You are on page 1of 4

© Copyright by International OCSCO World Press. All rights reserved.

2006
VOLUME 18
ISSUE 1-2
September–October
2006
Short paper 275
of Achievements in Materials
and Manufacturing Engineering
of Achievements in Materials
and Manufacturing Engineering
Comparison of experimental and simulation
results of 2D-draw-bend springback
T. Da Sisva Botelho
,
, E. Bayraktar*, G. Inglebert
LISMMA, ISMEP/Supméca. 3 rue Fernand Hainaut, 93407 Saint-Ouen Cedex, France
* Corresponding author: E-mail address: bayraktar@supmeca.fr
Received 15.03.2006; accepted in revised form 30.04.2006
Analysis and modelling
AbstrAct
Purpose: This paper presents the results of experimental, analytical and numerical studies of draw-bend
springback on the response of steel blank strips.
Design/methodology/approach: Springback, the elastically-driven change of shape of a part after forming, has
been simulated with 2-D plane strain finite elements model (ABAQUS). Springback simulations compared with
experimental test results under some test conditions and also sidewall curls were further discussed in this paper.
Findings: The results validate the finite element approach as a trustworthy tool for predicting the springback
parameters for a given set of stamping conditions and material properties.
Research limitations/implications: This model is a simple and an efficient way to take apart the responsibilities
of the steel maker for the selection of steel to be used and sheet stamping designer for the selection of the
forming conditions to work in a safety zone in order to increase the reliability of the stamped structures.
Originality/value: Experimental and finite element analyses have been conducted on blank strips of Hadfield
steel (high strength steel) and Mild steel with similar operational conditions.
Keywords: Computational mechanics; 2D-bending; Springback; Mild steel; Hadfield steel
1. Introduction
The phenomenon of springback in sheet forming operations can
easily takes place during bending and unbending deformations of
the sheet. This phenomenon appears when the sheet first makes
contact with the tool surface at the stage of bending and then leaves
the tool surface and partially or completely loses its curvature at the
stage of unbending.
Accurate prediction of springback and sidewall curls is essential
for the design of tools mainly used in sheet stamping operations.
These deformation sets generate more complex stress-strain states
in the sheet resulting in the formation of sidewall curl after the sheet
is allowed to unload [1-8].
Mild steel (MS) and Hadfield steel (HFS) were used for this
analysis, of which the yield stresses varied from 170 to 450 MPa.
Springback, the elastically-driven change of shape of a part after
forming, was simulated with 2-D plane strain finite elements model
(ABAQUS) as many industrial sheet stamping operations are
realized in plane strain conditions [2,4-8].
This paper aims essentially the correlation and comparison
between experimental and simulation results of springback.
2. Experimental conditions
2.1. Materials
Hadfield steel and mild steel are used in this study. Useful data
of materials and mechanical properties are given in Table 1.
2.2. Tools, lubrication and geometry
An apparatus of 2D-draw bending tools was specially designed
according to the schematic drawing shown in Figure 1. The set up
consists of a die, a punch and a blank holder. The upper and lower
platens were separated from the blank holder by springs. The blank
holder can move up and down freely on the four rods attached to the
lower platen.
1.Introduction
2.Experimentalconditions
2.1.Materials
2.2.tools,lubricationandgeometry
Short paper 276
Journal of Achievements in Materials and Manufacturing Engineering
T. Da Sisva Botelho, E. Bayraktar, G. Inglebert
Volume 18 Issue 1-2 September–October 2006
Table 1.
Materials properties
Material
Sheet
thickness
(mm)
Yield
Strength
(MPa)
Fracture
Strength
(MPa)
Strain until
fracture
Strain hardening
Exp(n)
Constitutive equation
Mild Steel 0.78 170 312 0.450 0.235
( )
0.2585
565.32 0.007117
p
o c = +
Hadfield
Steel
1.20 420 ± 30 1350 0.50 ± 0.20 0.45-0.95
( )
194
0.257 exp exp 0.01
0.01
237
1161.exp
T
T
c
o c c
| |
|
\ .
| |
=
|
\ .

Fig. 1. Experimental stamping apparatus and schematic drawing
detail of system
The lower springs hold the blank holder at a small distance
from the die, so that the blank can be introduced easily.
The upper springs are used for the transmission of the blank
holder force. The die and lower platen were fixed to the cross
head after alignment with the punch which is attached to the load
cell. After carefully placing the blank on the die, the upper platen
is pressed against the blank holder, while monitoring the force.
Once the required blank holder force was reached, the upper
platen was tightened to the rods. Care was exercised to keep the
blank holder force as close to the specified value as possible.
Before each experiment both sides of the blank sheet surface
were wiped with a paper towel dipped in the special lubricant and
were then kept in a vertical position for 30 minutes for
homogenization.
A standard U-shaped geometry was selected. Figure 2 gives
the details of the geometry studied for the experimental and
numerical analyses. This type of geometry was chosen for
isotropic springback analyses. Typical measurement method of
springback values was explained formerly [4].
All of experimental study was conducted on the INSTRON
8501 machine.
The punch speed was 1.66 mm/sec. The punch travel was
selected at a specified distance of 70mm and throughout the
experiments, the blank holder forces (BHF) were maintained at
2.45 kN and 19.6 kN at the test temperature of 20 + 2°C for each
steel. The initial sizes of Hadfield and mild steel sheets are
1.2x35x350 mm and 0.78x35x350 mm respectively.
2D-draw bending tool was made from tool steel (D2) and
hardened to 55 HRc. The average roughness values of the
different tool parts were measured at 0.28 µm.
The average roughness values measured on the blank sheets
was 1.04 µm for Mild steel and 1.15 µm for Hadfield steel. The
clearance between the punch and the die was measured as 1.0 mm.
Fig. 2. Measurements parameters for the springback and side wall
curl measured and used later in the finite elements model [4]
3. Results and discussion
3.1. Experimental results
Based on experimental study, springback results were
evaluated and these results were used in order to validate the
proposed finite element calculation.
After unloading of the hat sections, the sidewall curls (ѽ ҏ ȡ, the
radius of a circle through a, b and c) and springback values (
1
u ,
angle between the ox and the [ab] segment and
2
u , angle between
the [ab] and [ef] segments) of the parts were determined.
Schematic description of measurement positions for
springback and sidewall curls are given in the Figure 3.
3.resultsanddiscussion
3.1.Experimentalresults
277
Analysis and modelling
Comparison of experimental and simulation results of 2D-draw-bend springback
Fig. 3. Schematic description of measurement positions for
springback and sidewall curls
The calculation of the springback and the sidewall curls were
performed according to Figure 2.
The mean values of experimentally measured parameters taken
on four measurements on each side of the blank sheets during the
tests were evaluated for Hadfield and Mild steel respectively. They
were prepared in the rolling direction. Only one test at the BHF of
19.6 kN has been carried out for two steel blank strips with an
orientation of 90° with respect to the rolling direction.
The results obtained for both of steel blank strips in rolling
direction were given in Table 2. There were no considerable
differences in the results between two orientations of 0° and 90°
from the rolling direction found for both of the steel blank strips.
This case can be explained for Hadfield steel because it is an
isotropic material but for Mild steel, anisotropy does not
contribute to springback in our experimental conditions.
This point will be clarified using finite element analysis in the
next section. The detail discussion on the experimental and finite
element results will be given in the full paper.
Table 2.
Experimentally measured parameters of springback
Material BHF
T
1
(°)
T
2
(°)
U
(mm)
MS
2.45
19.6
94
100
85
85
112
120
HFS
2.45
19.6
115
115
75
75
103
102
At the first approach, experimental results showed that there
is no significant difference between springback values of Hadfield
steel for two different blank holder forces. However, these values
in mild steel indicate a very strong tendency to increase at higher
blank holder forces.
It appears that, the blank holder force does not influence
springback values for Hadfield steel. This case can be explained
knowing that this steel work hardens very quickly as a function of
plastic deformation level. In all experiments performed for two steels,
sidewalls do not fail by tearing as blank holder force increases. In the
same way, higher springback values of Hadfield steel than that of
mild steel is due to yield property of Hadfield steel.
It is well known that this steel has higher yield strength and
also as very high toughness properties. Therefore its yield strain is
very high. As a result of unbending, springback values occur to be
significantly related to the yield strain level. In other words, the
higher the yield strength the greater the elastic return which
results in higher springback values.
A picture taken from U shaped Hadfield strip at the BHF of
19.6 kN is shown in Figure 4a. 2D-draw bending specimens from
Mild steel strips after removing from die are also given in Figures
4b and 4c.
a)
b) c)
Fig. 4. 2D-draw bending specimens after springback for a Mild
steel strip with a BHF of 19.6 kN (Fig. 4a) and for a Hadfield
steel strip with a BHF of respectively 19.6 kN (Fig. 4b) and
2.45 kN (Fig. 4c)
Springback parameters obtained from experimental results
were used to compare with those of final element analysis in order
to better understand the phenomenon of springback in these
materials that they will be detailed in the following section.
3.2. Finite elements analysis (FEA)
A 2D numerical analysis of the U-shaped bending process
was carried out for comparison with experimental results. In order
to truthfully validate materials parameters, both analyses were
accomplished with similar operational conditions. This consisted
of two constant blank holder forces of 2.45 and 19.6 kN for all the
blank strips and a constant punch speed of 1.66 mm/sec, (a quasi
static displacement was performed). In fact, the FEA was
expected to be independent of rolling direction angle due to the
yielding isotropy within the plane of the sheet. Some of the
experimental and finite element results are given in Table 3 and it
seems to show a good correlation. According to the dimensions of
the blank in the out-of-plane direction, a two-dimensional plane
strain model was used. The elastically - driven change of shape of
metallic sheets have been simulated with ABAQUS code. Due to
plane symmetry, only half of the process was modelled (Figure 5).
The problem consists of the surface contact between steel blank
strip and the tools such as the punch, die and blank holder that is a
basic aspect of the stamping operations. The tools can be
modelled as rigid surfaces because they are much stiffer than the
blank. Figure 5 shows the basic arrangement of the components
considered in FEM model. The blank strip is squeezed between
the blank holder and the die trough a normal load applied on the
3.2.Finiteelementsanalysis(FEA)
Short paper 278 READING DIRECT: www.journalamme.org
Journal of Achievements in Materials and Manufacturing Engineering Volume 18 Issue 1-2 September–October 2006
blank holder while the die remains always unmoving. This force,
in conjunction with the friction between the blank and blank
holder and the blank and die, controls how the blank material is
drawn into the die during the forming process. The contact
between the punch and the blank was supposed to be frictionless
whereas the contacts between respectively the blank and the die
and the blank and the blank holder were supposed to have
a coulomb friction law with a friction coefficient of 0.144.
Table 3.
Comparison of experimental and numerical results for a Mild
steel bent with BHF of 19.6 kN
Angle
Experimental
Mean values
Numerical
results (°)
Comparison (%)
T1 (°) 99 98.8 0.2020
T2(°) 85 86.9 2.2353
The FEM took into account the forming process then the full
unloading process, leading to a springback phenomenon. The
forming process was controlled by the displacement of the punch.
The unloading procedure consists in totally removing the punch,
suppress the BHF and remove the blank holder and the die.
For the sake of the simplicity of the FEA, the following
assumptions were considered:
x An elasto-plastic constitutive model was accepted with
material isotropy and with a non-linear strain hardening
condition.
x U-Shaped bending deformation was assumed to take place
under the plane strain condition.
Punch
Die
Symmetry
plane
Blank Holder
Sheet
R
d
R
p
Fig. 5. Geometrical description of the simulation model
The tendencies form the experimental and FEA are in well
agreement with tendency in the yield strength, it means that the
higher the yield strength the greater the elastic springback values.
4. Conclusion
Experimental and finite element analyses have been
conducted on blank strips of Hadfield steel (high strength steel)
and Mild steel with similar operational conditions. The finite
element results have been validated from experimental results.
These results validate the FE approach as a trustworthy tool for
predicting the springback parameters for a given set of stamping
conditions and material properties.
This model is a simple and an efficient way to take apart the
responsibilities of the steel maker for the selection of steel to be
used and sheet stamping designer for the selection of the forming
conditions to work in a safety zone in order to increase the
reliability of the stamped structures.
Acknowledgements
The authors are grateful to ARCELOR research (CED –
Arcelor Auto, Montataire-France) for their support in this project.
References
[1] W.D. Carden, L.M. Geng, D.K. Matlock, R.H. Wagoner:
“Measurement of springback”, Int. Journal of mechanical
sciences 44 (2002) 79–101.
[2] E. Bayraktar, F. A. Khalid, C. Levaillant: “Deformation and
fracture behaviour of high manganese austenitic steel”, Journal
of Materials Processing Technology 147 (2004) 145–154.
[3] C. Gomes, O. Onipede1, M. Lovell: “Investigation of
springback in high strength anisotropic steels”, Journal of
Materials Processing Technology 159 (2005) 91–98.
[4] E. Bayraktar, S. Altintas: “Square Cup Deep and bending
Drawing of Hadfield steel”, Journal of Materials Processing
Technology 60 (1996) 183-190.
[5] I. R. Lazim, J. A. Nemes: “Anisotropy and spring back in
draw-bending of stainless steel 410: experimental and
numerical study”, Journal of Materials Processing Technology
166 (1) (2005) 116–127.
[6] F. Pourboghrat, E. Chu: “Springback in plane strain stretch
/draw sheet forming”, Int. J. Mech. Sci. 36 (3) (1995) 327–341.
[7] M. Samuel “Experimental and numerical prediction of
springback and side wall curl in U-bending of anisotropic
sheet metals”, Journal of Materials Processing Technology
105 (2000) 382–393.
[8] H.B. Mullan “Improved prediction of springback on final
formed components”, Journal of Materials Processing
Technology 153–154 (2004) 464–471.
4.conclusions
references
Acknowledgements

Inglebert . Experimental results Based on experimental study. b and c) and springback values ( 1 . E.Journal of Achievements in Materials and Manufacturing Engineering Volume 18 Issue 1-2 September–October 2006 Table 1.6 kN at the test temperature of 20 + 2°C for each steel.20 Yield Strength (MPa) 170 420 ± 30 Fracture Strength (MPa) 312 1350 Strain until fracture 0.257 exp 194 T exp 0.1.32 0. Bayraktar.15 µm for Hadfield steel. Typical measurement method of springback values was explained formerly [4].2585 p 1161.45 kN and 19. Figure 2 gives the details of the geometry studied for the experimental and numerical analyses. xperimental results E 3.78 1. The punch travel was selected at a specified distance of 70mm and throughout the Fig. Experimental stamping apparatus and schematic drawing detail of system The lower springs hold the blank holder at a small distance from the die. Once the required blank holder force was reached.450 0. esults and discussion r 3.28 µm.235 0. the upper platen is pressed against the blank holder. the blank holder forces (BHF) were maintained at 2.45-0. The average roughness values measured on the blank sheets was 1.2x35x350 mm and 0. Schematic description of measurement positions for springback and sidewall curls are given in the Figure 3. After carefully placing the blank on the die. while monitoring the force.95 Constitutive equation 565.0 mm. This type of geometry was chosen for isotropic springback analyses. Materials properties Sheet Material thickness (mm) Mild Steel Hadfield Steel 0.01 experiments. The average roughness values of the different tool parts were measured at 0.20 Strain hardening Exp(n) 0. The clearance between the punch and the die was measured as 1. G. Results and discussion 3.007117 0. angle between the ox and the [ab] segment and 2 .exp 237 T 0.04 µm for Mild steel and 1. 276 Short paper T. the upper platen was tightened to the rods. springback results were evaluated and these results were used in order to validate the proposed finite element calculation. angle between the [ab] and [ef] segments) of the parts were determined. After unloading of the hat sections. the radius of a circle through a. Before each experiment both sides of the blank sheet surface were wiped with a paper towel dipped in the special lubricant and were then kept in a vertical position for 30 minutes for homogenization.66 mm/sec. so that the blank can be introduced easily.50 ± 0. All of experimental study was conducted on the INSTRON 8501 machine. 2. The punch speed was 1. Measurements parameters for the springback and side wall curl measured and used later in the finite elements model [4] 3. The die and lower platen were fixed to the cross head after alignment with the punch which is attached to the load cell. Da Sisva Botelho. The initial sizes of Hadfield and mild steel sheets are 1. Care was exercised to keep the blank holder force as close to the specified value as possible. The upper springs are used for the transmission of the blank holder force. 1. Fig. the sidewall curls ( .78x35x350 mm respectively.01 0. A standard U-shaped geometry was selected. 2D-draw bending tool was made from tool steel (D2) and hardened to 55 HRc.1.

Schematic description of measurement positions for springback and sidewall curls The calculation of the springback and the sidewall curls were performed according to Figure 2. 2D-draw bending specimens after springback for a Mild steel strip with a BHF of 19. It appears that.45 and 19. The elastically . 4c) Springback parameters obtained from experimental results were used to compare with those of final element analysis in order to better understand the phenomenon of springback in these materials that they will be detailed in the following section. Figure 5 shows the basic arrangement of the components considered in FEM model. According to the dimensions of the blank in the out-of-plane direction.6 kN has been carried out for two steel blank strips with an orientation of 90° with respect to the rolling direction. They were prepared in the rolling direction. only half of the process was modelled (Figure 5).6 kN (Fig. a two-dimensional plane strain model was used.6 (°) 94 100 115 115 1 b) c) Fig. The blank strip is squeezed between the blank holder and the die trough a normal load applied on the (°) 85 85 75 75 2 At the first approach. The problem consists of the surface contact between steel blank strip and the tools such as the punch.6 kN (Fig. die and blank holder that is a basic aspect of the stamping operations. The mean values of experimentally measured parameters taken on four measurements on each side of the blank sheets during the tests were evaluated for Hadfield and Mild steel respectively. Experimentally measured parameters of springback Material MS HFS BHF 2.66 mm/sec. In order to truthfully validate materials parameters. The detail discussion on the experimental and finite element results will be given in the full paper.6 2. The results obtained for both of steel blank strips in rolling direction were given in Table 2. In other words. In the same way. the higher the yield strength the greater the elastic return which results in higher springback values. a) Fig.driven change of shape of metallic sheets have been simulated with ABAQUS code.6 kN for all the blank strips and a constant punch speed of 1. Finite elements analysis (FEA) (mm) 112 120 103 102 A 2D numerical analysis of the U-shaped bending process was carried out for comparison with experimental results.2. springback values occur to be Comparison of experimental and simulation results of 2D-draw-bend springback 277 . 2D-draw bending specimens from Mild steel strips after removing from die are also given in Figures 4b and 4c. 4a) and for a Hadfield steel strip with a BHF of respectively 19.45 19.6 kN is shown in Figure 4a. anisotropy does not contribute to springback in our experimental conditions.2. There were no considerable differences in the results between two orientations of 0° and 90° from the rolling direction found for both of the steel blank strips. The tools can be modelled as rigid surfaces because they are much stiffer than the blank. However. Due to plane symmetry. This case can be explained for Hadfield steel because it is an isotropic material but for Mild steel. Table 2. In all experiments performed for two steels.45 kN (Fig. In fact. both analyses were accomplished with similar operational conditions. A picture taken from U shaped Hadfield strip at the BHF of 19. This consisted of two constant blank holder forces of 2. This point will be clarified using finite element analysis in the next section. the blank holder force does not influence springback values for Hadfield steel. these values in mild steel indicate a very strong tendency to increase at higher blank holder forces. 3. higher springback values of Hadfield steel than that of mild steel is due to yield property of Hadfield steel. experimental results showed that there is no significant difference between springback values of Hadfield steel for two different blank holder forces. As a result of unbending. 3. 4b) and 2. (a quasi static displacement was performed). It is well known that this steel has higher yield strength and also as very high toughness properties. the FEA was expected to be independent of rolling direction angle due to the yielding isotropy within the plane of the sheet. Some of the experimental and finite element results are given in Table 3 and it seems to show a good correlation. 4. inite elements analysis (FEA) F 3. Therefore its yield strain is very high. Only one test at the BHF of 19.Analysis and modelling significantly related to the yield strain level.45 19. sidewalls do not fail by tearing as blank holder force increases. This case can be explained knowing that this steel work hardens very quickly as a function of plastic deformation level.

Lazim. Journal of Materials Processing Technology 60 (1996) 183-190. D.2020 1 (°) 85 86. suppress the BHF and remove the blank holder and the die.144. E. it means that the higher the yield strength the greater the elastic springback values. L.journalamme. Nemes: “Anisotropy and spring back in draw-bending of stainless steel 410: experimental and numerical study”. Mullan “Improved prediction of springback on final formed components”. E. Mech. Levaillant: “Deformation and fracture behaviour of high manganese austenitic steel”. For the sake of the simplicity of the FEA.8 0.6 kN Experimental Numerical Angle Comparison (%) Mean values results (°) 99 98. F. Geometrical description of the simulation model The tendencies form the experimental and FEA are in well agreement with tendency in the yield strength. These results validate the FE approach as a trustworthy tool for predicting the springback parameters for a given set of stamping conditions and material properties. 4. The unloading procedure consists in totally removing the punch. Bayraktar. Lovell: “Investigation of springback in high strength anisotropic steels”. onclusions c 4. Table 3. Journal of Materials Processing Technology 159 (2005) 91–98. 36 (3) (1995) 327–341.K. R.M. [8] 278 Short paper READING DIRECT: www. Altintas: “Square Cup Deep and bending Drawing of Hadfield steel”. S. Carden.D.org . in conjunction with the friction between the blank and blank holder and the blank and die. Wagoner: “Measurement of springback”. the following assumptions were considered: An elasto-plastic constitutive model was accepted with material isotropy and with a non-linear strain hardening condition. Chu: “Springback in plane strain stretch /draw sheet forming”.Journal of Achievements in Materials and Manufacturing Engineering Volume 18 Issue 1-2 September–October 2006 blank holder while the die remains always unmoving. Bayraktar. Khalid. Journal of Materials Processing Technology 147 (2004) 145–154. I. leading to a springback phenomenon. Int. references References [1] [2] [3] W. Journal of Materials Processing Technology 105 (2000) 382–393. Gomes. R. Onipede1.2353 2(°) The FEM took into account the forming process then the full unloading process.B. Comparison of experimental and numerical results for a Mild steel bent with BHF of 19. Journal of Materials Processing Technology 153–154 (2004) 464–471. Punch Rp Rd Blank Holder Die [4] [5] Sheet Symmetry plane [6] [7] Fig. H. Journal of mechanical sciences 44 (2002) 79–101. Montataire-France) for their support in this project. C. J. 5. M. O.H. This force. F. The contact between the punch and the blank was supposed to be frictionless whereas the contacts between respectively the blank and the die and the blank and the blank holder were supposed to have a coulomb friction law with a friction coefficient of 0. controls how the blank material is drawn into the die during the forming process. Sci. The finite element results have been validated from experimental results. U-Shaped bending deformation was assumed to take place under the plane strain condition. Int. Geng. A. Acknowledgements Acknowledgements The authors are grateful to ARCELOR research (CED – Arcelor Auto.9 2. J. Journal of Materials Processing Technology 166 (1) (2005) 116–127. Samuel “Experimental and numerical prediction of springback and side wall curl in U-bending of anisotropic sheet metals”. This model is a simple and an efficient way to take apart the responsibilities of the steel maker for the selection of steel to be used and sheet stamping designer for the selection of the forming conditions to work in a safety zone in order to increase the reliability of the stamped structures. The forming process was controlled by the displacement of the punch. C. Conclusion Experimental and finite element analyses have been conducted on blank strips of Hadfield steel (high strength steel) and Mild steel with similar operational conditions. Matlock. Pourboghrat. E. M. A.