J. Mater. Sci. Technol., Vol.23 No.

6, 2007 817

Effect of Welding Parameters on Dilution and Weld Bead

Geometry in Cladding
M.Nouri, A.Abdollah-zadeh† and F.Malek
Department of Materials Engineering, Tarbiat Modares University, P.O. Box 14115-143, Tehran, Iran
[Manuscript received January 9, 2007, in revised form June 9, 2007]

The effect of pulsed gas metal arc welding (GMAW) variables on the dilution and weld bead geometry in
cladding X65 pipeline steel with 316L stainless steel was studied. Using a full factorial method, a series of
experiments were carried out to know the effect of wire feed rate, welding speed, distance between gas nozzle
and plate, and the vertical angle of welding on dilution and weld bead geometry. The findings indicate that
the dilution of weld metal and its dimension i.e. width, height and depth increase with the feed rate, but
the contact angle of the bead decreases first and then increases. Meantime, welding speed has an opposite
effect except for dilution. There is an interaction effect between welding parameters at the contact angle. The
results also show forehand welding or decreasing electrode extension decrease the angle of contact. Finally,
a mathematical model is contrived to highlight the relationship between welding variables with dilution and
weld bead geometry.
KEY WORDS: Pulsed gas metal arc welding (GMAW); Cladding; Weld Bead; Dilution;
Full factorial design

1. Introduction was argon+2.5% CO2 at a flow rate of 15 L/min.

To increase corrosion resistance, a layer of alloy
is placed on the base metal during weld cladding[1] .
Submerged arc welding (SAW)[2,3] , laser welding[4] ,
electroslag welding[5] , gas tungsten arc welding
(GTAW)[6] and gas metal arc welding (GMAW) are
some of the processes used for cladding.
In GMAW process, pulsed current makes it possi-
ble to have spray transfer on average currents, lower
than the simple GMAW. Therefore, dilution and spat-
ter will be lower with the possible positional welding.
A successful cladding needs a weld bead with a
good profile and minimum dilution. By this way, a
thorough understanding is essential as to how weld-
ing variables affect bead geometry and dilution[7–9] .
It will be of immense help to have a mathematical
model to predict welding variables effects, especially
in the automated welding.
The present paper has attempted to study the ef-
fect of pulsed-GMAW parameters on bead geometry
and dilution of weld metal. To obtain apparent re-
sults about main and interaction effects of variables,
systematic experiments have been conducted in ac-
cordance with a full factorial design method. These
results would be helpful in developing mathematical
models of welding variables effects.
2. Experimental
A 316L stainless steel wire with 1.2 mm in di-
ameter was used as surfacing material. The base
metal specimens were 15 mm thick by 100 mm wide
by 200 mm long plates of X65 pipeline steel. Ta-
ble 1 indicates the chemical analysis of these materi-
als. Pulsed synergic GMAW process was applied to
clad the surfacing material onto base metal. The ex-
perimental set-up is shown in Fig.1. The shielding gas
† Prof., Ph.D., to whom correspondence should be addressed,
E-mail: zadeh@modares.ac.ir.
Apart from using electrode positive polarity, it
adopted String technique for the bead on plate weld-
ing, as well.
The whole experiment was accomplished in 3
blocks where the welding parameters were wire feed
rate and welding speed, which are listed in Table 2.
Each of these parameters was set at 3 levels, so total
welding state in each of the blocks was 9. To study the
stability of the parameters, consistency of the results
as well as the interaction effects between parameters,
each of the weld conditions were repeated twice, just
in the case of block 1. Consequently, testing speci-
mens in block 1 became 18. In addition to wire feed
rate and welding speed, two other parameters were
studied as “block parameters”, which are distance be-
tween gas nozzle and plate, and the vertical angle of
the welding (see Table 3). Samples were sectioned,
polished and etched in 2% nital. Digital photographs
were taken from the polished surface of the samples
and then bead width (w), bead height (h), depth of
penetration (t), angle of contact (Z) and dilution (D)
were calculated by using computer software. Figure 2
shows an example of the macrostructure of weld pro-
file. The schematic illustration of bead geometry and
dilution is also shown in this figure.
3. Results and Discussion
Tables 4–6 show the measured values of weld bead
parameters in 3 blocks. A thorough look on these
tables can easily certain the graphical relations be-
tween welding parameters and measured parameters.
Again, Figs.3–7 show the main effects of wire feed
rate whereas, and Figs.8–12 indicate the main effects
of the travel speed on bead geometry and dilution.
According to Fig.3, width (w) increases with in-
creasing wire feed rate from 4.5 to 8 m/min in all 3
blocks. Higher wire feed rate level causes additional
molten wire in welding pool leading to the bigger bead
geometry. In Fig.8, width decreases with increasing
818 J. Mater. Sci. Technol., Vol.23 No.6, 2007

Table 1 Chemical composition of welding wire and base metal (wt pct)
C Si Mn P S Cr Mo Ni Al Co Cu
316L steel 0.035 0.42 1.60 0.014 0.012 19.73 2.83 11.80 0.01 0.07 0.14
X65 steel 0.06 0.32 1.50 0.009 0.004 0.004 0.12 0.005 0.03 0.005 0.14
Nb Ti V W Sn B As Zr Pb Fe
316L steel 0.48 0.005 0.08 0.03 – – – – – balance
X65 steel 0.05 0.02 0.04 <0.015 <0.002 – 0.005 0.003 – balance

Fig.1 Experimental set-up used for cladding

Table 2 Welding parameters in blocks and their levels

Level Wire feed rate/(m/min) Travel speed/(cm/min)
1 (low) 4.5 30
2 (medium) 6.25 50
3 (high) 8 70

Table 3 Block parameters and their levels

Block Distance between gas nozzle and plate/mm Angle between gas nozzle and plate/deg.
1 10 0
2 5 0
3 10 20 (forehand)

Table 4 Measured values of bead dimension and dilution in block 1

No. Feed rate(m/min) Travel speed/(cm/min) w/mm h/mm t/mm D/% Z/deg.
1 4.5 30 7.25 2.84 1.56 21.24 64
2 4.5 30 7.49 2.88 1.52 21.69 57
3 4.5 50 5.74 2.22 1.28 25.25 56
4 4.5 50 5.56 2.25 1.30 23.74 56
5 4.5 70 5.41 2.14 1.02 23.02 54
6 4.5 70 4.41 2.22 1.09 23.85 46
7 6.25 30 9.98 3.19 2.59 26.03 55
8 6.25 30 9.65 3.06 2.43 25.04 51
9 6.25 50 6.96 2.54 1.85 27.44 55
10 6.25 50 7.04 2.45 2.10 29.31 45
11 6.25 70 6.11 2.17 1.49 28.76 39
12 6.25 70 6.37 2.30 1.61 30.70 48
13 8 30 11.63 3.52 3.02 27.17 52
14 8 30 10.69 3.58 3.30 26.46 75
15 8 50 8.13 2.43 2.23 30.26 55
16 8 50 8.87 2.87 2.67 30.85 51
17 8 70 7.29 2.56 2.17 34.78 52
18 8 70 8.09 2.86 2.22 29.74 47
 J. Mater. Sci. Technol., Vol.23 No.6, 2007 819
Table 5 Measured values of bead dimension and dilution in block 2
No. Feed rate/(m/min) Travel speed/(cm/min) w/mm h/mm t/mm D/% Z/deg.
1 4.5 30 7.51 2.62 1.48 25.35 59
2 4.5 50 6.98 2.23 1.11 23.99 46
3 4.5 70 6.06 1.66 1.17 32.88 49
4 6.25 30 11.56 2.94 2.90 34.49 46
5 6.25 50 8.55 2.19 2.14 33.88 45
6 6.25 70 6.99 1.84 1.61 35.67 38
7 8 30 11.72 3.13 2.86 33.97 54
8 8 50 9.23 2.42 2.53 36.62 41
9 8 70 8.06 2.12 2.40 38.05 43

Table 6 Measured values of bead dimension and dilution in block 3

No. Feed rate/(m/min) Travel speed/(cm/min) w/mm h/mm t/mm D/% Z/deg.
1 4.5 30 8.39 2.63 1.53 22.09 60
2 4.5 50 6.66 2.02 1.21 23.96 39
3 4.5 70 5.03 1.83 1.05 23.99 50
4 6.25 30 9.44 3.25 1.96 27.12 58
5 6.25 50 8.67 2.32 1.87 29.17 50
6 6.25 70 7.58 2.08 1.36 28.66 31
7 8 30 11.31 3.26 2.86 27.67 74
8 8 50 10.67 2.47 2.42 30.57 40
9 8 70 9.16 2.35 2.03 29.76 34

Fig.4 Effect of wire feed rate on means of h in 3 blocks

Fig.2 (a) Weld profile macrostructure of sample 18 in

Table 4, (b) schematic illustration of bead geom-
etry and dilution

Fig.5 Effect of wire feed rate on means of t in 3 blocks

distance here means the decrease in electrode exten-

Fig.3 Effect of wire feed rate on means of w in 3 blocks
welding speed from 30 to 70 cm/min. Faster welding
speed causes fewer melted wire deposited in the unit
length and therefore smaller bead geometry. In both
cases, width of bead increases with decreasing dis-
tance between gas nozzle and plate. The decrease in
sion. This finding is in accordance with some previous
studies[2,3,10] . Meantime, it is observed in the process
that bead width increases in block 3 compared with
block 1. In other words, forehand welding increases
bead width.
Figures 4 and 5 show the increase in bead height
(h) and depth of penetration (t) in all 3 blocks, with
the increase in wire feed rate from 4.5 to 8 m/min.
820 J. Mater. Sci. Technol., Vol.23 No.6, 2007

Fig.6 Effect of wire feed rate on means of D in 3 blocks Fig.7 Effect of wire feed rate on means of Z in 3 blocks

Fig.8 Effect of welding travel speed on means of w in 3 Fig.9 Effect of welding travel speed on means of h in 3
blocks blocks

Fig.10 Effect of welding travel speed on means of t in 3 Fig.11 Effect of welding travel speed on means of D in 3
blocks blocks

Fig.12 Effect of welding travel speed on means of Z in 3 Fig.13 Interaction effect of wire feed rate and travel speed
blocks on w
 J. Mater. Sci. Technol., Vol.23 No.6, 2007
Fig.14 Interaction effect of wire feed rate and travel speed
on D
821
tact. In Fig.13, it is clear that width has the same
trends by changes in welding speed in different lev-
els of feed rate. In the other words, all three lines
have approximately similar shape. Also, distance be-
tween two lines is almost constant in the course of the
length. Therefore, no interaction effects are seen be-
tween welding parameters, in the case of width. Fig-
ure 14 indicates that interaction effects on D. In this
graph, the three rows are approximately similar and
therefore, there seem no interaction effects in this sit-
uation either. Meanwhile, there appear to be different
levels of feed rate in Fig.15, and thus, Z has different
trends with increasing welding speed. It is evident
that three graph lines have different trends and dis-
tance between each of two lines is up-and-down at
different points.
Each of the weld bead parameters (w, h, t, Z and
D) can be expressed as an mathematical function of
f (wire feed rate) and s (welding speed)[3,10–12] :

y = θ(f, s)

For this purpose, the quadratic regression model was

used. They can be expressed thus:

Y = β0 + β1 f + β2 s + β11 f 2 + β22 s2 + β12 f s

where β0 is a free term of the regression equation, β1

and β2 are linear term, β11 and β22 are the quadratic
terms and, β12 is the interaction term.
The measured values of weld bead parameters are
Fig.15 Interaction effect of wire feed rate and travel speed
on Z
listed in Tables 4–6. From these tables, mathemat-
ical models were developed for three blocks through
computer calculation as follows:
The behavior of h and t against welding travel speed Block 1:
is similar to w (see Figs.9 and 10). According to these
figures, bead height decreases in block 2 and 3 com- w = 5.48 + 0.90f − 0.08s − 0.14(0.57f − 3.57)2 +
pared with block 1. Therefore, it can be said that at ³s
a particular feed rate and travel speed, melting metal 5 ´2 ³s 5´
0.81 − − 0.25(0.57f − 3.57) −
is constant in all 3 blocks and as such, width increases 20 2 20 2
with decreasing the height. Meanwhile, the depth h = 2.44 + 0.16f − 0.02s + 0.08(0.57f − 3.57)2 +
penetration behavior is different in all the blocks. The ³s
depth in block 2 is more than those of block 1 but at 5 ´2 ³s 5´
0.32 − − 0.04(0.57f − 3.57) −
the same time block 3 has less depth. Thus, a fore- 20 2 20 2
hand welding has a wider bead with less depth com- t = 0.63 + 0.37f − 0.02s − 0.06(0.57f − 3.57)2 +
pared with block 1. ³s
According to Figs.6 and 11, dilution (D) increases 5 ´2 ³s 5´
0.10 − − 0.12(0.57f − 3.57) −
with increasing both wire feed rate and welding speed. 20 2 20 2
In both situations, it seems that quantity of heat that D = 11.81 + 1.92f + 0.10s − 1.38(0.57f − 3.57)2 −
affected the base metal or in the other words, arc ³s
power increases leading to the melting of more base 5 ´2 ³s 5´
1.27 − + 0.87(0.57f − 3.57) −
metals. While comparing the variations of dilution in 20 2 20 2
3 experimental blocks, it is clear that blocks 1 and 3 Z = 63.06 + 0.04f − 0.28s + 6.58(0.57f − 3.57)2 +
roughly have the same dilution but dilution in block ³s
2 encounters different situations. Welding specimens 5 ´2 ³s 5´
0.33 − − 0.88(0.57f − 3.57) −
show that block 2 has more dilution than the other 20 2 20 2
in average about 6%. However, electrode extension in Block 2:
this particular block is lesser than blocks 1 and 3.
Figures 7 and 12 indicate angle of contact (Z) that w = 4.24 + 0.80f + 0.03s − 0.77(0.57f − 3.57)2 −
decreases with decreasing welding speed, but in feed ³s
rate, firstly it decreases to a minimum value and then 5 ´2 ³s 5´
2.62 − + 0.06(0.57f − 3.57) −
increases once again. Comparisons of the two above 20 2 20 2
cases show that specimens in block 1 have greater an-
gle of contacts than those in blocks 2 and 3 in average
h = 1.67 + 0.11f + 0.01s + 0.04(0.57f − 3.57)2 −
about 7◦ .
Figures 13, 14 and 15 show the interaction effects ³s 5 ´2 ³s 5´
of welding parameters on w, dilution and angle of con- 0.82 − − 0.07(0.57f − 3.57) −
20 2 20 2
822 J. Mater. Sci. Technol., Vol.23 No.6, 2007

t = −0.04 + 0.38f + 0.005s − 0.29(0.57f − 3.57)2 − (2) Width, height, depth and the contact angle
³s of weld metal decrease with increasing welding speed
5 ´2 ³s 5´
0.59 − + 0.05(0.57f − 3.57) − from 30 to 70 cm/min, but dilution increases.
20 2 20 2 (3) Wire feed rate and welding speed appear to be
D = 22.53 + 2.51f − 0.10s − 2.87(0.57f − 3.57)2 + have interaction effects on the angle of contact.
³s (4) Dilution increases in average about 6% and
5 ´2 ³s 5´
2.24 − + 1.87(0.57f − 3.57) − angle of contact decreases in average about 7◦ with
20 2 20 2 decreasing the electrode extension. It seems that fore-
Z = 57.91 − 1.52f + 0.02s + 5.67(0.57f − 3.57)2 − hand welding has no apparent effect on dilution but
³s reduces the angle of contact in average about 6◦ .
5 ´2 ³s 5´
9.33 − + 0.25(0.57f − 3.57) −
20 2 20 2
Block 3: Acknowledgements
The authors would like to express their gratitude for
w = 5.18 + 1.05f − 0.06s − 0.03(0.57f − 3.57)2 − the financial support of this research by the Pars Oil & Gas
³s Company. The authors are also thankful to Iran Technical
5 ´2 ³s 5´ & Vocational Training Organization, Instructor Training
0.18 − + 0.30(0.57f − 3.57) −
20 2 20 2 Center (ICT) of Karaj and Sadid Pipe and Equipments
Company for their support.
h = 2.60 + 0.15f − 0.02s − 0.12(0.57f − 3.57)2 +
³s 5 ´2 ³s 5´
0.30 − − 0.03(0.57f − 3.57) − REFERENCES
20 2 20 2
t = 0.45 + 0.33f − 0.01s + 0.12(0.57f − 3.57)2 − [1 ] W.Lucas: Weld. Metal. Fabr., 1994, 62(2), 6.
[2 ] U.D.Mallya and H.S.Srinivas: Weld. J., 1989, 68(12),
³s 5 ´2 ³s 5´ 30.
0.03 − − 0.09(0.57f − 3.57) − [3 ] N.Murugan, R.S.Parmar and S.K.Sud: J. Mater. Pro-
20 2 20 2
cess. Technol., 1993, 37(1-4), 767.
D = 16.21 + 1.71f + 0.05s − 1.98(0.57f − 3.57)2 − [4 ] H.Gedda: Ph.D. Thesis, Lulea University of Technol-
³s 5 ´2 ³s 5´ ogy, 2004.
1.35 − + 0.05(0.57f − 3.57) − [5 ] Y.K.Oh, J.H.Devletrian and S.J.Chen: Weld. J.,
20 2 20 2 1990, 69(8), 37.
Z = 73.51 − 0.10f − 0.64s + 3.17(0.57f − 3.57)2 + [6 ] C.S.Dimbylow and K.M.Chipperfield: Weld. Met.
³s Fabr., 1983, 51(5), 229.
5 ´2 ³s 5´ [7 ] P.K.Ghosh, P.C.Gupta and V.K.Goyal: Weld. J.,
8.17 − − 7.5(0.57f − 3.57) −
20 2 20 2 1998, 77(7), 307s.
[8 ] M.Amin: Metal Const., 1981, 13(6), 349.
[9 ] W.Lucas: FWP J., 1985, 25(6), 7.
4. Conclusions [10] N.Murugan and P.K.Palan: J. Inst. Eng., 2004,
84(March), 49.
(1) Width, height, depth and dilution of weld [11] T.Kannan and N.Murugan: J. Mater. Process. Tech-
metal increase with increasing the wire feed rate from nol., 2006, 176(1-3), 230.
4.5 to 8 m/min, but the contact angle decreases first [12] N.Murugan and V.Gunaraj: J. Mater. Process. Tech-
and again increases after reaching a minimum value. nol., 2005, 168(3), 478.