Journal of Materials Processing Tech.

254 (2018) 283–293

Contents lists available at ScienceDirect

Journal of Materials Processing Tech.

journal homepage: www.elsevier.com/locate/jmatprotec

Formability of sheet metal flowing through drawbead–an experimental T

investigation
⁎
Junyi Kea, Yuqi Liua, Hongchuan Zhub, Zhibing Zhanga,
a
State Key Laboratory of Materials Processing and Die & Mould Technology, Huazhong University of Science and Technology, Wuhan 430074, PR China
b
Research and development center of Wuhan iron and steel (group) corp, Wuhan 430074, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: A phenomenon that when the sheet metal flows through drawbead, the limit strains of sheet metal lie above the
Sheet metal forming forming limit diagram indicating fracture in theory, while it is safe in practice, violates the normal forming limit
Formability curve has been found. In order to explain the phenomenon, a series of tests were conducted. In the first part, the
Drawbead conventional uniaxial tensile test was employed as a reference. In the second part, specimens were pulled
Pre-strain
through various drawbead inserts several times, and then cut to the uniaxial tensile test specimens and tested by
the conventional uniaxial tensile test. In the third part, specimens were pulled through drawbead inserts many
times until the fracture occurs. The results indicate that the percentage of elongation of pre-strained specimens
with several times are smaller than that of the conventional uniaxial tensile test. The transition process from the
plane strain path to the uniaxial tensile strain path leading to the additional strain hardening is the main reason
for the reducing formability of sheet metal. While the percentage of elongation of specimens pulled through
drawbead inserts many times until the fracture occurs are evidently higher than that of the conventional uniaxial
tensile test and could reach close to 100%. Bending and unbending under tension has a significant effect on the
increased percentage of elongation.

1. Introduction
The formability of sheet metal is the capability which undergoes
plastic deformation to a given shape without defects (Banabic et al.,
2010). The most commonly used tool in assessing the formability of
sheet metal is the forming limit diagram (FLD) developed by Keeler and
Backofen (1963) and Goodwin (1968), which is composed of the
maximum values of the principal strains determined by measuring the
strains at failure, widely used in factories and research laboratories
(Stoughton and Zhu, 2004). The forming limit curve (FLC) is the line
consisting of limit principal strains separating the FLD into two regions.
The deformed sheet metal is defined as failure if strain combinations lie
above the FLC, otherwise the deformed sheet metal is treated as safe
(Nurcheshmeh and Green, 2016), as shown in Fig. 1.
However, a phenomenon that when the sheet metal flows through
drawbead, the limit strains of sheet metal lie above the forming limit
diagram indicating fracture in theory, while it is safe in practice, vio-
lates the normal forming limit curve has been found after testing a large
number of automobile panels. The test was done by stamping engineers
in Faw Mould Manufacturing co.LTD. The mostly used material of the
test is BH180 and the percentage of elongation is 29%. One of the test
pieces is the A-side of Audi A6 fender. When the test piece was stamped
especially during the process of passing through the drawbead set, the
major strains on the A surface of the piece were bigger than areas where
the material has not passed through drawbead, such as the flatter part
of the test piece shown in Fig. 2. The final thinning in thickness di-
rection in that area where the sheet metal flows through drawbead
inserts is close to 50%, which means the thinning in thickness direction
is very serious, but the sheet metal is very safe in practice. By numerical
simulating the forming process of the sheet metal, the forming limit
strains of that area is also very high, above the normal FLC.
As we know, the FLC is only valid under following conditions: (1)
straight or proportional strain path; (2) absence of bending; (3) without
through-thickness shear; (4) no normal stress or through-thickness
stress (Emmens and van den Boogaard, 2009). When the conditions
above are not satisfied, the FLC is not always effective (Stoughton,
2000).
In terms of strain paths, Laukonis and Ghosh (1978) conducted an
experiment on aluminum-killed steel and 2036-T4 aluminum by pre-
straining in balanced biaxial tension. The FLC of aluminum-killed steel
decreased while the position of FLC of 2036-T4 aluminum just moved
horizontally with the pre-strain increasing. Graf and Hosford (1993)

⁎
Corresponding author.
E-mail address: zhangzb@hust.edu.cn (Z. Zhang).

https://doi.org/10.1016/j.jmatprotec.2017.11.051
Received 2 July 2017; Received in revised form 22 November 2017; Accepted 24 November 2017
Available online 26 November 2017
0924-0136/ © 2017 Elsevier B.V. All rights reserved.
J. Ke et al. Journal of Materials Processing Tech. 254 (2018) 283–293

punch curvature on the stretching limit of AKQD steel and the limit
strains increased with the increasing punch curvature at constant ma-
terial thickness. Tharrett and Stoughton (2003) conducted a series of
stretch bending tests with various curvature punches and the limit
strains of small radius were larger and the position of FLC was higher.
Kitting et al. (2009) determined the forming limit of H340LAD, CP800,
and DP800 under a range of punch radius and deduced that bending
had an influence on the forming limit. With the decreasing of punch
radius, the forming limit became higher, especially when the punch
radius was below 10 mm. Emmens and van den Boogaard (2008) in-
troduced the bending into the conventional tensile test on DC04 and
DC06 with different thicknesses and discovered the percentage of
elongation was increased greatly. Fictorie et al. (2010) conducted the
Nakazima test with four kinds of radii of punches on DC06 and AA5051,
and the forming limit increased with decreasing of radius of punch,
especially in the plane strain region. The effect of bending on forming
limit of sheet metal cannot be neglected.
Fig. 1. Schematic diagram of the forming limit curve. In addition, the forming limit can be increased if the through-
thickness stress or through-thickness shear involves. Smith et al. (2003)
put forward a new sheet metal formability model considering the
carried out an experiment on Al 2008-T4 pre-strained in uniaxial,
through-thickness stress and proved that the formability increased with
biaxial and plane-strain tension parallel and perpendicular to the
the through-thickness stress. Assempour et al. (2010) predicted the
rolling direction and demonstrated that pre-straining in biaxial tension
forming limit diagram taking into account the normal stress and ver-
lowered the entire FLC, while pre-straining in uniaxial tension raised
ified that the FLD shifted up when the normal stress increased. Zhang
the right side of FLC. Pre-straining in plane strain tension raised both
et al. (2014) established constitutive models with M-K methods and
sides away from the minimum of the FLC. Then Graf and Hosford
different yield criterions to predict the effect of through-thickness stress
(1994) continued an experiment on Al 6111-T4 and found that if the
on the FLC and the FLC enhanced with increased through-thickness
directions of the principal strains were rotated, pre-straining in uniaxial
stress. Allwood and Shouler (2009) proposed a new generalized
or plane strain tension lowered the FLC besides the aforementioned.
forming limit diagram allowing for the normal stress and through-
That is to say, changing strain paths has an influence on the formability
thickness shear stress and demonstrated that the forming limits can be
of sheet metal.
increased significantly by both normal compressive stress and through-
The most commonly used method in experiment determination of
thickness shear. What’s more, Eyckens et al. (2009) established a model
the FLD is Nakazima test (out-of-plane) and Marciniak test (in-plane),
considering the effect of through-thickness shear with M-K method by
acquiring strains from uniaxial tensile test to plane strain and then to
investigating the incremental forming, and the results showed that
biaxial tension by testing specimens with different widths (Panich et al.,
formability was increased for all in-plane strain modes if through-
2013). However, the bending effect is ignored to meet the condition of
thickness shear was present.
membrane forces in the tests. Ghosh and Hecker (1974) compared the
On the other hand, the drawbead has been widely used in sheet
in-plane stretching with the out-of-plane stretching and found that the
metal forming processes, especially in automotive industry and elec-
out of plane stretching produced larger limiting strains under identical
trical appliances (Kim et al., 1997). By setting appropriate geometric
degree of biaxial tension. Charpentier (1975) investigated the effect of

Fig. 2. The phenomenon that when the sheet metal flows through drawbead, the forming limit strains of the sheet metal are above the FLC, which illustrates serious fracture in theory,
while it is very safe in practice. (a) the test piece (A-side of Audi A6 fender) in practice. (b) the forming limit strains of the test piece are above the normal forming limit curve. (c) the
simulation result of the stamping process of the test piece (A-side of Audi A6 fender).

284
J. Ke et al. Journal of Materials Processing Tech. 254 (2018) 283–293

Table 1
Chemical composition of WDQ steel by weight percentages.

C Si Mn P S Cu ALT ALS

0.0032 0.0061 0.145 0.01 0.0088 0.016 0.0289 0.0266

Ni Cr Ti Mo V Nb N B
0.006 0.01 0.0677 0.001 0.001 0.003 0.0029 0.0032

Table 2
Fig. 3. Schematic diagram of the sheet metal flowing through semicircle drawbead in- Mechanical properties of WDQ steel.
serts. Symbol meaning: R: radius of male bead, H: height of male bead, r1: entrance radius
of contra bead, r2: exit radius of contra bead. Number 0−1 0−2 45−1 45−2 90−1 90−2

Yield strength σy(Mpa) 165 166 171 171 168 166

shapes and positions of drawbeads which control the restraining force Tensile strength σu(Mpa) 318 321 320 321 315 312
and the material flow, the sheet metal formability can be improved to a Anisotropy index r 2.025 2.074 2.068 2.014 2.669 2.62
large extent (Firat, 2007). Compared with the method of controlling the Hardening exponent n 0.24 0.241 0.233 0.234 0.234 0.233
Hardening coefficient K 571 578 570 571 561 565
blank holder, drawbeads are relatively easy to adjust and also im- Young’s modulus E(Gpa) 172 185 202 203 206 183
portant design parameters since strong restraining forces can prevent Total Elongation At(%) 44.4 43 43.6 43 45.4 45.6
the sheet metal from draw-in and free from the necking, wrinkling and
other quality problems (Sun et al., 2010). Note: (1) 0–1 means the first specimen in the rolling direction, 0–2 means the second
Drawbeads have different geometrics, such as semicircle, square, specimen in the rolling direction. (2) 45 means the diagonal direction to the rolling di-
rection, 90 means the perpendicular to the rolling direction.
triangle, and trapezoid (Ghoo and Keum, 2000). Semicircle drawbead is
the most frequently used type, as shown in Fig. 3, which consists of a
semi-cylinder protrusion on the blankholder called male bead and a 2.1. Conventional uniaxial tensile test (as a reference)
groove with quarter-cylinder shoulders on the external part of the die
called contra bead. When the sheet metal flows through the drawbead, The blank material is WDQ steel of 0.6 mm thickness with good
it will suffer from bending, unloading and unbending around the entry ductility, widely used automotive steel provided by Research and de-
groove shoulder (from point A to point C), and repeated at the male velopment center of Wuhan iron and steel (group) corp. The chemical
bead (from point C to point D) and at the exit groove shoulder (from composition of WDQ steel by weight percentages is listed in Table 1.
point D to point F). These bending and unbending deformations to- The mechanical properties of WDQ steel tested on Zwick Z050 elec-
gether with the frictional force constitute the drawbead restraining tronic tensile testing machine are given in Table 2.
force, which provides the deformation of sheet metal (Courvoisier et al.,
2003).
2.2. Drawbead test (nine groups of drawbead inserts, passing through one
The process of sheet metal flowing through drawbeads can be seen
time, two times and three times)
as the cycle of bending, unloading and unbending with tensile forces to
some extent (Samuel, 2002). Strong strain hardening has an effect on
The drawbead test machine of sheet metal forming was utilized to
the sheet metal, as well as other factors. As a consequence, the prop-
conduct the drawbead test, specially designed by the cooperation of
erties of sheet metal may be different from the initial state, including
Research and development center of Wuhan iron and steel (group) corp
the yield strength, tensile strength and percentage of elongation, be-
and Beijing university of aeronautics and astronautics. As shown in
tween the specimens flowing through drawbead and the primitive
Fig. 4, the drawbead test machine mainly consists of three parts. The
specimens.
first part is the clamping unit, containing a serated jaw unit and a
Therefore, to understand the phenomenon of forming limit change
moving unit with force extensometer and displacement extensometer
when the sheet metal flows through drawbead inserts described above,
where the pulling force and displacement of specimen were recorded
a series of tests are conducted. In the first part, the conventional uni-
when the specimen was pulled through. Unlike the conventional uni-
axial tensile test was employed as a reference. In the second part,
axial tensile test with two sides of the specimen clamped, the specimen
specimens were pulled through various drawbead inserts several times,
was clamped with one side by the clamping unit in the drawbead test.
then cut to the uniaxial tensile test specimens and tested by the con-
The second part is the black box within which the drawbead grooves
ventional uniaxial tensile test. In the third part, specimens were pulled
including the male bead and the contra bead are only 70 mm width,
through two groups of drawbead inserts many times until the fracture
with one groove stationary and another moving when adjusting the gap
occurs.
between the male bead and the contra bead, driven by a set of hydraulic
transmission device, shown in Fig. 5. The processed drawbead inserts
2. Experiment procedures
were put into the grooves when tested. The third part is the computer
controlled unit. The pulling force, displacement, gap and velocity were
As described in the introduction, the experiment can be described
controlled by a series of electronic equipment and hydraulic equipment.
briefly below.
The maximum normal force of the machine is 14,000 N, while the
pulling force is 30,000 N, considering the resistance force is quite large
(1) conventional uniaxial tensile test (as a reference);
when the specimen flows through drawbead inserts, only mild steel and
(2) drawbead test (nine groups of drawbead inserts, passing through
aluminum with relatively small thickness can be selected, and that’s
one time, two times and three times);
why we choose WDQ steel. Of course, We have also tested a few of
(3) pre-strained specimens of (2) + conventional uniaxial tensile test;
materials (DC03, DC04) with the thickness of 1.0 mm and 1.2 mm and
(4) drawbead test (two groups of drawbead inserts selected, passing
the results are the same and valid for thicker materials certainly. Here
through many times until the fracture occurs).
we mainly demonstrate the results of WDQ with 0.6 mm thick.
Nine groups of semicircle drawbead inserts with three heights of
male bead and three radii of entrance and exit of contra bead, with
invariable radius of male bead, were processed, as the semicircle

285
J. Ke et al.
drawbead is the most widely used drawbead type, shown in Table 3.
The length of the specimen is designed 250 mm, with about 50 mm
griped by the clamping unit and the rest pulled through drawbead in-
serts, and the width of specimen is designed as 30 mm, as the width of
drawbead groove is only 70 mm.
In order to measure the strains of specimens flowing through var-
ious drawbead inserts precisely, 2.5 mm square grids were printed on
one surface of the specimen using electrochemical corrosion method.
Before that, the specimens were degreased and cleaned with alcohol.
The Auto Grid Com-smart device was used to take pictures of deformed
grids on the surface of the specimen when the specimen has been pulled
through drawbead inserts, and the length stain, width strain and
thickness strain were measured by software analysis. The principal of
the Auto-Grid Com-smart is the constant volume assumption.
In general, one group of drawbead inserts was put into the grooves
of the black box and one specimen with one surface printed 2.5 mm
square grids was gripped by the clamping unit of drawbead test ma-
chine. Adjusting the clamping unit down to the location about
30 ± 3 mm to the upper surface of the black box in Fig. 5(a) and
setting the gap between the male bead and the contra bead close to 1.1 t
to make sure the specimen was under bending. Setting the pulling speed
2.1 mm/s, and then pulling the specimen through drawbead inserts by
driving the hydraulic equipment. When the specimen was pulled
through drawbead inserts, the grids on the surface of the specimen were
deformed. The strains of the specimens were acquired by using the Auto
Grid Com-smart device. By changing the groups of drawbead inserts
and adjusting the number of times of the specimens passing through,
different pre-strained specimens were obtained.
2.3. Pre-strained specimens + Conventional uniaxial tensile test
The pre-strained specimens above were wire-electrode cut into the
uniaxial tensile test samples. Conventional uniaxial tensile test was
employed again on Zwick Z050 electronic tensile testing machine. Basic
mechanical properties, including yield strength, tensile strength, and
286
Journal of Materials Processing Tech. 254 (2018) 283–293
Fig. 4. The drawbead test machine of sheet metal forming equipment,
including the clamping unit, the black box as well as the hydraulic
transmission device and the computer controlled unit.
percentage of elongation were obtained.
2.4. Drawbead test (two groups of drawbead inserts selected, passing
through many times until the fracture occurs)
The test method was referring to the description of Section 2.2. The
variations were two folds. One was that two groups were selected from
the nine groups of drawbead inserts. Another was that the specimen has
to go through drawbead inserts many times until the fracture occurs
rather than several times.
Two problems occurred when the test was done. One was that the
grids on the surface of the specimens suffered from seriously abrasion
after many times, leading to the difficulty in measuring the strains by
Auto Grid Com-smart, as shown in Fig. 6(a). Another was that the
clamped part of the specimen ruptured limiting the further test, as
shown in Fig. 6(b).
To measure the length strains of specimens which suffered from
seriously abrasion, we measured the length of 30 grids, 40 grids and 50
grids manually, and then the final length strain was obtained by aver-
aging them. Moreover, the zone selected must be very careful because
the length strain is not uniform which will be described in the following
part. As shown in Fig. 6(c), the selected zone was almost in the middle
part of the deformed specimens. In order to solve the fracture of
clamped part of the specimen, about 5 mm or 10 mm were adjusted to
the normal position by moving up or down every time when the spe-
cimen was clamped to make sure the clamping part of the specimen is
not always at the same place.
3. Results and discussion
3.1. Drawbead test (one time, two times and three times)
The strains of specimens were measured by Auto Grid Com-smart,
including the length strain, width strain and thickness strain, as shown
in Fig. 7. It depicts that the width strain is almost unchanged during the
J. Ke et al. Journal of Materials Processing Tech. 254 (2018) 283–293

Fig. 5. The black box of the drawbead test machine.

(a): the enlarged view of black box; (b): the partial
enlarged detail of the male bead container and the
transmission shaft; (c): the partial enlarged detail of
the contra bead container and the male bead and the
contra bead; (d) the specimen with one side clamped
is pulled through the drawbead inserts.

Table 3 specimen going on, the resistance force tends to be relative stable,
Geometrical parameters of nine groups semicircle drawbead inserts. leading to the length strain and the thickness strain decreasing in some
degree. However, the length strain and the thickness strain are not
Name Male bead Contra bead
uniform in stable area because the lubrication is not uniform in the
Radius Height Entrance radius Exit radius operation resulting in the resistance force fluctuating.
To contrast the pre-strains caused by various drawbead inserts and
Symbol R H r1 r2 the number of times. Here we chose the average length strain of rela-
Value 6 4, 5, 6 1.5, 3, 4.5 1.5, 3, 4.5
tively stable area of the specimen as a criterion. The relationship be-
Note: (1) the units of radius, height, entrance radius and exit radius are mm; (2) the tween the pre-strains in length direction and the drawbead inserts
entrance radius and exit radius are the same of one group drawbead insert. (3) nine containing three groups of heights of male bead and three groups of
combinations (three groups of Heights and three groups of Entrance radius and Exit ra- radii of entrance and exit of contra bead and the number of times when
dius, namely, 3 × 3) are studied. (4) three repeat experiments are done for each test. the specimen was pulled through is shown in Fig. 8. It shows that with
the increase of height from 4 mm to 6 mm, the pre-strains are in-
process when the specimen is pulled through drawbead inserts, and creasing with no relation to the radius and the number of times. Under
compared with the length strain and thickness strain, the width strain is the conditions of same entrance and exit of contra bead and the number
very small, close to zero, which means the process of specimen flowing of times, the increase of height of male bead means the specimen has to
through drawbead inserts can be considered as plane strain deforma- go through drawbead inserts with more resistance forces causing more
tion. Moreover, in the beginning, the length strain and the thickness deformations. In the meantime, with the increase of radius, the pre-
strain increase fast as the resistance force starts to increase and with the strains decrease with no relation to the height and the number of times.

287
J. Ke et al.
Fig. 7. The schematic diagram of the length strain, width strain and thickness strain of
the specimen flowing through drawbead inserts (the height is 4 mm and the entrance
radius and exit radius are 3 mm, passing through three times) along the length measured
by the Auto Grid Com-smart equipment.
It is explained that under the conditions of same height of male bead
and the number of times, the greater the radius, the less resistance force
the specimen bears, resulting in less deformation. Moreover, with the
increase of number of times when the specimen is pulled through, the
pre-strains also increase. More times means the specimen goes through
drawbead inserts suffering more resistance forces and more deforma-
tions occurs.
When the height is 4 mm and 5 mm, the pre-strain of fillet radius
with a value of 1.5 mm approximates more than twice that of fillet
radius with a value of 4.5 mm, and the pre-strain of fillet radius with a
value of 3 mm is close to half as much again as that of the fillet radius
with a value of 4.5 mm. When the height is 6 mm, the pre-strain of fillet
radius with a value of 1.5 mm is less than two times than that of fillet
radius with a value of 4.5 mm, and the pre-strain of fillet radius with a
value of 3 mm is less than half as much again as that of the fillet radius
with a value of 4.5 mm, which means when a certain height is reached,
adjusting the height is not very effective in increasing more resistance
forces causing more deformations. Moreover, more pre-strains can be
obtained by increasing the radius rather than the height, which reveals
that adjusting the radii of entrance and exit of contra bead is more
efficient, very useful for engineers in adjusting the drawbead inserts to
provide more resistance to deform the sheet metal in practice.
288
Journal of Materials Processing Tech. 254 (2018) 283–293
Fig. 6. The problems in measuring the strains of the
deformed specimens. (a) the grids on the surface of
the specimens suffered from seriously abrasion. (b)
the clamped part of the specimen ruptured during
the test. (c) the length strain obtained by measuring
the length of 30 grids, 40 grids and 50 grids manu-
ally.
3.2. Comparison of mechanical properties
The pre-strained specimens were wire-electrode cut into the uni-
axial tensile test specimens and the conventional uniaxial tensile test
was employed to test the samples. Basic mechanical properties, in-
cluding the yield strength, tensile strength, and percentage of elonga-
tion were measured.
Fig. 9 shows the relationship between the yield strength and the
effective pre-strains of specimens passing through various drawbead
inserts and different number of times. Generally speaking, the yield
strength is increasing with the increase of effective pre-strain, with no
consideration of the number of times. Under the same number of times,
the yield strength increases with the increase of effective pre-strains
caused by different drawbead inserts, but the increasing trend of yield
strength slows down. In addition, the yield strength is different in spite
of the same effective pre-strain under different number of times. The
reason is that the specimens were pulled through different drawbead
inserts, bearing different resistance forces during the process of bending
and unbending.
Fig. 10 presents the relationship between the tensile strength and
the pre-strains of specimens flowing through various drawbead inserts
and different number of times. The variation trend and the reason for
the distinction are the same as that of the yield strength.
Fig. 11 depicts the relationship between the percentage of elonga-
tion of pre-strained specimens after the uniaxial tensile test and the
effective pre-strains of specimens flowing through various drawbead
inserts different times. The variation trend is opposite to that of the
yield strength.
Fig. 12 presents the engineering stress-engineering strain curve of
pre-strained specimens flowing through various drawbead types and
different number of times and the conventional uniaxial tensile test
specimen. The engineering strain are mostly located between 30% and
35%, even smaller than 30%, and occasionally, the engineering strain is
close to 40%. We can clearly see that the engineering strain is not ne-
cessarily increasing with the pre-strain at the same number of time and
even though the pre-strain was the same after different number of
times, the final engineering strain is still distinctive. But, it is quite
apparently that all the engineering strains of pre-strained specimens are
smaller than that of the conventional uniaxial tensile test curve, which
means the process of flowing through drawbead has a negative effect on
the formability of sheet metal.
If we assume the process of flowing through drawbead inserts as the
J. Ke et al.
Fig. 8. The relationship between the pre-strains in length direction and the drawbead
inserts containing three groups of heights of male bead and three groups of radii of en-
trance and exit of contra bead and the number of times of the specimens which were
pulled through. (a) one time. (b) two times. (c) three times.
Fig. 9. The relationship between the yield strength of pre-strained specimens after uni-
axial tensile test and the effective pre-strains of specimens after flowing through various
drawbead inserts with one time, two times and three times.
289
Journal of Materials Processing Tech. 254 (2018) 283–293
Fig. 10. The relationship between the tensile strength of pre-strained specimens after
uniaxial tensile test and the effective pre-strains of specimens after flowing through
various drawbead inserts with one time, two times and three times.
Fig. 11. The relationship between the percentage of elongation of pre-strained specimens
after uniaxial tensile test and the effective pre-strains of specimens after flowing through
various drawbead inserts with one time, two times and three times.
Fig. 12. The comparison of engineering stress-engineering strain curve between pre-
strained specimens flowing through various drawbead inserts and different number of
times and the conventional uniaxial tensile curve.
first strain path, namely, the plane-strain pre-strain, then the following
uniaxial tensile test as the second strain path, the results here shows
that the formability of sheet metal has reduced. As described in the
introduction, Graf and Hosford conducted a test on the Al 6111-T4 and
Al 2008-T4 to investigate the effect of changing paths on forming limits,
and they found that the plane-strain pre-strain raised both sides of the
FLC away from the minimum, which means the plane-strain pre-strain
raises the formability of sheet metal, as shown in Fig. 13.
However, the results here show that the left side of the FLC falls
down, contrary to the results of Graf and Hosford. The reasons may be
that, in the first place, the test material is different. Graf and Hosford
used the aluminum sheets, but the authors used the steel sheets. More
J. Ke et al.
Fig. 13. The forming limit strains of Al 6111-T4 after prestrained near plane-strain ten-
sion (Graf and Hosford, 1994).
important is that the process of pre-strain is different evidently. The
plane-strain pre-strain is obtained by stretching wide sheet under the
uniaxial tensile testing machine in Graf and Hosford’s test, while it is
acquired by pulling the specimen through the drawbead inserts in this
test. It is quite clear that the process of pulling through drawbead is
different from that of stretching wide sheet under the uniaxial tensile
testing machine, which involves bending and unbending under tension.
In order to explain more clearly the effect of passing through
drawbead on the formability of specimens. We construct an effective
stress − effective strain curve based on the uniaxial tensile curves of
pre-strained specimens, as shown in Fig. 14. From the figure, we can see
that the process of passing through drawbead makes the specimen
undergoes additional hardening compared to the single uniaxial tensile
strain path.
Then, if we suppose the instability condition:
dσ/dε = σ
known as the considère condition, premature necking is certain to take
place for the pre-strained specimens. Therefore, the transition process
from the plane-strain path to the uniaxial tensile strain path leading to
the additional strain hardening is the main reason for the reducing
formability of sheet metal.
3.3. Drawbead test (many times until the fracture occurs)
As described in Section 2.4, two groups of drawbead inserts were
selected from nine groups. The intention of the retested samples se-
lected was based on the pre-strain values which have been shown in
Section 3.1 Drawbead test. As the pre-strain of drawbead combination
(the height is 6 mm and the entrance radius and exit radius are 1.5 mm)
is the biggest under the same number of times when the samples were
Fig. 14. The comparison of effective stress − effective strain between the constructed
plane strain curve based on the uniaxial tensile curve after plane strain pre-strain and the
conventional uniaxial tensile curve.
290
Journal of Materials Processing Tech. 254 (2018) 283–293
Fig. 15. The relationship between the times of specimens after flowing through two
groups of drawbead inserts and the percentage of elongation of specimens.
pulled through. So we choose this drawbead combination. While the
pre-strain of drawbead combination (the height is 5 mm and the en-
trance radius and exit radius are 3 mm) is in the middle of all pre-strain
values among all the drawbead combinations. We choose this drawbead
combination as a contrast. Here we define the drawbead inserts with
the height of 6 mm and the radius of entrance and exit of 1.5 mm as
group A and the drawbead inserts with the height of 5 mm and radius of
entrance and exit of 3 mm as group B.
The relationship between the percentage of elongation of specimens
and the number of times of specimens pulling through two groups of
drawbead inserts is shown in Fig. 15. It depicts that with the increase of
the numbers of times, the percentage of elongation of specimens with
two drawbead inserts also increases almost linear. The trend of the
linear lines is that the resistance force is relatively invariable for any
group of drawbead inserts, when the pulling speed from the clamping
unit and the gap between the male bead and contra bead are fixed.
While the increment of the length strain is very related to the resistance
force. When the specimen was pulled through the drawebead inserts
one time, under the same resistance force, the increment of the length
strain, or the percentage of elongation, is relatively fixed. Moreover, the
gradient of lines of group A is higher than that of group B, as the pre-
strain of group A is bigger than the group B in Section 3.1 and the
increment of the length strain of group A is more than that of group B.
Therefore, we can deduce that the gradient between the lines propor-
tional to the change in pre-strain caused by different drawbead inserts.
Here we briefly explain the maximum number of times. In Section
2.4, two problems exist in the drawbead test (many times): one is the
measure of length strains when the surface grids on the sheet suffer
from abrasion. Another is the clamping part of the specimen ruptures
unexpectedly. Although the method of acquiring the length strains by
multiple measuring as relatively accurate as possible is adopted, the
number of times of group B is still restricted to 13 with almost no grids
recognized. Meanwhile, albeit we tried to adjust the distance of the
clamping part to keep the specimen not clamped in the same zone every
time, the result is not good, leading to the number of times of group A
only reaching 10.
However, ignoring the two problems, the percentage of elongation
of group A is still much larger than that of the conventional uniaxial
tensile test and reaches close to 100%. Then we have to ask why the
percentage of elongation can be so large when the specimen was pulled
through drawbead inserts than that in the conventional uniaxial tensile
test and why the percentage of elongation could be so large reaching
nearly 100%.
As described in the introduction, the process of passing through
drawbead can be seen as the cycle of bending, unloading and un-
bending with tensile forces. Here we only consider the effect of bending
and unbending on the deformation of the specimen.
During the process of passing through one round corner (the
J. Ke et al. Journal of Materials Processing Tech. 254 (2018) 283–293

Fig. 16. The schematic of pure bending deformation. Left: the specimen
with thickness t and length L0. Right: the bending moment M with radius
R.

Fig. 17. The strain distribution and stress distribution of the sheet under
pure bending.

Entrance corner or the Exit corner or the Height corner) of the spe- At any time, the tension force T per unit width and bending moment M
cimen, we can consider it as the process of bending under tension. A per unit width of the distance y from the neutral axial was calculated as
situation of rigid, perfectly plastic material with plane strain flow stress following:
S was assumed. For a sheet of thickness t, and length L0 that is bent to a
radius R, at the same time, a tension force per unit width T and a t /2

moment per unit width M is added. T= ∫ σx × 1 × dy

When the tension force T is zero, only the bending moment M −t /2

works, namely the pure bending situation, as shown in Fig. 16. The
strain distribution and stress distribution were presented in Fig. 17.
t /2
eb = y/R (neglect the second order effects)
And the stress is
M= ∫ σx × y × 1 × dy
−t /2

S y>0
σx = ⎧ When the β ≤ 1, the neutral layer has moved the distance u, and
⎨ −S y<0
⎩
1 1
Then when the tension force T increases, the neutral axis moves inside, u = (ea/ eb) × × t = βt
2 2
as shown in Fig. 18 and the strain distribution and the stress distribu-
tion were shown in Fig. 19.
At this time, the strain of the distance y from the neutral layer is −u t /2 −u t /2
e = ea + eb = (β + 1)eb = (β + 1)y/R (neglect the second order ef- T= ∫ σx × 1 × dy + ∫ σx × 1 × dy = ∫ −Sdy + ∫ Sdy=2Su
fects) −t /2 −u −t /2 −u

While β = ea/eb; = βSt

Fig. 18. The schematic of bending and tension deformation. Left: the
specimen with thickness t and length L0. Right: the bending moment M
and tension T with radius R.

291
J. Ke et al. Journal of Materials Processing Tech. 254 (2018) 283–293

Fig. 19. The strain distribution and stress distribution of the sheet under
bending and tension.

−u t /2 −u the deformation can be kept when the amount of material in com-

M= ∫ σx × y × 1 × dy + ∫ σx × y × 1 × dy = ∫ −S × y × 1 × dy pression decreases with increasing tension and as long as the zone
−t /2 −u −t /2
under compression exists. While in the drawbead test, the strain in
t /2
every process passing the round corner is very small (which can be seen
+ ∫ S × y × 1 × dy from Section 3.1) than the bending strain leading to the strain ratio
−u
1 1 β ≤ 1.
= 4 St 2 − Su2 = 4 St 2 (1 − β 2)
What’s more, we assume that during the process of passing through
When the β > 1, the neutral layer has moved out of the sheet, and drawbead, one neck was initiated and as described above, the strain in
this neck area was very small so that the defect cannot create the real
t /2 t /2
failure of the specimen. As we have explained that at any time, the
T= ∫ σx × 1 × dy = ∫ Sdy = St
deformation is concentrated into a small zone than in surrounding
−t /2 −t /2
areas. Then the next moment the neck area has moved out of the
t /2 t /2 t /2 bending or unbending zone and now that which was being bending is
M= ∫ σx × y × 1 × dy = ∫ S × y × 1 × dy = ∫ Sydy = 0 relatively the weakest point and the initiated neck may keep the status
−t /2 −t /2 −t /2 unchanged and when the next unbending (Fig. 21) occurs, the neck
The relationship between the tension force T per unit width and zone was strain hardened and became stronger again. So the defect
bending moment M per unit width and the strain ratio β was shown as cannot grow any further. As a result, mutilple necks developed during
Fig. 20. the process after passing through the drawbead many times, as shown
From Fig. 20, we can see that compared to the pure tension, if in Fig. 21, the specimen can still be stretched. As long as the region of
bending is involved, the sheet can be stretched with a lower force the specimen which is actually passing through the drawbead inserts
(Emmens and van den Boogaard, 2011). In other words, when the keeps safety without any severe necking or fracture, we can deduce that
specimen was pulled through the drawbead inserts, the tension force of the specimen could be passed through the drawbead repeated for a fi-
the zone which is under bending or unbending is less than any other nite number of passes more than now and the percentage of elongation
part resulting in the zone under bending or unbending deforming ahead can be much higher.
of other zones. In some sense, we believe that only the bending or
unbending area was deformed with other parts remain unchanged. 4. Conclusions
Consequently, the deformations are concentrated into the bending or
unbending zone when the specimen was pulled through the drawbead A phenomenon that when the sheet metal flows through drawbead,
inserts. the limit strains of sheet metal lie above the forming limit diagram
As we know, in the conventional uniaxial tensile test, facture is indicating fracture in theory, while it is safe in practice, violates the
initiated by a local neck and only the strain hardening can postpone the normal forming limit curve has been found. In order to explain the
neck process. In some way, the deformation can be kept stable if the phenomenon, a series of tests were done. The conclusions are as fol-
tension increases with the strain. In the drawbead test, from Fig. 20, the lowing:
tension increases with the strain ratio β when the β ≤ 1. In other words,
the tension is determined by the zone under compression, which means (1) The height of male bead and the entrance and exit radii of contra

Fig. 20. Left: the relationship between the tension force T and the strain
ratio β. Right: the relationship between the moment M and the strain ratio
β.

292
J. Ke et al.
bead and the number of times when the specimens were pulled
through drawbead inserts have an influence on the pre-strains of
sheet metal. In briefly, the higher the height of male bead, the
smaller the entrance and exit radii of contra bead, the more number
of times, the more pre-strains can be obtained. Meanwhile, ad-
justing the radius of entrance and exit of contra bead is more effi-
cient than the height of male bead.
(2) The yield strength and the tensile strength increase with the in-
crease of pre-strains, while the percentage of elongation decreases
with the increase of pre-strains. Under the condition of same pre-
strain, the yield strength, tensile strength and percentage of elon-
gation are not the same, which relates to the type of drawbead
inserts. Moreover, the percentage of elongation of pre-strained
specimens were evidently smaller than that of the conventional
uniaxial tensile test. The transition process from the plane-strain
path to the uniaxial tensile strain path leading to the additional
strain hardening is the main reason for the reducing formability of
sheet metal.
(3) When the specimens were pulling through drawbead inserts many
times until the fracture occurs, the percentage of elongation exceeds
that of the conventional uniaxial tensile test and reaches close to
100%. Bending and unbending under tension has a significant effect
on the increased percentage of elongation.
Acknowledgements
We express our sincere appreciation to the Research and
Development center of Wuhan iron and steel (group) corp for providing
the equipment and sheet metals. This research did not receive any
specific grant from funding agencies in the public, commercial, or not-
for-profit sectors.
References
Allwood, J.M., Shouler, D.R., 2009. Generalised forming limit diagrams showing in-
creased forming limits with non-planar stress states. Int. J. Plast. 25, 1207–1230.
Assempour, A., Nejadkhaki, H.K., Hashemi, R., 2010. Forming limit diagrams with the
existence of through-thickness normal stress. Comput. Mater. Sci 48, 504–508.
Banabic, D., Barlat, F., Cazacu, O., Kuwabara, T., 2010. Advances in anisotropy and
formability. Int. J. Mater. Form. 3, 165–189.
Charpentier, P.L., 1975. Influence of punch curvature on the stretching limits of sheet
metal. Metall. Trans. 5, 2161–2164.
Courvoisier, L., Martiny, M., Ferrom, G., 2003. Analytical modelling of drawbeads in
sheet metal forming. J. Mater. Process. Technol. 133, 359–370.
293
Journal of Materials Processing Tech. 254 (2018) 283–293
Fig. 21. Multiple necks developed along the specimen after passing
through the drawbead many times.
Emmens, W.C., van den Boogaard, A.H., 2008. Extended tensile testing with simultaneous
bending. Int. Deep Drawing Res. Group 1–12.
Emmens, W.C., van den Boogaard, A.H., 2009. An overview of stabilizing deformation
mechanisms in incremental sheet forming. J. Mater. Process. Technol. 209,
3688–3695.
Emmens, W.C., van den Boogaard, A.H., 2011. Cyclic stretch-bending: mechanics, sta-
bility and formability. J. Mater. Process. Technol. 211, 1965–1981.
Eyckens, P., Van Bael, A., Van Houtte, P., 2009. Marciniak–Kuczynski type modelling of
the effect of Through-Thickness Shear on the forming limits of sheet metal. Int. J.
Plast. 25, 2249–2268.
Fictorie, E., van den Boogaard, A.H., Atzema, E.H., 2010. Influence of punch radius in a
nakazima test for mild steel and aluminium. Int. J. Mater. Form. 3, 1179–1182.
Firat, M., 2007. Computer aided analysis and design of sheet metal forming processes.
Mater. Des. 28, 1311–1320.
Ghoo, B., Keum, Y., 2000. Expert drawbead models for sectional FEM analysis of sheet
metal forming processes. J. Mater. Process. Technol. 105, 7–16.
Ghosh, A.K., Hecker, S.S., 1974. Stretching limits in sheet metals?In-plane versus out-of-
plane deformation. Metall. Trans. A 6A, 1665–1669.
Goodwin, G., 1968. Application of Strain Analysis to Sheet Metal Forming Problems in the
Press Shop. SAE Technical Paper.
Graf, A., Hosford, W., 1994. The influence of changing strain path on forming limit
diagrams of Al 6111-T4. Int. J. Mech. Sci. 36, 897–910.
Graf, A., Hosford, W., 1993. Effect of changing strain path on forming limit diagrams of Al
2008-T4. Metall. Trans. A 24A, 2503–2512.
Keeler, S., Backofen, W., 1963. Plastic instability and fracture in sheets stretched over
rigid punches. ASM Trans. Q 56, 25–48.
Kim, C., Im, Y., Heo, Y., Kim, N., Jun, G., Seo, D., 1997. Finite-element analysis and
experimental verification for drawbead drawing processes. J. Mater. Process.
Technol. 72, 188–194.
Kitting, D., Koplenig, M., Ofenheimer, A., Pauli, H., Till, E.T., 2009. Application of a
concave-side rule approach for assessing formability of stretch-bent steel sheets. Int.
J. Mater. Form. 2, 427–430.
Laukonis, J.V., Ghosh, A.K., 1978. Effects of strain path changes on the formability of
sheet metals. Metall. Trans. A 9A, 1849–1956.
Nurcheshmeh, M., Green, D.E., 2016. Prediction of forming limit curves for nonlinear
loading paths using quadratic and non-quadratic yield criteria and variable im-
perfection factor. Mater. Des. 91, 248–255.
Panich, S., Barlat, F., Uthaisangsuk, V., Suranuntchai, S., Jirathearanat, S., 2013.
Experimental and theoretical formability analysis using strain and stress based
forming limit diagram for advanced high strength steels. Mater. Des. 51, 756–766.
Samuel, M., 2002. Influence of drawbead geometry on sheet metal forming. J. Mater.
Process. Technol. 122, 94–103.
Smith, L.M., Averill, R.C., Lucas, J.P., Stoughton, T.B., M, P.H., 2003. Influence of
transverse normal stress on sheet metal formability. Int. J. Plast. 19, 1567–1583.
Stoughton, T.B., Zhu, X., 2004. Review of theoretical models of the strain-based FLD and
their relevance to the stress-based FLD. Int. J. Plast. 20, 1463–1486.
Stoughton, T.B., 2000. A general forming limit criterion for sheet metal forming. Int. J.
Mech. Sci. 42, 1–27.
Sun, G., Li, G., Gong, Z., Cui, X., Yang, X., Li, Q., 2010. Multiobjective robust optimization
method for drawbead design in sheet metal forming. Mater. Des. 31, 1917–1929.
Tharrett, M., Stoughton, T.B., 2003. Stretch-bend Forming Limits of 1008 AK Steel. SAE
Technical Paper.
Zhang, F., Chen, J., Chen, J., 2014. Effect of through-thickness normal stress on forming
limits under Yld2003 yield criterion and M-K model. Int. J. Mech. Sci. 89, 92–100.