Materials and Design 49 (2013) 736745

Contents lists available at SciVerse ScienceDirect

Materials and Design

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

Research on the impact velocity of magnetic impulse welding of pipe tting

Xu Zhidan a, Cui Junjia a,, Yu Haiping a,b,, Li Chunfeng a
a
b

National Key Laboratory for Precision Hot Processing of Metals, Harbin Institute of Technology, Harbin 150001, China
The Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150001, China

a r t i c l e

i n f o

Article history:
Received 20 September 2012
Accepted 24 December 2012
Available online 2 January 2013
Keywords:
Aluminum alloy
Magnetic impulse welding
Wavy interface
Impact velocity

a b s t r a c t
Magnetic impulse welding, which is uniquely advantageous in welding heterologous pipe ttings, is a
new welding technology based on high-speed magnetic impulse shaping and solid-phase diffusion welding. The impact velocity of the welding points of AlFe heterologous pipe ttings was studied by combining numerical simulation and technological test with the assistance of constitutive relations of 3A21Al
alloy under a high strain rate. The momentary movement speeds when the outer tube (A1) impacts
the inner tube under four different voltages were analyzed to obtain the critical voltage for welding inner
and outer tubes. The speed of the welding points of the outer tube noticeably increased with the rise in
the discharge voltage. The weld interfaces of both the inner and outer tubes produced regular zigzag
waves when the impact velocity reached 350 m/s. The energy spectrum analysis revealed that pipes
undergo severe deformation under high-speed impact, and the increased temperature enhances the
activity of the atoms among other elements, thus producing a surface mass ow under strong impact
and granulated substances.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
Recently, the problem of lightweighting has gained more attention from around the world [1], with the increasing importance of
employing new measures to solving it to improve the application
of aluminum products. Notably, the application of the steelaluminum dual metallic welding structure has become the preferred
alternative to lightweighting in industrial production, which surely
involves the connection between aluminum and steel [2].
Since the 1960s, scholars from different countries have performed detailed research on the connection between the structures
of aluminum and steel. The welding of the aluminumsteel structure involves various methods used in the eld of welding [3], such
as explosion welding [4], friction welding [5], brazing welding [6],
and laser beam welding [7].
Some welding techniques such as electric resistance welding
(ERW) and brazing welding are carried out under high temperatures, which produces brittle intermetallic compound on the interface of the aluminumsteel structure, thus reducing the quality of
the welding point. On the other hand, explosion welding is a technique that is commonly applied in the solid phase connection between same metals (e.g. steelsteel, AlAl, CuCu) or different
metals (e.g. Alsteel, Custeel). However, this technique utilizes
explosives for its energy supply and the resulting powder charge
is very complicated, thus requiring a higher labor intensity and
Corresponding authors. Tel.: +86 451 86413970; fax: +86 451 86418753.
E-mail addresses: cuijunjia@hit.edu.cn (J. Cui), haipingy@hit.edu.cn (H. Yu).
0261-3069/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.matdes.2012.12.059

production costs, making it difcult to achieve automatic

production.
Magnetic pulse welding (MPW) technologies based on solid
state impact or collision welding can provide linear or circular
welds, which has been established as fast, reliable and cost-effective. Meanwhile, the magnetic pulse welding technology is characterized as a very short welding process that is accomplished within
a microsecond. Furthermore, the increased temperature caused by
the heat of the inner and outer tube impact is not enough to produce chemical compounds composed of Al and Fe. Therefore, no
metallic compounds will be produced in the transition zone, and
this method will not reduce the quality of the welding point significantly [8]. In addition, the use of magnetic pulse welding technology has many advantages, such as having controllable discharge
energy, high production efciency, and a unique advantage in tting welding.
A typical MPW system includes a power supply, a switching
system, and a coil. The parts to be welded are inserted into the coil
with the capacitor bank being charged, and then the switch is triggered in transient time, the current ows through the coil. After
that, the current is applied to the coil, a high-density magnetic ux
is created around the coil, and as a result an eddy current is created
in the workpieces. The eddy currents oppose the magnetic eld in
the coil and a repulsive force is created. This force can drive the
workpieces together at an extremely high rate of speed and creates
a type of weld.
Over the past decade, MPW has been successfully applied for
tube to tube impulse welding with both similar and dissimilar

737

Z. Xu et al. / Materials and Design 49 (2013) 736745

metals [9]. For example, Hokari et al. welded aluminum tube to

copper tube, and found it was more susceptible to deform outer
tube made of aluminum than copper [10]. Aizawa et al. investigated the welding process parameters and reported characteristics
of dissimilar materials (AlFe) [11]. Ben-Artzy et al. found that
intermetallic phases (IMP) of different compositions were created
during welding of the AlMg couple by rapid solidication of a thin
melted layer at the interface [12].
Previous studies on magnetic pulse welding were focused on
factors such as parameter optimization [13], electromagnetic force
[14], interface appearance [15], nite element simulation [16], and
organization, among many others [17]. Problems encountered during the welding process, such as the relationship between the
movement speed of the outer tube moving toward the inner tube
and the discharge voltage, the impact of movement speed of the
outer tube on welding strength, etc. were left out.
In the present study, numerical simulation mean and the constitutive relation of a 3A21Al alloy under dynamic load established
by the Hopkinson bar test were applied to study the voltage impact
on electromagnetic eld, magnetic eld force, and on the movement speed of the outer tube. Then, a high-speed camera system
was utilized to measure the movement speed of the contour of
the outer tube edge during magnetic impulse welding under different voltages. The results were compared with those obtained from
the numerical simulation. Finally, the impact of the movement
speed of the outer tube on the strength of the magnetic impulse
welding point and on the microstructure was analyzed.
2. Experimental material and method
2.1. Experimental material
Steels 20 and 3A21 (Al alloy) were selected as materials of the
inner and outer tubes, and the sizes of which are shown in Table 1.
The tubes were heated to the annealed condition to soften the outer material. The inner tube must have an angle of 4 in the joining
zones to establish a better welding effect. The mechanical properties of the 3A21 after the annealing process are shown in Table 2.

Table 2
Mechanical properties of 3A21 after annealing.
Yield strength (MPa)

Tensile strength (MPa)

Poissons ratio

42

97

0.33

coil
inner tube

field shaper
outer tube

Fig. 1. Test equipment and installation.

with the inner tube is the most important factor to measure. However, during measurement, the outer tube produces a different axial deformation that will cause other positions to collide and
produce sparks even before the measurement point comes into
contact with the inner tube. The measurement result will thus be
inuenced directly. Therefore, the inner tube was removed, and
only the outer tube was installed during measurement.
The pictures recorded by the high-speed camera were imported
into the AUTOCAD software, and the inner wall diameters of the
outer tube in different moments were measured using the label
function of the AUTOCAD software. To guarantee the accuracy of
the measurement, a circumference of 45 was taken as intervals,
and four directions were selected for the measurement. The average value was then obtained (Fig. 3).
2.3. Establishment of a nite element model for electromagnetic eld
analysis during magnetic impulse welding

2.2. Experimental method

2.2.1. Magnetic impulse welding equipment
A 20-kJ electromagnetic forming machine with a capacitance of
100 lF and a maximum charging voltage of 20 kV was used as the
test equipment. The coil was wound from the rectangular wire
with a cross-section measuring 7 mm  5 mm. Considering that
the overlap joint end is 15 mm long, and that the turns of the coil
are restricted, the eld shaper structure whose main material is red
copper was applied. Fig. 1 shows the assembly and the structure of
the eld shaper.
2.2.2. Measuring system
A high-speed camera system (FASTCAM SA5 1000 K-M2) was
used to measure the movement of the outer tube, as shown in
Fig. 2. The FASTCAM SA5 has a maximum shooting speed of one
million frames per second and the shortest exposure time of
370 ns. With respect to the speed of the pipe-pipe welding point,
the instantaneous speed before the outer tube comes into contact

Table 1
Sizes of inner and outer tubes.
Material

Outer diameter (mm)

Inner diameter (mm)

Thickness (mm)

3A21-O
Steel 20

20
15.2

18
8

1
3.2

2.3.1. Coupling method selection

The electromagnetic welding process involves multi-disciplinary crossing, and it is unable to establish and solve the coupled
equation between structural changes and electromagnetic elds
during deformation [18]. As a result, current studies are focused
mainly on electromagnetic eld analysis and workpiece deformation analysis [19]. This paper applied the loose coupling method
to establish the nite element model. Using the nite element software ANSYS as the platform, the electromagnetic eld was determined using the ANSYS/Multiphysics module, thus obtaining the
space-time distribution of the magnetic pressure. Furthermore,
the deformation of the outer tube was analyzed through the
ANSYS/LS-DYNA module, with the magnetic pressure serving as
the boundary condition. The ow chart of the simulation is shown
in Fig. 4.
2.3.2. Electromagnetic eld analytical model
In ANSYS/Multiphysics, a circuit can be directly connected to
the current supply in the nite element area through one share
node or a group of share nodes to achieve coupling. Coil, eld shaper, inner and outer tubes, and air section applied the SOLID117
unit with a circuit-magnetic coupling function for gridding
dividing. During the electromagnetic welding, air dielectric can
be regarded as an isotropic homogeneous medium with innite

738

Z. Xu et al. / Materials and Design 49 (2013) 736745

Fig. 2. Measuring system.

d'

a
b

c'

b'

d
a'

Fig. 3. Picture of measurement locations.

Fig. 5. Pearson coil and oscilloscope.

resistivity, permeability similar to that of a vacuum, and a relative

permeability of 1.
As the experiment applied the eld shaper structure, the inductance of the system included not only the inductance of the coil
and eld shaper but also the mutual inductance of the two. A Pearson coil and oscilloscope were applied for the measurement to obtain the accurate inductance value of the discharge circuit (Fig. 5).
The discharge voltage was 10 kV. The current wave of the discharge circuit was obtained through the oscilloscope (Fig. 6). The
inductance of the discharge circuit can be calculated at approximately 3.456 lH according to the computational
formula of the
p
discharge period of the RLC circuit: T 2p LC .
The nite element analytical model of magnetic impulse welding is shown in Fig. 7. Mapping gridding division was applied to the
coil with a regular shape as well as to the inner and tubes, whereas
for the eld shaper and the air eld, free gridding division was ap-

plied. The outer tube thickness was divided into three-layer units
considering the impact of skin effect. Taking the discharge current
from the equipment as the exciting condition, the form-wound coil
could serve as the current-carrying coil.
2.3.3. An analytical model for deformation during magnetic impulse
welding
Applying the dynamic analysis method is necessary in addressing instantaneous dynamic problems. The supporting dynamic
structural analysis and plastic dynamic response analysis
(ANSYS/LS-DYNA) can be used to solve instantaneous large strain
problem, high nonlinear problem, complicated contact-impact
problem, etc. [20].
The PLANE162 unit is used in two-dimensional modeling, which
is characterized as quick solving and simple modeling. The

Fig. 4. The ow chart of the simulation.

739

Z. Xu et al. / Materials and Design 49 (2013) 736745

r 92:5 114:2e0:2762 1 0:2062 ln e_  

60

The outer tube was made up of 3A21 aluminum alloy, and its
magnetic pressure originated from the solution result of the electromagnetic eld coupling analysis. For computational efciency,
half of the studied model was applied for modeling because the
model is an axial symmetrical model. The nite element model
established for deformation analysis is shown in Fig. 8.

40

Current (kA)

20

0
100

200

300

400

500

-20

3. Result and discussion

3.1. Electromagnetic eld and magnetic eld force

-40

Time (s)
Fig. 6. Current wave of the discharge circuit.

Fig. 7. Finite element model of electromagnetic welding and the view of a nite
element gridding division.

PLANE162 unit was selected to establish the axial symmetry model

because the magnetic impulse welding system belongs to the axial
symmetry structure. For the contact, a symmetrical penalty function algorithm was applied, and the ASS2D contact model that is
exclusively used by the PLANE162 unit was used.
Constitutive relation refers to the function that describes the
corresponding relation between stress and strain in a deformable
body [21]. A reasonable constitutive relation is the basis that guarantees that the simulation process can truly reect the objective of
the test [22]. As it is a high-speed impact process, the electromagnetic riveting process applied the JohnsonCook model [23]. GB/T
7314-2005 standard was adopted to quasi-static tests [24]. The
size of dynamic forming test samples is u10  8 mm, and experimental conditions were shown in Table 3. Given the absence of a
standard for dynamic compression test, the sample size of the
standard static test followed the standard sample size of the standard static test.
Based on the quasi-static test of the 3A21 aluminum alloy, the
JohnsonCook constitutive equation determined from the test data
under dynamic room temperature can be simplied as

The magnetic eld distribution during the rst half period of the
discharge time is shown in Fig. 9. During the rst half period of the
discharge time, the magnetic eld produced a trend wherein it increased rst and then decreased as time progressed, reaching its
peak in the rst quarter period of the discharge time. The magnetic
eld was mainly concentrated in the air between the outer tube
and the eld shaper within the rst quarter period of the discharge
time because of the electromagnetic shielding effect of the outer
tube. However, the magnetic eld decreased and penetrated the inner tube through the air surrounding the end face of the outer tube
after the rst quarter period of the discharge time. At this point,
the magnetic eld weakened. As a result, the existence of the inner
tube would not greatly affect the magnetic eld between the outer
tube and the eld shaper.
The distribution of the magnetic eld force vector of the outer
tube is shown in Fig. 10. The end face of the outer tube was simultaneously affected by a smaller axial magnetic eld force because
of the magnetic eld that was distributed in the air of the upper
end face. The end face of the outer tube had the smallest deformation resistance because of the lack of restriction of the outer metal,
thus making it the rst place of deformation. Driven by the deformation of the end metal and affected by the magnetic eld force,
the inner metal of the outer tube would cause radial shrinkage
deformation. The variation of the magnetic eld force with time
was found to be the same as that of the magnetic eld. The magnetic eld force reached the maximum value at the rst quarter
period.
The discharge voltage, which is easy to be changed and controlled, is the most important electrical parameter during a magnetic impulse welding procedure. The variation of the magnetic
eld force under different discharge voltages at the rst quarter
period is shown in Fig. 11. The outer tube suffered complicated
stresses and was affected by the axial and radial forces simultaneously. Furthermore, the outer tube had an uneven distribution,
with the maximum electromagnetic force observed in its center.
This result is caused by the location of the maximum magnetic

Outer tube

Coil

Joining Area

Table 3
Experimental conditions of dynamic forming test.
Dynamic experiment

Inner tube

Field shaper

Quasi static experiment

Ination pressure (MPa)

Strain rate (s1)

Strain rate (s1)

0.2
0.6
0.8

820
1441
2091

0.01

Fig. 8. Finite element model for deformation analysis.

740

Z. Xu et al. / Materials and Design 49 (2013) 736745

(a) 6 s

(b) 18 s

(c) 33 s

(d) 60 s

Fig. 9. Magnetic eld distribution in different periods.

(a) 6 s

(b) 18 s

(c) 33 s

(d) 60 s

Fig. 10. Magnetic eld force distribution at different periods.

(a) 9 kV

(b) 11 kV

(c) 13 kV

(d) 14 kV

Fig. 11. The magnetic eld force distribution at the rst quarter period under different voltages.

eld in the axial center of the air region between the eld shaper
and the outer tube. During magnetic impulse welding, the discharge energy is proportional to the square of the voltage under
xed capacitance. The larger the discharge energy is, the larger
the energy of the outer tube and the larger the impact velocity will
be. The simplest way to increase the quality of the magnetic impulse welding is to increase the discharge voltage.
3.2. Finite element analysis of the deformation eld
Fig. 12 shows the cloud picture of the outer tube velocity at different periods under a discharge voltage of 14 kV. The outer tube

velocity continuously increased as time progressed, and the maximum value of the velocity was reached when the outer tube made
contact with the inner tube. During the deformation process, the
end of the outer tube achieved the maximum movement velocity.
The joining region, which is far away from the end of the outer
tube, displayed the same movement velocity with the outer tube
because of the even distribution of electromagnetic eld and magnetic eld force. The velocity and time chart of the outer tube end
under different voltages are shown in Fig. 13. The gure shows that
the velocity signicantly increased as the voltage increased, and
that the deformation of the outer tube lasted the longest when
the discharge voltage value was 9 kV, reaching nearly 20 ls. At this

741

Z. Xu et al. / Materials and Design 49 (2013) 736745

(a) 3.14 s

(b) 9.14 s

(c) 12.148 s

(d) 15.446 s

Fig. 12. Cloud picture of the outer tube velocity at different periods.

point, the impact speed of the inner and outer tubes was 249 m/s.
When the discharge voltage value was 14 kV, the deformation
lasted for 14 ls, and the impact speed reached 368 m/s.
3.3. Measuring speed
During magnetic impulse welding, a high-speed impact is observed between the outer tube and the inner tube under the effect
400
9kV
11kV
13kV
14kV

350

Velocity (m/s)

300
250
200
150
100
50
0

10

15

20

Time ( s)
Fig. 13. The simulated velocity of the outer tube under different voltages.

of the magnetic impulsive force, forming a jet ow and accomplishing the welding process. As the interval between the inner
and outer tubes was set to 1.4 mm in the experimental condition,
the inner and outer tubes would impact with each other when the
inner diameter of the outer tube is shortened to 15.2 mm. Therefore, obtaining the impact velocities under different voltages based
on the speed-time curve is possible.
Fig. 14 presents the images of different periods under a discharge voltage of 14 kV collected by the high-speed camera. Based
on these images, the diameter of the pipe tting at different periods can be measured, and the movement speed variation curve
of the outer tube with time can be acquired by the average speed
between two images, as shown in Fig. 15. The deformation of the
outer tube accelerated at rst and then decelerated under the effect of the magnetic eld force. The deformation process comprised
a free contractive diameter and an instability wrinkling of the outer tube. The inner and outer tubes produced an impact during the
deformation process involving a free contractive diameter. Therefore, the different impact velocities were found to have been
caused by the different interval between the inner and outer tubes.
A large interval causes a large deformation of the outer tube, thus
producing instability wrinkling. However, small intervals will
make it difcult for the outer tube to obtain enough acceleration
time, resulting in a small impact velocity and making it difcult
to achieve a welding effect. In view of the inner diameter measurement of the outer tube at different periods (Fig. 14), the outer tube

742

Z. Xu et al. / Materials and Design 49 (2013) 736745

(a) 0 s

(b) 1.9 s

(c) 3.8 s

(d) 5.7 s

(e) 7.6 s

(f) 9.5 s

Fig. 14. Pictures of the high-speed deformation of the outer tube at different periods.

450

9kV
11kV
13kV
14kV

400

Velocity (m/s)

350
300
250
200
150
100
50
0
0

10

12

14

16

Time (s)

The velocity obtained by the outer tube completely comes from

the energy stored in the electromagnetic forming machine. Based
on the computational formula of energy of the electromagnetic
forming machine E 12 CU 2 , the energy stored in the electromagnetic forming machine is proportional to the square of the charging
voltage. The impact velocity of the inner and outer tubes increases
at about 40 m/s with every 1.5 kJ increase in the discharge energy
under such experimental condition. During the experiment, the
outer tube produced high-speed diameter shrinkage deformation
within a microsecond, and the air inside the tube was compressed
severely, thus bringing an outward support to the outer tube in the
radial direction. However, the impact of the air inside the tube was
neglected during the simulation, resulting in a smaller measured
value compared with the simulation value.

Fig. 15. The practical measuring speed of the outer tube under different voltages.

3.4. Scanning electron microscope and peeling test

began to wrinkle when the inner diameter was reduced to less
than 15.2 mm, before which the diameter shrinkage velocity of
the outer tube increased continuously. Therefore, the welding
achieved the best effect when the interval between the inner and
outer tubes was 1.4 mm.
As observed from Figs. 14 and 15, the outer tube was unable to
produce plastic deformation during the initial period of the magnetic eld force effect because of the small magnetic eld force.
However, as the magnetic eld force gradually increased with
the increase in time, the outer tube produced micro-deformation
under the effect of the magnetic eld force. At this point, the deformation time of the outer tube obtained through Fig. 14 was a little
shorter than the practical deformation time because the initial period of deformation of the outer tube during image collection is difcult to determine because of the small deformation in the initial
period and the errors in the inner diameter measurement. After
3.5 ls, the magnetic eld force is larger than the deformation resistance of the outer tube, making the outer tube able to produce
plastic deformation. The speed of the outer tube increases continuously until it comes into contact with the inner tube. The whole
deformation process lasts for 20 ls. Table 4 presents the simulation impact velocities and the practical impact velocities of various
directions under different voltages of Fig. 3.

The articial peeling test was carried out to the joint with over
227 m/s impact velocity. DIN EN ISO 10447-2007 standard was
adopted to test welding strength, and the test piece was cut into
strips axially at the joint [25]. The result of the peeling test is
shown in Fig. 16. As observed on the joint surface, one end of the
joint began to produce one strip of circumferential white impact
abrasion under an impact speed of 227 m/s. Moreover, the color
of the abrasion gradually became more distinct as the voltage increased, turning into a piece of a bonding belt approximately 1
2 mm wide. As this area corresponded to the end face of the working area of the eld shaper, the outer tube bore the largest stress,
thus causing a high-speed impact. Therefore, when the impact
velocity of the outer tube was larger than 200 m/s, the higher connection strength of the joint compared with the strength of the aluminum product was found but no alloy joint formed in the inner
and outer tubes.
When the impact velocity reached over 278 m/s, the aluminum
tube became more difcult to peel. When the impact velocity
reached 355 m/s, the strip aluminum tube was torn down while
the rest of the aluminum tubes were still adhered to the external
surface of the inner tube. The aluminum tube was peeled under
such impact velocity. Then, regional gray aluminum scraps and

743

Z. Xu et al. / Materials and Design 49 (2013) 736745

Table 4
Impact velocity of the inner and outer tubes.
U (kV)

E (kJ)

aa0 (m/s)

bb0 (m/s)

cc0 (m/s)

dd0 (m/s)

Vtest (m/s)

Vsimulate (m/s)

9
11
13
14

4
6
8.5
10

232
271
306
350

226
278
319
348

220
281
332
364

230
283
329
353

227
278
322
355

249
300
347
368

(a) 227 m/s

(b) 278 m/s

(c) 322 m/s

(d) 355 m/s

Fig. 17. Scanning picture of the connection interface morphology under magnication of 100 times.

Fig. 16. Peeling test of the connection joint under different impact velocities.

black substances zone left by the rigid separation of the aluminum

tube were observed on the external surface of the steel tube. The
peeling strength in this area reached the highest value. Compared
with other welding technologies such as friction welding, studies
have shown that only the aluminum element diffuses with the
steel inside, but other elements like Fe and Cr produce no obvious
diffusions. Furthermore, between the stainless steel particles and
aluminum, there were no obvious diffusions produced on the
welding position, and the fatigue performance of the joint was
about 30% lower than that with aluminum base [26]. In addition,
the peeling test witnessed a greater strength in the magnetic pulse
welding joint compared to that with a metal base. Therefore, as far
as the aluminumsteel structure welding is concerned, the magnetic pulse welding method can achieve better results than friction
welding. With regard to the explosion welding technology, Acarer
et al. applied appropriate heating processes to the explosion welding joint of the aluminumsteel structure to increase the welding
quality, thereby improving the welding joints toughness and
shearing strength [27]. Therefore, when the joint strength is lower
than that of the base metal, the magnetic pulse welding technology
can employ similar welding methods to improve the results.
A test sample with an impact velocity of 355 m/s was cut along
its axial joint for polishing. The sample was then corroded by 3%
nitric acid and alcoholic solution to observe the metallographic
phase of the carbon steel near the interface. As shown in Fig. 17,
a relatively zigzag wave-shaped interface was formed on the interface integrating two kinds of metals with a wave length of about
50 lm and a wave height of about 10 lm. This phenomenon is different from other welding technologies such as brazing welding
and fusion welding because these utilizes atomic diffusion, as produced by the melting of the welded metal under high temperatures, which consequently produce straight interfaces [28]. The
wave interface of a magnetic pulse welding joint is very similar
to the resulting explosion welding interface, except that the wave

height of the latter is relative higher about 10 times higher and 4

times wider than the wave of a magnetic pulse welding interface
[29]. Previous research demonstrated that the higher and wider
the waves in an interface are, the better the welding joints performance [30]. Therefore, the explosion welding joint shall possess
greater strength than the magnetic pulse welding joint. To signicantly increase the joint strength of magnetic pulse welding, a
higher energy utilization rate is required to obtain a greater impact
velocity. In the explosive welding study conducted by Acarer et al.
[29], the wavelength of the weld interface was approximately
400 lm, and the wave height was 80 lm. Based on the geometrical
parameters of the wave, the weld interfaces of both the magnetic
pulse and explosive welding were standard sine waves and their
k/h (where k represents the wavelength and h represents the wave
height) were valued at 5. Both of these two welding methods produced a high-velocity impact between two metals through external force and accomplished welding during impact, and thus, the
geometrical parameters of their interface are the same except that
the impact velocity between the two metals in explosive welding
was larger than that in magnetic pulse welding. Therefore, the
wavelength and wave height of the explosive welding interface
were signicantly larger compared with those of the magnetic
pulse welding interface. In light of the above test results, the joint
strength of magnetic pulse welding was already larger than that
with the aluminum base, given that the weak zone in the connection structure is found in the base metal. Therefore, the joint
strength can already meet the needs.
Meanwhile, it was found that granular intermediates were
gradually produced with increasing distance to the initial connection end (Fig. 18). Magnied 500 times, a wave-shaped gap between two pipe matrices was observed, and this gap signicantly
decreased the strength of the joint. When the joint was under
high-strength peeling condition, the aluminum tube was peeled
rigidly despite the similar microwave-shaped interface acquired.

744

Z. Xu et al. / Materials and Design 49 (2013) 736745

Table 5
Energy spectrum analysis result.
Element

Al
Fe

Mass fraction

Atom fraction

6.70
93.30

4.09
95.91

2.25
97.75

12.94
87.06

8.12
91.88

4.54
95.46

particles are Fe. Under such at and straight joint condition, the
outer tube was accelerated to the radial velocity with great amplitude within the range of 1.4 mm interval. The tube produced severe deformation under high-energy impact. Furthermore, the
increased temperature also increased the activity of the atoms
among some elements, thus producing a surface mass ow under
a strong impact and granular intermediates. Elements were diffused among the transition regions of welding.

Fig. 18. Granular intermediates.

4. Conclusions
The current study investigated impact velocities under different
voltages during magnetic pulse welding through simulations and
testing, compared the strengths of welded joints under different
impact velocities, and analyzed the appearance of the magnetic
pulse welding interface. Consequently, the following conclusions
were drawn:

Fig. 19. Energy spectrum analysis of particle points.

Therefore, an impact velocity of over 350 m/s was found to be the

essential condition for forming a wave-shaped interface during
magnetic impulse welding. In explosive welding, the explosion
supplies the energy for the impact of the two metals. Similar energies can be obtained from different positions of the yer plate as
the explosives were distributed on the yer plate evenly, consequently generating the same impact velocity on various points during the explosive welding. However, in the magnetic pulse welding
of the tube, the energies obtained by the outer tube are generated
by the stored energy of the capacitor in the electromagnetic former. Therefore, from the perspective of energy, the magnetic pulse
welding achieved less energy than the explosive welding. Furthermore, in magnetic pulse welding, given the absence of restriction
from the outer metal at the end of the outer tube (speed measuring
position in the test), it possesses a smaller deformation resistance
and moves faster than the other parts. The deformation of the outer tube is restricted by the undeformed tube when the distance to
the initial end reaches a certain extent, consequently reducing the
impact velocity and producing particles in the interface.
The granular clustering area was selected for energy spectrum
scanning. As seen in Fig. 19, the energy spectrum analysis was carried out for three points, discovering the existence of Fe and Al elements in the corresponding areas. The proportions of these two
elements are shown in Table 5, which reveals that almost all

(1) The magnetic eld force was mainly distributed in the center
of the tube (the connection zone), where the deformation
rst occurred because of the small deformation resistance
when the end of the outer tube was affected by the axial
magnetic eld force.
(2) The deformation process was accomplished within a microsecond, and the impact velocity of the inner and outer tubes
increased with the increase in voltage. Under such experimental conditions, the impact velocity of the inner and outer
tubes could reach over 350 m/s when the discharge voltage
was increased to 14 kV. The result of the simulation is found
to be in good accordance with the experimental result,
except that the former was slightly greater than the latter
because of the impact of air.
(3) When the impact velocity of the inner and outer tubes is larger than 200 m/s, the joint strength of the aluminum steel
was larger than the aluminum strength. When the impact
velocity reaches 355 m/s, a relatively regular zigzag waveshaped interface with a wave length of about 50 lm and a
wave height of about 10 lm was formed in the joint transition region.

References
[1] Haraga K. Strength properties of aluminum/aluminum and aluminum/steel
joints for lightweighing of automotive body. Weld World 2000;44(4):237.
[2] Larsson JK. Overview of joining technologies in the automotive industry. Weld
Res Abroad 2003;49(67):2945.
[3] Lewis MD. Joining of aluminum by methods other than fusion welding.
Aluminium Ind 1992;11(4):169.
[4] Findik F. Recent developments in explosive welding. Mater Des
2011;32(3):108193.
[5] Fukumoto S. Friction welding process of 5052 aluminium alloy to 304 stainless
steel. Mater Sci Technol 1999;15(9):10806.
[6] Roulin M, Luster JW, Karadeniz G. Strength and structure of furnace-brazed
joints between aluminum and stainless steel. Weld J 1999;78(5):1515.
[7] Schubert E. Light-weight structures produced by laser beam joining for future
applications in automobile and aerospace industry. J Mater Process Technol
2001;115(1):28.
[8] Xu ZD, Yu HP, Li CF. Interface microstructure of AlFe tubes joint by magnetic
pulse welding. J Iron Steel Res Int 2012;19(1):4425.

Z. Xu et al. / Materials and Design 49 (2013) 736745

[9] Zhang P, Kimchi M, Shao H, Gould JE, Daehn GS. Analysis of the
electromagnetic impulse joining process with a eld concentrator. Numiform
2004:12538.
[10] Hokari H, Sato T, Kawauchi K, Muto A. Magnetic impulse welding of aluminum
tube and copper tube with various core materials. Weld Int
1998;12(8):61926.
[11] Aizawa T, Kashani M, Okagawa K. Application of magnetic pulse welding for
aluminum alloys and SPCC steel sheet joints. Weld J 2007;86:11924.
[12] Ben-Artzy A, Stern A, Frage N, et al. Interface phenomena in aluminium
magnesium magnetic pulse welding. Sci Technol Weld Join 2008;13(4):4028.
[13] Kimchi M, Shao H, Cheng W, et al. Magnetic pulse welding aluminum tubes to
steel bars. Weld World 2004;48(34):1922.
[14] Shim JY, Kang BY, Kim IS, et al. A study on distributions of electromagnetic
force of the dissimilar metal joining in MPW using a FEM. Adv Mater Res
2010;8386:21421.
[15] Marya M, Marya S. Interfacial microstructures and temperatures in
aluminiumcopper electromagnetic pulse welds. Sci Technol Weld Join
2004;9(6):5417.
[16] Casalino G, Ludovico AD. Finite element simulation of high speed pulse
welding of high specic strength metal alloys. J Mater Process Technol
2008:3015.
[17] Stern A, Aizenshtein M. Magnetic pulse welding produces high-strength
aluminum welds. Sci Technol Weld Join 2002;7(5):33942.
[18] Marouani H, Ben Ismail A, Hug E, Rachik M. Numerical investigations on sheet
metal blanking with high speed deformation. Mater Des 2009;30:356671.
[19] Bahmani MA, Niayesh K, Karimi A. 3D simulation of magnetic eld distribution
in electromagnetic forming systems with eld-shaper. J Mater Process Technol
2009;209. 2295-2031.

745

[20] Viveka A, Kimb K-H, Daehn GS. Simulation and instrumentation of

electromagnetic compression of steel tubes. J Mater Process Technol
2011;211(5):84050.
[21] Lee OS, Kimb MS. Dynamic material property characterization by using split
hopkinson pressure bar (SHPB) technique. Nucl Eng Des 2003;22(6):1201.
[22] Naik NK, Venkateswara Rao Kavala, Veerraju Ch, Ravikumar G. Stressstrain
behavior of composites under high strain rate compression along thickness
direction: effect of loading condition. Mater Des 2010;31:396401.
[23] Lin YC, Chen XM, Liu G. A modied JohnsonCook model for tensile behaviors
of typical high-strength alloy steel. Mater Sci Eng 2010;527:69806.
[24] GB/T 7314-2005. Metallic materials compression testing at ambient
temperature. Standardization Administration of the Peoples Republic of
China; 2005.
[25] DIN EN ISO 10447-2007. Resistance welding-Peel and chisel testing of
resistance spot and projection welds. Austrian Standards Institute; 2007.
[26] Huseyin U. Friction stir welding of dissimilar Al 6013-T4 To X5CrNi18-10
stainless steel. Mater Des 2005;26(1):416.
[27] Acarer M, Glen B, Findik F. Investigation of explosive welding parameters
and their effects on microhardness and shear strength. Mater Des
2003;24(8):65964.
[28] Sayed ME. Structure and properties of carbon steelaluminium dissimilar
joints. Sci Technol Weld Join 2005;10(1):2731.
[29] Acarer M, Gulenc B, Findik F. The inuence of some factors on steel/steel
bonding quality on their characteristics of explosive welding joints. J Mater Sci
2004;39(21):645766.
[30] Zamani E, Liaghat GH. Explosive welding of stainless steelcarbon steel coaxial
pipes. J Mater Sci 2012;47:68595.