A Review of FEA Metal Forming Simulation Practices
in Automotive Industry
C. Q. Hu
2/18/2008
1
Contents
• Factors affecting stamping feasibility
• Material model
• Friction models
• Simulation methodology
• Parameters of current simulations and the
sensitivity
• Not quantified parameters
• Summary
2/18/2008
2
1. Factors affecting metal forming feasibility
Tooling (punch and die)
¾
¾
¾
¾
¾
Die face design (addendum)
Kinematics: tipping, binder wrap, punch contact progression
Draw beads design
Die material (hardness, rigidity & friction behaviour)
Blank location
Blank (sheet metal)
¾
¾
¾
¾
¾
¾
¾
¾
Mechanical properties ( r, n, r, YS, UTS values)
Gauge (thickness)
Variation of the mechanical properties and gauge
Blank size
Surface coating
Surface finish (topography)
Edge conditions
Aging and pre-strain conditions
2/18/2008
3
1. Factors affecting metal forming feasibility
Press
¾
¾
¾
Press type (hydraulic or mechanical)
Binder pressure (force)
Press speed
Lubrication
¾
¾
Lubrication ( type, viscosity, temperature and pressure sensitivity)
quantity
Process noise
¾
¾
¾
¾
¾
Blank location
Tool wear
Dirt
Ambient temperature and humidity
Press counter balance pressure
2/18/2008
4
2. Material models
An accuracy of FEA prediction is dependant on plastic models, in which different yielding
criteria and hardening
laws are suitable different materials and stamping process.
r
Autoform
Yielding Criteria
PamStamp 2G
Input requirements
current
new
current
new
Von Mises
(isotropic)
yes
yes
yes
yes
σy, r bar
Hill’s 48
(Planar anisotropic)
yes
yes
yes
yes
σy, r values at 0, 45 and
90 degrees
Hill’90 (introducing bi-axial
yield stress)
no
yes
yes
yes
σy, r values at more
than three angles or σy
at a bi-axial stress state
6 component Barlat
no
no
no
yes
Uni, bi-axial, and shear
tests
Vegter
no
no
no
no
Uni, bi-axial, and shear
tests
2/18/2008
5
2. Material models
Hardening Laws :
Kinematic hardening, not isotropic hardening, takes account of non-linear strain
path, such as bending/unbending, reverse drawing in a multi-stage forming
process
Autoform
PamStamp 2G
Hardening law
Input requirements
current
new
current
new
Isotropic
yes
yes
yes
yes
σy, UTS, n
Kinematic
no
no
yes
yes
σy, UTS, n
Hardening is represented by :
• A tensile test curve along the rolling direction
•
•
The power law (Ludwik equation or Krupkowski equation ) for steels
The voce equation for Al alloys
2/18/2008
6
3. Friction model
•
In practice, coefficient of friction is dependant on contact pressure,
slide speed (static and kinetic), blank coating, lubrication and tool
surfaces, etc.
•
Coulomb friction law, which assumes coefficient of fiction is
constant regardless kinetic factor, contact pressure and slide speed,
is the only option in both AutoForm and PamStamp 2G
•
Other general FEA code, such as LS/Dyna, HKS/ABAQUS,
provide more detailed friction models
2/18/2008
7
4. Simulation methodology
Product design
CAE
tool maker
Manufacture
CAE
Concept
feasibility
Tool Design
Press trial
2/18/2008
8
4. Simulation methodology
4.1 Simulation practices in concept stage
AutoForm for most panels:
¾ To develop binder surface, addendum components
¾ To study feasibility, using:
Von Mises (or Hill 45) yielding criterion, isotropic
hardening law
Blank holder load: decided using empirical formula,
Coefficient of friction: 0.14(steel), 0.12 or 0.17 (Al)
¾ Using more conservative feasibility criteria
PamStamp 2G only for A panels (skin) for validation, using:
¾
¾
¾
2/18/2008
Hill 45 yielding and isotropic hardening law
Coefficient of friction: 0.14(steel), 0.12 or 0.17 (Al)
Average mesh size : 10 mm, Adaptive mesh level, 2
9
4. Simulation methodology
4.2 Feasibility criteria in the concept stage
¾ For necking failure, allowing 20% safety margin
¾ For steel inner panels:
Maximum thinning <25%
¾ For steel skin panels:
Maximum thinning <25%, and minor strain≥1.0%, free of slid lines.
¾ For Aluminium skin panels:
¾
Maximum thinning <20%, and minor strain≥1.0%, free of slid lines.
¾ For wrinkling: using AutoForm default settings.
2/18/2008
10
4. Simulation methodology
4.3 The simulation practices in the tool maker sector
?
4.3 Feasibility criteria in the tool maker sector
?
2/18/2008
11
5. Parameters of current simulations and the sensitivity
5.1 Yielding criterion of Material model
•
•
•
•
•
For mild steels, Hill 48 appears to be accurate enough
Details
For HSS, Hill 90 is more accurate than Hill 48
For Aluminium, Barlat’s criterion is better than Hill 48 but need not only tensile
tests but also bi-axial and shear tests.
It has been claimed that Vegter criterion can describe yielding behaviour of HSS
and Al alloys, and material tests is easier compared with that for Barlat model.
More sophisticated and practical yield criteria for Al alloys has been expecting
5.2 Hardening law of Material model
•
Isotropic hardening for linear strain path
•
Kinematic hardening for nonlinear strain path
Details
2/18/2008
12
5. Parameters of current simulations and the sensitivity
5.3 Material properties
FLD0
•
Thickness: wrinkling tendency, FLD0, springback
•
YS: material distribution, hardening exponent, springback
•
UTS: (not a direct input parameter); hardening exponent, fracture strain
•
n value: hardening, distribution of strain, FLD0
•
r values: deep draw ability
•
Hardening curve: Krupkowski equation , Ludwik equation or converted tensile curve
•
Blank orientation
•
Direction of a tensile test for mechanical property input (0, 45, 90 degree)
•
Forming limit curve ( test methods)
•
Blank mesh quality and adaptive mesh level
Hardening
n effects
5.4 Tools
•
Tooling radius, size, punch type (flat or crown)
•
The clearance between die and punch: material flow, wrinkling and springback
•
Tool mesh design
2/18/2008
Tool geo
13
5. Parameters of current simulations and the sensitivity
5.4 Process
•
Blank holding pressure or load
•
Press type ( single or double action)
•
Operation stages
•
Artificial press speed setup including of type of curve and magnitude
•
Boundary conditions, especially for springback prediction
5.5 Friction
•
Using a constant coefficient of friction to quantify
different coating of sheet steels and Al alloys, and
other tribological factors
2/18/2008
Example
14
5. Not included Parameters
5.1 Material properties
•
Coating type (dipped and thickness or quantity
•
Surface finish (topology)
•
Material variation with a roll, or along the width of blank, especially for HSS
•
Strain rate effect
5.2 Lubrication
•Type (Fluid-film, solid, Extreme Pressure (EP) lubricants )
•Quantity and distribution
5.3 Process
•
Temperature
2/18/2008
15
6. Summary
•
•
•
•
•
FEA code and CAE experience have been developed to deal with for
convention mild steel panels
The accuracy of Wrinkling and springback prediction is considerablely
dependant on model setup
Regarding new materials, such as HSS and Al alloys, new material
model should be investigated
Due to confidential issues, the publication about bench mark test for
commercial FEA codes is rare
Simulation validation researches, dealing with new materials, different
tool and complex forming process, is required
2/18/2008
16
7. Appendix
7.1 Material models: Yielding criteria
Back
2/18/2008
17
7. Appendix
7.2 Material models: hardening laws
σ1
2σy0
σy0
σy1
σ2
σy1
kinematic
Isotropic
2/18/2008
Back
18
7. Appendix
7.3 Empirical formula: FLD0=min(n/0.24, 1.0)*(23.0+14.6t)%
Back
7.4 Krupkowski equation: σ=K*(ε0 + εp)n , and K=UTS*(e/n)n
Back
7.5 n effects on strain value
0.5
0.8
0.4
n=0.23
n=0.21
0.3
n=0.21
0.6
Floor strain
Peak strain
n=0.23
0.7
n=0.19
n=0.17
0.2
n=0.19
n=0.17
0.5
0.4
0.3
0.2
0.1
0.1
0
0
10
12
14
16
18
Punch depth (mm)
20
22
24
10
12
14
16
18
20
22
24
Punch depth (mm)
Back
2/18/2008
19
7. Appendix
7.6 Tool geometry effects
0.3
Eng. strain
Back
30
25
strain, crown
0.25
20
strain, flat
0.2
15
Crown geo
0.15
10
Flat geo
0.1
5
0.05
0
0
0
50
100
Punch Height (mm)
0.35
-5
200
150
Distance from centre (mm)
7.6 Friction effects
Strain along wall
Back
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
u=0.01
u=0.05
u=0.1
u=0.15
u=0.2
10
12
14
16
18
20
22
Punch depth (mm)
2/18/2008
20