Generally, the robotic arm Gas Metal Arc (GMA) welding process involves
sophisticated sensing and control techniques applied to various process
variables. This research work has been carried out to develop a response
surface model to establish a relationship between process parameters and
weld bead penetration in the robotic arm GMA welding process. To achieve
the objectives, partial-penetration and single-pass welds were fabricated
in 14 mm thick “Y” grooved modified SAE 1022 low carbon steel plate
using three different process parameters such as welding current, welding
speed, and land height. The experiments were conducted based on a three
factor, five-level, face-centered composite design matrix. After welding, the
joints were sectioned and bead profiles were scanned using stereo-zoom
macroscope. Empirical relationships were developed to predict bead profiles
such as penetration, weld width and reinforcement height incorporating the
significant main and interaction welding parameters alone. The developed
empirical relationships can be effectively used to predict the weld bead
profile of the robotic arm GMA welded modified SAE 1022 low carbon steel
plate joints.
KEYWORDS: Robotic arm GMA Welding, Weld bead profiles, Design of
Experiments, Response surface methodology, Optimization
INTRODUCTION
Welding is inherently complex due to the simultaneous
occurrence of chemical, physical, and mechanical phenomena
involving solid, liquid, gas, and plasma states within a small volume,
over short distances and time, and at high temperatures [1]. This
complexity makes sensing, measuring, and controlling welding
parameters challenging in conventional welding processes. To
manage this, automatic or robotic welding is widely adopted in
industries like automotive for better process control [2]. Although
several studies have modeled the relationship between process
parameters and bead geometry in semi-automatic arc welding
[3], limited work has focused on automatic or robotic arc welding
systems. Gas Metal Arc Welding (GMAW), commonly known
as MIG welding when using argon, helium, or their mixtures, is
also termed Metal Active Gas (MAG) welding when using pure
36
CO2
as the shielding gas [4]. A constant voltage direct current
power source is typically used in GMAW and Robotic GMAW
(R-GMAW), though constant current or alternating current
can also be employed as needed [5]. There are four main metal
transfer modes in GMAW globular, short-circuiting, spray, and
pulsed-spray each offering unique benefits and limitations
[6]. This study employs Response Surface Methodology to
optimize welding parameters for partial penetration, bead width,
reinforcement height, and dilution in robotic GMAW of Y-grooved
SAE 1022 low-carbon steel joints using 2.0 mm ER70S-6 filler
wire. ANOVA was conducted to identify significant main and
interaction effects. Optimizing these parameters improves weld
quality and reduces experimental costs [7].
METHODOLOGY
Two housing halves of a modified SAE 1022 low carbon steel,
with the chemical composition and electrode wire are shown in
Table 1. The specification of the welding wire depends upon the
material that is required to be welded. So, AWS A5.18ER 70S-6
welding wire of 2.0 mm diameter was selected. As the thickness
of the base metal was 14 mm, welding wire with a diameter of 2.0
mm was selected. The significant welding input parameters that
can affect the output response were identified and their range of
operation was selected as shown in Table 2.
Table 1 Chemical Composition (wt.%) of base metal
and filler wire
Compositions
C
Mn
Si
S (max)
P
(max)
Cu
(max)
Base
Metal
(Modified
SAE 1022)
0.17
1.38
0.27
0.003
0.02
0.006
Cr
V
Ti
Al
Nb
Ni
0.016
0.077
0.019
0.039
0.035
0.008
Filler
Metal
(ER70S-6)
0.19
1.63
0.98
0.025
0.025
0.025
Table 2 Levels of Welding Process Parameters
Parameters
Welding
Current
Welding
Speed
Land
Height
Units
Amps
mm/min
Mm
Range
450 –
525
350 – 450
4 – 7
Fig. 1 Robotic GMA Welded Modified SAE 1022 Low Carbon Steel
This study used an inverter-based robotic power source
LINCOLN Power Wave® S700 Advanced Process Welder
(K3279-1) paired with a FANUC ROBOT M201A/121 to join 14
mm "Y" grooved SAE 1022 low carbon steel housing halves. This
setup was chosen for its fast arc start, stick-out and crater control,
and Fresh Tip Treatment Technology (FTTT), which prevents
globule formation at the wire tip during weld stop [8]. Weld bead
geometry was examined using a stereo-zoom macroscope. The
welded “Y” groove butt joint is shown in Figure 1. In the Robotic GMAW process, process parameters were categorized into
two groups. The first group, consisting of variable parameters,
includes welding current, arc voltage, torch travel speed, and
groove land height. The second group, held constant during the
process, includes shielding gas flow and composition, torch angle,
stick-out, electrode extension, and electrode diameter (Table 1).
The constant parameters and their limits are listed in Table 3.With
this value (Table 2 & 3) a full design of experiment was performed
using three factors and five levels central composite design
matrix, thus the model consists of 20 treatments.
Table
3 Constant process variable and values
Variables
Units
Values
Arc voltage
Volts
32
Type of shielding gas
%
Argon
80%, CO2 20%
Shielding gas flow rate
LPM
26
Torch angle
Degree
90
Electrode specification
AWS A5.18
ER70S-6
Stick out
mm
20
Machine efficiency
%
85
Housing halves thickness
mm
14
Groove angle
Degree
46
RESULTS AND DISCUSSION
1. Development of an empirical relationship
The input parameters (variables) that were selected using the
four-factor and three levels face-centered composite design
matrix. Analysis of variance (ANOVA) method was used to
identify the significant main and interaction between welding
parameters and their level of significance (Table 6). Response
Surface Methodology has used to optimize the robotic arm
GMA welding process parameters to attain desired weld bead
characteristics such as optimum partial - penetration, weld
bead width, reinforcement height, and dilution. The levels of
the significant process variables are shown in Table 4.
Table 4 Levels of Welding Process Parameters
Parameters
Units
Levels
-1.68
-1
0
+1
+1.68
Welding
Current (I)
Amps
450
465.2
487.5
509.8
525
Welding
Speed (S)
mm/min
350
370.3
400
429.7
450
Land
height (H)
mm
4
4.6
5.5
6.4
7
2. Selection of a mathematical model for empirical relation
A convenient way to evaluate the performance of the
mathematical equation in presenting the system under
investigation is through the concept of response surface
methodology [9]. It is assumed that there exist some functional
relationship such as: - Y = f (I, S, H)
This defines the dependence of the response Y on the welding
parameters I, S, H where these parameters have already defined
in Table 4. The response Y may be any of the bead parameters i.e.
penetration (P), bead width (W), and bead height (R) etc. Assuming
a quadratic relationship in the first instance and taking into
account all possible interactions, Equation could be written in the
form of the following polynomial.
Y = b0
+ b1
I + b2
S + b3
H + b12
IS + b13
IH + b23
SH + b11
I2 +
b22
S2 + b33
H2
Or it can be also written as,
Y = b0
+ b1
I + b2
S + b3
H + b4
IS + b5
IH + b6
SH+ b7
I2 +
b8
S2 + b9
H2
These relations for predicting different responses mentioned
above are given below.
Penetration (P) = {+ 9.50 + 1.24I - 0.29S - 1.0H - 0.14IS - 0.39IH +
0.36SH + 0.69I2 - 0.069S2 - 0.39H2} mm
Width (W) = {+17.67 + 0.78I - 0.70S + 0.27H - 0.087IS - 0.037IH -
0.087SH + 0.51I2 + 0.036S2 - 0.017H2} mm
Reinforcement height (R) = {+2.74 + 0.046I - 0.28S + 0.30H +
0.063IS - 0.012IH + 0.063SH - 0.28I2 + 0.19S2 + 0.035H2} mm
The response Y may be any of the bead parameters i.e.
penetration (P), bead width (W), and bead height (R) etc. Assuming
a quadratic relationship in the first instance and taking into
account all possible interactions, Equation could be written in the
form of the following polynomial.
Y = b0
+ b1
I + b2
S + b3
H + b12
IS + b13
IH + b23
SH + b11
I2 +
b22
S2 + b33
H
Table
5 Experimental design matrix
Run
No.
Variables
Results
Current
(I)
Welding
Speed (S)
Land
Height (H)
Penet.
(P)
Bead
Width (W)
Rein.
Height (R)
Amps
mm/min
mm
mm
Mm
mm
1
465.2
370.3
4.6
10.2
18
2.5
2
509.8
370.3
4.6
14
19
2.7
3
465.2
429.7
4.6
10
16.7
2
4
509.8
429.7
4.6
12
18
2.5
5
465.2
370.3
6.4
9
18.5
2.8
6
509.8
370.3
6.4
10
20
3
7
465.2
429.7
6.4
9
17.5
2.6
8
509.8
429.7
6.4
10.7
18
3
9
450
400
5.5
9
17.2
2.2
10
525
400
5.5
14
21
1.8
11
487.5
350
5.5
10.5
19
4.2
12
487.5
450
5.5
9
16.5
2.5
13
487.5
400
4
12.5
17.2
2.2
14
487.5
400
7
8.8
18
3.6
15
487.5
400
5.5
9.5
17.7
2.75
16
487.5
400
5.5
9.4
17.7
2.75
17
487.5
400
5.5
9.5
17.6
2.75
18
487.5
400
5.5
9.5
17.7
2.7
19
487.5
400
5.5
9.5
17.6
2.75
20
487.5
400
5.5
9.6
17.7
2.7
Or it can be also written as,
Y = b0
+ b1
I + b2
S + b3
H + b4
IS + b5
IH + b6
SH+ b7
I2 +
b8
S2 + b9
H2
These relations for predicting different responses mentioned
above are given below.
Penetration (P) = {+ 9.50 + 1.24I - 0.29S - 1.0H - 0.14IS - 0.39IH +
0.36SH + 0.69I2 - 0.069S2 - 0.39H2} mm
Width (W) = {+17.67 + 0.78I - 0.70S + 0.27H - 0.087IS - 0.037IH -
0.087SH + 0.51I2 + 0.036S2 - 0.017H2} mm
Reinforcement height (R) = {+2.74 + 0.046I - 0.28S + 0.30H +
0.063IS - 0.012IH + 0.063SH - 0.28I2 + 0.19S2 + 0.035H2} mm
3. Analysis of Variance (ANOVA) for Robotic GMA
welding parameters
The following Table-6 represents the ANOVA results for
penetration, bead width, and reinforcement height.
Table
6 Analysis of Variance (ANOVA) Test Results
Penetration (P)
Bead Width (W)
Reinforcement Height (R)
Source
F Value
P Value
Prob>F
F Value
P Value
Prob>F
F Value
P Value
Prob>F
Model
29.9253
15.6175
4.9129
0.0102
I
120.910
58.9286
0.3010
0.5953
S
6.843
0.0258
46.5759
10.8086
0.0082
H
79.634
6.8518
0.0257
12.5745
0.0053
IS
0.8735
0.3720
0.4313
0.5262
0.3264
0.5804
IH
6.9376
0.0250
0.0792
0.7841
0.0131
0.9113
SH
6.0713
0.0334
0.4313
0.5262
0.3264
0.5804
I2
39.419
26.7844
0.0004
12.0763
0.0060
S2
0.4018
0.5404
0.1349
0.7210
5.6701
0.0385
H2
12.509
0.0054
0.0279
0.8708
0.1841
0.6770
R2
96%
-
93%
-
81.5%
-
Adj.R2
93%
-
87%
-
65%
-
Pred.R2
70.5%
-
48%
-
-39.7%
-
Model
Significant
-
Significant
-
Significant
-
The analysis of variance for the robotic arm arc welding process
parameter has done to find out the significant mains and
interaction of process parameters. From the results, it is observed
that the welding current is the most significant parameter of
having high F-value. It means that welding current is the most
effective in controlling penetration and similarly, the weld bead
width is mostly affected by the welding current and welding
speed. This may be due to the change in the metal deposition
rate. From the results, it is also observed that the land height
(H) is the most significant parameter which is having higher F
value. It means that land height is the most effective parameter in controlling the dilution. The most effective parameter of the
designed matrix can be easily understood from the perturbation
graphs as shown below in Figure 2.
Where, A = Welding current (I) in Amps; B = Welding speed (S) in
mm/min; D = Land height (H) in mm
4. Effect of process parameters on bead geometry
• Effect of Welding current (I)
The table and Figure 3 illustrate how welding current (I) affects
weld bead geometry penetration (P), bead width (W), and
reinforcement height (R). As shown, penetration and bead
width increase by 55.5% and 25%, respectively, likely due to
higher heat input and metal deposition with increased current.
Reinforcement height may decrease as deeper arcs cause
molten metal to spread sideways in the weld pool [10].
Fig. 3 Effect of welding current on weld bead geometry
• Effect of Welding speed (S)
The table and Figure 4 show the effect of welding speed (S)
on weld bead geometry penetration (P), bead width (W), and
reinforcement height (R). As welding speed increases from
350 to 450 mm/min, reinforcement height drops by 40.47%,
while bead width and penetration decrease by 17.5% and 15%,
respectively. These reductions are likely due to decreased metal
deposition rate at higher welding speeds.
Fig. 4 Effect of welding speed on weld bead geometry
• Effect of Land height (H)
Figure 5 illustrates how land height (H) influences weld
bead geometry, including penetration (P), bead width (W),
and reinforcement height (R). An increase in land height
significantly raises the reinforcement height by 63.6%, likely
due to a larger groove volume. However, penetration decreases
by around 29.6%, possibly because the heat input is insufficient
to fully fuse the increased land height with the deposited metal.
As a result, the molten metal, while filling the groove, fails to
entirely bond with the base metal and instead spreads outward,
increasing the overall bead width across the surface
Fig. 5 Effect of land height on weld bead geometry
• Microstructure
The Robotic GMA welded specimen of modified SAE 1022
low carbon steel joints were sectioned to the required size
from the welded joints transverse to the welding direction
and taken for metallographic examination. These specimens
were cut transversely as required size and macro - polished
with different grades (100, 320,400 and 600, 800 and 1000
grit size) of emery papers. Final polishing was done using the
alumina compound (3μm and 0.1μm particle size) on the disc
polishing machine for 5 minutes and then etched with a Nital
with 5 ml Nitric acid solution applied for 10 to 30 seconds as
per the ASTM specification E-381. The macrostructure of the
welds was studied and macro images are recorded using stereo
zoom macroscope as shown in figure 6. The microstructures
of various weld regions for optimized highest penetration are
recorded at 200X magnification using an optical microscope as
shown in figure 6.
Fig. 6 Macro and microstructure of R-GMA Welded
specimens on highest Penetration
CONCLUSIONS
In this paper, the effect of the process parameters for Robotic
Arm Gas Metal Arc Welding in the prediction of bead geometry
has been reported.
• Experimental results have shown that the process
parameters such as the welding current having more effect
on penetration i.e. the penetration increase with welding
current by 55.5% and bead width increased by 25 %
simultaneously.
• The results show that on increasing the welding speed,
decreases reinforcement height (R) by 40.47 % and
on increasing the land height, there is an increase in
reinforcement height (R) by 63.6% i.e. the land height having
a most significant effect on reinforcement height (R).
REFERENCES
1. I.S. Kim, K.J. Son, Y.S. Yang, and P.K.D.V. Yaragada, Sensitivity
Analysis for Process Parameters in GMA Welding Processes
Using a Factorial Design Method, Int. J. Mach. Tools Manuf.
43 (2003)763–769.
2. I. Kim, J. Son, S. Lee, and P.K.D. V Yarlagadda, Optimal Design
of Neural Networks for Control in Robotic Arc Welding,
Robot CIM-Int Manuf 20 (2004) 57–63.
3. D. Ramos-jaime, I. López-juárez, and P. Perez, Effect
of Process Parameters on Robotic GMAW Bead Area
Estimation, J. Mater. Process. Tech. 7 (2013) 398–405.
4. I.S. Kim, J.S. Son, I.G. Kim, J.Y. Kim, O.S. Kim, A study
on relationship between process variables and bead
penetration for robotic CO2 arc welding, J. Mater. Process.
Tech 136 (2003) 139–145.
5. S. Srivastava and R.K. Garg, Process Parameter
Optimization of Gas Metal Arc Welding on IS: 2062 Mild
Steel Using Response Surface Methodology, J. Manuf.
Process. 25 (2017) 296–305,
6. M.A. Wahab, M.J. Painter, and M.H. Davies, The Prediction
of the Temperature Distribution and Weld Pool Geometry in
the Gas Metal Arc Welding Process, J. Mater. Process. Tech
77 (1998) 233–239.
7. J. Xiong, G. Zhang, H. Gao, and L. Wu, Robotics and
Computer-Integrated Manufacturing Modeling of Bead
Section Profile and Overlapping Beads with Experimental
Validation for Robotic GMAW-Based Rapid Manufacturing,
Robot. Comput. Integr. Manuf. 29 (2013) 417–423.
8. M. Chaisri and P. Peasura, Effect of Gas Metal Arc Welding
Parameters on Mechanical Properties for Carbon Steel
ASTM A285 Grade A, Adv. Mat. Res. 342 (2012)16–20.
9. I.A. Ibrahim, S.A. Mohamat, A. Amir, and A. Ghalib, The Effect
of Gas Metal Arc Welding (GMAW) Processes on Different
Welding Parameters, Procedia Eng 41 (2012) 1502–1506.
10. Z. Yang and T. Debroy, Modeling Macro-and
Microstructures of Gas-Metal-Arc, Metall Mater Trans B 30
(1999) 483-493.
.png&w=3840&q=75)
