International Journal of Plasticity 27 (2011) 11651184

Contents lists available at ScienceDirect

International Journal of Plasticity

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

A new analytical theory for earing generated from anisotropic plasticity

J.W. Yoon a,d,, R.E. Dick b, F. Barlat c
a

Faculty of Engineering & Industrial Sciences, Swinburne University of Technology, P.O. Box 218, Hawthorn, VIC 3122, Australia
Alcoa Technical Center, 100 Technical Dr., Alcoa Center, PA 15069-0001, USA
Graduate Institute of Ferrous Technology, Pohang University of Science and Technology, San 31 Hyoja-Dong, Nam-Gu, Pohang, Gyeongbuk 790-784, South Korea
d
Center for Mechanical Technology & Automation, University of Aveiro, 3810-193 Aveiro, Portugal
b
c

a r t i c l e

i n f o

Article history:
Received 4 December 2009
Received in nal revised form 10 January
2011
Available online 21 January 2011
Keywords:
Earing
r-Value
Yield stress
Analytical approach
Cup height prole

a b s t r a c t
Commercial canmaking processes include drawing, redrawing and several ironing operations. It is experimentally observed that during the drawing and redrawing processes earing develops, but during the ironing processes earing is reduced. It is essential to
understand the earing mechanism during drawing and ironing for an advanced material
modeling. A new analytical approach that relates the earing prole to r-value and yield
stress directionalities is presented in this work. The analytical formula is based on the exact
integration of the logarithmic strain. The derivation is for a cylindrical cup under the plane
stress condition based on rigid perfect plasticity while force equilibrium is not considered.
The earing prole is obtained solely from anisotropic plastic properties in simple tension.
The earing mechanism is explained from the present theory with explicit formulae. It has
been proved that earing is the combination of the contributions from r-value and yield
stress directionalities. From a directionality (y-axis) vs. angle from the rolling (x-axis) plot,
the earing prole is generated to be a scaled mirror image of the r-value directionality with
respect to 90 (x = 90) and also a scaled mirror image of the yield stress directionality with
respect to the reference yield stress (y = 1). Three different materials (Al5% Mg alloy, AA
2090-T3 and AA 3104 RPDT control coil) are considered for verication purposes. This
approach provides a fundamental basis for understanding the earing mechanism. In practice, the present theory is also very useful for the prediction of the earing prole of a drawn
and iron cup and its related convolute cut-edge design for an earless cup.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Generally, r-value and stress directionalities are the key input parameters for phenomenological constitutive models.
These anisotropies are directly related to earing of a drawn cup. For example, Hills (1948) yield function (Hill, 1948) accepts
either r-values or yield stresses along 0, 45, 90 as anisotropy parameters, while Yld91 (Barlat et al., 1991a) uses only the
yield stress values for the balanced biaxial value as well as along the three major directions (0, 45, 90). The Yld2000 model
(Barlat et al., 2003) accommodates both r-value and stress directionalities for the three uniaxial and balanced biaxial directions. The Yld2004 model (Barlat et al., 2005) utilizes r-value and yield stress data every 15 from the rolling as well as a
biaxial datum. Thus, based on the combination of these directionalities, the Yld2004 model is able to predict more than four
ears in cup drawing as shown in Yoon et al. (2006). Characteristics for linear transformation yield functions are well summarized at Barlat et al. (2007). It has been proved that a good prediction of these material directionalities controls the overall
accuracy of the earing prole. However, the exact mechanism as to how much r-value and yield stress directionalities contribute to earing prole respectively has not been established.
Corresponding author.
E-mail address: jyoon@swin.edu.au (J.W. Yoon).
0749-6419/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijplas.2011.01.002

1166

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

Recently, there have been developed many advanced approaches to describe plastic anisotropy. The strain rate potential
is another concept that can describe plastic anisotropy (Barlat et al., 1993; Yoon et al., 1995; Chung et al., 1996; Kim et al.,
2008a,b; Rabahallah et al., 2009; Van Houtte et al., 2009; Cazacu et al., 2010). Non-associated ow plasticity was also implemented to the nite element method to predict plastic anisotropy (Civitanic et al., 2008; Taherizadeh et al., 2009). Experimentally it was shown that the yield surface shape can evolve in complex ways (Kuroda and Tvergaard, 2001; Kuwabara,
2007). For the corresponding modeling, it has been acknowledged that more advanced models should capture the distortion
of the yield surface (Wu et al., 2005; Holmedal et al., 2008; Aretz, 2008; Korkolis and Kyriakides, 2008; Stoughton and Yoon,
2009). Yield criteria to describe plastic anisotropy for complex hcp materials were proposed by Cazacu et al. (2006) and
Plunkett et al. (2008). Anisotropic hardening behaviors for the hcp materials were also investigated (Plunkett et al., 2007;
Nixon et al., 2009). A review on hardening models has been made by Chaboche (2008). Rousselier et al. (2009) also predicted
a complicated earing prole of AA 2090-T3 with a reduced polycrystal approach.
Compared to the above mentioned nite element method, there have been few studies on prediction of the earing prole
based on analytical approaches for a single step cup drawing. Hosford and Caddell (1983) and Chung et al. (1996) provided a
quantitative trend between the r-value anisotropy and the earing prole in a mild steel and an aluminum alloy, respectively.
Using a different approach, Barlat et al. (1991b) attempted to correlate the stress anisotropy (not r-values) to the earing
trends by applying the stress condition at the rim. Recently, an analytical approach considering the r-value directionality
as a main contributor to the earing prole was derived by Yoon et al. (2006). The method provides a simple tool for the prediction of the earing prole using, as input, basic information including the r-value directionality, the initial blank size and
the cup radius. However, the method did not consider the stress directionality. Yoon et al. (2008) simply combined Yoon
et al. (2006) and Barlat et al. (1991b) to consider both r-value and yield stress directionalities on earing prediction. Mulder
and Nagy (2009) further improved Yoon et al. (2008) considering non-uniform strain in the ange and the process effects.
A single cupping operation usually does not produce a cup deep enough for most rigid packaging applications. The can
diameter may be further reduced and the wall height increased by redrawing. Furthermore, if ironing (wall thinning) is employed, a more uniform wall thickness and increased cup height results. The ironing processes typically decreases the degree
of earing. Since the wall thickness in the valleys between ears is greater than at the eared portions, the increase in height due
to ironing is greater at the valleys. A more uniform height (i.e., less pronounced earing) results. Ironing is controlled by the
clearance between the punch and the ironing die land (wall). It increases the total punch force, but does not affect the drawability (Hosford and Caddell, 1983). During the manufacturing of two-piece beverage cans, multiple ironing steps increase
the wall height by about a factor of 3. There is, of course, a limiting reduction for a single ironing stage. Fig. 1 shows an ironing ring and the forces acting on an element of a cup wall during ironing. In the gure, the punch is moving down, while the
ring is stationary. It can be seen that there exists a difference in the way friction acts on the inner and outer walls of the can.
Friction on the ironing ring wall opposes the ow of material, whereas friction on the punch side trends to draw the material
in the direction of punch travel. Therefore, the opposing frictional forces produce severe transverse deformation through the
thickness which makes it difcult to use a normal FE analysis for the process. This is because of the difculties in handling
both anisotropy and double-sided contact with conventional solid elements under severe transverse shear deformation. A
large amount of CPU time is also required by using multi-layered solid elements. Therefore, no one has reported in detail
the evolution of the earing prole during the ironing processes using nite element method.
In this work, a new analytical method simultaneously uses both r-value and yield stress effects to predict earing prole
during both drawing and ironing operations. This method is relatively accurate and uses only seconds of CPU time. A new
closed form equation is based on the exact integration of the logarithmic strain. Especially, the earing mechanism has been
explicitly explained from the directionalities by demonstrating the contribution from each factor to the cup height prole.
For verication purpose, the method has been investigated using three examples from Al5% Mg alloy, AA 2090-T3, and the
AA 3104 control coil.

Fig. 1. Schematic view of ironing process.

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1167

2. Theory
For a quick assessment of the earing prole, it would be advantageous to use a simple analytic formula. It will be demonstrated in this section that the use of the r-value and yield stress directionalities can provide a reasonable approximation
of the earing prole.
The ange area, which being drawn into the inside cavity, can be viewed as a ring in Fig. 2a and the corresponding stress
states on the isotropic initial yield surfaces is shown in Fig. 2b, where von-Mises, Tresca, and Hosford isotropic function with
the exponent a = 8 are demonstrated. The stress state in the ange is pure compression at the outer radius (rhh rYh90 )
and evolves towards a shear state. In the middle of the ange, the stress state is roughly rrr 12 rYh and rhh  12 rYh90 , and
evolves to pure tension at the inside radius (rrr rYh ). For the case of isotropic material, it can be approximated as
rrr  rhh rY . For anisotropic materials, it is necessary to make assumptions to derive an explicit formula. In this work,
as shown in Fig. 2a, earing is assumed to be produced by the following two effects:
(1) Different levels of radial tensile stresses (rrr) during cup drawing generate the different radial velocity eld (called
yield stress effect). Tensile yield stresses (rYh ) are used to model the yield stress anisotropy. The stress mode at
the inner most ange is applied for the entire ange area.

Fig. 2. Deformation of an element on the ange: (a) stress states on the ange; (b) stress states on yield surface.

1168

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

(2) Different levels of compressive strains generate different ratios of the radial and thickness strain (called r value
effect). The r-values (rh+90) are used to model r-value anisotropy. The strain mode at the rim is applied for the entire
ange area. Due to the nature of the compressive strain, it is rotated by 90.
Then, the tensile and compressive contributions to the cup height are superimposed. Consideration of the r-value and
yield stress directionalities as the major source of earing is the main ideas of the method. It will be demonstrated in this
section that the use of the two directionalities can provide a reasonable approximation of the earing prole. The explicit formulae are derived from the above engineering assumptions.
2.1. Prediction of earing prole for a single cup drawing
2.1.1. Derivation of circumferential strain from yield stress directionality
As shown in Fig. 2a, when the material is isotropic, the isotropic circumferential strain for the cup wall is dened as:

 
Rc
for Rc 6 R 6 Rb
R

eISO
h

where the subscripts c and b refer to the cup wall and blank, respectively.
In Eq. (1), the effect of anisotropy on the circumferential strain is not included. The anisotropic contribution is derived
from the yield stress directionality in the rest of this section.
At the ange, it is assumed that rrh = 0. Then, the stress components in the ange with respect to the orthotropic material
coordinates are related to those in cylindrical coordinates by

rxx rrr cos2 h rhh sin2 h

ryy rrr sin2 h rhh cos2 h
rxy rrr  rhh sin h cos h

2-a
2-b
2-c

Taking a yield function of Frxx rrr; rhh ; h; ryy rrr; rhh ; h; rxy rrr; rhh ; h c , assuming yielding occurs completely around the
ange, and using the normality ow rule leads to

@F
@F
@F
drrr
drhh dh 0
@ rrr
@ rhh
@h

3-a

or
@F
@ rrr

@F
dh
@h
@F

rr
@ rhh

dr

drhh

1

3-b

where





@F
@F @ rxx
@F @ ryy
@F @ rxy
@ rxx
@ ryy
@ rxy dh
dh

dh e_ xx
e_ yy
2e_ xy
@h
@ rxx @h
@ ryy @h
@ rxy @h
@h
@h
@h
k_




@F
@F @ rxx
@F @ ryy
@F @ rxy
@ rxx
@ ryy
@ rxy drrr
drrr e_ xx
drrr

e_ yy
2e_ xy
@ rrr
@ rxx @ rrr @ ryy @ rrr @ rxy @ rrr
@ rrr
@ rrr
@ rrr
k_




@F
@F @ rxx
@F @ ryy
@F @ rxy
@ rxx
@ ryy
@ rxy drhh
drhh e_ xx
drhh

e_ yy
2e_ xy
@ rhh
@ rxx @ rhh @ ryy @ rhh @ rxy @ rhh
@ rhh
@ rhh
@ rhh
k_

3-c

In Eq. (3-c), the relationship of e_ ab k_ @@F

rab was applied.
By substituting Eq. (3-c) to Eq. (3-b) together with the strain tensor transformations of

e_ rr e_ xx cos2 h e_ yy sin2 h 2e_ xy sin h cos h

e_ hh e_ xx cos2 h e_ yy sin2 h  2e_ xy sin h cos h
e_ rh e_ yy  e_ xx sin h cos h 2e_ xy cos2 h  sin2 h

4-a
4-b
4-c

Eq. (3-b) leads to

2e_ rh rrr  rhh dh

1
e_ rr drrr e_ hh drhh

Under the assumption of the simple shear condition (e_hh e_ rr ; no thickness change) with the existence of e_rh , Eq. (5) reduces to

drrr  rhh
e_ rh
2 dh 2Cdh
rrr  rhh
e_ hh

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1169

In order to solve Eq. (6) explicitly, the deformation in the ange is assumed along a direction of h. As shown in Fig. 2a, it is
assumed that the radial tensile traction can be modeled from the yield stress directionality (and the compressive deformation is formulated from the r-value directionality in an uncoupled way discussed in Section 2.1.2). Then, the tensile and compressive contributions are superimposed. By assuming that the differences of the radial tensile stresses are attributed from
yield stress directionality, the stress state at the inner most ange is applied for the entire ange area to model yield stress
anisotropy. Then, rrr  rhh rYh by ignoring rhh. Then, Eq. (6) further reduces to

drYh

rYh

2Cdh

7a

Remark. If the compressive contribution is mainly considered from yield stress directionality, Eq. (6) reaches the following
relationship (Barlat et al., 1991b):

drYh90

rYh90

2Cdh

7b

We will discuss the results based on (7a) and (7b) later.

Now lets discuss kinematics to relate Eq. (7a). If vr, vh and vz represent components of the velocity vector at any point of
the ange with respect to cylindrical coordinates (Fig. 2a), the strain rate components can, in principle, be calculated and
related to the stress by the constitutive law. However, the stress state at any point of the ange is unknown and non-uniform. Therefore, two additional assumptions are necessary to calculate the earing prole. Although vh and vz are not equal to
zero, they are negligible when compared to the radial velocity. Moreover, it is assumed that the earing tendency is mainly
imposed by the radial velocity of points located at the rim of the ange. Except at the very end of the drawing operation,
these points are always part of the ange. Using the relationship between the velocity eld and strain rates for each point
on the ange, it is possible to show that (Barlat et al., 1991b)

dv r 2e_ rh
2 Cv r :

dh
e_ hh

Combination of Eqs. (8) and (7a) gives

dv r h drh

v r h rYh

9-a

dln v r h dln rYh

9-b

or

Therefore, it appears that the curve that represents the normalized variation of the radial velocity of the rim of the ange as a
function of angular position is the same as the one representing the normalized stresses variation. Integration of Eq. (9-b)
leads to

ln v r h C 1 ln rYh C 2

10-a

C 1  C 2 ln rYh  ln v r h

10-b

or

Boundary conditions are imposed to remove the constants C1  C2 as

C 1  C 2 ln rref  ln v ref

11

By substituting Eq. (11) to Eq. (10-b), the following relationship can be obtained

ln

v ref
rref
ln Y
v r h
rh

12

Next, consider the green part of Fig. 3 for the elongation from the stress directionality. As a rst approximation, the total cup
height after elongation can be obtained by multiplying the punch speed vp (=vref) by the time tf that the outer edge of the
ange needs to travel from the original position (r = Rb) to the punch radius. This leads to the following relationship between
the decrement of blank radius and the time increment necessary for the decrement:
ANI

h v p t f v p



v ref
Rb  Rc

R  R
v r h
v r h b c

13

In Eq. (13) and later, superscript ANI means the contribution from stress directionality. Then, the denition of the radial
strain is given as follows:

1170

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

Fig. 3. Initial blank and drawn cup: the deformation zones for analytical solution.

ANI

eANI
ln
r

v ref
h h
ln
Rb  Rc
v r h

14-a

rref
rYh

14-b

From Eq. (12),

eANI
ln
r

By applying the simple shear condition, the circumferential and radial strains from the contribution of the yield stress can be
written as

eANI
eANI
ln
h
r

rYh
rref

15

Eq. (15) shows the contribution of stress anisotropy on the circumferential strain. There is no thickness change for Eq. (15)
due to the simple shear assumption.
The procedure from Eqs. (2)(15) can be simply veried from a moment equilibrium by assuming no thickness change as

rYh to Lh rref to Lref

16-a

Then,

eANI
ln
r

Lh
Lref

rref
rYh

16-b

Eq. (16-b) is the same with Eq. (14-b). The average and directional circumferential strains are dened in Eqs. (1) and (15). By
merging the two equations, the total circumferential strain can be dened as
ANI
eh eISO
h beh

0
rYh
Rc
Rc
ln b ln
ln @
R
rref
R

rYh
rref

!b 1
A

17

In Eq. (17), a deceleration factor, b, is introduced. When the stress mode at the inner most ange is applied for the entire
ange (as shown in Fig. 2a), b = 1.0. If the linear distribution of the radial tensile stress is assumed, b = 0.5. The recommended
value of b is 0.5 6 b 6 1. For the isotropic materials, Eq. (17) reduces to Eq. (1).

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1171

2.1.2. Height considering yield stress and r-value directionalities

The normality rule for a circumferential strain is combined with Eq. (17) as

eh

0
@F
Rc
k
ln @
@ rh
R

rYh
rref

!b 1
A

18

where F is a yield function. Then, from Eq. (18), the plastic parameter k is dened as

rYh
rref

Rc
k ln @
R

!b 1 ,
@F
A
@ rh

19

By using the normality rule for er together with Eq. (19), the total radial strain becomes

0
@F
@F
Rc
@ rr
k

ln @
@ rr @@F
R
rh

er

rYh
rref

!b 1
A

20

@F
The quantity of @@F
rr = @ rh has the following ratio by using the normality rule:

@F
@ rr

@F
@ rh



  
@F
@F
er
k
k

@ rr
@ rh
eh

21

The ratio varies in the ange. However, in order to derive an explicit formula, it is assumed that the deformation at the rim
dominates the earing. As shown in Fig. 2, the rim behavior in the direction dened by h is controlled by the property of the
material in compression in the direction dened by h + 90. Assuming that, for a given direction, uniaxial tension and compression lead to identical r-values, these can be expressed as a function of the strains at the rim:

rh90

er
er

et
er eh

22

Here, the subscripts r, , t correspond to the radial, circumferential, and thickness directions, respectively. It is assumed that
the r-values are constant and the strains in Eq. (22) are plastic strain. From Eq. (21), the following relationship is obtained:

eh : er rh90 1 : rh90
er : et rh90 : 1

23-a
23-b

Then,

eh : er : et jh rh90 1 : rh90 : 1

23-c

In order to model r-value anisotropy, the quantity of

(23-a) at the rim, i.e.,

@F @F
=
@ rr @ rh

 

er
eh

at RRb


at RRb

@F
@ rr

= @@F
rh

can be dened as a constant term by using the relationship of Eq.

r h90
1 r h90

24

Then, substituting of Eq. (24) into Eq. (20), the total radial strain becomes

er 

r
rref

Y
h

r h90
ln @
1 r h90

!b

Rc A
ln @
R

rref
rYh

!b

RA
Rc

rh90
1rh90


25

rY

If rh 1, Eq. (25) reduces to Yoon et al. (2006).

ref
The thickness strain can be also derived from the relationship of Eq. (23-b) (the incompressibility condition) as

et

er
r h90

0
ln @

rref
rYh

!b

RA
Rc

1
1r h90


26

Eq. (26) is valid only for the region where et > 0.

Finally, the total height of a cup can be obtained from the logarithmic integral of Eq. (25) as follows:

Hcup h t o r c

Rb

Rc

where

exper dR

27-a

1172

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

Rb

exper dR

Rc

Rb

Rc

rref
rYh

!b

RA
Rc

rh90
1rh90


dR

27-b

In Eq. (27-b), the mathematical relationship of exp(In(a)) = a is applied. The integrated form of Eq. (27) can be expressed as

Hcup h t o rc



1
Bh Ah90
dAh90 
d
Ah90 1
Rb

28-a

where

Ah90
Bh

r h90
1 rh90
!b

28-b

rref
rYh

28-c

Rb
Rc
0:5 6 b 6 1
 Z 2p

rref
rYh dh =2p
d

28-d
28-e
28-f

for the quarter symmetry with every 15 of data

rref 1=12r0 2r15 r30 r45 r60 r75 r90

It is further interesting to derive the specic contributions from r-value and yield stress directionalities to the cup height. For
r
this purpose, an isotropic contribution can be derived with rh+90 = 1 and rref
Y 1 as

iso



Rb
1
to rc
d0:5 
d
1:5

29
r

The cup height from r-value directionality is obtained by excluding yield stress contribution with the assumption of rref
Y 1,
h
i.e.,

Hrv alue h t o rc



1
dAh90 
d
Ah90 1
Rb

30

In a similar approach, the cup height from yield stress directionality is obtained by using rh+90 = 1, i.e.,

Hstress h to rc



Rb
1
Bh 0:5
d0:5 
d
1:5

31

Then, the cup height contributions from r-value and yield stress directionalities can be derived as

DHrv alue h Hrv alue h  Hiso

32-a

DHstress h Hstress h  Hiso

32-b

and

Eq. (32) is useful to understand the explicit contributions from r-value and yield stress directionalities (discussed later).
2.2. Cup height elongation from ironing process
As can be seen in Fig. 4, H is the ironing starting height and the corresponding ironed thickness at H is also dened as tI ,
which is located on the cup wall. Usually, tI is constant and H is a weak function of h, i.e., nearly constant. In order to simplify the ironing process, it is assumed that H is constant and its value is dened from the process directly. Then, the current
work is concentrated on deriving the formula above H, because it difcult to accurately predict the thickness distribution
below H. Then, the corresponding initial radius of H is dened as Rh , which is located at Rx < Rh < Rb , where Rh is not constant with respect to h. Rh can be determined by an iterative procedure when H is given (will be discussed later).
By utilizing the relationship of Eq. (26), thickness strain is dened as

et

0
1

ln @
1 rh90

rref
rYh

!b

1
RA
:
Rc

33

Again, Eq. (33) is only valued for the region et > 0 (not accurate below H). Then, the wall thickness after a cup drawing can be
approximated as

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1173

Fig. 4. Analytical approximation of ironing process.

rref
rYh

tjh to expet jh t0 @

!b

1r 1 1
h90
RA
:
Rc

34

The additional thickness strain change from t|h to tI during the ironing process can be obtained as

eIt jh In

tI
tjh

<0

35

Unlike cup drawing and redrawing, ironing is characterized by the plane-strain ow, since eh = 0. The thickness reduction
directly contributes to the increase in the cup height as

I
r h

I
t h

e

0
 
to @
ln 
tI

rref
rYh

!b

1r 1 1
h90
RA
Rc

36

From Eqs. (25) and (36), the total radial strain during drawing and ironing is
Total
jh
r

cup
r jh

I
r jh

 
to
ln 
tI

rref
rYh

!b  
R
Rc

37

Then, the total cup height after ironing is

HIron h H

Rb

Rh

expeTotal
jh dR
r

Rb

Rh

 
to
tI

rref
rYh

!b  
R
dR
Rc

38

Integrating Eq. (38) leads to

0
H

Iron

h H @


rref
rYh

!b

1
to A R2b  R2
h
2Rc
t I

39

Alternatively, Eq. (39) can be derived from the volume constancy as follows:

Total
h

R
ln @
Rc

rref
rYh

!b 1
A

(from Eq. (17), since eh = 0 for ironing process) and

40

1174

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

eTotal
In
t

 
tI
to

41

Then,

eTotal
eTotal
 eTotal
r
h
t

to
ln @ 
tI

rref
rYh

!b

1
RA
Rc

42

Eq. (42) coincides with Eq. (37), which leads to the same equation of Eq. (39).
In Eq. (39), Rh is still unknown. This value can be determined from an iterative procedure as follows:

Fh Hcup Rh  H 0

43-a

!
  Ah90
Rh
Rc
to rc
  Bh Ah90  H 0
Ah90 1
Rc
Rh

43-b

or

Rh

Then, by the linear approximation,

f h


df h 
df h

 DRh 0 ! DRh f h

dRh
dRh

44-c

and

Rh Rh DRh

44-d

3. Verication
For the verication, cup drawing examples were evaluated using three materials i.e., Al5% Mg, AA 2090-T3, and AA 3104
control coil sheets. The yield stress and r-value data for the three materials are given in Fig. 5. In the gure, the anisotropic
data for AA 3104 control coil is optimized based on the Visco-Plastic Self Consistent model (Lebensohn and Tom, 1993)
(not from experiment) due to very limited elongation during uniaxial tension test. The sketch of the cup drawing process
is shown in Fig. 6. The specic dimensions of the tools are given as follows:
<Al5%Mg>
Punch diameter: Dp = 50 mm
Punch prole radius: rp = 5 mm
Die opening diameter: Dd = 52.8 mm
Die prole radius: rd = 5 mm
Blank radius: Db = 100 mm
Sheet thickness: t = 1.0 mm
<AA 2090-T3>
Punch diameter: Dp = 97.46 mm
Punch prole radius: rp = 12.70 mm
Die opening diameter: Dd = 101.48 mm
Die prole radius: rd = 12.70 mm
Blank radius: Db = 158.76 mm
Sheet thickness: t = 1.6 mm
<AA 3104 control coil>
Punch diameter: Dp = 35.560 mm
Punch prole radius: rp = 2.286
Die opening diameter: Dd = 36.576 mm
Die prole radius: rd = 2.286 mm
Blank radius: Db = 76.123 mm
Sheet thickness: t = 0.457 mm
3.1. Application to drawing
Based on the material characteristics in Fig. 5 and the cup geometry information in Fig. 6, the earing proles can be
predicted using the analytical equation (Eq. (28)), which involves the r values and stress ratios as well as the geometrical

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

AA 2090-T3

(a)

Al-5%Mg

1175

AA 3104

r-value

1.5

0.5

10

20

30

40

50

60

70

80

90

Angle from Rolling (Deg.)

(b)

AA 2090-T3

Al-5% Mg

AA 3104

Normalized yield stress

1.1
1.05
1

0.95
0.9
0.85
0.8
0.75
0

10

20

30

40

50

60

70

80

90

Angle from Rolling (Deg.)

Fig. 5. Anisotropic characteristics for Al5% Mg alloy, AA 2090-T3, and AA 3104 control coil: (a) r-value plot; (b) yield stress plot.

Fig. 6. Schematic diagram for circular cup drawing.

features of the cup drawing test. The cup radius, Rc, is located between Dp/2 6 Rc 6 Dd/2. Then, the drawing ratio, d = Rb/Rc,
can be selected between Db/Dp  d 6 d(Rb/Rc) 6 Db/Dd + d to t the best average cup height (d is a small number). Also, the
deceleration factor of b is selected as b = 1. A paramedic study for b is included in AA 2090-T3 sample.

3.1.1. Al5% Mg alloy

The binary Al5% Mg aluminum alloy sheet sample was tested after cold rolling to 1 mm gauge and annealing. Material
characterization and FE analysis for the alloy is given in Yoon et al. (2004). The main purpose of this example is to compare
the present theory with nite element predictions. Fig. 7 shows the comparisons of r-value and yield stress directionalities.
As can be seen in Fig. 7a, this alloy has low r-values with the maximum of 0.58 at 45. The r-values in the rolling and

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

(a)

EXPERIMENT (Yoon et al., 2004)

Yld2000-2d (Barlat at al., 2003)
Yld91 (Barlat et al., 1991a)
1

0.9

r-value

0.8

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

10

20

30

40

50

60

70

80

90

Angle from Rolling (Deg.)

(b)
Normalized yield stress

1176

EXPERIMENT (Yoon et al., 2004)

Yld2000-2d (Barlat et al., 2003)
Yld91 (Barlat et al., 1991a)
1.1

1.05

0.95
0

10

20

30

40

50

60

70

80

90

Angle from Rolling (Deg.)

Fig. 7. Comparison of directionalities for Al5% Mg alloy: (a) r-value plot; (b) yield stress plot.

Fig. 8. Al5% Mg alloy: cup height proles.

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1177

transverse directions are approximately equal. Yld91 predicts a monotonic increase of r-value. It will be investigated in the
next section how the r-value directionality inuences the earing prole. In Fig. 7b, the yield stress in the transverse direction
is relatively high compared to the yield stress in the rolling direction. Both Yld2000-2d and Yld91 predict very similar results.

(c)

Fig. 9. Cup height contributions from r-value and yield stress directionalities: (a) based on experimental tting; (b) based on Yld2000-2d tting; (c) based
on Yld91 tting.

1178

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

Fig. 8 shows the predicted earing proles from the present theory when Fig. 7 is used as the directionality input. As reference solutions, the gure includes the results obtained from the nite element analysis based on the Yld2000-2d and Yld91
models (Yoon et al., 2004). The cup height was normalized by the rolling direction value in order to see the relative trends of
earing. The present theory shows good correlation with the nite element results. Fig. 9 shows the cup height contributions
from r-value and yield stress directionalities. The gure is predicted from Eq. (28) in this work. It is clear that the yield stress
directionality contributes to monotonic decrease of the earing prole. Then, the cup height at 90 is lower than the one at the
rolling. It is worth mentioning that the earing prole is a scaled mirror image of r-value plot in Fig. 7a with respect to 90 and
is also a scaled mirror image of the yield stress plot in Fig. 7b with respect to the normalized yield stress (y = 1.0 in the plot).
Figs. 7 and 9c explains why the earing prole of Yld91in Fig. 8 shows the monotonic decrease.
In order to investigate the yield stress effect in further detail, the results from Yoon et al. (2006) and from the compression
stress mode at the rim Eq. (7b) were included in Fig. 10. The result from Yoon et al. (2006) based on only r-value contribution shows a larger earing amplitude at 45 and almost the same heights at 0 and 90. The result based on Eq. (7b) shows
an opposite earing prole at 0 and 90 (the cup height at 90 is higher than the height at 0, which is an incorrect trend).
Through this example, it is found that the yield stress contribution plays an important role in predicting the correct earing
trend.
The frictional effect on earing has been demonstrated through the results from the nite element method. Fig. 11 shows
the results for various friction coefcients. It can be shown that friction does not change the earing trend and only slightly
changes the cup height. Thus, the analytical formula in this work ignored the frictional effect.

3.1.2. AA 2090-T3 alloy

AA 2090-T3 shows the complete range of distribution of r-value and yield stresses shown in Fig. 5. It was reported that the
material produces six ears during cup drawing. Predictions of the cup height are displayed in Fig. 12a. Because symmetry
was not maintained during the experiment, the data from 0 to 360 are shown in the gure. The result from Yoon et al.

Fig. 10. Comparisons of cup height proles based on different theories for Al5% Mg alloy.

Fig. 11. Comparison of earing proles for Al5%Mg calculated using different coefcients of friction.

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1179

(a)

(b)

(c)

Fig. 12. AA 2090-T3: (a) cup height proles; (b) cup height contribution from r-value and yield stress directionalities; (c) effect of b on earing.

(2006) based on the nite element method using Yld2004-18p model and an r-value driven analytical solution based on
Yoon et al. (2006) are included in the plot. The present theory which considered both r-value and yield stress directionalities
shows a remarkable improvement compared to the r-value based solution. The present theory is consistent with the experiment and the prediction from Yld2004-18p. Fig. 12b shows the cup height contribution from the r-value and yield stress
directionalities. For this alloy, the contribution from the stress directionality is larger than r-value contribution. Therefore,
this effect should not be ignored. Fig. 12c shows the effect of b parameter on the earing prole. For b = 0.5, it gives a better
prediction for the small ears, while the case of b = 1.0 shows a better estimation for the big ears.

1180

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

Further investigation has been made with the present theory by using the directionalities predicted from Yld91 and Hills
(1948) criteria. Fig. 13 shows the r-value and yield stress directionalities predicted from the two criteria. Yld91 shows a large
deviation for the r-value plot and Hill (1948) overestimates the yield stress directionality. Fig. 14 shows the predicted earing
proles from the present theory when Fig. 13 is used as the directionality input. It is noticeable that Hills (1948) predicts
larger earing with the correct trend while Yld91 predicts a higher cup height at 90 compared to the rolling direction.
The predictions are consistent with the nite element results shown in Yoon et al. (2000). The earing mechanism has not
been explicitly explained yet. Now the reason can be explained from Fig. 15. In the gure, the contributions of r-value
and yield stress directionalities from Eq. (32) are correlated with Fig. 13 directly. From the two gures, we can conclude that
the earing prole is a scaled mirror image of r-value plot with respect to 90 and also a scaled mirror image of the yield stress
plot with respect to the normalized yield stress.
3.2. Application to drawing and ironing
A drawing and ironing example was considered using the AA 3104 control coil. It exhibits eight ears in the experiment. To
predict the evolution of earing proles during drawing and ironing processes, the additional information shown in Fig. 4 is
required. In this study, the target wall thickness and the starting height of ironing are set to tI 0:508 mm (0.02 in.) and
H = 12.7 mm (0.5 in.), respectively. The H (=12.7 mm) gives the Rh based on Eq. (43). The values are summarized as (unit:
mm)
R0

R15

R30

R45

R60

R75

R90

31.981

32.032

32.037

31.941

31.991

32.212

32.214

(a)

EXPERIMENT (Yoon et al., 2000)

Hill (1948)
Yld91 (Barlat et al., 1991a)

4.5

r-value

3.5
3
2.5
2
1.5
1
0.5
0
0

10

20

30

40

50

60

70

80

90

Angle from Rolling (Deg.)

Normalized yield stress

(b)

EXPERIMENT (Yoon et al., 2000)

Hill (1948)
Yld91 (Barlat et al., 1991a)
1.65
1.55
1.45
1.35

1.25
1.15
1.05
0.95
0.85
0.75
0

10

20

30

40

50

60

70

80

90

Angle from Rolling (Deg.)

Fig. 13. Comparison of directionalities for AA 2090-T3 alloy: (a) r-value plot; (b) yield stress plot.

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

Fig. 14. AA 2090-T3: comparison of cup height proles.

(a)

(b)

Fig. 15. Cup height contributions from r-value and yield stress directionalities: (a) based on Yld91 tting; (b) based on Hill (1948) tting.

1181

1182

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

The experimentally drawn and ironed cups are displayed in Fig. 16. As can be seen in Fig. 5, AA 3104 shows a complicated
r-value shape, but the yield stress directionality is small compared to the other two alloys in the gure. The anisotropic data
is not measured, but estimated from the VPSC model as mentioned before. In Fig. 17a, the earing prole after cup drawing
was predicted using the analytical equation, Eq. (32) and it was also compared with the analytical solution of Yoon et al.
(2006). In Fig. 17b, the contribution from the yield stress is much smaller than the one from the r-value. The results from
the present theory and Yoon et al. (2006) show similar predictions in Fig. 17a. In Fig. 18, the comparison of the cup height

Fig. 16. Experimentally obtained cups: (a) after cup drawing; (b) after ironing.

(a)

(b)

Fig. 17. AA 3104: (a) cup height proles; (b) cup height contribution from r-value and yield stress directionalities.

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

1183

Fig. 18. Cup height proles after ironing for AA 3014.

proles was made after ironing. The present theory shows excellent agreement with the experimental data. This is because
the r-value distribution does not inuence the cup height change for an ironing process as shown in Eq. (39). Unlike cupping
and redrawing, ironing does not benet from high r-values. This is because ironing is characterized by the same plane-strain
ow, eh = 0 (where Rc is cup radius). The only inuence of the r-value is its effect on the wall thickness changes before ironing. With a high r-value, there is less wall thickness in the initial drawing and redrawing steps, so less ironing is required to
achieve the same wall thickness. The metal ow in ironing can be regarded as similar to plane-strain drawing. Therefore,
ironing can be treated as a uniform redistribution of metal volume in the cup sidewall. Although the yield stress contribution
to the cup height is small as shown in Fig. 17b, it has a signicant contribution to the change of the earing prole which leads
to excellent agreement with the experiment. Fig. 19 shows the earing proles after drawing and ironing. It is worth mentioning that the overall earing magnitude was reduced to one third after ironing. Also, the cup height at the 45 is higher
than the height at 0 after ironing.

4. Conclusions
Analytical equations were derived in this work to provide an approximation of the earing prole of drawn and ironed
cups. The analytical model considers both the r-value and yield stress directionalities simultaneously. The earing mechanism
is explained based on the proposed theory. The theory is consistent with the results from FEM and experiments. It was
found that the yield stress contribution is important to predict the correct earing trend and the r-value has an indirect contribution to earing in the ironing process. The earing progression during cup drawing and ironing can be easily traced from

Fig. 19. Predicted earing proles and comparisons with experimental data after cup drawing and ironing.

1184

J.W. Yoon et al. / International Journal of Plasticity 27 (2011) 11651184

the analytical equations. The developed formula can be efciently used for a convolute cut-edge design by considering both
the drawing and ironing operations.
Acknowledgements
The authors are thankful to Dr. H. Mulder at Corus for discussion and invaluable comments on the formulation of Yoon
et al. (2008) published in the NUMISHEET2008 proceeding. The authors are also grateful to Drs. M.E. Karabin and G.H. Bray at
Alcoa Technical Center for their critical reviews. This work is also partially supported by the FCT project of PTDC/EME-TME/
109119/2008. The authors are very thankful for this support.
References
Aretz, H., 2008. A simple isotropic-distortional hardening model and its application in elasticplastic analysis of localized necking in orthotropic sheet
metal. Int. J. Plasticity 24, 14571480.
Barlat, F., Lege, D.J., Brem, J.C., 1991a. A six-component yield function for anisotropic metals. Int. J. Plasticity 7, 693.
Barlat, F., Panchanadeeswaran, S., Richmond, O., 1991b. Earing in cup drawing face-centered cubic single crystals and polycrystals. Metall. Trans. A 22, 1525.
Barlat, F., Chung, K., Richmond, O., 1993. Strain rate potential for metals and its application to minimum plastic work path calculations. Int. J. Plasticity 9,
5163.
Barlat, F., Brem, J.C., Yoon, J.W., Chung, K., Dick, R.E., Choi, S.H., Pourboghrat, F., Chu, E., Lege, D.J., 2003. Plane stress yield function for aluminum alloy sheets.
Int. J. Plasticity 19, 1297.
Barlat, F., Aretz, H., Yoon, J.W., Karabin, M.E., Brem, J.C., Dick, R.E., 2005. Linear transformation based anisotropic yield function. Int. J. Plasticity 21, 1009.
Barlat, F., Yoon, J.W., Cazacu, O., 2007. On linear transformation of stress tensors for the description of plastic anisotropy. Int. J. Plasticity 23, 876896.
Cazacu, O., Plunkett, B., Barlat, F., 2006. Orthotropic yield criterion for hexagonal close packed metals. Int. J. Plasticity 22, 11711194.
Cazacu, O., Ionescu, R., Yoon, J.W., 2010. Orthotropic strain rate potential for the description of anisotropy in tension and compression of metals. Int. J.
Plasticity 26, 887904.
Chung, K., Lee, S.Y., Barlat, F., Keum, Y.T., Park, J.M., 1996. Finite element simulation of sheet forming based on a planar anisotropic strain-rate potential. Int.
J. Plasticity 12, 93.
Chaboche, J.L., 2008. A review of some plasticity and viscoplasticity constitutive theories. Int. J. Plasticity 24, 16421693.
Civitanic, V., Vlak, F., Lozina, Z., 2008. A nite element formulation based on non-associated plasticity for sheet metal forming. Int. J. Plasticity 24, 646687.
Hill, R., 1948. A theory of the yielding and plastic ow of anisotropic metals. Proc. Roy. Soc. London A 193, 281297.
Holmedal, B., Van Houtte, P., An, Y., 2008. A crystal plasticity model for strain-path changes in metals. Int. J. Plasticity 24, 13601379.
Hosford, W.F., Caddell, R.M., 1983. Metal Forming: Mechanics and Metallurgy. Prentice-Hall, Inc., Englewood Cliffs, NJ.
Kim, D., Barlat, F., Bouvier, S., Rabahallah, M., Balan, T., Chung, K., 2008a. Non-quadratic anisotropic potentials based on linear transformation of plastic
strain rate. Int. J. Plasticity 23, 13801399.
Kim, J.H., Lee, M.G., Barlat, F., Wagoner, R.H., Chung, K., 2008b. An elasto-plastic constitutive model with plastic strain rate potentials for anisotropic cubic
metals. Int. J. Plasticity 24, 22982334.
Korkolis, Y.P., Kyriakides, S., 2008. Ination and burst of aluminum tubes. Part II: An advanced yield function including deformation-induced anisotropy. Int.
J. Plasticity 24, 16251637.
Kuroda, M., Tvergaard, V., 2001. A phenomenological plasticity model with non-normality effects representing observations in crystal plasticity. J. Mech.
Phys. Solids 49, 12391263.
Kuwabara, T., 2007. Advances in experiments on metal sheets and tubes in support of constitutive modeling and forming simulations. Int. J. Plasticity 23,
385419.
Lebensohn, R.A., Tom, C.N., 1993. A self-consistent anisotropic approach for the simulation of plastic deformation and texture development of polycrystals
application to zirconium alloys. Acta Metall. Mater. 41, 26112624.
Mulder, J., Nagy, G.T., 2009. Earing and wall thickness in cylindrical cup drawing. In: Onate, E., Owen, D.R.J. (Eds.), COMPLAS X Proceeding, Barcelona, Spain,
pp. 18.
Nixon, M.E., Cazacu, O., Lebensohn, R.A., 2009. Anisotropic response of high a-purity-titanium: experimental characterization and constitutive modeling.
Int. J. Plasticity. doi:10.1016/j.ijplas.2009.08.007 (available online).
Plunkett, B., Cazacu, O., Lebenson, R.A., Barlat, F., 2007. Elasticviscoplastic anisotropic modeling of textured metals and validation using the Taylor cylinder
impact test. Int. J. Plasticity. 23, 10011021.
Plunkett, B., Cazacu, O., Barlat, F., 2008. Orthotropic yield criteria for description of the anisotropy in tension and compression of sheet metals. Int. J.
Plasticity 24, 847866.
Rabahallah, M., Balan, T., Bouvier, S., Bacroix, B., Barlat, F., Chung, K., Teodosiu, C., 2009. Parameter identication of advanced plastic strain rate potentials
and impact on plastic anisotropy prediction. Int. J. Plasticity 25, 491512.
Rousselier, G., Barlat, F., Yoon, J.W., 2009. A novel approach for anisotropic hardening modeling Part I: Theory and its application to nite element analysis
of deep drawing. Int. J. Plasticity 25, 23832409.
Stoughton, T.B., Yoon, J.W., 2009. Anisotropic hardening and non-associated ow in proportional loading of sheet metals. Int. J. Plasticity 25, 17771817.
Taherizadeh, A., Green, D.E., Ghaei, A., Yoon, J.W., 2009. A non-associated constitutive model with mixed iso-kinematic hardening for nite element
simulation of sheet metal forming. Int. J. Plasticity 26, 288309.
Van Houtte, P., Yerra, S.K., Van Bael, A., 2009. The Facet method: a hierarchical multilevel modelling scheme for anisotropic convex plastic potentials. Int. J.
Plasticity 25, 332360.
Wu, P.D., MacEwen, S.R., Lloyd, D.J., Jain, M., Tugcu, P., Neale, K.W., 2005. On pre-straining and the evolution of material anisotropy in sheet metals. Int. J.
Plasticity 21, 723739.
Yoon, J.W., Song, I.S., Yang, D.Y., Chung, K., Barlat, F., 1995. Finite element method for sheet forming based on an anisotropic strain-rate potential and the
convected coordinate system. Int. J. Mech. Sci. 37, 733752.
Yoon, J.W., Barlat, F., Chung, K., Pourboghrat, F., Yang, D.Y., 2000. Earing predictions based on asymmetric nonquadratic yield function. Int. J. Plasticity 16,
1075.
Yoon, J.W., Barlat, F., Dick, R.E., Chung, K., Kang, T.J., 2004. Plane stress yield function for aluminum alloy sheet Part II: FE formulation and its
implementation. Int. J. Plasticity 20, 495522.
Yoon, J.W., Barlat, F., Dick, R.E., Karabin, M.E., 2006. Prediction of six or eight ears in a drawn cup based on a new anisotropic yield function. Int. J. Plasticity
22, 174193.
Yoon, J.W., Dick, R.E., Barlat, F., 2008. Analytical prediction of earing for drawn and ironed cups. In: Hora, P. (Ed.), NUMISHEET2008 Proceeding, Interlaken,
Switzerland, pp. 97100.