A Comparative Analysis of the Weld Pools Created with DC Single-, DC Double-, and PC Double-Electrode Configurations in Autogenous GTAW

Abstract

Three different Gas Tungsten Arc Welding methods—DC single electrode, DC double electrode, and PC double electrode—were analyzed using SS304 steel as the base material. Numerical models were developed to simulate the arc plasmas and calculate heat flux, current density, and wall shear stress on the surface of the workpiece. These data were used as input to simulate the weld pools across all three configurations. Experimental validation showed a good agreement with the numerical results. In the double-electrode setup, electromagnetic interaction caused the arcs to deflect, resulting in an 8% reduction in the maximum heat flux and a 4% decrease in the maximum current density. Marangoni stress had a notable effect on the weld pool shape, creating a -shaped pool with the stationary single-electrode setup, whereas the double-electrode setup produced a -shaped pool after 2 s. In the moving weld pool configurations, the sizes of the pools were maximum at the trailing electrodes. The pool was 1.7 mm deep and 5.6 mm wide in DC double- and 1.4 mm deep and 5.4 mm wide in PC double-electrode configurations. The pool depth and width were only 1.0 mm and 4.2 mm when a DC single-electrode setup was used. Comparing the three methods, the DC double-electrode setup produced the largest pool size. The findings of this research offer guidance for enhancing different arc settings and electrode arrangements to attain the intended welding quality and performance.

1. Introduction

Gas Tungsten Arc Welding (GTAW) is a welding technique that employs a non-consumable tungsten electrode to produce welds. The electrode and weld pool are protected from contaminants using an inert gas, commonly argon or helium. This method is typically employed for accurate and high-grade welding in sectors that require clean and durable welds. Pulsed Current Gas Tungsten Arc Welding (PC-GTAW) is a variation in the conventional direct current (DC) GTAW process in which the welding current alternates between a high (peak) and a low (background) current at a controlled frequency. PC-GTAW produces better-quality welds with improved strength and controlled heat input, preventing burn-through, especially in thin materials.

Double-electrode GTAW is an advanced welding process that uses two tungsten electrodes in a single welding torch, or two torches arranged one behind the other along the welding direction. This configuration is also known as a Twin-Torch or tandem GTAW. The primary objective of tandem welding is to improve productivity by increasing welding speed. Both the electrodes are independently powered and operate in the same weld pool. Usually, the leading electrode is used with a high welding current to achieve sufficient penetration and add filler metal more efficiently. The trailing electrode operates with a low current to manage the pool shape by maintaining it in a molten state for an extended duration, ensuring adequate time for reflow. This enhances the quality of the welding. Since both electrodes have their own power supply, they can be controlled independently, allowing for precise adjustment of the welding parameters for each arc. It is often used in automated or mechanized welding setups for applications requiring high productivity and quality, such as pipe and tube welding, cladding and overlay operations, and industries like aerospace, shipbuilding, and pressure vessel manufacturing.

Research has shown the significance of all these GTAW methods. Numerous numerical and experimental studies have been conducted to investigate various aspects of the conventional GTAW process, particularly the development of arc plasma and weld pools. Various researchers have developed numerical models to study arc behavior and have determined important parameters, including heat flux, current density, arc pressure, velocity, and gas shear stress, on the surface of the workpiece. Fan and Shi investigated the arc plasma temperatures in two-dimensional arc configurations [1], while Freton et al. examined them in three-dimensional [2] arc configurations. Wu and Gao [3], Du et al. [4], and Velázquez-Sánchez et al. [5] examined heat transfer and fluid flow in plasma arc columns. Sheath regions exist near the electrode and workpiece surfaces where non-equilibrium exists. They are thinly charged space layers through which current and heat transfer to the anode workpiece and cathode tungsten electrode. A significant voltage drop occurs in these regions. Sun et al. [6] and Uhrlandt et al. [7] have investigated this phenomenon by successfully modeling the voltage drop and heat loss in the sheath regions, allowing them to predict the plasma arc more accurately. Pang et al. [8] examined this same phenomenon in a micro-TIG welding process and examined heat transfer to the work part. Ilić et al. [9] compared two welding methods. In one method, they joined the root gap of the steel using Manual Metal Arc (MMA) and then proceeded with passes performed with Metal Active Gas (MAG). In the second method, they utilized Metal Inert Gas (MIG) welding for the root and the same MMA for the subsequent passes. For the MMA/MAG welding method, the fracture energy values of samples with a notch on the root pass side exceeded the values obtained from the test samples with the MIG/MAG weld. The fracture energy values were equal in both methods when measured in the heat-affected zone.

Extensive research exists on numerically predicting the development of the weld pool in DC-GTAW influenced by different welding parameters such as the electrode tip and torch angles [10,11,12], type of shielding gas [13,14,15], arc gap [16,17], welding current [18], and welding speed [19]. In summary, increasing the electrode tip angle from 30° to 90° produces deeper and concentric weld pools. Increasing the percentage volume of helium, hydrogen, or nitrogen in argon generates deeper weld pools. A short arc gap results in a narrow and deep molten pool. Conversely, a longer arc gap results in a wider and shallow molten pool. The size of the weld pool increases with the welding current and decreases with the welding speed.

A significant amount of research has also explored the PC-GTAW process. Parvez et al. [20], Tradia et al. [21], and Dutra et al. [22] numerically predicted that an increase in the pulse frequency leads to a smaller pool size at frequencies between 1 Hz and 10 Hz. However, Zhang et al. [23] showed that the weld pool become deeper as the pulse frequency increases from 100 Hz to 5000 Hz. Wang et al. [24] showed similar behavior with frequencies ranging from 10 kHz to 35 kHz. Experimental studies [25,26,27] indicate that the joint strength is higher when created with PC- in comparison to DC-GTAW. For example, in reference [28], the strength of the Inconel 178 steel was found to be 7.8% higher in PC- compared to DC-GTAW. The duty cycle in PC-GTAW is also an important parameter affecting the weld pool. Duty cycle indicates the percentage of time during which the peak current is active, followed by the background current. The duty cycle affects the microstructure, strength of the weld, and pool. Wang et al. [29] found through experiments that changing the duty from 30% to 60% raised the maximum tensile shear force to 4.4 kN in the lap joint of Invar. It is obvious and has been demonstrated by Jiang et al. [30] that a higher duty cycle enlarges the size of the weld pool.

The double-electrode (tandem) GTAW process has been studied by numerous researchers. A comparative numerical study between the single and tandem electrode GTAW was conducted by Ogino et al. [31]. The weld spot appeared circular in a single-torch configuration and elliptical in a tandem setup. The tandem configuration showed a considerable decrease in plasma shear stress when compared to the single-electrode arrangement. In another numerical comparison, Wu et al. [32] found that the two arcs attract one another, resulting in their expansion. Because of the high speed, the weld pool temperatures are much lower in the tandem configuration compared to those in the single setup, and a finer weld microstructure was obtained. Kim et al. [33] developed numerical models for the tandem arcs and weld pool, including the hot wire in the middle of the two electrodes. The hot wire was provided with reverse polarity, which consequently reduced the attraction between the two arcs. This raised the arc pressure, force, and heat flow on the surface of the workpiece. The weld pool’s depth and width were therefore increased. Ding et al. [34] also showed that the two arcs are attracted towards one another by developing a numerical model for the arc plasma with two electrodes.

Wang et al. [35] developed a unified coupled model for the arc and weld pool in double-electrode TIG welding. The arc and weld pool temperatures were calculated simultaneously. They concluded that the distribution of the current density, heat flux, and arc pressure lack rotational symmetry, which is why a Gaussian distribution cannot be used. With unequal welding currents of 80 A + 120 A, a significant difference in the behavior of the two arcs was observed; however, a slight difference was noted in the weld pool when compared to the one created with 100 A + 100 A currents. Qin et al. [36] analyzed the conventional and high-speed tandem TIG welding processes. Using a double-electrode setup with two heat sources, the duration for liquid metal was approximately 70% longer than with single TIG welding, allowing it to flow freely and fill any depressions that could be left behind the leading electrode. In a more recent work, Schilling et al. [37] experimentally investigated the effect of tandem TIG welding parameters on the arc and weld pool shape. More inclined torches produced wider beads. Large spacing between the electrodes led to instability in the arcs. Similarly, high welding speed also resulted in instability in the process.

Numerical research on tandem TIG welding of a titanium pipe was conducted by Wu et al. [38]. In the analysis, the currents of 160 A for the leading electrode and 70 A for the trailing electrode were used. The positive Marangoni stress caused the molten pool to move in the reverse direction, thereby increasing its depth. An arc efficiency of 80% was determined. The efficiency of the TIG welding process was examined by Leng et al. [39]. They discovered that under identical welding conditions, the arc voltage of each electrode in tandem GTAW is slightly lower than that of the convectional method, indicating that this approach is more efficient. Kobayashi et al. [40] employed tandem GTAW to join the PCLNG storage tank. It was determined that the tandem GTAW significantly shortens the completion time by employing a high welding speed. Jiang et al. [41] conducted experiments using three types of welding methods: DC-single, DC-tandem, and PC-tandem GTAW. The tensile strength of the joint produced by tandem PC-GTAW was slightly greater than that in tandem DC and single torch DC-GTAW and was close to that of the base material. The weld width increased as the peak current of the leading electrode increased. A greater duty cycle resulted in deeper penetration. An increased pulse frequency improved the ratio of depth to width of the molten pool surface depression. The welding speed was increased to 3 m/min, and even at this rate, a high-quality weld was produced using tandem PC-GTAW.

The current literature shows a limited number of studies focused on the combined investigation of traditional and tandem GTAW methods. While some research has addressed individual welding methods, there remains a significant gap in comprehensive evaluations that directly compare these techniques. No quantitative research has been identified that specifically investigates the behavior of arc characteristics within direct current (DC) and pulse current (PC) double-electrode configurations. The novelty of this work lies in the comparative analysis of arc plasma properties and weld pool formation across three alternative welding methods by developing numerical models. The objective was to examine the effect of each method on the weld pool. The findings of this study aim to assist researchers and professionals in the welding community by providing a clearer understanding of arc behavior and weld pool dynamics. This knowledge is critical for formulating improved welding strategies that enhance the overall quality, consistency, and performance of the welded joints.

2. Simulation Methodology

2.1. Modeling of the Arc Plasma

The simulation method is the same for single- and double-electrode configurations, hence why they are being discussed together in this section. The welding parameters used in the analyses are summarized in

Table 1. Welding parameters.

Parameter	Value
Peak current	130 A
Background current	52 A
Pulse frequency	1 Hz.
Argon flow rate	14 L/min
Tungsten electrode dia.	3.2 mm
Arc gap	2 mm
Electrode tip angle	60°
Electrode spacing	9 mm
Welding velocity	10 mm/s

. These parameters were chosen to produce a partially penetrated weld pool of sufficient size to compare the results across the three methods. As the welding electrodes are positioned perpendicular to the workpiece, the problem is planar symmetric; therefore, only half of the models were considered, as illustrated in Figure 1.

Although the arc and the workpiece domains are displayed together, they were solved individually using the software, Ansys CFX^®, Release 2019 R2. The dimensions of both computational domains in Figure 1a,b are the same. Altogether, eight simulation models were developed and analyzed, as summarized in

Table 2. Analyses conducted in this study.

S. No	Simulation	Weld Length	Analysis Type
1	Arc with single electrode	-	Steady state
2	Arc with double electrode	-	Steady state
3	Weld pool with single electrode	Spot	Transient, 2 s
4	Weld pool with double-electrode DC	Spot	Transient, 2 s
5	Weld pool with double-electrode PC	Spot	Transient, 2 s
6	Weld pool with single electrode	50 mm	Transient, 5 s
7	Weld pool with double-electrode DC	50 mm	Transient, 5 s
8	Weld pool with double-electrode PC	50 mm	Transient, 5 s

. The arc models were solved in a steady state. The arc domains were defined as argon. Argon turns electrically conductive when heated to a temperature higher than 7000 K. The arc domains were therefore initialized with a heat source to raise the temperature to that level. The heat source was subsequently removed after the argon became ionized.

The argon flow was assumed to be turbulent by applying the $\kappa -ϵ$ turbulent model. The total energy model was employed to determine heat distribution on the workpiece surface. Welding currents of 130 A and 52 A were applied to the tungsten electrodes to simulate the arc plasmas. The temperature of the tungsten electrode cannot go above 3000 K; therefore, it was set at this fixed temperature.

Ansys CFX^® models electric and magnetic fields by solving equations for the electric and magnetic vector potentials to calculate the current distribution on the workpiece surface. The Navier–Stokes equations together with the Maxwell equations simulate the arc plasma and calculate the arc velocity, pressure, temperature, wall shear stress, current, voltage, and other properties in the arc column. In this study, heat flux, current density, and wall shear were determined on the workpiece surface and used as input to simulate the weld pool. The heat flux Q was calculated using Equation (1). This approach is based on the study conducted by Parvez et al. [42].

$$Q=Q_c+J(\mathrm{\varnothing }+V_a+V_{th})$$

(1)

Q_c represents the convective heat transfer rate calculated in W/m² at the surface of the workpiece. J is the current density in A/m², which was also determined at the surface of the workpiece. $\mathrm{\varnothing }$ is the work function of the steel, and its value was taken as 4.3 V. V_a refers to the anode fall voltage at the transition region between the arc and the workpiece and is valued at 1 V. Vth is the electron thermal energy with a value of 0.58 V used in the analyses. The temperature-dependent properties of argon, including density, thermal conductivity, heat capacity, electrical conductivity, and viscosity, were taken from the work of Murphy and Tam [43].

2.2. Modeling of the Stationary Weld Pools

The stationary welds pool for single and double electrodes were simulated using a 10 mm thick, 70 mm square SS304 stainless steel plate. The thermophysical properties were taken from the handbook of properties for LMFBR Safety Analysis [44]. The whole SS304 workpiece domain was defined as liquid. A large viscosity of 1 × 10⁵ Pa. s was defined where the temperature was less than the melting temperature. Actual viscosity was specified where the temperature was above the melting temperature of 1723 K. The large viscosity treated the liquid domain as a solid. The variation in the density of the liquid steel was assumed to be negligible, and thus Boussinesq approximation was used to model the buoyancy-driven flow of the molten pool. The buoyancy reference temperature was set at 1723 K. The melt flow was assumed to be laminar. Marangoni stress plays a key role in defining the weld pool shape of SS304 steel, particularly in the presence of a high weight percentage of sulfur. Later in the discussion, we will observe the effect of Marangoni stress, which is notably important in this study. The Marangoni stress was modeled according to Equations (2) and (3).

$$M_x={\displaystyle \frac{\partial \gamma }{\partial T}}{\displaystyle \frac{\partial T}{\partial x}}$$

(2)

$$M_y={\displaystyle \frac{\partial \gamma }{\partial T}}{\displaystyle \frac{\partial T}{\partial y}}$$

(3)

M_x and M_y represent the Marangoni convections in the x-y plane on the workpiece surface. ${\displaystyle \frac{\partial \gamma }{\partial T}}$ is the surface tension gradient shown in Figure 2. The data was taken from the work of Sahoo et al. [45], corresponding to 0.03 wt.% of sulfur; however, the critical temperature was adjusted to 2700 K to match the “-shaped” experimental weld pool observed in this research. This critical temperature was chosen following several numerical experimentations that involved testing the temperature range from 2000 K to 3000 K with an increment of 100 K. The shape of the pool depth was considerably different when other critical temperatures were employed.

Additional details on the experimental weld pool shape are available in Section 4. ${\displaystyle \frac{\partial T}{\partial x}}$ and ${\displaystyle \frac{\partial T}{\partial y}}$ were evaluated during the simulation run, and M_x and M_y were determined.

The Marangoni stress, combined with the wall shear stress, was used to simulate the weld pools. The weld spots were made in 2.0 s. The analyses did not consider the depression of the liquid pool, and the surface was assumed to be flat. Since this is a comparative analysis, the error arising from this assumption will affect all findings uniformly, ensuring the consistency of the results and conclusions.

2.3. Modeling of the Moving Weld Pools

To simulate the moving weld pool with a single electrode, the parameters heat flux, current density, and wall shear stress derived from the single-electrode arc analysis were applied to the workpiece for 0.025 s, before advancing to next point at 0.25 mm in the welding direction. This process was iterated 200 times to cover the 50 mm distance. The weld started at the position w_o and finished at w_f, as shown in Figure 1a.

In the case of double-electrode DC GTAW, both the leading and trailing electrodes had the same current of 130 A. The two electrodes are 9 mm apart. The parameters obtained from the double-electrode arc analysis were used in the same way as discussed above. In double-electrode PC GTAW, the leading electrode was held at a current of 130 A throughout, while the trailing electrode fluctuated between 130 A and 52 A at 1 Hz. The 52 A current was selected as 40% of the peak current. The parameters calculated using the peak and background currents were applied. The process was repeated 20 times using the peak parameters, and then another 20 times using the background parameters to ensure 1 Hz frequency. The frequency cycle was repeated five times to cover the 50 mm weld length.

3. Boundary Conditions

The boundaries are defined by letters in Figure 1a showing three domains: electrode(s), arc, and workpiece. A General Grid Interface was defined between the interface areas of all three domains with conservative fluxes. Boundary g was defined as an interface between the arc and the workpiece during the arc simulation. The same boundary g was defined as a no slip wall during the weld pool simulation and heat flux, current density, Marangoni stress, and wall shear stress were applied to it. The remaining boundary conditions are given in

Table 3. Boundary conditions.

Boundary (Figure 1a)	Type	Description	Value
a	Wall	Tungsten electrode cross-section	130 A, 52 A
b	Symmetry	Tungsten electrode planar symmetry	-
c	Inlet	Argon flow	14 L/min
d	Symmetry	Arc planar symmetry	-
e, f	Opening	Open to the atmosphere	303 K, 1 atm.
h	Wall	Top surface of the workpiece with a heat transfer coefficient	25 W/m2 K
i	Symmetry	Workpiece planar symmetry	-
j	Wall	Sides of the workpiece with a heat transfer coefficient and magnetic potential	25 W/m2 K 0 T m
k	Wall	Bottom surface of the workpiece with a heat transfer coefficient and electric potential	25 W/m2 K 0 V

.

4. Model Validation

In this study, we only validate the stationary arc and weld pool with a single electrode. This same model was then adjusted to simulate the DC single-, DC double- and PC double-GTAW processes. The single-electrode, stationary arc was validated using prior data obtained by Parvez et al. [12]. The welding conditions were the same in both studies. The isotherms presented in Figure 3 were found to be almost similar regarding their distribution and temperature values, which shows the validity of the arc model.

The single-electrode, stationary weld pool was validated through experiments. Welding spots were created in 2.0 s using the Miller XMT 456 CC/CV multipurpose welding machine. Three tests were conducted with the same welding conditions using SS304 stainless steel plate to ensure significant experimental results. SS304 steel was chosen due to its ability to produce a clear macrograph after etching. Moreover, the surface depression of the molten pool is minimal during welding because of its high strength, enabling us to assume that it is a flat surface in our simulation. After the tests, the samples were cut at the symmetrical plane with a JINMA DK7736 wire-cut machine. The specimens were then polished and deep-etched with HNO3 (20 mL), HCl (40 mL), and Glycerol (40 mL) to obtain the weld pool shapes. The pool shapes were then photographed to compare the numerical results. In all three tests, a -shaped weld pool appeared after etching. A representative weld pool is depicted in Figure 4b.

Although we did not perform a chemical composition test on the workpiece material to determine the precise sulfur content, the -shaped weld pool indicates the presence of a critical temperature at which the surface tension gradient changes from positive to negative. In this analysis, we therefore used the surface tension gradient, which changes from positive to negative, as illustrated in Figure 2. This enabled us to achieve a more realistic weld pool shape, as shown in Figure 4a compared to the one in Figure 4b.

The width and depth of the numerical pool are 6.0 and 1.5 mm, whereas the width and depth of the experimental pool are 5.0 mm and 1.3 mm, respectively. The numerical model predicted the weld pool to be slightly bigger than the experimental one; however, the general trend remains consistent. As we are performing a comparative analysis, this discrepancy will not influence the findings of the study.

The velocity vectors in Figure 5 are observable in both inward and outward directions across the surface of the molten pool. The inward velocity vectors resulted from the positive surface tension gradient. The direction changed outward when the surface tension gradient became negative at a temperature above 2700 K. A -shaped weld pool was created where the opposing vectors converged and pushed down to deepen the pool at that location. This trend resembles the results investigated by Xu et al. [46], thereby confirming the numerical model. This same model was then modified, and various investigations were performed in this research.

5. Results and Discussion

5.1. The Arc Plasma

The arc plasma was simulated in a steady state with both single- and double-electrode configurations. The objective was to determine the heat flux, current density, and wall shear on the workpiece surface and use them to simulate the weld pool. The arc column with a single electrode is shown in Figure 3a. The arc columns with the double electrode are illustrated in Figure 6, which shows a typical arc deflection caused by electromagnetic interaction. Because the direction of the current flow was the same in both arcs, an electromagnetic force was produced that attracted the arcs towards each other. This arc attraction (or deflection) is shown in Figure 6a. The arc deflection diminishes when the welding current reduces [47]. This effect can be seen in Figure 6b. The simulation data of Figure 6a were used to model the weld pool with DC, whereas the data of Figure 6b were employed to model the weld pool with PC GTAW.

The heat flux on the workpiece surface is illustrated in Figure 7, which was calculated using Equation (1). A peak heat flux of 54.9 × 10⁶ W/m² was recorded using a single electrode, as shown in Figure 7a. Figure 7b shows the same distribution as a double-electrode setup using a welding current of 130 A on both the leading and trailing electrodes. The heat flux was determined to be 50.5 × 10⁶ W/m², which is lower. This lower heat flux was a result of the arc deflection, which slightly increased the gap between the arc peak and the workpiece surface. Because the arc deflection was minimal for 130 and 52 A current applications, this decrease was consequently minor, measured as 53.7 × 10⁶ W/m² generated by the leading electrode, as illustrated in Figure 7c. The trailing electrode produced a maximum heat flux of 24.7 × 10⁶ W/m².

Figure 8 shows the current density distribution on the workpiece surface in all three cases. The current density is used to generate the electromagnetic force within the molten pool, pushing it downwards, which subsequently enhances the depth of the weld pool [48,49]. However, in this study, we did not examine the individual impact of electromagnetic force. A symmetrical distribution of the current density was observed in the single-electrode configuration, reaching a peak of 6.9 × 10⁶ A/m², as illustrated in Figure 8a. The peak value was reduced to 6.6 × 10⁶ A/m² in the double-electrode setup (Figure 8b) because of the arc deflection. In Figure 8c, the maximum current densities were 6.8 × 10⁶ A/m² and 4.0 × 10⁶ A/m² measured with 130 A and 52 A welding currents, respectively.

The distribution of the wall shear stress on the workpiece surface is shown in Figure 9. The plasma arc layer moving over the molten pool creates wall shear stress. This contributes to an increase in the width of the weld pool investigated by Wang et al. [50] and Stadler et al. [51]. We did not examine this specifically; instead, we added it to the Marangoni stress to simulate the weld pool.

A maximum wall shear stress of 2.7 Pa was found with one electrode, whereas it was observed as 4.1 Pa with the double electrode at the same welding current of 130 A, as illustrated in Figure 9a,b. This increase resulted from the arcs being attracted to each other. Therefore, the wall shear is maximum inside the two electrodes. Figure 9c represents the wall shear on the workpiece surface at currents of 130 and 52 A. The 130 A current generated a maximum wall shear of 3.3 Pa, whereas the 52 A current resulted in an insignificant shear stress of 0.003 Pa. The figure does not depict contours at the specified scale.

5.2. The Stationary Weld Pools

Section 4 describes a stationary weld pool with a single-electrode configuration. This section discusses the stationary weld pool formed using a double electrode. Figure 10 shows the progression of the pool over time with a peak welding current of 130 A applied to each electrode. E₁ and E₂ denote the centers of the leading and trailing electrodes, respectively, which are 9 mm apart from each other. Point o represents the central distance between the two electrodes.

The pools began to grow beneath the corresponding electrodes, and the same -shaped pool was observed. Because of the Marangoni convection, the velocity vectors were directed inward on the surfaces of both molten pools. The pools started to merge after 1.0 s. As they began to reach the critical temperature, the velocity vectors changed direction, and the opposing vectors converged at o. The direction shifted downward, resulting in a deeper pool, as illustrated in the figure at time 2.0 s. The small arrows in each figure clarify this phenomenon. The maximum pool depth and width were 2.9 mm and 13.6 mm, respectively.

Figure 11 shows the development of the pool with welding currents of 130 A and 52 A applied to the leading and trailing electrodes, respectively. It is important to mention that the 52 A only maintained the glow of the arc during the background current cycle in PC-GTAW. It produced a small pool that was 0.2 mm deep and 1.9 mm wide after 2.0 s. The 130 A current created a 1.5 mm deep and 6.0 mm wide weld pool, which was the same size as that produced with the single electrode, as depicted in Figure 4a. However, the shape is slightly different because one side of the pool acquired a sudden depth because of the heat from the trailing electrode.

5.3. The Moving Weld Pools

This section addresses the development of the weld pool in DC single-, DC double-, and PC double-electrode autogenous GTAW processes. The weld pool moved 50 mm (from w_o to w_f in Figure 1a) at a speed of 10 mm/s. The remaining welding parameters were the same as those that are outlined in Table 1. Figure 12a displays a weld pool at 4.0 s that was fully developed after 1.0 s and maintained its shape and size until the process was completed. A maximum pool temperature of 2800 K was recorded. Figure 12b shows the weld pool formed using the DC double-electrode configuration. Although the currents were the same on both electrodes, the trailing pool was larger because of the heat generated by the leading electrode. In this case, the weld pool developed completely in 2.0 s, reaching a peak temperature of 3200 K. The pools did not fully merge at a welding speed of 10 mm/s. This shows that increasing the welding speed causes the two pools to move apart. Reducing the speed combines the two pools into a single pool.

Figure 12c illustrates the weld pool at 4.0 s generated using the PC double-electrode configuration. The background current was active at that time. The respective electrodes created separate weld pools. Owing to the background current on the trailing electrode, the depth and width of the pool were less than those of the leading pool. Shortly thereafter, the trailing electrode activated the peak current; the pool size increased, and it merged with the leading pool, as shown in Figure 12d. A maximum pool temperature of 3200 K was found, which is the same as that in the case of the DC double-electrode setup. The leading pool preserved its depth and width throughout the pulse cycles.

Figure 13 shows a comparison of the maximum pool depth and width for the same three electrode configurations. In the case of double-electrode setups, the sizes of the leading pools are the same, whereas the dimensions of the trailing pools differ and are larger than those of the leading pools. The measured pool depths were 1.0, 1.7, and 1.4 mm for the DC single, DC double, and PC double electrodes, respectively, indicating that the DC double configuration produced the deepest pool among the three. Correspondingly, the maximum pool widths were 2.1, 2.8, and 2.7 mm, respectively, with the DC double-electrode configuration achieving the largest width. Note that Figure 13 illustrates only half of the pool width due to the symmetry of the pool shape, which implies that the actual full widths are twice the values shown.

These results show that the DC double-electrode configuration generates the largest molten pool size, both in terms of depth and width, compared to the other configurations simulated. Employing a double electrode with a direct current source enhances energy input and heat distribution, thereby increasing the pool’s dimensions. The slightly smaller pool size observed with the PC double electrode reflects differences in the power delivery and electrode arrangement. Understanding these variations is critical for optimizing the welding or melting process parameters to achieve the desired pool characteristics for specific material-joining applications.

6. Conclusions

This study comprehensively investigated the behavior of arc plasma and weld pool dynamics in single- and double-electrode configurations in autogenous GTAW processes. The simulations of the arc plasma reveal that electromagnetic interactions in double-electrode setups cause arc deflection. As a result of the arc deflection, the maximum value of the heat flux in the DC double-electrode configuration was reduced by 8% compared to that in the DC single-electrode setup. This reduction was 4% in terms of current density. The wall shear stress was found to be the lowest with the DC single and the highest in the DC double-electrode arrangement.

In the stationary, single-electrode setup, the pool was formed in a shape with a maximum depth of 1.5 mm and width of 6.0 mm. In the stationary double-electrode configuration, the weld pools were affected by Marangoni convection and the thermal interaction between the leading and trailing pools, resulting in a shape. The maximum pool depth and width were found to be 2.9 mm and 13.6 mm, respectively. The pool size is nearly twice as large in the double-electrode configuration because of the heat generated by two arcs.

In the moving weld pool simulations, the DC double-electrode configuration produced the largest trailing pool with a depth of 1.7 mm and a width of 5.6 mm, whereas the PC double-electrode configuration generated a comparatively smaller trailing pool due to the lower background current. The maximum depth was 1.4 mm, while the width was 5.4 mm. The leading weld pools remained the same in both cases. In the case of the DC single-electrode setup, the pool was only 1.0 mm deep and 4.2 mm wide.

In conclusion, the largest weld pool was achieved with the DC double-electrode configuration, as expected by the amount of heat energy delivered to the workpiece. The findings highlight the use of the double-electrode configuration in GTAW and its effect on the weld pool size, shape, and thermal characteristics. The enhanced understanding of arc interactions and weld pool dynamics under different electrode and welding current setups provides valuable insights for optimizing welding parameters to achieve desired weld quality and performance in arc welding.