Sage Journals: Discover world-class research

Abstract

The bimetallic welds are frequently utilized for pipeline transport system of the nuclear power plants. The occurrences of welding defects generally depend on the filler electrode as well as the electrode coatings during shielded metal arc welding process. This study involves the design of austenitic stainless steel welding electrodes for SS304L–SA516 bimetallic welds. The objective of research work includes the novel design of Al₂O₃–TiO₂–CaO–SiO₂ coatings by combining two ternary phase systems using extreme vertices mixture design methodology to analyze the effect of key coating constituents on the weld metal chemistry and mechanical properties of the welds. The significant effect of electrode coating constituent CaO on weld metal manganese content is observed which further improves the toughness of bimetallic weld joints. Various regression models have been developed for the weld responses and multi objective optimisation approach using composite desirability function has been adopted for identifying the optimized set of electrode coating compositions. The role of delta ferrite content in promoting the favourable solidification mode has been studied through microstructural examination.

Keywords

Bimetallic welds ternary phase diagrams multi-objective optimisation regression models microstructure

Introduction

The bimetallic welds have applications in the steam generators and reactor pressure vessels of nuclear power plants. These welds represent the weld joints between the austenitic stainless steels and ferritic low alloy steels. The certain characteristics of austenitic stainless steels which include good mechanical behaviour at elevated temperatures make them most suitable to be used in certain sections such as superheaters and reheaters while the use of low alloy steels is preferred in the construction of reactor vessels particularly due to their low cost. The austenitic stainless steels show good weldability through the traditional arc welding processes like shielded metal arc welding (SMAW), submerged arc welding (SAW) etc.

In SMAW process, the design of welding electrode coatings is an area of research. The research work carried out by authors^1–6 mainly focused on similar welds using different coatings constituents. Mitra and Eagar^7–10 studied the transfer of alloying elements during slag-metal reactions in submerged arc welding process. Pandey et al.¹¹ analyzed the weld metal chemistry with different welding fluxes for SAW welded joints. The authors^12–15 used the statistical design of experiment (DoE) techniques to predict the effect of welding fluxes and process parameters on weld chemistry as well as mechanical properties of similar welds. Zhan et al.¹⁶ developed an intelligent welding procedure qualification based on artificial neural networks. The developed expert system could predict the welding process parameters along with the mechanical properties for a SMAW process.

The important findings of nickel (Ni) based welding consumables for the joining of stainless steels with high nickel alloys are reported by the researchers.^17–21 The major drawback of Ni-based welding consumables is their inferior weldability as compared to austenitic stainless steels. Further, the bimetallic welds with these consumables are also not immune to failure. The composition of bimetallic weld is largely dependent on the dilution ratios of base metals and the welding filler wire along with elemental diffusion across the weld.^22,23

Lee et al.²⁴ investigated the electrical and mechanical behaviour of friction welded Cu/Al bimetallic joints during annealing. The researchers applied two kinds of intermetallic compound layers and found the significant increase in electrical resistivity with the increase in thickness of intermetallic compound layers. It was also observed the remarkable decrease in tensile strength of bimetallic joints to near zero at the annealing condition of 773K. Zhang et al.²⁵ found good shear strength in explosive welded sheets of X65 pipe steel and 2205 stainless steel. The microstructural analysis highlighted the coarse columnar crystal formed in the melted zone near the interface of base materials. Ning et al.²⁶ examined the laser butt welded Ti-Steel bimetallic joint by using a copper interlayer. The purpose of the interlayer was to reduce the transition zone between base materials. Zhan et al.²⁷ studied the grain morphology along with the temperature field simulation of dissimilar joint between Ti-6Al-4V and 1050 aluminum alloy using laser welding-brazing technique. The authors also developed a 3D finite element model to find out the temperature field. The authors^28–31 explored the mixture design methodology for the design and development of electrode coatings for bimetallic welds. The researchers^32–35 investigated the different flux systems to design and develop SAW fluxes for various applications. Khan and Chhibber³⁶ studied the mechanical behaviour of filler metal on dissimilar welds.

The literature review shows that limited work is published on design of welding electrode consumables for bimetallic welds in SMAW process. There is lack of a scientific approach towards the design of electrode coatings for bimetallic welds involving studies on effect of electrode coating constituents on microstructural evolution and mechanical property enhancement. In the present work, the CaO–SiO₂–Al₂O₃–TiO₂ based electrode coatings with austenitic stainless steel fillers have been designed and developed for SMAW process. Total 21 electrode coating formulations have been designed using extreme vertices mixture design methodology and then the effect of various coating constituents on the weld chemistry and mechanical properties of bimetallic welds has been investigated. The paper also examines the role of delta ferrite content during solidification through metallurgical investigations.

Design approach

The extreme vertices mixture design methodology was used for the development of welding electrode coatings.³⁷ According to this method, for a mixture of k constituents with each constituent having lower limit (α_i) and upper limit (β_i) as constraint, the percentage composition (x_i) of ith constituent and the constrained mixture design can be mathematically represented as follows:³⁸

0 \leq α_{i} \leq x_{i} \leq β_{i} \leq 100

(1)

\sum_{i = 1}^{k} x_{i} = 100

(2)

The primary requirement for deciding the formulations of electrode coatings is that the composite melting temperature of the coating should be lower as compared to the core wire and the parent materials. So, the percentage compositions of various constituents are widely considered by analyzing the appropriate experimental phase diagrams. Depending upon the relevant previous research, the ranges of electrode coating constituents are further narrowed.²²

In present work, phase diagrams of CaO–SiO₂–TiO₂ and CaO–SiO₂–Al₂O₃ systems as shown in Figure 1(a) and (b) were used to decide the percentage composition of electrode coating constituents by selecting different points (points marked with X) on these phase diagrams. The ranges of electrode coating constituents (CaO, SiO₂, Al₂O₃ and TiO₂) were estimated while taking into consideration the other essential electrode coating constituents in the complete composition and are given below:

\begin{matrix} 15 \leq CaO \leq 35; 15 \leq Si O_{2} \leq 35; 8 \leq A l_{2} O_{3} \\ \leq 18; 10 \leq Ti O_{2} \leq 35 \end{matrix}

(3)

and

\sum_{i = 1}^{4} x_{i} = 75

(4)

Figure 1.

Ternary plots of (a) CaO–SiO₂–TiO₂³⁹ and (b) CaO–SiO₂–Al₂O₃ systems.⁴⁰

The Table 1 shows the 21 electrode coating formulations finalized on the basis of extreme vertices mixture design method. Each formulation represents the nature of point (vertex or centroid) on the constrained mixture design space as shown in Figure 2.

Table 1.

Electrode coating compositions.

S.No.	Type of pointin design space	Constituents of electrode coatings
S.No.	Type of pointin design space	CaO	SiO₂	Al₂O₃	TiO₂
1	Vertex	15	32	18	10
2	Centroid	23.17	23.17	8	20.67
3	Vertex	15	15	18	27
4	Vertex	15	17	8	35
5	Vertex	15	35	15	10
6	Vertex	22	35	8	10
7	Centroid	24.83	15	12.83	22.33
8	Centroid	25.67	25.67	13.67	10
9	Vertex	35	15	8	17
10	Vertex	35	15	15	10
11	Centroid	35	17.33	10.33	12.33
12	Centroid	15	24.83	12.83	22.33
13	Vertex	15	35	8	17
14	Centroid	20.67	20.67	18	15.67
15	Centroid	15.67	15.67	8.67	35
16	Centroid	22.17	22.17	11.83	18.83
17	Vertex	17	15	8	35
18	Centroid	17.33	35	10.33	12.33
19	Vertex	35	22	8	10
20	Vertex	15	15	10	35
21	Vertex	32	15	18	10

Figure 2.

Constrained mixture design space.

Experimentation

Electrode coating formulation

The various electrode coating constituents were weighed as described/formulated in design matrix (Table 1) and then mixed properly with dry mixer. After adding the suitable binder in the dry mixture and followed by further mixing in the wet mixer, the electrodes were extruded using SS308L and SS309L core wires having diameter 3.15 mm and length 350 mm. Each set of electrode coating formulation as given in Table 1 was used to develop two different types of coated welding electrodes that is, one with SS308L core wires and another with SS309L core wires.

For dissimilar welding, the base plates of SS304L and SA516 materials were used and weld coupons of dimensions as described in Figure 3 and Table 2 were prepared using developed coated electrodes. For welding, each set of developed SS309L core wire welding electrodes was used for buttering the SA516 base plates and then SS308L core wire electrodes having same type of electrode coating formulation were subsequently used to make the bimetallic welds by joining the buttered SA516 plates with SS304L plates.

Figure 3.

Weld coupon.

Table 2.

Weld coupon specifications.

Symbol	Size
W	80 mm
T _P	32 mm
T _S	6 mm
R	16 mm
O	7 mm
θº	60º

W: plate width; T_P: plate thickness; T_S: strip thickness; R: root gap; O: overlap; θ°: groove angle.

Welding parameters

To choose the welding parameters bead-on-plate tests were performed by varying the welding current (80–100 A) whilst keeping the welding speed at approximately 120mm/min. Base plates were butt welded using optimized welding process parameters.

Results and discussion

The chemical composition analysis of base plates along with core wires (Table 3) and of bimetallic welds between SS304L and SA516 base plates (Table 4) was determined using optical emmision spectrometer. The delta ferrite content (δ_f) which mainly prevents cracking in bimetallic welds during solidification, is represented by ferrite number as given in Table 5. It (δ_f) was measured using ferritescope and WRC 92 diagram.⁴¹

Table 3.

Elemental compositional analysis of base metals and core wires.

Material	%Cr	%Ni	%C	%Si	%Mn	%P	%S	%Mo	%Fe
SS304L	18.2	8.5	0.025	0.5	1.6	0.045	0.030	–	Remaining
SA516	–	–	0.19	0.3	1.1	0.03	0.030	0.08	Remaining
SS308L	20	9.9	0.02	0.4	1.5	–	–	–	Remaining
SS309L	23.5	13	0.02	0.4	1.7	–	–	–	Remaining

Table 4.

Chemical composition (wt%) of bimetallic welds.

Expt. no.	%Si	%Mn	%Cr	%Ni	%C
1	0.74	1.22	17.89	11.01	0.043
2	0.66	1.28	18.13	10.76	0.039
3	0.58	1.36	18.80	10.20	0.039
4	0.59	1.32	18.53	10.36	0.042
5	0.79	1.20	17.72	10.92	0.042
6	0.76	1.22	17.89	10.84	0.038
7	0.59	1.34	18.69	10.20	0.039
8	0.69	1.26	18.05	10.76	0.040
9	0.58	1.35	18.93	9.96	0.038
10	0.58	1.34	18.77	10.12	0.037
11	0.62	1.32	18.93	9.94	0.036
12	0.68	1.28	18.13	10.76	0.042
13	0.79	1.22	17.80	10.75	0.042
14	0.64	1.29	18.37	10.52	0.037
15	0.58	1.33	18.61	10.28	0.041
16	0.67	1.28	18.29	10.60	0.039
17	0.57	1.34	18.53	10.36	0.040
18	0.79	1.21	17.80	10.78	0.040
19	0.61	1.30	18.53	10.36	0.037
20	0.58	1.34	18.71	10.10	0.040
21	0.59	1.32	18.77	10.32	0.036

Table 5.

Results of ferrite number (FN).

Expt. no.	Cr/Ni equivalentratio	Ferrite number		Solidificationmode
Expt. no.	Cr/Ni equivalentratio	Ferritescope	WRC 92diagram	Solidificationmode
1	1.43	1.4	0.4	FA
2	1.50	2.8	2.0	FA
3	1.63	5.1	6.0	FA
4	1.57	4.2	3.5	FA
5	1.43	1.0	0.1	AF
6	1.47	1.2	2.0	FA
7	1.62	6.1	4.8	FA
8	1.49	3.3	1.8	FA
9	1.68	7.1	6.0	FA
10	1.64	6.3	5.5	FA
11	1.69	8.0	6.5	FA
12	1.48	2.1	1.2	FA
13	1.46	1.4	0.8	FA
14	1.55	3.7	3.0	FA
15	1.59	4.7	5.8	FA
16	1.53	2.9	3.9	FA
17	1.57	4.0	4.8	FA
18	1.46	1.5	0.8	FA
19	1.59	4.6	3.0	FA
20	1.63	4.9	5.6	FA
21	1.62	6.1	5.0	FA

FA: ferritic–austenitic; AF: austenitic–ferritic.

Different samples were cut from the welded coupons for testing of various properties such as tensile test, impact test and microhardness of bimetallic welds. The results of different properties are given in Table 6. The typical microstructures were obtained at 250X magnification from weld metal specimens.

Table 6.

Bimetallic weld properties.

Expt. no.	UTS (MPa)	Impacttoughness(N m)	Microhardness(VHN)
1	520	181	221
2	590	189	200
3	570	178	204
4	560	180	207
5	525	180	219
6	560	176	207
7	608	192	191
8	567	182	205
9	620	195	186
10	600	190	188
11	610	193	190
12	545	185	212
13	540	180	218
14	580	185	200
15	565	180	205
16	580	188	200
17	573	184	203
18	532	181	217
19	595	185	195
20	570	180	204
21	598	188	190

Regression models

The experimental data of various weld responses was used to formulate the regression models in terms of % composition of individual coating constituents and their binary mixtures. The various developed regression models were:

\begin{matrix} % Si = 0.004 CaO + 0.017 Si O_{2} - 0.01 A l_{2} O_{3} \\ + 0.003 Ti O_{2} - 0.0001 CaO . Si O_{2} \\ + 0.0005 CaO . A l_{2} O_{3} + 0.0001 CaO . Ti O_{2} \\ + 0.0003 Si O_{2} . A l_{2} O_{3} + 0.000045 Si O_{2} . Ti O_{2} \\ + 0.0004 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(5)

\begin{matrix} % Mn = 0.021 CaO + 0.016 Si O_{2} + 0.019 A l_{2} O_{3} \\ + 0.018 Ti O_{2} - 0.00009 (CaO . Si O_{2} - CaO . A l_{2} O_{3}) \\ - 0.00001 CaO . Ti O_{2} - 0.0001 Si O_{2} . A l_{2} O_{3} \\ - 0.00007 Si O_{2} . Ti O_{2} + 0.0001 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(6)

\begin{matrix} % C = 0.0008 CaO + 0.0004 Si O_{2} - 00003 A l_{2} O_{3} \\ + 0.0006 Ti O_{2} - 0.00002 CaO . Si O_{2} \\ + 0.00005 CaO . A l_{2} O_{3} - 0.000008 CaO . Ti O_{2} \\ + 0.00003 Si O_{2} . A l_{2} O_{3} + 0.000007 Si O_{2} . Ti O_{2} \\ + 0.000006 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(7)

\begin{matrix} % Cr = 0.32 CaO + 0.26 Si O_{2} + 0.29 A l_{2} O_{3} \\ + 0.26 Ti O_{2} - 0.002 CaO . Si O_{2} \\ - 0.002 CaO . A l_{2} O_{3} - 0.001 CaO . Ti O_{2} \\ - 0.002 Si O_{2} . A l_{2} O_{3} - 0.001 Si O_{2} . Ti O_{2} \\ + 0.0007 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(8)

\begin{matrix} % Ni = 0.06 CaO + 0.09 Si O_{2} + 0.25 A l_{2} O_{3} \\ + 0.13 Ti O_{2} + 0.003 CaO . Si O_{2} \\ - 0.0006 CaO . A l_{2} O_{3} \\ + 0.001 CaO . Ti O_{2} - 0.0003 Si O_{2} . A l_{2} O_{3} \\ + 0.0002 Si O_{2} . Ti O_{2} - 0.004 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(9)

\begin{matrix} Ferrite number = 0.52 CaO + 0.27 Si O_{2} + 0.19 A l_{2} O_{3} \\ + 0.14 Ti O_{2} - 0.02 CaO . Si O_{2} \\ - 0.01 CaO . A l_{2} O_{3} - 0.006 CaO . Ti O_{2} - 0.008 Si O_{2} . A l_{2} O_{3} \\ - 0.009 Si O_{2} . Ti O_{2} + 0.003 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(10)

\begin{matrix} Ultimate tensile strength = 6.08 CaO + 6.46 Si O_{2} \\ + 12.08 A l_{2} O_{3} + 4.55 Ti O_{2} + 0.08 CaO . Si O_{2} \\ - 0.02 CaO . A l_{2} O_{3} + 0.21 CaO . Ti O_{2} - 0.18 Si O_{2} . A l_{2} O_{3} \\ + 0.01 Si O_{2} . Ti O_{2} - 0.02 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(11)

\begin{matrix} Impact toughness = 2.56 CaO + 1.82 Si O_{2} - 0.95 A l_{2} O_{3} \\ + 0.99 Ti O_{2} - 0.02 CaO . Si O_{2} \\ + 0.06 CaO . A l_{2} O_{3} + 0.06 CaO . Ti O_{2} \\ + 0.09 Si O_{2} . A l_{2} O_{3} + 0.04 Si O_{2} . Ti O_{2} \\ + 0.06 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(12)

\begin{matrix} Microhardness = 3.12 CaO + 2.91 Si O_{2} + 2.72 A l_{2} O_{3} \\ + 3.42 Ti O_{2} - 0.02 CaO . Si O_{2} - 0.04 CaO . A l_{2} O_{3} \\ - 0.06 CaO . Ti O_{2} + 0.04 Si O_{2} . A l_{2} O_{3} + 0.01 Si O_{2} . Ti O_{2} \\ - 0.01 A l_{2} O_{3} . Ti O_{2} \end{matrix}

(13)

Analysis of variance (ANOVA)

The developed regression models have been analyzed using analysis of variance (ANOVA) that is, t-test and F-test. The significance of various regression coefficients have been tested by comparing their t-values with tabulated t-values at 95% confidance level. The regression analysis of weld chemistry and mechanical properties of bimetallic welds is presented in Tables 7 and 8 respectively. The Tables 9 to 11 depict the t-test analysis for significance of coefficients of regression models of various weld responses. The result analysis of regression models reveals that there are some insignificant terms in the models. These insignificant terms have been removed to improve the models. The summary of these tests recommends the adequacy of developed regression models.

Table 7.

Regression analysis of weld chemistry.

S. no.	Source	SS	DF	MS	F value	p Value	Status
%Si	Model	0.121884	9	0.013543	243.5456	<0.0001	ST
	Residual	0.000612	11	5.56E-05
	Total	0.122496	20	R ²	0.995007
	Std. dev.	0.007457		R ² _adj	0.990921
	Mean	0.652271		R ² _pred.	0.977657
	CV	1.143232		Adeq precision	41.65752
%Mn	Model	0.050658	9	0.005629	305.088	<0.0001	ST
	Residual	0.000203	11	1.84E-05
	Total	0.050861	20	R ²	0.99601
	Std. dev.	0.004295		R ² _adj	0.992745
	Mean	1.291719		R ² _pred.	0.984692
	CV	0.332523		Adeq precision	49.8981
%C	Model	7.29E-05	9	8.09E-06	9.782688	0.0004	ST
	Residual	9.1E-06	11	8.27E-07
	Total	8.2E-05	20	R ²	0.888938
	Std. dev.	0.00091		R ² _adj	0.79807
	Mean	0.03943		R ² _pred.	0.567264
	CV	2.306997		Adeq precision	9.440938
%Cr	Model	3.182276	9	0.353586	62.52626	<0.0001	ST
	Residual	0.062205	11	0.005655
	Total	3.244481	20	R ²	0.980827
	Std. dev.	0.0752		R ² _adj	0.965141
	Mean	18.37417		R ² _pred.	0.928873
	CV	0.409269		Adeq precision	22.43672
%Ni	Model	2.026763	9	0.225196	23.97828	<0.0001	ST
	Residual	0.103308	11	0.009392
	Total	2.130071	20	R ²	0.9515
	Std. dev.	0.096911		R ² _adj	0.911818
	Mean	10.47217		R ² _pred.	0.810802
	CV	0.925411		Adeq precision	15.63313
Ferrite number	Model	75.32857	9	8.369842	36.47992	<0.0001	ST
	Residual	2.523806	11	0.229437
	Total	77.85238	20	R ²	0.967582
	Std. dev.	0.478996		R ² _adj	0.941058
	Mean	3.747619		R ² _pred.	0.880296
	CV	12.78134		Adeq Ppecision	18.09278

SS: sum of squares; DF: degree of freedom; MS: mean squares; R²: determination coefficient; CV: coefficient of variation; ST: significant.

Table 8.

Regression analysis of mechanical properties.

S. no.	Source	SS	DF	MS	F value	p Value	Status
Ultimate tensile strength	Model	15892.1	9	1765.789	84.76826	<0.0001	ST
	Residual	229.1386	11	20.83078
	Total	16121.24	20	R ²	0.985787
	SD	4.564075		R ² _adj	0.974157
	Mean	571.8095		R ² _pred.	0.941889
	CV	0.798181		Adeq precision	32.57569
Impact toughness	Model	548.6184	9	60.9576	54.36458	<0.0001	ST
	Residual	12.33402	11	1.121274
	Total	560.9524	20	R ²	0.978012
	SD	1.058902		R ² _adj	0.960022
	Mean	184.381		R ² _pred.	0.916537
	CV	0.574301		Adeq precision	27.77454
Micro-hardness	Model	2185.334	9	242.8149	121.4938	<0.0001	ST
	Residual	21.98436	11	1.998578
	Total	2207.318	20	R ²	0.99004
	SD	1.413711		R ² _adj	0.981891
	Mean	202.9569		R ² _pred.	0.945979
	CV	0.696557		Adeq precision	35.28195

SS: sum of squares; DF: degree of freedom; MS: mean squares; R²: determination coefficient; CV: coefficient of variation; ST: significant.

Table 9.

Analysis of regression coefficients of weld responses (%Si, %Mn, %C) using t-test.

Regression coefficient	%Si		%Mn		%C
Regression coefficient	t-Value	Status	t-Value	Status	t-Value	Status
CaO	2.95	ST	26.63	ST	4.50	ST
SiO₂	12.75	ST	20.36	ST	2.76	ST
Al₂O₃	−1.95	ST	4.24	ST	−0.34	NST
TiO₂	2.69	ST	34.21	ST	5.55	ST
CaO.SiO₂	−2.03	ST	−3.01	ST	−1.86	ST
CaO.Al₂O₃	3.10	ST	−0.99	NST	0.26	NST
CaO.TiO₂	1.98	ST	−0.43	NST	−1.34	NST
SiO₂.Al₂O₃	1.81	ST	−1.83	ST	1.61	NST
SiO₂.TiO₂	0.69	NST	−2.30	ST	1.06	NST
Al₂O₃.TiO₂	2.43	ST	1.41	NST	0.29	NST

Regression coefficients significant if |t| > t_{α, DF(error)} (t_0.05,11 = 1.796); ST: significant; NST: not significant.

The F value of regression model is the ratio of mean square value of model to the mean square value of residual. It basically compares the model variance with the residual variance.⁴² Therefore, for F value close to one, there is lesser chance of model significance. At 95% confidence level, the p value less than 0.05 represents the significance of model. The F values and its p values of various weld responses as given in Tables 7 and 8 clearly show the significance of developed regression models.

The coefficient of determination (R²) values of regression models are ≥90% which indicate that predicted values of various weld responses are in close agreement with their actual values as shown in Figures 4(a) to (f) and 5(a) to (c). The coefficient of variation (CV) is calculated as the standard deviation divided by the mean.⁴² The low values of CV as given in Tables 7 and 8 indicate the good fit model. The summary of ANOVA recommends the adequacy of developed regression models.

Figure 4.

Comparison plots of predicted versus actual weld chemistry of bimetallic welds: (a) %C, (b) %Cr, (c) %Ni, (d) %Mn, (e) %Si and (f) ferrite number.

Figure 5.

Comparison plots of predicted versus actual mechanical properties of welds: (a) UTS, (b) impact toughness and (c) microhardness.

Role of electrode coating constituents on weld chemistry

The results show that percentage of silicon in weld gets significantly increased with the increase in SiO₂ content of coatings. SiO₂ gets dissociated into silicon and oxygen during slag-metal reactions and thereby raises the silicon content in the weld.⁴³ Al₂O₃ tends to decrease the weld metal silicon content as it reacts with the fluorides present in the electrode coatings to form AlF₃ compound which further reacts with silicon according to the following reaction:⁴⁴

4 Al F_{3} + 3 Si \to 3 Si F_{4} + 4 Al

(14)

The coating constituents CaO and TiO₂ show increasing effect on weld metal silicon content. The effect of binary mixture CaO.SiO₂ is to decrease the silicon content as the CaO has a tendency to decrease the SiO₂ activity in slag-metal reactions and thus decrease the silicon content in weld composition according to the following reaction:⁴

CaO \to C a^{2 +} + O^{2 -}

(15)

Si O_{2} + 2 O^{2 -} \to Si {O_{4}}^{4 -}

(16)

The individual effect of all the constituents of electrode coatings that is, CaO, SiO₂, Al₂O₃ and TiO₂ is increasing on the manganese content of weld metal. The transfer of manganese content in weld is given by Pandey et al.:¹¹

Mn + O \to MnO

(17)

The CaO increases the activity of MnO during slag-metal reactions thereby shifting reaction to the left side which is also confirmed by Sharma and Chhibber.³³ The effect of TiO₂ could be explained from the fact that it decreases the viscosity of slag during slag-metal reactions and hence promotes the transfer of manganese to weld. All the binary mixtures of SiO₂ show the reversing trend on manganese content of weld metal due to the increase in the concentration of MnO in slag resulting in the lower level of manganese content in the weld.⁷

The individual electrode coating constituents except Al₂O₃ show the increasing effect on the carbon content of weld metal. CaO.SiO₂ shows the single significant decreasing effect on weld metal carbon content whereas the interaction effects of other binary mixtures are not significant as shown in Table 10. The regression analysis reveals that the electrode coating constituents and the binary mixtures (CaO.SiO₂, CaO.TiO₂, SiO₂.TiO₂) increase the nickel content of the weld metal while the interaction effect of binary mixture Al₂O_3.TiO₂ is significantly decreasing.

Table 10.

Analysis of regression coefficients of weld responses (%Cr, %Ni, ferrite number) using t-test.

Regression coefficient	%Cr		%Ni		Ferrite number
Regression coefficient	t-Value	Status	t-Value	Status	t-Value	Status
CaO	22.52	ST	3.50	ST	5.95	ST
SiO₂	18.33	ST	5.19	ST	3.09	ST
Al₂O₃	3.51	ST	2.43	ST	0.38	NST
TiO₂	27.26	ST	10.63	ST	2.32	ST
CaO.SiO₂	−3.92	ST	4.72	ST	−5.26	ST
CaO.Al₂O₃	−1.37	NST	−0.30	NST	-0.97	NST
CaO.TiO₂	−2.27	ST	2.05	ST	−1.85	ST
SiO₂.Al₂O₃	−1.13	NST	-0.12	NST	−0.84	NST
SiO₂.TiO₂	−2.36	ST	2.99	ST	−2.74	ST
Al₂O_3.TiO₂	0.41	NST	−1.84	ST	0.31	NST

Regression coefficients significant if |t| > t_{α, DF(error)} (t_0.05,11 = 1.796); ST: significant; NST: not significant.

Role of electrode coating constituents on solidification behaviour

Electrode coating constituents promote the delta ferrite having vermicular morphology in welds as compared to sigma phase as is clear from the microstructures (as explained later in the microstructure analysis section of this paper). The results of ferrite number as given in Table 5 show that the solidification modes are either ferritic–austenitic (FA) or austenitic–ferritic (AF).

Role of electrode coating constituents on mechanical properties of weld

The regression analysis reveals that all electrode coating constituents have significant increasing effect on the ultimate tensile strength of weld (Table 11). The electrode coating constituents tend to promote inclusions in the form of metal sulfides and oxides in weld which may further assist in the ferrite nucleation at the interface between austenite matrix and inclusions. The plasticity of the material gets improved with the finely dispersed spheroid-shaped inclusions. Al₂O₃ tends to deoxidize the weld pool, thereby increasing the UTS of the weld. This result is in agreement with the research work reported by Jindal et al.³⁸ The effect of CaO.SiO₂ is to increase the UTS as the SiO₂ activity is reduced due to the formation of SiO₄⁴⁻ complex ion.⁴ Therefore, the excessive dissociation of SiO₂ gets reduced and the amount of dissolved oxygen remains limited. The binary mixture SiO₂.Al₂O₃ shows significant decreasing effect on ultimate tensile strength due to their combined increasing effect on O₂ in the weld.

Table 11.

t-Test analysis of regression coefficients of mechanical properties.

Regression coefficient	Ultimate tensile strength		Impact toughness		Microhardness
Regression coefficient	t-Value	Status	t-Value	Status	t-Value	Status
CaO	7.25	ST	13.16	ST	11.99	ST
SiO₂	7.70	ST	9.39	ST	11.23	ST
Al₂O₃	2.42	ST	-0.82	NST	1.77	NST
TiO₂	8.01	ST	7.57	ST	19.40	ST
CaO.SiO₂	2.39	ST	−2.18	ST	−1.81	ST
CaO.Al₂O₃	−0.15	NST	2.72	ST	−1.16	NST
CaO.TiO₂	6.44	ST	7.46	ST	−5.60	ST
SiO₂.Al₂O₃	−1.80	ST	3.84	ST	1.26	NST
SiO₂.TiO₂	0.31	NST	5.45	ST	1.26	NST
Al₂O₃.TiO₂	−0.17	NST	2.35	ST	−0.57	NST

Regression coefficients significant if |t |> t_{α, DF(error)} (t_0.05,11 = 1.796); ST: significant; NST: not significant.

The basicity index of electrode coatings is increased with increase in CaO constituent which further increases the impact toughness of welds as also reported in literature.⁴⁵ The TiO₂ constituent shows an increasing effect on the impact toughness of bimetallic welds. Kohno et al.⁴⁶ highlighted the grain refining property of TiO₂ as the primary cause for the increase in impact toughness of welds. The interaction effects of all binary mixtures except CaO.SiO₂ show the increasing trend on impact toughness. It is observed from regression analysis that the electrode coating constituents CaO, SiO₂ and TiO₂ increase the microhardness of weld while the effect of Al₂O₃ is not significant.

Contour plots

The contour surface plots of predicted values of chemical composition and mechanical properties of bimetallic welds for varying amounts of electrode coating constituents with constant TiO₂ = 18.83% are shown in Figures 6(a) to (f) and 7(a) to (c), respectively.

Figure 6.

Contour surface plots for: (a) %C, (b) %Cr, (c) %Ni, (d) %Mn, (e) %Si and (f) ferrite number.

Figure 7.

Contour surface plots for: (a) UTS, (b) impact toughness and (c) microhardness.

Microstructure analysis

The microstructures of various weld specimens are shown in Figure 8(a) to (c). The microstructures of specimens indicate the ferritic–austenitic (FA) as the primary solidification mode and the presence of austenite with varying amount of delta ferrite (δ_f) of vermicular ferrite (VF) type. The microstructure of weld specimen (coating formulation no. 1) shows the presence of some large size inclusions (I) on grain boundaries in Figure 8(a). This electrode coating formulation contains higher amount of SiO₂ and Al₂O₃ (Table 1) and their binary interaction promotes the large size inclusions which adversely affects the mechanical properties of the weld (Table 6). The delta ferrite may also transform to lathy ferrite (LF) in addition to vermicular ferrite as shown in the weld microstructure of coating formulation no. 2 (Figure 8(b)). This weld specimen has better mechanical properties as given in Table 6 due to the presence of small size spherical shaped inclusions (size 0.4 µm) inside the grains. This can be attributed to relative higher amount of TiO₂ and CaO in the electrode coating formulation which further assists in promoting the small size inclusions in weld. The inclusions of intermediate size 2 µm were observed in the microstructure of weld specimen of coating formulation no. 7 as shown in Figure 8(c). During SMAW process, the slag-metal reactions control the interaction of various electrode coating constituents with molten weld pool which further result in the variation of elements transfer in the weld. The variation in weld chemistry leads to the different mechanical properties and microstructures.

Figure 8.

Weld zone microstructures obtained from: (a) experiment no. 1, (b) experiment no. 2 and (c) experiment no. 7.

Validation of models

To validate the different regression models for various weld properties, three coating compositions were randomly chosen to confirm or validate with the experimental results. Table 12 shows the error percentage for different weld properties which is approximately ≤5% in most results.

Table 12.

Error percentage for different properties.

Electrode coating constituents				Predicted value			Actual value			Error (%)
CaO	SiO₂	Al₂O₃	TiO₂	UTS(MPa)	IT (N m)	MH(VHN)	UTS(MPa)	IT(N m)	MH(VHN)	UTS(MPa)	IT(N m)	MH(VHN)
20	20	15	20	577.5	186.9	201.4	595	195	190	3.02	4.35	5.65
30	15	15	15	608.2	191.8	188.2	580	182	198	4.64	5.11	5.21
15	20	10	30	556.9	183.1	209.2	585	192	219	5.05	4.87	4.70

Error (%) = (actual value-predicted value) × 100/(actual value); IT: impact toughness; MH: microhardness.

Optimisation methodology

Derringer and Suich⁴⁷ proposed a multi objective optimisation methodology in which the simultaneous optimisation of several output responses is reduced to a single response optimisation problem with the help of composite desirability. The desirability function transforms the response variable into a value (on 0 to 1 scale). The composite desirability combines the individual desirabilities and highlights the relative importance of responses. Depending upon the nature of the output response, the different weights are assigned in this optimisation method.^48,49 Table 13 shows the three multi objective optimized solutions with equal importance to mechanical properties along with the target value of ferrite number. The target value of ferrite number was chosen in accordance with Bystram.⁵⁰ The graphical demonstration of desirability of optimized solution no. 2 is shown in Figure 9. It clearly indicates that the desirability of all the electrode coating constituents along with ferrite number is one while the composite desirability of complete solution is 0.87.

Table 13.

Optimisation of different properties.

S. no.	Electrode coating constituents				UTS (MPa)	IT (N m)	MH (VHN)	FN	D
S. no.	CaO	SiO₂	Al₂O₃	TiO₂	UTS (MPa)	IT (N m)	MH (VHN)	FN	D
1	28.07	16.67	10.69	19.56	610	192	193	5.1	0.916609
2	32.73	17.72	9.39	15.16	610	190	193	5.5	0.875076
3	25.41	15.00	13.49	21.10	600	187	192	5.4	0.830636

IT: impact toughness; MH: microhardness; FN: ferrite number; D: desirability.

Figure 9.

Graphical demonstration of variation of desirability of optimized solution.

Conclusions

Welding electrodes have been successfully developed based on combined CaO–SiO₂–Al₂O₃ and CaO–SiO₂–TiO₂ phase diagrams using extreme vertices design method for bimetallic welds. SiO₂ increases the weld metal silicon content whereas the tendancy of Al₂O₃ is to decrease it.

TiO₂ promotes the transfer of manganese to weld by decreasing the viscosity of slag during slag metal reactions. Al₂O₃ releases O^-2 ions that react with carbon thereby reducing its composition. CaO tends to increase the manganese content of the weld by increasing the activity of MnO. The interaction effect of coating constituents on the weld metal chromium content is decreasing.

The mechanical properties of bimetallic welds significantly depend on the linear and binary interaction effects of electrode coating constituents. Binary effect of CaO.SiO₂ increases whereas SiO₂.Al₂O₃ significantly decreases the weld strength, while CaO and TiO₂ content increase the weld toughness. The significant increasing effect of electrode coating constituents CaO, SiO₂ and TiO₂ on microhardness has been observed.

The microstructures of the weld samples reveal the presence of inclusions of varying size and the delta ferrite (with vermicular morphology) at the dendrite cores. Higher amount of SiO₂ and Al₂O₃ and their binary interaction promote large size inclusions which drastically reduce the mechanical properties of the welds.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Hazlett

Parker

ER.

Effect of individual coating ingredients on surface tension of iron electrodes. Weld J 1956; 35(3): 113s–114s.

Hazlett

TH.

Coating ingredients influence on surface tension, arc stability and bead shape. Weld J 1957; 36(1): 18s–22s.

Tuliani

Boniszewski

Eaton

NF.

Carbonate fluxes for submerged arc welding of mild steel. Weld Met Fabr 1972; 40(7): 247–250.

Palm

JH.

How fluxes determine the metallurgical properties of submerged arc welds. Weld J 1972; 51(7): 358s–360s.

Ferrera

Olson

DL.

MnO–SiO₂–CaO welding flux system. Weld J 1975; 54(7): 211–215.

Davis

MLE

Coe

. The chemistry of submerged arc welding fluxes. The Welding Institute Research Report, 39/1977/M, London, pp. 1–61.

Mitra

Eagar

TW.

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

Mitra

Eagar

TW.

Slag–metal reactions during welding: Part I. Evaluation and reassessment of existing theories. Metall Trans B 1991; 22B: 65–71.

Mitra

Eagar

TW.

Slag–metal reactions during welding: Part II. Theory. Metall Trans B 1991; 22B: 73–81.

10.

Mitra

Eagar

TW.

Slag–metal reactions during welding: Part III. Verification of the theory. Metall Trans B 1991; 22B: 83–100.

11.

Pandey

Bharti

Gupta

SR.

Effect of submerged arc welding parameters and fluxes on element transfer behaviour and weld-metal chemistry. J Mater Process Technol 1994; 40: 195–211.

12.

Kanjilal

Pal

Majumdar

SK.

Prediction of submerged arc weld metal composition from flux ingredients with the help of statistical design of experiment. Scand J Metal 2004; 33: 146–159.

13.

Kanjilal

Pal

Majumdar

SK.

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

14.

Jindal

Chhibber

Mehta

NP.

Modeling flux chemistry for submerged arc weldments of high-strength low-alloy steel. Proc IMechE, Part B: J Engineering Manufacture 2014; 228(10): 1259–1272.

15.

Jindal

Chhibber

Mehta

NP.

Prediction of element transfer due to flux and optimization of chemical composition and mechanical properties in high-strength low-alloy steel weld. Proc IMechE, Part B: J Engineering Manufacture 2014; 229(5): 785–801.

16.

Zhan

Wei

, et al. The feasibility of intelligent welding procedure qualification system for Q345R SMAW. Int J Adv Manuf Technol 2016; 83(5–8): 765–777.

17.

Maruyama

Arc welding technology for dissimilar joint. Weld Int 2003; 17(4): 276–281.

18.

Sireesha

Albert

Shankar

, et al. A comparative evaluation of welding consumables for dissimilar welds between 316LN austenitic stainless steel and alloy 800. J Nucl Mater 2000; 279(1): 65–76.

19.

Srinivasan

Muthupandi

Dietzel

, et al. An assessment of impact strength and corrosion behaviour of shielded metal arc welded dissimilar weldments between UNS 31803 and IS 2062 steels. Mater Des 2006; 27(3): 182–191.

20.

Naffakh

Shamaniana

Ashrafizadeh

Dissimilar welding of AISI 310 austenitic stainless steel to nickel-based alloy inconel 657. J Mater Process Tech 2009; 209: 3628–3639.

21.

Jeng

Lee

Rehbach

, et al. Effects of Nb on the microstructure and corrosive property in the alloy 690-SUS 304L weldment. Mater Sci Eng A 2005; 397: 229–238.

22.

Bhandari

Chhibber

Arora

Effect of electrode coatings on diffusible hydrogen content, hardness and microstructures of the ferritic heat affected zones in bimetallic welds. Adv Mater Res 2012; 383–390: 4697–4701.

23.

Chhibber

Arora

Gupta

, et al. Use of bimetallic welds in nuclear reactors: associated problems and structural integrity assessment issues. J Mech Eng Sci C 2006; 220: 1121–1133.

24.

Lee

Bang

Jung

SB.

Effects of intermetallic compound on the electrical and mechanical properties of friction welded Cu/Al bimetallic joints during annealing. J Alloys Compd 2005; 390(1–2): 212–219.

25.

Zhang

Pei

Zhang

, et al. Study on the microstructure and mechanical properties of explosive welded 2205/X65 bimetallic sheet. Mater Des 2014; 64: 462–476.

26.

Ning

Zhang

Jiang

, et al. Narrow gap multi-pass laser butt welding of explosion welded CP-Ti/Q235B bimetallic sheet by using a copper interlayer. J Alloys Compd 2017; 701: 587–602.

27.

Zhan

Gao

, et al. Temperature field simulation and grain morphology on laser welding-brazing between Ti–6Al–4V and 1050 aluminum alloy. Mater Res Exp 2019; 6(5): 056551.

28.

Bhandari

Chhibber

Arora

, et al. Investigation of TiO₂–SiO₂–CaO–CaF₂ based electrode coatings on weld metal chemistry and mechanical behaviour of bimetallic welds. J Manuf Process 2016; 23: 61–74.

29.

Bhandari

Chhibber

Arora

, et al. Investigations on weld metal chemistry and mechanical behaviour of bimetallic welds using CaO–CaF₂–SiO₂–Ni based electrode coatings. Proc IMechE L 2019; 233(4): 563–579.

30.

Mahajan

Chhibber

Experimental investigations on P22/P91 dissimilar shielded metal arc welds for power plant applications. Proc IMechE L 2020; 234(10): 1313–1324.

31.

Mahajan

Chhibber

Investigations on dissimilar welding of P91/SS304L using Nickel-based electrodes. Mater Manuf Process 2020; 35(9): 1–14.

32.

Sharma

Chhibber

Study of weld bead chemical, microhardness & microstructural analysis using submerged arc welding fluxes for linepipe steel applications. Ceram Int 2020; 46(15): 24615–24623.

33.

Sharma

Chhibber

Design of CaO–SiO₂–CaF₂ and CaO–SiO₂–Al₂O₃ based submerged arc fluxes for series of bead on plate pipeline steel welds: effect on carbon and manganese content, grain size and microhardness. J Press Vessel Technol 2019; 141: 1–10.

34.

Zhang

Coetsee

Dong

, et al. Element transfer behaviors of fused CaF₂–TiO₂ fluxes in EH36 shipbuilding steel during high heat input submerged arc welding. Metall Mater Trans B 2020; 51(5): 1953–1957.

35.

Zhang

Coetsee

Yang

, et al. Structural roles of TiO₂ in CaF₂–SiO₂–CaO–TiO₂ submerged arc welding fluxes. Metall Mater Trans B 2020; 51(5): 1947–1952.

36.

Khan

Chhibber

Effect of filler metal on solidification, microstructure and mechanical properties of dissimilar super duplex/pipeline steel GTA weld. Mater Sci Eng A 2020; 803: 140476.

37.

Anderson

McLean

Design of experiments: a realistic approach. New York, NY: Marcell Dekker, 1974.

38.

Jindal

Chhibber

Mehta

NP.

Investigation on flux design for submerged arc welding of high-strength low-alloy steel. Proc IMechE, Part B: J Engineering Manufacture 2013; 227: 383–395.

39.

De Vries

Roy

Osborn

Phase equilibria in the system CaO–TiO₂–SiO₂. J Am Ceram Soc 1955; 38: 158–171.

40.

Eriksson

Pelton

AD.

Critical evaluation and optimization of the thermodynamic properties and phase diagrams of the CaO–Al₂O₃, Al₂O₃–SiO₂ and CaO–SiO₂–Al₂O₃ systems. Metall Mater Trans B 1993; 24: 807–816.

41.

Kotecki

Siewert

TA.

WRC-1992 constitution diagram for stainless steel weld metals: a modification of the WRC-1988 diagram. Weld J 1992; 71: 171s–178s.

42.

Aggarwal

Khangura

Garg

RK.

Parametric modeling and optimization for wire electrical discharge machining of Inconel 718 using response surface methodology. Int J Adv Manuf Technol 2015; 79: 31–47.

43.

Kou

Welding metallurgy. A John Wiley & Sons, Inc, 2002.

44.

Lau

Weatherly

McLean

Gas/metal/slag reactions in submerged arc welding using CaO–AI₂O₃ based fluxes. Weld Res 1986; Suppl: 31s–38s.

45.

De Rissone

NMR

Farias

De Souza Bott

, et al. ANSI/AWS A5.1-91 E6013 rutile electrodes: the effect of calcite. Weld J 2002: 113s–124s.

46.

Kohno

Takami

Mori

, et al. New fluxes of improved weld metal toughness for HSLA Steels Weld J 1982; 61(12): 373s–380s.

47.

Derringer

Suich

Simultaneous optimization of several response variables. J Qual Tech 1980; 12(9): 214–219.

48.

Harington

The desirability function. Ind Qual Control 1965; 21: 494–498.

49.

Castillo

Montgomery

McCrville

DR.

Modified desirability functions for multiple response optimization. J Qual Tech 1996; 28: 337–345.

50.

Bystram

MCT

. Consumable for fusion welding of dissimilar metals. Met Constr Br Weld J 1969; 1: 150.

Expt. no.	UTS (MPa)	Impacttoughness(N m)	Microhardness(VHN)
1	520	181	221
2	590	189	200
3	570	178	204
4	560	180	207
5	525	180	219
6	560	176	207
7	608	192	191
8	567	182	205
9	620	195	186
10	600	190	188
11	610	193	190
12	545	185	212
13	540	180	218
14	580	185	200
15	565	180	205
16	580	188	200
17	573	184	203
18	532	181	217
19	595	185	195
20	570	180	204
21	598	188	190

Expt. no.	UTS (MPa)	Impacttoughness(N m)	Microhardness(VHN)
1	520	181	221
2	590	189	200
3	570	178	204
4	560	180	207
5	525	180	219
6	560	176	207
7	608	192	191
8	567	182	205
9	620	195	186
10	600	190	188
11	610	193	190
12	545	185	212
13	540	180	218
14	580	185	200
15	565	180	205
16	580	188	200
17	573	184	203
18	532	181	217
19	595	185	195
20	570	180	204
21	598	188	190

Combining CaO–SiO 2 –TiO 2 and CaO–SiO 2 –Al 2 O 3 ternary phase systems for design of bimetallic welds

Abstract

Keywords

Introduction

Design approach

Experimentation

Electrode coating formulation

Welding parameters

Results and discussion

Regression models

Analysis of variance (ANOVA)

Role of electrode coating constituents on weld chemistry

Role of electrode coating constituents on solidification behaviour

Role of electrode coating constituents on mechanical properties of weld

Contour plots

Microstructure analysis

Validation of models

Optimisation methodology

Conclusions

Footnotes

Declaration of conflicting interests

Funding

References

Expt. no.	UTS (MPa)	Impacttoughness(N m)	Microhardness(VHN)
1	520	181	221
2	590	189	200
3	570	178	204
4	560	180	207
5	525	180	219
6	560	176	207
7	608	192	191
8	567	182	205
9	620	195	186
10	600	190	188
11	610	193	190
12	545	185	212
13	540	180	218
14	580	185	200
15	565	180	205
16	580	188	200
17	573	184	203
18	532	181	217
19	595	185	195
20	570	180	204
21	598	188	190