Materials Transactions, Vol. 49, No. 5 (2008) pp.

1112 to 1119
Special Issue on Platform Science and Technology for Advanced Magnesium Alloys, IV
#2008 The Japan Institute of Metals

Experimental Determination of Forming Limit Diagram

and Springback Characteristics of AZ31B Mg Alloy Sheets
at Elevated Temperatures
Hyung Jong Kim1 , Sun Chul Choi2 , Kyoung Taek Lee3 and Heon Young Kim1
1
Division of Mechanical Engineering & Mechatronics, Kangwon National University,
192-1 Hyoja 2-Dong, Chuncheon, Gangwon-Do 200-701, Korea
2
Graduate School, Kangwon National University, 192-1 Hyoja 2-Dong, Chuncheon, Gangwon-Do 200-701, Korea
3
Technical Center, AUSTEM Co., Ltd, 173-299 Kajwa-Dong, Seo-Gu, Incheon-Si 404-250, Korea

A warm/hot formability testing apparatus was designed and fabricated for this study. The forming limit curve (FLC) and springback
characteristics of AZ31B Mg alloy sheets at elevated temperatures of up to 300 C were investigated. The forming limit increased rapidly with
temperature up to 200 C, and increased slightly thereafter. In the range from 200 to 300 C, the FLC0 was found to be six to seven times higher
than at room temperature. In isothermal springback tests, the effect of the blank holding force (BHF) on springback differed slightly from that for
other metal sheets at low temperatures. At 200 C and above, negligible springback was observed over the range of BHF’s used. With increasing
temperatures, springback decreased rapidly up to 200 C, and very slowly afterward. In nonisothermal tests, considerable springback was
observed, even at temperatures of 200 C or higher; furthermore, it decreased almost linearly with increases in both the temperature and
BHF. [doi:10.2320/matertrans.MC200778]

(Received October 18, 2007; Accepted February 19, 2008; Published April 2, 2008)
Keywords: AZ31B magnesium alloy sheet, forming limit diagram, springback, punch stretching, 2D draw bending

1. Introduction characteristics at different temperatures. Kim et al.8) reported

Magnesium is a fairly strong and light-weight metal, with a
density that is one-fourth that of steel and two-thirds that of
aluminum. However, when alloyed with other metals, it has
an excellent strength-to-weight ratio and shows outstanding
performance in terms of machinability, vibration absorption,
and electromagnetic shielding. These features have drawn
attention to magnesium as a useful material for various
electronic parts and structural applications. Magnesium
alloys usually exhibit very poor workability and formability
at room temperature because of their hexagonal close-packed
(HCP) structure, but the forming limit can be considerably
increased at elevated temperatures.1) Conventionally, die
casting or squeeze casting is used for manufacturing
magnesium alloy parts. These processes, however, require
heating materials in the liquid or semisolid state, which may
result in casting defects and finally cause safety problems.
Recently, warm or hot press-forming technology for magne-
sium alloy sheets has been recognized as a promising
alternative. However, there still are difficulties involved in
controlling the process parameters such as the forming speed
and temperature since magnesium has a low heat capacity.
Most of the studies on magnesium alloy sheet forming at
elevated temperatures have focused on the formability, and
only few on the springback characteristics.
Zhang et al.,2) Yoshihara et al.,3) and Huang et al.4)
conducted experimental investigations and numerical analy-
ses on the formability and the effect of the blank holding
force (BHF) during a deep drawing process. Naka et al.5)
produced the forming limit diagram (FLD) of JIS-A5083
Al–Mg alloy for various temperatures and strain-rates. Chen
et al.6) evaluated the formability of AZ31B magnesium alloy
sheets through conical cup and V-bending tests. Bruni et al.7)
carried out air bending tests to investigate the springback
on the springback characteristics of AZ31B magnesium alloy
sheets in draw bend tests at room temperature, and Lee et al.9)
proposed a constitutive equation for the same material.
Recently the authors10) have performed some experiments to
investigate the change in the forming limit and the spring-
back characteristics of the same material with an increase in
temperature. Many studies on magnesium alloy sheets have
shown that it has significantly greater formability at temper-
atures above 200 C than at lower temperatures in various
sheet forming processes.
The limit strains are represented best by the concept of a
forming limit curve (FLC), which shows the onset of
localized necking over all possible combinations of surface
strains in the sheet.11) FLD’s are usually obtained by two
methods: the out-of-plane deformation test using a hemi-
spherical punch, proposed by Hecker;12) or the in-plane
deformation test using a flat-headed punch with a circular
cutout in the center, developed by Marciniak.13)
Springback occurs due to the elastic recovery after sheet
metal stamping operations, affecting dimensional accuracy,
assembly tolerance, and appearance to a great extent.
Therefore, it has been one of the most frequently studied
issues over the past scores of years.14–16) The 2D draw
bending test was adopted by NUMISHEET ’93 Conference17)
as a benchmark problem, aiming at the comparison between
numerical (finite element) prediction and experimental
measurement on the basis of the geometric definition of the
springback quantity. Since the blank is not constrained by its
configuration in this test, it undergoes considerable elastic
recovery. This test allows the application of various BHF’s.
In the present study, punch stretching tests and 2D draw
bending tests were carried out with AZ31B magnesium alloy
sheets to obtain the FLD and to investigate the springback
characteristics at elevated temperatures. It is known that
 Experimental Determination of Forming Limit Diagram and Springback Characteristics of AZ31B Mg Alloy Sheets
(a)
(b)
Fig. 1 Photo of hot formability testing equipment with (a) punch stretching
tools and (b) 2D draw bending tools.
at high punch speed of 200 mm/min the level of the FLD
for elevated temperatures is as low as that for the room
temperature.5) In this study, however, all the tests were
performed at very low deformation speed so that the effect
of strain-rate on the FLD and springback might not be
dominant. The goal of the current work is to set up an
experimental methodology and to provide reliable material
data for die and process designers.
2. Forming Limit Diagram
2.1 Punch stretching test
Figure 1 shows a double-acting press, which has two
hydraulic cylinders of 50-ton capacity each. An electric
furnace and a temperature control unit for heating the die set
and blank to the desired temperature were set up on the bed.
1113
Fig. 2 Schematic diagram of punch stretching test apparatus for FLC
measurements.
Fig. 3 Geometry and dimension of FLD test specimens.
Alternatively, heating cartridges of proper capacity were
inserted in each tool, enabling heating of the die set and blank
directly without the furnace. All the tests were conducted
using this direct heating method, which could maintain the
temperature more stably. A schematic of the hemispherical
punch stretching apparatus is illustrated in Fig. 2, showing
the location of the heating cartridges as well as the
dimensions of the tools. The main dimensions were the
same as those of the NUMISHEET ’9618) benchmark
problem.
FLD test specimens having the same length (in rolling
direction) of 175 mm and seven different widths in the range
from 25 to 175 mm were cut into the shapes shown in Fig. 3
from a 0.5-mm-thick AZ31B sheet. It was found helpful to
1114 H. J. Kim, S. C. Choi, K. T. Lee and H. Y. Kim
Fig. 4 Photos of FLD test specimens at 250 C and enlarged images showing local necks.
provide a circular cutout on both sides of the specimens
narrower than 75 mm in order to avoid wrinkling around the
punch and fracture in the clamping bead.10,11) All specimens
were gridded with 3  3 mm squares to permit strain
measurement. The grids were printed onto each specimen
using a silk-screen technique with a very accurate negative,
instead of an electrochemical etching method. This was
because etched grids might act as initial imperfections in the
thin sheets: several pretest specimens showed that etched
grids led to premature failure and, consequently, a drop in the
forming limit. No lubrication was used between the punch
and blank for most tests. To induce fracture in the balanced
biaxial mode, a water-soluble graphite lubricant was applied
on the punch surface prior to the test with the 175  175 mm
specimen. Tests were performed at temperatures of 25, 100,
200, 250, and 300 C and at a constant punch speed of
0.1 mm/s.
2.2 FLD test results and discussion
For reliable measurement of strains around localized
necks, ‘‘ASIAS’’19) was used; it is an automated surface-
strain measurement system developed by the authors by
using image processing and stereo vision techniques. Recent
improvements in ASIAS have enabled it to process high-
resolution images of eight megapixels or above, and to
measure strains to an accuracy of 0:17%.
Figure 4 is a photo of the test specimens formed at 250 C.
The enlarged photos show the fractures or localized necks
clearly. Figures 5(a)–(e) show the construction process of the
FLC’s for 25, 100, 200, 250, and 300 C, respectively, and
Fig. 5(f) compares them with each other. These figures show
that the higher the temperature, the higher the forming limit,
as can be predicted from the stress-strain curves for various
temperatures shown in Fig. 6.
The FLC0 of the AZ31B sheet increased rapidly up to a
temperature of 200 C, and thereafter increased slowly up to
300 C, at which it reached a value seven times that at room
temperature. It is known that the considerable increase in
formability at temperatures slightly higher than 200 C is due
to the thermal activation of the pyramid sliding planes in the
hexagonal structure.20)
The limit minor strains measured for specimens of even the
same width varied extensively with temperature. For exam-
ple, the minor strains obtained for the 25-mm-wide speci-
mens ranged from 0:15 at room temperature to 0:7 at
100 C, and those for the 175-mm-wide lubricated specimens
ranged from 0.04 at room temperature to 0.4 at 200 and
300 C. It is remarkable but not understood that the strain
state in the 25-mm-wide specimen at 100 C is far beyond the
uniaxial tension mode, and approaching the pure shear mode.
One possible explanation is that a change in the aggregated
structure of the AZ31B sheet around this temperature causes
its normal anisotropy to increase abruptly.
The graphite lubricant used for the warm/hot forming
processes did not exhibit an effect as good as that achieved by
a polyurethane membrane coated on both sides with mineral
oil or beef tallow, which is usually used for cold FLD tests.
This made it impossible to obtain the limit strains under
balanced biaxial tension.
In the tests at 25 and 100 C, brittle fracture occurred
suddenly during the forming processes, and it was very hard
to stop the punch at the very moment of local necking. This
gave rise to significant differences between the safe and
necking (or failure) zones in the obtained FLD’s. In the tests
at 200 C and above, on the other hand, it was comparatively
easy to find the point of local necking from the punch load–
displacement graph and also to stop the punch, enabling
accurate plotting of the FLC along the boundary between the
safe and necking zones. As shown in Figs. 5(c)–(e), the major
strains of necked grids are scattered over a band with a
maximum width of 0.2, indicating that the post-necking
strain was fairly large due to the increase in the strain-rate
sensitivity at elevated temperatures.
3. Springback Characteristics
3.1 2D draw bending test
To evaluate the springback characteristics of AZ31B
magnesium alloy sheets, 2D draw bending tests were con-
ducted as described in the NUMISHEET ’93 benchmark. The
test apparatus is illustrated in Fig. 7. Each test specimen was
rectangular in shape, with a length of 350 mm (in rolling
direction), width of 35 mm, and thickness of 1.0 mm.
The forming temperatures were set to 25, 100, 200, 250,
and 300 C—the same as in the FLD tests—using the direct
heating method. Tests were focused on the influence of the
 Experimental Determination of Forming Limit Diagram and Springback Characteristics of AZ31B Mg Alloy Sheets 1115

(a) (b)

(c) (d)

(e) (f)

Fig. 5 Test data and FLC’s for (a) 25 C, (b) 100 C, (c) 200 C, (d) 250 C, (e) 300 C, and (f) all.

BHF as well as temperature on the springback character-
istics. The BHF was determined to be 30 to 250 kgf from
preliminary tests. To control such small BHF’s, a screw-type
blank holding device and a load cell with a 2-ton capacity
were attached to the test apparatus, as shown in Fig. 1(b). The
punch speed and its final stroke were kept constant at
1.0 mm/s and 70 mm, respectively, for all the tests.
The tests were divided into two categories: the isother-
mal tests, wherein the punch, die, and blank holder were
heated to and maintained at the same temperature; and the
nonisothermal tests, wherein only the die and blank holder
were heated while the punch remained at a much lower
temperature.
3.2 Springback test results and discussion
The springback angles
1 and
2 , defined in Fig. 8,
quantify the extent of elastic recovery of a test blank. It can
be said that the larger the value of
1 and the smaller the value
of
2 , the greater the springback that occurs, and both angles
would remain right angles without springback. These angles
were measured using a CAD software from scanned images
of either side of a deformed specimen.
1116 H. J. Kim, S. C. Choi, K. T. Lee and H. Y. Kim

Fig. 6 Stress-strain curves at various temperatures.

3.2.1 Isothermal test

The isothermal tests were conducted after thermal equi-
librium at a given temperature was achieved by controlling
the heating units of the punch, die, and blank holder
simultaneously. Figure 9(a) shows a photo of the test spec- Fig. 7 Schematic diagram of 2D draw bending test apparatus for spring-
imens deformed at different temperatures. The springback back measurements.
angles for various temperature and BHF combinations are
tabulated in Table 1. Figure 10 illustrates the effect of the
BHF on the springback angles for different temperatures, and
Fig. 11 the relationship of the springback angles to the
temperature for various BHF’s. It is well known that as the
BHF increases, springback usually tends to be smaller, or in
other words,
1 is smaller and
2 is larger. However, the
magnesium alloy sheet used in this study exhibited character-
istics different from those of other metals, especially at low
temperatures. An increase in the BHF led to an increase in
both
1 and
2 at 25 C, but a decrease in both
1 and
2 at
100 C. This is similar to the results of a cold draw bend test Fig. 8 Definition of springback angles.

(a)

(b)

Fig. 9 Photos of 2D draw bending test specimens in (a) isothermal test and (b) nonisothermal test.
 Experimental Determination of Forming Limit Diagram and Springback Characteristics of AZ31B Mg Alloy Sheets 1117

Table 1 Measured data of springback angles in isothermal tests.

BHF, Angle,
25 C 100 C 200 C 250 C 300 C
kgf deg

1 130.2 113.5 90.7 90.1 89.9
30

2 70.3 78.1 87.9 89.9 91.0

1 128.9 113.6 90.8 90.7 90.2
50

2 70.6 78.1 88.8 90.6 89.8

1 131.4 114.6 90.2 89.9 90.1
120

2 71.7 77.9 89.0 90.2 91.2

1 132.8 112.6 90.7 90.4 89.1
250

2 74.5 75.8 89.2 89.5 91.3

(a)

Fig. 11 Variation of springback angles with temperature in isothermal

tests.

the decrease in flow stress with increasing temperature as

shown in Fig. 6.
3.2.2 Nonisothermal test
In the nonisothermal tests, the die and blank holder were
heated to the desired temperature but not the punch. This test
aimed to simulate the deep drawing process, wherein the
punch was cooled in order to improve formability. The
temperatures of the die and blank holder were set to 200,
250, and 300 C. In each case, the punch temperature was
maintained at 100, 110, and 120 C, respectively. Of course,
(b) a slight rise in the punch temperature due to thermal
conduction was inevitable, requiring a pause for cooling
the punch between two successive tests.
Figure 9(b) is a photo of the test specimens, and Table 2
lists the measured springback angles. The changes in the
springback angles with the BHF and temperature are dis-
played in Fig. 12 and Fig. 13, respectively. These figures
and the table clearly show that springback occurred even at
temperatures of 200 C or higher, unlike in the isothermal
tests: with increasing temperature and BHF, springback
decreased almost linearly. Therefore, when cooling or not
heating the punch to improve formability, the springback at
elevated temperatures must be properly taken into consid-
eration for the design of the forming process and tools.

Fig. 10 Variation of springback angles with BHF in isothermal tests.
by Kim et al.,8) wherein springback increased until the load
reached a certain value and decreased once the back-tension
exceeded the limit. On the other hand, at temperatures above
200 C, springback scarcely occurred, regardless of the BHF.
The effect of temperature on springback, as shown in Fig. 11,
was very similar to that on the forming limit; that is,
springback decreased rapidly up to 200 C, but subsequently
continued decreasing very slowly. The decrease in spring-
back with increasing temperature is mostly because of
4. Conclusion
A double-acting hydraulic press equipped with heating
and temperature control units was designed and fabricated
for this study. The forming limit and springback character-
istics of AZ31B magnesium alloy sheets at elevated temper-
atures were investigated extensively. The test results are
summarized as follows:
(1) The forming limit increased sharply with temperature
up to 200 C, around which the sliding planes were
activated; and thereafter, the increasing trend slowed
down. In the range from 200 to 300 C, the commonly
used range of forming temperatures for magnesium
1118 H. J. Kim, S. C. Choi, K. T. Lee and H. Y. Kim

Table 2 Measured data of springback angles in nonisothermal tests.

BHF, Angle, Die 200 C Die 250 C Die 300 C

kgf deg Punch 100 C Punch 110 C Punch 120 C

1 107.9 99.4 93.8
30

2 78.0 83.9 89.3

1 106.7 98.1 92.4
60

2 77.2 83.2 89.4

1 105.8 97.9 90.9
120

2 80.5 86.3 90.1

1 103.2 94.2 91.5
250

2 83.3 88.8 90.3

(a)

Fig. 13 Variation of springback angles with temperature in nonisothermal

tests.

(3) In the isothermal tests, where not only the die and blank
holder but also the punch were heated up to the same
temperature, an increase in the BHF led to an increase
in both
1 and
2 at 25 C, but a decrease in both
1 and

2 at 100 C. This feature differs slightly from that of
other metal sheets. At 200 C or above, springback was
observed to be ignorable in the range of BHF’s used. As
for the effect of temperature, springback decreased
rapidly up to 200 C, and very slowly afterward.
(4) In case of the nonisothermal tests, wherein the punch
(b) was not heated, considerable springback occurred even
at temperatures of 200 C or higher. With an increase
in the temperature and BHF, springback decreased
almost linearly. Therefore, in the case of cooling or not
heating the punch to improve formability, the spring-
back at elevated temperatures must be taken into
consideration for proper design of the forming process
and tools.

Acknowledgment

This study was supported by the Research Grant from

Kangwon National University.

REFERENCES

Fig. 12 Variation of springback angles with BHF in nonisothermal tests.
alloy sheets, the FLC0 was found to be six to seven
times higher than that at room temperature.
(2) The limit minor strains, despite being measured from
specimens of the same width, ranged over an extensive
area in the FLD’s, depending on the temperature. It is
remarkable but not understood that the limit strain state
in the 25-mm-wide specimen at 100 C was close to the
pure shear mode, far beyond the uniaxial tension mode,
unlike that at other temperatures.
1) H. Watanabe, H. Tsutsui, M. Kohzu, S. Tnabe and K. Higashi: Int. J.
Plasticity 17 (2001) 387–397.
2) S. H. Zhang, K. Zhang, Y. C. Xu, Z. T. Wang, Y. Xu and Z. G. Wang:
J. Mat. Proc. Tech. 185 (2007) 147–151.
3) S. Yoshihara, K. Manabe and H. Nishimura: J. Mat. Proc. Tech. 170
(2005) 579–585.
4) T. B. Huang, Y. A. Tsai and F. K. Chen: J. Mat. Proc. Tech. 177 (2006)
142–145.
5) T. Naka, G. Torikai, R. Hino and F. Yoshida: J. Mat. Proc. Tech. 113
(2001) 648–653.
6) F. K. Chen and T. B. Huang: J. Mat. Proc. Tech. 142 (2003)
643–647.
7) C. Bruni, A. Forcellese, F. Gabrielli and M. Simoncini: J. Mat. Proc.
Tech. 177 (2006) 373–376.
 Experimental Determination of Forming Limit Diagram and Springback Characteristics of AZ31B Mg Alloy Sheets
8) J. G. Choi, S. C. Choi, M. G. Lee and H. Y. Kim: Trans. Mat. Proc. 16
(2007) 302–305 (in Korean).
9) M. G. Lee, K. S. Chung and H. Y. Kim: Trans. Mat. Proc. 16 (2007)
234–238 (in Korean).
10) S. C. Choi, H. S. Lee, H. J. Kim, K. T. Lee and H. Y. Kim: Trans. Mat.
Proc. 16 (2007) 364–369 (in Korean).
11) S. S. Hecker: Proc. Symposium on Formability, ed. by A. K. Ghosh and
H. L. Gegel, (Metall. AIME, 1977) pp. 150–182.
12) S. S. Hecker: Sheet Metal Industries 52 (1975) 671–676.
13) Z. Marciniak: 1974, Aspects of Material Formability, McMaster Univ.,
(Hamilton, ON, Canada).
1119
14) F. Pourboghrat and E. Chu: J. Mat. Proc. Tech. 50 (1995) 361–374.
15) D. Zhou and R. H. Wagoner: J. Mat. Proc, Tech. 50 (1995) 1–16.
16) W. D. Carden, L. M. Geng, D. K. Matlock and R. H. Wagoner: Int. J.
Mech. Sci. 44 (2004) 79–101.
17) A. Makinouchi: Proc. 2nd Int. Conf. NUMISHEET ’93, ed. by
E. Nakamachi, E. Onate and R. H. Wagoner (1993).
18) J. K. Lee: Proc. 3rd Int. Conf. NUMISHEET ’96, ed. by G. L. Kinzel
and R. H. Wagoner (1996).
19) D. S. Kim and H. J. Kim: Trans. Mat. Proc. 8 (1999) 294–305
(in Korean).
20) E. Doege and K. Dröder: J. Mat. Proc. Tech. 115 (2001) 14–19.