Guidelines for Stretch Flanging Advanced High Strength
Steels
S. Sriram, J. Chintamani
Mittal Steel U. S. A
Research and Development
3001 E. Columbus Drive
E. Chicago, IN-46312
Abstract. Advanced High Strength Steels (AHSS) are currently being considered for use in closure and structural
panels in the automotive industry because of their high potential for affordable weight reduction and improved
performance. AHSS such as dual phase steels are currently being used in some vehicle platforms. From a
manufacturing perspective, stretch flanging during stamping is an important deformation mode requiring careful
consideration of geometry and the die process. This paper presents some geometric and process guidelines for stretch
flanging AHSS. Hole expansion experiments were conducted to determine the failure limit for a sheared edge condition.
Effects of punching clearance, prestrain and prior strain path on hole expansion were explored in these experiments. In
addition, dynamic explicit FE calculations using LS-DYNA were also conducted for a typical stretch flange by varying
some key geometric parameters. The experimental and FEA results were then analyzed to yield process and geometric
guidelines to enable successful stretch flanging of AHSS.
During stamping, typically after the draw
operation, the panel is trimmed and flanged before
assembly. Depending on the geometry, flanges can be
classified as shrink flange, stretch flange or a
combination of the two. During stretch flanging, the
trimmed edge is subjected to a tensile strain causing
splits in some cases. In a recent experimental study
[2], AHSS were found to be inherently more
susceptible to stretch flanging failures. Development
of guidelines for stretch flanging is necessary for
implementation of robust manufacturing processes for
parts stamped using AHSS.
INTRODUCTION
Advanced High Strength Steels (AHSS), such as
Dual Phase (DP) steels and Transformation Induced
Plasticity (TRIP) steels offer an attractive option for
lightweighting of automotive panels because of an
optimum combination of strength, formability and
cost. It is anticipated that in the near future AHSS will
comprise approximately 35% of the automotive body
structure replacing conventional High Strength Low
Alloy (HSLA) steels and bake hardenable steels [1].
For exposed panel applications, dual phase steels of
minimum ultimate tensile strength 500 MPa, also
known as DP500 are promising candidates for panel
lightweighting. In many cases in North America,
DP500 is being considered as a substitution for bake
hardenable steels of minimum yield strength of 210
MPa (BH210) for light weighting and improvement of
dent resistance of closure panels. To enable successful
implementation of AHSS in vehicle platforms, product
and process guidelines for different aspects of
manufacturing, such as stamping and welding are
being actively sought by the automotive industry.
The design of stretch flanges and trim-line
development has been explored in several
analytical/numerical studies [3-9].
One of the
common elements in these studies is the discovery of
the fact that the stress state at the edge of the stretch
flange is uniaxial tensile. Attempts to design the
flange or the trim-line were made for a "general"
flange, where the flange length, root radius, flange
angle and bend angle were treated as variables. In
these cases, the analytical work has mostly focused on
mild steels such as DQSK or IF steels, or in some
CP778 Volume A, Numisheet 2005, edited by L. M. Smith, F. Pourboghrat, J.-W. Yoon, and T. B. Stoughton
© 2005 American Institute of Physics 0-7354-0265-5/05/$22.50
681
cases Al alloys with very little published work on high
strength steels.
'LH
In this paper some guidelines for split-free stretch
flanging of a V-shaped flange are presented. Typical
formability FEA utilizes the FLC to determine the
occurrence of necking during analysis. However, after
shearing, the structure of the free edge is complicated
comprising of rollover, sheared and fractured zones
with the presence of a burr and is substantially
different from the bulk of the material. As a result, the
FLC is not an appropriate failure criterion to be used
for edge stretching predictions. In this study, a
combination of experimental and finite element
analyses were used to yield the stretch flanging
guidelines.
6KHHW
%LQGHU
3XQFK
Figure 1: FEA model of a typical stretch flange
Dynamic explicit FE analyses were conducted by
using a parametric representation of flange geometries
typically observed at some locations in door outer
panels. Geometrical parameters were varied using a
L8 DOE. The maximum strain determined from the
simulations were statistically analyzed to determine
the sensitivity of the strains to the geometric
parameters, and to develop regression equations. Hole
expansion testing was carried out to determine the
sheared edge stretching limits of sheet steels. Finally,
comparison of the maximum strain for a given flange
geometry to the failure limits determined by hole
expansion testing would provide information on the
feasibility of forming a given flange geometry. This
study has been focused on BH210 and DP500 steel
products.
T
)ODQJH5DGLXV5
)ODQJHZLGWKZ
Figure 2: Parameters of the stretch flange model
The values of the parameters were chosen after
measurement of selected areas that were subjected to
stretch flanging on some door outer panels.
TABLE 1. Values of parameters in the DOE
Variable
Low Value
High Value
R (mm)
5
25
110
170
θ (degrees)
W (mm)
3
10
METHODOLOGY
FE Simulations
Experiments
Dynamic explicit FE calculations were conducted
using LS-Dyna. Figures 1 and 2 show the model and
some of the parameters that were used to describe the
part geometry. A full factorial L8 DOE was used for
the parametric study of the model shown in Figures 1
and 2. Table 1 shows the values of the parameters
used in the study.
Determination of the sheared edge stretching limits
was conducted using the hole expansion test. Table 2
shows the mechanical properties of the materials in the
experiments as determined using standard ASTM
tensile testing in the L, T, and D directions. The
properties were averaged using standard expressions.
TABLE 2. Standard mechanical properties of the materials used in the study
Material
YS (MPa)
UTS (MPa)
UE (%)
TE (%)
n-bar
BH210
250
370
21
40
0.19
DP500
300
547
19
29
0.19
682
R-bar
1.5
0.9
Thickness (mm)
0.8
0.81
Figure 3 shows the geometry of the test specimen.
The test is conducted by punching a circular hole in a
sheet metal blank, and then subsequently stretching
the hole using a conical punch. The sheared edge
stretching limit is then given by the following
relationship:
% HE =
d f − d0
d0
h
203
305
10000mm
(1)
Hole expansion
specimen
w
Figure 4: Schematic of experimental panels used
for pre-straining samples.
RESULTS
Figure 3: Geometry of the hole expansion test
specimen.
FEA Simulations
Where d0 is the initial diameter of the hole (10mm),
and df the final formed diameter of the hole after
expansion. During production, typical clearance
used in trimming or punching operations is in the
range of 10 to 12% of the thickness of the sheet
metal per side. For the standard experiments in this
study, holes were punched using a cutting clearance
of 10% of metal thickness per side. The test piece is
formed with the burr side of the punched edge on the
opposite side of punch contact and is stopped when a
crack propagates through the thickness of the sheet.
To simulate the effect of wear on trim dies during
production, the effect of cutting clearance on sheared
edge stretching limits was evaluated by testing
specimens, where holes were punched at different
clearances.
Figure 5 shows a contour plot of the major true
strain for one of the simulations. As seen in Fig. 5,
the maximum strain occurs at the edge of the flange.
Analysis of the strain path at the edge of the flange
indicates a uniaxial tensile path as to be expected for
a free edge.
Maximum strain
The effect of prior work hardening history was
also evaluated where axisymmetric panels as shown
in Figure 4 were formed. The dimension “w” was
varied to produce strain paths of uniaxial tension,
plane strain and balanced biaxial tension. The
forming depth of the panel was changed to produce
equivalent plastic strain values of 0.05 and 0.1 on the
face of the panel. After forming, 100mm square
samples were cut from the panels, and holes punched
for hole expansion testing following the earlier
described procedure. To study the effect of prior
work hardening, the cutting clearance was set at 10%
of metal thickness.
DP500
Angle = 110o
Radius = 5mm
Width = 10mm
Figure 5: Contour plot of the true major strain for
one of the simulations.
Figure 6 shows a comparison of major strains
between BH210 and DP500 for the case shown in
Figure 5. The major strains on the free edge of the
flange are shown as a function of the undeformed
position of the node from the origin as marked in
Figures 6-9. As seen in Fig. 6, minor differences
were observed between the two materials. The effect
of the main variables θ, R, and w are shown in
Figures 7-9.
683
Statistical analysis was conducted to determine
the important variables and interactions governing
the magnitude of the maximum major strain at the
edge of the flange as a function of the different
variables. Tables 3 and 4 show the results of the
statistical analysis for BH210 and DP500
respectively.
Flange Angle = 110 Degrees
Flange Radius = 5mm
Flange Width = 10mm
120
Major Strain (%)
Origin
DP500
100
80
BH210
60
40
Table 3: Statistical analysis of maximum major true
strain for BH210.
20
0
0
5
10
15
20
BH210
25
Total Sum of Squares: 0.285831938
Factor
Sum of Squares
Angle
0.131220522
Radius
0.054321376
AngleXRadius
0.039674628
width
0.028295826
AngleXwidth
0.029575552
RadiusXwidth
0.002400552
Angle X Radius X width
0.000343482
Original Distance (mm)
Figure 6: Strain distributions for BH210 and DP500
for the case shown in Fig. 5
BH210
Flange Radius = 5mm
Flange Width = 10mm
100
Origin
90
dof
1
1
1
1
1
1
1
Factor SS/Total SS (%)
45.91
19.00
13.88
9.90
10.35
0.84
0.12
Major Strain (%)
80
Angle = 110
70
Table 4: Statistical analysis of maximum major true
strain for DP500.
60
50
40
Angle = 170
DP500
30
Total Sum of Squares: 0.342465516
Factor
Sum of Squares
Angle
0.146746531
Radius
0.068950411
AngleXRadius
0.047975629
width
0.034518781
AngleXwidth
0.03840329
RadiusXwidth
0.004623373
Angle X Radius X width
0.001247501
20
10
0
0
5
10
15
20
25
Original Distance (mm)
Figure 7: Effect of flange angle on the major strain
distribution
dof
1
1
1
1
1
1
1
Factor SS/Total SS (%)
42.85
20.13
14.01
10.08
11.21
1.35
0.36
BH210
Angle = 110 degrees
Flange width = 10mm
100
90
Major strain (%)
As seen in Tables 3-4, the flange angle has a very
significant influence on the magnitude of the
maximum major strain in the flange, as a single
variable as well as its interactions with the other
variables. Based on the FE and statistical analysis, a
regression model for the maximum major strain as a
function of the different flange variables was
developed.
Origin
80
R = 5mm
70
60
50
R=25mm
40
30
20
10
0
0
5
10
15
20
25
30
Original distance (mm)
Figure 8: Effect of flange radius on the major strain
distribution.
Experiments
BH210
Radius = 5mm
Angle = 110 Degrees
The results of the experimental work presented in
this section are used to propose a failure criterion for
edge stretching problems. Figure 10 shows the effect
of the punching clearance on the results of the hole
expansion test. The effect of increasing the punching
clearance is to increase the burr height and thus
degrade the quality of the sheared edge. As seen in
the Fig. 10, the increase in cutting clearance results
in a decrease in the % hole expansion. The decrease
is more significant up to ~ 30%clearance. Beyond
30% clearance, the % hole expansion reaches a
saturation value. Within the experimental limits of
this study, it was found that prior work hardening has
a small effect on the subsequent hole expansion, as
100
90
Origin
Major Strain (%)
80
70
w=10mm
60
50
40
w=3mm
30
20
10
0
0
5
10
15
20
25
Original Distance (mm)
Figure 9: Effect of flange width on the major strain
distribution.
684
composition of a given steel of interest. The hole
expansion test is simulative of the strain state and the
strain gradient in the deforming region of the flange
and is therefore a good practical first order
approximation of the edge stretching limits. As seen
in Fig. 10, there is no change in hole expansion for
clearances greater than 30%. Thus, using the hole
expansion value at 30% punching clearance in
conjunction with the regression model for maximum
strain described before could be used to determine
the feasibility of flanging a given part.
seen in Figure 11. Further testing at higher prestrains
is necessary to discern a broader trend. Also, it
appears that there is no effect of the prior strain path
on the hole expansion. From Figures 10 and 11, it
can be concluded that the sheared edge quality
dominates the edge stretching limits over prior work
hardening.
140
% Hole Expansion
120
BH210
100
80
60
Comparison with production trials
40
DP500
20
0
0
10
20
30
40
50
60
Punching Clearance (% of metal thickness)
Figure 10: Effect of punching clearance on the hole
expansion of BH210 and DP500
Some comparisons were conducted using the
predictive model with a stamping trial for a door
outer panel. Figure 12 shows the door outer and a
zoomed-in view of the area where splits during
flanging were observed for DP500.
140
% Hole Expansion
120
Angle
100
Radius
BH210
width
80
60
40
DP500
Plane Strain
DP500
Uniaxial Tension
20
BH210
Balanced Biaxial
Figure 12: Photograph of the door outer panel
showing the area of stretch flanging.
0
0
0.02
0.04
0.06
0.08
0.1
0.12
Equivalent Plastic Strain
Figure 11: Effect of prior work hardening on the hole
expansion of BH210 and DP500
As discussed earlier, use of a good failure
criterion in conjunction with FE analyses is an
important aspect of predicting manufacturing
feasibility of a given flange. In the literature, there is
a lack of suggested experimental techniques for
standard determination of edge stretching limits to be
used with FEA. In this study, the results from the
hole expansion test are used as failure criteria for
flange formability. Use of the hole expansion value
as a failure criterion is supported by the fact that the
free edge for the expansion test and for the flange
geometry in Figures 1-2 are subjected to uniaxial
tension. The strain gradient from the edge of the
flange to the edge of the die (flange width) is also
similar between the hole expansion test and a general
flange although the exact shapes of the gradients will
show differences. Furthermore, the edge stretching
properties of AHSS are significantly dependent on
the microstructure, processing and chemical
685
Measurements of the angle, radius and flange
width were undertaken for the two doors to yield the
maximum strain in the flange as predicted by the
regression model. Table 5 shows the results of this
comparison where the maximum strain predicted
from the model is compared with the hole expansion
value. Table 5 shows that the DP500 material would
have split at the flange location as the maximum
flange strain was higher than the material limit.
Experience during production indicated that this was
indeed the case. Also shown in the third column of
Table 5 is a scenario resulting in a safe DP500 part,
where for the same radius and angle, reducing the
flange width to 3mm would result in the maximum
strain being less than the material limit.
3. For BH210 and DP500, the hole expansion limit
is strongly dependent on the punching clearance.
Beyond a clearance of 30%, the effect of the
punching clearance was insignificant.
Table 5: Correlation of model predictions with
production experience for BH210 and DPS 00
I
Model Validation
Flange Radius (mm)
Flange Angle (degrees)
Flange width (mm)
Maximum strain predicted by model
Hole expansion @ 30% punching clearance
Flanging experience during trial
BH210
DP500
DP500 (safe)
14.29
110
6.16
14.29
115
6.29
14.29
115
3.0
42.05%
52.50%
Safe
42.51%
27.00%
Split
26.61%
27.00%
NA
4. In the range of the variables considered, the
magnitude and strain path of prior work
hardening did not have a strong influence on the
hole expansion limit.
5. The hole expansion limit at 30% hole punching
clearance was used as failure criteria to determine
success in a flanging operation.
DISCUSSION
ACKNOWLDEGEMENT
From the results presented in this paper, it can be
seen that the accuracy of the prediction is dependent
upon the failure criterion. In this paper, the hole
expansion value at 30% punching clearance was
used. During production, where trimming has to take
place along a complicated contour, it is difficult to
maintain the cutting clearance at an optimum value
of 10%. In addition, wear on the trim dies during
high volume production would result in an increase
in the clearance. For robustness of the flanging
operation, it is thus important to use a conservative
estimate for the failure limit. The generality of the
failure criterion for different flange geometries is yet
to be evaluated and will be undertaken for future
work.
The authors acknowledge Ray Rizzo for
conduction of the experimental work. Support and
encouragement from Mittal Steel USA's R&D
management is also acknowledged.
REFERENCES
1. Horvath, C., The Future Revolution in Automotive
High Strength Steel Usage, Great Designs in Steel
2004, AISI, Southfield, MI, Feb. 2004.
2. Sriram, S., Yan, B., Huang, M., Characterization of
Press Formability of Advanced High Strength Steels
Using Laboratory Tests, SAE Paper 2004-01-0506,
SAE, Warrendale, PA
3. Wang, N.-M., Wenner, M. L., An analytical and
experimental study of stretch flanging, Int. J. Mech.
Sci. 1974, vol. 16, pp. 135-143.
4. Wang, N.-M., Johnson, L. K., Tang, S. C., Stretch
flanging of V-shaped sheet metal blanks, J. Applied
Metalworking, vol. 3 No. 3, July 1984, pp. 281-291.
5. Dudra, S., and Shah, S., Stretch flanges: Formability
and trimline development, J. Materials Shaping
Technology, vol. 6, no. 2, 1988, pp. 91-101.
6. Demeri, M.Y., Tang, S. C., Computer simulation and
experimental validation of stretch flanging, J. Materials
Shaping Technology, (1991), vol. 9, pp. 241-251.
7. Wang, C. T., Kinzel, G., Allan, T., Failure and
"wrinkling criteria and mathematical modeling of shrink
and stretch flanging operations in sheet metal forming,
J. Materials Processing Technology, 53 (1995) pp. 759-
In general, for successful flanging of AHSS, the
width of the flange has to be decreased significantly
for aggressive flange configurations. The work
presented in this paper does not account for all
possible variations in general flanges but mainly
demonstrates an approach to provide guidelines for a
given type of stretch flange configuration as seen in
exposed panels. It is possible to extend this work to
other flange configurations by suitably providing a
parametric representation of the geometry.
CONCLUSIONS
1. A combination of FE modeling and simulative
laboratory formability tests can be used to
provide guidelines for successful stretch flanging
of AHSS parts. This approach was verified in an
actual stamped part.
780.
8. Worswick, M. J., and Finn, M. J., The numerical
simulation of stretch flange forming, Int. Journal of
Plasticity, 16 (2000) pp. 701-720.
9. Feng, X., Zhongqin, L., Shuhui, L., Weili, X., Study on
the influences of geometrical parameters on the
formability of stretch curved flanging by numerical
simulation, J. of Materials Processing Technology, 145
(2004), pp. 93-98.
2. The flange angle has the most significant
influence on the maximum strain at the edge of
the flange among the variables considered.
686