Journalof

Materials
Processing
Technology
ELSEVIER Journal of Materials Processing Technology 58 (1996) 174 183

Process design in multi-stage cold forging by the

finite-element method
Jin-Hee Lee a Beom-Soo Kang a,*, Jung-Hwan Lee b
~' ERC,/or Net Shape and Die Manujacturmg, Pusan National University, Pusan 609-735, South Korea
b Korea Institute of Machineo' and Metals', Changwon, Kyungnam 641-010, South Korea

Received 6 December 1994

Industrial summary

In this work, a process sequence for multi-stage cold forging is designed with the rigid-plastic finite element method to form a
constant-velocity joint (CVJ) housing with shaft. The material flow during the forming of a CVJ housing is axisymmetric until the
final forging process for the forming of the ball grooves. The numerical approach in this study treats the deformation as an
axisymmetric case. The main objective of the process-sequence design is to design preforms which satisfy the design criterion of
a near-net-shape product requiring less machining after forming without any defects. It also included investigation of velocity
distributions, effective strain distributions, and forging loads, which are useful information in process design. A systematic
approach to process-sequence design is established using the finite-element method for the cold forging of the CVJ housing.

Keywor&': Process design; Cold forging; Finite-element method

1. Introduction provide detailed information concerning the mechanics

Near-net-shape or net-shape manufacturing is be-
coming a trend in metal forming, especially in cold
forging, resulting in savings in material, energy and
machining requirements. Cold forging maintains
strength, dimensional and surface-finish specifications.
In practice, however, cold forging requires several pre-
forming operations to transform an initial simple billet
into a final complex product without defects. The
design of a forging-process sequence involves the deter-
mination of the number of preforms, and the determi-
nation of the shapes and preform dimensions. The
proper design of preforming operations to achieve an
adequate material distribution is one of the most im-
portant aspect in cold-forging processes. Traditionally,
forging-sequence design is carried out using mainly
empirical guidelines, experience and trial-and-error,
which results in a long time for process development,
and high cost of the products [1 3].
Computer-aided simulation techniques in metal
forming before actual tryout may reduce the cost and
time of process design. Many computer-aided ap-
proaches based on approximate analysis and empiri-
cally-established design rules have been published in the
literature [4-9]. These techniques do not always
* Corresponding author.
of the process. On the contrary the finite-element
method has been proven to provide more accurate and
detailed information, and is thus widely used for simu-
lating and analyzing various metal-forming processes.
Amongst various applications of the finite-element
method in the metal-forming industry is practically one
of the most important process-sequence design in multi-
stage forming processes [10-19].
The production quantity of constant-velocity joints,
which are indispensable for front-wheel-drive cars, has
been increasing. The mass manufacturing of constant-
velocity joints at an economical cost has thus become
an important issue for production engineering. Con-
stant-velocity joints are employed for torque transfer
between the gear and the drive wheels, in which the
steering and the suspension movements of the wheels
must be followed simultaneously. A constant-velocity
joint consists of a spider, an inner race and a housing
with a shaft. Amongst the components, the production
of the housing by machining is difficult because of its
irregular shape. There are two methods for the mass
production of housings, i.e., hot forging and cold forg-
ing. The conventional hot-forging process requires sub-
sequent machining operations involving an enormous
number of man-hours. The cold-forging process, how-
ever, makes it possible to produce net-shape housings
without any machining after forming, and effects save

Elsevier Science S.A.

SSD1 0924-0136(95)02093-2
 J.-H. Lee et al./Journal of Materials Processing Technology 58 (1996) 174 183
Frezvq
~ -- ~ upset
~extrusi°n
tingbackwardV
Annea in
initial forward ~ ~ deep backward
billet extrusion extrusion
c l o s e d die forging
Fig. 1. Suggested processes of cold forging to form a constant-veloc-
ity-joint housing [19.23].
material of up to 40%. However, great efforts in design-
ing a delicate process sequence are needed in cold
forging of the net-shape housings [20-22].
In this study, the process sequence in the cold forging
of a constant-velocity joint (CVJ) housing with a shaft
has been investigated and designed using the rigid-
plastic finite-element method. The cold forging process-
sequence to form the CVJ housing consists of four
operations: forward extrusion, closed-die forging, deep
backward-extrusion, and ironing. Here the practical
problems of the deformed shape of the cup opening, the
slope angle, the press capacity, and surface cracking are
considered, and a forging process-sequence is designed
which can produce a net-shape housing without defects
within a given press capacity. The information obtained
from the finite-element simulation, including the shapes
of preforms, effective strains, velocity fields, and die
load, is utilized in designing the new process sequence.
2. Process description and modeling
Fig. 1 is a schematic drawing showing the suggested
processes of cold forging for a CVJ housing with axial
symmetry [19,23]. Process sequence I in Fig. 1 is a
classical expert's solution [23]. The cylindrical billet is
extruded forwards to form a shift in the first operation,
and upset in the second operation. The head is com-
pressed by closed-die forging to form a flange with a
radius approaching the outer diameter of the cup. After
Fig. 2. Results of simulations of the conventional five-operation process [19].
175
annealing, so as to give the preform the required degree
of cold formability, the third operation of backward
extrusion is performed. During the operation, the flow
stress of the material becomes large and a great deal of
,~ compressive stress is applied to the tools. Thus, it is
essential to perform another annealing heat-treatment
on the preform (resulting from the third operation).
The fourth operation is deep backward-extrusion. Pre-
ironing
cise design of the punch and die shape is required in
order to obtain the required dimensional accuracy of
the final product. The part formed by the fourth opera-
tion has a shape close to the finished product, but it is
required to form the cup opening with the prescribed
dimensional accuracy. Usually an ironing operation
with axial bending is carried out as the last operation.
The actual punch in the ironing operation consists of
six radially-segmented components, like the slices of an
apple. Removal of the die from the final socket-shaped
product is carried out by retracting the central core and
collapsing the segments of the die inwards.
Fig. 2 shows the simulation results by Kang and
Kobayashi [19] to form a CVJ housing according to the
expert's solution of case I in Fig. 1. The possibility of
cold forging for this CVJ housing using a new process-
sequence has been suggested. By combining the upset-
ting with the backward-extrusion of case I, a new
process sequence is obtained with four forming opera-
tions and one annealing treatment, shown in Fig. 1 as
case II.
In this study, a process sequence of multi-stage cold
forging of a CVJ housing is designed by the rigid-plas-
tic finite-element method according to case II.
The dimensions and configurations of the initial billet
and the final cold-forged CVJ housing product are
illustrated in Fig. 3.
The material flow during forging of the CVJ housing
is axisymmetric before the deep backward-extrusion:
the flow then becomes three-dimensional. As mentioned
above, the socket-shaped product is formed by special
dies with six segments around a core. Assuming that
the difference between the actual three-dimensional pro-
cess and the axisymmetric case is relatively minor, it is
possible to consider the deformation of CVJ housing as
1
176 J.-H. Lee et al./ Journal of Materials Processing Technology 58 (1996) 174 183

assign new effective strains at the nodal points by linear

interpolation. The press capacity is limited to 1000
tonne in respect of the actual press capacity in industry.
(Unit : m m )
3. Simulation results and discussion

A process sequence for the multi-stage cold forging

I of a CVJ housing is simulated according to the process
i
of case II in Fig. 1. The main objective of the process-
[ \ "- i oi sequence design in this study is to obtain intermediate
preforms which produce a near-net-shape product. Also
i
q design conditions, such as the limit of the press capacity
and the avoidance of surface cracking, should be sa-
I
tisfied.
I The first operation shown in Fig. 4 is forward extru-
i

I sion to form the shaft of the CVJ housing. The reduc-

tion of area in the extrusion is 64% and the semi-cone

_ ~so.0 & angle of die is 35.5 ° . This extrusion is in preparation for

(a) (b)
Fig. 3. Dimensions and configurations of the CVJ housing: (a) initial
billet; (b) final product.
an axisymmetric problem with a socket-shaped product.
The maximum die load in the final operation is similar
to the value of the actual case, but the cup-opening
configuration of the product is influenced by the mod-
ification. When applying this result to die design in
industry, slight design-modification will be performed
to compensate for the difference between the actual
process and the simulated process.
Although three-dimensional finite-element-based
codes enable the modeling of large plastic deformation
[24], three-dimensional modeling is still much too com-
plicated and time consuming for industrial application.
Considering an axisymmetric cup configuration, the
simulations are performed with the two-dimensional
rigid-plastic finite-element method that can give sound
results for the material flow in the CVJ housing forging.
The fundamental of the rigid-plastic finite-element
method and the mathematical formulation are well
established and given in the literature [25,26]. The
material used for the simulations is a mild steel, AISI
1018, with the following flow-stress behavior: ~ =
0.037(1.0 + 50.0g) T M tonne/mm 2 (1 tonne mm 2). The
die-workpiece interface is characterized by the con-
stant-factor friction law, usually used for bulk metal-
forming problems: r = m k . Here, r is the frictional
shear stress, m is the friction factor, and k is the shear
flow stress.
The friction factor in this study is assumed to be 0.1
which is suitable for cold-forging processes. During the
simulations the mesh is distorted severely, and thus
should be remeshed to avoid negative Jacobians and to
the next operations. Figs. 4(a)-(c) show the initial
mesh, the deformed grid distortion at punch stroke of
60%, and the final grid distortion with the effective
strain distribution. The extrusion operation was
stopped when the extruding part was of the same final
dimensions as the shaft part of the CVJ housing.
The maximum extrusion load is 270 tonne, which is
within the press capacity of 1000 tonne. High effective
strains appear in the outside part of the extruded shaft,
whilst the workpiece in the container virtually does not
deform.
The simulation of the second operation is shown in
Fig. 5, which is called a closed-die forging, and com-
bines upsetting with the backward extrusion of the
conventional process sequence (see case I in Fig. 1).
Due to negative Jacobians resulting from severe mesh
distortion near to the part contacting with the punch,
two remeshings are required to complete the simulation
up to the die-filling stage. The number of nodes and
elements decrease slightly for each remeshing because a
coarser mesh is used in the shaft part. Fig. 5(a) shows
the preform after forward extrusion and the grid distor-
tion at a punch stroke of 88% whilst Fig. 5(b) shows
the grid distortion and the effective-strain distribution
at the end of the second operation. The maximum
forging load is 850 tonne: it is possible to reduce the
forging load by controlling the punch stroke in the
closed-die forging. Thus, the punch stroke at the finish-
ing step lies within the available press capacity. Since
this second operation is not a finishing operation and
does not influence significantly the shape of the final
product, complete filling is not required: A small cavity
appears at the corner between the flat part and spheri-
cal part of the punch. The preform obtained from the
second operation is assumed to be satisfactory as
preparation for the next operation. The preform is
called Preform PCDF, having the dimensions shown in
 J.-H. Lee et al./ Journal of Materutls Processing Technology 58 (1996) 174 183
_d- -t_
Die s e m i - a n g l e : 35.50
R e d u c t i o n of a r e a : 64%
(a)
Fig. 4. Simulation of forward extrusion: (a) initial billet; (b) stroke of 60%; and (c) final deformation and effective-strain distribution.
Fig. 5(c). The high value of effective strain near to
the punch tip reflects severe plastic deformation, as
shown in Fig. 5(b). In the next operation of deep
backward-extrusion, mainly the upper part of Y-shaped
Preform P C D F deforms. Complete filling in this opera-
tion can cause the final forging load exceeding the press
capacity so that the upsetting operation is stopped
before complete filling. It is necessary that the preform
be annealed to obtain sufficient formability prior to
deep backward-extrusion. Since the annealing treat-
ment of the workpiece means a strain-free state in
simulation using the finite-element method, Preform
P C D F is considered to be in the state of zero-strain at
the start of the next simulation. Remeshing is carried
out again for simulation of the next operation (see
Fig. 5(c)).
Fig. 6 shows the die configuration designed for the
following deep backward-extrusion process, hereinafter
the process being referred to as backward extrusion.
In this operation, the dies are designed carefully,
since the backward-extrusion process affects, directly,
the precision forming of the final product. In order to
obtain the net-shape final product, the preform from
the backward extrusion should be formed soundly and
precisely before being used in the last operation, an
ironing process.
Fig. 7 shows the dies for ironing which is conducted
by covering the male die of ball groove shape with the
workpiece and having the female die extrude the outer
rim of the workpiece upwards with bending.
When investigating the effects of the preform shape
derived from the backward extrusion on the forming of
177
(b) (e)
the final product, it would be helpful to be able to
obtain some knowledge of the metal flow involved in
the backward extrusion. Simulations are conducted us-
ing the trial dies with punch I and die I, as shown in
Fig. 6(a). The simulation results of the backward extru-
sion and the last operation of ironing process appear in
Figs. 8(a) and (b), respectively.
The left side of Fig. 8(c) displays the velocity field of
the preform (Preform PDBE-I) after the third opera-
tion, and the right side of the figure describes the final
product, with dimensions, after the ironing operation.
As shown in Fig. 8(b), the top surface of the cup
opening in the final product is not in close contact with
the punch after ironing, which indicates that sound and
precise forming at the top surface is not achieved. The
undesirable forming at the top surface can be assumed
to occur for the reason that an excessive amount of
material deforms radially outside at the top part in the
backward extrusion, as shown in the right side of Fig.
8(a). During the ironing operation, the thickness of the
cup bottom becomes less than that of Preform PDBE-I.
The dimensions in Fig. 8(c) differ from those of the real
product of the CVJ housing (see Fig. 3(b)). The slope
angle of the cup opening in the simulation is 33 ° , whilst
the angle in the real product is 25 ° .
The most stringent forging operation amongst all the
operations is the third operation. Surface cracks may
appear in the forming material near to the cup bottom
along the direction of the circumference. Fig. 9 shows
the relationship between the crack generation during
forging and the thickness of the cup bottom in the case
of CVJ forming with carbon steel [22].
178 J.-H. Lee et al./Journal of Materials Processing Technology 58 (I996) 174 183

(Unit : ram)

Preform PCDF

~95.0

~40.0

2 2!'5

1.51 2.0

2 (
~,o

(~) (b) (e)

Fig. 5. Simulation of closed-die forging: (a) initial set-up and stroke of 88%; (b) final deformation and effective-strain distribution; and (c)
dimensions of Preform PCDF.

(Unit : mm)

(a) (b) (c)

PUNCII I PUNCll II PUNCH III

"-.j/N/A ~,
"1 '5.

DIE I DIE II
Fig. 6. Dimensions and configurations of the dies for deep backward-extrusion.
 J.-H. Lee et al./Journal of Materials" Processing Technology 58 (1996) 174-183 179
(Unit : m m )
] 75.0

70.0

PUNCII 65.0

~- 2~ DIE +~ 60.0

"Q) 55,0
O DANGER ZONE / aJ C
>~ 50.0 /
DIE SAFE ZONE
//
/
/

11:3R]~ 1 0.0
0.0 l/ lo'.o 1 .o
/

. ~ ¢a6.0 F
Before ironing During i r o n i n g
Fig. 7. Dimension and configuration of the dies for the ironing
process.
The value calculated by the dimensions in Fig. 8(c) is
plotted as point A in Fig. 9, indicated to be inside the
danger zone where surface cracks may occur. This is
caused by a large amount of deformation being concen-
trated in the cup-bottom area of the forging. Based on
the simulation results, the preform derived from the
backward extrusion influences the final product signifi-
cantly. Especially, the control of the material flow near
to the cup opening affects the configuration of the final
product critically. In order to obtain the same dimen-
sions of the cup bottom and the same slope angle, and
to fill the undeformed part near to the cup opening, the
14'.o 16.0
Bottom thickness(ram)
Fig. 9. Forming limit for surface cracks in the inner surface near to
the cup bottom [22].
height and corner radius of the punch are reduced
compared to those of punch I. The newly designed
punch is shown in Fig. 6(b) and is called punch II,
based on the previous discussion.
Fig. 10 shows the simulation results using the die set
of punch II and die I, the third operation being shown
in Fig. 10(a), and the last ironing operation in Fig.
10(b). Also, Fig. 10(c) shows the velocity field of the
preform (Preform PDBE-II) formed in the third opera-
tion and the dimensions of the final product formed in
the last operation.
For the case using Preform PDBE-II, the amount of
material flow into the top part is reduced in comparison
with the case of Preform PDBE-I. The deformation of

undesirable
formation
(Unit : ram)

• i

(a) (b) (c)

Fig. 8. Simulations of deep backward-extrusion using the dies in Fig. 6(a) and the ironing process.
180 J.-H. Lee et al./Journal of Materials Processing Technology 58 (1996) 174-183

(Unit : ram)

Preform
_ 3 o ' ~

(a) (b) (c)

Fig. 10. Simulations of deep backward-extrusion using the dies in Fig. 6(b) and the ironing process.

the inward cup opening is almost complete and the
configuration of the final product in Fig. 10(c) is far
more desirable than the final configuration using Pre-
form PDBE-I: however, it still does not satisfy the
dimensions of the cup bottom thickness and slope angle
of the CVJ housing. A surface crack does not appear in
the forming material, the latter being represented by
point B in Fig. 9.
An attempt has been made to redesign the dies to
produce net-shape products without machining after
forming. Fig. 6(c) shows the finally-designed die set
consisting of punch III and die II for deep backward-
extrusion, the design of the die set reflecting the simula-
tion results in Fig. 10.
The left side of Fig. ll(a) shows the simulation
results of the third operation using the die set in Fig.
6(c), where the backward extrusion is carried out up to
the stage when the inside punch is filled completely.
Three stages of the last operation of ironing are shown
in the right side of Fig. l l(a) (c).
As shown in the right side of Fig. 1 l(c), there is no
undeforming part, but the final dimensions are not the
same as those of the product in Fig. 3(b). The slope
angle of the cup opening is 29 ° , which is larger than the
desired angle, but no surface cracks occur near to the
cup bottom: point C in Fig. 9. The maximum die load
in the backward extrusion is 1500 tonne at the final
stage with completion, which will be shown later in Fig.
15(c). Because the available press capacity is 1000
tonne, the process should be modified to reduce the
maximum die load. The present preform does not de-
form satisfactorily with respect to all of the final crite-
ria. The dimensions of the final product, however, are
similar to those of the real product. Thus, a slight

(Unit : m m )

(a) (b) (c)

Fig. 11. Simulations of deep backward-extrusion using the dies in Fig. 6(c) and the ironing process where the punch stroke is 100%.
 J.-H. Lee et al./Journal of Materials Processing Technology 58 (1996) 174-183 181

• unlilled
eavity

ro,orm=:

(a) (b) (e)

Fig. 12. As for Fig. 11, but for a punch stroke of 95%.

modification of the preform is made to keep the maxi-
mum die load during backward extrusion below 1000
tonne and to obtain dimensions identical to those of the
real product. In comparison with the real product in
Fig. 3(c), it is necessary to increase the cup bottom
thickness by 0.9 mm and to decrease the amount of
material which contributes to the formation of the
slope angle. For these purposes, the simulation of back-
ward extrusion is stopped at the stage when the punch
stroke is less than 100%, and the derived preform is
used for the ironing process. The dies in Fig. 6(c) are
used without modification during deep backward-extru-
sion. It is possible to maintain the forging load within
the press capacity because the load increases rapidly
near the end stroke.
Fig. 12 shows the simulation of backward extrusion
to a punch stroke of 95%, the darkened area in Fig.
12(c) indicating an unfilled cavity of very small volume.
The ironing operation is carried out using the preform
(Preform PDBE-III) in Fig. 12(c), as shown in Fig. 13.
Fig. 14(a) shows the velocity field and the effective-
strain distribution after deep backward-extrusion, and
Fig. 14(b) the dimensions and the effective-strain distri-
bution of the final product.
As shown in the shape and velocity field of Preform
PDBE-III, the material flow towards the outside of the
cup opening is smaller than that in the previous cases.
The maximum effective strain occurs near to the cup
bottom, as displayed in Fig. 14(a). Therefore, it is
known that should a surface crack occur, it will be
located near to the cup bottom. Also, an undeformed
part does not appear near to the cup opening, and the
cup bottom thickness remains unchanged during the
ironing process. The outside surface of the cup is
deformed mainly by the ring-typed die in the ironing
process. The dimensions of the final product in Fig.

L,

(a) (b) (c)

Fig. 13. Simulation of the ironing process using Preform PDBE-III.
182 J.-H. Lee et al. ,/'Journal of Materials Processing Teehnology 58 (1996) 174 183

(Unit : mm) 95%. The criterion for press capacity is thus satisfied
throughout all of the forming operations.
Preform
PDBE-III 0.8
o.
0.2
4. Conclusions

A process sequence for the multi-stage cold forging

of a CVJ housing with a shaft has been designed by the
rigid-plastic finite-element method. The cold-forging
~'~ 0.~
process sequence to form the CVJ housing consists of
I
¢i the four operations of forward extrusion, closed-die
forging, deep backward-extrusion, and ironing.
This study can be concluded with the following re-
marks.
1. A better process sequence to form a CVJ housing
has been designed by the rigid-plastic finite-element
(a) (b) method, solving actual problems such as the precise
deformed shape of the cup opening, the slope angle, the
Fig. 14. Showing: (a) distributions of effective strains for Preform
press capacity, and surface cracking.
PDBE-III; and (b) the final product.
2. The information obtained from the finite-element
simulation, including the shapes of preforms, the effec-
500 1000 tive strains, the velocity fields, and the die load, is
useful in designing a new process sequence.
400 0oo

~
i
3. A systematic approach to process-sequence design
o
300 60b [ has been established using the finite-element method for
< multi-stage cold forging to form a constant-velocity-
C~ 200 ~400
joint housing. This specific case can be considered as an
I00 2oo / / / ~ example of the application of the method and for
/ development of the sequence-design methodology in
% 20' 40' 6 o' co' 100 o 20' 40' 6 o' 80 ' 100
general.
STROKE(%) STROKE(Z)
(~) (b)
1800 500 [

References
400F
120C
~ ~o~ 3o0 [l] K. Lange, Closed-Die Forging of Steel (in German), Springer,
ooc Berlin, 1958.
200 [2] T. Altan, F.W. Boulger, J.R. Becket, N. Akgerman and H.J.
4OO
Henning, Forging Equipment, Materials and Practices, MC1C-
,00 HB-03, Battelle-Columbus Laboratories, Columbus, OH, 1973.
[3] J.A. Pale, R. Shivpuri and T. Altan, Recent developments in
o
2o 40 60 8o loo 20 40 60 8o 1o0 tooling, machines and research in cold forging of complex parts,
STZ~OKE(~) STROKE(%) J. Mater. Process. Technol. 33 (1992) 1 29.
(c) (d) [4] S.K. Biswas and W.A. Knight, Preform design for closed die
Fig. 15. Load-stroke curves for the four-operation process: (a) forgings: experimental basics for computer aided design, Int. J.
extrusion; (b) closed-die forging; (c) deep backward-extrusion using Math. Tool Des. Res., 15 (1975) 179 193.
the dies in Fig. 6(c); and (d) ironing. [5] G.B. Yu and T.A. Dean, A practical computer-aided approach
to mould design for axisymmetric forging die cavities, lnt. J.
Mach. Tool. Des. Res., 25(1) (1985) 1 3.
14(b) are almost the same as those of the real product [6] K. Sevenler, P.S. Raghupati and T. Altan, Forming sequence
design for multistage cold forging, J. Mech. Work. Technol., 14
of the CVJ housing. The slope angle obtained in 25 °,
(1987) 121 135.
which can be considered as almost the same as that in [7] K. Osakada, T. Kado and G.B. Yang, Application of AI-tech-
the real product. Also, a surface crack does not occur, nique to process planning of cold forging, Ann. C1RP, 37(1)
as indicated by point D in Fig. 9. (1988) 239 -242.
The load-stroke curves during the four operations [8] P. Bariani and W. Knight, Computer-aided cold-forging process
design: a knowledge based system approach to forming sequence
for the cold forging of the CVJ housing are displayed in
generation, Ann. CIRP, 37(1) (1988) 243 246.
Fig. 15. [9] K.R. Vemuri, S.I. Oh and T. Altan, BID: a knowledge-based
Point B in Fig. 15(c) is the maximum load of 980 system to automate blocker design, Int. J. Maeh. Tools Manu-
tonne for backward extrusion when the punch stroke is /act., 29(4)(1989) 505 518.
 J.-H. Lee et al./Journal of Materials Processing Te~4mology 58 (1996) 174 I83
[10] J.J. Park, N. Rebelo and S. Kobayashi, A new approach to
preform design in metal forming with the finite element method,
Int. J. Mach. Tool Des. Res., 23(1) (1983) 71 79.
[11] S.M. Hwang and S. Kobayashi, Preform design in disk forging,
Int. J. Mach. Tool Des. Res., 23(3) (1986) 231-243.
[12] D.Y. Yang and K.H. Kim, Rigid-plastic finite element analysis
of plane strain ring rolling, Int. J. Mech. Sci., 30(8) (1988)
571-580.
[13] B.S. Kang, N.S. Kim and S. Kobayashi, Computer-aided pre-
form design in forging of an airfoil section blade, Int. J. Mach.
Tool Manuf, 30(1) (1990) 43 52.
[14] N.S. Kim and S. Kobayashi, Preform design in H-shaped cross
sectional axisymmetric forging by the finite element method, Int.
J. Mach. Tool Manufact., 30(2) (1990) 243 268.
[15] B.S. Kang and S. Kobayashi, Preform design in ring rolling
processes by the three-dimensional finite element method, Int. J.
Mach. Tool Manufact., 31(1) (1991) 139-151.
[16] B.S. Kang, Process sequence design in a heading process, J.
Mater. Proc. Technol., 27 (1991), 213-226.
[17] T. Altan and M. Knoerr, Application of the 2D finite element
method to simulation of cold-forging process, J. Mater. Process.
Technol., 35 (1992) 275-302.
[18] B.S. Kang, J.H. Lee, B.M. Kim, C.C. Choi and S.M. Hwang,
183
Preform design in flashless forging of Rib/Web-shaped plane-
strain component, Accepted for publk'ation in J. Mater. Process.
Technol.,.
[19] B.S. Kang and S. Kobayashi, Process sequence design in cold
forging to form a constant velocity joint housing, Accepted for
publication in Int. J. Mach. Tool. Manuf., 1994.
[20] T. Maki, Trends in forging technology in field of constant-veloc-
ity joint, JIDOSHA-GIJUTSU(JSAE), 36 (1982) 898.
[21] KOMATSU Ltd., KOMATSU Warm and Cold Forging System:
Constant-Velocity Joint Forging Line, FESA-E-G-003, 1982.
[22] S. Fujikawa, H. Yoshioka and S. Shimamura, Cold and warm-
forging applications in the automative industry, J. Mater. Pro-
cess. Technol., 35 (1992) 317 342.
[23] Project Team of II-shin Cold Forging Company, Cold Forging
Process to Form a C.V. Joint Housing, Summary Report, II-shin
Cold Forging Company Ltd., Changwon, Korea, 1992.
[24] J.H. Yoon and D.Y. Yang, A three-dimensional rigid-plastic
finite element analysis of bevel gear forging by using a remeshing
technique, Int. J. Mech. Sci., 32 (1990) 277-291.
[25] O.C. Zienkiewicz, The Finite Element Method, McGraw-Hill,
New York, 1977.
[26] S.Kobayashi, S.I. Oh and T. Altan, Metal Forming and the
Finite-Element Method, Oxford University Press 1989.