- Email: [email protected]
Analysis of material flow in the sheet forming of friction-stir welds on alloys of mild steel and aluminum
Journal of Materials Processing Technology 226 (2015) 115–124
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Analysis of material flow in the sheet forming of friction-stir welds on alloys of mild steel and aluminum Tsutomu Tanaka a,∗ , Tomotake Hirata a , Naruaki Shinomiya b , Nobuhiko Shirakawa b a b
Metallic Materials Section, Technology Research Institute of Osaka Prefecture, 2-7-1 Ayumino, Izumi, Osaka, 594-1157, Japan Machining & Molding Section, Technology Research Institute of Osaka Prefecture, 2-7-1 Ayumino, Izumi, Osaka, 594-1157, Japan
a r t i c l e
i n f o
Article history: Received 12 September 2014 Received in revised form 26 June 2015 Accepted 27 June 2015 Available online 19 July 2015 Keywords: Friction stir welding Dissimilar welds Numerical simulation Deep drawing Aluminum alloy Steel
a b s t r a c t The deep drawability of and material flow in friction-stir welds of dissimilar steel/aluminum were investigated in this study. Steel/Al tailor-welded blanks (TWBs) were found to possess sufficient joint strength and deep drawability with limiting drawing ratios (LDRs) as low as 1.7. However, the drawability of these TWBs was lower than that of the two base materials. Fracturing was observed near the weld line on the flat bottom of the aluminum alloy side. In addition, the aluminum alloy near the fractured area in the TWB cup was subjected to significant strain near the forming limit, correlating to a significant loss in their deep drawability. Numerical simulation results indicated that slightly enhancing the friction coefficient between the punch and the blank significantly mitigated the significant strain near the fracture area. Deep drawing tests verified that the LDRs of TWBs obtained using rosin were increased by nearly 19% compared with that obtained using conventional press oil. Therefore, this approach of enhancing the friction coefficient can be effectively applied to improve the formability of dissimilar steel/aluminum TWBs. © 2015 Elsevier B.V. All rights reserved.
1. Introduction In the automotive industry, the development of new materials with superior mechanical properties, namely, strength, formability, and low weight is in high demand. In particular, the reduction of vehicle weight is extremely important. For example, Schubert et al. (2001) explained the importance of lightweight components in all applications that produce moving masses. Recently, thinner high-strength sheets and tailor-welded blanks (TWBs) have been widely utilized to address this problem. TWBs combine sheets of different materials that are welded together prior to the forming process; each base sheet has unique physical or material characteristics. TWBs suit the development of unique materials with locally different properties, namely, heterostructures; their use permits the reduction of the cost of the materials and assembly. In addition, TWBs are generally used in the production of complicated press-work, requiring high-quality mechanical properties and formability. Many experimental and numerical reports have discussed TWBs. Merklein et al. (2014) explained the potential practical uses and technical issues in tailor-welded blanks, patchwork blanks, tailor-rolled blanks, and tailor heat-treated blanks. Xu
∗ Corresponding author. Fax: +81 725 51 2749. E-mail address: t [email protected] (T. Tanaka). http://dx.doi.org/10.1016/j.jmatprotec.2015.06.030 0924-0136/© 2015 Elsevier B.V. All rights reserved.
et al. (2014) investigated the mechanical characteristics of highstrength steel TWBs. Panda et al. (2007) examined the formability of laser-welded interstitial-free steel blanks with different thicknesses via stretch-forming tests and finite element analysis. In addition, Choi et al. (2000) investigated the characteristics of weldline movement in TWBs during deep-drawing processes. However, these reports have primarily addressed TWBs made with steels of different thicknesses or strengths. In contrast, the automotive industry has begun using aluminum alloys to reduce vehicle weight. Therefore, interest in dissimilar TWBs of aluminum and steel has recently increased. However, the production of sound aluminum and steel welds is difficult using conventional fusion methods, because of the formation of considerable intermetallic compounds (IMC) by the high heat input. These IMCs are generally extremely hard, leading to fast rapture in joints. Therefore, a joining technology utilizing low-heat input, such as solid-state welding, is required to produce these dissimilar TWBs. Friction bonding is one typical technology used to achieve solidstate welding. Ikeuchi et al. (2005) examined the microstructures of the friction-bonded interfaces between steel and aluminum alloys, and indicated that the IMC layer thickness could be controlled by the friction time. Thomas et al. (1991) developed a new solid-state welding technology called friction stir welding (FSW). In FSW, joining is achieved by a combination of the heat generated by friction and the strong material flow induced by a rotating tool. FSW can
116
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
prevent heat generation during the welding process, for which it has attracted attention in the production of sound aluminum and steel welds. Fukumoto et al. (2004) examined the possibility of welding between a cast aluminum alloy and mild steel using FSW, establishing the basis for the dissimilar butt-welding technique. Kimapong and Watanabe (2004) reported adequate FSW conditions for butt-welding a plate of A5083 aluminum alloy to a mild steel plate. Uzun et al. (2005) investigated the microstructural, hardness, and fatigue properties of A6013 aluminum alloy friction-stir-welded to stainless steel. Lee et al. (2006) examined in detail the IMCs of dissimilar FSW joints between stainless steel and A6056 aluminum alloy, using transmission electron microscopy (TEM). Sun et al. (2013) reported on the formation of an amorphous structure along the joint interface between A6061 aluminum alloy and mild steel. Movahedi et al. (2013) examined the effect of annealing treatments on the joint strength of dissimilar A5083 aluminum alloy and steel friction-stir lap welds. Tanaka et al. (2009) reported on the relationships between welding temperature, IMC thickness, and bond strength of friction-stir welds between various aluminum alloys and mild steel. In addition, Tanaka et al. (2011) investigated the initiation and growth mechanisms of IMCs during FSW. As described, most studies concerning dissimilar friction-stir welds to date focused on their bond strengths and microstructures. Morita et al. (2009) reported on the deep drawability and bendability of friction-stir-welded TWBs of a commercial A6061 aluminum alloy and one variety of mild steel. However, the effects of combining different base materials on the formability of TWBs remain ambiguous. A thorough understanding of the formability of dissimilar friction-stir-welded TWBs is required to adopt them for industrial applications. This study was undertaken to clarify the formability of several different types of friction-stir-welded TWBs, consisting of mild steel and three different aluminum alloys. In addition, the material flow behaviors and severity of forming during deep drawing were investigated by both numerical analysis and experimental measurements. Finally, an effective improvement method of deep drawability was proposed to facilitate the use of dissimilar TWBs in the fabrication of wide varieties of complex parts.
2. Experimental procedures 2.1. Base materials and welding conditions Zinc-coated steel sheets and sheets of the three A1100-O, A5052-O, and A5182-O aluminum alloys were used in this study. The sheet thickness for all materials was 1 mm. A FSW tool with a screw thread probe made of SKD61 tool steel was used with specifications as follows: 10 mm shoulder diameter; 3 mm probe diameter; 0.9 mm probe length. The inclination angle during FSW was 2◦ . FSW was performed at a tool rotation speed of 1500 rpm and a weld travel speed of 300 mm/min; these welding conditions were determined from preliminary examinations. Similar welded blanks (A5052-A5052) and four dissimilar TWBs (A5182-A5052, steel-A1100, steel-A5052, and steel-A5182) were made by FSW. When the same types of materials were treated by FSW (A5052A5052 and A5182-A5052), the center of the moving tool conformed to the butt line, as shown in Fig. 1(a). Fukumoto et al. (2004) and Kimapong and Watanabe (2004) described the plunging of the rotation tool into the aluminum, successfully producing dissimilar welds by placing the aluminum alloy on the retreating side. The advancing side of welds is the side on which the rotation of the tool proceeds in the same direction as the motion of the tool itself; the opposite side is referred to as the retreating side. Therefore, the probe was placed directly into the aluminum alloy, fixed on the retreating side, for dissimilar FSW, as shown in Fig. 1(b). In
Fig. 1. Schematic of friction-stir welding: (a) similar welding and (b) dissimilar welding.
this study, the probe was offset from the steel’s faying surface by approximately 0.2 mm. 2.2. Tensile and deep drawing tests The tensile properties of the base materials were measured at angles of 0◦ , 45◦ , and 90◦ to the rolling direction (RD). The dimensions of the tensile specimens are shown in Fig. 2(a). The tensile specimen was cut using an electrical discharge machine; the yield stress, tensile stress, elongation, work-hardening exponent (n-value), and Lankford value (r-value) were determined to mechanically characterize the materials. The n-value was calculated from the slope of the true stress–true strain curve between the true strains of 0.1 and 0.15. The crosshead speed in the tensile tests for all cases was 3 mm/min. Tensile specimens of the frictionstir welds were also machined transverse to the welding direction, as shown in Fig. 2(b) and (c), and the tensile strengths of the joints were measured using these samples. Cylindrical deep drawing tests were performed to measure the formability of the base materials and the dissimilar FSWs. The tool geometries for the deep drawing testing were as follows: 40 mm punch diameter; 4 mm punch shoulder radius; 42.5 mm die hole diameter; and 10 mm die shoulder radius. Conventional press oil (Model Number: S-3) manufactured by HOKOKU OIL Co., Ltd. was used as the lubricant, with a kinetic viscosity at 313 K of approximately 85 mm2 /s. The punch speed and the blank holder force were 60 mm/min and 3 kN, respectively. The geometry of the initial blank was circular with the diameter ranging from 74 to 90 mm to reveal the limiting drawing ratio (LDR), calculated as: LDR =
Dmax dp
where Dmax is the maximum blank diameter drawn without fracture and dp is the punch diameter. In the TWBs’ deep drawing tests, the weld line was located at the center of the blank. 2.3. Metallurgical analysis of material flow during deep drawing The material flow behaviors during the deep drawing of both the base metals and the TWBs were investigated using electron backscatter diffraction (EBSD) analysis and the grid-marking method. For EBSD analysis, the deep-drawn cups of the A5052 base material and the steel-A5052 TWB were sectioned transverse to the rolling direction and to the weld line, respectively. A section of the flat bottom of each section was polished by an Ar ion beam. The EBSD analysis was performed by a field emission scanning elec-
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
117
Fig. 2. Dimensions of tensile specimens: (a) base material, (b) steel-aluminum weld, and (c) aluminum–aluminum weld.
tron microscope (FE–SEM) with an orientation imaging microscopy system produced by TexSem Laboratories, Inc. An analysis of the material flow by the grid-marking method was performed in the cup bottom itself during deep drawing. 2.4. Numerical simulation conditions Numerical simulation was conducted to examine the material flow behavior during the deep drawing of the TWBs and to predict appropriate forming conditions. The simulation of the deep drawing of the base materials and the TWBs was performed by a commercially available finite element code, LS-DYNA. The tools were modeled as rigid shell elements, while the blanks were modeled with four-noded shell elements with shell thicknesses of 1 mm. To reduce the computation time of this simulation, the geometry of the blanks was scaled to one-fourth for the base material and to one-half for the welds. The stress–strain relations of the base materials in the simulations, accounting for the strain hardening of the materials during plastic deformation, was described using their experimental stress–strain curves in the rolling direction. Hill (1948) proposed the quadratic yield criterion, a commonly used yield function in numerical simulations for metal forming. Barlat et al. (2003) suggested the orthotropic yield function. In this
Fig. 3. Top surface views at the weld area: (a) steel-A1100 weld, (b) steel-A5052 weld, (c) steel-A5052 weld, (d) A5182-A5052 weld, and (e) A5052-A5052 weld.
study, Barlat’s yield function was applied for the aluminum alloy because it successfully described the plastic behavior of materials with low r-values, like the aluminum alloys examined here. Meanwhile, the yield behavior of steel was described using Hill’s yield criterion. The friction at the tool-blank interface was defined as follows. The experimental punch load curves of all base materials during the deep drawing tests were measured. By numerical simulation, the punch load curves at several friction coefficients were estimated. The friction coefficient at which the simulated punch load curve matched well with the experimental curve was utilized as the interfacial friction in this study. The friction coefficients for the aluminum alloys and the steel were defined as 0.17 and 0.12, respectively.
118
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 4. Typical cross-sectional microstructure of the interface region of friction-stir-welded zinc-coated steel and A5052 aluminum alloy.
including the effect of alloy compositions on IMC formation will be needed in the future. 3.3. Deep drawabilities of dissimilar TWBs
Fig. 5. Higher-magnification view of the microstructure at the interface of the steelA5052 weld.
3. Results 3.1. Appearance and microstructure of dissimilar TWB produced by FSW The top surface appearance of the welds is shown in Fig. 3. All welds are successfully joined without porosity or defects. As shown in Fig. 3(a)–(c), the top surface of the dissimilar steel/aluminum welds is rough, with some flashes observed in the steel-A5052 and steel-A5182 weld. In addition, the bead width in the steel-A5052 weld is larger than that in the steel-A5182 weld, since the weld tool was inserted more deeply, because of the lower yield strength of A5052. Therefore, larger flashes occur in the steel-A5052 weld than in the steel-A5182 weld. The top surfaces of the A5182-A5052 and A5052-A5052 welds are smooth, as can be seen in Fig. 3(d) and (e). Fig. 4 shows a typical cross-section of the steel-A5052 weld; no void or cracks are observed at the joint interface or the weld zone. The highly magnified SEM image at the interface between steel and aluminum alloy is also shown in Fig. 5. The IMC thickness at the interface is less than 300 nm, as shown in Fig. 5. 3.2. Tensile properties of base materials and FSWs The tensile properties of the base materials and of the welds are shown in Tables 1 and 2, respectively. “Joint efficiency” is defined as the ratio of the strength obtained for the weld divided by the lower strength between those of the two base materials. The joint efficiency of the welds is approximately 100% in all cases, excepting the steel-A5182 weld, which exhibits a drop of about 15% in its tensile strength. Elliott and Wallach (1981a,b) reported that the Mg content of A5182 facilitated the formation of IMCs, resulting in the loss of joint strength. In addition, Yamamoto et al. (2005) showed that the oxide layer formed by the increased Mg content reduced the joint strength. Therefore, the lower joint efficiency of the steelA5182 weld in this study can be attributed to microstructural changes induced by A5182’s large Mg content. Further research
3.3.1. LDR of welded blanks and the effect of the mismatch in material strength on the LDR values Table 2 shows the deep drawabilities of the base materials and welds. “LDR efficiency” is defined as the ratio of the LDR for the weld divided by the lower LDR between those of the two base materials. From Table 2, dissimilar TWBs, including brittle IMCs, may be deepdrawn with LDRs as low as 1.7. However, the deep drawabilities of the dissimilar TWBs show a different trend from that of the tensile strength of the TWBs. The LDR values of all TWBs are consistently lower than those of their component base materials, despite the high joint efficiency in terms of tensile strength. A particularly significant decrease in the LDR is observed for the steel-A1100 weld. Panda et al. (2007) reported on a limiting dome height in steelonly TWBs between sheets of different thickness existing midway between the two sheets. This implies that steel–aluminum TWBs have different formability from steel-only TWBs. Fig. 6 shows the appearance of five weld cups drawn at their LDR. For dissimilar welds, the projection of the weaker materials is observed near the weld line in the form of flanges, indicated by circles in Fig. 6. The weld lines appear to have moved toward the stronger materials. In particular, a marginal shift in the weld line toward the steel side is observed for the steel-A1100 weld, as shown in Fig. 6(a). However, for the similar weld of two 5052 aluminum alloy sheets, shown in Fig. 6(e), the weld line at the flat bottom of the cup remains straight, and exterior anomalies in the drawing cup, such as flange projections and weld-line position shifts, are not observed. These are considered to result from the mismatches in material strengths and the differences in plastic stain ratios and material flow characteristics between the steel and aluminum alloys. 3.3.2. Failure of dissimilar TWBs The appearances of the base materials and two welds after their deformation to failure are shown in Fig. 7. As observed in Fig. 7(b) and (c), fractures in the A5052-A5052 weld and the A5052 base material cups occur at the punch radius, in accordance with the general trends observed in deep-drawing tests on standard nonwelded cups. For dissimilar TWBs, however, fractures are observed to occur parallel to the weld line on the side of the weaker material, at a short distance from the weld line itself. This demonstrates that the deformation behavior of dissimilar TWBs during deep-drawing differs from those of single materials and similar welds. 3.3.3. Material flow behavior during deep drawing Fig. 8 shows inverse pole figures (IPF) obtained by EBSD analysis for the single A5052 base material, its drawn cup, and the steelA5052 weld drawn cup. Analysis of the drawn cups was performed on sections of the cups’ flat bottoms, approximately 6 mm from the cups’ centers. The upper and lower rows of Fig. 8 show the IPFs
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
119
Table 1 Mechanical properties of base materials. Materials
Angle to RD
Yield strength (MPa)
Tensile strength (MPa)
Elongation (%)
n-value
r-value
r¯
r
A1100O
0 45 90
31.8 31.3 32.0
102.2 97.7 99.7
33.8 38.4 40.3
0.318 0.318 0.328
0.589 0.955 0.676
0.794
-0.323
A5052O
0 45 90
91.9 93.1 91.3
206.5 202.4 199.1
28.7 27.6 27.1
0.327 0.338 0.328
0.908 0.522 0.708
0.665
0.286
A5182O
0 45 90
136.9 135.3 138.2
283.1 287.3 288.9
26.9 30.3 28.0
0.394 0.387 0.386
0.636 0.837 0.754
0.766
-0.142
Zinccoated steel
0 45 90
293.0 304.4 302.7
355.2 366.8 356.5
41.2 38.2 43.0
0.260 0.246 0.253
1.069 0.799 1.539
1.052
0.505
Table 2 Tensile properties and deep drawability of friction-stir welds. Mater.1
Mater.2
Joint strength (MPa)
Joint efficiency (MPa)
Mater.1 LDR
Mater.2 LDR
TWB LDR
LDR efficiency (%)
Zinc-coated steel Zinc-coated steel Zinc-coated steel A5182-O A5052-O
A1100-O A5052-O A5182-O A5052-O A5052-O
99.5 198.5 244.9 205.7 199.5
100.2 97.9 85.4 101.5 98.4
2.175 2.175 2.175 2.150 2.100
2.075 2.100 2.150 2.100 2.100
1.700 1.850 1.925 2.050 2.125
81.9 88.1 89.5 97.6 101.2
Fig. 6. Appearance of TWBs deformed in the limiting drawing ratio (LDR) tests: (a) steel-A1100 weld, (b) steel-A5052 weld, (c) steel-A5182 weld, (d) A5182-A5052 weld, and (e) A5052-A5052 weld.
Fig. 7. Appearance of drawing cups deformed to failure: (a) steel-A1100 weld, (b) A5052-A5052 weld, and (c) A5052-O base material.
parallel and perpendicular, respectively, to the rolling direction for the single A5052 base material and its drawn cup and the weld line for the dissimilar TWB drawn cup. As shown in Fig. 8(a) and (b), the dominant crystal orientations parallel to the rolling direction in the single 5052 base material are <0 0 1> and <1 1 2>, while that perpen-
dicular to the rolling direction is <0 0 1>. As observed in Fig. 8(c) and (d), the dominant crystal orientation in the single A5052 drawn cup is nearly identical to that in the single A5052 base material shown in Fig. 8(a) and (b). This suggests that the material of the cup bottom is subjected to equi-biaxial tensile stresses and undergoes very little
120
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 8. Inverse pole figures of A5052 base metal, drawn A5052 cup, and drawn steel-A5052 weld cup.
Fig. 9. Strain distributions at the flat bottom of the drawn (a) A5052-O base material and (b) steel-A5052 TWB.
deformation during the deep-drawing test. Meanwhile, in Fig. 8(a), (e), and (g), the dominant crystal orientation parallel to the weld in the TWB drawn cup line is almost equivalent to that in the single A5052 base material. However, a new <1 1 1> texture peak perpendicular to the weld line is observed, as shown in Fig. 8 (f) and (h), where the <1 1 1> texture is one of the preferred orientations to the tensile axis of aluminum. This implies that the aluminum in the dissimilar TWB does not deform parallel to the weld line during deep drawing, but instead deforms perpendicular to the weld line near the fracture area; this corresponds to the plane strain state. The strain distribution at the flat bottoms of the cups of the single A5052 base material and the steel-A5052 weld was analyzed using the displacement of the 2 mm-interval lattice lines on the sample surface. Fig. 9(a) represents the strain both transverse and longitudinal to the rolling direction of the single A5052 base material cup, while Fig. 9(b) represents the strains transverse and longitudinal to the weld direction of the dissimilar TWBs. In addition, Fig. 10 presents forming limit diagrams (FLDs) for the flat bottoms of the two drawing cups, as calculated from the transverse and longitudinal strains. For the single A5052 base material cup shown in Figs. 9 (a) and 10, the strain mode is seen to be equibiaxial in tension, with strains below 5%. This matches reasonably well with the results shown in Fig. 8. As shown in Fig. 9(b), although both materials in the TWB drawn cup experienced very little longitudinal strain relative to the weld line, a significant transverse strain above 20% is observed at a small distance from the joint interface in the A5052 aluminum alloy. This implies that the significant
Fig. 10. Forming limit diagrams of the flat bottom of the drawing cups and of the A5052-O aluminum alloy.
transverse strain observed on the aluminum side results from the plane strain state. The forming limit in the plane strain state is well known to be substantially lower than those related to other strain states. Kohara et al. (1985) studied the FLD curves for the A5052O aluminum alloy, reporting that the major strain of failure in the plane strain state is approximately 20%. The FLD curve of A5052-O aluminum alloy is also included in Fig. 10. As observed in Fig. 10, many strain plots of the steel-A5052 TWB intersect with the FLD curve of the A5052-O aluminum alloy. This implies that the flat
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 11. Micro-hardness profile along the centerline of the cross-section of the steelA5052 weld.
bottom of the TWB cup is subjected to significant strain, near the forming limit, which results in the premature failure of drawn cups from dissimilar TWBs. Consequently, local deformation in the flat bottom on the aluminum side must be avoided in order to improve the drawability of dissimilar TWBs of steel and aluminum alloys. The simulation analysis and discussion henceforth were focused on the dissimilar TWBs of steel and aluminum alloys, in the interest of contributing to the improvement of the deep drawability of dissimilar TWBs. 3.4. Numerical simulation For a precise numerical analysis, the mechanical properties of the weld zone must be properly measured and incorporated into the simulation. For example, Padmanabhan et al. (2008) described the material behavior of the weld zone using a rule based on the mixture of aluminum and steel. Lee et al. (2009) quoted the experimental results reported by Saunders and Wagoner (1996), in which the mechanical properties of the weld zone were experimentally measured using small-sized tensile samples. In addition, Leitão et al. (2011) reported that the formability of dissimilar A5182A6016 TWBs was strongly influenced by the weld bead shape. In this study, micro-hardness tests were performed, as the mechanical properties of the weld zone can be difficult to measure because of the zone’s narrowness, as shown in Fig. 4. As observed in Fig. 11, the hardness of the weld zone only slightly exceeds that of the base aluminum alloy; it is approximately half that of the steel. Therefore, the change of mechanical properties in the weld zone is considered to have little influence on the formability of the TWBs. In fact, Ahmetoglu et al. (1995) and Panda and Kumar (2008) reported that the deformation behavior of TWBs could be well-simulated without incorporating the material properties of the weld zone; in these simulations, the yield strength ratio of the TWBs were 2.25 and 1.54, respectively. With the experimental verification of these prior conclusions, the change in the mechanical properties of the weld zone was neglected in this study. Top views of the drawing cups, as obtained from both experiments and simulations, are shown in Fig. 12. For the steel-A1100 weld in Fig. 12(a), a blank with a 68 mm diameter was drawn to a punch height of only 15 mm because of its extremely low formability. For the steel-A5052 and steel-A5182 welds in Fig. 12(b) and (c), blanks with diameters of 74 mm were drawn to a punch height of 20 mm. As observed in Fig. 12, the shapes predicted by the simulations mostly agree with the experimental results. However, slight differences of flange length between the experiments and sim-
121
ulations are observed. These are attributed to the non-uniform blank holder force. As shown in Fig. 3(a)–(c), rough weld beads and some flashes were observed in the steel–aluminum welds. The deep-drawing tests in this study were performed without removing flashes on the blank surface, because of the difficulty of machining thin sheets distorted by FSW. Therefore, the rough weld bead and flashes are considered to contribute to non-uniform blank holder force, resulting in the mismatch in flange length between the experiment and simulation. However, the deformation behavior at the flat bottom was studied more carefully, because this behavior is more closely linked with the premature failure of dissimilar TWBs. More precise numerical analysis of the flange behavior will be performed in the future. The major strain distributions of TWBs as simulated by LS-DYNA are shown in Fig. 13. The aluminum alloy side close to the weld line at the bottom of the cup is subjected to a larger strain than the steel side. A similar trend is observed in the experimental results, as shown in Fig. 9. This similarity verifies the utility of the numerical analysis using LS-DYNA in understanding the material flow behavior during the forming of TWBs. 4. Discussion The TWBs with high strength ratios in their base materials, such as the steel-A1100 TWB, failed at local deformation areas in the lower-strength material as a result of the materials’ strength mismatch. In addition, the plane strain state of the local deformation area, attributed to constraint by plastic deformation resulting from the weld line, increases the likelihood of premature fractures. Therefore, inhibiting this local deformation may improve the formability of TWBs. Many countermeasures restricting local deformation have been already proposed. Leitão et al. (2009) and Padmanabhan et al. (2008) reported that the adjustment of the location of the blank holder force could mitigate the localization of deformations observed in dissimilar TWBs. In addition, Panda and Kumar (2008) demonstrated that the limiting dome height of TWBs increased with the application of a counter pressure, while Heo et al. (2001a,b) reported that the movement of the weld line could be controlled by adequate draw-bead installation. In this study, the coefficient of friction between blanks and forming tools was of greatest interest; the effects of the friction conditions on the curvature of the weld line and on the local deformation during the deep drawing of TWBs were examined most closely. 4.1. Simulation analysis concerning the effect of friction conditions on the drawability Numerical simulation of deep drawing was performed in order to clarify the effect of friction and to suggest methods of inhibiting local deformation in the steel-A1100 weld, where the most noticeable shift in the weld line toward the steel side and the highest level of strain on the A1100 side was observed. Only the friction coefficient between the punch and the blank was varied in this simulation. The effect of the friction coefficient on the drawability was investigated in three different cases: variation of friction on only the A1100 side, on only the steel side, and over the whole area. The friction coefficient was set to 0.3, 0.45, and 0.6. Figs. 14–16 show the simulated major strain distributions of the steel-A1100 weld when the friction coefficient is varied on only the A1100 side, on only the steel side, and over the whole area, respectively. For comparison, the results shown in Fig. 13(a), i.e., those found using press oil as the lubricant, are included on the far left in Figs. 14–16. As shown in Fig. 14, the region of high strain on the A1100 side becomes wider and more prominent with the increase of the friction coefficient of the aluminum alloy. When the friction
122
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 12. Top views of the drawing cups obtained by experiment versus numerical simulation: (a) steel-A1100 weld, (b) steel-A5052 weld, and (c) steel-A5182 weld.
Fig. 13. Simulated major strain distributions of the TWBs for the (a) steel-A1100 weld, (b) steel-A5052 weld, and (c) steel-A5182 weld.
Fig. 14. Simulated major strain distributions of the steel-A1100 weld for friction coefficients on the A1100 side of (a) 0.17, (b) 0.30, (c) 0.45, and (d) 0.60.
Fig. 15. Simulated major strain distributions of the steel-A1100 weld for friction coefficients on the zinc-coated steel side of (a) 0.12, (b) 0.30, (c) 0.45, and (d) 0.60.
coefficient is enhanced on only the steel side and over the whole area, however, greater friction coefficients correlate to greater linearity of weld lines. This in itself corresponds to the more isotropic deformation behavior of the cup bottom during deep-drawing, as shown in Figs. 15 and 16. These results suggest that the enhancement of the friction coefficient on only the steel side or over the
whole blank area may improve the deep drawability of dissimilar TWBs. During the deep-drawing of dissimilar steel/aluminum TWBs, the stress required to draw into the cup section is lower on the aluminum alloy side than on the steel side; this stress is transferred from the flange to the flat bottom. Because of the higher required stress on the steel side, the aluminum alloy material on
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
123
Fig. 16. Simulated major strain distributions of the steel-A1100 weld for friction coefficients on the steel and A1100 sides of (a) 0.12 and 0.17, (b) 0.30 and 0.30, (c) 0.45 and 0.45, and (d) 0.60 and 0.60, respectively.
the flat bottom, punch radius, and cup wall is pulled toward the steel side of the TWB. The friction coefficient between the punch and the blank mainly affects the material flow at the punch radius. When the friction coefficient is enhanced only on the aluminum side, the material flow on the aluminum side toward the steel side is impeded at the punch radius by the high frictional force. Therefore, only the aluminum material on the flat bottom is pulled out toward the steel side, resulting in an expanded area of local deformation on the A1100 side, as shown in Fig. 14. When the friction coefficient is enhanced only on the steel side, the material flow is notably inhibited at the punch radius on the steel side. The stress on the steel side necessary to draw the material at the flange is not transferred to the flat bottom; therefore, the material flow toward the steel side on the A1100 side is inhibited, as shown in Fig. 15. When a high friction coefficient is in effect over the whole area, the material flow at the punch radius is inhibited on both sides of the TWB. Thus, the same trend observed in Fig. 15 is seen in Fig. 16. In Figs. 15 and 16, when the friction coefficient changes from 0.30 to 0.60, the strain distribution does not experience significant change. Consequently, local deformation can be restricted effectively by even slight enhancement of the friction coefficient to 0.3.
Fig. 17. Suitability of deep-drawing for a steel-A5052 weld under each friction condition.
4.2. Verification tests concerning the effect of friction conditions on the drawability Deep drawing tests were performed to determine the validity of the numerical predictions mentioned above. In the verification tests, different friction conditions were imposed by using rosin as a lubricant. The friction coefficients with the use of rosin were 0.25 for steel and 0.24 for the aluminum alloy, respectively, estimated by conforming the experimental punch load curve to the simulated curve as explained in Section 2.4. Rosin was applied only between the punch and blank, while press oil was used on the other contact surfaces, as in the numerical analysis. Fig. 17 shows the deep drawability of a steel-A5052 weld and the base materials under each friction condition. The vertical axis shows the initial blank diameter before the deep-drawing test. The double circle in this figure indicates the completion of the deep drawing test without failure; therefore, the maximum blank diameter with double circles corresponds to the blank diameter of the LDR value. The cross symbol indicates the failure of the blank during the deep-drawing test. As observed in Fig. 17, the enhancement of the friction coefficient by using rosin significantly improves the deep drawability of all blanks. The LDR of the TWB using rosin was approximately 19% higher than that of the TWB using press oil. In addition, although the LDR of the TWB using press oil was 88% of that of the weaker base material, the LDR using the rosin was nearly equal to that of the weaker material. The steel-A5052 cups drawn under the different friction conditions are shown in Fig. 18. As shown in Fig. 18, the cup height of the TWB drawn using rosin is significantly
Fig. 18. Steel-A5052 cups drawn under different friction conditions.
greater than that using the press oil; its weld line remains straight. These results validate the prediction of the numerical simulation described above. In summary, the enhancement of the friction coefficient between the punch and blank was revealed to be capable of improving the deep drawability of dissimilar steel/aluminum TWBs. 5. Conclusions In order to examine the formability performance of various dissimilar steel/Al TWBs produced by FSW, experiments and numerical simulations on deep-drawing were performed. In addition, possible methods for improving the formability of the TWBs
124
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
were explored by both numerical simulation and experimentation. The following conclusions were obtained. • The tensile strengths of the TWBs were nearly equivalent to those of the base materials. Although the TWBs had deep drawability with LDRs as low as 1.7, the LDR values were lower than those of the base materials; this tendency opposed that observed for steel-only TWBs. • After deep-drawing tests, the weld line of the TWBs was observed to shift toward the stronger side, and fracturing was observed near the weld line on the flat bottom of the aluminum alloy side. The degree of the weld line shift increased with the increasing difference in the strengths of the two base materials, resulting in the deterioration of the TWBs’ deep drawability. • Each type of analysis regarding material flow behavior, including inverse pole figures from EBSD measurements, the grid marking method, and numerical simulations, indicated that the aluminum alloy near the fractured area in the TWBs was locally deformed in the plane strain state. Major strain exceeding 20%, the forming limit of aluminum alloy in the plane strain state, was observed near the fracture area even prior to failure. • The effect of friction between the forming tools and the blanks on the local deformation of the aluminum alloy side was investigated by both numerical simulation and experimental tests. The simulation analysis indicated that the curvature of the weld line and the local deformation in the aluminum alloy would decrease with increasing the friction coefficient between the punch and blank. • Actual deep-drawing tests using rosin to increase the friction coefficient resulted in more uniform material flow on the flat bottom, and the LDR of the steel-A5052 TWB increased by nearly 19% relative to that drawn using press oil. Therefore, it could be concluded that enhancing the friction coefficient may be an effective solution to improve the formability of dissimilar TWBs. Acknowledgement The authors gratefully acknowledge financial support from the Amada Foundation for Metal Work Technology. References Ahmetoglu, M.A., Brouwers, D., Shulkin, L., Taupin, L., Kinzel, G.L., Altan, T., 1995. Deep drawing of round cups from tailor-welded blanks. J. Mater. Process. Technol. 53, 684–694. Barlat, F., Brem, J.C., Yoon, J.W., Chung, K., Dick, R.E., Choi, S.H., Pourboghrat, F., Chu, E., Lege, D.J., 2003. Plane stress yield function for aluminum alloy sheets – part I: theory. Int. J. Plast. 19, 1297–1319. Choi, Y., Heo, Y., Kim, H.Y., Seo, D., 2000. Investigations of weld-line movements for the deep drawing process of tailor welded blanks. J. Mater. Process. Technol. 108, 1–7. Elliott, S., Wallach, E.R., 1981a. Joining aluminum to steel Part 1-Diffusion bonding. Metal Constrution 13, 167–171. Elliott, S., Wallach, E.R., 1981b. Joining aluminum to steel Part 2-Friction bonding. Metal Construction 13, 221–225. Fukumoto, M., Tsubaki, M., Shimada, Y., Yasui, T., 2004. Welding between ADC12 and SS400 by means of friction stirring. Q. J. Jpn. Weld. Soc. 22, 309–314.
Heo, Y., Choi, Y., Kim, H.Y., Seo, D., 2001a. Characteristics of weld line movements for the deep drawing with drawbeads of tailor-welded blanks. J. Mater. Process. Technol. 111, 164–169. Heo, Y.M., Wang, S.H., Kim, H.Y., Seo, D.G., 2001b. The effect of the drawbead dimensions on the weld-line movements in the deep drawing of tailor-welded blanks. J. Mater. Process. Technol. 113, 686–691. Hill, R., 1948. A theory of the yielding and plastic flow of anisotropic metals. Proc. R. Soc. A193, 281–297. Ikeuchi, K., Yamamoto, N., Takahashi, M., Aritoshi, M., 2005. Effect of interfacial reaction layer on bond strength of friction bonded joint of Al alloys to steel. Trans. JWRI 34, 1–10. Kimapong, K., Watanabe, T., 2004. Friction stir welding of aluminum alloy to steel. Weld. J. 83, 277–282. Kohara, S., Mototsugu, K., Aoki, K., 1985. Forming limit curves of 5052 alloy sheets. J. Jpn. Inst. Light Metal 35, 700–704. Lee, W.B., Schmuecker, M., Mercardo, U.A., Biallas, G., Jung, S.B., 2006. Interfacial reaction in steel-aluminum joints made by friction stir welding. Scripta Mater. 55, 355–358. Lee, W., Chung, K.H., Kim, D., Kim, J., Kim, C., Okamoto, K., Wagoner, R.H., Chung, K., 2009. Experimental and numerical study on formability of friction stir welded TWB sheets based on hemispherical dome stretch tests. Int. J. Plast. 25, 1626–1654. Leitão, C., Emílio, B., Chaparro, B.M., Rodrigues, D.M., 2009. Formability of similar and dissimilar friction stir welded AA 5182-H111 and AA 6016-T4 tailored blanks. Mater. Des. 30, 3235–3242. Leitão, C., Zhang, B.K., Padmanabhan, R., Rodrigues, D.M., 2011. Influence of weld geometry and mismatch on formability of aluminium tailor welded blanks: numerical and experimental analysis. Sci. Technol. Weld. Join. 16, 662–668. Merklein, M., Johannes, M., Lechner, M., Kuppert, A., 2014. A review on tailored blanks-Production, applications and evaluation. J. Mater. Process. Technol. 214, 151–164. Morita, T., Sakamoto, A., Mabuchi, S., Iizuka, T., 2009. Strength and formability of friction stir welded aluminum alloy with cold rolled steel. J. Soc. Mater. Sci. Jpn. 58, 317–322. Movahedi, M., Kokabi, A.H., Seyed Reihani, S.M., Cheng, W.J., Wang, C.J., 2013. Effect of annealing treatment on joint strength of aluminum/steel friction stir lap weld. Mater. Des. 44, 487–492. Padmanabhan, R., Oliveira, M.C., Menezes, L.F., 2008. Deep drawing of aluminum-steel tailor-welded blanks. Mater. Des. 29, 154–160. Panda, S.K., Kumar, D.R., Kumar, H., Nath, A.K., 2007. Characterization of tensile properties of tailor welded IF steel sheets and their formability in stretch forming. J. Mater. Process. Technol. 183, 321–332. Panda, S.K., Kumar, D.R., 2008. Improvement in formability of tailor welded blanks by application of counter pressure in biaxial stretch forming. J. Mater. Process. Technol. 204, 70–79. Saunders, F.I., Wagoner, R.H., 1996. Forming of tailor-welded blanks. Metall. Mater. Trans. 27A, 2605–2615. Schubert, E., Klassen, M., Zerner, I., Wlaz, C., Sepold, G., 2001. Light-weight structures produced by laser beam joining for future applications in automobile and aerospace industry. J. Mater. Process. Technol. 115, 2–8. Sun, Y.F., Fujii, H., Takaki, N., Okitsu, Y., 2013. Microstructure and mechanical properties of dissimilar Al alloy/steel joints prepared by a flat spot friction stir welding technique. Mater. Des. 47, 350–357. Tanaka, T., Morishige, T., Hirata, T., 2009. Comprehensive analysis of joint strength for dissimilar friction stir welds of mild steel to aluminum alloys. Scripta Mater. 61, 756–759. Tanaka, T., Morishige, T., Hirata, T., 2011. Formation mechanism of intermetallic compound at the weld interface of dissimilar friction stir welded joint of pure aluminum and mild steel. Q. J. Jpn. Weld. Soc. 29, 101–106. Thomas M.W., Nicholas E.D., Needham J.C., Murch M.G., Templesmith P., Dawes C.J., 1991. GB Patent Applications No. 9,125,978.8, December 1991. Uzun, H., Donne, C.D., Argagnotto, A., Ghidini, T., Gambaro, C., 2005. Friction stir welding dissimilar Al 6013-T4 to X5CrNi18-10 stainless steel. Mater. Des. 26, 41–46. Xu, F., Sun, G., Li, G., Li, Q., 2014. Experimental investigation on high strength steel (HSS) tailor-welded blanks (TWBs). J. Mater. Process. Technol. 214, 925–935. Yamamoto, N., Takahashi, M., Aritoshi, M., Ikeuchi, K., 2005. Effect of interfacial layer on bond strength of friction-welded interface of Al–Mg 5083 alloy to mild steel. Q. J. Jpn. Weld. Soc. 23, 496–503.
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Analysis of material flow in the sheet forming of friction-stir welds on alloys of mild steel and aluminum Tsutomu Tanaka a,∗ , Tomotake Hirata a , Naruaki Shinomiya b , Nobuhiko Shirakawa b a b
Metallic Materials Section, Technology Research Institute of Osaka Prefecture, 2-7-1 Ayumino, Izumi, Osaka, 594-1157, Japan Machining & Molding Section, Technology Research Institute of Osaka Prefecture, 2-7-1 Ayumino, Izumi, Osaka, 594-1157, Japan
a r t i c l e
i n f o
Article history: Received 12 September 2014 Received in revised form 26 June 2015 Accepted 27 June 2015 Available online 19 July 2015 Keywords: Friction stir welding Dissimilar welds Numerical simulation Deep drawing Aluminum alloy Steel
a b s t r a c t The deep drawability of and material flow in friction-stir welds of dissimilar steel/aluminum were investigated in this study. Steel/Al tailor-welded blanks (TWBs) were found to possess sufficient joint strength and deep drawability with limiting drawing ratios (LDRs) as low as 1.7. However, the drawability of these TWBs was lower than that of the two base materials. Fracturing was observed near the weld line on the flat bottom of the aluminum alloy side. In addition, the aluminum alloy near the fractured area in the TWB cup was subjected to significant strain near the forming limit, correlating to a significant loss in their deep drawability. Numerical simulation results indicated that slightly enhancing the friction coefficient between the punch and the blank significantly mitigated the significant strain near the fracture area. Deep drawing tests verified that the LDRs of TWBs obtained using rosin were increased by nearly 19% compared with that obtained using conventional press oil. Therefore, this approach of enhancing the friction coefficient can be effectively applied to improve the formability of dissimilar steel/aluminum TWBs. © 2015 Elsevier B.V. All rights reserved.
1. Introduction In the automotive industry, the development of new materials with superior mechanical properties, namely, strength, formability, and low weight is in high demand. In particular, the reduction of vehicle weight is extremely important. For example, Schubert et al. (2001) explained the importance of lightweight components in all applications that produce moving masses. Recently, thinner high-strength sheets and tailor-welded blanks (TWBs) have been widely utilized to address this problem. TWBs combine sheets of different materials that are welded together prior to the forming process; each base sheet has unique physical or material characteristics. TWBs suit the development of unique materials with locally different properties, namely, heterostructures; their use permits the reduction of the cost of the materials and assembly. In addition, TWBs are generally used in the production of complicated press-work, requiring high-quality mechanical properties and formability. Many experimental and numerical reports have discussed TWBs. Merklein et al. (2014) explained the potential practical uses and technical issues in tailor-welded blanks, patchwork blanks, tailor-rolled blanks, and tailor heat-treated blanks. Xu
∗ Corresponding author. Fax: +81 725 51 2749. E-mail address: t [email protected] (T. Tanaka). http://dx.doi.org/10.1016/j.jmatprotec.2015.06.030 0924-0136/© 2015 Elsevier B.V. All rights reserved.
et al. (2014) investigated the mechanical characteristics of highstrength steel TWBs. Panda et al. (2007) examined the formability of laser-welded interstitial-free steel blanks with different thicknesses via stretch-forming tests and finite element analysis. In addition, Choi et al. (2000) investigated the characteristics of weldline movement in TWBs during deep-drawing processes. However, these reports have primarily addressed TWBs made with steels of different thicknesses or strengths. In contrast, the automotive industry has begun using aluminum alloys to reduce vehicle weight. Therefore, interest in dissimilar TWBs of aluminum and steel has recently increased. However, the production of sound aluminum and steel welds is difficult using conventional fusion methods, because of the formation of considerable intermetallic compounds (IMC) by the high heat input. These IMCs are generally extremely hard, leading to fast rapture in joints. Therefore, a joining technology utilizing low-heat input, such as solid-state welding, is required to produce these dissimilar TWBs. Friction bonding is one typical technology used to achieve solidstate welding. Ikeuchi et al. (2005) examined the microstructures of the friction-bonded interfaces between steel and aluminum alloys, and indicated that the IMC layer thickness could be controlled by the friction time. Thomas et al. (1991) developed a new solid-state welding technology called friction stir welding (FSW). In FSW, joining is achieved by a combination of the heat generated by friction and the strong material flow induced by a rotating tool. FSW can
116
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
prevent heat generation during the welding process, for which it has attracted attention in the production of sound aluminum and steel welds. Fukumoto et al. (2004) examined the possibility of welding between a cast aluminum alloy and mild steel using FSW, establishing the basis for the dissimilar butt-welding technique. Kimapong and Watanabe (2004) reported adequate FSW conditions for butt-welding a plate of A5083 aluminum alloy to a mild steel plate. Uzun et al. (2005) investigated the microstructural, hardness, and fatigue properties of A6013 aluminum alloy friction-stir-welded to stainless steel. Lee et al. (2006) examined in detail the IMCs of dissimilar FSW joints between stainless steel and A6056 aluminum alloy, using transmission electron microscopy (TEM). Sun et al. (2013) reported on the formation of an amorphous structure along the joint interface between A6061 aluminum alloy and mild steel. Movahedi et al. (2013) examined the effect of annealing treatments on the joint strength of dissimilar A5083 aluminum alloy and steel friction-stir lap welds. Tanaka et al. (2009) reported on the relationships between welding temperature, IMC thickness, and bond strength of friction-stir welds between various aluminum alloys and mild steel. In addition, Tanaka et al. (2011) investigated the initiation and growth mechanisms of IMCs during FSW. As described, most studies concerning dissimilar friction-stir welds to date focused on their bond strengths and microstructures. Morita et al. (2009) reported on the deep drawability and bendability of friction-stir-welded TWBs of a commercial A6061 aluminum alloy and one variety of mild steel. However, the effects of combining different base materials on the formability of TWBs remain ambiguous. A thorough understanding of the formability of dissimilar friction-stir-welded TWBs is required to adopt them for industrial applications. This study was undertaken to clarify the formability of several different types of friction-stir-welded TWBs, consisting of mild steel and three different aluminum alloys. In addition, the material flow behaviors and severity of forming during deep drawing were investigated by both numerical analysis and experimental measurements. Finally, an effective improvement method of deep drawability was proposed to facilitate the use of dissimilar TWBs in the fabrication of wide varieties of complex parts.
2. Experimental procedures 2.1. Base materials and welding conditions Zinc-coated steel sheets and sheets of the three A1100-O, A5052-O, and A5182-O aluminum alloys were used in this study. The sheet thickness for all materials was 1 mm. A FSW tool with a screw thread probe made of SKD61 tool steel was used with specifications as follows: 10 mm shoulder diameter; 3 mm probe diameter; 0.9 mm probe length. The inclination angle during FSW was 2◦ . FSW was performed at a tool rotation speed of 1500 rpm and a weld travel speed of 300 mm/min; these welding conditions were determined from preliminary examinations. Similar welded blanks (A5052-A5052) and four dissimilar TWBs (A5182-A5052, steel-A1100, steel-A5052, and steel-A5182) were made by FSW. When the same types of materials were treated by FSW (A5052A5052 and A5182-A5052), the center of the moving tool conformed to the butt line, as shown in Fig. 1(a). Fukumoto et al. (2004) and Kimapong and Watanabe (2004) described the plunging of the rotation tool into the aluminum, successfully producing dissimilar welds by placing the aluminum alloy on the retreating side. The advancing side of welds is the side on which the rotation of the tool proceeds in the same direction as the motion of the tool itself; the opposite side is referred to as the retreating side. Therefore, the probe was placed directly into the aluminum alloy, fixed on the retreating side, for dissimilar FSW, as shown in Fig. 1(b). In
Fig. 1. Schematic of friction-stir welding: (a) similar welding and (b) dissimilar welding.
this study, the probe was offset from the steel’s faying surface by approximately 0.2 mm. 2.2. Tensile and deep drawing tests The tensile properties of the base materials were measured at angles of 0◦ , 45◦ , and 90◦ to the rolling direction (RD). The dimensions of the tensile specimens are shown in Fig. 2(a). The tensile specimen was cut using an electrical discharge machine; the yield stress, tensile stress, elongation, work-hardening exponent (n-value), and Lankford value (r-value) were determined to mechanically characterize the materials. The n-value was calculated from the slope of the true stress–true strain curve between the true strains of 0.1 and 0.15. The crosshead speed in the tensile tests for all cases was 3 mm/min. Tensile specimens of the frictionstir welds were also machined transverse to the welding direction, as shown in Fig. 2(b) and (c), and the tensile strengths of the joints were measured using these samples. Cylindrical deep drawing tests were performed to measure the formability of the base materials and the dissimilar FSWs. The tool geometries for the deep drawing testing were as follows: 40 mm punch diameter; 4 mm punch shoulder radius; 42.5 mm die hole diameter; and 10 mm die shoulder radius. Conventional press oil (Model Number: S-3) manufactured by HOKOKU OIL Co., Ltd. was used as the lubricant, with a kinetic viscosity at 313 K of approximately 85 mm2 /s. The punch speed and the blank holder force were 60 mm/min and 3 kN, respectively. The geometry of the initial blank was circular with the diameter ranging from 74 to 90 mm to reveal the limiting drawing ratio (LDR), calculated as: LDR =
Dmax dp
where Dmax is the maximum blank diameter drawn without fracture and dp is the punch diameter. In the TWBs’ deep drawing tests, the weld line was located at the center of the blank. 2.3. Metallurgical analysis of material flow during deep drawing The material flow behaviors during the deep drawing of both the base metals and the TWBs were investigated using electron backscatter diffraction (EBSD) analysis and the grid-marking method. For EBSD analysis, the deep-drawn cups of the A5052 base material and the steel-A5052 TWB were sectioned transverse to the rolling direction and to the weld line, respectively. A section of the flat bottom of each section was polished by an Ar ion beam. The EBSD analysis was performed by a field emission scanning elec-
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
117
Fig. 2. Dimensions of tensile specimens: (a) base material, (b) steel-aluminum weld, and (c) aluminum–aluminum weld.
tron microscope (FE–SEM) with an orientation imaging microscopy system produced by TexSem Laboratories, Inc. An analysis of the material flow by the grid-marking method was performed in the cup bottom itself during deep drawing. 2.4. Numerical simulation conditions Numerical simulation was conducted to examine the material flow behavior during the deep drawing of the TWBs and to predict appropriate forming conditions. The simulation of the deep drawing of the base materials and the TWBs was performed by a commercially available finite element code, LS-DYNA. The tools were modeled as rigid shell elements, while the blanks were modeled with four-noded shell elements with shell thicknesses of 1 mm. To reduce the computation time of this simulation, the geometry of the blanks was scaled to one-fourth for the base material and to one-half for the welds. The stress–strain relations of the base materials in the simulations, accounting for the strain hardening of the materials during plastic deformation, was described using their experimental stress–strain curves in the rolling direction. Hill (1948) proposed the quadratic yield criterion, a commonly used yield function in numerical simulations for metal forming. Barlat et al. (2003) suggested the orthotropic yield function. In this
Fig. 3. Top surface views at the weld area: (a) steel-A1100 weld, (b) steel-A5052 weld, (c) steel-A5052 weld, (d) A5182-A5052 weld, and (e) A5052-A5052 weld.
study, Barlat’s yield function was applied for the aluminum alloy because it successfully described the plastic behavior of materials with low r-values, like the aluminum alloys examined here. Meanwhile, the yield behavior of steel was described using Hill’s yield criterion. The friction at the tool-blank interface was defined as follows. The experimental punch load curves of all base materials during the deep drawing tests were measured. By numerical simulation, the punch load curves at several friction coefficients were estimated. The friction coefficient at which the simulated punch load curve matched well with the experimental curve was utilized as the interfacial friction in this study. The friction coefficients for the aluminum alloys and the steel were defined as 0.17 and 0.12, respectively.
118
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 4. Typical cross-sectional microstructure of the interface region of friction-stir-welded zinc-coated steel and A5052 aluminum alloy.
including the effect of alloy compositions on IMC formation will be needed in the future. 3.3. Deep drawabilities of dissimilar TWBs
Fig. 5. Higher-magnification view of the microstructure at the interface of the steelA5052 weld.
3. Results 3.1. Appearance and microstructure of dissimilar TWB produced by FSW The top surface appearance of the welds is shown in Fig. 3. All welds are successfully joined without porosity or defects. As shown in Fig. 3(a)–(c), the top surface of the dissimilar steel/aluminum welds is rough, with some flashes observed in the steel-A5052 and steel-A5182 weld. In addition, the bead width in the steel-A5052 weld is larger than that in the steel-A5182 weld, since the weld tool was inserted more deeply, because of the lower yield strength of A5052. Therefore, larger flashes occur in the steel-A5052 weld than in the steel-A5182 weld. The top surfaces of the A5182-A5052 and A5052-A5052 welds are smooth, as can be seen in Fig. 3(d) and (e). Fig. 4 shows a typical cross-section of the steel-A5052 weld; no void or cracks are observed at the joint interface or the weld zone. The highly magnified SEM image at the interface between steel and aluminum alloy is also shown in Fig. 5. The IMC thickness at the interface is less than 300 nm, as shown in Fig. 5. 3.2. Tensile properties of base materials and FSWs The tensile properties of the base materials and of the welds are shown in Tables 1 and 2, respectively. “Joint efficiency” is defined as the ratio of the strength obtained for the weld divided by the lower strength between those of the two base materials. The joint efficiency of the welds is approximately 100% in all cases, excepting the steel-A5182 weld, which exhibits a drop of about 15% in its tensile strength. Elliott and Wallach (1981a,b) reported that the Mg content of A5182 facilitated the formation of IMCs, resulting in the loss of joint strength. In addition, Yamamoto et al. (2005) showed that the oxide layer formed by the increased Mg content reduced the joint strength. Therefore, the lower joint efficiency of the steelA5182 weld in this study can be attributed to microstructural changes induced by A5182’s large Mg content. Further research
3.3.1. LDR of welded blanks and the effect of the mismatch in material strength on the LDR values Table 2 shows the deep drawabilities of the base materials and welds. “LDR efficiency” is defined as the ratio of the LDR for the weld divided by the lower LDR between those of the two base materials. From Table 2, dissimilar TWBs, including brittle IMCs, may be deepdrawn with LDRs as low as 1.7. However, the deep drawabilities of the dissimilar TWBs show a different trend from that of the tensile strength of the TWBs. The LDR values of all TWBs are consistently lower than those of their component base materials, despite the high joint efficiency in terms of tensile strength. A particularly significant decrease in the LDR is observed for the steel-A1100 weld. Panda et al. (2007) reported on a limiting dome height in steelonly TWBs between sheets of different thickness existing midway between the two sheets. This implies that steel–aluminum TWBs have different formability from steel-only TWBs. Fig. 6 shows the appearance of five weld cups drawn at their LDR. For dissimilar welds, the projection of the weaker materials is observed near the weld line in the form of flanges, indicated by circles in Fig. 6. The weld lines appear to have moved toward the stronger materials. In particular, a marginal shift in the weld line toward the steel side is observed for the steel-A1100 weld, as shown in Fig. 6(a). However, for the similar weld of two 5052 aluminum alloy sheets, shown in Fig. 6(e), the weld line at the flat bottom of the cup remains straight, and exterior anomalies in the drawing cup, such as flange projections and weld-line position shifts, are not observed. These are considered to result from the mismatches in material strengths and the differences in plastic stain ratios and material flow characteristics between the steel and aluminum alloys. 3.3.2. Failure of dissimilar TWBs The appearances of the base materials and two welds after their deformation to failure are shown in Fig. 7. As observed in Fig. 7(b) and (c), fractures in the A5052-A5052 weld and the A5052 base material cups occur at the punch radius, in accordance with the general trends observed in deep-drawing tests on standard nonwelded cups. For dissimilar TWBs, however, fractures are observed to occur parallel to the weld line on the side of the weaker material, at a short distance from the weld line itself. This demonstrates that the deformation behavior of dissimilar TWBs during deep-drawing differs from those of single materials and similar welds. 3.3.3. Material flow behavior during deep drawing Fig. 8 shows inverse pole figures (IPF) obtained by EBSD analysis for the single A5052 base material, its drawn cup, and the steelA5052 weld drawn cup. Analysis of the drawn cups was performed on sections of the cups’ flat bottoms, approximately 6 mm from the cups’ centers. The upper and lower rows of Fig. 8 show the IPFs
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
119
Table 1 Mechanical properties of base materials. Materials
Angle to RD
Yield strength (MPa)
Tensile strength (MPa)
Elongation (%)
n-value
r-value
r¯
r
A1100O
0 45 90
31.8 31.3 32.0
102.2 97.7 99.7
33.8 38.4 40.3
0.318 0.318 0.328
0.589 0.955 0.676
0.794
-0.323
A5052O
0 45 90
91.9 93.1 91.3
206.5 202.4 199.1
28.7 27.6 27.1
0.327 0.338 0.328
0.908 0.522 0.708
0.665
0.286
A5182O
0 45 90
136.9 135.3 138.2
283.1 287.3 288.9
26.9 30.3 28.0
0.394 0.387 0.386
0.636 0.837 0.754
0.766
-0.142
Zinccoated steel
0 45 90
293.0 304.4 302.7
355.2 366.8 356.5
41.2 38.2 43.0
0.260 0.246 0.253
1.069 0.799 1.539
1.052
0.505
Table 2 Tensile properties and deep drawability of friction-stir welds. Mater.1
Mater.2
Joint strength (MPa)
Joint efficiency (MPa)
Mater.1 LDR
Mater.2 LDR
TWB LDR
LDR efficiency (%)
Zinc-coated steel Zinc-coated steel Zinc-coated steel A5182-O A5052-O
A1100-O A5052-O A5182-O A5052-O A5052-O
99.5 198.5 244.9 205.7 199.5
100.2 97.9 85.4 101.5 98.4
2.175 2.175 2.175 2.150 2.100
2.075 2.100 2.150 2.100 2.100
1.700 1.850 1.925 2.050 2.125
81.9 88.1 89.5 97.6 101.2
Fig. 6. Appearance of TWBs deformed in the limiting drawing ratio (LDR) tests: (a) steel-A1100 weld, (b) steel-A5052 weld, (c) steel-A5182 weld, (d) A5182-A5052 weld, and (e) A5052-A5052 weld.
Fig. 7. Appearance of drawing cups deformed to failure: (a) steel-A1100 weld, (b) A5052-A5052 weld, and (c) A5052-O base material.
parallel and perpendicular, respectively, to the rolling direction for the single A5052 base material and its drawn cup and the weld line for the dissimilar TWB drawn cup. As shown in Fig. 8(a) and (b), the dominant crystal orientations parallel to the rolling direction in the single 5052 base material are <0 0 1> and <1 1 2>, while that perpen-
dicular to the rolling direction is <0 0 1>. As observed in Fig. 8(c) and (d), the dominant crystal orientation in the single A5052 drawn cup is nearly identical to that in the single A5052 base material shown in Fig. 8(a) and (b). This suggests that the material of the cup bottom is subjected to equi-biaxial tensile stresses and undergoes very little
120
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 8. Inverse pole figures of A5052 base metal, drawn A5052 cup, and drawn steel-A5052 weld cup.
Fig. 9. Strain distributions at the flat bottom of the drawn (a) A5052-O base material and (b) steel-A5052 TWB.
deformation during the deep-drawing test. Meanwhile, in Fig. 8(a), (e), and (g), the dominant crystal orientation parallel to the weld in the TWB drawn cup line is almost equivalent to that in the single A5052 base material. However, a new <1 1 1> texture peak perpendicular to the weld line is observed, as shown in Fig. 8 (f) and (h), where the <1 1 1> texture is one of the preferred orientations to the tensile axis of aluminum. This implies that the aluminum in the dissimilar TWB does not deform parallel to the weld line during deep drawing, but instead deforms perpendicular to the weld line near the fracture area; this corresponds to the plane strain state. The strain distribution at the flat bottoms of the cups of the single A5052 base material and the steel-A5052 weld was analyzed using the displacement of the 2 mm-interval lattice lines on the sample surface. Fig. 9(a) represents the strain both transverse and longitudinal to the rolling direction of the single A5052 base material cup, while Fig. 9(b) represents the strains transverse and longitudinal to the weld direction of the dissimilar TWBs. In addition, Fig. 10 presents forming limit diagrams (FLDs) for the flat bottoms of the two drawing cups, as calculated from the transverse and longitudinal strains. For the single A5052 base material cup shown in Figs. 9 (a) and 10, the strain mode is seen to be equibiaxial in tension, with strains below 5%. This matches reasonably well with the results shown in Fig. 8. As shown in Fig. 9(b), although both materials in the TWB drawn cup experienced very little longitudinal strain relative to the weld line, a significant transverse strain above 20% is observed at a small distance from the joint interface in the A5052 aluminum alloy. This implies that the significant
Fig. 10. Forming limit diagrams of the flat bottom of the drawing cups and of the A5052-O aluminum alloy.
transverse strain observed on the aluminum side results from the plane strain state. The forming limit in the plane strain state is well known to be substantially lower than those related to other strain states. Kohara et al. (1985) studied the FLD curves for the A5052O aluminum alloy, reporting that the major strain of failure in the plane strain state is approximately 20%. The FLD curve of A5052-O aluminum alloy is also included in Fig. 10. As observed in Fig. 10, many strain plots of the steel-A5052 TWB intersect with the FLD curve of the A5052-O aluminum alloy. This implies that the flat
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 11. Micro-hardness profile along the centerline of the cross-section of the steelA5052 weld.
bottom of the TWB cup is subjected to significant strain, near the forming limit, which results in the premature failure of drawn cups from dissimilar TWBs. Consequently, local deformation in the flat bottom on the aluminum side must be avoided in order to improve the drawability of dissimilar TWBs of steel and aluminum alloys. The simulation analysis and discussion henceforth were focused on the dissimilar TWBs of steel and aluminum alloys, in the interest of contributing to the improvement of the deep drawability of dissimilar TWBs. 3.4. Numerical simulation For a precise numerical analysis, the mechanical properties of the weld zone must be properly measured and incorporated into the simulation. For example, Padmanabhan et al. (2008) described the material behavior of the weld zone using a rule based on the mixture of aluminum and steel. Lee et al. (2009) quoted the experimental results reported by Saunders and Wagoner (1996), in which the mechanical properties of the weld zone were experimentally measured using small-sized tensile samples. In addition, Leitão et al. (2011) reported that the formability of dissimilar A5182A6016 TWBs was strongly influenced by the weld bead shape. In this study, micro-hardness tests were performed, as the mechanical properties of the weld zone can be difficult to measure because of the zone’s narrowness, as shown in Fig. 4. As observed in Fig. 11, the hardness of the weld zone only slightly exceeds that of the base aluminum alloy; it is approximately half that of the steel. Therefore, the change of mechanical properties in the weld zone is considered to have little influence on the formability of the TWBs. In fact, Ahmetoglu et al. (1995) and Panda and Kumar (2008) reported that the deformation behavior of TWBs could be well-simulated without incorporating the material properties of the weld zone; in these simulations, the yield strength ratio of the TWBs were 2.25 and 1.54, respectively. With the experimental verification of these prior conclusions, the change in the mechanical properties of the weld zone was neglected in this study. Top views of the drawing cups, as obtained from both experiments and simulations, are shown in Fig. 12. For the steel-A1100 weld in Fig. 12(a), a blank with a 68 mm diameter was drawn to a punch height of only 15 mm because of its extremely low formability. For the steel-A5052 and steel-A5182 welds in Fig. 12(b) and (c), blanks with diameters of 74 mm were drawn to a punch height of 20 mm. As observed in Fig. 12, the shapes predicted by the simulations mostly agree with the experimental results. However, slight differences of flange length between the experiments and sim-
121
ulations are observed. These are attributed to the non-uniform blank holder force. As shown in Fig. 3(a)–(c), rough weld beads and some flashes were observed in the steel–aluminum welds. The deep-drawing tests in this study were performed without removing flashes on the blank surface, because of the difficulty of machining thin sheets distorted by FSW. Therefore, the rough weld bead and flashes are considered to contribute to non-uniform blank holder force, resulting in the mismatch in flange length between the experiment and simulation. However, the deformation behavior at the flat bottom was studied more carefully, because this behavior is more closely linked with the premature failure of dissimilar TWBs. More precise numerical analysis of the flange behavior will be performed in the future. The major strain distributions of TWBs as simulated by LS-DYNA are shown in Fig. 13. The aluminum alloy side close to the weld line at the bottom of the cup is subjected to a larger strain than the steel side. A similar trend is observed in the experimental results, as shown in Fig. 9. This similarity verifies the utility of the numerical analysis using LS-DYNA in understanding the material flow behavior during the forming of TWBs. 4. Discussion The TWBs with high strength ratios in their base materials, such as the steel-A1100 TWB, failed at local deformation areas in the lower-strength material as a result of the materials’ strength mismatch. In addition, the plane strain state of the local deformation area, attributed to constraint by plastic deformation resulting from the weld line, increases the likelihood of premature fractures. Therefore, inhibiting this local deformation may improve the formability of TWBs. Many countermeasures restricting local deformation have been already proposed. Leitão et al. (2009) and Padmanabhan et al. (2008) reported that the adjustment of the location of the blank holder force could mitigate the localization of deformations observed in dissimilar TWBs. In addition, Panda and Kumar (2008) demonstrated that the limiting dome height of TWBs increased with the application of a counter pressure, while Heo et al. (2001a,b) reported that the movement of the weld line could be controlled by adequate draw-bead installation. In this study, the coefficient of friction between blanks and forming tools was of greatest interest; the effects of the friction conditions on the curvature of the weld line and on the local deformation during the deep drawing of TWBs were examined most closely. 4.1. Simulation analysis concerning the effect of friction conditions on the drawability Numerical simulation of deep drawing was performed in order to clarify the effect of friction and to suggest methods of inhibiting local deformation in the steel-A1100 weld, where the most noticeable shift in the weld line toward the steel side and the highest level of strain on the A1100 side was observed. Only the friction coefficient between the punch and the blank was varied in this simulation. The effect of the friction coefficient on the drawability was investigated in three different cases: variation of friction on only the A1100 side, on only the steel side, and over the whole area. The friction coefficient was set to 0.3, 0.45, and 0.6. Figs. 14–16 show the simulated major strain distributions of the steel-A1100 weld when the friction coefficient is varied on only the A1100 side, on only the steel side, and over the whole area, respectively. For comparison, the results shown in Fig. 13(a), i.e., those found using press oil as the lubricant, are included on the far left in Figs. 14–16. As shown in Fig. 14, the region of high strain on the A1100 side becomes wider and more prominent with the increase of the friction coefficient of the aluminum alloy. When the friction
122
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
Fig. 12. Top views of the drawing cups obtained by experiment versus numerical simulation: (a) steel-A1100 weld, (b) steel-A5052 weld, and (c) steel-A5182 weld.
Fig. 13. Simulated major strain distributions of the TWBs for the (a) steel-A1100 weld, (b) steel-A5052 weld, and (c) steel-A5182 weld.
Fig. 14. Simulated major strain distributions of the steel-A1100 weld for friction coefficients on the A1100 side of (a) 0.17, (b) 0.30, (c) 0.45, and (d) 0.60.
Fig. 15. Simulated major strain distributions of the steel-A1100 weld for friction coefficients on the zinc-coated steel side of (a) 0.12, (b) 0.30, (c) 0.45, and (d) 0.60.
coefficient is enhanced on only the steel side and over the whole area, however, greater friction coefficients correlate to greater linearity of weld lines. This in itself corresponds to the more isotropic deformation behavior of the cup bottom during deep-drawing, as shown in Figs. 15 and 16. These results suggest that the enhancement of the friction coefficient on only the steel side or over the
whole blank area may improve the deep drawability of dissimilar TWBs. During the deep-drawing of dissimilar steel/aluminum TWBs, the stress required to draw into the cup section is lower on the aluminum alloy side than on the steel side; this stress is transferred from the flange to the flat bottom. Because of the higher required stress on the steel side, the aluminum alloy material on
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
123
Fig. 16. Simulated major strain distributions of the steel-A1100 weld for friction coefficients on the steel and A1100 sides of (a) 0.12 and 0.17, (b) 0.30 and 0.30, (c) 0.45 and 0.45, and (d) 0.60 and 0.60, respectively.
the flat bottom, punch radius, and cup wall is pulled toward the steel side of the TWB. The friction coefficient between the punch and the blank mainly affects the material flow at the punch radius. When the friction coefficient is enhanced only on the aluminum side, the material flow on the aluminum side toward the steel side is impeded at the punch radius by the high frictional force. Therefore, only the aluminum material on the flat bottom is pulled out toward the steel side, resulting in an expanded area of local deformation on the A1100 side, as shown in Fig. 14. When the friction coefficient is enhanced only on the steel side, the material flow is notably inhibited at the punch radius on the steel side. The stress on the steel side necessary to draw the material at the flange is not transferred to the flat bottom; therefore, the material flow toward the steel side on the A1100 side is inhibited, as shown in Fig. 15. When a high friction coefficient is in effect over the whole area, the material flow at the punch radius is inhibited on both sides of the TWB. Thus, the same trend observed in Fig. 15 is seen in Fig. 16. In Figs. 15 and 16, when the friction coefficient changes from 0.30 to 0.60, the strain distribution does not experience significant change. Consequently, local deformation can be restricted effectively by even slight enhancement of the friction coefficient to 0.3.
Fig. 17. Suitability of deep-drawing for a steel-A5052 weld under each friction condition.
4.2. Verification tests concerning the effect of friction conditions on the drawability Deep drawing tests were performed to determine the validity of the numerical predictions mentioned above. In the verification tests, different friction conditions were imposed by using rosin as a lubricant. The friction coefficients with the use of rosin were 0.25 for steel and 0.24 for the aluminum alloy, respectively, estimated by conforming the experimental punch load curve to the simulated curve as explained in Section 2.4. Rosin was applied only between the punch and blank, while press oil was used on the other contact surfaces, as in the numerical analysis. Fig. 17 shows the deep drawability of a steel-A5052 weld and the base materials under each friction condition. The vertical axis shows the initial blank diameter before the deep-drawing test. The double circle in this figure indicates the completion of the deep drawing test without failure; therefore, the maximum blank diameter with double circles corresponds to the blank diameter of the LDR value. The cross symbol indicates the failure of the blank during the deep-drawing test. As observed in Fig. 17, the enhancement of the friction coefficient by using rosin significantly improves the deep drawability of all blanks. The LDR of the TWB using rosin was approximately 19% higher than that of the TWB using press oil. In addition, although the LDR of the TWB using press oil was 88% of that of the weaker base material, the LDR using the rosin was nearly equal to that of the weaker material. The steel-A5052 cups drawn under the different friction conditions are shown in Fig. 18. As shown in Fig. 18, the cup height of the TWB drawn using rosin is significantly
Fig. 18. Steel-A5052 cups drawn under different friction conditions.
greater than that using the press oil; its weld line remains straight. These results validate the prediction of the numerical simulation described above. In summary, the enhancement of the friction coefficient between the punch and blank was revealed to be capable of improving the deep drawability of dissimilar steel/aluminum TWBs. 5. Conclusions In order to examine the formability performance of various dissimilar steel/Al TWBs produced by FSW, experiments and numerical simulations on deep-drawing were performed. In addition, possible methods for improving the formability of the TWBs
124
T. Tanaka et al. / Journal of Materials Processing Technology 226 (2015) 115–124
were explored by both numerical simulation and experimentation. The following conclusions were obtained. • The tensile strengths of the TWBs were nearly equivalent to those of the base materials. Although the TWBs had deep drawability with LDRs as low as 1.7, the LDR values were lower than those of the base materials; this tendency opposed that observed for steel-only TWBs. • After deep-drawing tests, the weld line of the TWBs was observed to shift toward the stronger side, and fracturing was observed near the weld line on the flat bottom of the aluminum alloy side. The degree of the weld line shift increased with the increasing difference in the strengths of the two base materials, resulting in the deterioration of the TWBs’ deep drawability. • Each type of analysis regarding material flow behavior, including inverse pole figures from EBSD measurements, the grid marking method, and numerical simulations, indicated that the aluminum alloy near the fractured area in the TWBs was locally deformed in the plane strain state. Major strain exceeding 20%, the forming limit of aluminum alloy in the plane strain state, was observed near the fracture area even prior to failure. • The effect of friction between the forming tools and the blanks on the local deformation of the aluminum alloy side was investigated by both numerical simulation and experimental tests. The simulation analysis indicated that the curvature of the weld line and the local deformation in the aluminum alloy would decrease with increasing the friction coefficient between the punch and blank. • Actual deep-drawing tests using rosin to increase the friction coefficient resulted in more uniform material flow on the flat bottom, and the LDR of the steel-A5052 TWB increased by nearly 19% relative to that drawn using press oil. Therefore, it could be concluded that enhancing the friction coefficient may be an effective solution to improve the formability of dissimilar TWBs. Acknowledgement The authors gratefully acknowledge financial support from the Amada Foundation for Metal Work Technology. References Ahmetoglu, M.A., Brouwers, D., Shulkin, L., Taupin, L., Kinzel, G.L., Altan, T., 1995. Deep drawing of round cups from tailor-welded blanks. J. Mater. Process. Technol. 53, 684–694. Barlat, F., Brem, J.C., Yoon, J.W., Chung, K., Dick, R.E., Choi, S.H., Pourboghrat, F., Chu, E., Lege, D.J., 2003. Plane stress yield function for aluminum alloy sheets – part I: theory. Int. J. Plast. 19, 1297–1319. Choi, Y., Heo, Y., Kim, H.Y., Seo, D., 2000. Investigations of weld-line movements for the deep drawing process of tailor welded blanks. J. Mater. Process. Technol. 108, 1–7. Elliott, S., Wallach, E.R., 1981a. Joining aluminum to steel Part 1-Diffusion bonding. Metal Constrution 13, 167–171. Elliott, S., Wallach, E.R., 1981b. Joining aluminum to steel Part 2-Friction bonding. Metal Construction 13, 221–225. Fukumoto, M., Tsubaki, M., Shimada, Y., Yasui, T., 2004. Welding between ADC12 and SS400 by means of friction stirring. Q. J. Jpn. Weld. Soc. 22, 309–314.
Heo, Y., Choi, Y., Kim, H.Y., Seo, D., 2001a. Characteristics of weld line movements for the deep drawing with drawbeads of tailor-welded blanks. J. Mater. Process. Technol. 111, 164–169. Heo, Y.M., Wang, S.H., Kim, H.Y., Seo, D.G., 2001b. The effect of the drawbead dimensions on the weld-line movements in the deep drawing of tailor-welded blanks. J. Mater. Process. Technol. 113, 686–691. Hill, R., 1948. A theory of the yielding and plastic flow of anisotropic metals. Proc. R. Soc. A193, 281–297. Ikeuchi, K., Yamamoto, N., Takahashi, M., Aritoshi, M., 2005. Effect of interfacial reaction layer on bond strength of friction bonded joint of Al alloys to steel. Trans. JWRI 34, 1–10. Kimapong, K., Watanabe, T., 2004. Friction stir welding of aluminum alloy to steel. Weld. J. 83, 277–282. Kohara, S., Mototsugu, K., Aoki, K., 1985. Forming limit curves of 5052 alloy sheets. J. Jpn. Inst. Light Metal 35, 700–704. Lee, W.B., Schmuecker, M., Mercardo, U.A., Biallas, G., Jung, S.B., 2006. Interfacial reaction in steel-aluminum joints made by friction stir welding. Scripta Mater. 55, 355–358. Lee, W., Chung, K.H., Kim, D., Kim, J., Kim, C., Okamoto, K., Wagoner, R.H., Chung, K., 2009. Experimental and numerical study on formability of friction stir welded TWB sheets based on hemispherical dome stretch tests. Int. J. Plast. 25, 1626–1654. Leitão, C., Emílio, B., Chaparro, B.M., Rodrigues, D.M., 2009. Formability of similar and dissimilar friction stir welded AA 5182-H111 and AA 6016-T4 tailored blanks. Mater. Des. 30, 3235–3242. Leitão, C., Zhang, B.K., Padmanabhan, R., Rodrigues, D.M., 2011. Influence of weld geometry and mismatch on formability of aluminium tailor welded blanks: numerical and experimental analysis. Sci. Technol. Weld. Join. 16, 662–668. Merklein, M., Johannes, M., Lechner, M., Kuppert, A., 2014. A review on tailored blanks-Production, applications and evaluation. J. Mater. Process. Technol. 214, 151–164. Morita, T., Sakamoto, A., Mabuchi, S., Iizuka, T., 2009. Strength and formability of friction stir welded aluminum alloy with cold rolled steel. J. Soc. Mater. Sci. Jpn. 58, 317–322. Movahedi, M., Kokabi, A.H., Seyed Reihani, S.M., Cheng, W.J., Wang, C.J., 2013. Effect of annealing treatment on joint strength of aluminum/steel friction stir lap weld. Mater. Des. 44, 487–492. Padmanabhan, R., Oliveira, M.C., Menezes, L.F., 2008. Deep drawing of aluminum-steel tailor-welded blanks. Mater. Des. 29, 154–160. Panda, S.K., Kumar, D.R., Kumar, H., Nath, A.K., 2007. Characterization of tensile properties of tailor welded IF steel sheets and their formability in stretch forming. J. Mater. Process. Technol. 183, 321–332. Panda, S.K., Kumar, D.R., 2008. Improvement in formability of tailor welded blanks by application of counter pressure in biaxial stretch forming. J. Mater. Process. Technol. 204, 70–79. Saunders, F.I., Wagoner, R.H., 1996. Forming of tailor-welded blanks. Metall. Mater. Trans. 27A, 2605–2615. Schubert, E., Klassen, M., Zerner, I., Wlaz, C., Sepold, G., 2001. Light-weight structures produced by laser beam joining for future applications in automobile and aerospace industry. J. Mater. Process. Technol. 115, 2–8. Sun, Y.F., Fujii, H., Takaki, N., Okitsu, Y., 2013. Microstructure and mechanical properties of dissimilar Al alloy/steel joints prepared by a flat spot friction stir welding technique. Mater. Des. 47, 350–357. Tanaka, T., Morishige, T., Hirata, T., 2009. Comprehensive analysis of joint strength for dissimilar friction stir welds of mild steel to aluminum alloys. Scripta Mater. 61, 756–759. Tanaka, T., Morishige, T., Hirata, T., 2011. Formation mechanism of intermetallic compound at the weld interface of dissimilar friction stir welded joint of pure aluminum and mild steel. Q. J. Jpn. Weld. Soc. 29, 101–106. Thomas M.W., Nicholas E.D., Needham J.C., Murch M.G., Templesmith P., Dawes C.J., 1991. GB Patent Applications No. 9,125,978.8, December 1991. Uzun, H., Donne, C.D., Argagnotto, A., Ghidini, T., Gambaro, C., 2005. Friction stir welding dissimilar Al 6013-T4 to X5CrNi18-10 stainless steel. Mater. Des. 26, 41–46. Xu, F., Sun, G., Li, G., Li, Q., 2014. Experimental investigation on high strength steel (HSS) tailor-welded blanks (TWBs). J. Mater. Process. Technol. 214, 925–935. Yamamoto, N., Takahashi, M., Aritoshi, M., Ikeuchi, K., 2005. Effect of interfacial layer on bond strength of friction-welded interface of Al–Mg 5083 alloy to mild steel. Q. J. Jpn. Weld. Soc. 23, 496–503.