1 Introduction

The first step of friction stir welding (FSW) was invented in 1991 at the Welding Institute in the UK [1]. The principal function is to create friction with stationary parts using a rotating and non-consumable tool that spins, moving along the joint between two components to produce high-quality butt or lap welds. This newly proposed technique has exponential applications in aerospace and rail car manufacturing, as well as with hard metals. After being primarily used to weld aluminum and a few other alloys [3, 4], friction welding has been employed to connect various types of related, similar, and dissimilar materials, including metals and non-metals [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20]. Regarding experimental results [21,22,23,24], FSW can be used on materials with different geometric forms [25,26,27,28,29,30], especially those with thicknesses below 1 mm, as well as on workpieces up to 75-mm thick. The low-density alloy combines weight reduction with lower assembly and maintenance costs, and when combined with advanced welding practices and aerostructure redesign, it offers up to 25% weight reduction [31]. Various types of combined loading, such as tension, compression, flexion, and fatigue, are essential parameters that should be considered in the design of structures, especially when they are welded using FSW. Several factors influence the quality and properties of the weld using the new FSW process, depending on (i) the machine and the tool used, (ii) the specimen, and (iii) the mode of fixation and connection. Most of these factors are related to pressure and friction time, pressure and forging time, rotational speed, oscillation frequency, oscillation amplitude, friction pressure, and their correlation with microstructure characterization, microhardness fluctuation, interfacial phase formation, optimal welding parameters, and mechanical properties [32,33,34,35]. Useful studies have characterized the mechanical properties of FSW aluminum alloys under static tensile loading [36, 37] and fatigue conditions [38,39,40], both before and after welding. Generally, 90% of failures in aluminum structures are initiated by fatigue phenomena. The FSW method improved fatigue performance by approximately 2.4 times compared to the single riveting method, ensuring mechanical properties and in-service fatigue life [41, 42]. The authors in [43] investigated the mechanical properties and residual stresses of FSW aluminum 5083 for tensile test specimens. The results confirm the sensitivity of the joint surface to thermal input rather than the mechanical deformation caused by the FSW tool. In [44, 45], it showed that static tensile shear strength increased with decreasing tool rotational speed and increasing tool holding time, while cross tension strength decreased with increasing both parameters. The results of microstructure confirmed the presence of two fracture modes: nugget shear fracture mode and mixed mode under tensile shear loading. Experimental studies [44,45,46,47,48,49,50,51,52,53,54,55,56] on friction stir spot welding for several types of aluminum have shown that rotational speed has a significant effect on the mechanical effectiveness of the joint weld. Akinlabi et al. [48] investigated tensile strength using different welding parameters such as tool size and rotational and traverse speed on Al/Cu specimens using SWP. The joint efficiency was 86% with a rotational speed of 950 tr/min and a feed rate of 50 mm/min using an 18-mm shoulder diameter tool. Nandan et al. [49] studied technical properties and residual constraints based on welding speed, showing that using a constant material property affects the model’s accuracy within the temperature domain. Riahi et al. [50] analyzed transient temperature and thermal constraints, where heat is generated due to friction between the tool and the workpiece, and the thermal behavior of the welded piece was considered an input heat for an elastoplastic model predicting residual stress. Additionally, heat distribution across the thickness varied and appeared asymmetrical. Veljić et al. [51] investigated heat generation during the plunge phase in FSW, where the influence of heat generated by friction surpassed plastic deformation. The sliding rate of the tool and the temperature relative to the workpiece are linked to heat generated by friction, and the material’s speed is linked to heat generated by plastic deformation. Al-Badour et al. [52] conducted investigations using FEM for FSW, taking into account adiabatic heating effect and estimating the maximum temperature equal to the material’s solidus temperature (583 °C) defined in the Johnson–Cook material model. Bussetta et al. [53] modeled the thermomechanical behavior of FSW, with both fluid and different solid formulas yielding the same results. Aziz et al. [54] studied the effect of rotational speed on heat generation, proving it to be more significant than the traverse speed. Sun et al. [55] predicted temperature peaks and residual stresses, demonstrating that the Fixed Shoulder FSW (SSFSW) process produced a narrower and more uniform weld nugget and heat-affected zone compared to conventional FSW. Garg and Bhattacharya [56] studied the influence of different tool pin shapes, tool rotation speed, and dwell time on shear strength around FSW, with the tool rotation speed and pin shape significantly influencing shear strength. Jain et al. [57] (studied temperature and stress distribution, predicting the temperature distribution on the workpiece and tool, with the maximum temperature observed higher on the upper surface of the piece. Maisonnette et al. [58] analyzed transient temperature and thermal stresses, often proposing assumptions and simplifications for FSW simulation, resulting in good agreement. The temperature continuously varied, initially increasing during heating and then decreasing during cooling without a pause at the highest temperature point (Palanivel et al. [59]).While valuable studies focusing on fatigue and microstructural issues in FSW aluminum structures have been undertaken, there is still a lack of studies on the characterization of material properties for FSW high-strength aluminum alloys, such as aluminum 2017AA, used for shipbuilding. Although several studies have addressed microstructural and fatigue aspects in FSW aluminum structures, there remains a dearth of research pertaining to the characterization of material properties specifically for high-strength aluminum alloys used in shipbuilding, such as aluminum 7000 and 7070.

The sampling direction in friction stir welding of aluminum alloys holds significant importance, alongside other crucial factors, for maximizing its application. Specimens can be sampled either perpendicular or longitudinal to the weld direction. Cutting the specimens longitudinally (parallel to the direction of the joint), FSW of aluminum alloys provides valuable information about the weld characteristics, aids in understanding the welding process, facilitates mechanical property evaluation, and contributes to process optimization in aluminum alloys. However, it is important to note that the specific novelty and significance of this approach may vary depending on the context and the research study or development being referred to.

The primary aim of this study is to offer a comprehensive analysis of the mechanical properties exhibited by friction stir welded aluminum alloys. Four crucial analyses have been conducted to assess the strength and durability of the welding joint. The findings and conclusions derived from this study are meticulously documented, along with a comprehensive description of the test database.

1.1 Method of analysis and experimental tests

In this section, the various experimental methods are described. Initially, the conducted mechanical tests and the procedures employed to determine mechanical behaviors under conditions such as tensile, flexural, and fatigue loading are presented.

1.1.1 Chemical composition and microstructural observations

In this study, a base material investigated is an alloy from the 2000 series (aluminum-copper). The alloy used for conducting welds exhibits a structural hardening known as “Age-hardening” with a designation of AU4G-1). The alloy was examined in a metallurgical condition known as T6, which corresponds to the as-received state of the sheet. Chemical analyses were conducted using X-ray fluorescence spectrometry. The chemical composition of the 2017AA alloy is provided in Table 1.

Table 1 Chemical composition of the base metal (2017AA) determined by X-ray

Plates measuring 100 mm in width and 250 mm in length were cut from a 6 mm and 10-mm thick 2017AA metal sheet using an industrial saw. All welding tools were manufactured from H11 steel (tool steel Z38CDV5.1), which has a higher melting temperature and hardness compared to aluminum. Table 2 presents the chemical composition of this steel.

Table 2 Chemical composition of tool steel Z38CDV5.1

Observations were conducted using an optical microscope of the brand “Leica DM 2500 M,” connected to a microcomputer equipped with “Leica Application Suite LAS” image acquisition software, which facilitates the transfer and processing of these images. To characterize the microstructure of the welded joints and identify different zones, a Keller reagent etching was performed. Prior to etching, the sample must be mechanically polished using 240-grade paper to achieve a mirror finish, followed by polishing with 3μm diamond paste. The metallographic etching involves immersing the sample in the Keller reagent (2.5 ml nitric acid HNO3, 1 ml hydrofluoric acid HF, 1.5 ml hydrochloric acid HCl, and 95 ml distilled water) for 10 s, followed by rinsing with distilled water. Micrographic examinations of the sample surfaces are conducted using a scanning electron microscope of the QUANTA 600FEI type, equipped with an infrared camera.

1.1.2 Mechanical characterization methods

Microhardness measurements HV0.5 were performed using an “HWDM-1” instrument connected to a computer equipped with “C.A.M.S” measurement and image acquisition software. The determination of microhardness values will be conducted based on these images. Measurements were carried out following the NF EN ISO 6507–1 standard with a 500-g load applied for 15 s. The spacing between two measurement points is 1 mm. Profiles taken at 1/4 thickness, midpoint, and 3/4 thickness positions in the 2017AA alloy allowed us to verify that the hardness values were of similar magnitude across the sheet thickness. As previously discussed, a Friction Stir Welding (FSW) joint consists of distinct microstructural zones, each with different characteristics. The mechanical properties of each zone are directly linked to their microstructure. Extracting micro-specimens to determine the mechanical properties of each joint zone is relatively challenging. Therefore, we analyze the overall behavior of the weld beads. Subsequently, we will analyze fracture surfaces to correlate this information with microstructural parameters. A comparative study of obtained weld joints was conducted, considering various parameters with the base material. This is essential for establishing a connection between mechanical performance results and process operational parameters. Tensile testing was conducted following the ASTM E08 standard to assess the mechanical properties of both the base material and FSW weld joints, particularly the static behavior law (stress–strain curve). Tensile tests were performed on specimens using a Zwick/Roell ZMART.PRO type universal testing machine with a maximum capacity of 200 kN. The tests were conducted at a crosshead speed of 4 mm/min. Test control and data acquisition were facilitated by the Test Xpert II software. Bending tests were carried out in accordance with the NF A 89–204 standard on carefully machined welded and non-welded specimens. Fatigue tests were performed under plane bending with a load ratio R =  − 1, at a frequency of 25 Hz. The objectives of these tests were, firstly, to compare the cycles to failure of different tested welds and, secondly, to study weld damage using non-linear acoustic techniques.

2 Results of mechanical tests on welds

2.1 Microhardness mapping

2.1.1 Microhardness mapping on 6 mm and 10 mm samples of thickness

To examine hardness variation across a cross-section of an FSW (friction stir welding) bead, a microhardness mapping was conducted. Measurements were taken at different depths from the bead surface (1.5 mm, 3 mm, and 4.5 mm). Figure 1 illustrates the microhardness curves for three FSW weld beads. The welds were created at a welding speed of 0.6 mm/s and three rotational speeds: 950 rpm (Fig. 1a), 1050 rpm (Fig. 1b), and 1250 rpm (Fig. 1c). The tests were repeated 5 to 7 times, depending on the availability of the test pieces with error values in a range of 8–12%. Figure 1 demonstrates the classic “W” shape where hardness is minimal in the HAZ (heat affected zone), the area that experienced over-aging. It also indicates that the minimum hardness value is nearly the same on both sides of the joint (AS and RS). The three measurement lines correspond to those presented in Fig. 1; the numbering of the lines starts from the closest to the upper surface of the joint (line 1/4 thickness). Hardness in the weld nugget is lower than the base metal but higher than that of the HAZ. This characteristic is typical for FSW joints in structurally hardened aluminum alloys. The microhardness map reveals that hardness does not vary across the thickness; the values measured on the three lines are very close. Welding causes softening and thus a decrease in hardness across the weld. This drop in hardness occurs within the HAZ, where values shift from 140 to 85 Hv over just a few millimeters. The thermo-mechanically affected zone (TMAZ) shows a minimum hardness, while within the weld nugget, it rises to an average value of 130 Hv. The hardness curve allows deducing the extent of the different zones; the dimensions of the various zones in the welded joint are summarized in Table 3. Welding 10-mm-thick plates presents a similar hardness profile to the 6-mm joint (Fig. 2). It is worth noting that hardness also varies across the thickness of the welded joints, with maximum values at the 1/4 thickness line of the sample. The values measured at half and 3/4 thickness (noted by T) are the lowest among the measurement lines. It can be observed that microhardness decreases slightly less in the depth of the welds at 1250 rpm compared to the welds at 1450 rpm. Figure 2b displays a decrease of about 25 HV between the top and bottom of the weld for the joints at 1450 rpm. The dimensions of the different zones in the welded joint are summarized in Table 3.

Fig. 1
figure 1

Illustration of the microhardness curves for three FSW weld beads: a 950 rpm, b 1050 rpm, and c 1250 rpm

Table 3 Summary of the approximate size of the zones in a joint welded by FSW on alloy 2017AA of 6-mm thick
Fig. 2
figure 2

Mapping of microhardness HV500g of the seal 10-mm thickness 2017AA alloy welded at rpm: a 1250 rpm and b 1450 rpm

2.1.2 Effect of rotation speed

To investigate the impact of rotation speed, measurements are taken along the midline of the sample (1/2 of the thickness), and microhardness curves are plotted for various speeds. The comparison of the microhardness profiles of the three weld beads, produced at rotation speeds of 950, 1050, and 1250 RPM, reveals that the rotational speed has no influence on the hardness in the 6-mm-thick samples (Fig. 3.a).

Fig. 3
figure 3

HV500g microhardness mapping of welded joint: a 6-mm thickness alloy 2017AA at different tool rotation speeds and b 10 mm

Along the weld plates produced by FSW, starting from the base material's microhardness value of 140 HV, the microhardness profile drops rapidly within the heat affected zone (HAZ) until reaching a minimum value of 85 HV within the HAZ close to the near-side HAZ at − 15 mm from the center of the joint (RS side) and at 12 mm from the center of the joint (AS side). The microhardness then increases within the thermo-mechanically affected zone (TMAZ), reaching a stable value of 130 HV in the nugget, which is close to that of the base material. For the 10-mm-thick samples, it can be observed that the microhardness trend obtained at a rotation speed of 1450 rpm is slightly higher than that obtained at a speed of 1250 rpm (Fig. 3b). In this case, the rotation speed indeed influences the microhardness of the 10-mm-thick joints.

2.1.3 Tensile tests

Several test specimens containing welds produced at different rotation speeds were subjected to tensile tests. These mechanical tests aimed to characterize the welds and define a range of speeds that would yield maximum tensile strength for plates of 6 and 10-mm thickness. The tests were conducted at room temperature. Table 4 presents the mechanical properties obtained from three trials.

Table 4 Mechanical properties obtained by the tensile test on weld joints

From the tensile tests conducted on the specimens outlined in Table 4, several observations can be drawn. According to these results, the values obtained after welding are consistently lower than those of the base material, whether in terms of yield strength, ultimate strength, or elongation at rupture. The welded joints at 950 and 1250 RPM exhibit the highest tensile strength and good ductility, while the joint welded at 1050 RPM maintains good tensile strength but undergoes a significant loss of ductility. The reduction in yield strength and tensile strength is primarily attributed to the decrease in joint hardness. As an example, Fig. 4 presents the stress–strain curves for the base metal and welded specimens of the 2017AA alloy, highlighting the three distinct phases of a stress–strain curve.

Fig. 4
figure 4

Stress–strain curves for 6-mm-thick specimens

Globally, it is generally observed that the rotational speed has little influence on the yield strength and tensile strength. As evident from Fig. 4, it is clear that better mechanical quality is achieved with a rotation speed of 1250 rpm. Additionally, it can be observed that the plastic zone is relatively narrow for the welded specimens, indicating a predominantly brittle fracture behavior. The fracture of the 6-mm-thick samples occurred at the center of the joint (Fig. 5a) for welds at 950 and 1050 rpm. For joints welded at 1250 rpm, the fracture is situated within the heat-affected zone on the recoil side (RS) (Fig. 5b). Notably, fractures are observed at the edge of the bead within the heat-affected zone, which exhibits lower hardness due to material softening. Regarding the welds on 10-mm plates, a significant decrease in mechanical properties is noticeable, encompassing both yield strength and maximum tensile strength. The loss of ductility in these joints is also substantial. These reductions can be attributed to the presence of defects within the joints, leading to diminished joint quality. Figure 6 provides an illustrative example of the stress–strain curves for the base metal and the welded specimens.

Fig. 5
figure 5

A Fracture within the joint for both 6-mm-thick samples and b fracture outside the welds at 1250 rpm

Fig. 6
figure 6

Stress–strain curves for 10-mm-thick specimens

Regarding the 10-mm-thick plates, the fracture occurs within the weld joint, whether it is for welds produced at 1250 rpm or at 1450 rpm. Despite the fact that FSW-welded assemblies still exhibit acceptable mechanical properties, it is worthwhile to determine the efficiency coefficient of the weld joints (Fig. 7).

Fig. 7
figure 7

Fracture in the weld joint for 10-mm thickness specimens

2.1.4 Weld joint efficiency coefficient for FSW

The weld joint efficiency coefficient is defined as the ratio of the tensile strength of the welded assembly to the tensile strength of the base metal. The efficiency coefficient of the joints is obviously dependent on both the welding parameters and the material. The calculation results are presented in Table 5. The following are the obtained weld joint efficiency coefficients:

Table 5 Weld joint efficiency coefficients for different welds

Joint Efficiency Coefficient = (R m Tensile Strength of Welded Joint) / (R m Tensile Strength of Base Metal).

First and foremost, the weld with the highest ultimate tensile strength (measuring 351.03 MPa) was performed at a rotation speed of 950 rpm. Its efficiency is evaluated at approximately 70.2% of the base material’s tensile strength, which is established at 500 MPa. When comparing this weld to those performed at rotation speeds of 1250 and 1050 rpm, its efficiency is relatively low. Similarly, considering the welds performed on 10-mm plates, the one executed at a rotation speed of 1450 rpm exhibits high efficiency. In conclusion, the tensile tests reveal that the best welds, with a maximum rupture strength (Rm) exceeding 315 MPa (Rm > 315), were achieved on 6-mm-thick plates regardless of the rotation speed used. In contrast, the welds performed on 10-mm plates exhibit the most degraded properties.

2.1.5 Fracture facies analysis

Microscopic observations were conducted on the fracture facies of each ruptured specimen. For all 6-mm-thick specimens welded at 1250 rpm, the fracture consistently occurs on the side opposite the welding direction at the boundary between the thermo-mechanically affected zone and the thermally affected zone of the weld joint. The obtained fracture facies are oriented at a 45° angle relative to the specimen surfaces. Visualization of the fracture facies was carried out using a scanning electron microscope (SEM). Figure 8 presents an SEM image taken perpendicular to the axis of the tensile specimen.

Fig. 8
figure 8

SEM observation (various magnifications) of a tensile fracture surface of the FSW-welded joint in 2017AA alloy at a rotation speed of 1250 rpm

In Fig. 8, the ductile nature of the fracture is evident. Additionally, the fracture facies reveal the presence of multiple dimples. These dimples elongate in the direction of tension, aggregate, and coalesce to expedite the fracture process. Figure 9 illustrates the observation using a scanning electron microscope of the fracture facies in a weld joint produced by FSW at a speed of 950 rpm. Cavities are noticeable in areas rich in precipitates, indicating a ductile mode of fracture in this case.

Fig. 9
figure 9

Fracture morphology of a welded specimen by FSW

However, for the welded joints at 1050 rpm, the analyses make it possible to distinguish mainly two areas: a ductile rupture zone where SEM observations reveal the presence of cupules (Zone 1, Fig. 10) and another area of sudden rupture. The presence of cavities is noted in precipitate-rich areas across the ductile rupture zone. In the sudden rupture, a smooth area (Zone 2) is observed in comparison to the ductile rupture zone (Zone 1).

Fig. 10
figure 10

Fracture surface of a weld specimen produced by FSW and tested under tension

2.2 Three-point bending test

All conducted tests are stopped as soon as the punch is no longer applied along the axis of the FSW weld. The results obtained during each test (maximum force, corresponding displacement at maximum force, and the angle after bending the Ф specimen) are recorded in Table 6 for both the 6 mm and 10-mm thicknesses. For the 6-mm thicknesses, an example of the samples after the three-point bending tests is shown in Fig. 10. After subjecting the samples to the three-point bending test, a visual inspection was carried out, revealing no cracks or tears on the 6-mm weld bead for the base metal and the specimens welded at 950 rpm. These results demonstrate excellent ductility of the specimens (Figs. 10a and 11). However, fractures in the weld joints are observable on the specimens welded at 1050 and 1250 rpm after bending. These results indicate that specimens with superior tensile mechanical characteristics, such as those welded at 1250 rpm, do not necessarily yield the same results in three-point bending (Fig. 10b,c). Samples after the three-point bending tests are presented in Fig. 11 for the 10-mm thickness. After subjecting the samples to the three-point bending test, a visual inspection was performed, revealing no cracks or tears on the 6-mm weld bead for the base metal and the specimens welded at 950 RPM. These results once again confirm the high ductility of the specimens (Fig. 11a). However, fractures in the weld joints are observable on the specimens welded at 1050 and 1250 rpm after bending. These results indicate that specimens with superior tensile mechanical characteristics, such as those welded at 1250 rpm, do not necessarily yield the same results in three-point bending (Fig. 11b,c).

Table 6 Mechanical properties obtained through the three-point bending test on weld joints
Fig. 11
figure 11

Three-point bending tests of the base metal and 6-mm welded samples at rotation speeds of a 950 rpm, b 1250 rpm, and c 1050 rpm

Regarding the 10-mm thickness, the absence of the plastic range was observed for the welded specimens. This is attributed to the brittleness of the joints and the presence of internal defects such as lack of penetration. For all the specimens subjected to three-point bending tests, fracture occurred at the midpoint of the joint. Fig. 12

Fig. 12
figure 12

Three-point bending tests of 10-mm samples welded at rotation speeds of 1250 rpm and 1450 rpm

2.2.1 Fatigue tests in plane bending

In order to demonstrate the fatigue resistance difference between a friction stir welded (FSW) joint and its associated base material, a comparison was conducted on several welds made from aerospace structural hardening alloys, specifically using the 2017AA alloy. Specimens with 6 mm and 10-mm FSW welds were tested in the as-welded condition using plane bending fatigue tests. Test conditions are provided in Section II.2. The results of the plane bending fatigue tests are presented in Table 7.

Table 7 Mechanical properties after fatigue tests in plane bending of weld joints

Figures 13a and b provide a comparison of the number of cycles to failure for the base metal and different FSW (friction stir welding) welds for the two thicknesses.

Fig. 13
figure 13

Comparison of the number of cycles to failure at 260 MPa for different friction stir welding (FSW) joints with a 6 mm and b 10-mm thickness

Based on the fatigue results, it was observed that the 6-mm-thick joint welded at a rotation speed of 1250 rpm has the best service life (1.47 × 105 cycles), with a decrease of approximately 21% compared to the base metal’s lifespan (1.87 × 105). The other two specimens welded at 950 and 1050 rpm respectively show service lives of 1.09 × 105 and 1.13 × 105 cycles. It was also noted that as the tool rotation speed increases, the service life of the obtained joint improves. The reduction in fatigue strength of the FSW (friction stir welding) joints compared to the base metal can be explained by the decrease in yield strength as observed in the tensile test. Additionally, the presence of tool groove traces is still evident even after surface treatment. For the 10-mm-thick specimens, the fatigue results confirm those obtained from the tensile and three-point bending tests. These results are very poor due to the presence of internal defects in the weld joints. In this context, we present an analysis of the fracture surfaces of the tested specimens. By focusing on the crack initiation location, it is evident that the potential rupture zones vary for each of the welding configurations.

The fracture facies of the test specimens were observed under a scanning electron microscope (SEM) in order to identify the zones and the crack initiation mechanisms that led to the failure of the joints welded by FSW. Fig. 14 illustrates the typical appearance of the fatigue fracture surfaces of the specimens. We note that there are three stages before the final rupture: crack initiation, a zone of propagation, and finally a zone of sudden rupture. Note that several initiation sites can be observed on the same fracture facies; this suggests that the fracture is associated with the coalescence of several cracks. In Fig. 15, fatigue “striations” can be seen on the fracture surfaces, which are typical of fatigue propagation.

Fig. 14
figure 14

Typical appearance of the plane fatigue rupture morphology of the FSW weld joint at 1050 rpm.

Fig. 15
figure 15

SEM observation of the fatigue streaks

3 Conclusion

The objective of the present study has been oriented on the collect test database on mechanical properties of friction stir welded of local aluminum alloys produce in the region. The advantages of the FSW techniques include reduced cost, improved quality, and flexibility. A series of tensile specimen’s tests, three points bending and fatigue, were undertaken on friction stir and fusion welded aluminum alloys as well as base alloys. All the tests were taken a care to analyze the microstructure and microhardness, and optical and SEM micrography have been conducted. Fractographic analysis indicated that all the joints exhibited a ductile fracture behavior when subjected to mechanical testing, further supporting the presence of favorable mechanical properties. Further study is required to obtain the optimal process parameters and tool geometry for improving the strength and fracture mechanical properties of the FSW joints of the proposal aluminum alloy. The findings highlighted the distinctive microstructural differences, variations in microhardness, ultimate tensile strength, and elongation between the FSW longitudinal and transverse joints. These insights contribute to a deeper understanding of the joint characteristics and performance, aiding in the development and optimization of friction stir welding processes for aluminum alloy joints.