MID
EXCL
100
$
MNPRES
SIGP1
$
$
IDAM
1
SIZFLG
0
DMGTYP
0
REFSZ
5
sigy
etan
fail
tdel
Standard material
Input (i.e. MAT_24)
vp
1.000
MXPRES
MNEPS
EFFEPS
VOLEPS
NUMFIP
NCS
SIGVM
MXEPS
EPSSH
SIGTH
IMPULSE
FAILTM
LCSDG
11
NAHSV
4
ECRIT
0.3
DMGEXP
2
DCRIT
0.2
FADEXP
2
LCREGD
12
Reference element size for which the failure

parameters were calibrated. This may be used
when mapping of failure parameters from finer
to coarser meshes is necessary.
Standard failure
GISSMO failure
parameters
Number of additional history variables written to d3plotdatabase. Please keep in mind, that this additional.
history variable are added to the standard variables of
the corresponding material model. I.e. the sum of both
history variable sets needs to be defined in
*DATABASE_EXTEND_BINARY. The number of the
history variables used in the material model (i.e.
MAT_24 in this example) is written to d3hsp-file.
46
Scherzugproben Kerbradius 1m m , 15
Sigma_technisch [GPa]
GISSMO vs. Gurson vs. MAT_24/81

Comparison of experiments and simulations
Mini-Flachzugproben ungekerbt
0,50
0,40
Versuch
0,30
GISSMO
0,20
Gurson
0,10
constant (v. Mises)
0,00
0,00
0,05
Kerbradius
1mm 1mm
NotchedMini-Flachzugproben
tensile specimen,
notch radius
Small tensile test specimen
0,10
0,6
0,50
0,50
0,5
0,40
0,40
0,4
0,30
0,3
0,20
0,2
0,10
0,1
0,0
0,2
0,4
0,6
0,8
0,0
-0,10
Flachzugproben
ENEN
10002
Tensile
specimenDIN
DIN
12001
0,30
0,20
0,10
0,10
0,30
0,50
0,00
0,00
0,50
10
0,40
0,40
0,30
0,30
0,20
0,20
0,10
0,10
0,00
0,30
0,40
0,00
0,15
0,20
Arcan
0,50
0,20
0,10
12
0,60
0,10
0,05
Scherzugproben
Kerbradius 1mm, 0
Shear
test, straight
0,60
0,00
0,00
0,20
Scherzugproben
Kerbradius
Shear test, inclined
15 1mm, 15
0,60
0,00
0,15
eps_technisch
0,05
0,10
0,15
0,20
0,0
0,5
1,0
1,5
47
Gurson vs. GISSMO regularized

Regularization of element size dependency
Gurson
Resultant Failure Strain constant
Failure energy depending on el. size
Identification of damage parameters
is difficult
GISSMO
Failure Strain constant
Fracture energy constant
Identification of Damage Parameters
is more straight-forward
48
Summary
Use of existing material models and respective parameters
Constitutive model and damage formulation are treated separately
Allows for the calculation of pre-damage for forming and
crashworthiness simulations
Characterization of materials requires a variety of tests
Offers features for a comprehensive treatment of damage
in forming simulations and allows simply carrying aver to crash
analysis
49
Thank you for your attention!
50

Haufe GISSMO

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Haufe GISSMO

Hochgeladen von

Copyright:

Verfügbare Formate

GISSMO Material Modeling with

a sophisticated Failure Criteria

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Technological challenges in the automotive industry

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Technological challenges in the automotive industry

Design to the point

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Closing the process chain: Standard materials / state of the art

Hill based models

v. Mises or Gurson model

Anisotropiy of yield surface

Strain rate dependency

State of the art: Failure by FLD

NEW: Computation of damage

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Some thoughts on damage formulations

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Isotropic (scalar) damage

J. Lemaitre, A Continuous Damage

Overall Section Area

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Plane stress condition

Typical discretization with shell elements:

Definition of stress triaxiality:

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Haigh-Westergaard coordinates in principle stress space

Definition of stress triaxiality:

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

A toy to visualize stress invariants

Download the PDF-file

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

A toy to visualize stress invariants

Page 2 of the set may be added for further

Final shape of toy

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Plane stress parameterised for shells

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

How to define the accumulation of damage ?

Small elastic deformations: 1

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

How to define the accumulation of damage ?

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

How to define the accumulation of damage ?

Equivalent plastic strain:

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Failure models in the plane of principal strain

Failure strain under

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Failure models in the plane of major strain vs. b

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Failure models in the plane equivalent plastic strain vs. b

Calibrating different criteria

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Failure models: equivalent plastic strain vs. triaxiality

For uniaxial and biaxial tension

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Johnson-Cook criterion (Hancock-McKenzie )

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe

Parametrized for 3D stress space

LS-Dyna Developer Forum 2011 DYNAmore Stuttgart A. Haufe