Optimization of CO2 Laser Beam Welding Process Parameters to Attain Maximum Weld Strength in Dissimilar Metals

Optimization of CO2 Laser Beam Welding Process Parameters to Attain Maximum Weld Strength in Dissimilar Metals

Available online at www.sciencedirect.com ScienceDirect Materials Today: Proceedings 5 (2018) 6607–6616 www.materialstoday.com/proceedings IMME17 ...

1MB Sizes 0 Downloads 21 Views

Available online at www.sciencedirect.com

ScienceDirect Materials Today: Proceedings 5 (2018) 6607–6616

www.materialstoday.com/proceedings

IMME17

Optimization of CO2 Laser Beam Welding Process Parameters to Attain Maximum Weld Strength in Dissimilar Metals M.P.Prabakaran1, G.R.Kannan2 * 1 2

Department of Mechanical Engineering, NPR College of Engineering and Technology, Dindigul-624401, Tamilnadu, India Department of Mechanical Engineering, PSNA College of Engineering and Technology, Dindigul-624622, Tamilnadu, India

Abstract In this study, influence of CO2 laser beam welding (LBW) process parameters such as laser power (P), welding speed (S), focal distance (F) on dissimilar metals of low carbon steel (AISI 1018) and austenitic stainless steel (AISI 316) was examined by using central composite design (CCD) based response surface methodology (RSM). Experimental investigations were performed on 4 kW CO2 laser (TRUMPF TLC1005) with the laser beam spot diameter of 0.15mm and a focal length of 15-25mm. Analysis of variance (ANOVA) was used to analyze the interaction effect of different parameters on the maximum weld strength. A ratification experiment was also carried out in order to validate the optimal process parameters values. Based on the investigation, the maximum weld strength of 458.21 N/mm2 was observed for dissimilar metal butt joint with the optimized laser power of 2600W, the focal distance of 25 mm and a welding speed of 1.5 m/min. Microstructure studies revealed cellular austenite in the form of dendrite and columnar type features accredited to the higher cooling rates concerned with laser welding process. Energy dispersive X-ray analysis (EDX) is carried out to measure the nature of the chemical element distribution in the weld interface. Microhardness tests showed that the maximum hardness was produced in weldment 151-458 Hv and fusion boundary of austenitic stainless steel 193-434 Hv, fusion boundary of low carbon steel 156-468 Hv. © 2017 Published by Elsevier Ltd. Selection and/or Peer-review under responsibility of International Conference on Emerging Trends in Materials and Manufacturing Engineering (IMME17). Keywords: CO2 welding; dissimilar metals; RSM; Microstructure Analysis; EDX; Microhardness.

* Corresponding author. Tel.: +91-9894429432; E-mail address: [email protected] 2214-7853 © 2017 Published by Elsevier Ltd. Selection and/or Peer-review under responsibility of International Conference on Emerging Trends in Materials and Manufacturing Engineering (IMME17).

6608

Prabakaran et al., / Materials Today: Proceedings 5 (2018) 6607–6616

Nomenclature RSM Response Surface methodology CCD Central Composite Design ASS Austenitic Stainless Steel LCS Low Carbon Steel HAZ Heat Affected Zone FZ Fusion Zone WZ Weld Zone ANOVA Analysis of variance 1. Introduction The laser beam welding has many advantages like low and precise heat input with narrow heat affected zone and low distortion. Also, it is easily amenable to automation with high-speed welding. Laser welding has many potential benefits compared to another welding process with unique features like deeper penetration, high welding speed, high precision, reliability, high efficiency and high productivity. The dissimilar metals like Austenitic stainless steel with low carbon steel are widely used in power generation applications, owing to attributes of good mechanical properties. The adoption of dissimilar metal combinations provides possibilities for the flexible design of the product by using each material efficiently. The laser beam welding can also use complex structure and complex joints of the thin material down to thick material [1]. The LBW process gains more advantages from high power density and welding speed, such as a small heat affected zone, low distortion, high joint strength and high production rate, compared with traditional welding process [2]. Curcio et al. [3] analyzed the laser welding parameters of laser power, welding speed, shielding gas and the process of focusing used among various materials. Torkamany et al. [4] represented the effect of dissimilar laser welding mode on microstructure and mechanical performance of low carbon to austenitic stainless steel and added that the size of fusion zone and dilution percentage of low carbon steel is highly dependent on the transition of laser welding mode from conduction to the keyhole. Li et al. [5] investigated the relationships between the redistribution of elements and the as welded microstructure in the dissimilar fusion zone of AISI304L and AISI12L13. They developed a technique to avoid solidification cracking and micro-fissuring in the fusion zone. Anawa et al. [6] studied the dissimilar LBW of stainless steel and low carbon steel by using statistical methods and reported that the welding speed and laser power have a significant effect on weld bead characteristics. Problems in dissimilar welding arise from the differences of physical and chemical properties between the welding count parts and possible formation of intermetallic brittle phases resulting in the degradation of mechanical properties of the welds [7]. Mai et al. [8] investigate the laser welding of dissimilar metals without filler materials using Nd: YAG laser. Response surface methodology (RSM) is employed to develop mathematical relationships between the welding process parameters and the output variables of the weld joint to determine the welding input parameters that lead to the desired weld quality. RSM has been used widely to predict the weld joint properties and find the optimum of the responses of interest in many welding processes. Amit et al. [9] investigated the microhardness of austenitic stainless steel (AISI 304) with low alloy steel (AISI 1021) weld interface was found to be maximum values. It was also found that the higher hardness values were found to be at austenitic stainless steel side for all the samples. This might be perhaps due to carbon migration from the low alloy steel to austenitic stainless steel. When the hardness was checked across the weld interface, it was noticed that the hardness decreased as one proceeds towards the parent metal. It was also observed that the value of hardness increases when one proceeds from the core of the interface towards the outer periphery of the joint during hardness evaluation along the weld interface. In this study, ASS AISI 316 with LCS AISI 1018 can be joined by using CO2 laser welding without using filler materials. The influence of laser beam welding process parameters like laser power, welding speed and focal distance on the weld strength, microstructure, and chemical composition of the joint and microhardness has been investigated.

Prabakaran et. al., / Materials Today: Proceedings 5 (2018) 6607–6616

6609

2. Experimental Procedure 2.1 Experimental Design Laser welding process parameters such as laser power, welding speed, shielding gas and focal distance are affecting the quality of weld samples. In this study, three process parameters namely, laser power (P), welding speed (S) and focal distance (F) were considered for the purpose of analysis [10]. Experiments were designed by the central composite design (CCD) was composed of three columns and 20 rows. This design was chosen based on three LBW welding process parameters with three levels each. Before the experiment, preliminary trials were made to achieve the visually acceptable joints. The process parameters employed in the experiment are listed in Table 1. Statistical software, Design-Expert, was used to code the variables and design the complete set of twenty experiments based on three-factors-three-levels CCD method. Table 1 LBW Process parameters and their levels S. No

Process Parameters

1 2 3

Units

Laser Power Focal Distance Welding Speed

W Mm (m/min)

Notation P F S

-1 2000 15 1.5

Levels 0 2300 20 2

+1 2600 25 2.5

2.2 Materials In this investigation, low carbon steel AISI 1018 to austenitic stainless steel AISI316 with the dimension of 100 X 100 X 2 mm was used to get square butt joint in single pass weld. The base material plates are cut by wire cut electro discharge machining; the surface edges are fine finished and cleaned with suitable chemicals before the welding process. The chemical composition of these base metals is listed in Table 2 respectively which was found in the spectroscopy test. Table 2 Chemical composition of base metals Wt%

C

Si

Mn

P

S

Cr

Mo

Ni

Fe

ASS AISI 316

0.024

0.28

1.44

0.041

0.017

16.95

2.06

10.09

Balance

LCS AISI 1018

0.12

0.10

0.60

0.015

0.012

-

-

-

Balance

2.3 CO2 laser welding The experiments were performed with a 4kW CO2 laser (TRUMPF TLC1005) with the spot diameter of the laser beam was 0.15mm, focused by a 15-25mm focal length lens. The plates were properly positioned and clamped using the special fixture. Pure argon gas with a flow rate of 20 lit/min was provided to all the experiments to prevent oxidation of the molten metal. The experiments are carried out according to the design matrix in a random order to avoid any systematic error in the experiment. 2.4 Response Surface Methodology Response surface methodology (RSM) is utilized to determine the relationship between various process parameters and welding characteristics of weld strength. A second-order response surface model is used to evaluate the parametric effects on the tensile strength of the joint. RSM is a good way to describe the process and to find the optimum value of the considered response. It concerns a set of mathematical and statistical tools that can be used to predict the response influenced by the considered input variables in order to optimize this response [11].

Prabakaran et al., / Materials Today: Proceedings 5 (2018) 6607–6616

6610

A second-order response surface model can be written as 𝒒 𝒒 𝒚𝒊 = 𝜷𝟎 + ∑𝒋=𝟏 𝜷𝒋𝒙𝒋 + ∑𝒋=𝟏 𝜷𝒋𝒋𝒙𝒋𝟐 + ∑ ∑𝟑𝒊<𝑗 𝜷𝒊𝒋 𝒙𝒊 𝒙𝒋

(1)

Where yi is a corresponding response, i.e., Tensile strength; xj represents laser power, focal distance and welding speed β0, βj, βjj, and βij represent the constant quadratic and interaction terms, respectively. The weld strength obtained from experimental results for the different combination of parameters is given as input to the design expert software; second order mathematical was developed to predict the weld strength. Weld Strength (WS) = 218.75705 + 0.15203×(P)-2.22929×(F)+89.16274×(S)+1.39667E-4×(PF)-3.15333E-3×(PS)(2) 0.87850×(FS)-3.35717E-5×(P2)+0.10122×(F2)-17.15182×(S2) 2.5 Tensile testing The tensile test was carried out in 30 kN, computer controlled universal testing machine (UTM). Welded joints were sliced by using wire cut Electro Discharge Machining (EDM) and twenty transverse tensile specimens were prepared to required dimensions as per ASTME8M-04. The specimen is loaded as per ASTM specifications so that the tensile specimens undergo deformation. Tensile strength obtained from the specimens varied from 448.256 to 458.214 N/mm2. The experiments were carried out according to the design matrix and the experimental value and predicted values of tensile strength is given below table 3. Table 3 Experimentally measured response with successive sequences of the order Coded values Run

Actual values

P

F

S

P

1

0

1

0

2300

2

0

0

0

2600

3

1

-1

1

4

0

0

0

5

0

0

6

1

0

7

-1

1

F 20

Experimental Weld Strength (N/mm2)

S

Predicted values (N/mm2)

2

453.214

453.2479

25

1.5

458. 214

458.5518

2300

20

2

453.521

453.2479

2300

20

2

453.861

453.2479

0

2000

15

1.5

452.291

451.8867

0

2600

20

2

455.776

456.5067

1

2300

15

2

453.447

454.3139

25

2

456.446

456.1521

8

1

-1

-1

2300

9

0

0

0

2600

15

1.5

457.441

457.0651

25

2.5

454.848

455.1091

10

-1

0

0

2600

11

0

0

1

2000

25

2.5

450.752

450.9847

20

1.5

451.568

452.5481

12

0

0

0

2300

13

0

0

0

2300

20

2

453.101

453.2479

20

2.5

449.688

449.2809

2

453.212

453.2479

14

0

0

-1

2300

15

-1

1

-1

2300

20

16

1

1

1

2000

20

2

452.013

451.8553

17

1

1

-1

2000

15

2.5

448.265

448.458

18

-1

-1

1

2600

15

2.5

454.524

454.2444

19

-1

-1

-1

2000

25

1.5

454.562

454.6984

20

0

-1

0

2300

20

2

453.724

453.2479

Prabakaran et. al., / Materials Today: Proceedings 5 (2018) 6607–6616

6611

3. Result and Discussion 3.1 ANOVA Test ANOVA is applied to judge the adequacy of the fitted models. It compares the variation due to change in the combination of variable levels along with variation caused by random errors in measurement of the responses. The significance of the regression for predicting the responses can be evaluated from the comparison [12]. In the present study, test for significance of the regression models, the F-test for significance on individual model coefficients, and the Lack-of-Fit test were all performed using the Design Expert same software package. Resulting ANOVA table 5 for all the responses along with significant model terms and R2 values are shown in corresponding tables. The Model F-value of 32.03 implies the model is significant. There is only a 0.01% probability that a larger F-value might occur due to noise. Values of "Prob > F" less than 0.0500 indicate the model terms significantly. In this case, A, B, C, A2, B2, C2 are significant model terms. Values greater than 0.1000 indicate the model terms are not significant. The "Lack of Fit F-value" of 7.08 implies the Lack of Fit is significant. There is only a 2.55% chance that a "Lack of Fit F-value" this large could occur due to noise. The "Pred R-Squared" of 0.7730 is in reasonable agreement with the "Adj R-Squared" of 0.9363; i.e. the difference is less than 0.2." Adeq Precision" measures the signal to noise ratio. Table 5 ANOVA for Response Surface Quadratic model (Analysis of variance table [Partial sum of squares - Type III]

Source

Sum of Squares

Df

Mean Square

F Value

P-value Prob >F

Model

112.3673

9

12.48525

32.02776

<0.0001

P-Laser Power

54.0888

1

54.0888

138.7511

<0.0001

F-Focal Distance

8.447448

1

8.447448

21.66979

0.0009

S-Welding Speed

26.68649

1

26.68649

68.45744

<0.0001

PF

1.381122

1

1.381122

3.542919

0.0891

PS

0.184832

1

0.184832

0.47414

0.5067

FS

0.040612

1

0.040612

0.104181

0.7535

P2

2.394311

1

2.394311

6.142

0.0326

F2

10.83661

1

10.83661

27.79858

0.0003

2

S

14.97319

1

14.97319

38.40994

0.0001

Residual

3.89826

10

0.389826

Lack of Fit

3.415829

5

0.683166

7.080454

0.025483

Pure Error

0.482431

5

0.096486

Cor Total

116.2655

19 R-Squared- 0.966471

Adj R-Squared - 0.936295 Adeq Precision - 0.772959

Fig.1. shows the relationship between the actual and predicted values of square butt joint weld strength respectively. This figure also points out that the developed models are adequate and predicted results are in a proper agreement with measured data. The relationship between actual and predicted values and the normality plot are shown in Fig.2. In figure testify that the models are adequate as the predicted responses are close to the actual ones. Even the normality plots, which basically show the residuals versus respective cumulative frequency percent,

Prabakaran et al., / Materials Today: Proceedings 5 (2018) 6607–6616

6612

indicate that residuals fall almost on a straight line [13]. Now it can be inferred that errors are normally distributed and models are sufficient to describe the factor-response relationship. 3.2 Response Surface Graphs The response surfaces clearly indicate the optimal response point and the different colored surfaces show that the value of tensile strength obtained for the corresponding values of input parameters. Response surface plots and contour plots for the weld strength obtained from the regression model [14]. Fig (3.a&b) shows the combined effect of laser power and focal distance on weld strength. The pure argon shielding gas kept at 20 lit/min all the twenty experiments [15]. It is clear that the weld strength is high at high laser power and high focal distance. Weld strength gradually increases with the increase of laser power and maximum weld strength observed in the range of approximately 2300 to 2600W. Weld strength also increases with the increase of focal distance in the range of 23 to 25 mm. The experimental results revel that comparatively high laser power and high focal distance are found to be favorable higher weld strength. The combination effect of welding speed and laser power on weld strength has been shown in Fig (3.c&d). The response surface plot reflects that the welding speed has a reasonable effect on weld strength. Weld strength increasing with the increase of laser power and decrease of welding speed. At high assisted laser power and low welding speed ranges towards from 1.5 to 1.9 m/min and 2400 to 2600W respectively with a constant shielding gas flow rate 20 lit/min. Welding speed is the factor which has the greater influence on tensile strength [16]. The optimum weld strength can be achieved at a favorable value with a suitable combination of laser power and welding speed.

Fig 1 Plot for Predicted vs. Actual response of weld strength

Fig .2 Scatter diagram and normality plot for the weld strength

Fig (3.e&f) shows in terms of interaction between the welding speed and focal distance on weld strength. It is evident that weld strength tends to increase with slow welding speed and focal distance. This is due to the fact that increase of focal distance with a decrease of welding speed results in an increase of heat input and hence a good bond is produced for that reason joint strength increases. Weld strength can be achieved by combining either low power with low welding speed or high power with high welding speed. In the case of low welding speed required by the relatively low laser power, the heat transfer comes into play here more as the heat conduction losses have a greater impact at these slow speeds.

Prabakaran et. al., / Materials Today: Proceedings 5 (2018) 6607–6616

6613

(a)

Response surface plot showing the influence of laser power (P) and welding speed (S) on weld strength (WS)

(b)

Response contour plot showing the influence of laser power (P) and welding speed (S) on weld strength (WS)

(c)

Response surface plot showing the influence of laser power (P) and focal distance (F) on weld strength (WS)

(d)

Response contour plot showing the influence of laser power (P) and focal distance (F) on weld strength (WS)

(e)

Response surface plot showing the influence of focal distance (F) and welding speed (S) on weld strength (WS)

(f)

Response contour plot showing the influence of focal distance (F) and welding speed (S) on weld strength (WS)

Fig. 3 Response surface plot

6614

Prabakaran et al., / Materials Today: Proceedings 5 (2018) 6607–6616

3.3 Microstructural Analysis In order to observe the microstructure under the optical microscope, specimens were removed from the weld by wire cut EDM process and samples were prepared using conventional metallographic methods

Fig 4.a) Microstructure of ASS AISI 316, b) Microstructure of LCS AISI 1018, c) Microscopic structure of HAZ in austenitic stainless steel, d) Microscopic structure of HAZ in low carbon steel, e) Microstructure revealed cellular austenite in the form of dendrites are seen, f) EDX Result for weld interface

Prabakaran et. al., / Materials Today: Proceedings 5 (2018) 6607–6616

6615

Transversely located weld sample was embedded into a Bakelite mould in order to facilitate surface preparation of grinding and polishing in accordance with ASTM E3 standard and etched as per the ASTM E407 procedure. The weld bead macro examination result had evinced the full penetration without any noticeable defects like lack of penetration, solidification cracks and porosity in the fusion zone. The weld bead formation during the laser beam welding was regulated by the laser process parameters like power, welding speed and focal distance. The absence of welding defects, full penetration, a smooth and uniform welded surface with sound face and root bead were the criteria for selecting the functional ranges. The microstructure observed from this condition 10% oxalic acid electrolytic etching with 100x magnification. Fig (4 a&b) illustrates the base material microstructure of low carbon steel and stainless steel, which is predominantly of equiaxed austenitic (γ) grain and medium sized ferritic (F) grain respectively. Fig (4.c) shows an optical micrograph of the optimal ASS AISI316 to LCS AISI 1018 butt-joint interface. The stainless steel and fusion zone can be easily recognized in the established joint. The apparent transition zone and heat affected zone (HAZ) between the ASS and weld was difficult to observe. Because during welding sufficient heat is not generated to change the grain structure in the ASS side. A narrow HAZ exists between the fusion zone and the stainless steel substrate. The microstructure in the fusion zone is mainly columnar dendrites growing from the interface to the center of the weld bead in the direction opposite to the heat transfer direction [17]. Fig (4.d) appears to indicate solid state transformation without noteworthy dissimilarity in the ferritic and pearlite structure distribution and there are two different morphologies in the HAZ, the grains are bigger nearer to the fusion line than the ones away from the fusion line and this is because of high input during welding. It is revealed from the Fig. (4.e) that the microstructure in the fusion zone contains columnar dendrites growing from the interface to the center and equiaxed dendrites in the center. In addition, the laser-welded joint contains γ-Fe as the major phase and δ-Fe as the minor phase. The continuation of dendrite structure is predominantly due to the high cooling rate observed in the laser welded melt pool [18, 19]. Furthermore, the columnar dendrites are distributed symmetrically along the center of the weld bead owing to the symmetrical distribution of the heat input developed by the laser energy. The microstructures at the center of the fusion zone are equiaxed dendrites because of the slightly smaller thermal gradient at the center of the welding pool. Fig (4.f) the distribution of dissimilar elements across weld interface and element distribution in diverse parts are measured by optimum parameters using Energy Dispersive X-ray (EDX). The LBW is done for the two dissimilar materials and the EDAX analysis is done to examine the distribution and effect of chemical composition between LCS/ASS at the weld interface. In this EDX analysis, Fe, Cr, Ni, Cu and Zn were detected. The EDX result shows weld zone with more amounts Fe with an intermediate level of Cr and Ni, it provides sufficient mechanical properties and high corrosion resistance of the joint. 3.4 Microhardness The micro hardness profiles were performed by ASS AISI 316 to LCS AISI 1018 laser beam welded joint. The microstructures of the fusion zone must contain a variety of complex austenitic and ferritic structures were observed. Micro hardness (HV 0.5) measured in the transverse section of dissimilar (austenitic stainless steel with low carbon steel) joints. The hardness was measured at the base metals, heat affected zone on either side and weld zone. AISI 1018 indicates lesser hardness than the AISI 316. This decrease in hardness may be attributed to recrystallization process taking place at the heat affected zone towards the low carbon steel. It has also been observed that the maximum hardness was being obtained at the weld zone for all the joints. The peak hardness of laser beam welded joints increases with the increase in laser power. In addition to that the higher values of hardness at the weld interface were probably due to the oxidation process which takes place during laser beam welding. The higher hardness values were found to be at weld zone for all samples. 4. Conclusion This study investigates the optimization of LBW process parameters through RSM based on CCD. LBW process parameters like laser power, focal distance and welding speed on the maximum weld strength of dissimilar metal joints were determined in this study. A confirmation experiment was also conducted in order to validate the optimal process parameters values. The developed relationship can be effectively used to predict the weld strength of laser beam welded joints at 95% confidence level. The maximum weld strength of the dissimilar butt joint is 458.214

Prabakaran et al., / Materials Today: Proceedings 5 (2018) 6607–6616

6616

N/mm2 on laser power 2600W, focal distance 25mm and welding speed 1.5m/min. The exceptional weld appearance and mechanical properties indicate that CO2 welding of austenitic stainless steel AISI 316 to low carbon steel AISI 1018 sheet is feasible in power generation applications. A primarily homogenous microstructure and well-mixed fusion zone were produced with specific point energy of 2 mm thick dissimilar metal joint. Microhardness at the weld zone was found to be maximum level. It was also found that the higher hardness values the samples. This might be perhaps due to carbon migration from the low alloy steel to AISI316. When the hardness was chequered transversely the weld zone, it was noticed that the hardness decreased as one taking in the direction of the parent metal References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

X.B.Liu, M. Pang, Z.G. Zhang, W.J. Ning, C.Y. Zheng,G. Yu, Optics and laser engineering.45(9) (2007) pp 929-934. M. Li, Z. Li, Y. Zhao, H. Li, Y. Wang, J Huang, Advances in Materials Science and Engineering.2011(doi:10.1155/2011/767260) F. Curcio, G. Daurelio, F. MemolaCapeceMinutolo, Journal of Material Processing Technology. 175(2006) 83-89. M. Torkamany, J. Sabbaghzadeh, M. Hamedi, Mater Des. 34 (2012) 666-672. Z. Li, G. Fontana, J Mater Process Technol. 74 (1998) 174-182. E. Anawa, A.G. Olabi, Opt Laser Technol. 40 (2008) 379-388. E.M. Anawa, A.G. Olabi, J Mater Process Technol. 204 (2008) 22-33. T. A. Mai, A. C. Spowage, Journal of Material Science and Engineering.374 (2004) 224-233. Amit Handa, Vikas Chawla, Int J Adv Manuf Technol. DOI 10.1007/s00170-014-6238-0. U. Reisgen, M. Schleser, O. Mokrov, and E. Ahmed, Opt. Laser. Technol. 44 (2012) 92-101. A. Ruggiero, L. Tricarico, A. G.Olabi, K. Y. Benyounis, Optics & Laser Technology. 43 (2011) 82-90. K. Bandyopadhyay, S.K. Panda, and P. Saha, JMEPEG.25 (2016) 2462-2477. G. Padmanaban, V. Balasubramanian, Trans. Nonferrous Met.Soc. China. 21 (2011) 467-476. Arindam Ghosal, Alakesh Manna, Optics & Laser Technology. 46 (2013) 67-76. PaulrajSathiya, M. Y. Abdul Jaleel, Int J AdvManufTechnol. 54 (2011) 525-535. G. Padmanaban, V. Balasubramanian, Optics & Laser technology. 42 (2010) 1253-1260. S.M. Chowdhury, D.L. Chen, S.D. Bhole, E. Powidajko, D.C. Weckman, Y. Zhou, Metall Mater Trans. 42 (2011):1974-89. M. Nekouie Esfahani, J. Coupland, S. Marimuthu, Journal of Materials Processing Technology. 214 (2014) 2941-2948 Marya M, Edwards GR. Sci Technol Weld Join. 7 (2007) 286-293.