Journal of Signal and Information Processing, 2011, 2, 159-164
doi:10.4236/jsip.2011.23020 Published Online August 2011 (http://www.SciRP.org/journal/jsip)
Copyright © 2011 SciRes. JSIP
159
Study on Evaluation Method of Aluminum Alloy
Pulse MIG Welding Stability Based on Arc Voltage
Probability Density
Jing Nie1, Xiao-Feng Meng 1, Yu Shi2
1School of Instrumentation Science & Optoelectronics Engineering, Beihang University, Beijing, China; 2Key Laboratory of Non-
Ferrous Metal Alloys and Processing of Ministry of Education, Lanzhou University of Technology, Lanzhou, China.
Email: niky711@163.com
Received February 24th, 2011; revised April 16th, 2011; accepted April 25th, 2011.
ABSTRACT
Bring forward a new analytica l method in order to evaluate the stability of the process of a luminum alloy pulsed MIG
welding. The ratio of the first and the second peak in arc voltage signal probability density was selected to evaluate
aluminum alloy pulse MIG welding stability. By calculating the arc voltage signal probability density from 80 sets of
welding experiments, the ratio of the two peaks in arc voltage probability in every set was captured. And the evaluation
system of aluminum alloy pulse MIG welding stability was established. The smaller the ratio of peaks in arc voltage
signal probability density is, the better the stability of the welding will be; the bigger the ratio of peaks in arc voltage
signal probability density is, the poorer the stability of the welding will be.
Keywords: Aluminum Alloy, Pulse MIG, Welding Stability, Probability-Density, Arc Voltage
1. Introduction
The process of aluminum alloy pulse MIG welding is
characterized by its non-linearity, strong time-varying,
strong coupling between the parameters, and so forth [1].
The reason why its process is unstable has not been well
explained until now. Becaus e of the lack of the parameter
matched stability evalu ation indicators, the control of the
aluminum alloy pulse MIG welding by changing the
welding parameters is very difficult. The fail of matching
the parameter will probably lead to the instability in the
welding process or even worse. In recent years, many
scholars at home and abroad have studied the stability in
the welding process: reference [2] put forward an on-line
evaluation model of the process stability by use of the
statistical analysis and partial least-squares regression.
Reference [3] an d [4] found that the approximate entropy
of the arc signal varied with the change of the parameter.
Reference [5] and [6] analyzed the prob ability density of
the corresponding welding current and voltage and fi-
nally evaluated the welding stability.
This paper analyzed the probability density distribu-
tion of the 80 sets of voltage signal under different weld-
ing speed, wire-feeder speed and duty cycle matched pa-
rameters.
2. Experimental System
The experiment applied DELEX VIRIO MIG-400L
welding machine to pure Ar gas shield bead-on-plate
welding. The designation of the welding wire was 5356,
diameter 1.2 mm. The material of the workpiece was
aluminum50581-H321 with the thick of 6mm. The torch
was fixed during welding. And its distance between the
workpiece was unchanging. The workpiece moved on the
worktable. The initial wire extension was 15 mm. Data
acquisition card PCL812PG was applied to the welding
and its frequency was 1 kHz. The current sensor was
CSM400FA/100 mA. The time between the arc igniting
and ending was not less than 45 s. Video acquisition
captured the condition of the arc and molten pool to
judge the stability of the welding. Figure 1 was the ex-
perimental system.
Because of the finding that the duty cycle of the pulse
current, the wire-feeder speed and welding speed impact
the welding stability greatly, the experiment combined
the duty cycle, the wire-feeder speed and the welding
speed to weld. Th e base valu e o f the weld ing curren t was
25 A, the peak value 180 A, pulse frequency 40 Hz,
welding voltage 18 V, and shielding airflow 18 L/min.
Because the aluminum alloy pulse MIG welding was
Study on Evaluation Method of Aluminum Alloy Pulse MIG Welding Stability Based on Arc
160 Voltage Probability Density
x-axis
y
-axis
PC
Dual-axis table
PCL-812PG
Acquisition
Motion
Control
Card
Video
Capture
Card
Power supply
Hall sensor
CCD
MIC
Isolation circuit
Stepping
Motor
Stepping
Motor
Torch
Figure 1. Experimental system.
very strict to its parameter matching, the welding pa-
rameters in Table 1 were fixed through the welding ex-
periments. These parameters could be combined ran-
domly. The maximum and minimum of single parameter
was the upper and lower limits of the matching. 5 × 4 × 4
sets of welding experiments were conducted and 80 sets
of arc voltage signals were acquired.
3. Voltage Signal Probability Distribution
Analysis
With regard to time array
x
twhen amplitude falls
on the range
,
x
xx , the total time is 1
k
xi
T
t
.
When observation time tends to infinity, array X
TT is
the probability of case

x
xt x x

, which can be
written as limt xxT

x
T
Px xT


the prob-
ability density function can be defined as:

0
00
1
Px xtxx
lim x
11
lim limlim
T
xx
k
xi
xTx i
x
P
Tt
xT T


 













(1)
Probability density function constantly equals
to real-value non-negative function. The data must be
standardized before analyze its probab ility density. Ev ery
variable minus the mean and then divided by its standard
deviation, supposing the collected good-running data ma-
trix is

px
np
X
R
, each column corresponds to a variable
Table 1 .Welding parameters.
Welding speed
cm/min Wire feed speed
m/min Duty cycle
%
11 6 50
15 6.5 44
20 7 40
25 7.3 37.5
31
and each line corresponds to a sample. Standardized the
as follows:

12
11 1
11 1diag,,,
T
s
p
XX P
s
ss


 



(2)
P—the mean of variable X,
12
111
,,,
p
P
s
ss
, S—the standard deviation of the va-
riable, 12
,,,
p
Sss s
. Then, normalize the standar-
dized data. Finally, calculate and analyze the probability
density.
Analyze the probability density of 4 sets of voltage
signal selecting from 80 sets of welding experiments
arranging from good to bad. Every set of voltage signal
cut from the 10th second and 30000 p oints wer e captur ed,
namely 30 seconds. The results are shown in Figure 2.
According to Figure 2, two peak values were found in
each set of probability density distribution figure and the
distance between the two peaks increase. In order to get
the accurate variation of the two peaks from 4 sets of
experiments, the ratio of the first peak and the second
peak was selected to reflect the variation of the distance
between the two peaks. The variation of the ratio was
shown in Figure 3.
The ratio of the two voltage probability density peaks
gradually increase in the four sets of experiments. After
comparing the corresponding weld bead, it was found
that the ratio of peaks was small when the weld bead
shaping was good, and th e opposite was big. The second
set and the fourth set were shown in Figure 4. Obviously,
the weld bead of the second set was better than that of
the fourth. And the ratio of peaks of the second set was
smaller than that of the fourth. Therefor e, the ratio of the
two peaks in voltage probability can be applied to evalu-
ate the stability of aluminum allay MIG welding.
4. Evaluation System of Welding Stability
According to the result above, 80 sets of welding voltage
signal was calculated for probability density. By making
use of the ratio of the two peaks in each set of voltage
signal probability density, the evaluation system of alu-
minum alloy pulse MIG welding stability was established.
The result is shown in Figures 5-7. What is shown in
Figures 5-7 is the fluctuation of the ratio of peaks in arc
voltage signal probability density under different pa-
rameter. The figure well reflects the aluminum alloy
pulse MIG welding stability under different welding
speed, wire feed speed and duty cycle. In aluminum alloy
pulse MIG welding, the smaller the ratio of peaks in arc
voltage signal probability denity is, the bet ter the stabi lity s
Copyright © 2011 SciRes. JSIP
Study on Evaluation Method of Aluminum Alloy Pulse MIG Welding Stability Based on Arc
Voltage Probability Density
Copyright © 2011 SciRes. JSIP
161
(a) (b)
(c) (d)
Figure 2. Probability density of voltage. (a) NO.1; (b) NO.2; (c) NO.3; (d) NO.4.
(a)
(b)
Figure 4. The weld of different parameters. (a) NO.1; (b)
O.4.
Figure 3. Changes in the ratio of peaks in probability den-
sity. N
Study on Evaluation Method of Aluminum Alloy Pulse MIG Welding Stability Based on Arc
162 Voltage Probability Density
(a) (b)
(c) (d)
Figure 5. The match of welding and wire feed speeds on different duty cycle. (a) duty cycle 50%, (b) duty cycle 44%, (c) duty
cycle 40%, (d) duty cycle 37.5%.
(a) (b)
(c) (d)
Figure 6. The match of welding speeds and duty cycle on different wire feed speeds. (a) wire feed speed 6 m/min; (b) wire feed
speed 6.5 m/min; (c) wire feed speed 7 m/min; (d) wire feed speed 7.3 m/min.
Copyright © 2011 SciRes. JSIP
Study on Evaluation Method of Aluminum Alloy Pulse MIG Welding Stability Based on Arc 163
Voltage Probability Density
(a) (b)
(c) (d)
Figure 7. The match of wire feed speeds and duty cycle on different welding speeds. (a) welding speed 11 cm/min; (b) welding
speed 15 cm/min; (c) welding speed 20 cm/min; (d) welding speed 25 cm/min.
Table 2. The ratio of peaks in voltage probability density and weld forming under different welding parameters.
Number Wire feed speed m/minDuty cycle % Welding speed
cm/min Ratio of peaks in
probability density Stability Welding forming
1 6 50 11 1.3333 Stable Good
2 6 44 15 1.24 Stable Good
3 6 37.5 20 2.6111 Unstable Bad
4 6.5 44 25 2.1 Unstable Bad
5 6.5 40 15 4 Unstable Bad
6 6.5 37.5 20 3.0625 Unstable Bad
7 7 44 11 2.7778 Unstable Bad
8 7 40 25 2.35 Unstable Bad
9 7.3 50 20 0.8649 Stable Good
of the welding will be. The opposite is the instability of
the welding.
Table 2 shows the ratio of peaks in voltage probability
density and weld forming under different welding pa-
rameters, through which the ratio of peaks in arc voltage
signal probability density and the welding stability can
be compared more objectively.
5. Conclusions
1) By analyzing the ratio of peaks in arc voltage signal
probability density and comparing the welding bead, the
ratio of peaks in aluminum alloy pulse MIG welding arc
voltage signal probability density is suitable to be applied
to evaluate the welding stability.
Copyright © 2011 SciRes. JSIP
Study on Evaluation Method of Aluminum Alloy Pulse MIG Welding Stability Based on Arc
164 Voltage Probability Density
2) According to the fluctuation condition of the ratio
of peaks in arc voltage signal probability density, the
evaluation system of aluminum alloy pulse MIG welding
stability is established. The smaller the ratio of peaks in
arc voltage signal probability density is, the better the
stability of the welding will be; the bigger the ratio of
peaks in arc voltage signal probability density is, the
poorer the stability of the welding will be.
6. Acknowledgements
This paper is supported by the National Natural Science
Foundation of China under Grant No.50675093 and Na-
tional Natural Science Foundation of International Co-
operation under Grant No.50710105060. Thanks for help
from Key Laboratory of Non-ferrous Metal Alloys and
Processing of Ministry of Education, Lanz hou University
of Technology.
REFERENCES
[1] S. Yamada and K. Masubuchi, “Advanced Welding
Technology Keeps Japan’s High-Speed Trains on Trac,”
Welding Journal, Vol. 79, No. 11, November 2000, pp.
48-53.
[2] Y. Cai, H. L. Yang, K. Xu and Y. X. Wu, “An On-line
Evaluation Model for Process Stability of Short-Circuit
Arc Welding,” Journal of Shanghai Jiaotong University,
Vol. 39, No. 7, July 2005, pp. 1038-1041.
[3] R. T. Charles, A. L. Randall and B. S. Herschel, “Is There
Evidence of Determinism in Droplet Detachment within
the Gas Metal Arc Welding Process,” 6th International
Conference on Trends in Welding Research, April 2002.
[4] B. Cao, X. Q. Lv, M. Zeng and S. S. Huang, “Analysis of
Welding Conditions in Short-Circuiting Transition Base
on Approximate Entrop y,” Chinese Journal of Mechanical
Engineering, Vol. 43, No. 10, October 2007, pp. 50-54.
doi:10.3901/JME.2007.10.050
[5] Y. Zhang, X. Xu and Y. Cai, “Collection and Analysis
System of Welding Parameters Based on LabView,”
Transactions of the Chi na Welding Institu tion, Vol. 24, No.
4, August 2003, pp. 43-46.
[6] F. S. Wu, Z. L. Lu and Z. F. Liu, “Study on the Process
Parameter Monitoring and Analysis System of SAW
Submarine Shell,” China Mechanical Engineering, Vol.
15, No. 7, April 2004, pp. 621-624.
Copyright © 2011 SciRes. JSIP