Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02272953 1999-OS-21
' WO 98/22247 PCT/FR97/02070
Pulsed-current arc welding process and apparatus
The present invention relates to pulsed-current
arc welding processes and apparatuses, and more
particularly to gas-shielded pulsed-current arc welding
processes and apparatuses using consumable wire.
In general, gas-shielded arc welding processes
using consumable wire consist in striking an electric
arc between the sheets to be welded and a consumable
welding wire. The consumable wire undergoing a
translational movement is progressively melted by the
arc and thus contributes to the formation of a weld
bead.
The fact of being able to feed the wire at
different speeds gives rise to the existence of several
modes of operation, the welding characteristics of
which are completely different. Depending on the value
of the current and of the voltage, the way in which the
molten metal is transferred from the wire to the liquid
pool varies considerably.
At low currents (80 to 150 A), the Joule effect
and the heat released by the arc melt the end of the
wire, thus forming a drop of molten metal. Since the
wire continues to advance, the drop comes into contact
with the weld pool. There is then a short circuit
(extinction of the arc) and it is at that moment that
the drop is detached, and then transferred into the
pool. When the transfer has taken place, the arc is
reignited and the process may thus start all over
again. This constitutes the short-circuit transfer
(short arc) mode.
At higher currents (150 to 250 A), the transfer
mode is globular. This transfer mode is not used very
much in practice since it results in erratic transfer
and in considerable spatter. In this mode, the drop may
grow inordinately and either explode before any contact
with the sheet or be transferred by short-circuiting
incorrectly, often outside the pool.
CA 02272953 1999-OS-21
- 2 -
At very high currents and about 30 to 35 V,
depending on the shielding gas used, the welding enters
the axial spray (spray-arc) mode in which a string of
small drops of molten metal regularly escape from the
end of the molten wire.
The current in this case is high. enough to
expel the drops formed before any contact with the
sheet.
There is a fourth, more artificial, transfer
mode which, from the standpoint of the average welding
energy, lies within the globular range: this is the
pulsed mode. In this case, welding is carried out using
a pulsed current, the pulse parameters being chosen in
such a way that, for each pulse, transfer is of the
spray-arc type with a single drop per pulse. This
requires special generators in which the current
waveform is controlled. This mode is often used as it
allows good transfer within an average energy range in
which conventional DC welding is difficult to put into
practice. However, this mode is very noisy and it is
not unusual to reach levels of acoustic noise close to
100 dBA (at 40 cm from the arc), which quite easily
exceed the thresholds defined in the legislation
concerning the regulation of noise emitted by machines.
In pulsed mode, the pulse may have various waveforms:
the trapezoidal waveform is most used in practice, but
it is possible, by means of transistorized current
transformers, to give the pulse various waveforms,
while still controlling quite well the parameters which
define it.
The procedure for establishing a pulsed-current
welding programme consists in finding the most suitable
situation between the pulse parameter values and the
desired welding result. That is to say, for each wire,
shielding gas and wire feed speed, it is necessary to
determine the values of these parameters which best
satisfy the criteria according to which the welding
operation is judged. In general, a good pulsed-current
welding programme must be one in which only one drop is
CA 02272953 1999-OS-21
- 3 -
detached per pulse, with drop diameters of about the
diameter of the wire. For a given wire feed speed, this
therefore determines a pulse frequency. Moreover, the
common belief is that the rise and fall slopes of the
current should be as high as possible so as to obtain
stiff and highly directional arcs which, for most
operators, are the best way of producing a weld bead in
pulsed mode. The other parameters are generally chosen
empirically.
Thus, the document JP-A-57,124,572 describes a
pulsed arc welding process using a consumable electrode
and a shielding gas consisting of an inert gas.
According to this document, the amplitude of the
current must be at least 300 A, and the duration
extending between the start of the current increase and
the start of the current fall, that is to say including
the peak time during which the current is at its
maximum value, is between 0.2 ms and 4 ms, this being
the case for triangular or rectangular pulses.
However, the problem of noise pollution caused
during an arc welding operation in pulsed mode has not.
hitherto been solved.
The object of the invention is therefore to
provide an arc welding process in pulsed mode having an
acceptable level of acoustic noise (of less than
approximately 90 dBA) and good welding characteristics.
For this purpose, the subject of the invention
is a pulsed-current arc welding process, characterized
in that each electric current pulse has rise and/or
fall slopes of between 50 Alms and 1000 Alms, apart
from triangular pulses.
Preferably, each current pulse has a peak time
of non-zero duration.
Within the context of the present invention,
the expression "pulse having a peak time of non-zero
duration" should be understood to mean a pulse whose
representation, visualization or measurement shows that
there is a non-zero duration separating the end of the
rise of the current and the start of the fall of the
CA 02272953 1999-OS-21
- 4 -
said current. Over the period of this non-zero duration
of the peak time, the amplitude of the current may,
depending on the case, be maintained approximately
constant or, in contrast, may undergo slight
variations, as will be explained in detail below.
Depending on the case, the process may include
one or more of the following characteristics:
- the rise and/or fall slopes are between
100 Alms and 600 Alms, preferably between 150 Alms and
500 Alms;
- the said arc welding process is a consumable-
wire process;
- the pulse frequency is adjusted so as to obtain
drops of molten metal with a diameter, at the moment of
their release, of between 1 and 1.4 times the diameter
of the said consumable wire;
- the pulse frequency is adjusted so as to obtain
drops of molten metal with a diameter, at the moment of
their release, of approximately 1.2 times the diameter
of the said consumable wire;
- the pulses are trapezoidal;
- the pulses are composed of a combination of
several identical, similar or different patterns chosen
from trapezoidal, sinusoidal, triangular, rectangular
and square patterns;
- the welding process is carried out under a flow
of shielding gas that includes an active gas; and
- the active shielding gas comprises argon and/or
helium, and at least one compound chosen from CO2, H2,
OZ and mixtures thereof.
The present invention furthermore relates to a
welding apparatus capable of implementing the process
according to the invention and, in particular, to a
pulsed-current arc welding apparatus that includes
means allowing electric current pulses to be obtained
which have rise and/or fall slopes of between 50 Alms
and 1000 Alms, apart from triangular pulses.
CA 02272953 1999-OS-21
- 5 -
Preferably, the apparatus of the invention
allows pulses to be obtained which furthermore have a
peak time of non-zero duration.
The invention will be more clearly understood
on reading the description which follows, given solely
by way of example, and with reference to the drawings
in which:
- Figure 1 is a diagrammatic view of a gas-
shielded consumable-wire arc welding set;
- Figure 2 is a diagrammatic view, with cut-away,
of a process for welding sheets carried out using such
a set;
- Figure 3 is a graph showing a trapezoidal
electric pulse which has the characteristics of the
invention;
- Figure 4 shows diagrammatically the
decomposition of a pulse of any shape;
- Figures 5 and 6 are graphs showing,
respectively, sinusoidal pulses according to the
invention and triangular pulses outside the scope of
the present invention;
- Figure 7 shows a trapezoidal pulse shape
according to the invention; and
- Figures 8 to 10 show a distorted trapezoidal
shape.
The arc welding set in Figure 1 essentially
comprises a transistorized electric current generator
10, a gas supply 20, a welding torch 30, a consumable-
wire supply reel 40 and a consumable-wire drive feed
unit 50. The generator 10 is connected via a first
cable 60 to the sheets 70 to be welded, and via a
second cable 80 to the consumable-wire drive feed unit
50 so as to deliver the desired electrical signal
between the consumable wire and the sheets to be
welded. A pipe 100 allows the torch to be supplied with
gas via the drive feed unit 50. Moreover, a control
cable 90 connects the generator 10 to the consumable-
wire drive feed unit 50 so as to tailor the current
amplitude to the wire feed speed.
CA 02272953 1999-OS-21
- 6 -
Figure 2 shows diagrammatically the operation
of welding sheets by means of the torch 30. The figure
shows, in the end nozzle 31 of the torch, the terminal
part of the consumable wire 32, a contact tube 33 which
guides this wire and supplies it with the electric
current, a column of gas 34 shooting forth from the
torch towards the sheets 70 to be welded, and the weld
pool 35 which comes from the melting of the sheets and
the consumable wire and which, after cooling, forms a
weld bead 36.
Figure 3 shows a graph of a trapezoidal
electric pulse, this being the pulsed current waveform
most used in pulsed-current arc welding. In this graph,
the time t is plotted on the x-axis and the current I
is plotted on the y-axis. The parameters defining the
signal shown in Figure 3 are the period 1/f, the peak
current Ih, the background current Ib, the rise time
tm, the fall time td, the peak time (segment BC), the
pulse time Tp and the background time Tb. Here, the
rise slope (segment AB) and the fall slope (segment CD)
of the pulse are approximately constant over tm and
over td, respectively, and are defined by (Ih-Ib)/tm
and (Ih-Ib)/td, respectively.
In general and thus as shown in Figure 4, a
pulsed-current wavetrain is a periodic succession of 4
elements, namely a background current level, a
continuous and non-monotonic current rise from the
background current level to the peak current level, a
peak current level and a continuous and non-monotonic
fall from a peak current level to a background current
level. Whatever the shape of this wavetrain, each of
these elements may be decomposed into a series of
current segments of length Did and of duration ~t~, as
shown in Figure 4.
The current rise is therefore defined by a
succession of n current segments of positive slope.
Likewise, the current fall is defined by a succession
of k current segments of negative slope.
CA 02272953 1999-OS-21
The current rise starts when the value of the
slope of a significant number of consecutive current
segments moves away from 0 in the positive direction.
The current rise stops when the value of the slope of a
significant number of current segments is close to 0
(the peak current level).
The current fall starts when the value of the
slope of a significant number of current segments moves
away from 0 in the negative direction. It terminates
when the value of the slope of a significant number of
current segments approaches 0 (the background current
level ) .
The total values of the rise and fall slopes
are therefore defined as follows.
Di+~ = i+~ - i+~-i
~t+~ = t+~ - t+~_1
2 0 P+~ = 0i+~ /Ot+i
The rise slope (Pm) is therefore given by the
formula
1 "
2 5 Pm = - ~ P+p
n
Likewise:
Di-~ = i-~ - i-~-i
Ot-~ = t-~_1 - t
P-~ = Di-~ /0t p
The fall slope (Pd) is therefore given by the
formula:
CA 02272953 1999-OS-21
k
Pd = k ~ p~_
l
_ g _
Thus, it is always possible, from a signal of
any (but not triangular) shape and preferably with a
peak time of non-zero duration, to determine the rise
and/or fall slopes of this signal.
Figure 5 shows a wavetrain of sinusoidal shape
which can be decomposed as described above and shown in
Figure 4, so as to determine therefrom the rise and/or
fall slopes of the current.
Figure 6, given by way of comparison, shows a
triangular electric pulse not in accordance with the
invention. In other words, the peak time (Th) of the
pulses in Figure 6 is zero.
The parameters defining the signals in Figures
5 and 6 are identical to those in Figure 3.
The Applicant has evaluated the influence of
the rise and/or fall slopes and of the pulse frequency
on the noise level. Thus, the ranges of values in which
these parameters allow a significant reduction in noise
to be obtained, while maintaining good weld bead
production, have been determined. The resulting set
current was delivered by a transistorized generator
conventionally used in arc welding in pulsed mode.
Surprisingly, the Applicant has found that, for
non-triangular pulses and for a given wire feed speed,
the use of low fall and rise slopes of between 50 Alms
and 100 Alms, and more particularly of between 100 Alms
and 500 Alms, as well as the use of low pulse
frequencies, i.e. corresponding to drops with a
diameter, at the moment of their release, of between 1
and 1.4, and more particularly approximately 1.2, times
the diameter of the wire, result in a considerable
reduction in the noise levels generated, while still
maintaining good operating weldability.
The experimental results are given in the
tables below, in which Pm and Pd denote, respectively,
the rise slope and the fall slope which are
CA 02272953 1999-OS-21
_ g _
approximately constant over the times tm and td,
respectively, for each test presented.
Table 1
Wire feed speed 2 4 6
(m/min)
IH (A) 360 390 390
IB (A) 30 30 110
tm (ms) 0.5 0.6 0.6
td (ms) 0.9 0.8 0.6
Pm (A/ms) 660 600 467
Pd (A/ms) 367 450 467
Th (ms) 1.6 1.5 1.5
Tb (ms) 14 5 3.1
f (Hz) 57 127 171
noise (dBA) 86.5 89.5 90
Table 1 shows, for different wire feed speeds,
using a 1.2 mm diameter steel wire and a gas mixture
(Ar + 3o COZ + to 02) conventionally used in pulsed MAG
(Metal Active Gas) welding, the typical values of the
parameters of a trapezoidal pulse according to the
invention (see Figure 3). The noise levels indicated
were measured at 40 cm from the arc.
CA 02272953 1999-OS-21
- 10 -
Table 2
Wire feed speed 2 4 6
(m/min)
IH (A) 330 330 330
IB (A) 50 110 170
tm (ms) 1.5 1.5 1.5
td (ms) 1.5 1.5 1.5
Pm (A/ms) 187 147 107
Pd (A/ms) 187 147 107
Th ( ms ) 2 2 2
Tb (ms) 23 12.8 6.1
f (Hz) 35.7 56 90
noise (dBA) 81 83 81
Table 2 shows, under the same conditions, the
values taken by the same parameters with reduced rise
and fall slopes and reduced pulse frequencies.
This very clearly shows the reduction in noise
provided by the use of lower slopes and pulse
frequencies according to the invention, namely a saving
of between 5 dBA and 9 dBA for this range of wire feed
speeds. The operating weldability is still good,
although the arcs are not so stiff as in the case of
Table 1 ("splayed" arcs).
In other words, by comparing the results given
in Tables 1 and 2 above, it may be seen that a
reduction in the slope values leads to an even more
significant reduction in acoustic noise.
CA 02272953 1999-OS-21
- 11 -
Table 3
Wire feed speed 1.3 3
(m/min)
IH (A) 340 360
IB (A) 30 30
tm (ms) 0.6 0.6
td (ms) 0.8 0.8
Pm (A/ms) 517 550
Pd (A/ms) 388 413
Th (ms) 1.3 1.3
Tb (ms) 23.8 10.2
f (Hz) 38 77
noise (dBA) ~ 85 87.5
Table 4
Wire feed speed 1.3 3
(m/min)
IH (A) 340 360
IB (A) 30 30
tm (ms) 1.5 1.5
td (ms) 1.5 1.5
Pm (A/ms) 207 200
Pd (A/ms) 207 200
Th (ms) 1.4 2
Tb (ms) 36.7 20.4
f (Hz) 24 39
noise (dBA) 83.3 84.5
CA 02272953 1999-OS-21
' - 12 -
Tables 3 and 4 show the results of a welding
programme, carried out this time on a 1.2 mm diameter
stainless steel wire and with a gas mixture (Ar + 3~
COZ + to Hz) conventionally used in pulsed MAG welding.
Table 3 corresponds to a first trapezoidal signal
according to the invention (see Fig. 3) and Table 4
corresponds to a trapezoidal signal with reduced rise
and fall slopes and reduced pulse frequency. Here too,
it may be seen that there is a reduction in noise when
gentler rise and fall slopes and lower frequencies are
used.
The process according to the invention may be
applied to any type of pulsed-current arc welding, with
or without a consumable wire, and using pulses,
especially trapezoidal pulses or pulses of various
composed shapes, such as combinations of trapezoidal,
triangular, sinusoidal, rectangular and square
patterns.
As mentioned above, the trapezoidal shape is
the pulse shape most used in pulsed-current arc
welding. Nevertheless, due to the regulating of the
current inherent in the technology of the welding set
used, a given current waveform, in particular of
trapezoidal shape, may be substantially modified and/or
distorted, that is to say the current regulated by the
welding set may not exactly follow the set current
which is delivered thereto.
However, in such cases, it is possible to
reconstruct the set current waveform from the
"distorted" shape, i.e. the shape observed by measuring
the current. To do this, the current waveform is
firstly determined by means of an oscilloscope
connected to a current measurement sensor whose
bandwidth is at least 10 kHz.
Usually, the measurement sensor is placed in
the welding circuit, for example in the torch system,
and the measurement is carried out during the welding
operation; nevertheless, care should be taken to ensure
that the operation of the generator during measurement
CA 02272953 1999-OS-21
- 13 -
is not precluded by the manufacturer in the instruction
manual for the said generator.
Next, the set shape is reconstructed from the
measured shape of the pulse (so-called "distorted"
shape), as shown diagrammatically in the appended
Figures 7 to 10. .
More specifically, Figure 7 shows the diagram
of a trapezoidal pulse (set shape) and Figures 8 to 10
show diagrammatically certain distortions likely to
affect the trapezoidal set shape, such as those
appearing on the screen of the oscilloscope.
In particular, Figures 8 and 10 especially show
diagrammatically the phenomenon known in power
electronics as "overshooting", which corresponds to the
current exceeding the set current at one or more
moments, due to the response time of the generator.
This phenomenon gives the peak time (Th) and/or
the background time (Tb) of the current a saw-tooth,
i.e. irregular, appearance.
When there is a "distorted" pulse, such as
those shown in Figures 8 and 10, the trapezoidal set
pulse shown in Figure 7 may be obtained by calculating
the mean value of the current.
When there is a "distorted" pulse, such as that
shown in Figure 9, the trapezoidal set pulse is
obtained by calculating the mean value of the current
from the trace of the peak time (Th) plateau and from
the tangent (Tgm) at the start of the rise and the
tangent (Tgd) at the start of the fall of the current.
Furthermore, it is found that programmes
established with low pulse frequencies and gentle
slopes generally lead to higher mean current amplitudes
and therefore to hotter weld pools. This is an
advantage, especially in the case of downhand welding
of stainless steels which, in the presence of a
slightly oxidizing gas, are relatively difficult to
wet.
