The influence of welding parameters on flux consumption in submerged arc
welding
Fateh Pal Singh
Department of Mechanical Engineering
Lovely Professional University, Jalandhar
Delhi G.T Road– 144402, Punjab (India)
#Corresponding Author, E-mail address: [email protected]
Tel.: +91-9876632400
#
Abstract
Welding flux contributed a major part (about 50 %) towards total welding cost in submerged arc
welding. Besides economics consideration, flux consumption influences the pick up or loss of
alloying elements by weld metal, thus the mechanical and metallurgical properties. In this
research work the effect of welding parameters on flux consumption has been investigated. Four
welding parameters wire feed rate, arc voltage, travel speed and contact tip to work distance
were selected. Flux consumption in Kg/Kg of weld metal deposited was measured.
Mathematical model was developed from the data generated using two level half factorial
technique. The significant of coefficient and adequacy of developed model was checked by ‘ t ‘
test and F test respectively. The main effect of welding parameters on flux consumption has been
presented in graphical form for better understanding. It was observed that flux consumption
decreases with increase in wire feed rate and travel speed. Flux consumption increased with
increase in arc voltage.
Introduction :
Since the development of submerged arc welding attempts have been made by researchers and
engineers to decrease welding cost. Flux contributes about 50% to the total welding cost in
submerged arc welding. Flux consumption not only affects the welding cost but also governs the
chemistry of weld metal which in turn influences metallurgical and mechanical properties [1].
Flux contributes about 50% to the total welding cost in submerged arc welding. Flux
consumption affects the chemistry of weld metal which in turn influences metallurgical and
mechanical properties. The flux consumption in submerged arc welding is mainly dependent on
physical properties of flux such as melting point, density, thermal and electrical conductivity.
Submerged arc welding fluxes basically performed same function as the coating of manual
electrode, additionally; it must satisfy certain special conditions demanded by nature of the
process[2]. Buttler et al. [3] found that flux consumption decreased as the melting temperature of
flux increased. Thermal conductivity of molten flux also has pronounced effect on flux
consumption. Higher the thermal conductivity of molten fluxes higher the flux consumption [4].
They further observed that the flux consumption was highest in acidic flux than neutral and basic
fluxes. Gupta et al. [5, 6] reported that welding parameters have a significant effect on flux
consumption. They found that flux consumption increased with increase in arc voltage and
decreased with increase in welding current. Visvanath [7] indicates that rate of flux consumption
is more in fused flux compared with agglomerated fluxes. A comparative study done by Srinath,
H. [8] reported that flux consumption increases with increase in arc voltage. Wittstock [9] also
found that arc voltage controls the amount of fused flux.
All the above researchers used commercial available fluxes in their investigations. It appears that
almost no consideration has so for been made on the influence of welding parameters on flux
consumption and establish relationship between flux consumption and welding parameters.
Plan Of Investigation :
1. Identifying the important process control variables and finding their upper and lower limits.
2. Developing the design matrix.
3. Conducting the experiments as per the design matrix.
4. Recording the responses.
5. Developing the mathematical models.
6. Calculating the coefficients of the polynomials.
7. Checking the adequacy of the models developed.
8. Calculating the significance of the coefficients and arriving at the final mathematical
models.
9. Conducting the conformity test.
10. Presenting the effects of the process variables on flux consumption.
Identification of process variables and their working range:
Welding current (I), arc voltage (V), travel speed (S) and nozzle to plate distance (N). The
working range was decided upon by inspecting the bead for a smooth appearance without any
visible defects such as surface porosity, under cut, pock marks. The upper limit of a factor was
coded as (+1) and lower limit as (-1) or simply (+) and (-) respectively. The decided values of
process parameters with their units and notations are given in Table 1.
Table 1: Welding parameters and their working range
PARAMETERS
SYMBOL
Welding current, amp.
Arc voltage, volts
Travel speed, m/min
Nozzle to plate
distance, mm
Development of design matrix:
Lower limit
300
30
0.3
30
I
V
S
N
LIMITS
Upper limit
500
36
0.6
40
The design matrix developed to conduct the eight trials of 24-1 (=8) two level half factorial design
is shown in Table 2.
Table 2: Design Matrix
Trial No.
1.
2.
3.
4.
5.
6.
7.
8.
I
1
+1
-1
+1
-1
+1
-1
+1
-1
V
2
+1
+1
-1
-1
+1
+1
-1
-1
S
N
3 4=123
+1
+1
+1
-1
+1
-1
+1
+1
-1
-1
-1
+1
-1
+1
-1
-1
Conducting the experiments:
Beads on mild steel plates having size 10×50×170 mm were deposited as per design matrix using
3.15 mm diameter conforming to AWS A5.17- 69, EL-08 wire. Electrode positive reverse
polarity was used. A constant potential transformer-rectifier type power source with a current
capacity of 1200 amperes at 60% duty cycle and 900 amperes at 100 % duty cycle, an OCV of
26 to 44 volts was used. For flux consumption beads on plate were deposited. The weight of
plate before and after deposition bead were recorded. Flux consumed for each bead was noted.
Then flux consumed in kg/kg of weld metal deposited was calculated. The experiments were
performed in random manner to avoid any systematic error. The complete set of eight trials was
repeated thrice for the sake of determining the variance of parameters and variance of adequacy
for the model.
Recording of responses:
The experiments were performed in random manner to avoid any systematic error. A measured
quantity (1kg) of flux was used for each bead. The initial and final weight of base plate before
and after deposition of weld bead was noted and recoded in Table 3 and 3.1 for two sets of
experiments. Then flux consumption in kg/kg of weld metal deposited was calculated for each
trial run.
Table 3: Flux consumption (kg/kg metal deposited) for set-I
Trail
No
I
V S N
W1
(initial
wt. of
plate)
1.
2.
3.
4.
5.
6.
7.
8.
+
_
+
_
+
_
+
_
+
+
_
_
+
+
_
_
1.252
1.250
1.246
1.112
1.264
1.252
1.256
1.262
Trail
No
I
V S N
W1
(initia
l wt.
of
plate)
+
_
+
_
+
_
+
+
_
_
+
+
1.260
1.252
1.266
1.240
1.254
1.270
+
+
+
+
_
_
_
_
+
_
_
+
_
+
+
_
W2
t,
Metal
f1
f2
f
(wt. of sec deposition (initial (wt. of
(gm/sec)
plate
rate
wt. of flux
after
g/sec
flux
retained)
welding)
used)
1.282
1.266
1.276
1.124
1.330
1.282
1.316
1.290
18
17
14
14
35
33
33
30
1.8
0.94
2.21
0.85
1.88
0.9
1.82
0.93
1000
1000
1000
1000
1000
1000
1000
1000
976
974
985
986
932
948
968
966
1.33
1.529
1.071
1.00
1.943
1.575
0.969
1.133
Table 3.1: Flux consumption (kg/kg metal deposited) for set-II
1.
2.
3.
4.
5.
6.
+ +
+ _
+ _
+ +
_ _
_ +
W2
t,
Metal
(wt. of sec deposition
plate
rate rate
after
g/sec
welding)
1.292
1.266
1.296
1.254
1.316
1.302
19
16
15
17
33
35
1.68
0.88
2.1
0.82
1.87
0.91
f1
(initia
l wt.
of
flux
used)
1000
1000
1000
1000
1000
1000
f2
(wt. of
flux
retained
)
f
(gm/se
c)
976
977
988
984
942
942
1.263
1.438
0.8
0.94
1.75
1.657
7. + _ _ + 1.242
1.300
31
1.86
1000
970
0.968
8. _ _ _ _ 1.250
1.280
32
0.94
1000
965
1.094
Development of mathematical models:
The response function representing any of the weld bead dimensions could be expressed as
Y = f(I, V, S, N)
Assuming a linear relationship in the first instant and taking into account all the possible two
factor interactions only, the above expression could be written as
Y = b0 + b1 I + b2 V + b3 S + b4 N + b12 IV + b13 IS + b14 IN + b23 VS + b24 VN + b34 SN
After confounding the model can be rewritten as
Y = b0 + b1 I + b2 V + b3 S + b4 N + b5 (IV + SN) + b6 (IS + VN) + b7 (IN + VS)
Evaluation of the coefficients:
The regression coefficients of the selected model were calculated using equation.
∑N
i=j Xji Yi
bj =
, j = 0,1 … . . , k
N
Developed models: Substituting the values of the coefficients
F = 1.064 − 0.375I + 0.225V − 0.0515S − 0.0059N − 0.0347( IV + SN) − 0.015(IS + VN)
− 0.033(IN + VS)
Testing significance of the coefficients:
The statistical significance of the coefficients was tested by student ‘t’ test. The value of ‘t’ from
the standard table for eight degree of freedom and 95% confidence level is 2.3 . The calculated
‘t’ values for the coefficients are given in Table 4. The coefficients having ‘t’ value less than 2.3
are insignificant hence dropped in the final models.
Table 4: Coefficient of the model and their significant
Coefficient
b0
b1
b2
b3
b4
b5
b6
b7
Value
1.065 -0.376 0.226 -0.5156 -0.0059 -0.0347 -0.015 -0.033
‘ t ‘ Value 59.48
21
12.62
2.88
0.332
1.938
0.821
1.83
Significant YES
YES
YES
YES
NO
NO
NO
NO
Conducting the conformity test:
The conformity test is carried out to check and analysis the model in order to determine, whether
the model fulfill the requirements or not.
Table 5: Conformity test
Trial
No.
1
2
3
4
5
6
7
8
Predicted Value of
flux consumption
0.8566
1.6184
0.4184
1.1566
0.9714
1.76096
0.5096
1.2714
Actual value of flux
consumption (3rd Set)
0.82
1.625
0.43
1.18
0.985
1.765
0.525
1.25
Development of final models
The final mathematical models after dropping insignificant coefficients are
𝐅𝐜 = 𝟏. 𝟎𝟔𝟒 − 𝟎. 𝟑𝟕𝟓𝐈 + 𝟎. 𝟐𝟐𝟓𝐕 − 𝟎. 𝟎𝟓𝟏𝟓𝐒 − 𝟎. 𝟎𝟎𝟓𝟗𝐍
Checking adequacy of the developed models:
The adequacy of developed models was then tested by the analysis of variance technique. As per
this technique if the calculated value of model’s F-ratio does not exceed its tabulated value for a
desired level of confidence, then the model is adequate. The calculated ‘F’- ratio of the models
were compared with the corresponding ‘F’- ratio from the standard table. In the present study,
tabulated F- ratio for 95% confidence level at 3 degree of freedom of the variance of adequacy
and 8 degree of freedom of variance of optimization respectively. It has been found that the
developed model is adequate within 95% level of confidence as shown in Table 6.
Response
Parameter
Table 6: Analysis of variance
Degree of
freedom
2
S2y Sad
Variance Std. devi.
of
of
optimn. coefficient
parameter
Sbj
2
Sy
0.002564
0.0179
Vari. of
adequacy
2
S ad
FRatio
Model
Fm
FRatio
from
table
Ft
Model
whether
adequate
Fm< Ft
Flux
8
3
0.003418
1.33
4.12
Yes
consumption
Result and Discussion:
Effect of welding current on flux consumption:
Figure1 indicates the effect of welding current on flux consumption kg/kg of metal deposited. It
can be observed that flux consumption decreases linearly from 1.439 to 0.689 kg/kg of weld
metal deposited with an increase in welding current from 300 to 500 ampere. This decrease in
flux consumption is due to the fact that with the increase in welding current, the melting rate of
electrode increases and there by the metal deposition rate increases. It was further observed that
flux consumption in kg/sec decreases with increase in welding current; however rate of metal
deposition kg/sec is more than rate of flux consumption in kg per second. Therefore the ratio of
flux consumption kg/kg of metal deposited decreased.
Effect of arc voltage on flux consumption:
Figure2 indicates the effect of arc voltage on flux consumption. The flux consumption increased
from 0.8 to 1.3 kg/kg of metal deposited with an increase in arc voltage from 30 to 36 volts. The
increase in flux consumption with increase in arc voltage can be attributed to the fact that the
increase in arc voltage increases arc length, thereby increasing the spread of arc and hence higher
amount of flux coming in contact with the arc. Due to increased arc length arc strikes on a larger
surface area causing wider and flatter bead hence increase in contact area between molten pool
and flux, resulting more flux to melt.
Effect of travel speed on flux consumption:
Figure3 indicates the effect of travel speed on flux consumption. It can be observed that flux
consumption decreases linearly from 1.1155 to 1.0125 kg/kg of weld metal deposited with an
increase in welding speed from 0.3 to 0.6 m/min. As welding current and arc voltage is constant
increase in travel speed decreases heat input per unit length of weld. Less amount of heat input
per unit length of weld generates smaller size of molten pool. Due smaller bead width smaller
area of contact between molten metal and flux caused less amount of flux to melt. Hence flux
consumption decreased with increase in travel speed.
Effect of nozzle to plate Distance on Flux Consumption Rate:
Figure 4 indicates the effect of nozzle to plate distance on flux consumption. It can be observed
that flux consumption decreases linearly from 1.0699 to 1.0581 kg/kg of weld metal deposited
with an increase in nozzle to plate distance from 30 to 40 mm. As nozzle to plate distance
increases a digging arc is obtained affecting weld penetration. More heat arc is utilized for
melting of base metal. Due to constricted and digging arc, area of contact between molten metal
and flux decreases leading to decrease in flux consumption.
Flux consumption, kg/kg
weld metal deposited
1.5
1.3
1.1
0.9
0.7
0.5
-1
-0.5
0
0.5
1
Welding current, coded values
Flux consumption, kg/kg
weld metal deposited
Figure1: Effect of welding current
1.5
1.35
1.2
1.05
0.9
0.75
-1
-0.5
0
0.5
1
Arc voltage, coded values
Flux consumption, kg/kg
weld metal deposted
Figure2 : Effect Of Arc Voltage
1.14
1.12
1.1
1.08
1.06
1.04
1.02
1
-1
-0.5
0
0.5
1
Travel speed, coded values
Flux consumption, kg/kg
weld metal deposited
Figure3 : Effect Of Travel Speed
1.072
1.068
1.064
1.06
1.056
-1
-0.5
0
0.5
Nozzle to plate distance, coded values
Figure4: Effect Of Nozzle To Plate Distance
1
Conclusion:
1. Two level factorial design techniques could be employed for developing the model for
predicting the flux consumption in submerged arc welding process.
2. Arc voltage has profound effect on flux consumption, flux consumption increased from
0.839 to 1.289 kg/kg of metal deposited with an increase in arc voltage 30 to 36 volts.
3. It was further observed that flux consumption decreased from 1.439 to 0.689 kg/kg of
metal deposited by increasing welding current from 300 to 500 amperes.
4. Flux consumption decreases linearly from 1.1155 to 1.0125 kg/kg of weld metal
deposited with an increase in travel speed from 0.3 to 0.6 m/min.
5. Flux consumption decreases linearly from 1.0699 to 1.0581 kg/kg of weld metal
deposited with an increase in nozzle to plate distance from 30 to 40 mm.
6. Slag detachability and arc stability both were satisfactory.
References:
[1]
Polar, A., Indacochea, J.E. and Blander, M., Fundamentals of the chemical behavior of
selecting welding fluxes, Welding Journal, 70(1991), 51s-18s.
[2]
Ferrera, K.P. and Olson, D.L., Performance of MnO-SiO2-CaO system as a welding flux,
Welding Journal, 54(1975), 211s-215s.
Butler, C.A., and Jackson, C.E., Submerged arc welding characteristics of CaO-TiO2SiO2 system, Welding Journal, 46(1967), 448s-456s.
Renwick, B.G. and Patchett, B.M., Operating characteristics of submerged arc process,
Welding Journal, 55(1976), 69s-76s
[3]
[4]
[5]
Gupta, S.R., Gupta, P.C. and Reddy, P.K., Investigation into flux consumption in
submerged arc welding, Indian Welding Journal, 21(1988), 365-370.
[6]
Gupta, S.R., Gupta, P.C. and Reddy, P.K., Influence of flux basicity on flux consumption
and melting rate in submerged arc welding, in Proc. of National Welding Seminar,
organizes by Indian Institute of Welding, New Delhi (1989), 10.1-10.10
[7]
[8]
Visvanath, P.S., Submerged arc welding fluxes, Indian Welding Journal, 15(1982), 1-11.
Srinath, H., Some aspects of submerged arc welding process, Indian Welding Journal,
12(1975), 67-73.
[9]
Wittstock, G.G., Selecting submerged arc fluxes for carbon and low alloy steels, Welding
Journal, 55(1976), 733-741.