Study on the arc behavior and mechanical properties of energy-efficient hybrid CMT-pulsed gas metal arc narrow gap mild steel welds

The novel hybrid cold metal transfer (CMT)–pulsed gas metal arc welding (P-GMAW) process was used to perform the narrow gap welding of mild steel plates. A systematic approach was followed to select the working values of numerous process parameters using high-speed images of the welding arc in synchronization with the welding current and voltage. Furthermore, emphasis was given to understanding the influence of pulse frequency on the complex arc and metal transfer behavior in the narrow gap and its subsequent effect on the side wall fusion and mechanical properties of weld joints. The root pass was deposited using the CMT process, while the filling and closing passes were deposited using the P-GMAW process to eliminate the incomplete fusion between the layers while keeping the overall heat input lower. A decrease in the narrow gap distance and an increase in the arc oscillation amplitude and welding voltage resulted in severe arc climbing over the facing surface, an inadequate fusion between the layers, and incomplete side wall fusion. Simultaneously decreasing the number of passes and welding speed engendered the overhead flow of the molten pool and hindered the heat transfer from the arc to the already deposited layer resulting in the lack of fusion. Pulse current and duration directly affect the welding arc lengths, resulting in higher welding arc deflections to the side walls in the case of lower pulse frequencies. The controlled welding arc deflections, molten metal transfer, and inter-pulse cooling displayed a noticeable effect on the mechanical properties of the weld joint.


Introduction
Thick plates are used to build massive advanced structures in heavy fabrication industries.The joining of thick plates (≥ 5 mm) is generally performed using conventional onesided single V or U butt joints and double-sided V or U butt joints, which entail high costs of workpiece edge preparation, filler material, and weld time.In contrast, narrow gap welding (NGW) is a high-efficiency welding technique with distinctly low heat input and filler material, resulting in a narrow heat-affected zone (HAZ), low thermal stresses, and minimum distortions.The higher heat input inherent to the standard GMAW process contributes to a larger HAZ, consequently diminishing the mechanical properties of the weld [1].Moreover, the decreased arc stability during welding leads to increased spatter formation and susceptibility to welding defects.In contrast, the P-GMAW process offers a relatively lower heat input, reducing HAZ and improving mechanical properties.Notably, the spray mode of droplet detachment in P-GMAW enhances arc stability, thereby reducing weld spatter generation [1][2][3].The P-GMAW process is commonly used in the NGW to optimize the heat input to achieve quality weld joints [4].The recently developed advanced short circuit variant of the GMAW process, cold metal transfer (CMT), can further reduce the heat input and improve the arc stability over the P-GMAW for a given deposition [5][6][7][8].However, the lower heat input associated with the CMT process leads to the lack of fusion defects [9].On the contrary, adopting the hybrid CMT-P-GMAW allows optimization of the heat input, which can help achieve good quality welds with minimum weld distortions.The hybrid CMT-P-GMAW process in the present work refers to depositing a few passes in the narrow gap joint using the CMT process and the remaining through the P-GMAW process.However, if not appropriately controlled, the complex arc behavior and metal transfer phenomenon in CMT or P-GMAW result in welding defects like improper side wall fusion, porosity, inclusion, incomplete fusion, and center line cracking.Additionally, the presence of sidewalls in NGW can substantially impact arc deposition.The proximity of the sidewalls to the welding arc creates a conductive medium that can influence the arc stability and the distribution of heat and weld bead formation [10].Moreover, the interplay between the arc and sidewalls can substantially impact aspects such as fusion, penetration, and the overall shape of the bead.Furthermore, numerous process parameters associated with CMT and P-GMAW processes enhance the complexity in realizing its influence on weld quality.A systematic study on the effect of the CMT and P-GMAW process parameters on the behavior of arc and metal transfer phenomenon in the NGW and its subsequent control on the weld quality is needed to successfully adopt the Hybrid CMT-P-GMAW NGW technology.
Anirudh et al. [11] reported a comprehensive study on real-time or in situ weld defect detection, which plays a crucial role in identifying defects during welding.Various methods, including NDT, acoustic signals, optical sensors, welding current and voltage signals, and multi-sensor-based approaches, were explored.These methods were classified based on the signal types, and their prediction accuracies were discussed.Algorithms used for defect detection, covering both traditional signal processing and modern machine learning techniques, were also categorized.Bevans et al. [12] employed data obtained from an in situ acoustic sensor to identify arc instabilities that lead to defects such as porosity, variations in line width, and spatter.They employed an in situ acoustic sensor alongside a wavelet-integrated graph theory approach to monitor and identify flaws during wirebased directed energy deposition.This method involves extracting a feature known as the Fiedler number from the acoustic signature and utilizing it to detect the early stages of flaw formation.Adopting this approach, they detected various flaw types with a false alarm rate of less than 2%.
Agarwal et al. [13] employed the P-GMAW method to conduct NGW with a single seam per layer.A hypothetical dimensionless factor (ø) defined as the ratio of the product of base current amplitude, frequency, and base current duration to the pulse current amplitude was proposed to study the influence of pulse current, base current, base current duration, and frequency on the quality of the weld joint.Better mechanical characteristics of the weld joint were reported at a higher mean current and ø at a lower heat input.It was also observed that the joint quality could be improved by single seam per layer multipass narrow gap P-GMAW by altering the pulse parameters without changing the heat input [4].Anant et al. [14] used the P-GMAW method with a newly developed GMAW narrow torch nozzle to create an ultranarrow multipass weld.The impact of gas flow rates on weld joint quality was investigated.Krampit et al. [15] investigated the pulsed arc welding parameters' impact on forming a narrow gap's root layer.It was reported that the increasing pulse current time and decreasing pulse amplitude and frequency resulted in deeper penetration.Zhang et al. [16] investigated the arc properties in the narrow gap GMAW process.The decrease in the arc length was explained while moving from the center pass to the weld passes near the left and right faying surfaces.
Wang et al. [17] employed an arc rotation technique to enhance sidewall penetration during NGW.They observed that increasing the rotation speed improved the melting rate of the wire within the rotation arc narrow gap GMAW process.This rotation speed increases enhanced penetration into the groove sidewalls while reducing the bead's sectional thickness.Consequently, the overall weld shape was significantly improved, and the occurrence of finger-like penetration into the bottom plate was effectively eliminated.Silveira et al. [18] investigated metal-cored arc welding (MCAW) with a rotating electrode, examining both beadon-plate and narrow gap joint configurations.The impact of rotation frequency and rotation diameter on the weld metal's resulting microstructure and mechanical properties are discussed.Interestingly, higher penetration levels were observed at 1500 rpm rotation frequency regardless of the rotation diameter.
Furthermore, increased rotation diameter and frequency led to an extension of the heat-affected zone (HAZ).The hardness exhibited greater uniformity at a rotation diameter of 3 mm, with a minor increase noted at 5000 rpm rotation frequency.It is noteworthy that the adequate grain size remained consistent across all conditions.However, a notable shift in the frequency of high-angle grain boundaries (HAGB) was evident at 1500 rpm.Additionally, grain average misorientation (GAM) maps revealed that, at 5000 rpm, most grains were deformed with high-strain energy levels.
Silveira and colleagues suggested that a rotation frequency of 1500 rpm and a 3 mm diameter were suitable parameters for welding joints with a thickness of 10 mm.Guo et al. [19] described the flow of molten metal in the direction of the weld width during the rotating arc narrow gap horizontal welding process.It was reported that the temperature distribution towards the sidewall was more significant in the rotating arc than in the non-rotational arc, which aids in complete sidewall fusion.In another study, Guo et al. [20] explored the influence of rotating arc frequency on droplet transfer stability.They noticed that a wire rotation frequency of 5-20 Hz results in steady droplet transfer, whereas a frequency of 50 Hz results in unstable droplet transfer.
Liu et al. [21] investigated the effect of active gases on the molten electrode droplet transfer and arc behavior in NG-GMAW.They observed that adding CO 2 to the Ar shielding gas resulted in projected spray transfer and that adding O 2 resulted in slag deposition on the weld surface.Similarly, Cai et al. [22] studied the effect of shielding gas composition in NG-GMAW and found that helium had a significant role in expanding the weld profile and improving sidewall penetration.In another study, Sun et al. [23] observed that increasing the N 2 percentage of an Ar-based shielding gas improved weld penetration and weld area.It was also observed that using N 2 as the shielding gas might reduce porosity in laser welds.Similarly, Zhao et al. [24] and Wong et al. [25] found that lowering the CO 2 level in the ternary shielding gas (Ar + CO 2 + O 2 ) resulted in a wider arc and a more stable pulsed streaming spray in P-GMAW.
In various approaches to enhance the NGW quality, Zhu et al. [10] introduced flux strips to reduce the arc climbing effect in ultra-NGW, and Xu et al. [26] performed tandem NG-GMAW and explained the influence of wire distance and bent angle of contact tip to get defect-free joint.Wang et al. [27] proposed an alternative magnetic field to deflect the pulsed current, allowing a more significant proportion of the arc profile to remain attached to the sidewall to avoid incomplete side wall fusion.
In summary, the application of the hybrid CMT-P-GMAW process for the narrow gap joining of thick steel plates is not reported in the open literature.Furthermore, the studies on the individual effects of the GMAW process parameters on the arc behavior and joint quality in a narrow gap joint are not reported.In the present work, systematic studies are undertaken to realize the influence of the CMT and P-GMAW process parameters on the behavior of arc and metal transfer phenomenon in the narrow gap joint of mild steel plates and its subsequent impact on the weld bead profiles.Emphasis was given to understanding the effect of pulse frequencies in the hybrid CMT-P-GMAW process on the microstructure and mechanical properties of the narrow gap joints of mild steel plates.

Experimental setup
Mild steel plates of 200 mm × 65 mm × 12 mm are used as the base metal, and ER70S-6 wire of diameter 1.2 mm as filler wire in the present study.Table 1 presents the base metal and filler wire chemical composition.Fig. 1a shows the GMA welding setup consisting of a current sensor and high-speed camera integrated into the welding power source and data acquisition (DAQ) system.The above arrangement yields real-time current and voltage in synchronization with high-speed arc images during the welding operation.Fig. 1b shows the joint configuration of the welding plates with copper backing.The circular groove in the copper backing ensures complete fusion during the root pass.
Fig. 2 depicts the systematic representation of the procedure followed in the present work to select the working values (as presented in Table 2) of the numerous process parameters in hybrid CMT-P-GMAW NGW of mild steel material.The impact of process parameters on the weld arc stability and the joint quality is studied using the synchronously recorded instantaneous welding current, voltage, and high-speed arc images.The related details are presented in Section 3. Wire cut EDM process is used to extract metallography samples from the weld plates.Subsequently, these samples are mirror polished using the grit papers of 80 to 2000 size in sequence, followed by cloth polishing in the diamond suspension of particle size 0.25 µm.Furthermore, these polished samples are etched using a 2% Nital solution with a dwell time of 10 s.Macrographs are recorded using a stereomicroscope at 12.5× magnification.Micro images are captured at 500× magnification for every 1 × 1 mm array distance within the weld bead.The volume fractions of different phases in the weld joint are calculated using the point count method following the ASTM E562 standard [28].The weld mechanical behavior is studied from microhardness and tensile tests following ASTM E384 and ASTM E8M standards, respectively [29,30].
Heat input is determined by using recorded I-V characteristics for all the experiments.The average power and heat input of each cycle are computed as per Eqs. 1 and 2, respectively.A process efficiency (η) of 0.8 is used in the present work. (1) where I, V, and are welding current, voltage, and process efficiency, respectively.

Results and discussion
Fig. 3a shows the influence of GMAW variants-CMT and P-GMAW-on the arc behavior and weld bead formation during the NGW of mild steel plates.Fig. 3b depicts the real-time recorded current-voltage waveform of the CMT process.In this process, three primary phases are observed: boost, burn, and short-circuiting [5][6][7].During the boost phase, the electrode undergoes melting while concurrently moving forward.This forward motion persists as the electrode progresses towards the molten pool during the burn phase.Subsequently, the short-circuiting phase initiates, marked by the detachment of the molten droplet due to the retraction of the wire away from the molten pool, and the arc re-establishes.This unique sequence of events characterizes the CMT process, allowing for minimal heat input through the deposition of molten metal facilitated by the forward and backward motion of the electrode [31].The arc images during the boost phase (point 1 in Fig. 3b) and short-circuiting phase (point 2 in Fig. 3b) during the root pass welding are shown in Fig. 3c and d, respectively, and the corresponding macrograph is presented in Fig. 3e.Similarly, the arc images during the boost and short-circuiting phase during the second pass deposition are shown in Fig. 3f and g, and Fig. 3h depicts the corresponding macrograph.Fig. 3 Influence of GMAW variants on the weld joint quality Fig. 3i illustrates the real-time recorded current-voltage waveform of the P-GMAW process during the NGW of mild steel material.The arc images taken during the peak current phase, marked as "point 1" in Fig. 3i, depict the moment when the electrode undergoes melting and droplet formation due to the substantial heat input generated by the high-intensity arc pulse.Correspondingly, the base current phase, "point 2" in Fig. 3i, captures the stage when droplet detachment occurs.These events occur during the root pass welding.The respective visual representations of these phases can be observed in Fig. 3j and k.The corresponding macrograph is shown in Fig. 3l.Similarly, the arc images during the peak current and base current phases during the second pass deposition are shown in Fig. 3m and n, while Fig. 3o displays the corresponding macrograph indicating the fusion defects in specific cross sections.It is worth noting, however, that complete fusion is evident at different cross sections of the same welded sample, as illustrated in Fig. 3p.This particular P-GMAW process is noteworthy for its capability to reduce fusion defects significantly, garnering attention for its effectiveness.
A complete sidewall fusion was evident in the case of root pass welding for both CMT and P-GMAW, and this can be related to the enhanced flow of arc and molten pool to the side walls due to the circular groove in the backing plate.Incomplete side wall fusion is evident from the second pass deposited with the CMT process (Fig. 3h), while improved side wall fusion is observed with the P-GMAW weld bead (Fig. 3p).The observed incomplete side wall fusion in the CMT process can be attributed to its lower heat input and reduced exposure of the welding arc to the faying surface due to the short-circuiting period.The improved sidewall fusion with P-GMAW is due to the enhanced exposure of the welding arc to the faying surface.Guided by the insights gained from Fig. 3 and the advantage of lower heat input associated with the CMT process, it is decided to employ CMT for the root pass.Furthermore, the lack of fusion concerns in the present work is detailed in Figs. 3, 4, 5, 6, and 7 with a systematic variation of the process parameters.
Fig. 4 shows the influence of narrow gap distance on the welding arc behavior and the weld bead formation during the NGW.As per the observation from Fig. 3, the root pass is deposited using the CMT process, and the filling and closing passes are deposited using the P-GMAW process (Fig. 4a).The narrow gap distance is changed at three levels-5 mm, 6.5 mm, and 8 mm (Fig. 4a).Fig. 4 b-g show the arc images in the narrow gap maintained at a distance of 5 and 8 mm, respectively.The deflected arcs during the electrode oscillation to the left and the right faying surface are shown in Fig. 4b and d for a 5-mm narrow gap distance and Fig. 4e and g for 8 mm.The welding arc is stable in the narrow gap center (Fig. 4c and f), while the arc climbs the faying surface as it approaches the side wall (Fig. 4b, d, e, and g).The sudden jump of the arc from the narrow gap center to the side walls results in the loss of arc stability.The increase in the unstable arc region with the decrease in the narrow gap distance diminishes the side wall fusion in the weld joint, as shown in Fig. 4h and i. Fig. 4j depicts the weld joint with no defects with a larger narrow gap distance.However, the increase in the narrow gap distance enhances the volume of filler material, resulting in higher heat input and a reduction in the weld joint mechanical properties.Also, it increases the total welding time, directly affecting the production rates.This work considers the narrow gap distance of 5 mm for further studies.
Fig. 5 shows the influence of torch/arc oscillation on the welding arc behavior and the weld bead formation during the NGW.As per the observations from Figs. 3 and 4, the root pass is deposited using the CMT process, and the filling and closing passes are deposited using the P-GMAW process.Furthermore, the narrow gap distance is maintained constant at 5 mm.The torch/arc oscillation varies at three levels-0.0,1.0, and 2.0 mm (Fig. 5a).Fig. 5b shows the welding arc at 0 mm or no arc oscillation condition where a stable welding arc is positioned at the center of the narrow gap.The corresponding welding macrograph (Fig. 5c) depicts the incomplete fusion between the layers and the side wall.It is due to the non-reachability of the welding arc to the sidewall and the absence of high-intensity arc exposure to the corners.Fig. 5 d and e show the arc images near the right and left faying surfaces during the 1-mm arc oscillation condition.Fig. 5f explains the improvement in the interlayer melting and side wall fusion as the arc oscillations help move the arc from its mean position resulting in higher arc interactions with the corners.The increase in the arc oscillation amplitude further to 2 mm results in the uncontrollable climbing of the arc over the faying surfaces (Fig. 5g-i) causing incomplete side wall fusion and interlayer melting.For a 5-mm narrow gap, 1-mm arc oscillation amplitude shows a stable arc with minimum weld defects.Fig. 6 shows the impact of welding voltage on the welding arc behavior and the weld bead formation during the NGW.As per the observations from Figs. 3, 4, and 5, the root pass is deposited using the CMT process, and the filling and closing passes are deposited using the P-GMAW process.Furthermore, the narrow gap distance and arc oscillation amplitude are maintained constant at 5 mm and 1 mm, respectively.The welding voltage was varied at three levels-20 V, 25 V, and 30 V (Fig. 6a).Fig. 6b shows the welding arc at 20 V condition, where a high-intensity welding arc with a lower arc length is positioned at the center of the narrow gap.The corresponding welding macrograph (Fig. 6e) depicts the incomplete side wall fusion.It is due to the non-reachability of the welding arc to the sidewall and the absence of high-intensity arc exposure to the corners.Fig. 6d shows the arc image at 30 V welding condition, and the corresponding macrograph (Fig. 6g) depicts the incomplete fusion with the side wall.The long and expanded arc associated with higher welding voltage conditions causes the weld bead to widen at the top of the faying surface and reduces the welding arc exposure to the narrow gap corners.Fig. 6c shows the arc image for the 25 V welding condition.The corresponding macrograph shown in Fig. 6f explains the improvement in the side wall fusion as the high-intensity welding arc is equally distributed to all the regions.Fig. 7 shows the influence of the number of passes on the welding arc behavior and the weld bead formation during the NGW.As per the observations from Figs. 3, 4, 5, and 6, the root pass is deposited using the CMT process, and the filling and closing passes are deposited using the P-GMAW process.Furthermore, the narrow gap distance, arc oscillation amplitude, and welding voltage are maintained constant at 5 mm, 1 mm, and 20 V, respectively.The number of weld passes was varied at four levels-1, 2, 3, and 4 (Fig. 7a).A constant wire feed speed (WFS) with increased welding speed requires more passes to fill the joint.The set of welding speeds and the number of passes considered in the present work are presented in Table 3. Fig. 7 b and c show the arc images and molten metal in the narrow gap at the welding conditions (4.2 mm/s, 2 passes; and 9 mm/s, 4 passes), respectively.The overhead flow of the molten pool was noticed in Fig. 7b, which acts as a hindrance for the arc to strike the previously deposited layer.This results in the lack of side wall and interlayer fusion.The absence of overhead flow of the molten pool was observed with the simultaneous increase in the welding speed and number of Fig. 4 Weld arc behavior and macrographs for narrow gap distance of 5-, 6.5-, and 8-mm joint gap passes (Fig. 7c), minimizing the lack of fusion defects.Fig. 7  d-g show the weld macrographs of a single-pass, two-pass, three-pass, and four-pass weld joint, respectively, where the lack of side wall and interlayer fusion defects are reduced and can be directly correlated to the minimization of the overhead flow of the molten weld pool.
To adopt the hybrid CMT-P-GMAW process to join the thick plates, it is recommended to choose the CMT process to deposit the root pass and the filling and closing pass with the P-GMAW process.Furthermore, the welding voltage, narrow gap distance, arc oscillation amplitude, and the number of passes are recommended to be maintained constant at Fig. 5 Influence of weld arc oscillation on the stability of the arc and weld joint quality for a 5-mm narrow gap Fig. 6 Influence of welding voltage on the arc lengths and the weld joint quality ~25 V, 5 mm, 1 mm, and four passes, respectively.The weld joint properties can be improved by choosing the suitable pulse frequency corresponding to P-GMAW.Emphasis is given subsequently to understanding the effect of P-GMAW pulse frequencies in the hybrid CMT-P-GMAW process on the arc behavior, molten metal transfer from the electrode, microstructure, and mechanical properties of the narrow gap joints of mild steel plates.Three different pulse frequencies and the related welding conditions, as presented in Table 2, are used to perform the experiments.
Fig. 8 shows the synchronized high-speed arc images (1-14) corresponding to the points on instantaneous current and voltage waveforms for the higher pulse frequency of 180 Hz to demonstrate the arc behavior and molten electrode metal transfer phenomenon in the NGW of 12-mm thick steel plates.At point 1, the pulse current is 435 A, and the voltage is 29 V with a welding arc of high intensity.The arc is deflected towards the left faying surface of the wall where the welding electrode is at extreme left during torch oscillation.From point 2, the welding current and voltage dip to 235 A and 24 V in the background phase, where the deflection of the welding arc to the side wall decreases due to the relatively shorter arc associated with the lower voltages and the electrode movement towards the center.The one drop per pulse phenomenon can be noticed in points 3-5.The sequence of observations follows the melting of the electrode in point 3, the formation of the droplet pendant in point 4, and the one drop per pulse detachment in background current at point 5, followed by the completion of droplet transfer into the weld pool at point 6.Similar repeated phenomena are observed with every pulse cycle from points 7 to 14, during which the electrode shifts towards the right faying surface.The undeflected or straight welding arc can be observed at points 7-9, where the welding electrode is at the center during torch oscillation.
Observations from Fig. 8 are schematically represented in Fig. 9 for better visualization and understanding of the welding arc and molten metal transfer behavior for 180 Hz pulse frequency, while the electrode moves from the left to the right side of the wall during the torch oscillation.Fig. 10 shows the synchronized high-speed arc images (1-9) corresponding to the points on instantaneous voltage and current waveform for the pulse frequency of 55 Hz to demonstrate the arc behavior in the narrow gap joint.At point 1, the pulse current is 360 A, and the voltage is 31 V with a welding arc of high intensity.The arc is deflected towards the left faying surface of the wall where the welding electrode is at extreme left during weld arc oscillation.It is noted that 360 A pulse current and its durations of 5 ms increase the voltage, which leads to higher arc lengths.During this region, the deflected arc is in direct contact with the left side wall, and the deflection to the side wall is much higher than that of the 180 Hz pulse frequency condition.The droplet detachment in the pulse current was noticed due to longer pulse current durations at point 2, contrary to the 180 Hz condition.From point 3, the welding current and voltage dip to 235 A and 26 V, respectively, in the background phase, where the deflection of the welding arc to the side wall decreases as the torch moves towards the center.The multiple droplets per pulse phenomenon can be observed from points 4 to 6.The observation follows the melting of the electrode and multiple droplet detachment at point 4 and a trace of droplets at point 5, followed by the low-intensity welding arc during the background current at point 6.Similar repeated phenomena are observed with every pulse cycle from points 7 to 9, during which the electrode shifts towards the right faying surface.
Similarly, observations from Fig. 10 are schematically represented in Fig. 11 for 55 Hz pulse frequency.Fig. 11a shows the welding arc deflected towards the left side wall and the droplet detachment to the weld pool at 5 ms.Fig. 11b shows the low-intensity welding arc during the background current region at 19 ms, where the welding arc shifts towards the center during the arc oscillation.The welding arc during the pulse current region where multiple droplet transfer can be observed at 23 ms is shown in Fig. 11c-followed by the welding arc at the background current region at the center of the faying surface at time 38 ms as shown in Fig. 11d.Fig. 11e shows the welding arc at time 70 ms, where the welding arc is deflected towards the right-side wall and the droplet detachment to the weld pool.The average arc length for a 55 Hz pulse frequency is 4.84 (±0.3) mm, 45% higher than that of the 180 Hz pulse frequency, resulting in higher deflections towards the side wall.
Fig. 12 shows the synchronized high-speed arc images (1-14) corresponding to the points on instantaneous voltage and current waveform for 30 Hz pulse frequency.At point 1, the pulse current is 385 A, and the voltage is 31 V with a welding arc of high intensity.The arc is deflected towards the left faying surface of the wall where the welding electrode is at extreme left during weld arc oscillation.It is noted that 385 A pulse current and its durations of 16.5 ms increase the voltage, which leads to higher arc lengths.During this region, the deflected is in direct contact with the left side wall, and the deflection to the sidewall is much higher compared to 180 Hz and 55 Hz pulse frequency conditions.Droplet formation due to high welding current and longer pulse phase is noticed at point 2, and pinching of the droplet from the wire, droplet detachment, and following the streaming mode of droplet detachment were seen in points 3-5.Small droplet detachment was visible in the background current durations at points 6-8, where the welding arc is at the center of the faying surfaces during torch oscillation.Similar repeated phenomena are observed with every pulse cycle from point 9 to 14, during which the electrode shifts towards the right faying surface.
Similar to previous explanations, Fig. 13 shows the schematic representation of the observations for 30 Hz pulse frequency.The average arc length at this frequency is measured as 4.95 (±0.6) mm, 48.6% and 3% higher than that of the 180 Hz and 55 Hz pulse frequency conditions, respectively.As a result, the arc deflection towards the side wall is relatively more, as shown in Fig. 13a and f.Fig. 13 b-d show the droplet's formation and detachment followed by the streaming mode of metal transfer during the pulse cycle at 15, 17, and 19 ms, respectively.During this period, the welding arc is at the center of the faying surfaces during arc oscillation.Fig. 13e shows the schematic representation of a low-intensity welding arc during background current duration at 28 ms, where the welding electrode starts shifting towards the right faying surface during torch oscillation.
The methodology used for measuring welding arc deflection angles is illustrated in Fig. 14.Fig. 14 a-c present a schematic portrayal of these angles under three pulse frequency conditions, 180 Hz, 55 Hz, and 30 Hz, highlighting the welding arc's deflection towards the sidewall, when it is at the leftmost extreme position during arc oscillation.In this context, AB represents the width of the welding arc, C denotes the tip of the welding electrode, and D represents the midpoint between AB.The angle formed between the electrode axis and line CD is defined as the arc deflection angle (θ).Fig. 14d illustrates the influence of pulse frequencies on the welding arc deflection angle.The arc images corresponding to ten oscillations of the welding arc within the narrow gap have been utilized to compute the welding arc deflection angles.It is observed that a more pronounced arc deflection angle (22.9±5.2°) is evident at 30 Hz pulse frequency.This increased arc deflection angle at 30 Hz can be linked to the concurrent elevation in arc lengths (4.95±0.6 mm), which, in turn, can be attributed to the extended duration of pulse current (16.5 ms) under this particular frequency.In the 55 Hz pulse frequency case, the arc deflection angles (10.4±4.9°) are comparatively lower than the 30 Hz pulse frequency condition.This reduction in arc deflection is associated with the relatively shorter arc lengths (4.84±0.3mm) at 55 Hz, which are related to a decreased pulse current duration (5 ms).
Conversely, the lowest deflection angle (5.6±2.6°) is observed at the pulse frequency of 180 Hz.This can be attributed to this frequency condition's smaller arc lengths Fig. 15 depicts the influence of pulse frequencies on the average droplet radius and the number of droplets per cycle.The calculations are based on arc images corresponding to ten cycles of the welding current and voltage waveforms.The determination of the average droplet size involves an analysis of these arc images using AUTOCAD software.Remarkably, the 180 Hz pulse frequency condition showcases the largest droplet radius accompanied by a lower droplet count (0.71±0.05 mm, with only 1 droplet per cycle).Following this, the 55 Hz pulse frequency condition is observed (0.56±0.18 mm, with 4 droplets per cycle), while the 30 Hz pulse frequency condition displays the smallest droplet radius in conjunction with a higher droplet count (0.53±0.19 mm, with 11 droplets per cycle).It is worth emphasizing the notable variation in droplet radius for the 55 Hz and 30 Hz pulse frequency conditions, highlighted by the standard deviation bars.This variance is attributed to multiple droplet detachments characterized by diverse sizes.Please note that the droplet transfer phenomenon as a function of the pulse frequencies is already explained in Figs. 8, 9, 10, 11 and 12. Fig. 16 a-c show the transverse weld macrographs of the narrow gap weld joint produced using the pulse frequency of 180 Hz, 55 Hz, and 30 Hz, respectively.The complete side wall fusion is evident for all the pulse frequencies considered in the present work.The weld penetration dimensions are calculated from a fixed reference, i.e., from the top surface of the plate.The penetration of pass 2 (~ 8.5 mm) and pass 3 (~ 6.2 mm) for all the pulse frequencies is the same.However, there is a noticeable change in the final pass weld penetration, width, and reinforcement.The 30 Hz, pulse frequency condition, produced 20% higher weld penetration, 20% less reinforcement, and 14% less weld reinforcement width than the other two frequencies in the final pass.The higher penetration in 30 Hz pulse frequency may be owing to the relatively long pulse duration, resulting in more penetration in the last pass, which contrasts from passes 2 and 3, where the peak current heat is deflected to faying surfaces.
Fig. 17a indicates the ferrite (white) and perlite (dark) in a base metal microstructure, and the same is depicted in the SEM picture for distinct visualizations of the phases present, as shown in Fig. 17b.The weld metal microstructures displayed in Fig. 17c and d contain grain boundary ferrite (α Gb ), Widmanstatten ferrite (α w ), polygonal ferrite (α P ), and acicular ferrite (α A ).The martensitic microstructure with prior austenite grain boundaries is noticed in the HAZ region, as shown in Fig. 17e and f.The pits observed in Fig. 17   etching pits rather than the pores, as they are not discernible the etching procedure.Fig. 18 depicts the influence of pulse frequencies on the volume fraction of ferrite phases formed in the weld pool.It is observed that the volume fraction of the acicular ferrite phase is higher (α A = 86.8%,α Gb = 11.3%, and α W = 1.9%) in the case of 30 Hz pulse frequency condition followed by 55 Hz pulse frequency condition (α A = 80.6%, α Gb = 16.8%, and α W = 2.6%) and then 180 Hz pulse frequency condition (α A = 77.6%,α Gb = 20.8%, and α W = 1.6%).The higher   volume fraction of the acicular ferrite phase in 30 Hz is due to longer background current duration (16.5 ms).The liquid metal was undercooled in this region when the heat input was suddenly reduced.This helps in localized heterogeneous nucleation, which leads to grain refinement and easy formation of the acicular ferrite phase [32,33].The same effect is also noticed in the 55 Hz pulse frequency condition, but a relatively less volume fraction of the acicular ferrite phase is observed when compared to the 30 Hz pulse frequency due to lower background current durations (13 ms).The further reduction in the background current duration (5 ms) associated with 180 Hz pulse frequency decreases the volume fraction of the acicular ferrite phase to the minimum compared to the other welding conditions.
Hardness profiles were taken at three levels in the thickness direction, i.e., 2, 6, and 10 mm, from the top surface of the base plate.Fig. 19 a-c shows the hardness profiles for the 180 Hz, 55 Hz, and 30 Hz pulse frequency conditions in narrow gap weld joints.The profile covers the weld metal, heat-affected zone (HAZ), and base metal.It can be seen that the microhardness values are higher in HAZ for all the pulse frequencies.The weld metal average microhardness values for 180 Hz, 55 Hz, and 30 Hz pulse frequency conditions are 213 (±12) HV, 223 (±9) HV, and 221 (±10) HV, respectively.The hardness distribution shown across the weldment at a distance of 6 mm from the top surface of the base metal is lower when compared to the 2-mm and 10-mm regions.This is due to the exposure of this 6-mm region to multiple thermal cycles (passes 3 and 4).The HAZ hardness is higher than weld metal due to the martensitic structure formed in the microstructures (Fig. 17e and f).
Fig. 20 depicts the influence of pulse frequencies on weld joint tensile properties.The maximum yield strength, tensile strength, and lower percentage elongation (474 MPa, 596 MPa, 15%) were noticed in the 30 Hz pulse frequency condition.In contrast, minimum yield strength, tensile strength, and higher percentage elongation (349 MPa, 504 MPa, and 19%) were noticed in the 180 Hz pulse frequency condition.This is attributed to a higher percentage of acicular ferrite (86.8%) at 30 Hz and a comparatively lower percentage of acicular ferrite (77.6%) in the case of the 180 Hz pulse frequency condition.Similarly, these mechanical properties (373 Mpa, 534 Mpa, and 18%) for the 55 Hz pulse frequency fall between 30 and 180 Hz due to the intermediate acicular ferrite (80.6%) formation.

Conclusions
The present work reports the NGW of mild steel plates using novel hybrid CMT-P-GMAW and studies to understand the influence of associated process parameters on arc behavior, molten metal transfer phenomenon, and weld quality.The crucial observations are summarized as follows: 1.The achievement of sound weld joints with overall minimal heat input was realized through the hybrid CMT-P-GMAW process.This approach uses CMT for the root pass while P-GMAW for the filling and closing passes.
2. The welding arc's stability inside the narrow gap is primarily determined by process parameters such as narrow gap distance, welding arc oscillation amplitude, welding voltage, and the number of passes.Reducing narrow gap distance, increasing arc oscillation amplitude, and increasing welding voltage result in severe arc climbing over the faying surface, insufficient fusing between the layers, and incomplete side wall fusion.The overhead flow of the molten pool formed while simultaneously decreasing the number of layers and welding speed restricts heat transfer from the arc to the previously deposited layer, resulting in a lack of fusion.
3. For a constant wire feed rate and heat input, a higher pulse current and a shorter pulse time result in one drop per pulse.Increasing the pulse current duration (5 ms) results in the formation of multiple droplets as the pinching effect continues.Further enhancing the pulse current duration results in the streaming mode of metal transfer.Pulsing frequencies of 180 Hz, 55 Hz, and 30 Hz result in one drop per pulse, multiple droplets per pulse, and streaming mode, respectively.
5. Additionally, in terms of droplet sizes and number of droplets per cycle, the largest droplet radius is evident at the pulse frequency of 180 Hz, paired with a lower droplet count (0.71±0.05 mm, involving only 1 droplet per cycle).Subsequently, at the pulse frequency of 55 Hz, a droplet radius of 0.56±0.18mm is observed, accompanied by 4 droplets per cycle.In contrast, the pulse frequency of 30 Hz displays the smallest droplet radius, coupled with a higher droplet count (0.53±0.19 mm, involving 11 droplets per cycle).
6.The imposition of several thermal cycles in filling pass weld region reduces hardness compared to the root and cap passes.The higher hardness in the HAZ can be associated with the martensitic weld structure.A higher percentage of acicular ferrite is formed at a 30 Hz pulse frequency followed by 55 Hz resulting in higher weld joint tensile properties.Superior joint tensile strength is observed at 30 Hz pulse frequency followed by 55 Hz pulse frequency, while the joint tensile strength was relatively low at 180 Hz pulse frequency.
7. For a mild steel material of 12-m thickness, a sound weld joint with optimal heat input is accomplished by employing CMT for the root pass and P-GMAW for the filling and closing passes with a narrow gap distance of 5 mm, weld arc oscillation amplitude of 1 mm, welding voltage of 25 V, four number of weld passes (including the root pass), and 30 Hz pulse frequency.These parameters can be adapted to the narrow gap joining of mild steel material of any plate thickness by varying the number of passes.

Fig. 1 aFig. 2
Fig. 1 a Experimental GMAW setup and b schematic representation of weld joint configuration Fig. 9 a and b show the deflection of the welding arc towards the left side wall at 8 ms, followed by the welding arc moving to the center during the arc oscillation at 30 ms.Fig. 9 c and d show the welding arc and droplet (one drop) detachment at 31 ms and the welding arc deflected towards the

Fig. 7 3 Fig. 8 Fig. 9
Fig. 7 Influence of the number of passes on the flow of molten pool inside the narrow gap

Fig. 10
Fig. 10 Arc images depicting the welding arc and molten metal transfer phenomenon corresponding to real-time welding current and voltage for 55 Hz pulse frequency

Fig. 11
Fig. 11 The diagram for the 55 Hz pulse frequency condition

Fig. 12
Fig. 12 Arc images depicting the welding arc and molten metal transfer phenomenon corresponding to real-time welding current and voltage for 30 Hz pulse frequency

Fig. 13
Fig.13 The schematic diagram for the 30 Hz pulse frequency condition are the

Fig. 14 Fig. 15
Fig. 14 The schematic diagram of welding arc deflection angles for a 180 Hz, b 55 Hz, c 30 Hz welding condition, and d arc deflection angles plots

Fig. 16
Fig. 16 Weld macrograph of NGW joint for pulse cycle frequencies a 180 Hz, b 55 Hz, and c 30 Hz

Fig. 17
Fig. 17Optical micrographs: a base metal microstructure, b base metal SEM image, c weld metal microstructure, d weld metal SEM image, e HAZ region, and f HAZ SEM image Fig. 17Optical micrographs: a base metal microstructure, b base metal SEM image, c weld metal microstructure, d weld metal SEM image, e HAZ region, and f HAZ SEM image

Fig. 18
Fig. 18 Influence of pulse frequencies on the volume fraction of ferrite phases in the weld pool

Table 1
Base metal and filler wire composition (wt %)