Comparisons of microstructure and mechanical properties of MAG joints welded under 2G and 3G conditions

MAG welding of 07MnMoVR steel was performed at the 2G and 3G positions, and weld formation, microstructure, residual stress, and tensile properties were compared. In this study, welds without defects were obtained at the 2G and 3G positions. The results showed that a larger distortion of the weld at the 3G position was present because of the higher heat input and that the perlage morphology was related to the introduction of the arc weaving process. In addition, the grain size of the filling pass was coarser than that of the cap pass because of the repeated heating process, and the grain sizes of the filling and cap passes increased by approximately 33% for the weld at the 3G position compared with that at the 2G position. In this case, the weld at the 3G position showed a larger residual stress and lower yield and tensile strengths, and the elongation rates and microhardness of the weld at the 3G position were lower than were those of the weld at the 2G position, regardless of the root pass, filling pass, or cap pass.


Introduction
Pumped storage power plants are of particular interest as clean energy sources.Among these, high-strength and low-alloyed (HSLA) steel is widely used because of its high strength and good toughness, and it is the preferred material for pumped-storage power plants [1][2][3].Welding is one of the main joining technologies for large-scale structures, particularly multi-pass welding for thick-walled structures in engineering projects.However, owing to the actual working conditions, welding should be performed at various positions, such as the 2G and 3G positions for penstock steel pipes [4][5][6][7].
At various welding positions, owing to the effect of gravity on the molten metal, the selection of welding parameters and corresponding weld formation are quite different [8].By comparing the welding of S700 steel at the 1G and 2G positions, Guo et al found that the equilibrium of the surface tension, gravity, and vapour pressure was easier to obtain and that welding defects, including undercut and bead sag, could be weakened [9].In narrow gap GMAW employing a 3G position, Ni et al found that humps are easily formed at high welding speeds, which increases the occurrence of a lack of fusion [10].In the CMT welding of X65 steel employing the 3G position, Chen et al investigated the relationship between the heat input and peak current time, which can be used to predict the penetration condition [11].In variable-polarity plasma arc welding employing the 2G position, Chen et al found that the adverse effects of gravity on the molten flow can be weakened by adjusting the wire position [12].
According to the current research related to different welding positions, the stability of the molten pool and weld formation have mainly been investigated, whereas comparisons of the microstructure and mechanical properties employing various welding positions have rarely been reported.Establishing a relationship between the microstructure and mechanical properties by combining the process characteristics at the 2G and 3G positions is one of the most important issues.Thus, MAG welding of 07MnMoVR steel was carried out at the 2G and 3G positions, the microstructure evolution was investigated, and the mechanical properties were discussed.This work is significant for the engineering construction of pumped-storage power plants and can deepen the understanding of welding characteristics at various welding positions.

Materials and methods
The base metal used in this study was 07MnMoVR steel with a length of 500 mm, width of 200 mm, and thickness of 30 mm.The welding wire was ER80S-G with a diameter of 1.2 mm.Prior to welding, a V-groove with each side at 30°was prepared and clamped to a gap of 5 mm.
The welding source was an Artsen Pro 500 with a protective gas of 82%Ar+18%CO 2 .The strategy of a low heat input without the weaving process was employed for welding at the 2G position, whereas the strategy of a larger heat input for welding at the 3G position and the weaving process was also introduced to overcome the adverse effect of the gravity of the molten pool on the weld formation.A lower welding speed and wire-filling rate were employed for welding at the 3G position than at the 2G position.The multipass distribution and corresponding welding parameters are shown in figure 1 and listed in table 1.After welding, metallographic and tensile specimens were prepared.The metallographic specimens were sanded, polished, and etched with 5% nital for 8 s.Then, their macromorphologies and micromorphologies were observed using an optical microscope.Electron backscattered diffraction (EBSD) tests were performed using a Sigma 300 field-emission scanning electron microscope.The micro-hardness was tested by SVD-432TS with load of 10 kg for 15 s and an interval of 0.5 mm within the center of the root pass, filling pass, and cap pass.The distribution of residual stress was tested using an iXRD MG40 x-ray diffractometer, and the results were calculated from the residual strain using Hooke's law.All-weld tensile tests were performed using a microcomputer-controlled electronic testing machine according to ASTM E8(E8M)-11.

Weld formation
As shown in figure 2, for welding at the 2G or 3G position, the welds were free of defects, based on the optimised welding parameters obtained by a vast number of pre-experiments.The plate welded at the 3G position exhibited a larger distortion because of the higher heat input than that observed at the 2G position for each pass.Moreover, the perlage morphology was obvious for the cap passes at the 3G position, owing to the introduction of the weaving process.Regardless of the distortion or perlage morphology, the special welding position is the main factor, which can be illustrated as follows: In this study, a smaller heat input and more welding passes were employed at the 2G position without introducing the weaving process, whereas a larger heat input, fewer passes, and a weaving process were  employed at the 3G position, as shown in figure 3. Generally, the gravity of the molten metal was verticaldownward, which was vertical to the welding direction at the 2G position.A downward flow was observed for the molten metal along the sidewall of the groove, although the molten metal was supported by the groove and previous welding passes.In this case, the downward flow is more severe with a larger heat input and more molten metals.Moreover, the weaving process was not suitable for the 2G position because the weaving direction was the same as or opposite to the direction of gravity, which accelerated the downward flow of the molten metal.However, when welding at the 3G position, although the gravity of the molten metal was maintained vertically downward, it was parallel with and opposite to the welding direction.The downward flow of the molten metal occurred more easily because the supporting effect of previous welding passes was absent.In this case, the molten metal flowed to the back of the molten pool, which deteriorated the welding stability, and hump defects formed easily at the 3G position [13].If a weaving process was introduced at the 3G position, which was vertical to the direction of  welding and the gravity of the molten metal, the molten pool widened while the wire filling rate remained the same, which accelerated the cooling rate of the molten metal.As a result, the backward melt flow was restricted, and hump defects were inhibited [10,14].Even if a larger heat input is employed with the formation of more molten metal, the downward flow of the molten metal and hump defects can be eliminated.Thus, with the introduction of the weaving process, a larger heat input and fewer welding passes can be employed at the 3G position.

Microstructure characteristics
Figure 4 shows the microstructural characteristics of the welds at the 2G and 3G positions.At the 2G and 3G positions, the microstructure of the fusion zone was mainly composed of proeutectoid and acicular ferrite, whereas that in the HAZ was mainly composed of granular and lath bainite.For the root pass, the microstructures remained similar for the 2G and 3G welds because of similar heat inputs.For the other passes and HAZ, the microstructure was coarser for the weld at the 3G position than that at the 2G position, owing to the larger heat input.
During the solidification process, austenitic transformation was initially performed on the weld metal.Owing to the presence of alloying elements in high-strength, low-alloy steel, the pearlitic transformation was strongly inhibited, even at a low cooling rate, whereas acicular ferrite, which was non-equiaxial with high-density dislocation by the phase transition of the shear mechanism, tended to form because of the high forming temperature.Generally, ferrite formed with a cooling rate larger than 1.5 °C s −1 , whereas acicular ferrite formed with a cooling rate larger than 3 °C s −1 .The weld metal was composed of ferrite and acicular ferrite because of the non-equilibrium heating and cooling processes.The nucleation of the acicular ferrite tended to occur in the dislocation substructure within the original austenite grains, with lower requirements for energy and concentration fluctuations than those at the grain boundaries.In this case, the acicular ferrite was fine and intertwined with different orientations.In contrast, bainite tended to form at temperatures that were lower than those of ferrite, which was present in the HAZ.
For a clear characterisation of the grain size in each region, EBSD tests were performed, and figure 5 shows the results.For the root pass, the grain sizes for the welds at 2G and 3G positions were 5.13 μm and 5.99 μm, respectively, as shown in figures 5(a) and (b).Under the welding conditions at the 3G position, owing to the effect of gravity, the molten metal tended to flow to the back of the molten metal, which concentrated the heat, and the grain size was larger because of the prolonged holding time at high temperatures.
Compared with the weld at the 2G position, the grain size of the filling pass for the weld at the 3G position increased from 4.22 μm to 5.66 μm, that of the cap pass increased from 3.43 μm to 4.54 μm, and that of the HAZ increased from 20.74 μm to 36.29 μm.Overall, the grain size of the filling and cap passes increased by approximately 33%, whereas that of the HAZ increased by 75%, owing to the larger heat input of the weld at the 3G position.It should be noted that regardless of the position, the grain size at the filling pass was coarser than that at the cap pass because of the repeated heating process.

Mechanical properties 3.3.1. Residual stress
Residual stress is a detrimental factor in pipeline failure [15][16][17].By comparing the residual stress of the joint at the 2G and 3G positions, as shown in figure 6, it was found that the residual stress was compressive and that the 3G joint had a larger stress than did the 2G joint because of the larger heat input.At a distance of 50 mm from the weld centre, the residual stress decreased to 0 for the 2G joint, whereas a value of 200 MPa was maintained for the 3G joint.

Tensile properties
Figure 7 shows the stress-strain curves of the 2G and 3G joints with all-weld tensile samples, and table 2 lists the tensile results.For the root pass, compared with the 2G joint, the yield strength (YS) and tensile strength (TS) of the 3G joint were similar to those of the 2G joint at approximately 650 and 709 MPa, respectively.The YS and TS values for the filling pass of the 3G joint were 625 and 685 MPa, respectively, whereas those of the 2G joint   increased to 730 and 773 MPa, respectively.Moreover, the elongation rates of the 2G joint were obviously higher than those of the 3G joint in different regions, indicating the better toughness of the 2G joint.

Microhardness
Figure 8 shows that the microhardness of the joints at the 2G position were obviously higher than those of the joints at the 3G position, regardless of the root pass, filling pass, or cap pass.Moreover, the average hardness of the root, filling, and cap passes was approximately 290 HV 10 for the joints at the 2G position, whereas those for the joints at the 3G position were 252, 263, and 250 HV 10 , respectively.The difference in hardness between the 2G and 3G positions can be attributed to the finer grain size owing to the lower heat input at the 2G position, which promoted a higher hardness.When welding at the 3G position, a weaving process was employed to overcome the adverse effects of gravity, and a larger heat input with more filling wires was optimised to ensure sufficient filling within the weaving range.The difference in hardness among the root, filling, and cap passes can be attributed to the reheating effect of the latter passing over the previous pass.

Conclusions
(1) The perlage morphology was obvious for the cover passes at the 3G position because of the introduction of the weaving process; however, a larger distortion was observed because of the higher heat input.
(2) Compared to the weld at the 2G position, the grain sizes of the filling and cap passes increased by approximately 33% for the weld at the 3G position, whereas those of the HAZ increased by 75%.Moreover, the grains at the filling pass were coarser than were those at the cap pass because of the repeated heating process.
(3) Overall, joint welding at the 2G position resulted in improved mechanical properties.Compared with the weld at the 2G position, the weld at the 3G position exhibited a larger residual stress, lower yield strength and tensile strength, and poorer elongation rate.Moreover, the microhardness of the joint at the 3G position was lower than that at the 2G position, regardless of the root pass, filling pass, or cap pass.

Figure 5 .
Figure 5. EBSD results, where GS denotes grain size: (a) weld metal of root pass at 2G, (b) weld metal of root pass at 3G, (c) weld metal of filling pass at 2G, (d) weld metal of filling pass at 3G, (e) weld metal of cap pass at 2G, (f) weld metal of cap pass at 3G, (g) HAZ at 2G, (h) HAZ at 3G.

Figure 6 .
Figure 6.Distribution of residual stresses on the weld surface.

Figure 7 .
Figure 7. Stress-strain curves of the root pass and filling pass.

Table 1 .
Welding parameters for 2G and 3G positions.

Table 2 .
Tensile properties of the joints.