- Email: [email protected]
On the influence of tool path in friction stir spot welding of aluminum alloys
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
journal homepage: www.elsevier.com/locate/jmatprotec
On the influence of tool path in friction stir spot welding of aluminum alloys G. Buffa, L. Fratini ∗ , M. Piacentini Dipartimento di Tecnologia Meccanica, Produzione e Ingegneria Gestionale, Universita` di Palermo, Viale delle Scienze, 90128 Palermo, Italy
a r t i c l e
i n f o
a b s t r a c t
Article history:
Friction stir spot welding (FSSW) has been proposed as an effective technology to spot weld
Received 3 September 2007
the so-called “difficult to be welded” metal alloys. In the paper, a variation of the FSSW
Received in revised form
process has been considered. A tool path is given after the sinking phase nearby the ini-
2 January 2008
tial penetration site; in this way a larger welding spot is obtained and more material is
Accepted 4 January 2008
involved in the bonding process. The process mechanics of such modified FSSW process is highlighted and the joint strength undergoing tensile tests is considered at the varying both of the assigned tool path and of a few process parameters. Macro- and micro-analyses
Keywords:
are made in order to analyze the local material microstructure evolution. It is found that
Welding
improved performances, with respect to the “traditional” FSSW process, are obtained for all
Spot
the considered case studies.
Microstructure
1.
Introduction
Spot welding can be considered a very common joining technique in automotive and generally in transportation industries. Apart from the utilized joining technology and in particular from the heat source (Joule effect, electric arc and so on) it permits to obtain effective lap joints with short process times. What is more it is easily carried out through robots and automated systems as shown in Chae et al. (2002), Hemmingson (1996) and Darwish and Al-Dekhial (1999). It should be observed that in the past decades mechanical fastening techniques, such as clinching and riveting, also have been widely utilized, since they permit to obtain the same advantages of the former welding processes and they do not suffer from the insurgence of typical welding defects due to the fusion of the base material. For instance Larsson (1994), Lennon et al. (1999) and Riches et al. (1995) highlighted the benefits given by mechanical fastening techniques in several
∗
© 2008 Elsevier B.V. All rights reserved.
applications such as automotive assemblies and constructions. Friction stir welding (FSW) has been presented and proposed at the beginning of the nineties as an innovative technique able to join also the so-called un-weldable or difficult to be welded light weight alloys which are very common materials in automotive or aerospace industries. Actually classic welding processes, i.e. TIG or laser, determine quite weak joints due to a strong increase of the average grain size observed in the melted zone and to quite large thermally affected zones. Furthermore strong precautions have to be taken during classic welding processes in order to avoid inclusions and other typical defects in the joint core. In turn, FSW is a solid state welding process in which a specially designed rotating pin is first inserted into the adjoining edges of the sheets to be welded with a proper tilt angle and then moved all along the joint. Such tool produces frictional and plastic deformation heating in the welding zone; actually no melting of material is observed during FSW. Then, as the tool moves,
Corresponding author. Tel.: +39 091 6657051; fax: +39 091 6657039. E-mail addresses: [email protected] (G. Buffa), [email protected] (L. Fratini), [email protected] (M. Piacentini). 0924-0136/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2008.01.001
310
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
material is forced to flow around the tool in a quite complex flow pattern determining a very particular microstructure characterized by continuous dynamic recrystallization phenomena. Several authors have been investigating FSW in the past years; for the purposes of the present research interesting information related to the occurring metallurgy, to the material flow and to the mechanical performances of the joints can be found out in Liu et al. (2003), Rhodes et al. (1997) and in Shigematsu et al. (2003). As a natural evolution of the FSW process, friction stir spot welding (FSSW) has been recently proposed for the development of lap joints also by Pan et al. (2004) and by Addison and Robelou (2004). In such a process, directly based on the FSW process mechanics, a rotating tool with a probe pin is introduced in the two overlapped blanks to be jointed supported by a proper back-plate. The rotating tool generates friction heat in the specimens and at the same time a material flow is determined. The heated and softened material close to the tool plastically deforms and a bond is made between the bottom surface of the upper sheet and the top surface of the lower sheet. No linear movement is given to the tool which is retracted from the workpiece when the stirring process is completed. As far as the process parameters are regarded, both geometrical and technological aspects have to be considered: both the former and the latter affect the material flow during the process and the generated heat flux. In this way the tool shoulder and the pin shapes and geometries are very important since they influence the circumferential material flow. What is more, the tool sinking into the blanks, i.e. the applied normal load, determines, together with the chosen tool rotating speed, the friction forces work and, consequently, the generated heat flux. As far as technological parameters are considered, also the process times, namely the tool descending time, the process time length and finally the tool ascending time, play an important role in the FSSW process. In the paper, a variation of the FSSW process has been considered: after the sinking phase, the tool is fed along a prescribed path close to its initial print. In this way a wider spot weld is obtained and a larger material stirred zone is obtained; at the end of the process, as the tool is released from the welded joint, a hole reproducing the tool pin is left in the welded sheets right at the center of the path, just like the one observed in “traditional” FSSW processes. An improvement in the material flow during the process is reached, which
enhances the joint resistance. The differences, in terms of process mechanics and local material microstructure evolution between such process and the “traditional” FSSW are highlighted in the next sections through macro- and microanalyses. The maximum load in tensile tests was considered as performance index of the joints. The influence of the most relevant process parameters on the mechanical performances of the joints is considered.
2. “Traditional” and modified FSSW processes As briefly described before, FSSW is aimed to obtain lap joints utilizing a cylindrical tool with a pin tip centered on the bottom circular face; in the “traditional” process, the tool rotates on its axis and it is inserted into the specimens to be jointed with a normal force. In other words the tool is pushed down into the two overlapped blanks until a fixed level of reaction force is reached. A back-plate is utilized on the bottom side of the specimens to support the applied load and no linear translation is given to the tool, which is retracted after an assigned time length. As the rotating tool is inserted into the sheets (Fig. 1A) a local backward extrusion mechanics is observed and a full contact between the upper sheet and the tool shoulder is reached. Then, keeping the tool into the sheets, a heat flux is generated by the friction forces work and plastic deformation work decaying into heat (Fig. 1B). Finally, after a proper time length, the tool is moved up and the joint is released (Fig. 1C). The observation of the material flow and of its locally reached microstructure leads to the full understanding of the process mechanics. Considering a section of the joint (Fig. 2), the backward extrusion mechanics is highlighted for the two overlapped sheets. Due to the absence of tool feed rate no asymmetry in the metal flow, typical of the FSW processes, is observed. In particular, the interface surface of the lower specimen is deformed upwards and such material acts as a mechanical anchor between the two jointed sheets, as shown in the 60× magnification of Fig. 2. As in FSW processes, a detailed observation of the material microstructure in the joint section allows finding out a few different areas, as shown in the next Fig. 3, where a 250× image of the material zone on the lateral surface of the pin imprint is reported (zone 2 in Fig. 2).
Fig. 1 – The FSSW process.
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
311
Fig. 2 – The FSSW joint transverse section highlighting the anchoring effect.
In particular, as in FSW, moving from the periphery of the joint towards the welding line, i.e. towards the tool pin surface, first of all the parent material, in which no metallurgical modification is observed after the welding process, is found. Then a simple heat affected zone (HAZ, C) is reached, in which the material has undergone a thermal cycle which has modified the microstructure and the mechanical properties. Continuing towards the welding spot, a zone in which the material has been plastically deformed by the stirring action of the tool is encountered. It should be observed that in such zone the effects of the induced heat flux are also observed with an enlargement of the material average grain size. In this way, this zone is a thermo-mechanically affected one (TMAZ, B). Finally the so-called nugget is found (A). The latter is a recrystallized area in which the original grain and subgrain boundaries appear to be replaced with fine, equiaxed recrystallized grains characterized by a nominal dimension of few m as described by Jata and Semiatin (2000), Su et al. (2003) and Fratini and Buffa (2005). It should be observed that the peculiarities of the obtained microstructures strongly determine the joint effectiveness and its mechanical behavior. In the considered modified FSSW process, a tool path is given after the completion of the sinking phase (Fig. 1A); in other words, the tool is not left in its position but it is moved along a peculiar trajectory determining a definitively larger stirred zone in the material. Once completed the assigned tra-
Fig. 3 – The FSSW joint microstructure.
jectory, the tool leaves the sheets to be welded. In particular four different spiral movements have been chosen with the aim to enlarge both the thermo-mechanically affected area and the nugget (Fig. 4), and a few combinations of advancing speed and rotating speed have been investigated. It should be observed that all the considered tool trajectories are characterized by the same overall dimension as it will be clarified in the next section. What is more, macro- and micro-observations on the welded joints permitted to highlight the metallurgical effect of the used tool paths on the material, justifying the mechanical performances of the joints.
3.
The developed experiments
In order to carry out the FSSW tests a properly designed clamping fixture was utilized to fix the specimens to be welded on a CNC milling machine. The steel plates composing the fixture were finished at the grinding machine to assure a uniform pressure distribution on the fixed specimens. What is more proper sheet supports were utilized in order to obtain the lap joints. As far as the utilized tool is regarded, it was made in H13 steel quenched at 1020 ◦ C, characterized by a 52 HRc hardness. The shoulder was 15 mm in diameter and a 40◦ conical pin was adopted, with a major diameter of 7 mm and a minor diameter of 2.2 mm; the pin height was 2.6 mm. The above geometrical characteristics were chosen on the basis of the knowledge acquired on FSW and FSSW and on preliminary tests developed by the authors and presented in Buffa et al. (2007): such tests have been performed utilizing a circular tool path, and no quantitative evaluation of the stirred volume, the TMAZ area and the process time was performed; no further investigation was made at the varying geometry of the tool. The utilized base material was AA6082-T6 aluminum alloy, 1.5 mm in thickness. The material was characterized by an average yield stress of 280 MPa and an ultimate tensile stress (UTSb) of 319 MPa. The material showed a microhardness equal to 120Hv and an average grain size of about 80 m. The influence of a few process variables on the mechanical performance of the welded lap joints was investigated in this paper; in particular the tool rotation speed (R) and the advancing speed for the prescribed tool path (Vf ) were taken into account. Each of the above indicated variables was varied in a
312
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Fig. 4 – Sketch of the assigned tool axis paths with respect to the tensile tests applied force direction.
proper range assuming different levels and the effects on the joining process in terms of joint resistance and of local material microstructure were observed. The tool sinking into the specimens during the process (h = 2.85 mm) and the sinking speed (equal to 0.003 mm/rev) were kept constant for all the tests. What is more, as briefly described before, four different spiral movements were given to the tool utilizing a CAD-CAM software. In particular a circle path (C), a square path (Sq), an ellipse-x (Ex) and an ellipse-y (Ey) path, i.e. with the major axis respectively orthogonal and parallel to the applied load (F) direction in the tensile tests, were considered (see again Fig. 4). The circle diameter, the square edge and the ellipses major axis were chosen equal to 10 mm; in this way the same overall dimension was obtained for the final tool print left on the welded specimen at the end of the process. A constant pitch of the tool path equal to 1.8 mm was assumed for all the investigated case studies. Starting from the periphery of the given path, the rotating tool enters the sheets and, after the completion of the sinking phase, the tool axis starts moving along the prescribed path with different values of advancing velocity. In the following Table 1 the investigated values of the considered process variables are reported. In particular three different values of rotational speed (R), namely 300, 500 and 700 rpm, and three different values of advancing speed (Vf ), namely 0.2, 0.4 and 0.6 mm/rev were chosen: a total of nine different tests were
Table 1 – The utilized process plan R (rpm)
Advancing velocity (mm/rev) Vf = 0.2 (mm/min)
300 500 700
60 100 140
Vf = 0.4 (mm/min) 120 200 280
Vf = 0.6 (mm/min) 180 300 420
thus carried out both for the “traditional” FSSW process and the modified one, and each test was repeated three times. The investigated levels and ranges of each parameter were chosen on the basis of preliminary tests (Buffa et al., 2007) and literature data (Pan et al., 2004; Addison and Robelou, 2004). From each joint specimen for either tensile tests macroand micro-observations were made. Macro-observations were aimed to analyze the material area involved in the process mechanics and the possible presence of macrodefects; through the micro-observations the different material zones, i.e. the thermally affected zone, the thermo-mechanically affected ones, and the so-called nugget zone, were observed and the differences between the “traditional” and modified FSSW techniques were highlighted. In order to obtain such results the specimens were properly treated with Keller reagent and observed by a light microscope. As far as the tensile tests are regarded, they have been developed on specimens presenting an average width of 30 mm.
4.
Results and discussion
4.1.
Comparison between the tested tool paths
First of all a comparison between the chosen tool paths in terms of maximum load transferred by the welded joints was performed: Fig. 5 shows the average tensile tests results for an advancing velocity of 0.6 mm/rev. It should be observed that no comparison in terms of stress can be made for spot welding or spot mechanical fastening, namely clinching or riveting, since the joints are not characterized by a uniform transferring load section, as for instance in the welding processes of butt joints. For each configuration the best performance is obtained at the lowest considered value of the rotational speed R, and a decrease in the maximum load is observed at the increase of R, but a few differences should be highlighted on the basis of the obtained results: both the square and the two ellipse paths show a lower sensitivity to R with respect to the circle path, resulting in larger maximum load values obtained at higher
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Fig. 5 – Comparison of the maximum loads for the investigated tool paths at the varying of R (Vf = 0.6 mm/rev).
Fig. 6 – TMAZ area, tool path area and joint failure load at the varying of the tool path.
rotational speeds. In particular, with R = 300 rpm, the square path case study presents about the same maximum load value found for the circle path case study, i.e. the best overall performance between all the considered case studies. At the increase of R, the best joints are the ones welded utilizing the square path, while the circle path case study eventually becomes the less effective (R = 700 rpm). As far as the two elliptical paths are regarded, they show about the same behavior, with the maximum load for the ellipse-x case study slightly higher, due to the larger section in the direction orthogonal to the applied load, as better explained in the following. Fig. 6 shows a quantitative analysis of the stirred zone (namely the TMAZ and the HAZ areas) for all the investigated case studies. In particular the aforementioned stirred areas, measured in a transverse section orthogonal to the applied load direction, are reported at varying tool path for the operative parameters set that gave the best results, i.e. R = 300 rpm and Vf = 0.6 mm/rev (as highlighted in Fig. 5). Such areas have been correlated to the results of the tensile tests and to the areas of the external geometrical figure described by the tool axis in a top view of the joint. In other words the areas of the geometrical figures described by the tool during its path have been considered (for example, the square path, 10 mm in edge, has an area of 100 mm2 , see again Fig. 4). It has to be noticed that such an area is strictly correlated to the stirred volume because the area – and consequently the volume – inside such paths is also involved in the stirring of material by the tool pin,
313
Fig. 7 – Influence of the tool path on the measured TMAZ area in the considered transverse section.
due to the spiral movement and thus represents an important indicator of the quality of the bond. As far as the TMAZ area is regarded, the highest value is obtained for the Sq case study, while lower values are found for the C, Ex and Ey case studies, respectively. A first consideration can be made on the Ey case study: in such a section the smallest area is measured, with a sensible drop off with respect to the others; that was easily predictable remembering that the measured section (TMAZ) is orthogonal to the applied load direction (see again Fig. 4). On the other hand, the difference found out between the other case studies can be explained looking at Fig. 7. In the AA section, the width of the C, Sq and Ex profiles is the same, but the way the tool reaches such section is different for the three case studies (see black arrows in Fig. 7); in particular, for the Sq case study, the maximum width of 10 mm is reached from the beginning of the “vertical” edge, thus enhancing the material flow and resulting in a wider area involved in the material stirring, while for the C and Ex case studies such maximum value is reached only in the AA section. The area of the geometrical profile of each tool path, i.e. the area carrying the applied load in a horizontal section, is also reported and compared to the joint resistance. As already noticed, the Sq and C case studies show about the same failure load, but the larger stirred area found for the Sq case study results in the higher reliability showed at varying rotational speed R (see again Fig. 5). The good correlation found between the stirred areas, the “path area” and the joint resistance supports the assumption that enlarging the stirred zone directly improves the joint performances.
Fig. 8 – Modified process results at the varying of R and Vf (circle path case study).
314
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Fig. 9 – Filling defects for the circle path case study (R = 700 rpm, Vf = 0.4 mm/rev).
4.2.
Effect of the process parameters
The effect of the operative parameters on the joint performances is highlighted in the next Fig. 8 for the circle path case study: the average values of the maximum load transmitted by the joints are shown at varying tool feed rate (Vf ) and for different values of the tool rotating speed (R). It can be argued from the figure that the rotational speed (R) plays a significant role in the maximum load transmitted by the joints: the overall performances of the joints obtained with a smaller R value are in fact better than those obtained with larger R values. What is more, for each of the considered R values, the joint resistance decreases at the increase of the advancing velocity (Vf ), becoming the slope of such curves larger as R increases. As a consequence, the best performance, in the investigated ranges of the considered variables, is obtained with R = 300 rpm and Vf = 60 mm/min, and, utilizing such R value, the joint failure load is almost constant at the varying Vf . Similar trends were obtained for the other tool paths investigated. It can be thus stated that large rotational speed values result in poorer performances for all the four considered case studies, for which an unsatisfactory material flow is obtained: from the performed macro-observations, filling defects, sim-
Fig. 10 – “Traditional” vs. modified-circle path-FSSW process for R = 300 rpm.
ilar to the so-called “tunnel defects”, well known in FSW of butt joints, are found out (Fig. 9). Looking at a few reference results obtained for FSW (Barcellona et al., 2004), it arises that the proper range of rotational velocities for the modified FSSW process is lower than the optimal one for FSW of butt joints: this is due to the absence of the tilt angle in the FSSW process that limits the possibility of the material to “follow” the tool movement, especially when the rotational speed – and
Fig. 11 – Top view (a) and etched AA section (b) of the modified process joint; etched section (c) of the “traditional” process joint (R = 300 rpm, Vf = 60 mm/min, a and b: circle path).
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
315
Fig. 12 – 60× rebuilt image of the cross-section of the modified process joint (R = 300 rpm, Vf = 60 mm/min, circle path), (a) middle of the joint and (b) bottom of the joint.
consequently the advancing speed, with the latter defined in mm/rev in this study – is high.
4.3.
Comparison with the “traditional” FSSW process
Fig. 10 presents a comparison between the modified (“tool path”—circle path case study) and the “traditional” (“no tool path”) process in terms of joints failure load. In particular, an R value of 300 rpm, which gave the best results for both the configurations, has been selected, and the results in terms of maximum load transmitted by the joints are plotted. A significant difference in the maximum load values is immediately observed, with the modified process being more effective for all the considered Vf values. What is more, the curve referred to the modified process, as highlighted before, is less sensitive to the Vf variation, and the maximum delta between the two curves is thus observed for the highest Vf value considered (Vf = 0.6 mm/rev). Such differences can be explained looking at the next Fig. 11. Both the top view (Fig. 11a) and the transverse section observation (Fig. 11b), obtained from a joint welded utilizing the most effective set of operative parameters previously found, reveal that the material flow occurring during the modified process generates a stirred zone dramatically larger than the one obtained with “traditional” FSSW processes (Fig. 11c). Consequently, the bonding between the two sheets involves larger areas and in this way more effective mechanical performances are reached. A deeper observation of the transverse section of the welded joint is shown in the next Fig. 12a and b. The 60× rebuilt image 12a, centered at 1.5 mm from the bottom of the weld, i.e. at the interface between the two sheets, clearly shows that a perfect bonding is obtained even far from the center of the weld, i.e. from the hole left by the tool pin, with the separation line between the welded sheets being visible just at the right end of the picture. Besides, a mostly homogeneous microstructure is observed from the tool pin hole to the above mentioned separation line, indicating that a large nugget area is obtained, resulting in dramatically enhanced joint performances. From Fig. 11b the “wavy” profile of the stirred zone at the bottom of the joint, due to the spiral movement of the tool, can be seen. The effect of the tool path pitch (kept constant and equal to 1.8 mm, as already mentioned) on the joint per-
formance has not been taken into account in this study, but it can be considered for further developments. In Fig. 13 the observed stirred zones – see dotted lines in Fig. 11b and c – and the corresponding joint failure loads are reported for all the investigated case studies and compared to the ones obtained by “traditional” FSSW (R = 300 rpm, Vf = 0.6 mm/rev). The significant increase in the stirred zones extension is mirrored by the joint resistance, though the failure load does not increase proportionally to the measured area; for instance, the increases in stirred area and joint failure load between the “traditional” FSSW and the Ey case study are 262% and 93%, respectively, while the same increases are 601% and 112% for the Sq case study. Finally, it should be observed that maintaining good mechanical properties with high processing speed is always a desirable effect for a process that is likely to be used for industrial applications, for which production time is a key factor; in fact, the additional time required for the tool path completion is small if compared to the total processing time (basically due to clamping and sinking), and it is partially counterbalanced both by the elimination of the dwelling time, during which the tool stays in the sheets before being retracted, and by the relatively high advancing velocity reachable maintaining acceptable performances. Fig. 14 shows a quantitative correlation between the process times and the joint failure loads, together with a comparison between a “traditional” FSS welded joint and the Sq case study.
Fig. 13 – TMAZ area and joint failure load at the varying of the tool path – circle path – and comparison with the “traditional” FSSW results.
316
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
velocity reachable maintaining acceptable performances. A good compromise between process time and joint resistance is found. On the basis of the above observations, the investigated technology appears a promising joining technique in order to obtain effective lap joints. Future work will be focused on the development of a specific FEM model aimed to highlight and further clarify the complex material flow occurring during the process and its effect on the microstructural properties. Further developments include the definitive optimization of the operative and geometrical parameters of the process. Fig. 14 – Process time and joint resistance, Sq case study, Vf = 0.6 mm/rev.
Although the best joint performance is reached for an R value equal to 300 rpm and a Vf value equal to 180 mm/min, an acceptable compromise between process time and joint effectiveness is obtained for R = 500 rpm and Vf = 300 mm/min; for such a test, a dramatic drop off of the process time is observed and, at the same time, only a slight reduction of the failure load is measured with respect to the best performance.
5.
Conclusions
In the paper, a variation of the FSSW process was considered. In particular, four different spiral tool paths were assigned to the welding tool, after the sinking phase, in order to improve the material flow and consequently the joint resistance. The following considerations can be drawn: • For each tested configuration the best performance is obtained at the smallest considered value of the rotational speed R, but the sensitivity to R and Vf changes depending on the chosen tool path. In particular, the Sq and C case studies show about the same maximum failure load, but the larger stirred area found for the Sq case study results in the higher reliability showed at varying operative parameters. • The proper range of rotational velocities for the modified FSSW process is lower than the optimal one for FSW of butt joints: in fact, the absence of the tilt angle in the FSSW process limits the possibility of the material to flow around the rotating tool, especially with high velocities. • The good correlation found between the stirred areas, the “path area” and the joint resistance supports the assumption that the enlargement of the stirred zone directly affects the joint performances and explains the significant difference observed in the joint mechanical response between the “traditional” FSSW and the modified process proposed in this study. • The modified FSSW requires a larger process time with respect to the traditional one, but such additional time is almost neglectable if compared to the total processing time, and it is partially counterbalanced both by the elimination of the dwelling time and by the relatively high advancing
Acknowledgement This work was made using MIUR (Italian Ministry for University and Scientific Research) funds.
references
Addison, A.C., Robelou, A.J., 2004. Friction stir spot welding: principal parameters and their effects. In: Proceedings of Fifth International Symposium on Friction Stir Welding, Metz, France, ISBN 1-903761-04-2. Barcellona, A., Buffa, G., Fratini, L., 2004. Process parameters analysis in friction stir welding of AA6082-T6 sheets. In: Keynote paper of the VII ESAFORM Conference, Trondheim, Norway, pp. 371–374. Buffa, G., Fratini, L., Piacentini, M., 2007. Tool path design in friction stir spot welding of AA6082-T6 aluminum alloys. Key Eng. Mater. 344, 767–774. Chae, S.W., Kwon, K.Y., Lee, T.S., 2002. An optimal design system for spot welding locations. Finite Elem. Anal. Des. 38 (3), 277–294. Darwish, S.M., Al-Dekhial, S.D., 1999. Statistical models for spot welding of commercial aluminum sheets. Int. J. Mach. Tools Manuf. 39 (10), 1589–1610. Fratini, L., Buffa, G., 2005. CDRX modelling in friction stir welding of aluminium alloys. Int. J. Mach. Tools Manuf. 45/10, 1188–1194. Hemmingson, E., 1996. New robot improves cost-efficiency of spot welding. Fuel Energy Abstr. 37 (4), 273. Jata, K.V., Semiatin, S.L., 2000. Continuous dynamic recrystallization during friction stir welding of high strength aluminum alloys. Scripta Mater. 43, 743–749. Larsson, J.K., 1994. Clinch-joining—a cost effective joining technique for BIW assembly. In: Proceedings of the International Conference on Body Engineering Clinch-joining, vol. 10. Advanced Technologies and Processes, pp. 140–145. Lennon, R., Pedreschi, R., Sinh, B.P., 1999. Comparative study of some mechanical connections in cold formed steel. Constr. Build. Mater. 13, 109–116. Liu, H.J., Fujii, H., Maeda, M., Nogi, K., 2003. Tensile properties and fracture locations of friction-stir-welded joints of 2017-T351 aluminum alloy. J. Mater. Process. Technol. 142, 692–696. Pan, T.Y., Joaquin, A., Wilkosz, D.E., Reatherford, L., Nicholson, J.M., Feng, Z., Santella, M.L., 2004. Spot friction welding for aluminum joining. In: Proceedings of Fifth International Symposium on Friction Stir Welding, Metz, France, ISBN 1-903761-04-2.
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Rhodes, C.G., Mahoney, M.W., Bingel, W.H., Spurling, R.A., Bampton, C.C., 1997. Effects of friction stir welding on microstructure of 7075 aluminum. Scripta Mater. 36 (1), 69–75. Riches, S.T., Westgate, S.A., Nicholas, E.D., Powell, H.J., 1995. Advanced joining technologies for lightweight vehicle manufacture. In: Proceedings of the Materials for Lean Weight Vehicles Conference, Warwick, UK, pp. 137–146.
317
Shigematsu, I., Kwon, Y.J., Suzuki, K., Imai, T., Saito, N., 2003. Joining of 5083 and 6061 aluminum alloys by friction stir welding. J. Mater. Sci. Lett. 22, 343–356. Su, J.Q., Nelson, T.W., Mishra, R., Mahoney, M., 2003. Microstructural investigation of friction stir welded 7050-T654 aluminium. Acta Mater. 51, 713–729.
journal homepage: www.elsevier.com/locate/jmatprotec
On the influence of tool path in friction stir spot welding of aluminum alloys G. Buffa, L. Fratini ∗ , M. Piacentini Dipartimento di Tecnologia Meccanica, Produzione e Ingegneria Gestionale, Universita` di Palermo, Viale delle Scienze, 90128 Palermo, Italy
a r t i c l e
i n f o
a b s t r a c t
Article history:
Friction stir spot welding (FSSW) has been proposed as an effective technology to spot weld
Received 3 September 2007
the so-called “difficult to be welded” metal alloys. In the paper, a variation of the FSSW
Received in revised form
process has been considered. A tool path is given after the sinking phase nearby the ini-
2 January 2008
tial penetration site; in this way a larger welding spot is obtained and more material is
Accepted 4 January 2008
involved in the bonding process. The process mechanics of such modified FSSW process is highlighted and the joint strength undergoing tensile tests is considered at the varying both of the assigned tool path and of a few process parameters. Macro- and micro-analyses
Keywords:
are made in order to analyze the local material microstructure evolution. It is found that
Welding
improved performances, with respect to the “traditional” FSSW process, are obtained for all
Spot
the considered case studies.
Microstructure
1.
Introduction
Spot welding can be considered a very common joining technique in automotive and generally in transportation industries. Apart from the utilized joining technology and in particular from the heat source (Joule effect, electric arc and so on) it permits to obtain effective lap joints with short process times. What is more it is easily carried out through robots and automated systems as shown in Chae et al. (2002), Hemmingson (1996) and Darwish and Al-Dekhial (1999). It should be observed that in the past decades mechanical fastening techniques, such as clinching and riveting, also have been widely utilized, since they permit to obtain the same advantages of the former welding processes and they do not suffer from the insurgence of typical welding defects due to the fusion of the base material. For instance Larsson (1994), Lennon et al. (1999) and Riches et al. (1995) highlighted the benefits given by mechanical fastening techniques in several
∗
© 2008 Elsevier B.V. All rights reserved.
applications such as automotive assemblies and constructions. Friction stir welding (FSW) has been presented and proposed at the beginning of the nineties as an innovative technique able to join also the so-called un-weldable or difficult to be welded light weight alloys which are very common materials in automotive or aerospace industries. Actually classic welding processes, i.e. TIG or laser, determine quite weak joints due to a strong increase of the average grain size observed in the melted zone and to quite large thermally affected zones. Furthermore strong precautions have to be taken during classic welding processes in order to avoid inclusions and other typical defects in the joint core. In turn, FSW is a solid state welding process in which a specially designed rotating pin is first inserted into the adjoining edges of the sheets to be welded with a proper tilt angle and then moved all along the joint. Such tool produces frictional and plastic deformation heating in the welding zone; actually no melting of material is observed during FSW. Then, as the tool moves,
Corresponding author. Tel.: +39 091 6657051; fax: +39 091 6657039. E-mail addresses: [email protected] (G. Buffa), [email protected] (L. Fratini), [email protected] (M. Piacentini). 0924-0136/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2008.01.001
310
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
material is forced to flow around the tool in a quite complex flow pattern determining a very particular microstructure characterized by continuous dynamic recrystallization phenomena. Several authors have been investigating FSW in the past years; for the purposes of the present research interesting information related to the occurring metallurgy, to the material flow and to the mechanical performances of the joints can be found out in Liu et al. (2003), Rhodes et al. (1997) and in Shigematsu et al. (2003). As a natural evolution of the FSW process, friction stir spot welding (FSSW) has been recently proposed for the development of lap joints also by Pan et al. (2004) and by Addison and Robelou (2004). In such a process, directly based on the FSW process mechanics, a rotating tool with a probe pin is introduced in the two overlapped blanks to be jointed supported by a proper back-plate. The rotating tool generates friction heat in the specimens and at the same time a material flow is determined. The heated and softened material close to the tool plastically deforms and a bond is made between the bottom surface of the upper sheet and the top surface of the lower sheet. No linear movement is given to the tool which is retracted from the workpiece when the stirring process is completed. As far as the process parameters are regarded, both geometrical and technological aspects have to be considered: both the former and the latter affect the material flow during the process and the generated heat flux. In this way the tool shoulder and the pin shapes and geometries are very important since they influence the circumferential material flow. What is more, the tool sinking into the blanks, i.e. the applied normal load, determines, together with the chosen tool rotating speed, the friction forces work and, consequently, the generated heat flux. As far as technological parameters are considered, also the process times, namely the tool descending time, the process time length and finally the tool ascending time, play an important role in the FSSW process. In the paper, a variation of the FSSW process has been considered: after the sinking phase, the tool is fed along a prescribed path close to its initial print. In this way a wider spot weld is obtained and a larger material stirred zone is obtained; at the end of the process, as the tool is released from the welded joint, a hole reproducing the tool pin is left in the welded sheets right at the center of the path, just like the one observed in “traditional” FSSW processes. An improvement in the material flow during the process is reached, which
enhances the joint resistance. The differences, in terms of process mechanics and local material microstructure evolution between such process and the “traditional” FSSW are highlighted in the next sections through macro- and microanalyses. The maximum load in tensile tests was considered as performance index of the joints. The influence of the most relevant process parameters on the mechanical performances of the joints is considered.
2. “Traditional” and modified FSSW processes As briefly described before, FSSW is aimed to obtain lap joints utilizing a cylindrical tool with a pin tip centered on the bottom circular face; in the “traditional” process, the tool rotates on its axis and it is inserted into the specimens to be jointed with a normal force. In other words the tool is pushed down into the two overlapped blanks until a fixed level of reaction force is reached. A back-plate is utilized on the bottom side of the specimens to support the applied load and no linear translation is given to the tool, which is retracted after an assigned time length. As the rotating tool is inserted into the sheets (Fig. 1A) a local backward extrusion mechanics is observed and a full contact between the upper sheet and the tool shoulder is reached. Then, keeping the tool into the sheets, a heat flux is generated by the friction forces work and plastic deformation work decaying into heat (Fig. 1B). Finally, after a proper time length, the tool is moved up and the joint is released (Fig. 1C). The observation of the material flow and of its locally reached microstructure leads to the full understanding of the process mechanics. Considering a section of the joint (Fig. 2), the backward extrusion mechanics is highlighted for the two overlapped sheets. Due to the absence of tool feed rate no asymmetry in the metal flow, typical of the FSW processes, is observed. In particular, the interface surface of the lower specimen is deformed upwards and such material acts as a mechanical anchor between the two jointed sheets, as shown in the 60× magnification of Fig. 2. As in FSW processes, a detailed observation of the material microstructure in the joint section allows finding out a few different areas, as shown in the next Fig. 3, where a 250× image of the material zone on the lateral surface of the pin imprint is reported (zone 2 in Fig. 2).
Fig. 1 – The FSSW process.
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
311
Fig. 2 – The FSSW joint transverse section highlighting the anchoring effect.
In particular, as in FSW, moving from the periphery of the joint towards the welding line, i.e. towards the tool pin surface, first of all the parent material, in which no metallurgical modification is observed after the welding process, is found. Then a simple heat affected zone (HAZ, C) is reached, in which the material has undergone a thermal cycle which has modified the microstructure and the mechanical properties. Continuing towards the welding spot, a zone in which the material has been plastically deformed by the stirring action of the tool is encountered. It should be observed that in such zone the effects of the induced heat flux are also observed with an enlargement of the material average grain size. In this way, this zone is a thermo-mechanically affected one (TMAZ, B). Finally the so-called nugget is found (A). The latter is a recrystallized area in which the original grain and subgrain boundaries appear to be replaced with fine, equiaxed recrystallized grains characterized by a nominal dimension of few m as described by Jata and Semiatin (2000), Su et al. (2003) and Fratini and Buffa (2005). It should be observed that the peculiarities of the obtained microstructures strongly determine the joint effectiveness and its mechanical behavior. In the considered modified FSSW process, a tool path is given after the completion of the sinking phase (Fig. 1A); in other words, the tool is not left in its position but it is moved along a peculiar trajectory determining a definitively larger stirred zone in the material. Once completed the assigned tra-
Fig. 3 – The FSSW joint microstructure.
jectory, the tool leaves the sheets to be welded. In particular four different spiral movements have been chosen with the aim to enlarge both the thermo-mechanically affected area and the nugget (Fig. 4), and a few combinations of advancing speed and rotating speed have been investigated. It should be observed that all the considered tool trajectories are characterized by the same overall dimension as it will be clarified in the next section. What is more, macro- and micro-observations on the welded joints permitted to highlight the metallurgical effect of the used tool paths on the material, justifying the mechanical performances of the joints.
3.
The developed experiments
In order to carry out the FSSW tests a properly designed clamping fixture was utilized to fix the specimens to be welded on a CNC milling machine. The steel plates composing the fixture were finished at the grinding machine to assure a uniform pressure distribution on the fixed specimens. What is more proper sheet supports were utilized in order to obtain the lap joints. As far as the utilized tool is regarded, it was made in H13 steel quenched at 1020 ◦ C, characterized by a 52 HRc hardness. The shoulder was 15 mm in diameter and a 40◦ conical pin was adopted, with a major diameter of 7 mm and a minor diameter of 2.2 mm; the pin height was 2.6 mm. The above geometrical characteristics were chosen on the basis of the knowledge acquired on FSW and FSSW and on preliminary tests developed by the authors and presented in Buffa et al. (2007): such tests have been performed utilizing a circular tool path, and no quantitative evaluation of the stirred volume, the TMAZ area and the process time was performed; no further investigation was made at the varying geometry of the tool. The utilized base material was AA6082-T6 aluminum alloy, 1.5 mm in thickness. The material was characterized by an average yield stress of 280 MPa and an ultimate tensile stress (UTSb) of 319 MPa. The material showed a microhardness equal to 120Hv and an average grain size of about 80 m. The influence of a few process variables on the mechanical performance of the welded lap joints was investigated in this paper; in particular the tool rotation speed (R) and the advancing speed for the prescribed tool path (Vf ) were taken into account. Each of the above indicated variables was varied in a
312
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Fig. 4 – Sketch of the assigned tool axis paths with respect to the tensile tests applied force direction.
proper range assuming different levels and the effects on the joining process in terms of joint resistance and of local material microstructure were observed. The tool sinking into the specimens during the process (h = 2.85 mm) and the sinking speed (equal to 0.003 mm/rev) were kept constant for all the tests. What is more, as briefly described before, four different spiral movements were given to the tool utilizing a CAD-CAM software. In particular a circle path (C), a square path (Sq), an ellipse-x (Ex) and an ellipse-y (Ey) path, i.e. with the major axis respectively orthogonal and parallel to the applied load (F) direction in the tensile tests, were considered (see again Fig. 4). The circle diameter, the square edge and the ellipses major axis were chosen equal to 10 mm; in this way the same overall dimension was obtained for the final tool print left on the welded specimen at the end of the process. A constant pitch of the tool path equal to 1.8 mm was assumed for all the investigated case studies. Starting from the periphery of the given path, the rotating tool enters the sheets and, after the completion of the sinking phase, the tool axis starts moving along the prescribed path with different values of advancing velocity. In the following Table 1 the investigated values of the considered process variables are reported. In particular three different values of rotational speed (R), namely 300, 500 and 700 rpm, and three different values of advancing speed (Vf ), namely 0.2, 0.4 and 0.6 mm/rev were chosen: a total of nine different tests were
Table 1 – The utilized process plan R (rpm)
Advancing velocity (mm/rev) Vf = 0.2 (mm/min)
300 500 700
60 100 140
Vf = 0.4 (mm/min) 120 200 280
Vf = 0.6 (mm/min) 180 300 420
thus carried out both for the “traditional” FSSW process and the modified one, and each test was repeated three times. The investigated levels and ranges of each parameter were chosen on the basis of preliminary tests (Buffa et al., 2007) and literature data (Pan et al., 2004; Addison and Robelou, 2004). From each joint specimen for either tensile tests macroand micro-observations were made. Macro-observations were aimed to analyze the material area involved in the process mechanics and the possible presence of macrodefects; through the micro-observations the different material zones, i.e. the thermally affected zone, the thermo-mechanically affected ones, and the so-called nugget zone, were observed and the differences between the “traditional” and modified FSSW techniques were highlighted. In order to obtain such results the specimens were properly treated with Keller reagent and observed by a light microscope. As far as the tensile tests are regarded, they have been developed on specimens presenting an average width of 30 mm.
4.
Results and discussion
4.1.
Comparison between the tested tool paths
First of all a comparison between the chosen tool paths in terms of maximum load transferred by the welded joints was performed: Fig. 5 shows the average tensile tests results for an advancing velocity of 0.6 mm/rev. It should be observed that no comparison in terms of stress can be made for spot welding or spot mechanical fastening, namely clinching or riveting, since the joints are not characterized by a uniform transferring load section, as for instance in the welding processes of butt joints. For each configuration the best performance is obtained at the lowest considered value of the rotational speed R, and a decrease in the maximum load is observed at the increase of R, but a few differences should be highlighted on the basis of the obtained results: both the square and the two ellipse paths show a lower sensitivity to R with respect to the circle path, resulting in larger maximum load values obtained at higher
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Fig. 5 – Comparison of the maximum loads for the investigated tool paths at the varying of R (Vf = 0.6 mm/rev).
Fig. 6 – TMAZ area, tool path area and joint failure load at the varying of the tool path.
rotational speeds. In particular, with R = 300 rpm, the square path case study presents about the same maximum load value found for the circle path case study, i.e. the best overall performance between all the considered case studies. At the increase of R, the best joints are the ones welded utilizing the square path, while the circle path case study eventually becomes the less effective (R = 700 rpm). As far as the two elliptical paths are regarded, they show about the same behavior, with the maximum load for the ellipse-x case study slightly higher, due to the larger section in the direction orthogonal to the applied load, as better explained in the following. Fig. 6 shows a quantitative analysis of the stirred zone (namely the TMAZ and the HAZ areas) for all the investigated case studies. In particular the aforementioned stirred areas, measured in a transverse section orthogonal to the applied load direction, are reported at varying tool path for the operative parameters set that gave the best results, i.e. R = 300 rpm and Vf = 0.6 mm/rev (as highlighted in Fig. 5). Such areas have been correlated to the results of the tensile tests and to the areas of the external geometrical figure described by the tool axis in a top view of the joint. In other words the areas of the geometrical figures described by the tool during its path have been considered (for example, the square path, 10 mm in edge, has an area of 100 mm2 , see again Fig. 4). It has to be noticed that such an area is strictly correlated to the stirred volume because the area – and consequently the volume – inside such paths is also involved in the stirring of material by the tool pin,
313
Fig. 7 – Influence of the tool path on the measured TMAZ area in the considered transverse section.
due to the spiral movement and thus represents an important indicator of the quality of the bond. As far as the TMAZ area is regarded, the highest value is obtained for the Sq case study, while lower values are found for the C, Ex and Ey case studies, respectively. A first consideration can be made on the Ey case study: in such a section the smallest area is measured, with a sensible drop off with respect to the others; that was easily predictable remembering that the measured section (TMAZ) is orthogonal to the applied load direction (see again Fig. 4). On the other hand, the difference found out between the other case studies can be explained looking at Fig. 7. In the AA section, the width of the C, Sq and Ex profiles is the same, but the way the tool reaches such section is different for the three case studies (see black arrows in Fig. 7); in particular, for the Sq case study, the maximum width of 10 mm is reached from the beginning of the “vertical” edge, thus enhancing the material flow and resulting in a wider area involved in the material stirring, while for the C and Ex case studies such maximum value is reached only in the AA section. The area of the geometrical profile of each tool path, i.e. the area carrying the applied load in a horizontal section, is also reported and compared to the joint resistance. As already noticed, the Sq and C case studies show about the same failure load, but the larger stirred area found for the Sq case study results in the higher reliability showed at varying rotational speed R (see again Fig. 5). The good correlation found between the stirred areas, the “path area” and the joint resistance supports the assumption that enlarging the stirred zone directly improves the joint performances.
Fig. 8 – Modified process results at the varying of R and Vf (circle path case study).
314
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Fig. 9 – Filling defects for the circle path case study (R = 700 rpm, Vf = 0.4 mm/rev).
4.2.
Effect of the process parameters
The effect of the operative parameters on the joint performances is highlighted in the next Fig. 8 for the circle path case study: the average values of the maximum load transmitted by the joints are shown at varying tool feed rate (Vf ) and for different values of the tool rotating speed (R). It can be argued from the figure that the rotational speed (R) plays a significant role in the maximum load transmitted by the joints: the overall performances of the joints obtained with a smaller R value are in fact better than those obtained with larger R values. What is more, for each of the considered R values, the joint resistance decreases at the increase of the advancing velocity (Vf ), becoming the slope of such curves larger as R increases. As a consequence, the best performance, in the investigated ranges of the considered variables, is obtained with R = 300 rpm and Vf = 60 mm/min, and, utilizing such R value, the joint failure load is almost constant at the varying Vf . Similar trends were obtained for the other tool paths investigated. It can be thus stated that large rotational speed values result in poorer performances for all the four considered case studies, for which an unsatisfactory material flow is obtained: from the performed macro-observations, filling defects, sim-
Fig. 10 – “Traditional” vs. modified-circle path-FSSW process for R = 300 rpm.
ilar to the so-called “tunnel defects”, well known in FSW of butt joints, are found out (Fig. 9). Looking at a few reference results obtained for FSW (Barcellona et al., 2004), it arises that the proper range of rotational velocities for the modified FSSW process is lower than the optimal one for FSW of butt joints: this is due to the absence of the tilt angle in the FSSW process that limits the possibility of the material to “follow” the tool movement, especially when the rotational speed – and
Fig. 11 – Top view (a) and etched AA section (b) of the modified process joint; etched section (c) of the “traditional” process joint (R = 300 rpm, Vf = 60 mm/min, a and b: circle path).
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
315
Fig. 12 – 60× rebuilt image of the cross-section of the modified process joint (R = 300 rpm, Vf = 60 mm/min, circle path), (a) middle of the joint and (b) bottom of the joint.
consequently the advancing speed, with the latter defined in mm/rev in this study – is high.
4.3.
Comparison with the “traditional” FSSW process
Fig. 10 presents a comparison between the modified (“tool path”—circle path case study) and the “traditional” (“no tool path”) process in terms of joints failure load. In particular, an R value of 300 rpm, which gave the best results for both the configurations, has been selected, and the results in terms of maximum load transmitted by the joints are plotted. A significant difference in the maximum load values is immediately observed, with the modified process being more effective for all the considered Vf values. What is more, the curve referred to the modified process, as highlighted before, is less sensitive to the Vf variation, and the maximum delta between the two curves is thus observed for the highest Vf value considered (Vf = 0.6 mm/rev). Such differences can be explained looking at the next Fig. 11. Both the top view (Fig. 11a) and the transverse section observation (Fig. 11b), obtained from a joint welded utilizing the most effective set of operative parameters previously found, reveal that the material flow occurring during the modified process generates a stirred zone dramatically larger than the one obtained with “traditional” FSSW processes (Fig. 11c). Consequently, the bonding between the two sheets involves larger areas and in this way more effective mechanical performances are reached. A deeper observation of the transverse section of the welded joint is shown in the next Fig. 12a and b. The 60× rebuilt image 12a, centered at 1.5 mm from the bottom of the weld, i.e. at the interface between the two sheets, clearly shows that a perfect bonding is obtained even far from the center of the weld, i.e. from the hole left by the tool pin, with the separation line between the welded sheets being visible just at the right end of the picture. Besides, a mostly homogeneous microstructure is observed from the tool pin hole to the above mentioned separation line, indicating that a large nugget area is obtained, resulting in dramatically enhanced joint performances. From Fig. 11b the “wavy” profile of the stirred zone at the bottom of the joint, due to the spiral movement of the tool, can be seen. The effect of the tool path pitch (kept constant and equal to 1.8 mm, as already mentioned) on the joint per-
formance has not been taken into account in this study, but it can be considered for further developments. In Fig. 13 the observed stirred zones – see dotted lines in Fig. 11b and c – and the corresponding joint failure loads are reported for all the investigated case studies and compared to the ones obtained by “traditional” FSSW (R = 300 rpm, Vf = 0.6 mm/rev). The significant increase in the stirred zones extension is mirrored by the joint resistance, though the failure load does not increase proportionally to the measured area; for instance, the increases in stirred area and joint failure load between the “traditional” FSSW and the Ey case study are 262% and 93%, respectively, while the same increases are 601% and 112% for the Sq case study. Finally, it should be observed that maintaining good mechanical properties with high processing speed is always a desirable effect for a process that is likely to be used for industrial applications, for which production time is a key factor; in fact, the additional time required for the tool path completion is small if compared to the total processing time (basically due to clamping and sinking), and it is partially counterbalanced both by the elimination of the dwelling time, during which the tool stays in the sheets before being retracted, and by the relatively high advancing velocity reachable maintaining acceptable performances. Fig. 14 shows a quantitative correlation between the process times and the joint failure loads, together with a comparison between a “traditional” FSS welded joint and the Sq case study.
Fig. 13 – TMAZ area and joint failure load at the varying of the tool path – circle path – and comparison with the “traditional” FSSW results.
316
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
velocity reachable maintaining acceptable performances. A good compromise between process time and joint resistance is found. On the basis of the above observations, the investigated technology appears a promising joining technique in order to obtain effective lap joints. Future work will be focused on the development of a specific FEM model aimed to highlight and further clarify the complex material flow occurring during the process and its effect on the microstructural properties. Further developments include the definitive optimization of the operative and geometrical parameters of the process. Fig. 14 – Process time and joint resistance, Sq case study, Vf = 0.6 mm/rev.
Although the best joint performance is reached for an R value equal to 300 rpm and a Vf value equal to 180 mm/min, an acceptable compromise between process time and joint effectiveness is obtained for R = 500 rpm and Vf = 300 mm/min; for such a test, a dramatic drop off of the process time is observed and, at the same time, only a slight reduction of the failure load is measured with respect to the best performance.
5.
Conclusions
In the paper, a variation of the FSSW process was considered. In particular, four different spiral tool paths were assigned to the welding tool, after the sinking phase, in order to improve the material flow and consequently the joint resistance. The following considerations can be drawn: • For each tested configuration the best performance is obtained at the smallest considered value of the rotational speed R, but the sensitivity to R and Vf changes depending on the chosen tool path. In particular, the Sq and C case studies show about the same maximum failure load, but the larger stirred area found for the Sq case study results in the higher reliability showed at varying operative parameters. • The proper range of rotational velocities for the modified FSSW process is lower than the optimal one for FSW of butt joints: in fact, the absence of the tilt angle in the FSSW process limits the possibility of the material to flow around the rotating tool, especially with high velocities. • The good correlation found between the stirred areas, the “path area” and the joint resistance supports the assumption that the enlargement of the stirred zone directly affects the joint performances and explains the significant difference observed in the joint mechanical response between the “traditional” FSSW and the modified process proposed in this study. • The modified FSSW requires a larger process time with respect to the traditional one, but such additional time is almost neglectable if compared to the total processing time, and it is partially counterbalanced both by the elimination of the dwelling time and by the relatively high advancing
Acknowledgement This work was made using MIUR (Italian Ministry for University and Scientific Research) funds.
references
Addison, A.C., Robelou, A.J., 2004. Friction stir spot welding: principal parameters and their effects. In: Proceedings of Fifth International Symposium on Friction Stir Welding, Metz, France, ISBN 1-903761-04-2. Barcellona, A., Buffa, G., Fratini, L., 2004. Process parameters analysis in friction stir welding of AA6082-T6 sheets. In: Keynote paper of the VII ESAFORM Conference, Trondheim, Norway, pp. 371–374. Buffa, G., Fratini, L., Piacentini, M., 2007. Tool path design in friction stir spot welding of AA6082-T6 aluminum alloys. Key Eng. Mater. 344, 767–774. Chae, S.W., Kwon, K.Y., Lee, T.S., 2002. An optimal design system for spot welding locations. Finite Elem. Anal. Des. 38 (3), 277–294. Darwish, S.M., Al-Dekhial, S.D., 1999. Statistical models for spot welding of commercial aluminum sheets. Int. J. Mach. Tools Manuf. 39 (10), 1589–1610. Fratini, L., Buffa, G., 2005. CDRX modelling in friction stir welding of aluminium alloys. Int. J. Mach. Tools Manuf. 45/10, 1188–1194. Hemmingson, E., 1996. New robot improves cost-efficiency of spot welding. Fuel Energy Abstr. 37 (4), 273. Jata, K.V., Semiatin, S.L., 2000. Continuous dynamic recrystallization during friction stir welding of high strength aluminum alloys. Scripta Mater. 43, 743–749. Larsson, J.K., 1994. Clinch-joining—a cost effective joining technique for BIW assembly. In: Proceedings of the International Conference on Body Engineering Clinch-joining, vol. 10. Advanced Technologies and Processes, pp. 140–145. Lennon, R., Pedreschi, R., Sinh, B.P., 1999. Comparative study of some mechanical connections in cold formed steel. Constr. Build. Mater. 13, 109–116. Liu, H.J., Fujii, H., Maeda, M., Nogi, K., 2003. Tensile properties and fracture locations of friction-stir-welded joints of 2017-T351 aluminum alloy. J. Mater. Process. Technol. 142, 692–696. Pan, T.Y., Joaquin, A., Wilkosz, D.E., Reatherford, L., Nicholson, J.M., Feng, Z., Santella, M.L., 2004. Spot friction welding for aluminum joining. In: Proceedings of Fifth International Symposium on Friction Stir Welding, Metz, France, ISBN 1-903761-04-2.
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 8 ( 2 0 0 8 ) 309–317
Rhodes, C.G., Mahoney, M.W., Bingel, W.H., Spurling, R.A., Bampton, C.C., 1997. Effects of friction stir welding on microstructure of 7075 aluminum. Scripta Mater. 36 (1), 69–75. Riches, S.T., Westgate, S.A., Nicholas, E.D., Powell, H.J., 1995. Advanced joining technologies for lightweight vehicle manufacture. In: Proceedings of the Materials for Lean Weight Vehicles Conference, Warwick, UK, pp. 137–146.
317
Shigematsu, I., Kwon, Y.J., Suzuki, K., Imai, T., Saito, N., 2003. Joining of 5083 and 6061 aluminum alloys by friction stir welding. J. Mater. Sci. Lett. 22, 343–356. Su, J.Q., Nelson, T.W., Mishra, R., Mahoney, M., 2003. Microstructural investigation of friction stir welded 7050-T654 aluminium. Acta Mater. 51, 713–729.