Sage Journals: Discover world-class research

Abstract

This study explores the elongation of submerged arc-welded joints by using developed agglomerated fluxes. The response surface methodology technique was used to conduct the experiments. MnO, CaF₂, NiO, MgO, and Fe-Cr were added to the fluxes comprising CaO, SiO₂, and Al₂O₃ as the main constituents. The effect of the addition was studied by means of mechanical performances of the welds in terms of elongation. The elongation of weld metal increases with the increase in MnO, MgO, and Fe-Cr constituents, but CaF₂ and NiO constituents did not reveal favorable sign unlike the other three constituents mentioned here. The recommended optimal conditions were verified by conducting confirmation experiments.

Keywords

Response surface methodology submerged arc welding agglomerated fluxes elongation

Introduction

The submerged arc welding (SAW) is a versatile process and is used for a wide range of applications that include the critical applications, for example, joining of pressure vessels, thick plates, ship hulls, and so on. In SAW, the arc is covered by a flux. The constituents, present in the flux, control arc stability, bead shape, and the mechanical and chemical properties of weld metal.¹ The SiO₂ content, for example, in the synthetic fluxes improves the welding performance.

As is determined through metallographic examination, the presence of a metallic mist in the matrix of the flux and its amount get increased with the addition of silica contents.² Colvin³ in his study propelled that the basic fluxes are more or less associated with the improved metallurgical properties of weld than the acidic fluxes. The study of commercial acidic fluxes by Bennett⁴ has correlated the slag detachability and weld bead surface appearance to their chemical composition. Several other studies have revealed the effect of various flux constituents onto the transfer of elements to the weld metal. As a result, the mechanical and metallurgical properties of the weld metal get affected.^5–20 This work studies the elongation of low-carbon steel weld joint, which is the most important property to be imparted to the weld metal.

Selection of flux constituents and their ranges

In the study under consideration, the fluxes have been designed on the basis of binary and ternary phase diagrams for different oxide and fluoride systems.²¹ In these fluxes, varying amount of oxides and fluorides was added to study the impact of these constituents onto the elongation of the weld metal. After ascertaining the compositions and constituents of fluxes, they were prepared by agglomeration method using the potassium silicate as its binder.

The selection of flux constituents is generally based on the composition of base plate and electrode employed to welding. CaO-Al₂O₃-SiO₂–based flux systems have been selected for the study, as these are the most widely used fluxes at the commercial level. The details of basic constituents, alloying constituents, and their percentage are reported in Table 1.

Table 1.

Constituents and their range.

Basic constituents and alloying constituents	CaO	SiO₂	Al₂O₃	MnO	CaF₂	MgO	NiO	Fe-Cr
Amount in %	20	35	10	5–10	10–20	5–15	0–10	0–10

Response surface methodology technique

The response surface methodology (RSM) was used to design and point out the relationship between the response of interest and the variables (alloying constituents of flux) as well as to determine the conditions of these variables that were actually able to optimize this response. In this study, five constituents of flux taken as a process variable are used as levels that maximize the yield (Y) of a process as shown in equation (1)

Y = f (x_{1}, x_{2}, x_{3}, x_{4}, x_{5}) \pm ε

(1)

where Y is the response, f is the response function, ε represents the experimental error, and x₁, x₂, x₃, x₄, and x₅ are the independent constituents in the developed agglomerated flux. The expected response (elongation) is defined by E(Y) = f(x₁, x₂, x₃, x₄, x₅) = η, where η is called the response surface. To establish the relationship between response and the independent variable, the first-order model is employed and the above equation can be written as shown in equation (2)

Y = β_{0} + β_{1} x_{1} + β_{2} x_{2} + \dots + β_{5} x_{5} \pm ε

(2)

The behavior of the system can be described by the following second-order polynomial as shown in equation (3)

Y = β_{0} + \sum_{i = 1}^{5} β_{i} x_{i} + \sum_{i = 1}^{5} β_{ii} x_{i}^{2} + \sum β_{ij} x_{i} x_{j} \pm ε

(3)

where Y is the response, $β_{0}$ is the interception coefficient, $β_{i}$ are the linear terms, $β_{ii}$ are the quadratic terms, $β_{ij}$ are the interactions terms, and $x_{i}$ and $x_{j}$ are the coded levels of the independent variable taken in the form of flux constituents.^22–26 The design matrix for the experiments was planned on the basis of the coded values. It was generated through RSM technique, which has been mentioned along with consecutive response (percentage elongation) in Table 2.

Table 2.

The design matrix of agglomerated fluxes constituents and response.

Flux numbers	Run	MnO	CaF₂	MgO	NiO	Fe-Cr	Percentage elongation
1	23	−1	−1	−1	−1	1	28.56
2	3	1	−1	−1	−1	−1	29.07
3	1	−1	1	−1	−1	−1	26.77
4	32	1	1	−1	−1	1	27.03
5	9	−1	−1	1	−1	−1	26.52
6	21	1	−1	1	−1	1	27.03
7	8	−1	1	1	−1	1	29.32
8	15	1	1	1	−1	−1	29.58
9	26	−1	−1	−1	1	−1	27.28
10	6	1	−1	−1	1	1	29.1
11	2	−1	1	−1	1	1	27.28
12	4	1	1	−1	1	−1	27.54
13	7	−1	−1	1	1	1	27.77
14	13	1	−1	1	1	−1	28.07
15	20	−1	1	1	1	−1	27.79
16	19	1	1	1	1	1	29.07
17	18	−2	0	0	0	0	28.56
18	30	2	0	0	0	0	29.12
19	28	0	−2	0	0	0	27.54
20	5	0	2	0	0	0	27.79
21	22	0	0	−2	0	0	27.54
22	11	0	0	2	0	0	27.79
23	10	0	0	0	−2	0	27.54
24	17	0	0	0	2	0	28.56
25	14	0	0	0	0	−2	27.54
26	16	0	0	0	0	2	28.05
27	25	0	0	0	0	0	27.79
28	31	0	0	0	0	0	27.89
29	12	0	0	0	0	0	27.84
30	29	0	0	0	0	0	27.64
31	24	0	0	0	0	0	27.48
32	27	0	0	0	0	0	27.54

Flux preparations

The agglomeration method is used for flux preparation by taking small batches of (2 kg) weighed quantities of powdered chemicals, with binder, which is in turn thoroughly mixed in a ball miller. The prepared powdered flux is dried and baked in an electrically heated furnace for 3 h at 800 °C. The dried flux is allowed to cool down to room temperature before storing it in a moisture-free box.

Selection of process parameters of SAW

The process parameters of SAW were selected on the basis of good slag detachability, weld appearance, arc initiation, and arc stability. The higher and lower values of welding parameters may lead to poor welding properties. The welding parameters were chosen on the basis of above properties, as reported in Table 3. The composition of base metal and electrode is given in Table 4.

Table 3.

Process parameters.

Parameter	Value
Arc voltage	36 V
Welding current	500 A
Speed	28 cm/min
Nozzle to plate distance	25 mm
Heat input	3.87 kJ/mm
Polarity	DCEP

DCEP: direct current electrode positive.

Table 4.

Composition of electrode and base metal (%).

Composition of electrode and base plate (%)	C	Mn	Si	Ni	Cr
Electrode (Auto Rod 12.08L 3.15 mm) EL-8 (ESAB)	0.08	0.5	0.05	—	—
Base plate	0.13	0.5	0.2	0.01	0.02

Experimentation

The weld joints were made on low-carbon steel plates of 300 × 300 × 20 mm (V-groove at 45°) size by using 32 developed agglomerated fluxes by SAW machine, with fixed process parameters mentioned in Table 3. The specimens for measuring elongation were prepared from the welded joints as per the standard specification shown in Figure 1. The elongation of prepared samples was measured with the help of Tensometer (KIPL-PC 2000) testing machine (highest limit 18 kN), with a 0.8 mm/min uniform test speed. The specimens before and after testing are shown in Figure 2. The responses of elongations for each welded joint are given in Table 2.

Figure 1.

Specification of specimen.

Figure 2.

Specimens (a) before and (b) after the elongation measurement.

Results and discussion (response surface modeling of percentage elongation)

The analysis of variance (ANOVA) and pooled ANOVA have been reported in Tables 5 and 6. It revealed that the model was actually significant. The f-value of 25.18 in pooled ANOVA and the p-value<0.0001 demonstrated that this regression was statistically significant at 99% confidence level. The value of R² was found to be 0.9115 and fulfilled the requirement of ANOVA. According to ANOVA, R² must lie between zero and one, the larger values in this range being more desirable.²⁵ The values of probability > f for the model less than 0.05 (i.e. α = 0.05, or 95% confidence) indicate that the model terms were significant. The lack-of-fit term is not significant in the model as per the ANOVA condition. In this study, the factors MnO, MgO, and Fe-Cr, the combined effect of constituent MnO with constituent Fe-Cr, the combined effect of constituent CaF₂ with constituent MgO, the combined effect of constituent NiO with constituent Fe-Cr, and second-order term of constituent MnO have significant effect altogether.

Table 5.

ANOVA table of precentage elongation.

Source	Sum of squares	Degree of freedom	Mean square	f-value	Probability > f
Model	17.61	20	0.88	12.76	<0.0001, significant
x₁-MnO	1.66	1	1.66	24.11	0.0005
x₂-CaF₂	0.093	1	0.093	1.34	0.2715
x₃-MgO	0.38	1	0.38	5.51	0.0387
x₄-NiO	0.18	1	0.18	2.56	0.1378
x₅-Fe-Cr	0.53	1	0.53	7.65	0.0184
x₁x₂	0.073	1	0.073	1.06	0.3261
x₁x₃	0.016	1	0.016	0.23	0.6435
x₁x₂	0.28	1	0.28	4.07	0.0687
x₁x₅	2.72	1	2.72	39.45	<0.0001
x₂x₃	8.64	1	8.64	125.24	<0.0001
x₂x₄	0.27	1	0.27	3.84	0.0758
x₂x₅	0.016	1	0.016	0.23	0.6435
x₃x₄	0.014	1	0.014	0.21	0.6567
x₃x₅	4.00E−04	1	4.00E−04	5.80E−03	0.9407
x₄x₅	0.4	1	0.4	5.84	0.0342
$x_{1}^{2}$	2.12	1	2.12	30.7	0.0002
$x_{2}^{2}$	0.017	1	0.017	0.25	0.6252
$x_{3}^{2}$	0.018	1	0.018	0.27	0.6165
$x_{4}^{2}$	0.15	1	0.15	2.16	0.1699
$x_{5}^{2}$	1.65E−03	1	1.65E−03	0.024	0.8799
Residual	0.76	11	0.069
Lack of fit	0.62	6	0.1	3.64	0.0887, not significant
Pure error	0.14	5	0.028
Cor total	18.37	31
Standard deviation = 0.26			R² = 0.9587
Mean = 27.94			Adjusted R² = 0.8835
CV % = 0.94			Predicted R² = 0.0954
PRESS = 16.62			Adequate precision = 13.807

CV: coefficient of varition.

Table 6.

Pooled ANOVA table of percentage elongation.

Source	Sum of squares	Degree of freedom	Mean square	f-value	Probability > f
Model	16.75	9	1.86	25.18	<0.0001, significant
x₁-MnO	1.66	1	1.66	22.52	<0.0001, significant
x₂-CaF₂	0.093	1	0.093	1.25	0.2753, not significant
x₃-MgO	0.38	1	0.38	5.14	0.0335, significant
x₄-NiO	0.18	1	0.18	2.39	0.1362, not significant
x₅-Fe-Cr	0.53	1	0.53	7.15	0.0139, significant
x₁x₅	2.72	1	2.72	36.84	<0.0001, significant
x₂x₃	8.64	1	8.64	116.97	<0.0001, significant
x₄x₅	0.4	1	0.4	5.46	0.0290, significant
x₁²	2.14	1	2.14	28.9	<0.0001, significant
Residual	1.63	22	0.074
Lack of fit	1.48	17	0.087	3.09	0.1109, not significant
Pure error	0.14	5	0.028
Cor total	18.37	31
Standard deviation = 0.27			R² = 0.9115
Mean = 27.94			Adjusted R² = 0.8753
CV % = 0.97			Predicted R² = 0.7701
PRESS = 4.22			Adequate precision = 20.225

CV: coefficient of varition.

The mathematical model that describes the final response equation for elongation is given in equation (4)

Percentage elongation = + 47.74258 - 1.69053 \times MnO - 1.15117 \times Ca F_{2} - 1.71367 \times MgO - 0.092667 \times NiO + 0.92233 \times Fe - Cr - 0.13200 \times MnO \times Fe - Cr + 0.11760 \times Ca F_{2} \times MgO + 0.025400 \times NiO \times Fe - Cr + 0.17075 \times Mn O^{2}

(4)

Figure 3 describes the normal probability plot of the residuals for percentage elongation. Figure 4 shows the values of percentage elongation predicted by the model and actual values obtained by the experiment. It is observed that the distribution of points around the line for this response is fit to linear model.

Figure 3.

Normal probability plot of the residuals for percentage elongation.

Figure 4.

Predicted versus actual percentage elongation value.

Figure 5 depicts the response surface for percentage elongation in relation to the flux constituents CaF₂ and MnO. In this figure, the percentage elongation marginally increases with increasing percentage of CaF₂ constituents because it lowers the high oxygen in submerged arc welds. The value of percentage elongation also increases with the increasing percentage of MnO constituents because it is a transition metal oxide.

Figure 5.

Effect of CaF₂ and MnO on percentage elongation.

The response surface of CaF₂ constituents and MgO constituents on percentage elongation is shown in Figure 6. This figure shows that the value of percentage elongation increases with CaF₂ constituent, the reason for which has been explained earlier. The value of percentage elongation also increases with the increasing percentage of MgO because it is very stable (nonreactive) with the combination of CaF₂, which reduces the oxide nature.

Figure 6.

Effect of MgO on elongation.

Figure 7 shows the response surface of behavior of CaF₂ and NiO constituents. The effect of CaF₂ on elongation was explained earlier, but the effect of NiO on percentage elongation was slightly constant because Ni increases the toughness of the weld metal, which imparts slightly constant elongation at the weld joint. The presence of NiO in the combination of CaF₂ is really helpful to increase the percentage elongation.

Figure 7.

Effect of NiO on elongation.

Figure 8 shows the response surface behavior of CaF₂ and Fe-Cr constituents. The effect of CaF₂ on elongation was explained earlier, but the effect of Fe-Cr on percentage elongation gave positive impact on percentage elongation because Fe and Cr are good transition metals.

Figure 8.

Effect of Fe-Cr on elongation.

Confirmation experiments

Since for validation, the confirmation tests must be performed, so the three confirmation experiments were performed for elongation. In the confirmation tests, experimental values were compared with the predicted value, obtained by the mathematical model equation of percentage elongation. The confirmation is given in Table 7. From the confirmation tests, it is observed that the calculated error is small and tolerable in the field of welding.

Table 7.

Confirmation tests and their comparison with the predicted value.

Experiment number	MnO	CaF₂	MgO	Fe-Cr	Experimental elongation (%)	Predicted elongation (%)	Error (%)
1.	10	20	15	5	27	29.17	−7.43
2.	10	20	15	5	27.5	29.17	−5.72
3.	10	20	15	5	28.49	29.17	−2.33

Conclusion

The ANOVA revealed that the factors MnO, MgO, and Fe-Cr have the positive impact on elongation but CaF₂ and NiO give little impact on elongation.

The value of percentage elongation at the optimum parameters MnO (10), CaF₂ (20), MgO (15), NiO (0), and Fe-Cr (5) is 29.17%.

The confirmation tests show that the errors between experimental and predicted values of elongation are −7.43%, −5.72%, and −2.33%.

Footnotes

Acknowledgements

The authors would like to acknowledge the support of Departments of Mechanical and Chemical Engineering, Noida Institute of Engineering and Technology, Greater Noida, Department of Manufacturing Processes and Automation Engineering, Netaji Subhas Institute of Technology, New Delhi, and Department of Mechanical Engineering, Jamia Millia Islamia, New Delhi.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Declaration of conflicting interests

The authors declare that there is no conflict of interest.

References

Jackson

. Fluxes and slags in welding. Bulletin 190, December 1973. New York: Welding Research Council.

Butler

Jackson

. Submerged arc welding characteristics of the CaO-TiO₂-SiO₂ system. Weld J 1967; 46(10): 448s–455s.

Colvin

. Basic submerged arc welding fluxes. Metallurgia 1970; 45–50.

Bennett

. Using basic fluxes. Met Constr: Brit Weld 1970; 523–527.

Ferrera

Olson

. Performance of the MnO-SiO₂-CaO system as a welding flux. Weld J 1975; 54: 211s–215s.

Wittstock

. Selecting submerged arc fluxes for carbon and low alloy steels. Weld J 1976; 733–741.

North

Bell

Nowicki

. Slag/metal interaction, oxygen and toughness in submerged arc welding. Weld J 1978; 57(3): 63s–75s.

Chai

Eager

. Slag metal reactions in binary CaF₂-metal oxide welding fluxes. Weld J 1982; 229s–232s.

Terashima

Tsuboi

. Submerged arc flux for low oxygen and hydrogen weld metal. Met Constr: Brit Weld 1982; 648–654.

10.

Mitra

Eager

. Slag metal reactions during submerged arc welding of alloy steels. Metall Trans A 1984; 15: 217–227.

11.

Lau

Weatherly

McLean

. Gas/metal/slag reactions in submerged arc welding using Cao-Al₂O₃ based fluxes. Weld J 1986; 31s–38s.

12.

Tandon

Kaushal

Gupta

. Effect of flux characteristics on HAZ during submerged arc welding. In: International conference on welding technology, Department of Mechanical and Industrial Engineering, University of Roorkee, Roorkee, India, September 1988, pp.II-65–II-71.

13.

Burck

Indacochea

Olson

. Effects of welding flux additions on 4340 steel weld metal composition. Welding J 1990; 69(3): 115s–122s.

14.

Paniagua-Mercado

Lopez-Hirata

Munoz

. Influence of the chemical composition of flux on the microstructure and tensile properties of submerged-arc welds. J Mater Process Tech 2005; 169: 3–100.

15.

Kanjilal

Pal

Majumdar

. Combined effect of flux and welding parameters on chemical composition and mechanical properties of submerged-arc weld metal. J Mater Process Tech 2006; 171: 223–231.

16.

Kanjilal

Pal

Majumdar

. Prediction of element transfer in submerged arc welding. Welding J 2007; 86: 135s–146s.

17.

Adeyeye

Oyawale

. Mixture experiments and their applications in welding flux design. J Braz Soc Mech Sci 2008; XXX(4): 319–326.

18.

Adeyeye

Oyawale

. Weld-metal property optimization from flux ingredients through mixture experiments and mathematical programming approach. Mater Res 2009; 12(3): 1–9.

19.

Bang

K-s

Park

Jung

H-c

. Effects of flux composition on the element transfer and mechanical properties of weld metal in submerged arc welding. Met Mater Int 2009; 15(3): 471–477.

20.

Kumar

Batish

Bhattacharya

. Effect of process parameters on microhardness of heat affected zone in submerged arc welding. Proc IMechE, Part B: J Engineering Manufacture 2010; 225: 711–721.

21.

Arnulf

Osborn

. Phase equilibria among oxides in steel making. Massachusetts: Addison-Wesley, 1965.

22.

Kumar

. Modeling of weld bead geometry and shape relationship in submerged arc welding using developed fluxes. Jordan J Mech Ind Eng 2011; 5(5): 461–470.

23.

Box

GEP

Draper

. Evolutionary operation. New York: John Wiley, 1969.

24.

Khuri

Cornell

. Response surfaces design and analysis. 2nd ed. New York: Dekker.

25.

Prakash

Kumar

. Modeling for prediction of surface roughness in drilling MDF panels using response surface methodology. J Compos Mater 2011; 45(16): 1639–1646.

26.

Montgomery

. Design and analysis of experiments. 5th ed. New Delhi, India: Wiley, 2007.

Modeling and analysis of elongation for submerged arc-welded joints by using developed agglomerated fluxes

Abstract

Keywords

Introduction

Selection of flux constituents and their ranges

Response surface methodology technique

Flux preparations

Selection of process parameters of SAW

Experimentation

Results and discussion (response surface modeling of percentage elongation)

Confirmation experiments

Conclusion

Footnotes

Acknowledgements

Funding

Declaration of conflicting interests

References