You are on page 1of 8

Journal of Materials Processing Technology 210 (2010) 585592

Contents lists available at ScienceDirect


Journal of Materials Processing Technology
j our nal homepage: www. el sevi er . com/ l ocat e/ j mat pr ot ec
Asymmetric forming of aluminum sheets by synchronous spinning
Ichiro Shimizu

Graduate School of Natural Science and Technology, Okayama University, 3-1-1 Tsushima-naka, Okayama 700-8530, Japan
a r t i c l e i n f o
Article history:
Received 2 June 2009
Received in revised form 7 October 2009
Accepted 5 November 2009
Keywords:
Metal forming
Asymmetric spinning
Sheet metal
Incremental forming
Strain distribution
a b s t r a c t
Asymmetric conventional metal spinning was performed in this study to break through the limitation of
ordinaryaxisymmetric spinning. Asynchronous spinning machine was developedfor this purpose; inthis
case, themandrel motion, mandrel feed, androller feedweresynchronizedbypulsecontrol. Trial spinning
experiments were carried out on a 1-mm-thick aluminum sheet to obtain an asymmetric truncated-
elliptical-cone-shaped product. The roller pathwas traced using a control software by taking into account
the errors caused by the step pulse control. Asymmetric sheet metal forming was successfully carried out
by using the developed synchronous spinning method. The characteristics of asymmetric spinning were
discussed by comparing the shape of the product, strain distributions, and force components with those
obtained in the case of axisymmetric spinning.
2009 Elsevier B.V. All rights reserved.
1. Introduction
Spinning is one of the traditional metal forming processes used
to fabricate hollow products from metal sheets or tubes. Spinning
is expected to be advantageous for the small- and medium-lot pro-
duction of hollow metal products, as the equipment required is
inexpensive, and the forming load is very small, as be pointed out
by Wong et al. (2003).
In spinning processes, a metal sheet (workpiece) is supported
on a rotating mandrel, while a tool is used to deformthe workpiece
over the mandrel; therefore, only products with an axisymmet-
ric shape are obtained in this method. In the last few decades,
several researchers attempted to carry out asymmetric spinning
in order to break through the limitation of ordinary axisymmet-
ric spinning. Amano and Tamura (1984) were probably the rst
who succeeded the asymmetric elliptical cone spinning by means
of a conventional spinning machine equipped the combined mech-
anism of cams and links, which enabled the work roll to follow
an elliptic contour of mandrel. It was epoch-making achievement,
however, their methodrequiredthereplacements of cams andlinks
with the shape change of products. Gao et al. (1999) developed a
newmethodcalledellipse spinning, inwhichoff-axis rotationof a
metal sheet (workpiece) is carried out so that the roller movement
results in the formation of products with an elliptical shape. This
method was in contrast to the one by Amano and Tamura (1984);
the mandrel motion was reciprocated instead of the roller motion.
However, the adaptability of the method is basically limited to

Tel.: +81 86 251 8033; fax: +81 86 251 8266.


E-mail address: shimizu@mech.okayama-u.ac.jp.
elliptical-shape products. The trial of asymmetric spinning was also
performed by Awiszus and Meyer (2005). Although they succeeded
to form the tripod-shaped workpiece, they used the spring-loaded
spinning roller that was insufcient for the precise forming. Arai
(2005) designed a new apparatus for asymmetric spinning with
force feedback control to secure the workpiece (metal sheet) onto
a mandrel having an asymmetric shape. This apparatus can be used
to produce hollowproducts withvarious asymmetric shapes. How-
ever, the force feedback control system required for this purpose
is sophisticated and expensive. In addition, it is difcult to achieve
displacement control, which is crucial for precise metal forming.
Recently, the author has developed a new synchronous spin-
ning machine, in which mandrel rotation, mandrel feed, and roller
feed are precisely controlled by means of pulse-controlled actua-
tors and a stepping motor. The developed machine cansynchronize
the mandrel rotation and the roller motion, that is impossible by
standard spinning machines. The machine can be applied to var-
ious product shapes only by changing control software code, that
means the machine is much more exible than the previous ones
developed by Amano and Tamura (1984) and by Gao et al. (1999).
The objective of this study is to show the adaptability of the devel-
oped synchronous spinning machine to asymmetric products and
toclarifythecharacteristics of theasymmetric spin-formedproduct
by the developed method. Although synchronous control appears
to be a less intelligent approach than force feedback control, it
requires relatively simple equipment and is inexpensive; hence, it
is preferred for the small-lot production of hollow metal products.
As the rst trial using the developedmethod, the conventional one-
path spinning of pure industrial aluminum sheets was performed
to obtain asymmetric truncated-elliptical-cone-shaped products.
In conventional spinning, the outer diameter of the nal product
0924-0136/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmatprotec.2009.11.002
586 I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592
Table 1
Specications of stepping motor and electric actuators employed in the synchronous spinning machine.
Holding torque (Nm) Max. thrust force (N) Resolution of motion
Stepping motor for mandrel rotation (Oriental Motor CSK299) 4.4 0.9

Electric actuator for mandrel feed (Oriental Motor PWA100) 1000 0.005mm
Electric actuator for roller feed (Oriental Motor DRL60PA4) 300 0.004mm
becomes less than that of the starting workpiece. The shape of the
resulting product, strain distribution, and force components were
compared with those obtained in the case of axisymmetric trun-
cated cone spinning. The important characteristics of asymmetric
synchronous spinning are discussed on the basis of the results of
this comparison.
2. Synchronous spinning machine
The spinning machine developed in the present study is based
on synchronous control of the mandrel rotation, mandrel feed, and
roller feed. Synchronous control is relatively simple to achieve, and
it facilitates the formation of hollowmetals with various asymmet-
ric shapes. Fig. 1shows theschematic of thedevelopedsynchronous
spinning machine. The following components are connected to
the motion controller (Interface Co. Ltd., PCI-7414V): a stepping
motor for controlling the mandrel rotation (Oriental Motor Co. Ltd.,
CSK299 type), a high-power electric actuator for controlling the
mandrel feed in the Z-direction (Oriental Motor Co. Ltd., PWA100
type), and an electric actuator for controlling the roller feed in the
X-direction (Oriental Motor Co. Ltd., DRL60PA4 type). The motion
controller is then connected to a personal computer. The specica-
tions of the stepping motor and electric actuators are summarized
in Table 1. The rotation of the stepping motor is transmitted to the
mandrel axis via a synchronous belt with a transmission ratio of
2:3 so that the mandrel rotation per pulse is 0.6

. The synchronous
motion of the stepping motor and the actuators is controlled by a
software code, and this enables exible control of the roller path on
the workpiece. Loadcells are attached to the ends of both the elec-
tric actuators for measuring the force components. Fig. 2 shows the
photographof the synchronous spinning machine developedinthis
study.
3. Experimental conditions
3.1. Synchronous spinning
Synchronous one-path spinning, which involves the following
steps, is carriedout after the desiredshape of the product is chosen:
Fig. 1. Synchronous spinning machine in which mandrel rotation, mandrel feed,
and roller feed are synchronized.
(1) Identication of the shapes of the mandrel and metal sheet
(workpiece) that are suitable for obtaining the desired product.
(2) Determinationof theroller pathandthecorrespondingworking
conditions; here, the roller path is controlled by combination
of the mandrel rotation rate, mandrel feed rate, and roller feed
rate.
(3) Appropriate adjustment of the mandrel rotation rate and the
mandrel and roller feed rates for minimizing errors caused by
the stepping motion.
The abovementioned steps are discussed in detail in the follow-
ing sections.
3.2. Shapes of the mandrel and workpiece
Inaconventional spinningprocess, thenal shapeof theproduct
is almost identical to that of the mandrel used. In this study, a man-
drel havingtheshapeof atruncatedelliptical coneis usedintherst
trial experiment on asymmetric spinning so that the experimental
results canbeeasilycomparedwiththoseobtainedinaxisymmetric
cone spinning. The shape and dimensions of the asymmetric man-
drel (truncated elliptical cone), axisymmetric mandrel (truncated
cone), androller are showninFig. 3. The half-cone angle of the man-
drel, , is set to 60

and one-path spinning was successfully carried


out. The roller is equipped with a ball bearing at its center to reduce
friction generated during its rotation in the spinning process.
Cold-rolled commercially available pure aluminum sheets
(composition: Si 0.11wt%, Fe 0.51wt%, Cu 0.12wt%, and Ti
0.03wt%) with a thickness of 1mm were used in this study. The
shapes of the metal sheets (workpieces) used for asymmetric and
axisymmetric spinning are shown in Fig. 4. The long axis of the
elliptical workpiece was aligned parallel to the rolling direction.
Prior to their use in the spinning experiments, the workpieces were
annealed at 673K for 1h for elimination of residual stress. The
mechanical properties of the materials obtained by carrying out
uniaxial tensile tests on three specimens, whose axis are aligned at
0

, 45

, and 90

to the rolling direction are summarized in Table 2.


Fig. 2. Photograph of synchronous spinning machine.
I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592 587
Fig. 3. Shapes and dimensions of the mandrel and roller. (a) Asymmetric elliptical-cone-shaped mandrel; (b) axisymmetric cone-shaped mandrel; and (c) roller.
Fig. 4. Aluminum workpieces with circular marks on their surfaces (unit: mm). Initial thickness of the workpieces is 1mm. The x-axis is aligned parallel to the rolling
direction. (a) Asymmetric spinning and (b) axisymmetric spinning.
3.3. Roller path and working conditions
Fig. 5 shows the roller position in the one-path spinning process
carried out using the elliptical mandrel. The xyz coordinate sys-
tem is dened with respect to the asymmetric mandrel, while the
XYZ coordinate system is dened with respect to the spinning
machine. Here, the z-axis coincides with the Z-axis. The mandrel
rotates about the Z (z)-axis; the roller is assumed to move along
the X-axis, while maintaining contact with the mandrel at point C
(x
c
, y
c
) (the clearance between the mandrel and the roller will be
Table 2
Mechanical properties of pure industrial aluminum plate.
Angle between tensile and rolling directions, 0.2% proof stress (MPa) Tensile strength (MPa) Elongation (%) Strain-hardening coefcient r-Value
0

25.0 85.2 35.0 0.25 1.05


45

24.0 91.6 44.0 0.30 0.39


90

22.0 82.6 41.0 0.24 1.31


588 I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592
Fig. 5. Schematic showing the roller in contact with the mandrel.
discussed later). The roller path is determined by the relative posi-
tion of the roller center (point R (x
R
, y
R
)) in the machine coordinate
system. The coordinates of point R in the xy plane are given by
_
x
R
= x
c
+ r cos
y
R
= y
c
+ r sin
, (1)
where is the angle between line RC and the x-axis; r, the length
of line RC obtained fromthe angle of rotation of the mandrel, , and
the radius of the roller, r. Geometrically, r is expressed as
r = r
_
sin
2
_

_
+cos
2
cos
2
_

_
(2)
The relation between x
R
and y
R
is expressed as
y
R
x
R
= tan . (3)
Further, is a function of time t and can be written as
=
2
T
t, (4)
Fig. 6. Accumulated error in the roller feed owing to the stepping motion for
t =1/60s, 1/90s, and 1/120s during each step of the process.
Fig. 7. Representative elliptical-cone-shaped workpiece formed by asymmetric
spinning.
where T is the period of rotationof the mandrel. Substitutionof Eqs.
(1) and (4) into Eq. (3) yields
tan
_
2
T
t
_
=
y
c
+ r sin
x
c
+ r cos
. (5)
Meanwhile, the equation for line RC is given as
y =
a
2
y
c
b
2
x
c
x +
_
1
a
2
b
2
_
y
c
, (6)
where a and b are one-half the lengths of the long and the short
axes of the ellipse, respectively. The tangent of angle is the slope
of line RC and can be written as
tan =
a
2
y
c
b
2
x
c
. (7)
Since point C (x
c
, y
c
) lies on the ellipse,
x
c
2
a
2
+
y
c
2
b
2
= 1 y
c
=
b
a
_
a
2
x
c
2
. (8)
BysubstitutingEqs. (7) and(8) intoEq. (5), the followingrelation
between t and x
c
is obtained:
Fig. 8. Normalized outer diameter of the elliptical-cone-shaped and circular-cone-
shaped workpieces.
I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592 589
Fig. 9. Strain distributions in the (a) circumferential, (b) radial, and (c) thickness directions after spin forming of the asymmetric elliptical cone and axisymmetric circular
cone.
t =
T
2
tan
1

b
_
a
2
x
c
2
_
ra
2
+b
_
b
2
x
c
2
+a
2
_
a
2
x
c
2
_
_
abx
c
_
rb +
_
b
2
x
c
2
+a
2
_
a
2
x
c
2
_
_

,
(9)
where r is calculated by substituting Eqs. (4), (7) and (8) into Eq.
(2). The values of x
c
and y
c
for each value of t are calculated using
Eqs. (2), (8) and (9) in a numerical analysis. Then, the coordinates
of point R (x
R
, y
R
) are estimated from Eq. (1). The coordinate of
the roller center, which determines the roller feed, is calculated by
using the following expression (Fig. 5):
X

R
=
_
x
R
2
+y
R
2
. (10)
During the spinning process, the coordinate X

R
gradually
changes with the mandrel feed rate. The clearance C between the
roller and the mandrel is added to X

R
, and thus, the coordinate of
the roller center, X
R
, at time t becomes
X
R
= X

R
+v
fR
t cos
_

2

_
+
C
cos
, (11)
where v
fR
is the relative rate of motion of the roller over the slope
of the mandrel. The roller path against a product with an arbitrary
asymmetric shape can be traced by following a procedure similar
to that explained above.
Inone-pathspinning, the average speeds of the roller moving on
the slope of the mandrel are set at 0.6mm/s and 2.4mm/s for asym-
metric spinning and axisymmetric spinning, respectively. Because
of the relatively low roller speed during asymmetric spinning,
observation of the roller movement becomes easy. The revolution
speeds of the mandrel in asymmetric spinning and axisymmetric
spinning are set at 2rad/s and 8rad/s, respectively. Therefore,
590 I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592
theroller feedrate, whichis oneof theimportant parameters affect-
ing the formability and forming quality, becomes 0.095mm/rad;
this value is thought to be appropriate (Zhan et al., 2007). C is set at
1.4mm, a value larger than the initial thickness of the workpiece,
in order to allow deformation in the thickness direction.
3.4. Adjustment of the mandrel rotation rate and the mandrel
and roller feed rates
In order to realize synchronous motion control of the mandrel
and roller, the number of pulse signals for the mandrel rotation,
mandrel feed, and roller feed are maintained constant for a given
time interval t under the abovementioned working conditions.
For simplicity, t is maintained constant throughout the process.
In this case, it is important to chose the appropriate value of t so
that the error in the roller feed, E
X
, is reduced. E
X
results from
the resolution of the actuator at a given roller feed rate and can be
expressed as
E
X
= X
S
R
X
R
, (12)
where X
S
R
is the actual coordinate of the roller center. Fig. 6 shows
the error accumulated in the roller feed,

E
X
, when t is 1/60s,
1/90s, and 1/120s. The accumulated error is within 5m when
t =1/60s, while it decreases further for t =1/90s and 1/120s.
Therefore, t is set to 1/60s in the present study.
3.5. Strain measurements
Circular marks each with a diameter of 2mm were printed on
the workpiece surface, as shown in Fig. 4. The circumferential and
radial strains,

and
r
, were evaluated from the change in the
dimensions of the circular marks after spinning was carried out
in the circumferential and radial directions, respectively. Measure-
ments were performed using a microscope equipped with a digital
camera; the maximum resolution of the microscope (4m) was
equivalent to a strain of 0.002. The thickness strain
t
was evalu-
ated by measuring the thickness of the workpiece before and after
spinning; this measurement was performedusingadigital microm-
eter with a maximum resolution of 1m. The logarithmic strain
measure was used throughout the study.
4. Results and discussions
4.1. Shape of the workpiece after spinning
A truncated-elliptical-cone-shaped product was successfully
formedby the synchronous spinning withanasymmetric elliptical-
cone-shaped mandrel whose half-cone angle was 60

(Fig. 3(a)).
Owing to springback in conventional spinning process, there was
a slight increase in the slope angle, particularly near the outer
edge of the workpiece as shown in Fig. 7. The nal shape of the
workpiece is evaluated by dividing its outer diameter by the ini-
tial diameter in each direction denoted by the angle . Fig. 8 shows
the results obtained for elliptical cone spinning (asymmetric); the
results obtained for circular cone spinning (axisymmetric) are also
shownfor comparison. The plottedmarks inFig. 8indicate the aver-
ages of the values obtained in three trials. In each of the three trials,
the maximum deviation of the observed result is less than 0.001,
and this conrms the excellent reproducibility of the proposed
method. The normalized outer diameter is always less than 1, since
the working conditions employed in this study are conventional
spinning. During circular spinning, a slight variation is observed in
the normalized outer diameter of the workpiece because of plas-
tic anisotropy. The diameter increases when approaches 0

, 90

,
and 180

and corresponds to a large r-value, as shown in Table 2.


Fig. 10. Schematic representation of the relative movement of the roller against the
workpiece. Soliddots represent the positions where roller was fedinthe z-direction.
The variation in the normalized outer diameter is markedly greater
in the case of elliptical spinning than in the case of circular spin-
ning. Owing to variations in the strain distribution, the diameter
of the workpiece increases near the long axis of the ellipse ( =0

and 180

), as will be shown in Section 4.2. This phenomenon is


considered typical to asymmetric spinning, particularly when the
clearance C is made greater than the initial thickness of the work-
piece.
4.2. Strain distributions
The local strain distributions in the workpiece after elliptical
cone spinning and circular cone spinning were found to be simi-
lar, as shown in Fig. 9. The circumferential strain slightly increased
near the shoulder of the workpiece andgradually decreasedtoward
the outer edge. The radial strain showed a transient increase near
the shoulder and increased further, reaching a maximum near the
center of the slope of the workpiece surface. In contrast, the thick-
ness strain decreased toward the center of the slope and increased
toward the outer edge of the workpiece. The radial tensile strain
and decrease in the wall thickness near the center of the slope
were caused when the roller displaces the ange in the axial (z-)
direction.
A typically observed phenomenon during elliptical cone spin-
ning is the change in the local strain with . In the long-axis
direction of the ellipse ( =0

), the radial strain increases, while the


thickness strain decreases. As mentioned above, the radial tensile
strain is caused by the roller feed in the z-direction. In the present
spinningprocess, theroller is fedinthez-directionfor every6

rota-
tionof the workpiece, as showninFig. 10. Inthis case, it is supposed
that the tensile stress generated by the roller feed is larger when
the roller is near the long axis of the elliptical cone than when the
roller is near the short axis. Thus, the resultant radial tensile strain
inthe long-axis directionbecomes larger thanthat inthe short-axis
direction. This implies that the local strain distribution can be con-
trolled by changing the timing of the roller feed in the z-direction.
The effect of the roller path control on the nal strain distributions
will be discussed in the future study.
I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592 591
Fig. 11. Variation in working forces during elliptical cone spinning and circular cone spinning.
4.3. Working forces
Fig. 11 shows the variations in the force components during
elliptical cone spinning and circular cone spinning, respectively.
The radial force is the force component in the radial direction of
the mandrel measured by a loadcell attached to the actuator end
for roller feed (Fig. 1). Meanwhile, the axial force is that in the
direction of rotation axis of the mandrel, measured by another
loadcell attached to the actuator end for mandrel feed. In both
elliptical cone spinning and circular cone spinning, the compres-
sive axial force increased toward the center of the slope and then
decreased toward the edge of the workpiece. This variation in the
axial force qualitatively corresponded to the radial strain distribu-
tion in the workpiece shown in Fig. 9(b). Since only the local area of
the workpiece that was incontact withthe roller underwent plastic
deformation during spinning, the radial strain increased with the
axial force. The variation in the radial force was indicative of the
difference between axisymmetric and asymmetric spinning. The
radial force was compressive during circular cone spinning, while it
was alternately compressive andtensile during elliptical cone spin-
ning. Fig. 12shows thechangeinthenatureof theforcecomponents
between two mandrel revolutions during elliptical cone spinning.
The radial force changed periodically during every one-half revo-
lution of the mandrel. Synchronization of the mandrel rotation and
periodic changes in the radial force revealed that a compressive
radial force was generated when the roller moved from the long
axis to the short axis, while a radial tensile force was generated
Fig. 12. Variation in working forces between two mandrel revolutions during ellip-
tical cone spinning.
when the roller moved from the short axis to the long axis. Fig. 13
is the schematic showing the change in force components with the
roller movement. In the left gure, the roller moves in the negative
X-direction, so that the roller presses the workpiece to the man-
drel and the radial force is compressive. Meanwhile, in the right
gure, the roller moves in the positive X-direction and the roller
unbends the ange part of the workpiece, thus, the radial force
becomes tensile. In the latter case, it is supposed that the radial
Fig. 13. Change in force components with the roller movement.
592 I. Shimizu / Journal of Materials Processing Technology 210 (2010) 585592
Fig. 14. Photographs of product with different asymmetric shapes formed by the
synchronous spinning method. (a): Elliptical-cone-shaped product with side wall
and (b) product with a truncated pyramidal shape.
tensile force is not working to deform the workpiece into elliptical
cone shape. The deformationof the workpiece was inducedprimar-
ily by the axial force components under the conditions employed
in this experiment.
4.4. Trial experiments for obtaining products with other
asymmetric shapes
Fig. 14 shows the trial synchronous spinning experiments car-
ried out to obtain a truncated-elliptical-cone-shaped product with
a sidewall (a) and a product with a truncated pyramidal shape (b).
Products with these shapes were successfully formed by the syn-
chronous spinning method, in which the roller path is accurately
controlled in the manner similar to that described in Section 3.3. It
should be noted that the application of the synchronous spinning
method to those products are achieved only by changing the soft-
ware code and no part of the machine is replaced. This is a great
advantage of the synchronous spinning comparing with the meth-
ods of Amano and Tamura (1984) and Gao et al. (1999). Although
detailed investigation of these results will be carried out in a future
study, the present results indicate that the synchronous spinning
method can be applied to products with a wide variety of asym-
metric shapes.
5. Conclusions
A synchronous spinning method in which the mandrel motion,
mandrel feed, and roller feed were synchronized by pulse con-
trol was used for asymmetric spin forming. The one-path elliptical
spinning was performed on an aluminum sheet, as the rst trial
experiment. The roller path was estimated and controlled in accor-
dance with the shape of the elliptical-cone-shaped mandrel. The
nal shape of the workpiece, local strain distributions, and force
components were evaluated and compared with those obtained
in axisymmetric circular cone spinning; the characteristics of
asymmetric spinning were conrmed fromthe results of this com-
parison. The results obtained could be summarized as follows:
(1) An asymmetric truncated-elliptical-cone-shaped product was
successfully formed by the proposed synchronous spinning
method.
(2) The normalized outer diameter of the workpiece was found
to increase when the workpiece was near the long axis of the
ellipse, as the radial strain in the long-axis direction was larger
than that in the short-axis direction.
(3) Periodical changes in the radial force were observed during
elliptical cone spinning; the radial force was compressive when
the roller moved from the long axis to the short axis, while it
was tensile when the roller moved from the short axis to the
long axis.
(4) The successful application of synchronous spinning method
for the formation of an asymmetric truncated-elliptical-cone-
shaped product with a sidewall and a truncated-pyramid-
shaped product conrmed the exibility of this method.
Acknowledgements
The author thanks Prof. Naoya Tada of Okayama University for
valuable discussions and Mr. Shuichi Tanaka, Graduate Student,
Okayama University, for his assistance in the spinning tests.
References
Amano, T., Tamura, K., 1984. The study of an elliptical cone spinning by the trial
equipment. In: Kobayashi, M. (Ed.), Proceedings of the Third International Con-
ference on Rotary Metalworking Processes. Kyoto, Japan, pp. 213224.
Arai, H., 2005. Robotic metal spinning-Forming asymmetric products using force
control. In: Proceedings of 2005 IEEE International Conference on Robotics and
Automation, Barcelona, Spain, pp. 27022707.
Awiszus, B., Meyer, F., 2005. Metal spinning of non-circular hollowparts. In: Bariani,
P.F. (Ed.), Proceedings of the Eighth International Conference on Technology of
Plasticity. Verona, Italy, pp. 353354.
Gao, X.-C., Kang, D.-C., Meng, X.-F., Wu, H.-J., 1999. Experimental research on a new
technology-ellipse spinning. J. Mater. Process. Technol. 94, 197200.
Wong, C.C., Dean, T.A., Lin, J., 2003. A review of spinning, shear forming and ow
forming processes. Int. J. Mach. Tools Manuf. 43, 14191435.
Zhan, M., Yang, H., Zhang, J.H., Xu, Y.L., Ma, F., 2007. 3D FEM analysis of inuence of
roller feed rate on forming force and quality of cone spinning. J. Mater. Process.
Technol. 187188, 486491.

You might also like