1364

Materials Transactions, Vol. 51, No. 7 (2010) pp.
1364 to 1366
#2010 The Japan Institute of Metals
EXPRESS RAPID PUBLICATION
Bauschinger Eect on Springback of Clad Sheet Metals in Draw Bending

K. Yilamu* , R. Hino, H. Hamasaki and F. Yoshida
Department of Mechanical System Engineering, Hiroshima University, Higashi-Hiroshima 739-8527, Japan
For clad sheet metals, springback behavior is directly related to the elastic-plastic properties such as Youngs modulus, the yield strength,
workhardening and the Bauschinger eect of each metal layer. The aim of this study is to understand the springback characteristics of the clad
sheet metals in draw bending based on the accurate FE simulation corresponding to experimental data on stainless-steel clad aluminum sheets. In
the discussion, the special emphasis is placed on the inuence of the Bauschinger eect. For that purpose, in the simulation, two types of material
models, one is the classical isotropic hardening model (no Bauschinger eect included) and the other Yoshida-Uemori kinematic hardening
model which describes the Bauschinger eect of materials properly, were used and these numerical results were compared.
[doi:10.2320/matertrans.M2010058]
(Received February 18, 2010; Accepted April 7, 2010; Published May 19, 2010)
Keywords: clad sheet metals, draw bending, springback, Bauschinger eect, nite element simulation, constitutive model
Introduction
Many types of clad sheet metals are increasingly used in a

wide area of industries because of their excellent mechanical
and functional properties.1) However, press forming of clad
sheet metals is a dicult task, because their elasto-plastic
behavior is dierent from that of their individual component
sheets.25) Especially, their bending and springback characteristics are quite dierent from those of monolithic sheets
because their mechanical properties change in thickness
direction.610)
The aim of this study is to understand the springback
characteristics of clad sheet metals in draw bending based on
the FE simulation. In the discussion, special emphasis is
placed on the inuence of the Bauschinger eect on the
springback. For the accurate description of the Bauschinger
eect, Yoshida-Uemori kinematic hardening model (Y-U
model11,12)) was employed.
Since draw bending is a process of cyclic plasticity
deformation induced by stretch bending-unbending, springback behavior of clad sheet metals is much more complicated
compared to a case of simple V-shaped air bending.10) In
draw bending the Bauschinger eect appears both in the
processes of stretch unbending and the subsequent springback, while in air bending it will takes place only during
springback but it is negligible in bending process.
2.
Materials and Experimental Procedures
The experimental data on the springback of stainless-steel

clad sheets after draw bending reported previously by the
present authors10) are used for the discussion. The mechanical
properties of the sheets and the experimental procedures are
summarized as follows.
Two types of stainless-steel (430SS) clad aluminum
(A1100) sheet metal, ts =t 0:32 and 0.49, were used as
shown schematically in Fig. 1. In order to determine the
mechanical properties of each component metal layer, two
types of experiments, i.e., uniaxial tension and tension*Graduate
Student, Hiroshima University
t
ts /t
1.47 0.32
1.55 0.49
Fig. 1
Aluminum layer (A1100)
ts
Stainless-steel layer (430SS)
Schematic illustration of clad sheet metal (in mm).
600
430SS
400
True stress, /MPa
1.
200
A1100
0
Experiment
Experiment
-200
-400
Permanent
stress off-set
-600
-0.05
(Rule of mixture)
Y-U model
IH model
Transient
Bauschinger effect
0.05
0.1
True strain,
0.15
Fig. 2 Stressstrain curves for the component materials of the clad sheet
metals.
compression tests, were conducted in the present work for the

clad sheet metals and the stainless steel sheets taken from the
clad sheets. Using these experimental data, the stress-strain
responses of the aluminum layer was calculated by using
the rule of mixture. The obtained stress-strain responses of
each metal layer are shown in Fig. 2. As it is seen in this
gure, both aluminum and stainless steel exhibit apparent
Bauschinger eect. Especially for stainless steel, it should
be noted that the ow stress in the transient Bauschinger
region is much lower than that under monotonic deformation
(= IH model prediction). Y-U model can simulate it
accurately.
As schematically shown in Fig. 3, one end of a clad strip is
pulled downward keeping the blank-holding force constant.
In this process, the specimen is drawn over a die corner
(its radius was 6 mm), which is a process of bending and
unbending under a constant stretching force. After spring-
(a)
Stretching
force
Stretch-bending
Stretch-unbending
Punch
Specimen
res
Side-wall
of specimen
15
Die
80
Blank holder
Die
(b)
Fig. 3
Schematic illustrations of the draw bending test.
back, the residual curvature res 1=res is determined from

the measurement of the curl.
The draw bending experiments were performed for both
the cases of Alin /SSout (i.e., aluminum layer is located at the
inside of the bent clad) and SSin /Alout (i.e., stainless-steel
layer is located at the inside of the bent clad) for several
levels of stretching force.
Die radius: 6mm

ts /t: 0.32
0.004
0.003
0.002
0.001
Y-U model, SS in /Al out

Y-U model, Al in /SSout
Experiment, SS in /Al out
Experiment, Al in /SSout
0
-0.001
0.005
0.2
0.4
0.6
0.8
Stretching force/Yield force
Die radius: 6mm
ts /t: 0.32
0.004
0.003
0.002
0.001
IH model, SS in /Alout
IH model, Al in /SSout
Experiment, SS in /Alout
0
-0.001
0.2
0.4
0.6
0.8
3.1 FE model
The nite element simulation of draw bending and the
subsequent springback was performed, where the planestrain four-node quadrilateral element with four Gaussian
integration points was used. For FE meshing, the sheet is
divided into 12 layers in the thickness direction and the total
number of elements was 13368 or 13746. Since the two
layers of the clad sheet metals were rmly bonded and no
delamination at the bonded interface was found in a whole
process of draw bending and springback, a FE model of
simply xed two-layers was used. As for the friction, the
Coulomb law with the friction coecient of 0.1 was
assumed, as an empirical value for lubricated sheet metals,
both stainless steel and aluminum layers. The static implicit
FE code MARC was chosen for the simulation.
3.2 Material models
To discuss the inuence of the Bauschinger eect of
materials on the springback of clad sheet metals, we
employed two types of material models in the FE simulation,
i.e., one is the classical isotropic hardening model (IH model)
that cannot take into account the Bauschinger eect, and the
other is Yoshida-Uemori kinematic hardening model (Y-U
model11,12)) that describes the Bauschinger eect properly
(see Fig. 2). This model also includes a model of the plasticstrain dependent Youngs modulus.13) Details of modeling
are described in Refs. 1113). For the FE simulation, Y-U
model had been implemented in code MARC using the user
subroutine HYPELA2.
Results and Discussions
Figures 4(a) and (b) show the results of residual curvature

vs. stretching force for the clad sheet of ts =t 0:32, where
the calculated results by Y-U model and IH model are
depicted in (a) and (b), respectively. For this clad sheet, when
(b)
Residual curvature, res /mm-1
FE Simulation of Springback
(a)
4.
1365
Fig. 4 Residual curvature vs. stretching force (ts =t 0:32). Predictions by

(a) Y-U model; and (b) IH model, together with the experimental results.
3.
0.005
res = 1/res
15
Blank-holding force
Bauschinger Eect on Springback of Clad Sheet Metals in Draw Bending
0.007
Y-U model, SS in /Al out

Y-U model, Al in /SSout
0.006
0.005
0.004
0.003
0.002
0.001
00
Die radius: 6mm

ts /t: 0.49
0.2
0.4
0.6
0.007
0.8
IH model, SS in /Al out

IH model, Al in /SSout
0.006
0.005
0.004
0.003
0.002
0.001
00
Die radius: 6mm

ts /t: 0.49
0.2
0.4
0.6
0.8
Fig. 5 Residual curvature vs. stretching force (ts =t 0:49). Predictions by

(a) Y-U model; and (b) IH model, together with the experimental results.
Alin /SSout , springback decreases signicantly with increasing stretching force, but under the opposite sheet-set
condition, SSin /Alout , the eect of stretching force is very
weak. The amount of the residual curvature in the case of
SSin /Alout is much larger than that of Alin /SSout . As it is seen
in Fig. 4(a), Y-U model captures well these characteristics of
springback behavior. In contrast, as shown in Fig. 4(b), the
calculated results by IH model are not accurate enough,
especially for the case of SSin /Alout , the calculated residual
curvature increases with the increase of stretching force,
whereas in the experiment it slightly decreases.
Figures 5(a) and (b) show the results on the clad sheet of
ts =t 0:49. The experimental results of residual curvatures
of both Alin /SSout and SSin /Alout , for a given stretching
1366
K. Yilamu, R. Hino, H. Hamasaki and F. Yoshida
(a)
(b)
res =1/res =1/
M
F
M=0
F=0
Fig. 6 Process of stretch bending/unbending and springback. (a) Stretch

bending ! unbending; (b) Springback.
force, are almost the same, and they decrease monotonically

with the increase of stretching force. The calculated results
by Y-U model for these two cases of sheet-set conditions are
close to each other, and they agree fairly well with the
experimental data (see Fig. 5(a)). Contrary to this, IH model
gives the results showing that the springback of SSin /Alout is
apparently larger than that of Alin /SSout (see Fig. 5(b)).
To understand the mechanism of springback of clad
sheet metals in draw bending, here, we simply assume that a
sheet becomes completely at ( 0) at the nal stage of
draw bending, and springback takes place fully elastically.
Note that these assumptions are only for the following
qualitative discussion, but the present FE simulation was not
restricted by them (e.g., springback-induced plastic deformation was taken into account in the simulation). For elastic
springback, the stress change is given by the following
equation:
x
E
"o z
1 2
where, "o and denote the changes of longitudinal strain

and the curvature at the middle plane (z 0) of the clad
sheet, as schematically illustrated in Fig. 6. The residual
curvature after springback, res , is determined by
solving the following equations:
Z
Z
Z
E
E
dz

zdz 2
F x dz "o
1 2
1 2
Z
Z
Z
E
E
M x zdz "o
zdz

z2 dz
2
1
1 2
3
where, F and M are the stretching force and bending moment
at the nal stage of draw bending before springback. As it is
seen from these equations, the springback of a clad sheet is
directly related to both the stretching force F and bending
moment M. Moreover, Youngs moduli E are completely
dierent layer by layer, and this strongly aects the springback behavior.
Since draw bending is a process of tension-compression
cyclic deformation, the stresses induced by stretch unbending
are greatly inuenced by the Bauschinger eect of materials.
Especially for a two-ply clad sheet, neutral surface shifts to
the stronger layer side, cyclic strain range of its stronger layer
becomes considerably smaller than that of its weaker layer
(refer to Ref. 14)), and consequently, springback is greatly
inuenced by the transient Bauschinger eect behavior of
the stronger layer. As shown in Fig. 2, the ow stress in the

transient Bauschinger eect region is considerably lower
than that under monotonic deformation, the strength dierential between the stronger layer (stainless steel) and the
weaker layer (aluminum) in the stretch unbending process is
smaller than that predicted by IH model. Since the shift of
neutral axis is remarkable for a combination of a thin strong
layer and a thick weak layer, especially for such a case,
accurate prediction of springback would not be possible
unless the Bauschinger eect is taken into account (see
Fig. 4(b) for ts =t 0:32 in the case of SSin /Alout ). For a clad
sheet metal, the re-yielding is more likely to occur during
springback in the weaker layer (aluminum, in the present
work), since the driving force of the springback (bending
moment M and stretching force F) is mainly determined by
the strength of the stronger layer, which is extremely large for
the weaker layer. Consequently, the Bauschinger eect will
appear during the unloading (springback) process, and it also
aects the amount of springback.
5.
Concluding Remarks
Springback behavior of two-ply clad sheet metals in draw

bending, especially the eect of stretching force, is greatly
inuenced by the transient Bauschinger eect of the
component materials. This eect appears more clearly for a
combination of a thin strong layer and a thick weak layer,
since in such a case cyclic strain range of the stronger layer
becomes considerably smaller than that of the weaker layer
because of the shift of the neutral surface during bendingunbending. Hence, the use of a material model for the
Bauschinger eect is of vital importance for an accurate
prediction of springback of clad sheet metals.
REFERENCES
1) J. K. Kim and T. X. Yu: J. Mat. Process. Technol. 63 (1997) 3342.
2) R. Hawkins and J. C. Wright: J. Inst. Metal. 99 (1971) 357371.
3) N. L. Semiatin and H. R. Piehler: Metall. Trans. A 10 (1979) 1107
1118.
4) T. Mori and S. Kurimoto: J. Mater. Process. Technol. 56 (1996) 242
253.
5) F. Yoshida and R. Hino: J. Mater. Process. Technol. 63 (1997) 6671.
6) H. Verguts and R. Sowerby: Int. J. Mech. Sci. 17 (1975) 3151.
7) S. A. Majlessi and P. Dadras: Int. J. Mech. Sci. 25 (1983) 114.
8) R. McMurray, A. Leacock and D. Brown: Key Eng. Mater. 340341
(2007) 853858.
9) K. Yilamu, R. Hino, H. Hamasaki and F. Yoshida: J. Mater. Process.
Technol. 210 (2010) 272278.
10) R. Hino, Y. Goto and F. Yoshida: J. Mater. Process. Technol. 139
(2003) 341347.
11) F. Yoshida and T. Uemori: Int. J. Plastic. 18 (2002) 661686.
12) F. Yoshida and T. Uemori: Int. J. Mech. Sci. 45 (2003) 16871702.
13) F. Yoshida, T. Uemori and K. Fujiwara: Int. J. Plastic. 18 (2002) 633
659.
14) F. Yoshida, M. Urabe, R. Hino and V. V. Toropov: Int. J. Plastic. 18
(2003) 21492170.

1364

Uploaded by

Copyright:

Available Formats

1364

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1364

Uploaded by

Copyright:

Available Formats

Materials Transactions, Vol. 51, No. 7 (2010) pp.

EXPRESS RAPID PUBLICATION

Bauschinger Eect on Springback of Clad Sheet Metals in Draw Bending

Many types of clad sheet metals are increasingly used in a

Materials and Experimental Procedures

The experimental data on the springback of stainless-steel

Student, Hiroshima University

Aluminum layer (A1100)

Stainless-steel layer (430SS)

Schematic illustration of clad sheet metal (in mm).

compression tests, were conducted in the present work for the

Schematic illustrations of the draw bending test.

back, the residual curvature res 1=res is determined from

Die radius: 6mm

Y-U model, SS in /Al out

Figures 4(a) and (b) show the results of residual curvature

Residual curvature, res /mm-1

Fig. 4 Residual curvature vs. stretching force (ts =t 0:32). Predictions by

Residual curvature, res /mm-1

Residual curvature, res /mm-1

Bauschinger Eect on Springback of Clad Sheet Metals in Draw Bending

Y-U model, SS in /Al out

Die radius: 6mm

IH model, SS in /Al out

Die radius: 6mm

Fig. 5 Residual curvature vs. stretching force (ts =t 0:49). Predictions by

K. Yilamu, R. Hino, H. Hamasaki and F. Yoshida

res =1/res =1/

Fig. 6 Process of stretch bending/unbending and springback. (a) Stretch

force, are almost the same, and they decrease monotonically

where, "o and denote the changes of longitudinal strain

the stronger layer. As shown in Fig. 2, the ow stress in the

Springback behavior of two-ply clad sheet metals in draw

You might also like