Symmetrical Transition Waveform Control on Double-Wire MIG Welding

Abstract

To reduce unsteady and undercut phenomena during double-wire MIG welding process, a symmetrical transition waveform control method, which was based on analyzing the effect of the electromagnetic force, was proposed. A symmetrical transition period was added when the welding current was switched between peak value and base value. Three variables were introduced, which were leading transition current ISL, double wires transition time Ts and trailing transition current IST. Corresponding experiments were conducted to compare the different phenomena between employing this period and not employing the period. Orthogonal experiments were conducted to study the influence by the three variables. An optimum combination of ISL, Ts, IST could be obtained by means of range analysis, and corresponding experiments verified the combination. All the experiments showed that the symmetrical transition period can reduce the large variation of welding current, make the welding process steadier and the welding seam much more polished, and obtain welding seam with higher quality.

Full text

Journal
of
Materials
Processing
Technology
229
(2016)
111–120
Contents lists available at ScienceDirect
Journal
of
Materials
Processing
Technology
journal homepage: www.elsevier.com/locate/jmatprotec
Symmetrical
transition
waveform
control
on
double-wire
MIG
welding
Yao
Pinga,
Xue
Jiaxiangb,
Zhou
Kangc,∗,
Wang
Xiaojuna,
Zhu
Qiangb
aCollege
of
Electromechanical
Engineering,
Guangdong
Polytechnic
Normal
University,
Guangzhou
510635,
China
bSchool
of
Mechanical
and
Automotive
Engineering,
South
China
University
of
Technology,
Guangzhou
510640,
China
cState
Key
Laboratory
of
High-temperature
Gas
Dynamics,
Institute
of
Mechanics,
Chinese
Academy
of
Sciences,
Beijing
100190,
China
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
6
May
2015
Received
in
revised
form
27
August
2015
Accepted
29
August
2015
Available
online
11
September
2015
Keywords:
Double-wire
MIG
welding
Symmetrical
transition
Current
waveform
control
Orthogonal
experiment
a
b
s
t
r
a
c
t
To
reduce
unsteady
and
undercut
phenomena
during
double-wire
MIG
welding
process,
a
symmetrical
transition
waveform
control
method,
which
was
based
on
analyzing
the
effect
of
the
electromagnetic
force,
was
proposed.
A
symmetrical
transition
period
was
added
when
the
welding
current
was
switched
between
peak
value
and
base
value.
Three
variables
were
introduced,
which
were
leading
transition
current
ISL,
double
wires
transition
time
Tsand
trailing
transition
current
IST.
Corresponding
experiments
were
conducted
to
compare
the
different
phenomena
between
employing
this
period
and
not
employing
the
period.
Orthogonal
experiments
were
conducted
to
study
the
influence
by
the
three
variables.
An
optimum
combination
of
ISL,
Ts,
IST could
be
obtained
by
means
of
range
analysis,
and
corresponding
experiments
verified
the
combination.
All
the
experiments
showed
that
the
symmetrical
transition
period
can
reduce
the
large
variation
of
welding
current,
make
the
welding
process
steadier
and
the
welding
seam
much
more
polished,
and
obtain
welding
seam
with
higher
quality.
©
2015
Elsevier
B.V.
All
rights
reserved.
1.
Introduction
Double-wire
MIG
welding
is
an
efficient
welding
method
that
has
been
increasingly
employed
in
the
welding
industry.
In
the
1950s,
Ashton
(1954)
and
Steinert
(1954),
separately
proposed
the
systematic
structure
of
double-wire
MIG
welding,
which
laid
foun-
dations
for
current
double-wire
MIG
power
source
application.
Person
and
Ruzek
(1956)
improved
the
double-wire
MIG
welding
system,
and
proposed
a
tandem
system
of
arc
welding.
Because
two
wires
simultaneously
provide
heat
energy
to
the
base
metal,
compared
to
single-wire
welding
technology,
tandem
welding
can
change
the
heat
distribution,
and
effectively
avoid
undercut
and
has
various
other
relative
disadvantages.
Hence,
tandem
welding
can
improve
the
travel
speed
for
automatic
welding
and
yield
welding
seams
with
high
quality
(Li
and
Zhang,
2010).
As
a
highly
productive
welding
technique,
double-wire
MIG
welding
appeals
a
lot
of
focuses.
Reis
et
al.
(2015)
focused
on
the
effect
of
different
elements
on
arc
interruption
in
double-wire
MIG
welding,
and
concluded
that
the
proper
selections
of
certain
param-
eters,
such
as
a
little
delay
between
trailing
current
and
leading
∗Corresponding
author.
+86
10
82545985.
E-mail
addresses:
ypsunny@163.com
(P.
Yao),
zhoukang326@126.com,
zhoukang@imech.ac.cn
(K.
Zhou).
current,
can
lower
the
possibility
of
the
appearance
of
arc
inter-
ruption.
Niu
et
al.
(2010)
focused
on
the
effect
of
the
peak
pulse
voltage
of
the
leading
current
on
the
welding
process
and
weld
quality.
Shi
et
al.
(2014a,b)
established
a
model
to
simulate
the
double-wire
MIG
welding
process,
and
respectively
obtained
the
effects
of
arc
length
and
welding
current
on
the
process.
Wen
et
al.
(2010)
analyzed
the
different
effects
of
synchronous
versus
alter-
native
changes
in
host
and
accessory
pulse
currents.
Because
the
double-wire
MIG
welding
process
involves
many
different
control
variables
and
complex
techniques,
current
researches
have
mainly
focused
on
a
case
in
which
the
leading
current
and
trailing
current
have
the
same
frequency
but
inverse
phase
(SFIP).
Many
previous
studies
have
concluded
that
under
the
SFIP
current,
the
noise
of
the
output
electrical
arc
is
low
and
the
welding
process
is
steady;
how-
ever,
the
welding
penetration
is
relatively
small(Andersson
et
al.,
2006;
Bagus
Yudharto
Bharotokusumo
et
al.,
2008).
Conversely,
Ueyama
et
al.
(2005)
conducted
an
experiment
with
the
leading
current
and
trailing
current
having
the
same
frequency
and
same
phase
(SFSP),
and
then
they
found
that
when
the
peak
period
of
the
trailing
current
was
delayed
by
0.5
ms
more
than
that
of
the
leading
current,
a
steady
welding
process
together
with
a
satis-
factory
welding
seam
could
be
obtained.
Moreover,
by
delaying
the
pulse
end
timing
of
the
trailing
arc
by
0.4–0.5
ms
from
that
of
the
leading
arc,
Ueyama
et
al.
(2009)
work’s
effectively
reduced
the
arc
interference
in
tandem
pulsed
gas
metal
arc
welding.
http://dx.doi.org/10.1016/j.jmatprotec.2015.08.031
0924-0136/©
2015
Elsevier
B.V.
All
rights
reserved.

112
P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
Additionally,
Motta
et
al.
(2007)
showed
that
an
inverse
phase
waveform
is
not
a
necessary
condition
for
obtaining
satisfactory
welding
seams.
Wang
et
al.
(2011)
discovered
that
deceasing
the
ratio
between
the
leading
current
and
the
trailing
current
could
reduce
the
possibility
of
arc
interruption.
Previous
studies
have
examined
different
aspects
of
double-
wire
MIG
welding
using
various
methods,
and
have
drawn
some
interesting
conclusions.
However,
the
current
waveform
of
double-
wire
MIG
welding
that
is
typically
employed
is
a
traditional,
rectangular
type.
Because
the
difference
between
the
base
value
and
the
peak
value
of
the
rectangular
signal
is
very
large,
unsteady
welding
process
may
be
induced,
which
can
make
the
variation
in
energy
during
the
different
welding
stages
difficult
to
prop-
erly
control.
As
the
processor
capacity
and
welding
machine
digital
technologies
have
been
improved
in
past
decade,
the
new
current
waveform
control
method
was
more
and
more
considered,
so
that
the
molten
droplet
transferring
can
be
well
controlled,
as
well
as
the
welding
products
with
high
quality
can
be
correspondingly
obtained.
Lincoln
Electric
Co.
developed
a
surface
tension
trans-
fer
(STT)
process,
which
accurately
controlled
the
current
in
seven
points
of
each
waveform
cycle
period.
This
method
substantially
reduced
spatter
and
lowered
heat
delivery
on
thin-gauge
material
(DeRuntz,
2003).
Miller
Electric
Mfg.
Co.
introduced
a
new
weld-
ing
technology
process
called
Regulated
metal
deposition
(RMD)
process.
The
technology
was
based
on
an
advanced
software
appli-
cation
for
modified
short
circuit
transfer
GMAW
(MIG
welding)
that
monitored
the
electrode
current
in
each
step
of
the
short
cir-
cuiting.
The
RMD
approach
was
illustrated
in
different
steps
as
follows:
Wet,
Pinch,
Clear,
Blink,
Ball,
Background,
Pre-short,
which
can
predict
future
arc
conditions
and
controls
the
droplet
transfer
accordingly
(Peterson,
2009).
Apart
from
welding
line
pipe
carbon
steel
materials,
the
technology
can
also
be
used
in
welding
nickel
alloy
(Petro,
2011).
Kemppi
Company
developed
a
modified
short
arc
circuit
welding
process,
which
was
called
WiseRoot
Process.
In
this
process,
the
power
source
was
monitored
by
the
wave
of
the
current,
which
can
be
analyzed
in
two
main
parts:
the
short
cir-
cuit
and
the
arc
period
(Uusitalo,
2007).
This
process
was
capable
of
achieving
similar
welds
with
5–25%
less
heat
delivery
than
the
conventional
short
arc
and
can
maintain
the
same
heat
delivery
as
a
laser
welding
process
(Peltola
et
al.,
2010).
Fronius
Company
combined
special
wave
control
features
and
an
assistance
back-
drawing
force,
and
then
developed
a
cold
metal
transfer
(CMT)
process
which
can
be
used
in
aluminium
alloy
materials
welding
(Schierl,
2005).
Also,
Feng
et
al.
(2009)
employed
the
sensing
and
image
method
to
analyze
the
wave
control
characteristics
and
its
droplet
transfer
process
in
CMT
method,
the
results
showed
that
it
can
realize
no-spatter
welding
and
low
heat
input
during
welding
process.
Daihen
Corporation
(Tong
and
Ueyama,
2011)
introduced
a
method
of
low
frequency
modulated
type
pulsed
MIG
welding
process,
which
can
improve
the
welding
quality
by
switching
the
high
frequency
and
low
frequency
pulse
in
real
time.
Yao
et
al.
(2009)
explored
the
application
of
this
control
method
on
alu-
minium
alloy
materials
welding,
and
found
that
strong/weak
pulse
peak
current
and
high-frequency
have
great
influence
on
the
sta-
bility
of
welding
process
and
spatters.
Peters
(2013)
introduced
a
power
supply
which
includes
a
waveform
type
selector
that
selects
a
desired
shape
for
an
output
waveform,
the
operator
is
given
a
choice
between
an
advanced
“crisp”
square
wave,
a
“soft”
square
wave,
a
sine
wave,
and
a
triangle
wave,
which
can
meet
the
require-
ments
in
different
welding
occasions.
Kawagoe
and
Suzuki
(2014)
developed
a
power
source
which
can
control
the
welding
current
in
real
time.
A
current
increasing
phase
and
a
current
decreasing
phase
were
added
in
the
latter
of
pulse
peak
phase
in
this
process.
The
current
increasing
can
melt
the
workpiece.
Hence,
this
method
can
obtain
a
satisfactory
welding
performance
by
low
energy
delivery.
Though
there
were
many
current
waveform
methods
proposed
to
improve
the
welding
technique,
they
mainly
focused
on
the
single-wire
MIG
application
(Kah
et
al.,
2013).
It
is
because
that
the
current
double-wire
MIG
facility
usually
employed
two
sepa-
rated
control
parts,
and
then
the
communication
between
two
parts
is
necessary
to
conduct
online
corporative
control.
However,
real
time
control
must
be
guaranteed
in
welding
process
control,
which
needs
high-speed
Central
Processing
Unit
(CPU).
Also,
achieving
waveform
coordinated
control
for
the
two
parts
were
very
complex.
Hence,
there
were
less
relative
research
about
double-wire
MIG
current
waveform
control
process
in
reality.
Currently,
the
process-
ing
capacity
of
CPU
has
been
largely
increased,
whose
capacity
can
be
capable
of
processing
control
actions
in
two
circuits.
Under
this
circumstance,
the
current
waveform
control
process
is
proposed
in
this
paper
so
as
to
improve
the
performance
of
double-wire
MIG
technique
and
the
steady
of
the
process.
In
this
work,
according
to
the
current
double-wire
tandem
welding
technology,
a
different
current
waveform
control
pro-
cess,
which
was
symmetrical
transition
waveform
control
method,
was
proposed
to
improve
the
double-wire
MIG
welding
qual-
ity
and
operation.
A
significant
characteristic
of
this
method
was
that
a
symmetrical
transition
period
was
newly
introduced,
which
induced
that
the
magnitude
of
energy
variation
was
reduced
during
the
switching
process
between
peak
value
and
base
value
of
the
cur-
rent.
After
fine
tuning
the
relative
parameters
in
this
period,
much
steadier
welding
process
can
be
obtained.
This
method
was
fur-
ther
studied
step
by
step
by
means
of
a
series
of
experiments,
such
as
experiments
which
employed
or
not
employed
the
symmetri-
cal
transition
period,
orthogonal
experiments
to
seek
the
affecting
discipline
of
parameters
in
the
transition
period,
as
well
as
verifi-
cation
experiment
of
optimum
parameters
combinations
obtained
by
range
analysis.
Final
results
showed
that
a
much
steadier
weld-
ing
process
and
much
more
polished
welding
seam
can
be
obtained
using
this
proposed
method.
2.
Principle
of
double-wire
MIG
welding
The
double-wire
MIG
welding
technique
involves
placing
two
wires
on
a
specially
designed
welding
torch
in
a
predetermined
pattern;
the
two
wires
are
fed
by
individual
power
sources
and
insulated
from
each
other.
Almost
all
of
parameters
of
the
two
wires
including
the
wire
feed
speeds
are
independent
from
each
other
in
reality.
Both
of
the
diameters
and
materials
of
the
wires
may
be
different.
For
example,
one
wire
can
be
fed
by
an
impulse
signal,
and
a
different
signal
type
can
feed
the
other
wire.
During
the
process,
the
electrical
arc
can
be
generated
in
the
terminates
of
the
wires,
and
then
the
heat
generated
by
electrical
arc
discharg-
ing
can
melt
the
wire
and
base
metal
together,
finally,
a
weld
seam
will
formed
after
the
heat
removed.
Hence,
to
obtain
satisfactory
performance,
the
electrical
arc
should
be
well-controlled
using
this
design
and
some
satisfactory
performances,
such
that
the
mutual
interferences
of
the
two
wires
should
be
low
(Goecke
et
al.,
2011),
can
be
correspondingly
obtained.
Fig.
1
shows
a
schematic
diagram
of
double-wire
MIG
welding.
In
double-wire
MIG
welding,
the
distance
between
the
two
direct
current
(DC)
electrical
arcs
is
very
small.
Under
this
cir-
cumstance,
the
magnetic
interference
between
the
electrical
arcs
may
induce
an
unsteady
welding
process.
Apart
from
the
concern
regarding
the
metal
transfer
process
of
each
wire,
the
mutual
effects
from
each
wire
should
also
be
seriously
considered.
Hence,
con-
trolling
of
the
current
waveform
of
each
wire
is
very
important
for
guaranteeing
the
efficiency
and
steadiness
of
the
double-wire
MIG
welding
process.
In
this
system,
the
two
wires
are
independent
from
each
other
and
metal
transfer
occurs
in
the
same
molten
pool.
However,

P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
113
Fig.
1.
Schematic
figure
of
double-wire
MIG
welding.
because
the
arc
is
both
a
heat
source
and
a
force
source,
the
mutual
interference
between
the
two
wires
may
affect
the
welding
pro-
cess,
which
induce
more
metal
transfer
characteristics
than
that
in
single-wire
MIG
welding.
To
address
this
issue,
Halmoy
(1980),
Chandel
(1987,
1988)
and
Suban
and
Tuˇ
sek
(2001)
studied
the
melting
speed
and
believed
that
melting
was
induced
by
both
arc
heat
and
resistance
heat.
Tuˇ
sek
(2000)
established
an
equation
for
calculating
the
melting
rate:
M
=
a0+
a1I
+
a2
I2L
d2(1)
where
a0,
a1,
and
a2are
constants.
The
units
of
the
parameters
in
Eq.
(1)
can
be
expressed
in
the
following
units:
M
(kg/h),
I
(A
per
wire),
L
(mm),
and
d
(mm).
It
can
be
observed
that
the
melting
rate
M
is
determined
by
the
welding
current
intensity
I,
wire
exten-
sion
length
L
and
wire
diameter
d.
When
other
conditions
remain
unchanged,
a
high
welding
current
indicates
a
high
melting
rate
and
a
high
welding
efficiency.
The
metal
transfer
process
is
affected
by
surface
tension,
the
electromagnetic
shrinkage
force,
the
plasma
fluid
force
and
grav-
ity.
Because
the
distance
between
the
two
wires
in
double-wire
MIG
welding
is
very
small,
a
mutually
attracting
electromagnetic
force
can
be
induced
between
the
two
wires,
which
may
induce
the
magnetic
arc
blow
phenomenon.
This
phenomenon
causes
the
arcs
to
present
an
oblique,
conical
shape
or
bundle,
and
the
arc
stiffness
becomes
poor.
The
electromagnetic
force
may
affect
the
electromagnetic
shrinkage
force
and
the
plasma
fluid
force,
which
may
change
the
arc
during
welding.
If
the
change
is
significantly
large,
arc
interruptions,
irregular
metal
transfer
or
considerable
expulsion
may
be
induced.
The
magnitude
of
the
electromagnetic
force
is
related
to
the
current
of
the
wire,
as
well
as
the
angle
and
the
distance
between
the
two
wires.
Ueyama
et
al.
(2007)
established
an
equation
for
calculating
the
electromagnetic
force
between
two
wires.
F
=
FML =
FMT =IT
␲r2
T×0IL
2␲DE(2)
where
FMT and
FML are
the
electromagnetic
force
(Lorentz
force),
IL
is
the
leading
current,
ITis
the
trailing
current,
0is
the
permeabil-
ity
of
free
space,
DEis
the
distance
between
the
two
wires,
and
rTis
the
radius
of
the
trailing
arc
column.
Eq.
(2)
indicates
that
the
prod-
uct
of
the
currents
in
the
two
wires
determines
the
electromagnetic
force
when
the
other
physical
conditions
remain
unchanged.
Hence,
to
reduce
the
mutual
interference
of
the
two
wires,
the
method
of
decreasing
the
magnitude
of
the
welding
current
can
be
employed.
However,
decreasing
the
welding
current
can
also
reduce
the
amount
of
energy
delivered,
the
melting
rate,
and
the
welding
efficiency,
which
may
cause
the
double-wire
MIG
welding
to
lose
its
merit
as
a
highly
efficient
process.
In
this
case,
the
method
of
waveform
control
is
proposed
to
improve
the
stability
of
the
double-wire
MIG
welding
process.
3.
Proposed
current
waveform
control
method
In
general
case,
three
control
strategies
are
employed
in
double-
wire
MIG
welding,
as
shown
in
Fig.
2(a)–(c).
The
number
of
control
variables
is
very
large
because
of
the
com-
plexity
of
the
current
waveform
for
double-wire
pulse
welding.
The
key
control
variables
are
shown
in
Fig.
2(a):
the
leading
peak
cur-
rent
ILp,
leading
base
current
ILb,
leading
peak
time
TLp,
leading
base
time
TLb,
trailing
peak
current
ITp,
trailing
base
current
ITb,
trailing
peak
time
TTp,
trailing
base
time
TTb,
leading
current
frequency
fL
and
trailing
current
frequency
fT.
All
variables
are
independently
adjusted
and
controlled.
Welding
Time(/ms)
tnerruC(/A)
Lea
ding
current
Trailing curren
t
T
S
I
SL
I
ST
f
L
f
T
f
L
=f
T
Fig.
3.
Symmetrical
transition
double-wire
waveform.
Lea
ding current
Lea
ding current
Trailing current
Trailing current
Welding Time(/ms)
Welding Time(/ms)
ab
T
Lp
T
Lb
T
Tb
T
Tp
I
Lb
I
Lp
I
Tb
I
Tp
f
L
f
L
f
T
f
L
=f
T
f
L
=f
T
f
T
Lea
ding
current
Trailing curr
en
t
Welding Time(/ms)
c
f
L
f
T
f
L
≠
f
T
tnerruC(/A)
Current(/A)
Current(/A)
Fig.
2.
(a)
Current
waveform
(same
frequency
and
inverse
phase
pulse);
(b)
current
waveform
(same
frequency
and
same
phase
pulse);
(c)
current
waveform
(different
frequency
and
different
phase
pulse).

114
P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
Fig.
4.
Flowchart
of
the
experimental
procedures.
Fig.
2(b)
shows
the
current
waveform
under
the
same
frequency
and
same
phase.
Under
this
circumstance,
the
electromagnetic
force
is
large
because
the
phases
of
the
two
wires
are
identical,
which
induces
significant
mutual
interference:
ˇ
=1
␲r2
T×0
2DE(3)
F1=
ˇ
×
(aLILav ×
aTITav)(aT,
aL>
1)
(4)
F1=
ˇaLaT×
(ILav ×
ITav)
(5)
ILp =
aLILav (6)
ITp =
aTITav (7)
where
ILav is
the
mean
value
of
leading
current,
ITav is
the
mean
value
of
trailing
current;
aLand
aTare
the
coefficients
of
the
peaks
of
the
leading
current
and
trailing
current,
respectively.
As
shown
in
Fig.
2(c),
when
the
current
waveform
is
under
a
ran-
dom
mode
involving
different
frequencies
and
different
phases,
the
mutual
interference
may
also
be
very
large
because
the
phases
of
the
two
wires
may
be
identical
during
the
welding
process.
More-
over,
the
welding
process
is
difficult
to
control
because
the
phases
of
the
two
wires
are
random,
which
provides
no
rules
for
metal
transfer.
Under
these
conditions,
the
largest
electromagnetic
force
is
identical
to
that
in
Fig.
2(b).
On
the
other
hand,
as
shown
in
Fig.
2(a),
when
the
current
wave-
form
exhibits
the
same
frequency
with
a
180◦phase
difference,
meaning
that
there
are
regular
and
alternative
phases
of
the
two
arc
effects,
the
effect
of
the
force
on
the
metal
transfer
is
very
low
and
the
mutual
inference
can
be
solved
for.
However,
reducing
the
current
from
the
large
peak
value
to
the
small
base
value
may
easily
induce
an
unsteady
welding
process
or
arc
interruption.
Hence,
it
can
be
concluded
that
during
the
double-wire
MIG
welding
pro-
cess,
difference
of
phases
and
values
between
the
current
peak
and
current
base
can
directly
affect
the
steadiness
of
the
welding
process.
Under
the
same
melting
rate,
decreasing
the
mutual
force
between
two
wires
benefits
the
arc
stability,
and
reduces
the
amount
of
heat
delivered.
Hence,
the
mutual
interference
between
two
wires
can
be
reduced
by
the
pulse
welding
method:
F2=
ˇaLbT×
(ILav ×
ITav)(bT<
1) (8)
where
bTis
the
scaling
factor
between
the
base
value
and
the
average
value
of
the
trailing
current.
Clearly,
when
the
phases
of
two
wires
are
inverted,
the
electromagnetic
force
F2is
lower
than
that
observed
when
the
phases
are
the
same,
F1in
Eqs.
(4),
(5)
and
(8).According
to
this
analysis,
when
the
current
values
are
not
very
large,
a
satisfactory
welding
seam
can
be
obtained
under
the
waveform
with
the
same
frequency
and
the
same
phase,
and
the
possibility
of
the
appearance
of
arc
interruption
can
also
be
simulta-
neously
reduced.
However,
because
the
current
value
is
insufficient
for
supplying
enough
energy,
the
welding
efficiency
is
low.
To
reduce
the
effect
of
the
electromagnetic
force
and
improve
the
welding
efficiency,
the
symmetrical
transition
waveform
control
method
is
proposed
as
shown
in
Fig.
3.
It
can
be
shown
in
Fig.
3
that
in
the
proposed
method,
a
transition
period
is
introduced
when
the
current
phase
is
changed.
Hence,
the
entire
pulse
waveform
can
be
described
by
three
components:
peak
period,
transition
period
and
based
period.
Correspondingly,
three
variables
are
added:
the
leading
tran-
sition
current
ISL,
double
wires
transition
time
Tsand
trailing
transition
current
IST.
These
variables
can
directly
determine
the
effects
of
the
energy
delivery
and
electromagnetic
force
on
metal
transfer
during
the
transition
period.
The
electromagnetic
force
in
the
symmetrical
transition
period
can
be
calculated
as
follows:
FS=
ˇsLsT×
(ILav ×
ITav)
(9)
ISL =sL×
ILav (10)
IST =sT×
ITav (11)
where
sLis
the
leading
current
transition
coefficient,
and
sTis
the
trailing
current
transition
coefficient.
sT,
sLmay
be
greater
or
smaller
than
1.
When
sL=
sT=
1,
ISL =
ILav,
IST =ITav,
and
the
sym-
metrical
transition
period
is
added,
according
to
Eq.
(1),
I
remains
unchanged,
and
then
the
mean
value
of
the
current
and
melting
rate
remain
unchanged;
however,
because
aL×
bTmay
be
greater
than
sL×
sT,
the
value
of
electromagnetic
force
in
the
symmetrical
transition
period,
which
is
marked
as
Fs,
may
be
lower
than
that
observed
when
the
phases
of
the
two
arcs
are
inverted.
By
adjusting
these
variables
and
the
relationship
between
sLand
sT,
the
electromagnetic
force
under
the
same
phase
mode
of
the
two
wires
can
approach
or
even
be
less
than
that
under
the
inverse
phase
mode.
Hence,
according
to
above
analysis,
the
existence
of
a
symmetrical
transition
period
can
make
the
welding
process
much
steadier
and
controllable,
and
then
a
more
satisfactory
welding
quality
can
be
obtained
(Yao
et
al.,
2012).
4.
Experimental
setup
and
method
4.1.
Experimental
setup
To
verify
the
proposed
method,
actual
welding
experiments
were
conducted.
In
the
experiment,
a
DSP
integrated
double-wire
MIG
power
source,
a
walking
controller,
a
welding
experiment
platform,
a
welding
arc
dynamic
wavelet
analysis
instrument,
a
double-wire
pulse
MIG
welding
inverter
with
a
software
switch,
a
wire
feeder,
a
double-wire
welding
torch,
a
tank
and
other
rela-
tive
facilities
were
employed.
During
the
experimental
process,
the
welding
arc
dynamic
wavelet
analysis
instrument
was
employed
to
collect
and
analyze
the
waveform,
while
the
integrated
double-

P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
115
Table
1
Control
parameters
used
in
the
current
waveform.
Parameter
340
88
300
75
5.2
12.5
56
Item
ILp (A)
ILb (A)
ITp (A)
ITb (A)
Tp(ms)
Tb(ms)
fL,
fT(Hz)
wire
pulse
MIG
welding
inverter
was
used
to
control
the
actual
waveform.
The
basic
conditions
used
in
the
double-wire
symmetrical
tran-
sition
current
waveform
control
experiment
were
as
follows:
the
workpiece
was
Q235
steel
with
a
thickness
of
8.0
mm.
The
weld-
ing
wires
were
composed
of
H08Mn2SiA
and
measured
1.0
mm
in
diameter,
and
the
shielding
gas
was
pure
argon
(15
L/min
flow).
The
distance
between
the
two
terminals
of
the
wires
was
8.0
mm.
In
addition,
flat
welding
was
used.
The
control
parameters
used
in
the
current
waveform
are
shown
in
Table
1.
4.2.
Experimental
program
To
comprehensively
explore
the
operational
performance
of
symmetrical
transition
waveform
control
method,
three
experi-
ments
were
employed
as
follows:
1)
The
first
experiment
was
to
compare
the
different
phenom-
ena
that
occurred
when
the
symmetrical
transition
period
was
employed
and
not
employed,
the
goal
of
the
experiment
was
to
observe
the
effect
of
the
symmetrical
transition
period
on
the
process
and
performance.
2)
The
second
experiment
was
the
orthogonal
experiment
employ-
ing
three
parameters
with
three
different
values
under
three
different
levels.
Then
the
range
analysis
was
employed
to
further
study
the
effect
of
the
parameters
in
the
symmetrical
transition
period
on
the
welding
quality.
3)
Different
parameters
combinations
were
employed
in
the
exper-
iments
so
as
to
verify
the
effectiveness
of
the
range
analysis
and
obtain
an
optimum
parameters
combination.
To
clearly
show
the
procedures,
Fig.
4presents
a
flowchart
of
the
above
experiments.
5.
Experiments
and
analysis
5.1.
Experiment
of
symmetrical
transition
period
First,
two
experiments,
one
employed
the
transition
period
and
the
other
did
not
employ
the
period,
were
conducted
and
com-
pared
to
explore
the
effect
of
symmetrical
transition
period
on
the
welding
process.
The
data
used
in
these
experiments
are
shown
in
Table
2.
The
average
value
of
welding
current
could
be
obtained
as
fol-
lows:
when
the
transition
period
was
not
employed,
the
average
Fig.
5.
The
current
waveform
when
the
transition
phase
is
not
employed.
Fig.
6.
Welding
seam
when
the
transition
phase
is
not
employed.
Fig.
7.
The
current
waveform
and
seam
when
the
transition
phase
is
employed.

116
P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
Table
2
Experimental
data
of
using/not
using
symmetrical
transition
period
experiment.
Item
Peak
timeTp(ms)
Base
timeTb(ms)
Transition
parameters
Walking
velocity(m/min)
No
using
transition
period
5.2
12.5
NA
0.9
Using
transition
period 5.2
12.5 TS=
3.2
ms 1.0
ISL =
130
A
IST =
220
A
values
of
leading
current
ILav and
the
trailing
current
ITav could
be
obtained
using
Eqs.
(12)
and
(13),
respectively:
ILav =ILp ×
TLp +
ILb ×
TLb
TLp +
TLb
(12)
ITav =ITp ×
TTp +
ITb ×
TTb
TTp +
TTb
(13)
when
the
transition
period
was
employed,
corresponding
average
values
of
two
currents
ISLav and
ISTav could
be
calculated
by
Eqs.
(14)
and
(15):
ISLav =ILp ×
TLp +
ILs ×
TLs +
ILb ×
TLb
TLp +
TLs +
TLb
(14)
ISTav =ITp ×
TTp +
ITs ×
TTs +
ITb ×
TTb
TTp +
TTs +
TTb
(15)
Then
according
to
Table
1
and
2
together
with
Eqs.
(12)–(15),
ILav and
ITav were
162.0
A
and
141.1
A,
when
the
transition
period
was
introduced,
the
corresponding
currents
values
ISLav and
ISTav
were
170.9
A
and
139.4
A.
By
means
of
Eqs.
(1)
and
(10)–(11),
melting
rate
was
increased,
which
induced
the
wire
feed
speed
could
be
increased,
therefore
the
welding
efficiency
was
improved.
The
observed
welding
current
waveform
and
welding
seam
when
the
transition
phase
was
not
employed
are
respectively
shown
in
Figs.
5
and
6.
As
shown
in
Fig.
5,
when
the
transition
phase
was
not
employed
and
the
current
was
directly
switched
between
the
base
current
and
peak
current,
arc
interruption
could
be
easily
induced
because
of
the
large
variation
in
current.
As
shown
in
Fig.
5,
at
14
ms,
the
current
waveform
exhibited
a
distinct
jut,
which
may
have
easily
induced
arc
interruption.
Additionally,
the
same
feature
could
be
observed
in
the
welding
seam.
Due
to
the
unsteady
welding
current,
the
shape
of
the
welding
seam
was
irregular,
and
some
sunken
areas
appeared.
After
three
sunken
areas
formed,
the
seam
showed
the
formation
of
a
grain,
which
may
have
been
induced
by
restarting
the
arc,
as
well
as
the
expulsion
marks
were
clearly
observed
in
the
workpiece.
Additionally,
the
arc
length
was
irregular,
and
an
exploding
sound
was
heard
during
the
process.
Figs.
7
and
8
show
the
current
waveform
and
corresponding
seam
when
the
transition
phase
was
employed.
It
can
be
observed
that
the
overall
welding
seam
was
regular.
The
arc
length
was
consistent,
and
the
length
could
be
increased
during
the
welding
process.
Moreover,
no
distinct
arc
interruption
occurred,
and
the
unsteady
tendency
could
be
observed.
How-
ever,
it
could
be
immediately
adjusted
to
the
normal
state.
The
sound
heard
during
the
experiment
was
different
from
that
heard
which
occurred
during
the
ordinary
double-wire
MIG
welding
pro-
cess,
which
was
“woon.”.
The
sound
head
during
the
experiment
was
also
much
gentler,
range
between
“woon.”
and
“zi.”.
Further-
more,
the
shape
of
the
seam
was
regular,
and
the
surface
gloss
is
much
better
than
that
observed
when
the
transition
phase
was
not
employed.
The
crest
of
the
seam
was
also
different
from
that
in
an
ordinary
double-wire
welding
seam;
the
comparison
is
shown
in
Fig.
9.
In
an
ordinary
double-wire
MIG
welding
seam,
the
crests
appeared
as
fish-scale
grains
with
only
a
few
defects,
as
shown
in
the
left
panel
of
Fig.
9.
Conversely,
there
was
an
obvious
transition
zone
in
the
seam
obtained
by
the
symmetrical
transition
current
waveform
control
method.
This
zone
made
the
seam
crest
surface
gloss
much
better
than
that
of
the
ordinary
seam.
Additionally,
the
fish
scale
grain
below
the
transition
zone
was
much
more
regular
and
uniform,
as
shown
in
the
right
panel
of
Fig.
9.
According
to
the
corresponding
analysis,
these
results
occurred
during
the
transition
period
because
a
proper
combination
of
leading
and
trailing
transition
currents
filled
the
gap
when
the
lead-
ing
current
and
trailing
current
were
switched.
Hence,
the
flute
between
the
fish-scale
grains
was
filled
with
melting
filler
dur-
ing
the
transition
period,
such
that
more
better
formed
welding
seam
surface
crest
was
obtained.
It
was
because
melting
rate
was
increased
by
introducing
the
transition
period,
so
that
more
liquid
metal
can
fill
the
gap
between
fish-scale
grains.
5.2.
Experiments
of
parameters
affecting
discipline
5.2.1.
Orthogonal
experiments
Orthogonal
experimental
design
is
an
important
designing
method
to
study
the
system
effects
that
involves
multiple
fac-
tors
and
multiple
levels.
In
this
method,
certain
parameters
are
chosen
as
different
factors
for
conducting
experiments
based
on
orthogonality.
This
selection
can
reduce
the
workload
and
involves
corresponding
methods
of
analyzing
the
experimental
results,
so
as
to
yield
more
reliable
conclusions
(Keppel,
1991;
Taguchi
et
al.,
1987).
To
further
explore
the
effect
of
the
transition
period
on
double-
wire
MIG
welding
quality
and
obtain
an
optimum
operational
parameters
combination,
orthogonal
experiments
were
conducted
in
this
work.
In
this
experiment,
the
effects
of
the
variation
of
three
impor-
tant
parameters
on
welding
quality
were
respectively
explored;
the
parameters
were
leading
transition
current
ISL,
trailing
transi-
tion
current
IST and
current
transition
times
Ts.
First,
the
variation
ranges
and
levels
of
these
three
parameters
could
be
confirmed
using
extreme
experiments,
and
then
an
experimental
scheme
was
designed
using
an
orthogonal
experiment
table
L9(34).
Table
3
shows
the
parameters
values
employed
corresponding
three
levels
in
the
experiments.
Three
repeated
experiments
were
conducted.
The
waveform
was
collected
and
analyzed
using
the
welding
arc
dynamic
wavelet
analysis
instrument.
The
corresponding
welding
elec-
trical
signal
was
analyzed
via
the
transient
current
waveform,
the
voltage–current
probability
density
distribution
and
corre-
sponding
U–I
graph,
the
transient
energy
graph
and
other
related
methods.
Additionally,
the
welding
quality
obtained
from
differ-
ent
combinations
of
key
parameters
could
be
assessed
based
on
the
observed
phenomena
and
corresponding
welding
seam
shapes.
The
welding
quality
could
be
scored
using
a
previously
reported
quality
scoring
method
(Yao
et
al.,
2014a,b)
and
the
data
obtained
from
three
repeated
experiments
using
the
same
parameters.
The
detailed
scheme
and
welding
quality
scores
are
shown
in
Table
4.
In
the
nine
experiments
described
above,
the
mean
current
val-
ues
of
the
two
wires
were:
ISLav =
154–180
A
and
ISTav =
131–151
A.
It
was
observed
that
welding
could
be
steadily
conducted
with
an
unchanging
walking
velocity,
with
the
other
variables
held
at
their
standard
values.
No
arc
interruption
or
inability
to
ini-
tiate
arc
phenomena
was
observed;
hence,
the
majority
of
the

P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
117
Fig.
8.
Welding
seam
when
the
transition
phase
is
employed.
Fig.
9.
Comparison
between
ordinary
seam
and
seam
obtained
from
symmetrical
transition
current
waveform
control
method.
Table
3
Parameters
and
levels.
Levels
Leading
transition
current
(ISL/A)
Transition
time
(TS/A)
Trailing
transition
current
(IST/A)
1
130
0.9
120
2
200
2
170
3
250
3.2
220
Table
4
Orthogonal
test
program
and
results.
Index
Parameters
Leading
transition
current
(ISL/A)
Transition
time
(TS/A)
Trailing
transition
current
(IST/A)
Welding
quality
1
130
0.9
120
79
2
130
2
170
68
3
130
3.2
220
76
4
200
0.9
170
74
5
200
2
220
81
6
200
3.2
120
76
7
250
0.9
220
74
8
250
2
120
85
9
250
3.2
170
88
Table
5
Results
of
the
range
analysis.
Index
1
(ISL)
2
(TS)
3
(IST)
Ij/kj74.3333
75.6667
80
IIj/kj77
78
76.6667
IIIj/kj82.3333
80
77
Dj80
4.3333
3.3333
experiments
could
yield
the
products
with
satisfactory
quality.
Therefore,
the
stable
operating
point
range
was
so
large
that
the
parameters
could
be
easily
matched.
As
shown
in
Table
4,
the
lowest
welding
quality
was
obtained
from
experiment
2,
whereas
the
highest
welding
quality
was
obtained
from
experiment
9.
The
current
waveforms
and
welding
seams
obtained
from
these
experiments
are
shown
in
Figs.
10–13
.
First,
the
stability
of
the
welding
process
could
be
analyzed
using
current
waveform.
According
to
Fig.
10,
which
showed
the
cur-
rent
waveform
graph
of
experiment
2,
this
experiment
showed
no
distinct
arc
interruption
or
short
circuit.
Additionally,
the
weld-
ing
process
was
unsteady
and
juts
appeared.
Moreover,
explosion
could
occur
during
the
process
and
some
expulsions
may
occur.
In
Fig.
11
which
shows
the
welding
seam,
although
the
start-
ing
arc
process
was
smooth,
the
welding
seam
was
irregular,
and
undercut
and
arc
interruption
could
occur.
However,
due
to
the
existence
of
a
transition
period,
the
process
could
independently
retrieve
a
normal
state.
The
appearance
of
the
welding
seam
was
regular.
Moreover,
a
transition
zone
existed
in
the
welding
seam,
but
the
zone
was
interrupted
when
the
tendency
or
arc
interruption
appeared.
On
the
other
hand,
the
current
waveform
of
experiment
9
shown
in
Fig.
12
was
regular
and
steady,
with
distinct
transition
period.
The
corresponding
welding
seam
shown
in
Fig.
13
was
regular
with-
out
undercut
and
showed
a
distinct
tendency
for
arc
interruption.
The
welding
seam
was
smooth,
and
the
transition
zone
observed
in
the
welding
seam
crest
was
well-formed
with
silver
gloss.
5.2.2.
Range
analysis
results
In
general
case,
the
range
analysis
method
and
the
variance
method
were
used
to
analyze
the
orthogonal
experimental
process.
The
following
results
could
be
obtained
from
the
range
analysis:
(1)
The
effects
of
each
parameter
on
the
experimental
results.
(2)
The
variation
tendency
of
the
experimental
results
for
each
parameter.
To
directly
observe
the
variation
tendency,
the
cal-
culated
results
could
also
be
described
in
figures.

118
P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
Fig.
10.
Current
waveform
of
experiment
2.
Fig.
11.
Welding
seam
of
experiment
2.
Fig.
12.
Current
waveform
of
experiment
9.
Fig.
13.
Welding
seam
of
experiment
9.
(3)
The
optimum
parameters
combination
that
could
achieve
the
best
performance.
In
this
work,
the
range
analysis
method
was
employed.
The
anal-
ysis
results
are
shown
in
Table
5.
In
this
table,
the
data
in
Ijand
IIj,
IIIjrepresent
the
sum
of
the
experimental
results
at
level
1,
level
2
and
level
3;
kjis
the
number
of
replicates
of
the
same
level
in
the
jth
column,
in
this
work,
kj=
3.
Ij/kj,
IIj/kj,
IIIj/kjare
the
mean
val-
ues
of
the
experimental
results
corresponding
to
each
level
in
the
jth
column;
Djis
the
range,
which
is
calculated
using
the
following
equation:
Dj=
max{Ij
kj
,IIj
kj
,
.
.
..
.
.}
−
min{Ij
kj
,IIj
kj
,
.
.
..
.
.}
(16)
According
to
Table
5,
the
effect
of
the
leading
transition
current
ISL on
welding
quality
was
the
most
significant,
then
followed
by
that
of
the
transition
time
Ts,
and
the
effect
of
IST was
relatively
the
1
2
3
70
73
76
79
82
85
Level
erocS
I
SL
T
S
I
ST
Fig.
14.
Parameters
effect
curves.

P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
119
Fig.
15.
Current
waveform
under
the
optimized
parameters.
Fig.
16.
Weld
surface
under
the
optimized
parameters.
smallest.
Fig.
14
shows
the
effects
of
these
three
parameters
on
the
welding
process.
According
to
Fig.
14,
the
welding
quality
was
much
better
as
increasing
ISL and
Ts,
while
the
welding
quality
was
the
best
when
the
value
of
IST was
smallest,
and
larger
value
of
the
IST,
worse
welding
quality
could
be
obtained.
In
addition,
when
the
cur-
rent
in
transition
period
was
not
very
large,
relative
satisfactory
welding
quality
could
be
obtained.
Because
the
variation
differ-
ence
of
current
was
lowered
by
means
of
existence
of
transition
period,
a
much
steadier
welding
process
could
be
obtained.
Hence,
increase
the
transition
period
can
improve
the
welding
quality
in
practice.
5.3.
Experiment
verification
According
to
the
criteria
of
evaluating
welding
seam
quality,
in
this
orthogonal
experiment,
high
experimental
index
scores
indi-
cate
high
welding
quality.
Hence,
the
optimal
operating
condition
is
a
combination
of
the
parameters
that
can
increase
the
welding
quality
scores
in
each
level.
As
shown
in
Table
5,
the
average
score
of
ISL in
level
3
was
highest,
the
corresponding
value
was
82.333;
the
highest
average
scores
of
Tsfor
each
level
also
appeared
in
level
3,
the
value
was
80;
the
highest
average
scores
of
IST for
each
level
corresponded
to
the
level
1.
Then
the
optimum
parameters
com-
bination
can
be
obtained
using
each
level
which
corresponded
the
highest
score
for
each
parameter.
According
to
Tables
3
and
5,
the
optimum
parameters
combination
can
be
shown
as
follows:
Leading
transition
current
ISL at
level
3:
250
A;
Transition
time
Tsat
level
3:
3.2
ms;
Trailing
transition
current
IST at
level
1:
120
A.
Thus,
a
corresponding
experiment
could
be
conducted
to
val-
idate
the
above-described
outcome.
The
current
waveform
and
welding
seam
are
shown
in
Figs.
15
and
16.
As
shown
in
Fig.
15,
the
current
waveform
was
highly
regu-
lar
with
distinct
symmetrical
transition
period,
and
the
waveform
appeared
was
very
steady
without
any
short
circuit
or
arc
inter-
ruption.
The
sound
heard
during
the
experiment
was
a
gentle
sound
ranging
between
“woon”
and
“zi.”.
According
to
Fig.
16,
the
welding
seam
shape
was
also
regular,
and
no
undercut,
arc
interruption,
jut
or
other
defects
appeared.
Additionally,
the
surface
was
much
cleaner
than
that
observed
without
a
transition
period.
5.4.
Discussion
According
to
the
experimental
results,
the
effects
of
ISL,
Tsand
IST
on
the
welding
quality
presents
some
certain
rules.
However,
the
current
researches
were
based
on
the
individual
varying
of
these
elements,
instead
of
concerning
the
effects
when
all
of
three
param-
eters
vary
at
the
same
time.
Due
to
the
complexity
of
the
welding
process,
the
effect
when
all
the
parameters
vary
on
the
welding
quality
may
have
large
differences.
Hence,
the
future
work
will
concern
this
point
and
explore
the
effect
in
depth.
In
addition,
the
workpiece
in
this
work
is
steel.
In
the
future
work,
the
optimum
combination
of
control
variables
for
current
waveform
control
in
the
transition
period
will
be
further
examined
under
workpieces
with
other
materials
and
corresponding
other
currents.
Furthermore,
the
double-wire
MIG
process
is
very
complex.
To
decrease
the
complexity
of
parameters
matching
and
improve
the
intelligent
control
level
of
the
process,
the
relations
between
differ-
ent
parameters
will
be
further
explored.
The
analytical
equations
will
be
obtained
through
further
theoretical
and
experimental
anal-
ysis,
and
then
an
intelligent
parameter
matching
library
will
be
established
to
serve
different
applying
occasions.
6.
Conclusions
According
to
analyze
the
effect
discipline
of
electromagnetic
force
on
double-wire
MIG
welding,
a
symmetrical
transition
wave-
form
control
method
was
proposed
to
improve
the
welding
quality,
relative
experiments
were
conducted,
then
some
conclusions
were
drawn
as
follows:
(1)
The
method
was
based
on
current
waveform
control.
The
exper-
imental
results
showed
that
it
could
improve
the
welding
quality,
and
make
the
welding
process
steadier,
as
well
as
more
polished
welding
seam
could
be
obtained.
(2)
The
effects
of
three
important
parameters,
which
were
ISL,
Ts
and
IST on
the
performance
were
obtained
in
detail
through

120
P.
Yao
et
al.
/
Journal
of
Materials
Processing
Technology
229
(2016)
111–120
orthogonal
experiment
and
range
analysis.
It
could
be
con-
cluded
that
the
effect
of
ISL on
welding
quality
was
the
most
important
significant,
then
followed
by
Ts,
whereas
the
effect
of
IST was
relatively
small.
Tscould
improve
the
stability
of
weld-
ing
process
within
a
proper
range,
and
when
the
ISL and
IST had
small
difference
from
their
original
values,
the
welding
quality
could
be
improved.
(3)
The
optimum
parameters
combination
could
be
obtained
through
the
results
of
range
analysis
for
the
experiments
using
different
combinations.
The
experimental
results
showed
that
more
satisfactory
performance
can
be
obtained
when
the
opti-
mum
parameters
combination
was
employed.
Acknowledgments
The
authors
would
like
to
thank
the
Foundation
for
Dis-
tinguished
Young
Teachers’
Training
of
Guangdong
(Grant
No.
Yq2013106),
the
Characteristics
of
the
Guangdong
Province
Ordi-
nary
University
Innovation
Project
(Grant
No.
2014KTSCX145)
and
Natural
Science
Foundation
of
Guangdong
Province
(2015).
References
Andersson,
J.,
Erik,
T.,
Joakim,
H.,
2006.
The
fundamental
stability
mechanism
in
tandem
MIG/MAG
welding,
and
how
to
perform
implementation.
In:
Proceedings
of
the
IIW
International
Conference,
Quebec
City,
Canada.
Bagus
Yudharto
Bharotokusumo,
Y.,
Hermans,
Marcel
Joseph
Alarie,
Richardson,
I.M.,
2008.
Process
stability
analysis
during
tandem
wire
arc
welding.
Rev.
Soldagem
Insp.,
13.,
pp.
82–90,
ISSN
0104-9224.
Chandel,
R.S.,
1987.
Mathematical
modeling
of
melting
rates
for
submerged
arc
welding.
Weld.
J.
66,
135s–140s.
DeRuntz,
B.,
2003.
Assessing
the
benefits
of
surface
tension
transfer
welding
to
industry.
J.
Ind.
Technol.
19,
55–62.
Feng,
J.,
Zhang,
H.,
He,
P.,
2009.
The
CMT
short-circuiting
metal
transfer
process
and
its
use
in
thin
aluminium
sheets
welding.
Mat.
Des.
30,
1850–1852.
Goecke,
S.,
Berlin,
F.U.B.T.,
Hedegård,
J.,
Joining,
S.I.M.R.,
Ab,
E.W.E.,
2011.
Tandem
Mig/Mag
Welding,
56
(2–3.
A
Welding
Review
Published
by
Esab,
pp.
24–28.
Halmoy,
E.,
1980.
Wire
melting
rate,
droplet
temperature
and
effective
anode
melting
potential.
Arc
Phys.
Weld.
Pool
Behav.,
49–57.
Kawagoe,
T.,
Suzuki,
K.,
2014.
Method
and
apparatus
for
arc
welding
by
controlling
welding
current,
U.S.
Keppel,
G.,
1991.
Design
and
Analysis:
A
researcher’s
Handbook.
Prentice-Hall,
Inc.
Li,
K.,
Zhang,
Y.,
2010.
Interval
model
control
of
consumable
double-electrode
gas
metal
arc
welding
process.
IEEE
Trans.
Autom.
Sci.
Eng.
7,
826–839.
Motta,
M.F.,
Dutra,
J.C.,
Gohr
Jr.,
R.,
Scotti,
A.,
2007.
A
study
on
out-of-phase
current
pulses
of
the
double
wire
MIG/MAG
process
with
insulated
potentials
on
coating
applications—part
I.
J.
Braz.
Soc.
Mech.
Sci.
Eng.
29,
202–206.
Niu,
Y.,
Xue,
H.,
Li,
H.,
Zeng,
Z.,
2010.
Effect
of
peak
pulse
voltage
on
metal
transfer
and
formation
of
weld
in
double-wire
pulsed
MIG
welding
(in
Chinese).
Trans.
China
Weld.
Inst.
31,
50–54.
Kah,
P.,
Suoranta,
R.,
Martikainen,
J.,
2013.
Advanced
gas
metal
arc
welding
processes.
Int.
J.
Adv.
Manuf.
Technol.
67,
655–674.
Peltola,
T.,
Kumpulainen,
J.,
Veikkolainen,
M.,
2010.
Novel
tailored
welding
arcs
help
welders
meet
quality
and
productivity
demands.
In:
Innovation
and
Business
Development.
Kemppi
Oy,
Lahti,
pp.
15–18.
Person,
J.A.,
Ruzek,
J.M.,
1956.
High
speed
tandem
arc
working,
US
patent:
2756311.
Peterson,
N.,
2009.
New
technology
doubles
contractor’s
pipe
welding
output.
Weld.
J.
88,
56–58.
Chandel,
R.S.,
1988.
Wire
melting
rate
in
mild
steel
MIG
welding.
Metal
Constr.
20,
214–216.
Petro,
S.J.,
2011.
Effect
of
Interpass
Temperature
on
the
Structure
and
Properties
of
Multipass
Weldments
in
High
Performance
Nickel
Alloys.
Colorado
State
University.
S.R.
Peters,
2013.
Arc
welding
with
waveform
control
function,
U.S.
patent:
US20140251967A1.
Reis,
R.P.,
Souza,
D.,
Ferreira
Filho,
D.,
2015.
Arc
interruptions
in
tandem
pulsed
gas
metal
arc
welding.
J.
Manuf.
Sci.
Eng.
137
(1),
011004.
Schierl,
A.,
2005.
The
CMT
process
a
revolution
in
welding
technology.
International
Conference,
Benefits
of
New
Methods
and
Trends
in
Welding
to
Economy,
Productivity
and
Quality.
Shi,
C.,
Zou,
Y.,
Zou,
Z.,
Wu,
D.,
2014a.
Twin-wire
indirect
arc
welding
by
modeling
and
experiment.
J.
Mater.
Process.
Technol.
214,
2292–2299.
Shi,
C.,
Zou,
Y.,
Zou,
Z.,
Zhang,
H.,
2014b.
Physical
characteristics
of
twin-wire
indirect
arc
plasma.
Vacuum
107,
41–50.
Steinert,
E.F.,
1954.
Twin
arc
welding
system,
US
patent:
2673915.30.
Suban,
M.,
Tuˇ
sek,
J.,
2001.
Dependence
of
melting
rate
in
MIG/MAG
welding
on
the
type
of
shielding
gas
used.
J.
Mater.
Process.
Technol.
119,
185–192.
Ashton,
T.,
1954.
Twin-arc
submerged
arc
welding.
Weld.
J.
33,
350–355.
Taguchi,
G.,
Clausing,
D.,
Watanabe,
L.T.,
1987.
System
of
Experimental
Design:
Engineering
Methods
to
Optimize
Quality
and
Minimize
Costs.
UNIPUB/Kraus
International
Publications,
White
Plains,
NY.
Tong,
H.,
Ueyama,
T.,
2011.
Feature
of
low
frequency
modulated
type
pulsed
MIG
welding
process.
Weld.
Join.
11,
33–35.
Tuˇ
sek,
J.,
2000.
Mathematical
modeling
of
melting
rate
in
twin-wire
welding.
J.
Mater.
Process.
Technol.
100,
250–256.
Ueyama,
T.,
Ohnama,
T.,
Yamazaki,
K.,
Tanaka,
M.,
Ushio,
M.,
Nakata,
K.,
2005.
High-speed
welding
of
steel
sheets
by
the
tandem
pulsed
gas
metal
arc
welding
system.
Trans.
JWRI
34,
11–18.
Ueyama,
T.,
Ohnawa,
T.,
Tanaka,
M.,
Nakata,
K.,
2007.
Occurrence
of
arc
interaction
in
tandem
pulsed
gas
metal
arc
welding.
Sci.
Technol.
Weld.
Join.
12,
523–529.
Ueyama,
T.,
Uezono,
T.,
Era,
T.,
Tanaka,
M.,
Nakata,
K.,
2009.
Solution
to
problems
of
arc
interruption
and
arc
length
control
in
tandem
pulsed
gas
metal
arc
welding.
Sci.
Technol.
Weld.
Join.
14,
305–314.
Uusitalo,
J.,
2007.
Modified
short
arc
process-
new
way
of
welding
root
passes.
Weld.
World
(Lond.)
51,
283.
Wang,
F.,
Hua,
X.M.,
Ma,
X.L.,
Wu,
Y.,
Cao,
N.,
Qian,
W.,
2011.
Analysis
of
arc
interference
and
interruption
in
double-wire
GMAW
welding.
Trans.
China
Weld.
Inst.
32
(7),
109–112.
Wen,
Y.,
Huang,
S.,
Wu,
K.,
Miu,
Z.,
2010.
Effects
of
current
phase
relations
on
weld
bead
formation
for
twin-wire
co-pool
pulsed
MAG
welding
(in
Chinese).
Trans.
China
Weld.
Inst.
31,
17–20.
Yao,
P.,
Xue,
J.,
Meng,
W.,
Zhu,
S.,
2009.
Influence
of
processing
parameters
on
weld
forming
in
double
pulse
MIG
welding
of
aluminum
alloy
(in
Chinese).
Trans.
China
Weld.
Inst.
30,
69–72.
Yao,
P.,
Xue,
J.,
Ma,
Q.,
Chen,
H.,
Chen,
X.,
2012.
Symmetrical
transition
waveform
control
process
of
double
wire
MIG
welding
(in
Chinese).
Trans.
China
Weld.
Inst.
33,
21–24.
Yao,
P.,
Xue,
J.,
Wang,
L.,
Zhu,
Q.,
2014a.
Fuzzy
evaluation
of
weld
quality
based
on
Matlab.
China
Weld.
23,
67–71.
Yao,
P.,
Xue,
J.,
Zhou,
K.,
Wang,
X.,
2014b.
Sample
entropy-based
approach
to
evaluate
the
stability
of
double-wire
pulsed
MIG
welding.
Math.
Prob.
Eng.
2014,
1.