O
Journalof ManufacturingProcesses
Vol. 8/No. 2
2006
Sensitivity Analysis for Process Parameters in
Cladding of Stainless Steel by Flux Cored
Arc Welding
P.K. Palani, Faculty of Mechanical Engineering, Government College of Technology, Coimbatore, India.
E-mail: pkpalaniku @yahoo.com
N. Murugan, Professor of Mechanical Engineering, Coimbatore Institute of Technology, Coimbatore, India.
E-mail: drmurugan @yahoo.com
Abstract
by the process parameters applied during cladding.
Also, engineering components in many industrial
applications are subjected to wear and corrosion,
which dictate frequent maintenance and jeopardize
reliability. The replacement cost of many of these
components is extremely high; consequently, extension of service life can result in significant savings
(Alam et al. 2002; Heston 2000; Missori, Murdolo,
and Sili 2004).
Cladding is a process of depositing a relatively
thick layer of filler material on a carbon or low alloy
steel base metal (Alam et al. 2002; Murugan and
Panner 1995). It is possible to achieve high economic
gains by fabricating components from the stainless
steel surfaced low carbon steel for use in important
applications in chemical, fertilizer, thermal, and
nuclear power industries. Based on these considerations, low carbon structural steel (IS: 2062) is
cladded with austenitic stainless steel (317L). The
former is extensively employed as a construction
material in nuclear power plants and fertilizer industries and the latter for depositing buffer layers
(Murugan 1993; Stewart 1981). Rolling, explosive
welding, or fusion welding are commonly employed
for cladding. Fusion welding is readily accepted by
the engineering industry owing to its easy and versatile application and no legal implications of safety,
pollution, and noise (Rajasekaran 2000). Among the
fusion welding processes, flux cored arc welding
(FCAW) has been widely used for cladding due to
several advantages, such as high deposition rate,
high-quality weld metal deposits, low fume generation, excellent weld appearance (smooth, uniform
welds and excellent contour of horizontal fillet
Austenitic stainless steel cladding is generally used to
attain better corrosion resistance properties to meet the requirements of petrochemical, marine, and nuclear applications. The quality of cladded components depends on the
weld bead geometry and dilution, which in turn are controlled
by the process parameters. In this investigation, the effect of
cladding parameters such as welding current, welding speed,
and nozzle-to-plate distance on the weld bead geometry was
evaluated. The objective of controlling the weld bead geometry can easily be achieved by developing equations to predict these weld bead dimensions in terms of the process
parameters. Mathematical equations were developed by using the data obtained by conducting three-factor five-level
factorial experiments. The experiments were conducted for
317L flux cored stainless steel wire of size 1.2 mm diameter
with IS:2062 structural steel as a base plate. Sensitivity analysis was performed to identify the process parameters exerting the most influence on the bead geometry and to know the
parameters that must be most carefully controlled. Studies
reveal that a change in process parameters affects the bead
width, dilution, area of penetration, and coefficient of internal
shape more strongly than it affects the penetration, reinforcement, and coefficient of external shape.
Keyword$: Cladding, Flux Cored Arc Welding, Weld Bead
Parameters, Coefficient Weld Shapes, Response Surface
Methodology, Sensitivity Analysis
Introduction
In many critical industries such as the power and
petrochemical industries, bi-layer components in the
form of cladded plates are used due to their superior
environmental/mechanical properties (Khodadad
Motarjemi and Kocak 2001). The mechanical and
metallurgical characteristics of these corrosion-resistant layers are not only controlled by the chemistry of the stainless steel wire but to a greater extent
90
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
the process variables (Kim et al. 2003; McGlone
1982; Allen et al. 2002). The Welding Institute and
Chandel and Bala (1986) pioneered in attempting
these types of modeling. The results show that the
mathematical models so derived from experimental
results can be used to predict the bead geometry
(Kim, Son, and Jeung 2001; Kim et al. 2003).
Also, it has been proved by several researchers
that efficient use of statistical design of experimental techniques allows the development of an empirical methodology, which incorporates a scientific
approach in establishing a welding procedure (Kim
et al. 2003; Allen et al. 2002; Chandel and Bala 1986;
Marimuthu and Murugan 2005; Subramaniam et al.
1999; Murugan and Parmer 1994).
In this work, investigations were carried out to
study the effect of the process parameters on bead
formation and their sensitivity. The qualitative and
quantitative effectiveness of process parameters can
be determined using sensitivity analysis. By this
analysis, critical parameters can be identified and
ranked by their order of importance. This will help
plant engineers to select the process parameters efficiently and to control the bead geometry effectively
without much trial and error, resulting in savings of
time and materials.
The study was carried out in two steps. In the first
step, experiments were conducted with different process parameters using design of experiments to develop statistical models for the prediction of weld
bead geometry'. In the second step, sensitivity analysis was carried out based on the empirical equations
developed.
The chemical compositions of the low carbon
smactural steel, IS: 2062, substrate and the austenitic stainless steel type, AISI 317L, filler material used
in this study are given in Table 1.
welds), relatively high electrode metal utilization,
relatively high travel speeds, gasless variations that
can be used outdoors, the possibility of welding in
all positions, and reduced distortion compared to
shielded metal arc welding (Cary 2002; Raja, Rohh-a,
and Samidas 1999; Sakai et al. 1989; Cornu 1988).
Principal applications of FCAW include steel fabrication, public works (e.g. bridges), naval works,
boiler making, tube/pipe welding, heterogeneous
assemblies, and so on.
The selection of the welding procedure must be
specific to ensure that an adequate clad quality is
obtained (Kim, Son, and Jeung 2001). Further, it is
essential to have complete control over the relevant
process parameters to obtain the required bead geometry (Figure 1) and shape relationships on which
the integrity of a weldment is based (Chandel and
Bala 1986). It has also been reported by some researchers that in FCAW process quality can be represented by the bead shape, and the weld pool
geometry plays an important role in determining the
mechanical properties of the weld (Kang et ai. 2003;
Kim, Rhee, and Park 2002: Chen et al. 2000: Juang
and Tarng 2002). Therefore, it is very important to
select and control the welding process parameters to
obtain optimal clad geometry.
Numerous attempts have been made to develop
mathematical models relating the process variables
and clad geometry for the selection and control of
Width (W)
Reinforcement (R)
Penetration (P)
- - ~ ' -
t
Aera A is added metal
Area B is base metal melted
% Dilution = [B/(A+B)] x 100
Experimental Procedure
The independently controllable process parameters
were identified. They are welding current (/), weld-
Figure 1
Weld Bead Geometry
Table 1
Chemical Composition of Materials Used
S.
No.
1
2
Materials
Used
317L
(flux coredwire)
IS: 2062
Mn
1.38
Element,%weight
P
S
Cr
Mo
0.016 0.007 18.46 3.18
C
0.021
Si
0.89
0.180
0.180 0.980 0.016 0 . 0 1 6
91
.
.
Ni
13.10
.
.
N~_
Cu
0.057 0.007
.
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
Table 2
Process Variables and Their Bounds
Process
Variables
Welding
current
Welding
speed
Nozzle-toplate
distance
Units
Notation
amps
I
cm/min.
S
mm
N
FILLERW I R E ~
Factor Levels
-1.682 - I
0
1 1.682
176 190 210 230 244
26
15
29
17
34
20
39
23
WIRE
COOLINGWATER
42
SHIELDING
25
!
TORCH
WELDCONTROLLER
~]
-
CURRENT
REGULATOR
c I
Figure 2
Schematic of Experimental Setup for FCAW Process
ing speed (S), and nozzle-to-plate distance (N). It
was found that the wire feed rate is directly proportional to the welding current. The relation is found
to be W/= -6.92 + 0.0860"/, where I is the welding
current in amps. Wf is the wire feed rate in m/min.
and hence was treated as a dependent variable.
The working range was decided by conducting
trial runs and by inspecting the bead for smooth appearance and the absence of any visible defects. For
deciding the working range, several trial welds were
made. For determining the range of one variable,
the other two variables were kept constant during
trial runs. For example, to find the working range
for the welding current, the welding speed and
nozzle-to-plate distance were kept initially constant,
and the current was varied from the lower value to
higher values. The beads were inspected for smooth
appearance and absence of ally visible defects.
All of the variables, notation, and units used in
this paper are shown in the Appendix.
A similar procedure was adopted for determining
the upper and lower limits for the welding speed and
nozzle-to-plate distance. Also, trial welds were made,
keeping the values of all the parameters both at their
minimum and maximum values to verify quality of
the weld bead.
After determining the working range of the process parameters, the upper limit was coded as +1.682
and the lower limit as -1.682. The coded values of
the intermediate levels were calculated from the relationship Xi = 1.682"[2X - (Xm~ + X,~)] / (Xm~, Xm~), where X; is the required coded value of a variable X, and X is any value of the variable from Xmt,
to Xm~x.Xmmis the lower level of the variable; Xm~,is
the upper level of the variable. The selected values
of the process parameters together with their units
and notations are given in Table 2.
Experiments were carried out using a Unimacro
Esseti 501 Synergic MIG welding machine available
at the C o i m b a t o r e Institute of T e c h n o l o g y ,
Coimbatore, India. Twenty experimental runs were
conducted as per the central composite rotatable
design matrix at random to avoid any systematic error creeping into the system. A single bead of 150
mm length was laid on structural steel plates using
317L stainless steel flux cored wire (AWS: A5-2295; EN 12073) of 1.2 mm diameter under a shield of
95% Ar and 5% CO2 gas mixture supplied at the rate
of 16 L/min. A DCEP with electrode-to-work angle
of 90 ° was maintained throughout the study. The
schematic experimental setup is shown in Figure 2.
The cladded plates were cross sectioned at their
mid-points to obtain test specimens of 25 mm wide.
These specimens were ground, polished, and etched
with 2% nital. The weld bead profiles were traced
by using an optical profile projector, and the bead
dimensions viz. width (W), penetration (P), and reinforcement (R) were measured. With the help of a
digital planometer (Super PLANIX o~ by Tamaya
Technics Inc.), the areas of the parent metal melted
(AP) and the metal forming the reinforcement were
measured, and percent dilution (D) was calculated.
Coefficients of shape of welds (coefficient of external shape, q0e = W/R; coefficient of internal shape, q~o
= W/P) were also determined (Cornu 1988). Figures
3 and 4 show the typical weld bead cross sections
and weld bead geometry traces, respectively.
Development of Mathematical Models
The response function representing any of the weld
bead dimensions can be expressed as (Murngan and
Parmer 1994, 1995; Cochran and Cox 1957; Gunaraj
and Murugan 1999; Montgomery and Runger 1999;
Walpole, Myers, and Myers 1998; Cheremisinoff and
Ferrante 1987; Khuri and Cornell 1996; Montgomery 2001):
92
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
(a)
(b)
Figure 3
Typical Weld Bead: (a) Specimen for experimental runs 1--4, (b) Cross sections for experimental runs 4, 6, a n d 7.
/~-'~"'~'"
MIO
"
/
...... k ..... . . _ . . j . . _ . L _ _ . .
Figure 4
Ty'pical Weld Bead Geometry Traces (magnification: ×10)
Y =f(I,S,N)
where b 0 is the free term of the regression equation, the coefficients bl, b2..... b~ are linear terms,
the coefficients bjj, b22 . . . . . bt~ are the quadratic
terms, and the coefficients bl2 , b,3 ..... b~._l~are the
interaction terms.
For three factors, the selected polynomial could
be expressed as given below:
(1)
where
Y
1
S
N
is
is
is
is
the
the
the
the
response (penetration, bead width, etc.)
welding current, amps
welding speed, cm/min.
nozzle-to-plate distance, mm
The second-order polynomial (regression) equation used to represent the response surface for K factors is given by
Y = bo + b,I + bzS + b3N +
bulS + bl31N + b23SN +
k
v=b0+
bIiI 2 + b22 $2 + ]933N2
b,x,+
,
E b,jx,L
i,j=l
(3)
~
(2)
The coefficients of the polynomial in Eq. (3) are
calculated using the following formulae with usual
notations (Cochran and Cox 1957):
i=l
93
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
Table 3
ANOVA for the Models Developed
Bead
Geometry
Sum-of-Squares
Regression
Residual
Degrees of
Freedom
Regression
Residual
Mean-Square
Regression
Residual
F-ratio
R2
Adjusted
R2
Remarks
Penetration (P)
0.269
0.014
9
10
0.030
0.0014
20.614
0.95
0.90
Passed 95%
F-ratio test
Reinforcement (R)
1.238
0.148
4
15
0.309
0.0100
31.371
0.89
0.87
Passed 95%
F-ratio test
Bead width (W)
37.644
4.699
5
14
7.529
0.3360
22.433
0.89
0.85
Passed 95%
F-ratio test
% Dilution (D)
34.086
4.496
8
11
4.261
0.4090
10.425
0.88
0.80
Passed 95%
F-ratio test
Area of
penetration (AP)
13.658
2.618
6
13
2.276
0.201
11.304
0.81
0.73
Passed 95%
F-ratio test
Coefficient of
internal shape of
welds (¢p.)
32.754
8.851
6
13
5.459
0.681
8.018
0.79
0.69
Passed 95%
F-ratio test
Coefficient of
external shape of
of welds (,p,)
1.158
0.222
6
13
0.193
0.017
11.300
0.84
0.77
Passed 95%
F-ratio test
Tabulated values of F: F~5' 14.0.05~= 3.02; F<4 ~5,0.05~= 3.06; F<5.J4,0.05~= 2.96; F~s' H, 0.05>= 2.95;
b0 = 0.166338ZY - 0.056791ZZ ( X y )
b~=O.073224Z(Xy)
bii = 0.062500E (XiiY) +
0.006889~Z(Xi, Y ) - 0.056797ZY
:0125000z(x#)
(4)
F r, 13, 0,05) =
2.92.
Checking the Adequacy of the Model Developed
The estimated coefficients obtained above were
used to construct models for the response parameters. The adequacy of the models so developed was
then tested by using the analysis of vmiance technique (ANOVA). Using this technique, it was found
that calculated F ratios were larger than the tabulated values at a 95% confidence level; hence, the
models are considered to be adequate (Ramasamy,
Gould, and Workman 2002).
Two more criterions that are commonly used to illustrate the adequacy of a fitted regression model are
the coefficient of determination (R'). For the models
developed, the calculated R 2 and adjusted R e values
were above 80% and 70%, respectively. These values
indicate that the regression models are quite adequate
(Ramasamy, Gould, and Workman 2002). The results
of the ANOVA are given in Table 3.
The validity of regression models developed were
further tested by drawing scatter diagrams. Typical
scatter diagrams for P and D are shown in Figures 5
and 6, respectively. The observed values and predicted values of the responses are scattered close to
the 45 ° line, indicating an almost perfect fit of the
developed empirical models (Kim et al. 2003).
The final mathematical models, with parameters
in coded form as determined by the above procedure, are presented below:
(5)
(6)
(7)
Using the results obtained from experiments, the
values of the coefficients of the above polynomial
were calculated with the help of commercial statistical software, Systat ®, Version 10.2. As a first step, a
complete model was developed that contained all of
the variables. Then, a 'stepwise' procedure was used
to remove the insignificant variables, one at a time.
Using this procedure, the variables with F values
greater than or equal to the standard tabulated value
are retained in the model and the variables with F
values less than or equal to the standard tabulated
value are removed from the model one at a time automatically. After determining the significant coefficients, the final model was constructed using only
these significant coefficients, without affecting the
accuracy of the model.
94
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
1.2i
11
A
1.1
g
-#
9
==7
~ o.9t
>~
!
~ 0.8
L
~ 0.7
o.ol
3
0.6
0.7
0.8
0.9
1
1.1
4
5
6
7
8
9
10
11
Observed values of % dilution
1.2
Observed Values of PenetraUon (mm)
Figure 6
Scatter Diagram for Percent Dilution lVh)del
Figure 5
Scatter Diagram for Penetration Model
the bead parameters were measured using the same
procedure described in the previous section. The
results obtained were quite satisfacto~" and the details are presented in Tables 4a and 4b.
P = 0.971 + 0.093 * I - 0.062 * S - 0.016 * N O.047* I * I +O.02* S* S-O.039* N * N -
(8)
0.03 * I * S -0.042 * I * N -0.042 * S * N
R = 4.417+0.143"I-0.253"
S+
0.038"N+0.067"S*S
W = 10.531 + 1 . 5 3 9 " I - 0 . 4 6 " S + 0 . 0 6 9 " N +
0.3* N * N-O.357* I* N
Sensitivity Analysis for Bead Geometry
(9)
From the above developed mathematical equations [Eqs. (8)-(14)] to be used for the estimation of
bead geometry, the sensitivity equations are obtained
by differentiating them with respect to process parameters of interest, such as welding current (I),
welding speed (S), and nozzle-to-plate distance (N)
(Kim et al. 2003; Marimuthu and Murugan 2005).
The sensitivity equations for welding current were
obtained by differentiating Eqs. (8)-(14) with respect
to welding current mad are given below.
(lO)
D = 7.533 + 0.039" 1 + 0.001" S + 0.266" N 0 . 2 7 5 * S* S - 0.237" N * N -
(11)
1.22* I* N-O.375* I* S - 1 . 5 1 2 * S* N
AP = 2.956+0.592* I-O.156* S-O.094* N 0 . 7 8 " I * N + 0 . 1 5 3 " S * 1 - 0 . 6 0 4 " S* N
(12)
dP/dI = 0.093 - 0.047 * 2 * I 0.03 * S - 0.042 * N
(p, = 10.534+0.502"1+0.325"S+0.392"N+
0.933* S* N +O.809* I * I +O.869* N * N
(Pe = 2.319+0.269" I +0.032" S - 0 . 0 0 2 " N +
O.053*I*I+O.07*N*N-O.082*I*S
Confirmation
(13)
(15)
dR/dl =0.143
(16)
d W / d l = 1.539-0.357 * N
(17)
dD/dl = 0.039 - 1.22 * N - 0.375 * S
(18)
dAP/dl = 0.592 - 0.78 * N + 0.153 * S
(19)
d(p,/dl = 0.502 + 0.809" 2" I
(20)
d(p~/dl = 0.269 + 0.053" 2" I - 0 . 0 8 2 " S
(21)
(14)
Experiments
Experiments were conducted to verify the above
d e v e l o p e d regression equations [Eqs. (8)-(14)].
Three weld runs were made using different values
of current, welding speed, and nozzle-to-plate distance other than those used in the design matrix, and
95
Journal o f Manufacturing Processes
Vol. 8/No. 2
2006
Table 4a
Results of Confirmation Experiments for P, B~ and R
Expt.
No.
I
(A)
CON 1
CON2
CON3
230
200
220
Parameters
S
N
(cm/min.) (mm)
34
32
45
23
21
18
P
Actual Values
W
R
Predicted Values Using
Regression Model
P
W
R
P
% Error
W
R
5.4
5.3
4.6
-1
1
-4.8
2
-1.1
--0.5
(mm) (mm) (mm) (mm) (mm) (mm)
0.97
0.98
0.87
12.2
10
9.9
4.5
4.45
4.22
0.92
0.93
0.83
12.08
9.9
10.4
4.59
4.5
4.24
Table 4b
Results of Confirmation Experiments for AP, D, ~ , and ~p,
Expt.
No.
CON1
CON2
CON3
D
Actual Values
AP
¢Pa
(%)
(mm2)
6.4
7
8.9
2.8
2.75
4.15
where % error is given by % error =
12.58
10.2
11
%
2.68
2.24
2.3
D
Predicted Values Using
Regression Model
AP
¢p,
(%)
(ram2)
6.3
6.6
9.36
2.67
2.6
4.32
13.1
10.4
10.4
%
D
2.63
2.18
2.5
1.6
6
--4.9
% Error
AP
%,
%
4.9
4.6
3.94
2
2.8
-8
4
2
5
Actual Value -Predicted Value x 100
Predicted Value
The sensitivity equations for welding speed were
obtained by differentiating Eqs. (8)-(14) with respect
to welding speed and are given below.
dW/dN =O.O69+O.3*2*N-0.357* I
dD/dN = 0 . 2 6 6 - 0.237" 2 * N 1.22"I-1.512"S
dP/dS = -0.062 + 0.02 * 2 * S 0.03 * S - 0 . 0 4 2 * N
(31)
(32)
(22)
dAP/dN =-O.O94* N-0.78* I-0.604 * S
(33)
dR/dS = -0.253 +0.067 * 2 * S
(23)
d%/dN ---0.392 + 0.933" S + 0.869" 2 " N
(34)
dW/dS=-0.46
(24)
d%/dN = --0.002 + 0.07 * 2 * N
(35)
dD/dS = 0.001- 0.275 * 2 * S 0.375"I-1.512"N
Sensitivities of welding current, welding speed,
and nozzle-to-plate distance on bead geometry are
presented in Tables 5-.11 and are shown in Figures
7-12. Positive values of sensitivities mean that the
dimension of bead geometry, increases with the corresponding increase in the values of process parameters, and n e g a t i v e v a l u e s m e a n that the b e a d
geometry decreases with the corresponding increase
in the values of process parameters.
(25)
dAP/dS = - 0 . 1 5 6 + 0.153" I - 0.604" N
(26)
d%/dS = 0.325 + 0.933 * N
(27)
dtPe/dS= 0.032-0.082 * I
(28)
The sensitivity equations for no~le-to-plate distance were obtained b y differentiating Eqs. (8)-(14)
with respect to welding speed and are given below.
Sensitivity of Welding Current on
Bead Geometry
dP/dN = -0.016 - 0.039" 2 * N 0.002 * I - 0.042 * S
dR/dN = 0.038
Figure 7 shows the sensitivity of welding current
on bead geometries (viz., P, R, W, D, AP, q%, and %).
O f all the weld bead parameters, the coefficient of
internal shape (%) is more sensitive to the welding
(29)
(30)
96
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
Table 9
Table 5
Depth of Penetration Sensitivities of Process P a r a m e t e r s
Percent Dilution Sensitivities of Process P a r a m e t e r s
Welding Speed (S) = 34 cm/min.;
Nozzle-to-Plate Distance (N) = 20 mm
Welding Speed (S) = 34 cm/min.;
Nozzle-to-Plate Distance (N) = 20 mm
Welding Current (1), amps
176
190
210
230
244
dP/dl
0.2511
0.1870
0.0930
-0.0010
-0.0651
dP/dS
-0.01
-0.03
-0.06
-0.09
-0.11
Welding Current (I), amps
dP/dN
0.0546
0.0260
-0.016
-0.058
-0.087
176
190
210
230
244
dD/dl
0.039
0.039
0.039
0.039
0.039
dD/dS
dD/dN
0.6318
2.3180
0.3760
1.4860
0.0010 0.2660
-0.3740 -0.9540
-0.6298 -1.7860
1"able 6
Table 10
R e i n f o r c e m e n t Sensitivities of Process P a r a m e t e r s
Coefficient of Internal S h a p e Sensitivities of Process P a r a m e t e r s
Welding Speed (S) = 34 cm/min.;
Nozzle-to-Plate Distance (N) = 20 mm
Welding Speed (S) = 34 cm/min.;
Nozzle-to-Plate Distance (N) = 20 mm
Welding Current (1), amps
dR/dl
dR/dS
176
190
210
230
244
0.143
0.143
0.143
0.143
0.143
-0.253
-0.253
-0.253
-0.253
-0.253
Welding Current (/), amps
dR/dN
0.038
0.038
0.038
0.038
0.038
176
190
210
230
244
&Pa/dl &Pa/dS
0.325
0.325
0.325
0.325
0.325
-2.2195
-1.1160
0.5020
2.1200
3.2235
d~a/dN
0.392
0.392
0.392
0.392
0.392
Bead Width Sensitivities of Process P a r a m e t e r s
Table 11
Coefficient of External Shape Sensitivities of Process P a r a m e t e r s
Welding Speed (S) = 34 cnffmin.;
Nozzle-to-Plate Distance (N) = 20 mm
Welding Speed (S) = 34 cm/min.;
Nozzle-to-Plate Distance (N) = 20 nma
Table 7
Welding Current (I), amps
176
190
210
230
244
dW/dl
1.539
1.539
1.539
1.539
1.539
dW/dS
-0.46
-0.46
-0.46
-0.46
-0.46
Welding Current (/), amps
176
190
210
230
244
dW/dN
0.669
0.426
0.069
-0.288
4).532
Welding Speed (S) = 34 cm/min.;
Nozzle-to-Plate Distance (N) = 20 mm
dAP/dl
0.592
0.592
0.592
0.592
0.592
dAP/dS
-0.101
0.003
0.156
0.309
0.413
&peldN
-0.002
-0.002
-0.002
-0.002
-0.002
From Figure 7, it can be observed that the sensitivity of welding current increases steadily with an
increase in current for coefficient of external shape
(%), whereas for the penetration its value is positive
for lower current values, and it turns to negative as
the value of current is increased after a certain level
(say, beyond 225 amps).
Table 8
A r e a of Penetration Sensitivities of Process P a r a m e t e r s
Welding Current (1), amps
176
190
210
230
244
& p e / d l d~e/dS
0.0907
0.169
0.1630
0.114
0.2690
0.032
0.3750
-0.050
0.4473
-0.106
dAP/dN
1.218
0.686
-0.094
-0.874
-1.406
Sensitivity of Welding Speed on Bead Geometry
Figure 8 depicts the sensitivity of welding speed
on the weld bead geometry. It is evident from these
figures that for a given welding current the sensitivity of welding speed on bead parameters (except the
dilution and reinforcement) is lower than the sensitivity of welding current, which means that any
changes in the welding current affect the penetration, bead width, area of penetration, coefficient of
internal shape, and coefficient of external shape more
strongly than any changes in welding speed. The
depth of penetration has negative sensitivity, the
current than others. Also. it is interesting to note that
the welding current sensitivity for coefficient of internal shape is negative at lower values of current
and changes to positive when the welding current is
increased beyond a certain level. It can be observed
from Figure 7 that though the bead width, reinforcement, and dilution are sensitive to welding current,
its value remains unchanged for all welding currents
when welding speed and nozzle-to-plate distance are
kept at a constant level.
97
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
S = 34¢m I m i n N = 2Omm
S. 34cm/mln
N . 20 mm
2.1
2.7'
1.6
1.1
1.7'
fiR
riW
0.7"
riD
-0.3"
nAP
r1~a
-1.3 "
="(be
0.6
IR
~lW
0.1
BD
-0.4
-0.9
-1.4
-1.9
-2.3 "
176
190
210
230
176
244
0.5
0.3
0.1
.0.1
IP, iP.
rip
mR
riW
laD
~AF
.0.3
[3(l~
.0.5
Bee
_
176
190
_
210
230
244
parameters. Also, the sensitivity on the dilution is
positive at lower values of current, but it turns to
negative if the current value is at higher regions, say,
beyond 214 amps.
Similar to the dilution, the sensitivity of nozzleto-plate distance on the penetration, bead width, and
area of penetration becomes negative when the current is increased above a certain level. It remains
unchanged for the reinforcement and coefficient of
internal shape and is positive for all levels of current. The sensitivity of nozzle-to-plate distance is not
very significant for depth of penetration, reinforcement, and coefficient of internal shape, which means
that these parameters are least affected by any change
in the values of N.
Figures 10-12 show some typical sensitivity plots.
From these figures, it can be observed that the change
in sensitivity of penetration is more pronounced for
changes in welding current, whereas changes in sensitivity of dilution and bead width are more prominent for c h a n g e s in n o z z l e - t o - p l a t e d i s t a n c e
compared to the changes in welding current and
welding speed.
S = 34cm / rain N = 2 0 r a m
-0.7
210
Figure 9
Sensitivity Analysis Results of Nozzle-to-Plate Distance on
P, R, W, D, AP, q~, a n d ~.
Figure 7
Sensitivity Analysis Results of Welding C u r r e n t on P, R, W,
D, AP, ~., and ~
0.7 -~
190
Welding Current (1~ Amps
WeildlwJCurrent(I), Amp=
i
230
244
Welding Current (I), Amp=
Figure 8
Sensitivity Analysis Results of Welding Speed on P, R, W,
D, AP. ~., a n d ~p.
magnitude of which increases with the current; that
is, penetration decreases with increasing welding
speed, and this effect is more pronounced at higher
values of current. The sensitivity of welding speed
for dilution and coefficient of external shape is positive at lower current regions and turns negative at
higher current regions, whereas, it is constant for coefficient of internal shape for all levels of welding current. The sensitivity of welding speed for reinforcement
is negative and is constant for all levels of current.
Conclusions
The following conclusions were made from the
above investigations:
Sensitivity of Nozzle-to-Plate Distance on Bead
Geometry
1. The relationships between process parameters
for flux cored arc welding (FCAW) of 317L
flux cored wire on structural steel plate have
been established. The response surface methodology was adopted to develop the regression
models, which were checked for their accuracy
and found to be satisfactory.
Figure P shows the sensitivity of nozzle-to-plate
distance on weld bead geometry. Percent dilution is
more sensitive to nozzle-to-plate distance as compared to the other parameters, and it means that the
variation in nozzle-to-plate distance causes large
changes of dilution and small changes of other bead
98
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
S = 34 cm/min N = 20 mm
S = 34 cm/min N = 20 mm
3
--
=
0.3
0.25
0.2
0.15
0.1
0.050
-oo5
_ .I w'"
,~
I--e--Welding
~
Current
~
~
'~
I
°1
.--S.--Nozzle-toplate
L distance
~
[
~
~
Current
I--e--Welding
speed
-o.1
-0.15
.
176
190
210
230
.
.
.
.
.
.
.
i= Welding
-
-°
244
Welding Current (I), Amps
Welding Current (I), Amps
Figure 10
Sensitivity of Penetration
Figure 11
Sensitivity of Dilution
S = 34 cm/min N= 20 mm
2
2. Sensitivity analysis was performed to identify
process parameters exerting the most influence
on the bead geometry.
3. Changes in welding current have more significant effect on P, W, AP, q~,, and % than does
welding speed.
4. Sensitivity analyses have indicated that W, AP,
D, and q~, are more strongly affected by change
of process parameters compared with the other
bead parameters.
5. Because the sensitivity of welding current on
bead width is higher than that of welding speed
and nozzle-to-plate distance, the change o f
welding current is more useful in controlling
bead width.
6. The height of reinforcement (R) is more sensitive to changes in welding speed than the other
process paranaeters. Hence, it is reasonable to
control the welding speed to get the desired
value of R.
7. The effect of change in welding current on
coefficient of internal shape (q~,) is more significant and, therefore, is more useful to control the welding current to obtain a desirable
value of q~,,.
8. Percent dilution can easily be controlled with
minimal change in the value of N, while the other
process parameters are kept at a desh-ed level.
9. The external shape coefficient (%) is not very
sensitive to change in the process parameters;
however, its value can effectively be controlled
by changing the levels of the welding current
because the sensitivity of welding current on
% is more than the other process parameters.
¸
! --e.-- Welding
Current
1.5
i'
--f--Welding
speed
"6 0.5
'|
u)
o
.0.51
190
2u10
Itt"
- 2~
.L Nozzle-t(>
plate
distance
7u,
-1
Welding Current (I), Amps
Figure 12
Sensitivity of Bead Width
Acknowledgments
The authors wish to thank the All India Council
for Technical Education, New Delhi, and University
Grant Commission, New Delhi, tbr financial support for procuring the equipment and materials. The
authors also wish to thank M/S B6hler Thyssen
Welding, Austria, for sponsoring the 317L wire to
carry out this investigation.
References
Alam, N.; Jarvis, L.; Harris, D.; and Solta, A. (2002). "Laser cladding
for repair of engineering components." Australasian Welding Journal (v47, 2nd qtr), pp38-47.
Allen, T.T.; Richardson, R.W.; Tagliabue, D.P.; and Maul. G.P. (2002).
"Statistaical process design for robotic GMA welding of sheet
metal." Welding Journal (v81, n5), pp69s-76s.
Cary, Howard B. (2002). Modern Welding Technology, 5th ed.
Englewood Cliffs, NJ: Prentice-Hall, p154.
Chandel, R.S. and Bala, R.S. (1986). "Effect of welding parameters
and groove angle on the soundness of root beads deposited by the
SAW process." Proc. of Trends in Welding Research conf. (May
1986). pp379-385.
Chen, S.B., Lou, Y.J.; Wu, L.; and Zhao, B.D. (2000). "Intelligent
methodology for sensing, modeling and control of pulsed GTAW:
Part I - bead-on-plate welding." Welding Journal (v79, n6),
pp151s-163s.
99
Journal of Manufacturing Processes
Vol. 8/No. 2
2006
Cheremisinoff, Nicholas E and Ferrante, Louise (1987). Practical Statistics Jor Engineers and Scientists. Technomic Publishing Co. Inc.
Cochran, W.G. and Cox, G.M. (1957). Experimental Designs, 2nd ed.
New York: John Wiley & Sons Inc., p350.
Cornu, Jean (1988). Advanced Welding Systems (v2). IFS (Publications) Ltd., pp97-114.
Gunaraj, V. and Murugan, N. (1999). "Application of response surface methodology for predicting weld quality in submerged arc
welding of pipes." Journal of Materials Processing I~chnology
(v88), pp266-275.
Heston, Tim (2000). "Cladding operations doubles life of boiler tubing." Welding Journal (v79, n7), pp45-47.
Juang, S.C. and Tarng, Y.S. (2002). "Process parameter selection for
optimizing the weld pool geometry in the tungsten inert gas welding of stainless steel." Journal of Materials Processing Technology
(v122), pp33-37.
Kang, M.J.; Kim, Y.; Ahn, S.; and Rhee, S. (2003). "Spatter rate
estimation in the short circuit transfer region of GMA~\" Welding
Journal (v82, n9), pp238s-247s.
Khodadad Motarjemi, A. and Kocak, M. (2001). "Fracture assessment of weld repaired cladded steel wide plates by SINTAP defect
assessment procedure." Welding #~ the World (v45), Special Issue,
pp27-33.
Khuri, A.I. and Cornell, J.A. (1996). Response Sulfaces - Designs
and Analyses. New York: Marcel Dekker Inc.
Kim, Dongcheol; Rhee, Sehun; and Pal'k, Hyunsung (2002). "Modeling and optimization of a GMA welding process by genetic
algorithm and response surface methodology." lnt'l Journal of
Production Research (v40, n7), pp1669-1711.
Kim. Ill-Soo; Son, Joon-Sik; and Jeung, Young-Jae (2001). "Control
and optimisation of bead width for nmlti-pass welding in robotic
arc welding processes." Australasian Welding Journal (v46, 3rd
qtr), pp43-46.
Kim. I.S.; Son, K.J.; Yang, Y.S.; and Yaragada, P.K.D.V. (2003).
"Sensitivity analysis for process parameters in GMA welding process using factorial design method." lnt'l Journal of Machine
?bols & Manufacture (v43), pp763-769.
Marimuthu, K. and Murugan, N. (2005). "Sensitivity analysis of
process parameters in PTA hardfacing of valve seats using response surface methodology." Materials Science and Technology
(v21, n8), pp941-947.
McGlone, J.C. t1982). "Weld bead geometry prediction - a review."
Material Construction (July 1982), pp378-384.
Missori, S.; Murdolo, E; and Silk A. (2004). "Single-pass laser beam
welding of clad steel plate." Welding .loumal (v83, n2), pp65s-71s.
Montgomery, D.C. (2001). Design and Analysis ~/ ILrperiments, 5th
ed. New York: John Wiley & Sons Inc.
Montgomery, D.C. and Runger, George C. (1999). Applied Statistics
and ProbabiliO, for Engineers, 2nd ed. New York: John Wiley &
Sons Inc,
Murugan, N. (1993). "Some studies on metal surfacing." PhD thesis.
Delhi: Indian Institute of Technology.
Murugan, N. and Parmer, R.S. (1994). "Effects of MIG process parameters on the geometry of the bead in the automatic surfacing
of stainless steel." Journal of Materials Processing Technology
(v41), pp381-398.
Murugan, N. and Parmer, R.S. (1995). "Mathematical models for
bead geometry prediction in austenitic stainless steel surfacing by
MIG welding." lnt'l Journal .for the Joining of Materials (v7, n2/
3), pp71-80.
Raja. A.; Rohira, K.L.; and Samidas, S. (1999). "Flux cored arc welding - selection of consumables, equipments and applications."
WRI Jounal (v20, n4), pp124-129.
Rajasekaran, S. (2000). "Surface topography of pulsed current gas
metal arc clads." Smface Engg. (v16, n6), pp495-500.
Ramasamy, S.; Gould, J.; and Workman, D. (2002). "'Design of experiments study to examine the effect of polarity on stud welding." Welding Journal (v81, n2), pp19s-26s.
Sakai, Y.; Aida, G.; Suga, T.; and Nakano, T. (1989). "Development
of various flux cored wires and their applications in Japan." IIW
document XII-1131-89.
Stewart, J.P. (1981). ?'he Welder's Handbook. Restor Publishing Co.
Inc., pp163-219.
Subramaniam, S.; White, D.R.; Jones, J.E.; and Lyons, D.W. (1999).
"Experimental approach to selection of pulsing parameters in pulsed
GMAW." Welding Journal (v78, n5), pp166s-172s.
Walpole, R.E.; Myers, R.H.; and Myers, S.L. (1998). Probability and
Statistics for Engineers and Scientists, 6th ed. Englewood Cliffs,
NJ: Prentice-Hall.
Appendix
Variables
Notation
Units
I
S
N
~
P
R
W
D
AP
~Pa
amps
cm/min.
mm
m/rain.
mm
mm
mm
dimensionless
mm'dimensionless
q~
dimensionless
Welding current
Welding speed
Nozzle-to-plate distance
Wire feed rate
Penetration
Reinforcement
Bead width
% Dilution
Area of penetration
Coefficient of internal
shape of welds
Coefficient of external
shape of welds
Authors' Biographies
P.K. Palani is currently working as senior lecturer at the Government College of Technology, Coimbatore, India. He obtained his BE
from the National Institute of Technology, Thiruchirapally, in 1987
and his ME from the Government College of Technology, Coimbatore,
in 1999. He is presently doing his PhD in the field of welding at Anna
University, Chennai, India. His research interests are in welding and
cladding. He is a member of the Institution of Engineers (India),
Indian Welding Society (IWS), and Indian Society for Technical Education (ISTE). He is a recipient of the President of India's Prize
(English) for a technical paper published in the Joun~al of tile hzstitution of Engineers (India).
Dr. N. Murugan is a professor of mechanical engineering at the
Coimbatore Institute of Technology. Coimbatore, India. He received
his BE in mechanical engineering and ME in production engineering
from Madras University during 1981 and 1983, respectively. He
obtained his PhD from the Indian Institute of Technology, Delhi, in
1994. He is a member of the American Welding Society and a fellow
of the Institution of Engineers (India). His research interests are in the
areas of welding, process modeling, optimization, color metallography, hardfacing, cladding, and corrosion studies. He received the
prestigious McKay-Helm Award for the best contribution for the
advancement of knowledge in the field of stainless steel surfacing
from the American Welding Society for the year 1998.
100