**2. Experimental Set-Up and Materials**

NGW were fabricated using both GMAW and the cold wire gas metal arc welding (CW-GMAW) process. Figure 1 shows a schematic of the geometry of the narrow groove used in this work and the detail of cold wire positioning regarding the wire electrode and cold wire feeding angle. The grooves were welded using a Lincoln R500 welding power source linked to a Fanuc ArcMate 120i robotic arm. The size of the joints was 140 mm (length) × 115 mm (width) × 15 mm (thickness). Figure 1 shows groove gap and the root face. To determine the reproducibility, three replicates were manufactured for each welding condition. It is important to mention that the welds were performed in constant voltage mode, where no synergic controls were used during welding to adaptively control the arc dynamics.

To manufacture the welds, ER100S-G in the diameter of 1.2 mm was used as electrode, while the cold wire had a diameter of 1.0 mm. API X80 [9] was selected as base metal. The nominal compositions of the electrodes and of the base metal are given in Table 1. Moreover, no weaving or preheating was used during welding.

Cold wire feed rates are expressed here as a fraction of the electrode mass feed rate, due to the fact that two wires of different diameters were used for the electrode and cold wire, respectively. To quantify the cold wire feed rate as function of the electrode wire, the mass percentage as a fraction of the electrode was used. The mass feed rate was calculated from the wire density and cross-sectional area.

To evaluate the welding process, a NGW joint design used in heavy welding applications such as pipeline welding was considered. This configuration was chosen to demonstrate an immediate application of the process, which is critical to welding of thick structures, since it decreases the heat-induced detrimental effects of welding passes.

**Figure 1.** Schematics of the groove geometries used in this work: (**a**) schematic of the narrow groove, showing the cross-section; and (**b**) detail of CW positioning and feeding angle.


During welding, the current and voltage signals were acquired at the sampling frequency of 20 kHz for 2 s with synchronized high speed imaging at 5000 fps with shutter speed of 25 ms, an aperture of f/22, and a narrow band pass filter of 900 ± 10 nm. The videography was performed with the camera in parallel to the groove longitudinal line to record the metal transfer inside the groove. The high speed images shown in this work were selected to adequately represent the arc dynamics and metal transfer, when in a stable condition. The welding parameters used are reported in Table 2. For all welds, the shielding gas mixture used was Ar-15%CO2 at a flow rate of 17 L/min, and the contact tip to work-piece distance (CTWD) was constant and equal to 17 mm. The welding parameters were set to apply the same heat input using both processes. The quantities of average voltage (*Uaverage*), average current (*Iaverage*), and average power (*Paverage*) were calculated using the average instantaneous algorithm according to Equations (1)–(3):

$$
\mathcal{U}\mathcal{U}\_{\text{average}} = \sum\_{1}^{i} \mathcal{U}\_{i} \tag{1}
$$

$$I\_{\text{average}\text{\(\sigma\)}} = \sum\_{1}^{i} I\_{i} \tag{2}$$

$$P\_{\text{average}} = \sum\_{1}^{\bar{i}} \mathcal{U}\_{\bar{i}} \times I\_{\bar{i}} \tag{3}$$

The arc arc stability was also assessed through cyclogrammes which are voltage versus current plots, and are used to study the events occurring in the arc electric. They are useful since they show the amount of short-circuits and account for the general arc stability of the process [13]. For a more

thorough discussion on cyclogrammes and their respective zones, the reader should refer to prior work [14].


**Table 2.** Welding parameters.

Once the experiments were completed, the specimens were subjected to standard metallographic procedures and etched with 5% Nital to reveal the macrostructure. The cross-sections showing the passes sequence were taken from the start, middle, and end of the joints. The cross-sections showing the complete joint were taken from the middle of the joints. The hardness map was performed with a 200 g load and 10 s dwell time. The distance between indentations was 0.3 mm and the distance between lines was 0.3 mm.
