3.4.1. Geometry of Weld Bead

The change of waveforms had remarkable influence on weld geometry. It is of great significance to study the weld bead geometry of S-GMAW under different wave forms. Transverse sections are shown in Figure 19 for beads on plate deposition which contains partial penetration (the thickness of base material was 4 mm) as well as fully penetration (the thickness of base material was 2 mm), respectively.

**Figure 19.** The geometry of the weld beads with different waveforms (wire feed rate: 3 m/min, voltage: 19 V).

As shown in Figure 19, the area of base metal fusion of Cold Arc was smaller than that of the other two waveforms. As mentioned in Section 3.1, the average heating powers of LSC and Cold Arc to the base material were similar. The difference of weld pool behaviors was the main factor caused the difference of the area of bead deposit.

The oscillation intensified the convection of the weld pool, which led to the metal in the hot part to flow to the pool boundary, which intensified the melting of the base material. The impact of the conventional process and LSC on the weld pool was much larger than that of Cold Arc, so the area of base metal fusion for the Cold Arc was the smallest. Due to the oscillation amplitude of the fully penetrated pool was larger than that of the partly penetrated pool, as well as the bad heat dissipation of thin plate, the area of full penetration base metal fusion was much larger than that of the partial penetration.

The influence of electric explosion on the weld pool changed the depth and width of weld pool. Figure 20 listed the variation of weld geometry with increasing of wire feed rate. Due to the narrow range of weld parameters of the full penetration pool, this paper only listed the weld forming parameters of the partial penetration pool with different wire feed rate.

**Figure 20.** Effect of wire feed rate on geometry of fusion (partial penetration) in bead on plate deposition of S235JR: (**a**) average welding power; (**b**) area of penetration; (**c**) width; and (**d**) depth.

As shown in Figure 20, the main factor that determined weld geometry was welding heat input. The influence of weld pool behavior on weld geometry and penetration area was not obvious. Low wire feeding speed resulted in smaller peak current, lower energy input difference, lower effect of electric explosion impact on the pool, and more consistent weld formation of three waveforms. The difference of weld forming was more obvious with the increase of wire feeding speed.

#### 3.4.2. Microstructure of Weld Metals

The waveform of S-GMAW has great influence on the microstructure of weld bead. It determines the dynamic behavior of the weld pool, which further influences the solidification behavior of the weld pool. In a large part, solidification behavior determines microstructure of weld metal. This microstructure of weld metals is probably due to the complex interactions between weld thermal cycle, cooling rate, and the prior austenite grain size [22]. In this study, the weld metal composition remained almost constant under the conditions that the same base metal, filler metal, and shielding gas were used in all experiments, therefore, the change of microstructure is related to the change of arc heat input and weld pool behavior.

For the same welding condition, the weld pool thermal behaviors between the three waveform processes were completely different, which resulting in obvious and different weld microstructure, as shown in Figures 21 and 22. The proeutectoid ferrite (PF) volume fraction precipitated at the initial austenite grain boundary were related to the welding heat cycle. The lower the cooling rate of weld bead, the greater the amount of PF formation. Poor heat dissipation can aggravate this phenomenon, as the full penetration weld was prepared with 2 mm steel plate, which inhibited the heat transfer and reduced the cooling rate, it caused thicker PF compared with the partly penetrated weld pool.

**Figure 21.** The microstructures of the weld beads with different waveforms (magnification: 50 times).

**Figure 22.** The microstructures of the weld beads with different waveforms (magnification: 200 times).

The heat input of conventional process was larger than the LSC and the Cold Arc, resulting in larger PF microstructure. The heat input of LSC was close to Cold Arc, but the oscillating behavior of the weld pool aggravated the heat exchange between the metal liquid in the high temperature zone and the low temperature zone, resulting in the slow cooling rate and the precipitation of more PF in the initial austenite grain boundary.

PF usually precipitates at the initial austenite grain boundary, and the distribution of PF reflected the initial austenite grain boundary [23]. The directivity of PF structure in partly penetrated weld bead was obvious, which indicated that most of the initial austenite grain were columnar. In the fully penetrated weld bead of conventional process and LSC, the structure of PF microstructure no longer had direction, which indicated that the solidification and growth process of initial austenite grains could be changed by the oscillation of the weld pool.

In the weld microstructure, acicular ferrite (AF) can effectively increase the mechanical properties (particularly toughness) of the weld. Therefore, the volume fraction of acicular ferrite in Figure 21 was measured, image processing software was used to measure the area fraction of different phases in the Figure. The area fraction of AF and PF phases in the metallographic picture were approximately equal to their respective volume fractions, and the results were shown in Figure 23. The results show that the volume fractions of AF were significantly different, due to the change of the weld heat gradient within the different waveforms. The heat gradient of the weld pool of the fully penetrated weld bead was smaller than that of the partial penetration, which resulted in a decrease of the volume fraction of AF. The weld pool oscillated continuously in the traditional process and LSC, the weld pool was agitated and the thermal gradient declined, resulting in the decrease of AF content. The AF content in the weld microstructure of the traditional process and LSC was lower than that of Cold Arc.

**Figure 23.** Acicular ferrite and proeutectoid ferrite volume fraction in weld deposits: (**a**) partial penetration; (**b**) full penetration.

The AF microstructures of different bead-on-welds as shown in Figure 24 reveal different grain size. The average grain sizes of AF grain were calculated using intercept method (as per ASTM E112-10). In general, the grain size of weld metals can be typically correlated with the heat input or the cumulative effect of weld parameters. As mentioned in Table 3, the heat input of traditional process was obviously larger than that of LSC and Cold Arc, which caused the grain size of conventional process was larger than the others. The heat input of LSC was approximately equal with the Cold Arc, the size of acicular ferrite precipitated in the weld of LSC was larger than that of Cold Arc. In the case when the heat input was approximately equal, the change in AF size also verified the difference of thermal gradient of welding pools between different waveforms. The decrease of thermal gradient resulted in AF grain grow.

**Figure 24.** Optical micrographs of acicular ferrite (**a**) conventional, (**b**) LSC, and (**c**) cold Arc.

The weld microstructure of low carbon steel was the result of solidification and solid phase transformation of the weld pool in non-equilibrium state. The columnar-to-equiaxed transition caused by the oscillation of the weld pool had not been observed, but the oscillation homogenized the temperature distribution of the weld pool, reduced the temperature gradient of the weld pool, and resulted in the coarsening of the weld structure in the process of solid phase transformation.
