Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 17 (2014) 686 691

Variety Management in Manufacturing. Proceedings of the 47th CIRP Conference on Manufacturing

Systems

Conditions for Bending of Thin Metal Sheet

by Thermal Strain in Electrical Discharge Machining
Katsushi Furutani*, Norimichi Yoshida
Toyota Technological Institute, Nagoya 468-8511 Japan
* Corresponding author. Tel.: +81-52-809-1796; fax: +81-52-809-1796. E-mail address: furutani@toyota-ti.ac.jp

Abstract

This paper deals with the conditions for bending of sheet metals by electrical discharge machining (EDM). The thermal strain caused by the
tensile stress after EDM was used to bend a sheet metal. Plate and wedged electrode were used to bend the strips simply, and a pipe was used to
form a convex shape. The negative electrode polarity was preferable to reduce the removal rate. The warp of the strip in the lateral direction
disturbed the bending performance. The incremental forming by shifting the discharge position was also tried to avoid the influence of the warp.

2014
2014The Authors.
Elsevier B.V.Published byopen
This is an Elsevier B.V.article under the CC BY-NC-ND license
access
Selection and peer-review under responsibility of the International Scientific Committee of The 47th CIRP Conference on Manufacturing
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Systems
Selectionin thepeer-review
and person of theunder
Conference Chair Professor
responsibility Hoda ElMaraghy.
of the International Scientific Committee of The 47th CIRP Conference on
Manufacturing Systems in the person of the Conference Chair Professor Hoda ElMaraghy
Keywords: thermal strain; electrical discharge machining; forming; wedge; bending angle; warp

1. Introduction
Today, mechanical and electronic parts have been
miniaturized with high precision. Small mechanical parts are
often made of sheet metals which are generally formed by
mechanical methods such as stamping or bending [1]. It is
difficult to adjust a bending angle accurately because of
spring-back. In prototyping or small lot production, many
forming method without the dies have been proposed to
reduce the cost of the dies and some have been practically
used.
In shot peening, small media such as steel, ceramics, glass
or resin beads are blasted onto a sheet metal over a velocity of
several tens of meters per second. Their kinetic energy is
changed into the residual compression stress on the sheet
metal, so that a convex shape is formed without the die [2]. A
rubber mask is used to restrict the shot area. The blast
deteriorates the surface roughness, and sometimes induces the
work hardening which disturbs the forming performance.
Electromagnetic forming is a very high speed forming
method without one side of stamping dies. Because a current
pulse with very large peak and short duration is supplied to
generate magnetic field, equipment becomes large.
Laser forming uses laser pulses with a very short duration
to generate thermal expansion by heating or impulsive force
by the laser abrasion. Then the compression stress deforms the
sheet metal. Plasma jet also induces the compression stress for
forming [5]. Laser equipment itself and its running cost are
generally expensive.
The thermal strain or transformation generates a tensile
stress during electrical discharge machining (EDM) [6].
Because they frequently cause the distortion of a workpiece
and crack on its surface, they should be avoided in general
EDM process.
On the other hand, these phenomena have been positively
used to deform a sheet metal. Otsu et al. have proposed spark
forming in air to bend a sheet metal [7]. They generated spark
with a capacitance discharge circuit and tested the bending
performance under various conditions. The discharge point
was shifted by a constant interval. In addition, the temperature
and stress distributions have been simulated by the finite
element method. Yeo et al. have demonstrated EDM forming
in oil [8]. The maximum bending angle was 90 deg. with
wedged electrodes under a set of machining conditions.
However, the performance was not investigated in detail.

2212-8271 2014 Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer-review under responsibility of the International Scientific Committee of The 47th CIRP Conference on Manufacturing Systems
in the person of the Conference Chair Professor Hoda ElMaraghy
doi:10.1016/j.procir.2014.01.051
 Katsushi Furutani and Norimichi Yoshida / Procedia CIRP 17 (2014) 686 691 687

In this paper, the bending performance by EDM is was measured with an electric balance with a resolution of 0.1
experimentally investigated with various electrodes. Then an mg after ultrasonic cleaning in acetone for 10 min.
incremental forming with a thin rod is tried to avoid a warp of
a sheet metal. Finally, a pipe electrode is used to form a Thickness Thickness
convex surface. Electrode Z 3 mm Electrode
3 mm
Y X
2. Machining conditions Clamp 45 Clamp
Workpiece 45 Workpiece
a b
Figure 1 illustrates a setup for bending experiments. A
general die-sinking EDM machine (M35KC7G70 by Thickness Thickness
Electrode 1 mm Electrode
Mitsubishi Electric Corp.) was used for experiments. An 3 mm
electrode was mounted on the quill with a holder or chuck. An
Clamp Clamp
end of a workpiece was fixed with a clamp on the table. 30
Workpiece Workpiece
c d
Z Quill of
die-sinking
machine Fig. 2. Shape and arrangement of electrodes (a) 45-wedge (fixed side); (b)
Y X 45-wedge (free side); (c) 30-wedge; (d) 1-mm plate.

Electrode Chuck
Table 1. Machining conditions for wedge and plate electrodes.
Workpiece Working oil Open-circuit voltage V 80
Clamp Servo voltage % 25
Discharge current A 3
Magnet chuck
Table Pulse duration s 32
Working tank Pulse interval s 77
Machined depth mm 0.03
Fig. 1. Experimental setup. Jump up velocity mm/min 240

Workpieces were a 30400 stainless steel strip with Figure 3 illustrates a possible deformation mechanism. (1)
dimensions of 60 mm in length, 5 mm in width and 0.05 or A small area on a strip melts by a discharge. The surface
0.1 mm in thickness, and weighed 0.12 and 0.24 g, temperature rises to the melting point and heat conducts to the
respectively. They were cut out from a large sheet by wire interior. (2) The heat affected zone (HAZ) expands and
EDM. Their longitudinal direction corresponded with the compression stress acts there. Because the metal is easily
rolling direction. yielded at higher temperature, HAZ plastically deforms
Figure 2 shows the shapes and arrangements of wedged or though the rest zone deforms elastically. The plastic
plate copper electrodes with a width of 20 mm. The taper side deformation by the thermal stress at a high temperature
faced either the fixed or free side. The influence of a wedge depends on the stress level but is independent from the time
angle which restricts the maximum bending angle due to their duration of the applied stress [9]. (3) The molten zone is
geometry was tested for small angles by using the removed. (4) The HAZ is contracted by cooling. A tensile
arrangements shown in Figs. 2 (a), (b) and (c). A copper plate stress acts on the strip, so that the strip deforms concavely. A
with a thickness of 1 mm shown in Fig. 2 (d) was also used to steep temperature gradient to almost a half of the thickness is
test the influence of discharge area. The relative position in preferable for the bending process to yield in the HAZ and
the longitudinal direction was measured from the center of generate the compression stress by the elasticity.
discharge area which was often wide for the plate electrode.
In addition, a copper rod and pipe were used in experiments. Z Arc
Table 1 shows machining conditions for the wedge and column Molten
Arc Molten zone
plate electrodes, which are similar to ones described in Ref. column
Y X
zone
[8]. In order to reduce the removal rate, the discharge current
HAZ
was set to 3 A with a transistor discharge circuit. Because the Sheet metal
HAZ Expansion
servo voltage corresponding to the threshold of the average (1) (2)
gap-voltage was low, the delay time was often long. In the Removal Contraction
case of using a jump motion of the electrode, each motion and
machining duration took 0.3 and 0.25 s and the net machining
(3) (4)
time was 45% of the total one. Synthetic high-speed EDM oil
(Metal Work EDF-K2 by JX Nippon Oil & Energy Corp.)
Fig. 3. Deformation mechanism (HAZ: heat affected zone); (1) Discharge; (2)
was used for working fluid. Expansion and plastic deformation in HAZ and elastic deformation in rest
The median of 3 bending angles measured with an optical zone, and bending convexly; (3) Removal of molten zone; (4) Contraction by
microscope was adopted. A warp in the lateral direction and cooling in HAZ and bending concavely.
curvature were measured with a surface profiler. The weight
688 Katsushi Furutani and Norimichi Yoshida / Procedia CIRP 17 (2014) 686 691

By assuming that the surface temperature is the melting polarity. A side of a 0.1-mm strip was slit by the removal at
point and heat is conducted in one direction, the temperature T the positive polarity. On the contrary, a 0.05-mm strip was
at a depth of x after a time of t is calculated as: frequently slit at both polarities. In the following experiments,
the electrode polarity was mainly set to the negative to avoid
the heavy removal.
x
T Tm  T0 1  erf T
0
(1)
2 t c

where Tm, T0, , , c and erf() are the melting point (1673 K), Z X
room temperature (300 K), heat conductivity (16.7 W/(mK)), Y
density (7.93103 kg/m3), specific heat (590 J/(kgK)) and
error function, respectively. Figure 4 shows calculated a b
temperature distributions at various pulse durations. The
thermal stress possibly acts below 20 m from the surface. 1 mm
Z 1 mm
2000
Y X
Melting point
1500
Temperature K

pulse duration: 32 s
1000
16 s Fig. 5. Results of bending of 0.1-mm strips; (a) positive; (b) negative.

500 8 s
0 Figure 6 shows warps in the lateral direction. They were
0 10 20
Depth m
Fig. 4. Calculated temperature distribution at various pulse duration.
3. Experiments
3.1. Polarity of tool electrode
Strips with a thickness of 0.05 or 0.1 mm were bent to 90
in the arrangement shown in Fig. 2 (a). Table 2 shows the
bending angle, removal amount and removal rate. The
machining time containing the jump time at the negative
polarity for 90 bending was much longer than that at the
positive polarity. The removal rate at the negative polarity
was smaller than that at the positive polarity. The workpiece
3.2. Warp in lateral direction
Room temperature
30 40 50
measured at 1.5, 3.0 and 4.5 mm apart from the middle of the
discharge line toward the clamp. The wedged electrode was
arranged as shown in Fig. 2 (a). A 0.05-mm strip was warped
at near side from the discharge line and slanted a little. The
warp in a 0.1-mm strip was smaller than that in the 0.05-mm
one. The tensile stress act on the machined surface and in the
lateral direction formed the strip into a V-shape as described
in Chapter 2, and warped and sometimes twisted gently,
respectively. Consequently, the height of both sides was
higher than the center, and a higher side was removed as
shown in Fig. 5 (a).
Z
Height 10 m/div

4.5 mm 3.0 mm
was sometimes lifted together with the tool electrode during Y X
short-circuit at both polarities in the case that the molten 1.5 mm
surface of the strip by the discharge adhered to the electrode
and were resolidified. Because the restoring force due to
stiffness of the 0.1-mm strip is larger than that of the 0.05-mm
strip, the machining time for the 0.05-mm strip was often long. 0 1 2 3 4 5
The machining time has the same order regardless of the Position mm
thickness of the strips.
Fig. 6. Warp of 0.05-mm strip bent with wedged electrode.
Table 2. Influence of polarity of tool electrode.
Thickness Electrode Bending Removal Machining Removal
3.3. Progress of bending
mm polarity angle amount time s rate
deg. mg mg/min
Figure 7 shows the relationship between the bending angle
0.1 + 92.6 1.3 221 0.35
and machining time. The wedged electrode was arranged as
0.1 90.1 0.4 1890 0.013 shown in Fig. 2 (a). The height at 5 mm apart from the
0.05 + 90.0 0.2 132 0.091 discharged line was measured by the electrical contact to the
0.05 87.1 0.1 2493 0.002 tool electrode every 180 s, and the bending angle was
calculated. The measured point was alternated to 1 or 2 mm
Figure 5 shows results of bending of 0.1-mm strips. The for the large one. The bending angle was steeply increased
sharp bending edge was observed in both cases of electrode over 15. The machining time for 90 bending for 0.1 mm was
 Katsushi Furutani and Norimichi Yoshida / Procedia CIRP 17 (2014) 686 691 689

shorter than that for 0.05 mm. When a workpiece was warped
in the lateral direction, the stiffness in the longitudinal
direction was increased. Then, the discharge was not
distributed to the whole area and the tensile stress in the
longitudinal direction was disturbed. By decreasing the
thickness on the discharged line due to the removal and
increasing the tensile stress in the lateral direction, the
stiffness in the longitudinal direction was decreased over 15, 1 mm
so that the strip was bent easily. Unstable discharge due to the
adhesion also caused the longer bending time for the 0.05-mm
strip as described in Section 3.1.

90 Fig. 8. Result of bending with 30 wedge.

0.1 mm Table 4 shows the bending angle machined with the plate
60
electrode. Because the discharge was dispersed in the bottom
Angle deg.

0.05 mm area of the tool electrode and the residual stress on the strip
30 surface became small, the bending angle was much smaller
than 90.
0
Table 4. Bending angle with 1-mm plate electrode.
0 1000 2000 3000 4000 5000
Electrode Machining time Bending angle
Machining time s Thickness mm
polarity s deg.
Fig. 7. Relationship between bending angle and machining time. + 221 30.9
0.1
1890 42.8
3.4. Influence of wedge shape
+ 132 7.6
0.05
To investigate the influence of wedge shape, the 45 2493 6.4
asymmetrical and 30-symmetrical tool electrodes shown in
Figs. 2 (a), (b) and (c) were used. A 0.05-mm strip was bent Figure 9 (a) shows a result of bending of a 0.1-mm strip.
for 5 to avoid the geometrical restriction due to the Discharge occurred on the area of workpiece projected below
arrangement or the tip angle of the tool electrode. The the tool electrode, and the area gradually spread. Then the gap
bending angle was measured by the electrical contact of the length on the free side became shorter. Therefore, there were
tool electrode to the strip every 10 s. more craters on the free side. Because the discharge was
Table 3 shows result of bending. The bending time and rate distributed and many did not contribute bending, the bending
were almost the same for all conditions. Therefore, the angle was smaller than that with the wedged electrodes.
arrangement of the electrode did not affect the when its tip Figure 10 shows warps bent with a plate electrode in the
was sharp. lateral direction. They were smaller than those with a wedged
electrode because the discharge distributed in the longitudinal
Table 3. Influence of electrode arrangement on bending angle.
direction and bending force in the lateral direction was larger.
Arrangement Machining Bending
Sample No. Rate s/deg.
in Fig. 2 time s angle deg.
1 70 4.9 14.3
(a)
2 40 4.2 9.4
1 50 5.3 9.5
(b)
2 55 6.4 8.6
1 40 6.1 6.5 a b
(c)
2 50 5.8 9.0

1 mm 1 mm
Figure 8 shows the machining result by the arrangement
shown in Fig. 2 (c). The discharge occurred only the tip of the
tool electrode and the craters were distributed in line. Because
the machining time was short and removal amount was small,
the appearances were also very similar to the others.
Fig. 9. Results of bending with plate electrode; (a) 0.1 mm; (b) 0.05 mm.
To investigate the influence of the discharge distribution in
the longitudinal direction, the plate electrode shown in Fig. 2
The discharge was concentrated at a side and no crater was
(d) was used. The machining time was set to the same as that
observed at the center area on a 0.05-mm strip. The bending
for each polarity and thickness as shown in Table 2.
angle was small due to a warp in the lateral direction.
690 Katsushi Furutani and Norimichi Yoshida / Procedia CIRP 17 (2014) 686 691

With the wedge electrode, the discharge area was
concentrated in the longitudinal direction and spread in the
lateral direction. On the contrary, with the plate electrode, the
discharge area spread also in the longitudinal direction.
Consequently, the bending angle with a wedge electrode was
larger than that with a plate electrode.
that of a 0.1-mm strip. Because the gap length was
automatically adjusted during EDM even if the strip was
warped, the discharge area was not restricted to the side with
the rod electrode. Therefore, repeating of short-time
machining is preferable for uniform bending. The jump
motion did not affect the bending angle.

Table 5. Bending angle by incremental forming (w: with, w/o: without).

Machining Bending
Number of Thickness Electrode
Height 10 m/div

4.5 mm Jump time / point angle

machining mm polarity
3.0 mm s deg.
+ w/o 7 12.9
1.5 mm 1 0.05
w/o 126 12.3
+ w/o 1 35.0
w/o 25 13.1
0 1 2 3 4 5 5 0.05
Position mm + w 3 37.1
w 55 17.6
Fig. 10. Warp of 0.05-mm strip bent with plate electrode. + w 1.5 43.9
0.05
w 28 17.3
3.5. Incremental forming by sequential EDM 10
+ w 3 24.8
0.1
w 21 8.4
The warp of the strip restricted the discharge area in the
case of using wedge electrodes. To form various shape on a
wide strip, an incremental forming by EDM with a thin rod Figure 12 shows a result of bending of a 0.05 mm-strip
was tried. with the jump motion. The machining sequence was repeated
Figure 11 illustrates a machining sequence. A copper rod 5 times at the negative polarity. The discharge was uniformly
electrode with a diameter of 1 mm was shifted with 0.5-mm occurred on the whole area.
intervals during EDM under the same electrical conditions as
shown in Table 1. This sequence was repeated once, five or
ten times. The net machining time for each set of the
conditions was the same as Table 2.

Electrode
Z
Y
1 mm
X Series machining
Specimen

5 mm Fixed end
Free end
A Machined points Fig. 12. Result of bending with rod electrode.

Machined points It is preferable for the discharge area to be restricted in the

longitudinal direction and to spread in the lateral direction
Y
uniformly, based on the results with the wedge, plate and rod
5 mm

Intervals

Workpiece
0.5 mm

electrodes.
Z X
3.6. Forming with pipe electrode
A

Fig. 11. Sequence of incremental forming by EDM with thin rod.
Table 5 shows the bending angle. Because the longitudinal
length of the discharge area was the same as that with the 1-
mm plate electrode, the bending angle was almost the same as
that with the plate electrode. The bending angle became larger
by repeating the sequence. The discharge distribution
uniformly generated the tensile stress on the whole discharge
area. The bending angle of a 0.05-mm strip was larger than
The tensile stress acting omnidirectionally was applied to
form a convex surface. Copper pipes with outer diameters of
5.0 or 2.5 mm were used as an electrode. Figure 13 shows an
appearance of the 5-mm electrode which inner diameter was
2.75 mm. An end of each electrode was chamfered with a drill
tip with an angle of 120 in order to sharpen the edge.
Table 6 shows the machining conditions for forming with
the pipe electrodes. To reduce the machining time, the
electrode polarity was set to positive.
 Katsushi Furutani and Norimichi Yoshida / Procedia CIRP 17 (2014) 686 691 691

Progrram for Form ming Strategicc Research Infrastructure

I for
Privaate Universitiees, Smart MMachine Technology Reseaarch
Centrre, a matchin ng fund subsiidy by Ministtry of Educattion,
Cultuure, Sports, Sccience and Tecchnology, Jap
pan.

Figg. 13. 5-mm coppper pipe electrodee.

Tablee 6. Machining coonditions for form

ming by EDM witth pipe electrodess.
Open-circuit voltage V 80
Dischaarge current A 4
Pulse duration
d us 16
Pulse interval
i us 2048
Electro
ode polarity +
Machiining time s 120

Fig. 14. Curv

ved surface formeed with 5-mm pip
pe electrode.
Fiigure 14 show ws an appearannce of a curveed surface form med
by EDM
E with thee 5-mm pipe electrode. More M craters w
were
obserrved on the freee side. L
Longitudinal

Height 20 m/div
Fiigure 15 shoows profiles of the curved surface. The
curvaature radii weere 89 and 1661 mm in the longitudinal and
lateraal directions, respectivelyy. Burs were formed on the
bounndary. Because the workpiiece was not a square andd its
size was
w not relatiively large aggainst the diam meter of the ppipe Lateraal
electrrode, the sides were warpedd. Consequen ntly, the curvaature
radiu
us in the lateeral direction was larger than that in the 0 2 4 6 8 10
longiitudinal directtion. Positioon mm
A curved surfaace was formeed also with the 2.5-mm ppipe
electrrode. Its curvvature radius was 110 and d 56 mm in the Fiig. 15. Profiles off curved surface fformed with 5-mm
m pipe electrode.
longiitudinal and laateral directionns, respectivelly.
Referrences
4. Co
onclusions
[1] Voollertsen F, Hu Z, Z Niehoff HS, Theiler C. Statee of the art in microm
In
n this paper,, the bendinng performan nce by EDM
M is forrming and investigations into m micro deep drawiing. J Mater Proocess
Teechnol 2004;151:7 70-9.
experrimentally innvestigated with
w wedge, plate and ppipe [2] Miiao Hy, Larose S, Perron C, Deemers D, Levesq que M. Experimental
electrrodes. Concluusions can be drawn
d as follo
ows. stu
udy of shot peeniing and stress peeen forming. J Mater
M Process Tecchnol
2010;210:2089-102 2.
x Affter comparinng the machhining resultss with the pplate [3] Psy yk V, Risch D, Kinsey
K BL, Tekkakayaa AE, Kleiner M. Electromagnnetic
electrode with that with the rod one, the discharge shoould forrming a review.. J Mater Processs Technol 2011;2 211:787829.
[4] Daaniel FW, Sameerr V. Bending of T Titanium Sheet Using
U Laser Formming.
bee dispersed in the lateral diirection and be concentratedd in J Manuf
M Process 20000;2:258-69.
th
he longitudinall direction forr a large bendiing angle. [5] Maale AT, Pan C, Chen YW, Li PPJ, Zhang YM. Processing P Effects in
x Allthough the bending anggle at the po ositive electrrode Plaasma Forming off Sheet Metal. CIR RP Ann Manuf Technol
T 2000;49::213-
po
olarity was larrger than thatt at the negatiive, the electrrode 6.
po
olarity was ppreferably neggative in order to reduce the [6] Ekkmekci B, Tekka aya AE, Erdencc A. A semi-em mpirical approachh for
ressidual stresses in electric dischargge machining (EDDM). Int J Mach Tool
removal of the wworkpiece. Maanuf 2006;46:858 8-68.
x Foorming curveed surfaces with a pipee electrode was [7] Otsu M, Wada T, Osakada K. Mi Micro-bending of thin spring by laser
deemonstrated. forrming and spark forming.
f CIRP AnAnn 2001; 50: 1411-4.
[8] Yeeo S H; Tan K H. A novel appr proach in microfooil bending usinng an
Ackn
nowledgemen
nts eleectrodischarge machine.
m Proc Insst Mech. Eng., PtP B: J Engg Manuf
M
2004; 218: 1403-7.
[9] Booley BA, Weiner JH. Theory of T Thermal Stress. New
N York, NY, USA: U
Th
he authors wwish to thankk Guest Proff. K. Kondohh of Johhn Wiley & Sons; 1960.
Toyoota Technologgical Institute for his valuab
ble advice. A ppart
of th
his study waas financiallyy supported by the Suppport