Journal of Materials Processing Technology 240 (2017) 354361

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

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

Forming limit in thermal cruciform biaxial tensile testing of titanium

alloy
Rui Xiao, Xiao-Xing Li , Li-Hui Lang, Qiu Song, Kang-Ning Liu
School of Mechanical Engineering and Automation, Beihang University, Beijing 100191, China

a r t i c l e i n f o a b s t r a c t

Article history: In this study, a new cruciform biaxial tensile testing technology was employed to evaluate the thermal
Received 5 July 2016 limit strains of a TA1 titanium alloy, which is widely used as an aircraft structural component, so as to
Received in revised form 12 October 2016 prevent fracture defects during thermoplastic processing. The effect of specimen shape on the forming
Accepted 15 October 2016
limit and its measurement range is discussed. The results of this study show that a design that reduces
Available online 19 October 2016
the semispherical thickness results in localized necking and fracture at the center of the specimen when
normal stress is applied to it. The forming limit diagrams obtained at different angles with respect to
Keywords:
the material rolling direction suggested that the limit strain was anisotropic in nature. The experimental
Thermal limit strains
Cruciform biaxial tensile tests
results were in good agreement with the substituted values obtained from the nite element analysis
Titanium alloy and Oyane ductile fracture criterion. Finally, the efcacy of the related test and prediction methods was
Finite element analysis corroborated.
Oyane ductile fracture criterion 2016 Elsevier B.V. All rights reserved.

1. Introduction for materials having a high phase-transition temperature, such as

A forming limit diagram (FLD) is used to evaluate the strain (or
stress) at the onset of localized necking of sheet metals under com-
plicated loading. It is composed of two principal strains (or stresses
obtained from different strain ratios tests) that act as the verti-
cal and horizontal coordinates in a plane. Depending upon how the
force is applied, FLDs can be commonly obtained by two biaxial ten-
sile methods: one method uses a cylindrical at punch to deep draw
(in-plane), whereas the other uses a hemispherical punch to bulge
(out-of-plane). In addition, sheet forming using thermoplastic
processes such as hot stamping has been widely implemented.
Kotkunde et al. (2016) reported the FLD for a Ti6Al4 V alloy in
a thermal environment based on the bulging caused by a hemi-
spherical punch. Unfortunately, such methods require changing
the length-to-width ratios of specimens to obtain different strain
ratios, which causes the experimental veracity to be narrowly
related to the precision of the mold, thus making the process costly.
Consequently, Yoshida and Kuwabara (2007) proposed tube expan-
sion testing (also named hydroforming) to acquire the limit stress
and strain of materials. However, the hot-pressure medium (e.g.,
oil or gas) hardly met the experimental temperature requirements
Corresponding author.
E-mail address: 88229015@qq.com (X.-X. Li).
titanium alloys.
In this paper, a thermal cruciform biaxial tensile method is pro-
posed to realize different strain ratio tests by changing the stroke
ratios of the orthogonal axes. Hannon and Tiernan (2008) published
a useful summary on cruciform biaxial tensile techniques, which
widely used by researchers to investigate the plastic-deformation
properties of materials. The test requirements are adhered to when
the center section of the specimen reaches the plastic-deformation
stage and a low strain level. However, in order to study the form-
ing limit in the current study, the center section of the specimen
needed to reach necking or fracture. Therefore, one of the most
formidable technical barriers has been to address issues related to
specimen design. Merklein et al. (2008) employed a laser beam to
heat specimens, whereas Abu-Farha et al. (2009) applied a heat gun
as an exterior heating device. However, these techniques result in
defects similar to those found in hydroforming, and they fail to
meet the temperature requirements. Kulawinski et al. (2015) was
the only exception in this regard, as they investigated the effect of
the multiaxial stress state on the fatigue life of superalloys in the
temperature range of 400650 C via the isotropic theory.
To predict the sheet metal forming limit, Hill (1952) proposed
the tensile instability theory, which was based on diffuse and local-
ized necking. Stren and Rice (1975) determined the limit strains
in the whole region of strain paths using a simplied constitutive
model of a pointed vertex on subsequent yield loci. However, the
predicted FLD does not always t the measured values very well.

http://dx.doi.org/10.1016/j.jmatprotec.2016.10.016
0924-0136/ 2016 Elsevier B.V. All rights reserved.
 R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361
Thus, Marciniak and Kuczynski (M-K) (1967) proposed the widely
used damage instability and groove theory. The limit strains near
the balanced biaxial tension predicted using the M-K analysis are
extremely large as compared to those obtained experimentally.
Tadros and Mellor (1978) later modied the M-K theory. Although
these modications have optimized the M-K theory, experimental
verication suggests that the theory is expected to have a cer-
tain gap with the test results. Another hypothesis to predict the
limit strain is the ductile fracture criterion, which is derived from
deformation energy calculations and ductile damage models and
is estimated by the macroscopic stress and strain during form-
ing. Oyane et al. (1980) assumed that the history of hydrostatic
stress played an important role in the occurrence of ductile fracture.
On the basis of nite element simulations and hardening models,
Takuda et al. (2000) successfully applied this criterion to the pre-
diction of the forming limit of aluminum alloy sheets, although the
value still needs to be veried at both room and elevated temper-
atures.
In this work, a special specimen was designed for investi-
gating the forming limit. Xiao et al. (2016a,b) developed a new
high-temperature biaxial tensile testing machine and a digital-
image correlation (DIC) measurement system to conduct studies on
the mechanical properties of nickel-based superalloys in previous
study. The TA1 titanium alloy limit strains at elevated tempera-
tures and different stroke ratios were calculated by using these
technologies. The strains for localized necking were predicted by
the combination of the nite element analysis and ductile fracture
criterion. The objective of the present work was to evaluate the
plasticity of TA1 and to determine its FLD at different temperatures.
2. Cruciform specimen
Many researchers have investigated the optimized shapes of
cruciform specimens, including those used at elevated tempera-
tures. However, the standard geometry remains a topic of ongoing
research. The design prerequisite for the forming limit is a center-
section fracture or necking of the specimen subjected to normal
stress. Solutions using slots in the specimen arms and thickness
reduction at the test section are reported to have more adherents.
Based on this design concept, Zidane et al. (2010) investigated the
forming limit of aluminum alloys. However, there is some skepti-
cism regarding the determined values: rst, as shown in Fig. 1, a
convex llet located between the specimen arms plays a role sim-
ilar to that of a reinforcing rib, which lowers the strain level at
the center section; second, based on nite element simulation and
experiments, Xiao et al. (2016a,b) suggested that the failure origi-
nates at the endpoints of the slots close to the center area because
of shear concentration and unavoidable small cracks caused by
machining defects. Further, Smits et al. (2006) reported that a cir-
cular groove can increase the deformation in the central area and
Fig. 2. Cruciform specimen with circular grooves.
355
Fig. 1. Cruciform specimen with slots and thickness reduction.
avoid premature arm fracture during experiments. Unlike other
researchers, Ognedal et al. (2012) employed a semispherical thick-
ness reduction of the central area. In the current study, the authors
discussed the inuence of the thickness reduction methods on
stress distribution using nite element analysis. Kuwabara (2014)
regulated the dimensions of the specimen standard, shown in Fig. 2,
recommended by the biaxial tensile test international standard
ISO16842-2014 with a length of 240 mm and a width of 50 mm.
In the following discussion, the rolling and transverse directions
are dened as the X and Y directions, respectively. The length of the
clamped area was 42 mm.
A quarter-model with solid elements (SOLID164) with an edge
length of 1 mm, a simulation time of 0.1 s, the von Mises yield cri-
terion with an elastic modulus of 206 GPa, a Poisson ratio of 0.28, a
yield strength of 145 MPa, and a tangent modulus of 5800 MPa was
built using the pre-processing software HyperMesh. The calcula-
tion and post-processing were done in ANSYS solver. Fig. 3 shows
the calculation results. For model A, the stress was concentrated at
the border between the lleted corner and traditional straight sur-
face, which resulted in premature fracture in this region. For model
B, the stress was concentrated at the center area of the arc surface.
Because of the non-uniform thickness of the center area, some
data of the nodes along the center line in the X direction were
extracted in order to further verify the feasibility of model B and
the location of the strain measurement in the test, as shown in
Fig. 4, as a top partial viewport of the specimen center area. The
distance between the adjacent nodes was uniform (1 mm). Node
0 and element (Ele.) 0 were located at the heart of the specimen.
Fig. 5a shows the straintime curves of nodes 05. At the beginning
of the deformation, the elastic strain of every node showed a linear
increase with similar values. With increasing plastic deformation,
wide variations in the elastic strains were observed. The maximum
356 R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361

Fig. 3. Stress distribution of the specimens during the biaxial tensile test.

Fig. 4. Position of the sample nodes.

Fig. 5. Data extracted from the nodes.

strain decreased with increasing distance from the nodes to the tension test using the von Mises criterion. The relationship can be
specimen heart. As compared to node 5, nodes 04 showed smaller expressed as
gaps. Fig. 5b compares the principal and shear stress of nodes 04.  2
Dawicke and Pollock (1997) investigated the relationship between xx + yy xx yy
1or2 = + xy 2 , (1)
the principal, normal, and shear stress obtained during a biaxial 2 2
where  1or2 is the principal stresses,  xx and  yy are the normal
stresses in the X and Y directions, respectively, and  xy is the shear
 R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361 357

Fig. 6. Instant of fracture.

stress. Eq. (1) indicates that the deformation is mainly affected by
the normal stress and that the effect of the shear stress is negligible.
Therefore, in this study, a measurement range of 4 mm in the central
area of the specimen was considered to be optimum.
3. Experimental procedure
3.1. Biaxial tests
Xiao et al. (2016a,b) described a biaxial tensile testing machine
used in our study in detail in previous study. The displacement of
each axis was achieved through an independent and closed-loop
proportional control system. The displacement control error of a
single axis was less than 0.02 mm, and the synchronous error of
the opposing axis was less than 0.02 mm. The heating furnace could
generate and maintain a constant temperature from 20 C to 800 C.
The main difculty in biaxial experiments lay in detecting the
onset of localized necking. Han and Kim (2003) gave the most
immediate way: the strain path slope would suddenly increase
when the material deformed from localized necking to failure. In
this study, the strain around the fracture area was measured by
digital-image correlation (DIC) measurement technology instead of
by using traditional grids, which suffer from limited data and pre-
cision. Xiao et al. (2016a,b) also introduced the process of strain
measurement using DIC in a recent study. In the current study,
the gauge length can be determined using pixels: the diameter of
the center area of the initial specimen is 40 mm, and the pixels at
the two endpoints can be obtained. The number of pixels corre-
sponding to the diameter of the delineation area can be calculated
based on the basis of proportion. As shown in Fig. 6, a tiny crack
emerged in the fracture area. The strain at any position near this
crack area could be measured precisely at the instant immediately
before fracture with the help of images recorded at a frequency
of 50 Hz. Then, the strains at the onset of localized necking were
acquired in accordance with the variation of the strain path slope.
The specimens used in this study had a thickness of 1.0 mm. The
chemical composition of the alloy is listed in Table 1. Prior to the
application of tension, the specimens were heated to the test tem-
perature and maintained at that temperature for 5 min to ensure
a homogeneous temperature distribution throughout the material.
The experiments were performed at three different temperatures:
20, 400, and 600 C. Despite the effect of the tensile speed on the
forming limit capacity of the material in the thermal environment,
a quasi-static experimental process was used in this study in order
to simplify the discussion. To cover the whole forming limit dia-
gram, the stroke ratios were 3:1, 3:2, 3:3, 2:3, and 1:3 (strain rates
along the X-axis and strain rates along the Y-axis) at the base of
3:0.0025 s1 .
3.2. Uniaxial tests
The uniaxial tests were conducted using standard MTS810 rect-
angular dog-bone-shaped samples at 20, 400, and 600 C in three
different directions (0 , 45 , and 90 with respect to the rolling ori-
entation). The variation of the strain with the thickness was used to
estimate the anisotropy. The uniaxial test materials were identical
to those used in the biaxial tests, and the samples were deformed
at a constant strain rate of 0.0025 s1 .
4. Results
4.1. Uniaxial testing results
The equivalent stressstrain () relationship obtained during
the uniaxial testing at different temperatures and along different
directions can be approximated by
 = K n , (2)
where  is the orientation with respect to the material rolling direc-
tion and K is the stress coefcient. This relationship was used for the
simulations described later. Table 2 shows the tensile properties of
the materials; the initial yield stresses were  0 and  90 .
The r-value of metal sheets can be expressed as
w
r = , (3)
t
where w and t are the width and thickness strains of the uniaxial
tests, respectively. In spite of the increase in the anisotropy coef-
cient with an increase in the major strain, Kuwabara et al. (2002)
reported the r values were measured at a plastic strain of = 0.1 on
Table 2
Tensile properties of TA1 at elevated temperatures.

Table 1 Temperature K0 (MPa) K90 (MPa) n0 n90  0 (MPa)  90 (MPa)

The chemical composition (wt%) of TA1. ( C)

20 450.96 426.79 0.133 0.136 123.43 131.29

Alloy Fe Si C N H O Ti
400 391.11 403.55 0.16 0.16 79.66 81.02
TA1 0.3 0.15 0.1 0.05 0.015 0.2 balance 600 333.26 336.21 0.16 0.155 55.60 54.43
358 R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361

Table 3
TA1 anisotropy coefcients at elevated temperatures.

Temperature ( C) r0 r45 r90 rave

20 1.81 2.40 2.22 2.2075

400 1.18 1.77 1.60 1.58
600 0.76 1.28 0.93 1.0625

the basis of experience. Moreover, the r values of the metal should

vary in response to the changes in the temperature. Table 3 shows
the TA1 r values obtained at room and elevated temperatures. In
addition, the averaged r values, which were obtained from the
following equation, were required for the subsequent calculations:

rave = (r90 + 2r45 + r0 ) /4, (4)

4.2. Forming limit diagrams

In an FLD, a conventional forming limit curve (FLC) represents a

plot of the major and minor strains along the vertical and horizon-
tal coordinates, respectively. However, the anisotropy of the metal
sheet brings about changes in the major strain at different angles
along the rolling direction. Hakoyama and Kuwabara (2014) devel-
oped a new form: the limit strains along the rolling direction were
on the horizontal coordinate, whereas those along the transverse
direction were on the vertical coordinate. In this paper, this form is
adopted as shown in Fig. 7. is the strain ratio or the strain path,
given as

= x /y , (5)

where x and y are the strains in the X and Y directions, respec-

tively. For isotropic materials, = 1under equal biaxial tension,
= 0.5 under uniaxial tension along the transverse direction, and
= 2 under uniaxial tension along the rolling direction. In addi-
tion, proting from the DIC measurement system, the actual strain
paths of the biaxial tests could be obtained from the combination
of the x time and y time curves. The uniaxial test results and
the strain paths are also described below. Moreover, in order to
obtain the plane strain, the balance stroke test results were taken
as symmetric points and interpolated contours were determined in
the form of a parabola by the quadratic interpolation method. The
uniaxial test results and intersection point of the coordinate axis
and contours were connected to straight lines. The FLCs related to
anisotropy were identied at different temperatures.

5. Discussion

5.1. Finite element simulation

In order to determine the threshold from the ductile fracture

criterion, the biaxial and uniaxial tension processes were simulated
by nite element analysis in Ansys/Ls-Dyna, which was suggested
by Zhang and Dong (2015) as an explicit numerical code, popularly
used to analyze a variety of impact problems. The selection of the
yield criterion and hardening model determines the accuracy of
the simulation. Because of the evident anisotropy of the materials
at elevated temperatures, Hill48 was adopted under a plane stress
condition, which is given by
Fig. 7. The forming limit diagrams for TA1 at different temperatures.
2 2
(G + H)xx 2Hxx yy + (F + H)yy + 2Nxy = 1, (6)

where r0 = H/G, r45 = (2N-F-G)/G, r90 = H/F.

For the uniaxial tension along two different orientations with
respect to the rolling direction, the work-hardening models (K0 , n0
and K90 , n90 ) and r-values (r0 and r90 ) were used for the simulation. els were K0 and n0 and the r-value was rave . Because of a symmetry
For the biaxial tension, the parameters of the work-hardening mod- model, the displacement ratios were 3:1, 3:2, and 3:3.
R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361 359

Fig. 8. The threshold I and sheet thickness reduction curves.

360 R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361

Fig. 9. Comparison between the simulation and measured results.

Table 4 Table 5
Limit strain values. Material constants C1 and C2 at elevated temperatures.

20 C 400 C 600 C Material constants 20 C 400 C 600 C

f (uniaxial tension in X direction) 0.283 0.389 0.620 C1 (from = 2 to = 1) 11.284 10.922 8.620
f (plane-strain tension in X-axis) 0.227 0.270 0.511 C2 (from = 2 to = 1) 2.387 3.541 5.039
f (uniaxial tension in Y direction) 0.275 0.383 0.613 C1 (from = 0.5 to = 1) 12.028 9.656 8.018
f (plane-strain tension in Y-axis) 0.197 0.297 0.527 C2 (from = 0.5 to = 1) 2.476 3.071 4.620

5.2. Comparison with experimental results 

2(2 + rave )(1 + rave )
= , (11)
Oyane et al. (1980) researched compressive plastic constitutive 1 3(1 + 2rave )
equations of porous materials with consideration of the hydrostatic
where 1 is the major strain. The solutions indicate that the ratios
pressure and suggested the density of the material decreases dur-
 m / and /1 are only related to the anisotropy coefcient. From
ing deformation and that fracture occurs when the density of the
= 2 to = 1, f is from plane-strain and uniaxial tension in the X
material reaches a certain value. The ductile fracture model is given
direction. From = 0.5 to = 1, f is from plane-strain and uniaxial
as
 f
tension in the Y direction. The material constants C1 and C2 in the
m two regions with symmetry with = 1 are given in Table 5 so as to
( + C1 )d = C2 , (7)
0
 exploit the effects of anisotropy on the forming limit.
We dene the threshold I by modifying Eq. (7) as
where  m is the hydrostatic pressure, is the equivalent strain, and  f
C1 and C2 are the material constants. Table 4 shows the limit strain 1 m
I= ( + C1 )d, (12)
values, f , derived from the X and Y directions for uniaxial tension C2 
0
and those calculated from the intersections of the forming limit
contours and the coordinate axis for plane-strain tension. The physical meaning of the threshold is the condition of
For the obvious anisotropy sheet, Takuda et al. (1997) proved the specimen localized necked when I 1. Using the values of
the relationship between the r-value and the ratios of hydrostatic obtained from the simulation results, the threshold I for each node
stress to equivalent stress,  m /, and between the r-value and the could be calculated at any time.
ratios of equivalent strain to strain in the tensile direction, f /1 . The thickness reduction and threshold I were calculated and
The terms in Eq. (7) can be expressed for uniaxial and plane-strain compared. For the biaxial tests, the data were obtained from Ele.
tension states as 0 (see Fig. 4). In the simulated results of displacement ratios of 3:1
(uniaxial) and 3:2, to distinguish the test stroke ratios, the material constants
 derived from = 0.5 to = 1 and from = 2 to = 1 were sub-
m 1 2(2 + rave ) stituted into Eq. (12). Fig. 8 shows the results. It can be seen that
= , (8)
 3 3(1 + rave ) with increasing temperature, the thickness reduction and plastic
 strain decreased simultaneously. Under the same conditions, the
2(2 + rave ) trend shown by each threshold curve on the left was consistent
= , (9)
1 3(1 + rave ) with that shown by the corresponding thickness reduction curve
at the right. The time increased with increasing temperature when
(plane-strain) the threshold I reached 1. Additionally, the tolerance of the thick-
 ness reduction and plastic deformation capacity also increased as
m 1 2(2 + rave )(1 + 2rave ) the strain path changed from = 0.5 to = 1 (or from = 2 to
= , (10)
 3 3(1 + rave ) = 1).
 R. Xiao et al. / Journal of Materials Processing Technology 240 (2017) 354361
It was concluded that the neck occurred at the heart of the speci-
men when the threshold reached 1. Subsequently, the limit strains,
x and y , were obtained from the calculated strains in Ele. 0 (see
Fig. 4). The tted curves obtained by the simulation and ductile
fracture criterion are shown in Fig. 9, along with the experimental
data. It was found that, for the uniaxial tests, the predictive limit
strains were in good agreement with the measured ones, whereas
considerable differences, although acceptable, existed in the biaxial
tests. In short, the FLDs for fracture at different temperatures could
be predicted by the combination of the simulation and the Oyane
ductile fracture criterion.
6. Conclusions
Thermal biaxial tensile testing technology and a digital-image
correlation measurement system were adopted to analyze the
forming limit of TA1 at high temperatures. The main conclusions
drawn are as follows:
(1) The broken specimens proved that the semispherical thickness
reduction design played a key role in the center-section local-
ized neck and fracture that occurred as a consequence of the
application of normal stress.
(2) The FLDs described by anisotropy illustrated that the limit
strains at high temperatures were affected by the orientation
of the tension with respect to the rolling direction.
(3) The experimental results were successfully predicted by the
combination of the nite element simulation and ductile frac-
ture criterion. It is thought that the present approach will enable
the prediction of forming limits at both room and elevated tem-
peratures.
This project has already been completed; however, the forming
limit stress will be studied in the future.
References
Abu-Farha, F., Hector, L.G., Khraisheh, M., 2009. Cruciform-shaped specimens for
elevated temperature biaxial testing of lightweight materials. Jom J. Miner.
Metals Mater. Soc. 61, 4856.
Dawicke, D.S., Pollock, W.D., 1997. Biaxial testing of 2219-t87 aluminum alloy
using cruciform specimens. In: NASA Langley Technical Report Server.
Hakoyama, T., Kuwabara, T., 2014. Forming limit analyses of cold rolled if steel
sheet using differential work hardening model. Procedia Eng. 81, 12461251.
361
Han, H.N., Kim, K.-H., 2003. A ductile fracture criterion in sheet metal forming
process. J. Mater. Process. Technol. 142, 231238.
Hannon, A., Tiernan, P., 2008. A review of planar biaxial tensile test systems for
sheet metal. J. Mater. Process. Technol. 198, 113.
Hill, R., 1952. On discontinuous plastic states, with special reference to localized
necking in thin sheets. J. Mech. Phys. Solids 1, 1930.Kotkunde, N., Srinivasan,
S., Krishna, G., Gupta, A.K., Singh, S.K., 2016. Inuence of material models on
theoretical forming limit diagram prediction for Ti6Al4 V alloy under warm
condition. Trans. Nonferr. Metals Soc. China 26, 736746.
Kulawinski, D., Weidner, A., Henkel, S., Biermann, H., 2015. Isothermal and
thermo-mechanical fatigue behavior of the nickel base superalloy WaspaloyTM
under uniaxial and biaxial-planar loading. Int. J. Fatigue 81, 2136.
Kuwabara, T., Van Bael, A., Iizuka, E., 2002. Measurement and analysis of yield
locus and work hardening characteristics of steel sheets wtih different
r-values. Acta Mater. 50, 37173729.
Kuwabara, T., 2014. 1.06 Biaxial stress testing methods for sheet metals. In:
Yilbas, S.H.F.B.J.V.T. (Ed.), Comprehensive Materials Processing. Elsevier,
Oxford, pp. 95111.
Marciniak, Z., Kuczynski, K., 1967. Limit strains in the processes of stretch-forming
sheet metal. Int. J. Mech. Sci. 9, 609620.
Merklein, M., Huntter, W., Geiger, M., 2008. Characterization of yielding
behavior of sheet metal under biaxial stress condition at elevated
temperatures. CIRP Annals Manufacturing Technology 57, 269274.
Ognedal, A.S., Clausen, A.H., Polanco-Loria, M., Benallal, A., Raka, B., Hopperstad,
O.S., 2012. Experimental and numerical study on the behaviour of PVC and
HDPE in biaxial tension. Mech. Mater. 54, 1831.
Oyane, M., Sato, T., Okimoto, K., Shima, S., 1980. Criteria for ductile fracture and
their application. J. Mech. Work. Technol. 4, 6581.
Smits, A., Van Hemelrijck, D., Philippidis, T.P., Cardon, A., 2006. Design of a
cruciform specimen for biaxial testing of bre reinforced composite laminates.
Compos. Sci. Technol. 66, 964975.
Stren, S., Rice, J.R., 1975. Localized necking in thin sheets. J. Mech. Phys. Solids 23,
421441.
Tadros, A.K., Mellor, P.B., 1978. An experimental study of the in-plane stretching of
sheet metal. Int. J. Mech. Sci. 20, 121133.
Takuda, H., Mori, K., Fujimoto, H., Hatta, N., 1997. Fracture prediction in stretch
forming using nite element simulation combined with ductile fracture
criterion. Arch. Appl. Mech. 67, 143150.
Takuda, H., Mori, K., Takakura, N., Yamaguchi, K., 2000. Finite element analysis of
limit strains in biaxial stretching of sheet metals allowing for ductile fracture.
Int. J. Mech. Sci. 42, 785798.
Xiao, R., Li, X.-X., Lang, L.-H., Chen, Y.-K., Yang, Y.-F., 2016a. Biaxial tensile testing of
cruciform slim superalloy at elevated temperatures. Mater. Des. 94, 286294.
Xiao, R., Li, X.-X., Lang, L.-H., Chen, Y.-K., Ge, Y.-L., 2016b. Design of biaxial tensile
cruciform specimen based on simulation optimization. Adv. Eng. Res. 1,
178186.
Yoshida, K., Kuwabara, T., 2007. Effect of strain hardening behavior on forming
limit stresses of steel tube subjected to nonproportional loading paths. Int. J.
Plast. 23, 12601284.
Zhang, H., Dong, X., 2015. Finite element analysis of multiple solid particles erosion
in cermet coating. Surf. Coat. Technol. 262, 184190.
Zidane, I., Guines, D., Lotoing, L., Ragneau, E., 2010. Development of an in-plane
biaxial test for forming limit curve (FLC) characterization of metallic sheets.
Meas. Sci. Technol. 21, 055701.