- Email: [email protected]
Simulation of material flow in friction stir processing of a cast Al–Si alloy
Materials and Design 40 (2012) 415–426
Contents lists available at SciVerse ScienceDirect
Materials and Design journal homepage: www.elsevier.com/locate/matdes
Simulation of material flow in friction stir processing of a cast Al–Si alloy S. Tutunchilar a,⇑, M. Haghpanahi a, M.K. Besharati Givi b, P. Asadi b, P. Bahemmat a a b
School of Mechanical Engineering, Iran University of Science and Technology, Tehran, Iran School of Mechanical Engineering, College of Engineering, University of Tehran, Tehran, Iran
a r t i c l e
i n f o
Article history: Received 4 February 2012 Accepted 1 April 2012 Available online 19 April 2012 Keywords: Friction stir processing Finite element simulation Material flow Stir zone shape
a b s t r a c t The objective of the current paper is using DEFORM-3D software to develop a 3-D Lagrangian incremental finite element method (FEM) simulation of friction stir processing (FSP). The developed simulation allows prediction of the defect types, temperature distribution, effective plastic strain, and especially material flow in the weld zone. Three-dimensional results of the material flow patterns in the center, advancing and retreating sides were extracted using the point tracking. The results reveal that the main part of the material flow occurs near the top surface and at the advancing side (AS). Material near the top surface was stretched to the advancing side resulting in a non-symmetrical shape of the stir zone (SZ). Furthermore, macrostructure and temperature rise were experimentally acquired to evaluate the accuracy of the developed simulation. The comparison shows that the stir zone shape, defect types, powder agglomeration, and temperature rise, which were predicted by simulation, are in good agreement with the corresponding experimental results. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The Al–Si alloy is a well-known casting Aluminum alloy with low coefficient of thermal expansion, fine wear, corrosion resistances and good mechanical properties at the wide temperature range. Among the effective features, the microstructure and alloying elements are the decisive factors, which dictate the mechanical properties of this alloy. Coarse grains, dendritic arm spacing (DAS), shrinkage porosity, and needle-like morphology of silicon are some of the undesirable microstructural features which aggravate the mechanical properties of cast Al–Si alloys including the tensile strength, ductility, etc. [1–3]. Friction stir processing based on friction stir welding (FSW) is a solid state process for microstructural modifications [4], grain refinement [5], and fabrication of surface layer composites [6,7]. Despite the simple concept of FSW/FSP, simulation and analysis of material behavior including material flow during the process is of great complexity. A substantial effort has been done by several works to experimentally and numerically investigate the material flow as a dictating factor in the joint quality [8], some of which are present in literature. Early work in predication of material flow during FSW/FSP can be attributed to Colligan [9]. Afterwards, material flow during ⇑ Corresponding author. Tel.: +98 914 3526852; fax: +98 21 77240488. E-mail addresses: [email protected], [email protected] (S. Tutunchilar). 0261-3069/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.matdes.2012.04.001
FSW/FSP was experimentally investigated by employing tracer tracking [10–13], welding of dissimilar materials [14–17], and X-ray tomography [10,18]. Schmidt et al. [10] investigated the material flow by embedding a copper strip in the workpiece and using X-ray tomography. They detected a repeated deposition of copper marker in the weld zone due to a sliding/sticking condition. Also they estimated that the average velocity of material in the transition zone is 0.1–0.3 fractions of the tool rotational speed. Morisada et al. [11] investigations showed that the mechanism of onion rings formation is attributed to the convectional flow produced by the tool shoulder. In 2010, Lorrain et al. [13] by analyzing macrographs of the weld joints concluded that marker material is deposited near the top surface in the advancing side and near the bottom surface in the retreating side of the weld zone. Guerra et al. [15] employed a copper strip embedded at the center line of the welding sheets as well as overlapped friction stir welding of dissimilar aluminum alloys, from which the material flow in the vertical direction was examined. Li et al. [14] observed complex swirls and vortex-like flow patterns in dissimilar FSW of 2024 and 6061 Al alloys, which were revealed by optical micrographs as a result of the different etchings of the alloys. Within the last decade, a number of researchers [19–25] have been working on the thermo-mechanical modeling of FSW/FSP. However, very few works have been done in the simulation of material flow. In 2005, Zhang et al. [19] presented a two-dimensional simulation based on arbitrary Lagrangian–Eulerian (ALE) description using the ABAQUS commercial FEM code. Results
416
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Table 1 Chemical composition of LM13 (wt%).
Table 2 Thermal characteristics of the LM13 workpiece and H13 tool.
Element
Si
Cu
Mg
Ni
Fe
Mn
Al
Weight (%)
11.35
1.49
1.46
1.15
0.37
0.02
Bal.
Property
LM13
FSP tool H13
Backplate
Heat capacity (N/mm2 °C) Emissivity Conductivity (N/s °C) Heat transfer coefficient between tool and billet (N/°C s mm) Heat transfer coefficient between backing plate and billet (N/°C s mm) Heat transfer coefficient between tool/ workpiece and air (N/mm s °C)
2.57 0.7 117 11
3.24 0.7 24.5 11
– – – –
1
–
1
0.02
0.02
–
Fig. 1. Schematic illustration of force measurement setup during FSP.
Fig. 3. Simulation and experimental axial force.
showed that the shoulder has a significant effect on material flow pattern at the top surface, and this effect decreases by increasing the distance from the top surface. Liechty and Webb [21] in 2008 investigated various frictional boundary conditions on the material flow behavior. Their results showed that the shear model is superior to a sticking condition. Although a significant progress has been made in finding material flow during FSW/FSP, no systematic study has been carried out to numerically find all the main features of FSW/FSP. The current paper presents a fully coupled incremental Lagrangian thermo-mechanical simulation of FSP via employing the DEFORM-3D as commercial FEM software to predict the material flow pattern, temperature distribution, and effective plastic strain in the weld zone. Furthermore, temperature rise induced by the process, stir zone shape, the defect types and powder agglomeration were found experimentally and the results obtained by numerical simulation were compared with those acquired via experiment in order to validate the accuracy of the developed simulation. 2. Experimental procedures
Fig. 2. Flow stress–strain curves of base metal under different conditions.
demonstrated that the material has a revolving motion around the pin at the advancing side, whereas at the retreating side (RS) material does not revolve around the pin. In 2008, Zhang and Zhang [20] developed a fully coupled thermo-mechanical simulation and
The LM13 eutectic Al–Si as-cast alloy was subjected to FSP. The dimensions of plates were 100 mm in length, 50 mm in width and 7 mm in thickness. The chemical composition of LM13 alloy is given in Table 1. In order to recognize the material flow, graphite powder (with 5 lm particle size) was embedded in a groove with the depth and width of 3 and 1 mm, respectively. The FSP tool was a H13 hot working steel with a cylindrical pin. The shoulder and pin diameter, and pin height were 18, 6, and 4 mm, respectively. The tool rotational and the traverse speeds were also 900 rpm and 50 mm/min, respectively. The tilt angle was fixed at 3°. Subsequently, specimens were cut from friction stir processed (FSPed) samples and polished according to standard metallo-
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
417
Fig. 4. Schematic illustration of tool and workpiece.
machine bed and the back plate in this research, one charge amplifier was employed to convert the dynamometer charge signals to output voltages balanced to the forces obtained. The dynamometer was mounted on the bed of the machine using screws and a supporting block. The force data were obtained via a data acquisition system and plotted with respect to the time step of the process (Fig. 1). To attain the flow stress of the LM13 at various deformation conditions, hot compression test was also conducted at the temperature ranges of 300–500 °C and strain rates of 0.01 and 1 s1. Dimensions of compressive test samples were 15 mm in height and 10 mm in diameter (in accordance with ASTM E209 [26]). Fig. 5. Comparison of simulation and experimental temperature history at 9 mm from the center of weld and 2 mm under the top surface.
3. Model description Deform-3D software, which enables simulation of processes with large plastic deformation and point tracking, was employed to simulate the FSP. Moreover, to simplify the problem, the following assumptions were made in this study: (1) Material model of the workpiece is rigid-visco-plastic (RVP); (2) FSP tool and backing plate are rigid; (3) Friction factor is constant; (4) Thermal properties of the workpiece and FSP tool are constant; (5) Free surfaces of the workpiece and FSP tool were under free convection at ambient temperature of 20 °C; (6) The backing plate and two sides of workpieces were fixed in all degrees of freedom. The rigid visco-plastic FEM is generally used in the metal forming simulation and derived from principle of variational stability, which results in nonlinear equations. These equations can be solved by iterative methods (such as Newton–Raphson or direct method). Further details in governing equations can be found in Ref. [22]. In the FE model, tool and workpiece geometry and parameters (translational and rotational speeds and tilt angle) used in experiments were employed. 3.1. Material model
Fig. 6. Initial positions of points that have been selected for material flow in the stir zone.
graphic procedure, aiming to microstructural analysis. A thermocouple was also inserted in the workpiece at the distance of 9 mm from the center line and depth of 2 mm from the top surface at the AS. To measure the axial force (in the plunging direction), one-component Kistler dynamometer was fitted between the
The flow stress of LM13 cast Al alloy is given as a function of plastic strain, strain rate and temperature:
r ¼ r ðe; e_ ; TÞ
ð1Þ
is the flow stress, e is the effective plastic strain, e_ is the where r effective strain rate, and T is the temperature. Flow stress values are interpolated in log strain, log strain rate, and linear temperature space. For data points at 0.0 plastic strain, linear interpolation is used between the 0 value and the first non 0 values, and log interpolation is used for subsequent values.
418
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 7. Center line points flow pattern.
The relationship between flow stress and strain for LM13 alloy at temperature range of 300–500 °C and strain rates of 0.01 and 1 s1 are shown in Fig. 2. Thermal properties of the LM13 workpiece and H13 hot working steel are summarized in Table 2. 3.2. Friction model
the deforming body. To determine the best approximation for friction factor, the axial forces of FE modeling were calculated for three different friction factors, 0.4, 0.5, and 0.6, under 100 mm/min transverse speed and 900 rpm rotational speed. Friction factor of 0.5 exhibited the best match based on the comparison between the axial forces obtained by simulation and experimental measurement. Results are shown in Fig. 3.
Constant shear friction is used mostly for bulk-forming simulations. The frictional force in the constant shear model is defined by:
3.3. Workpiece and tool models
f ¼ mk
FSP tool was assumed to be a rigid body and was meshed with 11,300 tetrahedral elements. The workpieces were modeled as a temperature and strain rate dependent RVP material. The workpiece was meshed with 39,000 tetrahedral elements and
ð2Þ
where, f is the frictional stress; k is the shear yield stress; and m the friction factor. Thus, the friction is a function of the yield stress of
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
0.8 mm minimum element size. Automatic remeshing with a nonuniform mesh was applied to the workpiece. The smallest elements are located beneath the shoulder and the element size is increased by increasing the distance from the tool shoulder. Workpiece and tool meshes are schematically illustrated in Fig. 4. 4. Results and discussion 4.1. Numerical simulation and experimental verification of temperature
419
In order to validate the accuracy of the simulation results, the temperature distribution found via simulation were compared with the experimental results, as shown in Fig. 5. Both simulation and experimental results were reported at a distance of 9 mm from the weld center at the AS and 2 mm below the top surface. The comparison shows that although the numerical temperature results are slightly higher (19 °C) than those acquired by experiment, the simulation results are acceptably in good agreement with the experimental data. 4.2. Material flow
Since the flow stress is dictated by the temperature rise, heat generation during the process, which is a function of the plastic deformation and friction, it has an imperative role in material flow during FSW/FSP.
Fig. 6 shows the initial position of points selected to study material flow pattern in the SZ. Points P1–P8 are located 0.5 mm from each other in the weld center. Similarly, points P9–P16 and
Fig. 7 (continued)
420
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 8. Applied plastic strain to the points P1–P8.
P17–P24 are located at the advancing and retreating sides with the distance of 2.5 mm from the center line. Fig. 7 also shows the material flow at the weld center. Besides, Fig. 7a illustrates the initial position of points located along the center line and Fig. 6b–e track the aforementioned points during the process. As could be seen, the points near the bottom of the SZ (P7 and P8) have a semicircular movement from the leading side toward the trailing side. However, the part of material which is
Fig. 11. (a) Tunneling cavity formation at the behind of the pin and (b) slot formation at advancing side behind the pin.
close to the top surface, in spite of the movement to the tool pin behind, is stretched toward the advancing side. Such a behavior was clearly expected due to the frictional shear force induced by the tool shoulder. An interesting observation was that the final position of the points except point P1 is similar to SZ shape which will
Fig. 9. Material velocity (a) velocity contour in the cross section of the weld and (b) velocity diagram of points P1–P8.
Fig. 10. (a) Tunneling cavity formation at the behind of the pin and (b) slot formation at advancing side behind the pin.
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
421
Fig. 12. Advancing side points flow pattern.
be discussed. Point P1 is an exception, which is stuck to the shoulder and rotates with the shoulder for an unknown time period. Fig. 8 also indicates the plastic strain applied to the points P1– P8. Similar to the other works present in the literature, the plastic strain for the different points is about 30. Moreover, according to Fig. 8, the amount of effective plastic strain decreases upon moving to the SZ root. The effective plastic strain of point P1, which is stuck to the shoulder, is considerably higher than that of the other points, which can be attributed to the accumulation of the effective strain during several successive rotations around the pin. Furthermore, Fig. 9a and b indicate the velocity of points P1–P8 and velocity contour in the cross section. As was expected, velocity decreases upon increasing the distance from the top surface. Fig. 9b also reveals that the velocity at the RS is higher than that at the AS, which validates the results reported by Morisada et al. [8]. They employed X-ray transmission examination to show that velocity at the RS is slightly higher than that at the AS. Nevertheless, the amount of velocity reported by them is higher than that
obtained by the developed simulation in the current paper, which can be most probably attributed to the different friction conditions. Many researchers have discussed about frictional boundary condition [27–30] but still it is not clear that the friction is sticky or slippery. At the AS near the trailing side, the pin velocity has its own minimum value and due to the pressure reduction, the material barely flows into this region. This is the main reason for formation of tunneling cavity and thus the cavity mostly is formed at the AS near the trailing side [31,32]. The probability of tunneling cavity formation increases by increasing the traverse speed. Fig. 10a demonstrates the cavity that formed in the higher traverse speed (200 mm/min). As shown in Fig. 10b, a slot was created at the AS of trailing side. This slot and tunneling cavity was also visible in the experiment (Fig. 11). Flow patterns of points P9–P16 and P17–P26 are also shown in Figs. 12 and 13, respectively. As shown in Fig. 12, points initially located at the AS revolve around the pin and finally come back to the AS. However, this behavior is not observed for the points located at
422
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 12 (continued)
the RS and they are scattered over the AS and RS after passing of the tool. Moreover, the points near the SZ root remain at the RS and upon decreasing the distance from the top surface, material stretching toward the AS diminishes. In other words, the part of material near the tool shoulder at the AS, RS, and center line moves toward the AS. On the other words, most of the SZ materials are finally located in the AS after passing of the tool. This phenomenon, clearly shown in Fig. 14, can explicate the powder agglomeration at the AS near the top surface in the case of FSP with powder. This was obviously seen in the experiment. Fig. 15 reveals the graphite agglomeration occurred at the AS near the top surface. The agglomeration of embedded powder or marker at the AS has been widely reported in other works [33,34]. The shape of SZ can be predicted through using the point tracking option of DEFOM-3D software. Thus, in order to achieve a better understanding of material flow during the process, more points were picked inside the SZ before the tool passing (Fig. 16a). The final position of these points after tool passing is shown in Fig. 16b. The shape of SZ predicted by simulation was projected on Fig. 15
(SZ shape resulted from experiment) and the complex is shown in Fig. 17. It is revealed that the simulated SZ shape is in good agreement with the experiment. Fig. 18 opens new perspectives towards a better understanding of how materials move into the SZ. According to Fig. 18a, nine points at the RS (P1–P9) and ten points at the AS (P10–P19) were picked. Fig. 18b and c represent the final positions of the aforementioned points. As is seen, the points at AS remain in contact with the tool shoulder and rotate several times during the process. Nevertheless, the points at RS never revolve around the tool and are scattered over the SZ. Thus, it can be concluded that the main part of material inside the marked zone probably comes from the zone located at a certain distance from the tool pin and near the top surface at RS. Furthermore, the material at the AS travels a longer path compared to the material at the RS and the SZ center. Thus, it can be concluded that shifting the tool toward the RS can lead to a better powder distribution during the process since the higher traveling pass may result in a better powder distribution and mixing with the matrix. Fig. 19 is provided to validate this claim. Fig. 19
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
423
Fig. 13. Retreating side points flow pattern.
shows the graphite powder distribution in the FSPed specimen with 2 mm shifted tool toward the RS. As could be seen, shifting the tool toward the RS considerably led to a better powder distribution. Additionally, Mahmoud et al. [35] reported the same conclusion by experiment. 5. Conclusions FSP of a cast Al–Si alloy was simulated based on Lagrangian incremental with the aim of achieving a better understanding of martial behavior during the process. The SZ shape, forming defects, and effective plastic strain were also scrutinized and the simulation results were compared with those acquired via experiment.
The following results were achieved: 1. The SZ shape was successfully predicted. Deep analysis of the material flow revealed that the material near the top surface was stretched toward the AS due to the shoulder frictional force and upon increasing the distance from top surface the final position of the material remains almost constant in the transverse direction regardless of rotating around the pin. 2. Formation of tunneling defect occurs at the AS and near the root of the SZ. While powder agglomeration occurs at the AS and near the top surface. 3. Both simulation and experiment indicated that shifting the FSP tool toward the RS can result in a better powder distribution
424
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 13 (continued)
Fig. 14. Aggregation of points at the AS in final position.
Fig. 15. Cross section macro-image of FSPed specimen.
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
425
Fig. 16. (a) Initial position of points picked for studying the SZ shape and (b) final position of points.
Fig. 17. Comparison of simulated and experimental SZ shape.
Fig. 19. Cross section macro-image of FSPed specimen with 2 mm tool shifting toward the RS.
Fig. 18. (a) Initial positions of points picked for specification of initial location of points which finally are located in the dashed line, (b) and (c) final position of points.
426
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
References [1] Wang Feng, Liu Huimin, Ma Yajun, Jin Yuansheng. Effect of Si content on the dry sliding wear properties of spray-deposited Al–Si alloy. Mater Des 2004;25:163–6. [2] Özyürek Dursun, Aktar Niyazi, Aztekin Hasan. Design and construction of a thixo forming unit and production of Al–Si alloys. Mater Des 2008;29:1070–4. [3] Dwivedi DK, Sharma A, Rajan TV. Influence of silicon morphology and mechanical properties of piston alloys. Mater Manuf Processes 2005;20:777–91. [4] Karthikeyan L, Senthilkumar VS, Padmanabhan KA. On the role of process variables in the friction stir processing of cast aluminum A319 alloy. Mater Des 2010;31:761–71. [5] Ma ZY, Pilchak AL, Juhas MC, Williams JC. Microstructural refinement and property enhancement of cast light alloys via friction stir processing. Scr Mater 2008;58:361–6. [6] Azizieh M, Kokabi AH, Abachi P. Effect of rotational speed and probe profile on microstructure and hardness of AZ31/Al2O3 nanocomposites fabricated by friction stir processing. Mater Des 2011;32:2034–41. [7] Sharifitabar M, Sarani A, Khorshahian S, Shafiee Afarani M. Fabrication of 5052Al/Al2O3 nanoceramic particle reinforced composite via friction stir processing route. Mater Des 2011;32:4164–72. [8] Morisada Y, Fujii H, Kawahito Y, Nakata K, Tanaka M. Three-dimensional visualization of material flow during friction stir welding by two pairs of X-ray transmission systems. Scr Mater 2011;65:1085–8. [9] Colligan K. Material flow behaviour during friction stir welding of aluminum. Welding J 1999;78:229–37. [10] Schmidt HNB, Dickerson TL, Hattel JH. Material flow in butt friction stir welds in AA2024-T3. Acta Mater 2006;54:1199–209. [11] Morisada Y, Fujii H, Nagaoka T, Nogi K, Fukusumi M. Fullerene/A5083 composites fabricated by material flow during friction stir processing. Composites A 2007;38:2097–101. [12] Reynolds AP. Flow visualization and simulation in FSW. Scr Mater 2008;58:338–42. [13] Lorrain O, Favier V, Zahrouni H, Lawrjaniec D. Understanding the material flow path of friction stir welding process using unthreaded tools. J Mater Process Technol 2010;210:603–9. [14] Li Y, Murr LE, McClure JC. Flow visualization and residual microstructures associated with the friction-stir welding of 2024 aluminum–6061 aluminum. Mater Sci Eng, A 1999;271:213–23. [15] Guerra M, Schmidt C, McClure JC, Murr LE, Nunes AC. Flow patterns during friction stir welding. Mater Character 2003;49:95–101. [16] da Silva AAM, Arruti E, Janeiro G, Aldanondo E, Alvarez P, Echeverria A. Material flow and mechanical behaviour of dissimilar AA2024-T3 and AA7075T6 aluminium alloys friction stir welds. Mater Des 2011;32:2021–7. [17] Liechty BC, Webb BW. The use of plasticine as an analog to explore material flow in friction stir welding. J Mater Process Technol 2007;184:240–50.
[18] Zettler R, Donath T, dos Santos JF, Beckman F, Lohwasser D. Validation of marker material flow in 4 mm thick friction stir welded Al 2024-T351 through computer microtomography and dedicated metallographic techniques. Adv Eng Mater 2006;8:487–90. [19] Zhang HW, Zhang Z, Chen’ JT. The finite element simulation of the friction stir welding process. Mater Sci Eng, A 2005;403:340–8. [20] Zhang Z, Zhang HW. A fully coupled thermo-mechanical model of friction stir welding. Int J Adv Manuf Technol 2008;37:279–93. [21] Liechty BC, Webb BW. Modeling the frictional boundary condition in friction stir welding. Int J Mach Tools Manuf 2008;48:1474–85. [22] Asadi P, Mahdavinejad RA, Tutunchilar S. Simulation and experimental investigation of FSP of AZ91 magnesium alloy. Mater Sci Eng, A 2011;528:6469–77. [23] Chen CM, Kovacevic R. Finite element modeling of friction stir welding— thermal and thermomechanical analysis. Int J Mach Tools Manuf 2003;43:1319–26. [24] Zhang Z, Zhang HW. Numerical studies on the effect of transverse speed in friction stir welding. Mater Des 2009;30:900–7. [25] Assidi Mohamed, Fourment Lionel, Guerdoux Simon, Nelson Tracy. Friction model for friction stir welding process simulation: calibrations from welding experiments. Int J Mach Tools Manuf 2010;50:143–55. [26] ASTM standard E 209-00. Standard practice for compression tests of metallic materials at elevated temperatures with conventional or rapid heating rates and strain rates. In: Annual book of ASTM standards, vol. 03.01(July); 2000. [27] Guerra M, Schmidt C, McClure JC, Murr LE, Nunes AC. Flow patterns during friction stir welding. Mater Charact 2002;49:95–101. [28] Schmidt HNB, Dickerson TL, Hattel JH. Material flow in butt friction stir welds in AA2024-T3. Acta Mater 2006;54:1199–209. [29] Gerlich A, Yamamoto M, North TH. Strain rates and grain growth in Al 5754 and Al 6061 friction stir spot welds. Metall Mater Trans A 2007;38A:1291–302. [30] Gerlich A, Avramovic-Cingara G, North TH. Stir zone microstructure and strain rate during Al 7075-T6 friction stir spot welding. Metall Mater Trans A 2006;37A:2773–86. [31] Arbegast William J. A flow-partitioned deformation zone model for defect formation during friction stir welding. Scr Mater 2008;58:372–6. [32] Mishra RS, Ma ZY. Friction stir welding and processing. Mater Sci Eng R 2005;50:1–78. [33] Lorrain Olivier, Favier Véronique, Zahrouni Hamid, Lawrjaniec Didier. Understanding the material flow path of friction stir welding process using unthreaded tools. J Mater Process Technol 2010;210:603–9. [34] Asadi P, Faraji G, Masoumi A, Givi MKB. Experimental investigation of magnesium-base nanocomposite produced by friction stir processing: effects of particle types and number of friction stir processing passes. Metall Mater Trans A 2011;42A:2820–32. [35] Mahmoud ERI, Takahashi M, Shibayanagi T, Ikeuchi K. Effect of friction stir processing tool probe on fabrication of SiC particle reinforced composite on aluminium surface. Sci Technol Weld Joining 2009;14:413–25.
Contents lists available at SciVerse ScienceDirect
Materials and Design journal homepage: www.elsevier.com/locate/matdes
Simulation of material flow in friction stir processing of a cast Al–Si alloy S. Tutunchilar a,⇑, M. Haghpanahi a, M.K. Besharati Givi b, P. Asadi b, P. Bahemmat a a b
School of Mechanical Engineering, Iran University of Science and Technology, Tehran, Iran School of Mechanical Engineering, College of Engineering, University of Tehran, Tehran, Iran
a r t i c l e
i n f o
Article history: Received 4 February 2012 Accepted 1 April 2012 Available online 19 April 2012 Keywords: Friction stir processing Finite element simulation Material flow Stir zone shape
a b s t r a c t The objective of the current paper is using DEFORM-3D software to develop a 3-D Lagrangian incremental finite element method (FEM) simulation of friction stir processing (FSP). The developed simulation allows prediction of the defect types, temperature distribution, effective plastic strain, and especially material flow in the weld zone. Three-dimensional results of the material flow patterns in the center, advancing and retreating sides were extracted using the point tracking. The results reveal that the main part of the material flow occurs near the top surface and at the advancing side (AS). Material near the top surface was stretched to the advancing side resulting in a non-symmetrical shape of the stir zone (SZ). Furthermore, macrostructure and temperature rise were experimentally acquired to evaluate the accuracy of the developed simulation. The comparison shows that the stir zone shape, defect types, powder agglomeration, and temperature rise, which were predicted by simulation, are in good agreement with the corresponding experimental results. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The Al–Si alloy is a well-known casting Aluminum alloy with low coefficient of thermal expansion, fine wear, corrosion resistances and good mechanical properties at the wide temperature range. Among the effective features, the microstructure and alloying elements are the decisive factors, which dictate the mechanical properties of this alloy. Coarse grains, dendritic arm spacing (DAS), shrinkage porosity, and needle-like morphology of silicon are some of the undesirable microstructural features which aggravate the mechanical properties of cast Al–Si alloys including the tensile strength, ductility, etc. [1–3]. Friction stir processing based on friction stir welding (FSW) is a solid state process for microstructural modifications [4], grain refinement [5], and fabrication of surface layer composites [6,7]. Despite the simple concept of FSW/FSP, simulation and analysis of material behavior including material flow during the process is of great complexity. A substantial effort has been done by several works to experimentally and numerically investigate the material flow as a dictating factor in the joint quality [8], some of which are present in literature. Early work in predication of material flow during FSW/FSP can be attributed to Colligan [9]. Afterwards, material flow during ⇑ Corresponding author. Tel.: +98 914 3526852; fax: +98 21 77240488. E-mail addresses: [email protected], [email protected] (S. Tutunchilar). 0261-3069/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.matdes.2012.04.001
FSW/FSP was experimentally investigated by employing tracer tracking [10–13], welding of dissimilar materials [14–17], and X-ray tomography [10,18]. Schmidt et al. [10] investigated the material flow by embedding a copper strip in the workpiece and using X-ray tomography. They detected a repeated deposition of copper marker in the weld zone due to a sliding/sticking condition. Also they estimated that the average velocity of material in the transition zone is 0.1–0.3 fractions of the tool rotational speed. Morisada et al. [11] investigations showed that the mechanism of onion rings formation is attributed to the convectional flow produced by the tool shoulder. In 2010, Lorrain et al. [13] by analyzing macrographs of the weld joints concluded that marker material is deposited near the top surface in the advancing side and near the bottom surface in the retreating side of the weld zone. Guerra et al. [15] employed a copper strip embedded at the center line of the welding sheets as well as overlapped friction stir welding of dissimilar aluminum alloys, from which the material flow in the vertical direction was examined. Li et al. [14] observed complex swirls and vortex-like flow patterns in dissimilar FSW of 2024 and 6061 Al alloys, which were revealed by optical micrographs as a result of the different etchings of the alloys. Within the last decade, a number of researchers [19–25] have been working on the thermo-mechanical modeling of FSW/FSP. However, very few works have been done in the simulation of material flow. In 2005, Zhang et al. [19] presented a two-dimensional simulation based on arbitrary Lagrangian–Eulerian (ALE) description using the ABAQUS commercial FEM code. Results
416
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Table 1 Chemical composition of LM13 (wt%).
Table 2 Thermal characteristics of the LM13 workpiece and H13 tool.
Element
Si
Cu
Mg
Ni
Fe
Mn
Al
Weight (%)
11.35
1.49
1.46
1.15
0.37
0.02
Bal.
Property
LM13
FSP tool H13
Backplate
Heat capacity (N/mm2 °C) Emissivity Conductivity (N/s °C) Heat transfer coefficient between tool and billet (N/°C s mm) Heat transfer coefficient between backing plate and billet (N/°C s mm) Heat transfer coefficient between tool/ workpiece and air (N/mm s °C)
2.57 0.7 117 11
3.24 0.7 24.5 11
– – – –
1
–
1
0.02
0.02
–
Fig. 1. Schematic illustration of force measurement setup during FSP.
Fig. 3. Simulation and experimental axial force.
showed that the shoulder has a significant effect on material flow pattern at the top surface, and this effect decreases by increasing the distance from the top surface. Liechty and Webb [21] in 2008 investigated various frictional boundary conditions on the material flow behavior. Their results showed that the shear model is superior to a sticking condition. Although a significant progress has been made in finding material flow during FSW/FSP, no systematic study has been carried out to numerically find all the main features of FSW/FSP. The current paper presents a fully coupled incremental Lagrangian thermo-mechanical simulation of FSP via employing the DEFORM-3D as commercial FEM software to predict the material flow pattern, temperature distribution, and effective plastic strain in the weld zone. Furthermore, temperature rise induced by the process, stir zone shape, the defect types and powder agglomeration were found experimentally and the results obtained by numerical simulation were compared with those acquired via experiment in order to validate the accuracy of the developed simulation. 2. Experimental procedures
Fig. 2. Flow stress–strain curves of base metal under different conditions.
demonstrated that the material has a revolving motion around the pin at the advancing side, whereas at the retreating side (RS) material does not revolve around the pin. In 2008, Zhang and Zhang [20] developed a fully coupled thermo-mechanical simulation and
The LM13 eutectic Al–Si as-cast alloy was subjected to FSP. The dimensions of plates were 100 mm in length, 50 mm in width and 7 mm in thickness. The chemical composition of LM13 alloy is given in Table 1. In order to recognize the material flow, graphite powder (with 5 lm particle size) was embedded in a groove with the depth and width of 3 and 1 mm, respectively. The FSP tool was a H13 hot working steel with a cylindrical pin. The shoulder and pin diameter, and pin height were 18, 6, and 4 mm, respectively. The tool rotational and the traverse speeds were also 900 rpm and 50 mm/min, respectively. The tilt angle was fixed at 3°. Subsequently, specimens were cut from friction stir processed (FSPed) samples and polished according to standard metallo-
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
417
Fig. 4. Schematic illustration of tool and workpiece.
machine bed and the back plate in this research, one charge amplifier was employed to convert the dynamometer charge signals to output voltages balanced to the forces obtained. The dynamometer was mounted on the bed of the machine using screws and a supporting block. The force data were obtained via a data acquisition system and plotted with respect to the time step of the process (Fig. 1). To attain the flow stress of the LM13 at various deformation conditions, hot compression test was also conducted at the temperature ranges of 300–500 °C and strain rates of 0.01 and 1 s1. Dimensions of compressive test samples were 15 mm in height and 10 mm in diameter (in accordance with ASTM E209 [26]). Fig. 5. Comparison of simulation and experimental temperature history at 9 mm from the center of weld and 2 mm under the top surface.
3. Model description Deform-3D software, which enables simulation of processes with large plastic deformation and point tracking, was employed to simulate the FSP. Moreover, to simplify the problem, the following assumptions were made in this study: (1) Material model of the workpiece is rigid-visco-plastic (RVP); (2) FSP tool and backing plate are rigid; (3) Friction factor is constant; (4) Thermal properties of the workpiece and FSP tool are constant; (5) Free surfaces of the workpiece and FSP tool were under free convection at ambient temperature of 20 °C; (6) The backing plate and two sides of workpieces were fixed in all degrees of freedom. The rigid visco-plastic FEM is generally used in the metal forming simulation and derived from principle of variational stability, which results in nonlinear equations. These equations can be solved by iterative methods (such as Newton–Raphson or direct method). Further details in governing equations can be found in Ref. [22]. In the FE model, tool and workpiece geometry and parameters (translational and rotational speeds and tilt angle) used in experiments were employed. 3.1. Material model
Fig. 6. Initial positions of points that have been selected for material flow in the stir zone.
graphic procedure, aiming to microstructural analysis. A thermocouple was also inserted in the workpiece at the distance of 9 mm from the center line and depth of 2 mm from the top surface at the AS. To measure the axial force (in the plunging direction), one-component Kistler dynamometer was fitted between the
The flow stress of LM13 cast Al alloy is given as a function of plastic strain, strain rate and temperature:
r ¼ r ðe; e_ ; TÞ
ð1Þ
is the flow stress, e is the effective plastic strain, e_ is the where r effective strain rate, and T is the temperature. Flow stress values are interpolated in log strain, log strain rate, and linear temperature space. For data points at 0.0 plastic strain, linear interpolation is used between the 0 value and the first non 0 values, and log interpolation is used for subsequent values.
418
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 7. Center line points flow pattern.
The relationship between flow stress and strain for LM13 alloy at temperature range of 300–500 °C and strain rates of 0.01 and 1 s1 are shown in Fig. 2. Thermal properties of the LM13 workpiece and H13 hot working steel are summarized in Table 2. 3.2. Friction model
the deforming body. To determine the best approximation for friction factor, the axial forces of FE modeling were calculated for three different friction factors, 0.4, 0.5, and 0.6, under 100 mm/min transverse speed and 900 rpm rotational speed. Friction factor of 0.5 exhibited the best match based on the comparison between the axial forces obtained by simulation and experimental measurement. Results are shown in Fig. 3.
Constant shear friction is used mostly for bulk-forming simulations. The frictional force in the constant shear model is defined by:
3.3. Workpiece and tool models
f ¼ mk
FSP tool was assumed to be a rigid body and was meshed with 11,300 tetrahedral elements. The workpieces were modeled as a temperature and strain rate dependent RVP material. The workpiece was meshed with 39,000 tetrahedral elements and
ð2Þ
where, f is the frictional stress; k is the shear yield stress; and m the friction factor. Thus, the friction is a function of the yield stress of
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
0.8 mm minimum element size. Automatic remeshing with a nonuniform mesh was applied to the workpiece. The smallest elements are located beneath the shoulder and the element size is increased by increasing the distance from the tool shoulder. Workpiece and tool meshes are schematically illustrated in Fig. 4. 4. Results and discussion 4.1. Numerical simulation and experimental verification of temperature
419
In order to validate the accuracy of the simulation results, the temperature distribution found via simulation were compared with the experimental results, as shown in Fig. 5. Both simulation and experimental results were reported at a distance of 9 mm from the weld center at the AS and 2 mm below the top surface. The comparison shows that although the numerical temperature results are slightly higher (19 °C) than those acquired by experiment, the simulation results are acceptably in good agreement with the experimental data. 4.2. Material flow
Since the flow stress is dictated by the temperature rise, heat generation during the process, which is a function of the plastic deformation and friction, it has an imperative role in material flow during FSW/FSP.
Fig. 6 shows the initial position of points selected to study material flow pattern in the SZ. Points P1–P8 are located 0.5 mm from each other in the weld center. Similarly, points P9–P16 and
Fig. 7 (continued)
420
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 8. Applied plastic strain to the points P1–P8.
P17–P24 are located at the advancing and retreating sides with the distance of 2.5 mm from the center line. Fig. 7 also shows the material flow at the weld center. Besides, Fig. 7a illustrates the initial position of points located along the center line and Fig. 6b–e track the aforementioned points during the process. As could be seen, the points near the bottom of the SZ (P7 and P8) have a semicircular movement from the leading side toward the trailing side. However, the part of material which is
Fig. 11. (a) Tunneling cavity formation at the behind of the pin and (b) slot formation at advancing side behind the pin.
close to the top surface, in spite of the movement to the tool pin behind, is stretched toward the advancing side. Such a behavior was clearly expected due to the frictional shear force induced by the tool shoulder. An interesting observation was that the final position of the points except point P1 is similar to SZ shape which will
Fig. 9. Material velocity (a) velocity contour in the cross section of the weld and (b) velocity diagram of points P1–P8.
Fig. 10. (a) Tunneling cavity formation at the behind of the pin and (b) slot formation at advancing side behind the pin.
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
421
Fig. 12. Advancing side points flow pattern.
be discussed. Point P1 is an exception, which is stuck to the shoulder and rotates with the shoulder for an unknown time period. Fig. 8 also indicates the plastic strain applied to the points P1– P8. Similar to the other works present in the literature, the plastic strain for the different points is about 30. Moreover, according to Fig. 8, the amount of effective plastic strain decreases upon moving to the SZ root. The effective plastic strain of point P1, which is stuck to the shoulder, is considerably higher than that of the other points, which can be attributed to the accumulation of the effective strain during several successive rotations around the pin. Furthermore, Fig. 9a and b indicate the velocity of points P1–P8 and velocity contour in the cross section. As was expected, velocity decreases upon increasing the distance from the top surface. Fig. 9b also reveals that the velocity at the RS is higher than that at the AS, which validates the results reported by Morisada et al. [8]. They employed X-ray transmission examination to show that velocity at the RS is slightly higher than that at the AS. Nevertheless, the amount of velocity reported by them is higher than that
obtained by the developed simulation in the current paper, which can be most probably attributed to the different friction conditions. Many researchers have discussed about frictional boundary condition [27–30] but still it is not clear that the friction is sticky or slippery. At the AS near the trailing side, the pin velocity has its own minimum value and due to the pressure reduction, the material barely flows into this region. This is the main reason for formation of tunneling cavity and thus the cavity mostly is formed at the AS near the trailing side [31,32]. The probability of tunneling cavity formation increases by increasing the traverse speed. Fig. 10a demonstrates the cavity that formed in the higher traverse speed (200 mm/min). As shown in Fig. 10b, a slot was created at the AS of trailing side. This slot and tunneling cavity was also visible in the experiment (Fig. 11). Flow patterns of points P9–P16 and P17–P26 are also shown in Figs. 12 and 13, respectively. As shown in Fig. 12, points initially located at the AS revolve around the pin and finally come back to the AS. However, this behavior is not observed for the points located at
422
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 12 (continued)
the RS and they are scattered over the AS and RS after passing of the tool. Moreover, the points near the SZ root remain at the RS and upon decreasing the distance from the top surface, material stretching toward the AS diminishes. In other words, the part of material near the tool shoulder at the AS, RS, and center line moves toward the AS. On the other words, most of the SZ materials are finally located in the AS after passing of the tool. This phenomenon, clearly shown in Fig. 14, can explicate the powder agglomeration at the AS near the top surface in the case of FSP with powder. This was obviously seen in the experiment. Fig. 15 reveals the graphite agglomeration occurred at the AS near the top surface. The agglomeration of embedded powder or marker at the AS has been widely reported in other works [33,34]. The shape of SZ can be predicted through using the point tracking option of DEFOM-3D software. Thus, in order to achieve a better understanding of material flow during the process, more points were picked inside the SZ before the tool passing (Fig. 16a). The final position of these points after tool passing is shown in Fig. 16b. The shape of SZ predicted by simulation was projected on Fig. 15
(SZ shape resulted from experiment) and the complex is shown in Fig. 17. It is revealed that the simulated SZ shape is in good agreement with the experiment. Fig. 18 opens new perspectives towards a better understanding of how materials move into the SZ. According to Fig. 18a, nine points at the RS (P1–P9) and ten points at the AS (P10–P19) were picked. Fig. 18b and c represent the final positions of the aforementioned points. As is seen, the points at AS remain in contact with the tool shoulder and rotate several times during the process. Nevertheless, the points at RS never revolve around the tool and are scattered over the SZ. Thus, it can be concluded that the main part of material inside the marked zone probably comes from the zone located at a certain distance from the tool pin and near the top surface at RS. Furthermore, the material at the AS travels a longer path compared to the material at the RS and the SZ center. Thus, it can be concluded that shifting the tool toward the RS can lead to a better powder distribution during the process since the higher traveling pass may result in a better powder distribution and mixing with the matrix. Fig. 19 is provided to validate this claim. Fig. 19
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
423
Fig. 13. Retreating side points flow pattern.
shows the graphite powder distribution in the FSPed specimen with 2 mm shifted tool toward the RS. As could be seen, shifting the tool toward the RS considerably led to a better powder distribution. Additionally, Mahmoud et al. [35] reported the same conclusion by experiment. 5. Conclusions FSP of a cast Al–Si alloy was simulated based on Lagrangian incremental with the aim of achieving a better understanding of martial behavior during the process. The SZ shape, forming defects, and effective plastic strain were also scrutinized and the simulation results were compared with those acquired via experiment.
The following results were achieved: 1. The SZ shape was successfully predicted. Deep analysis of the material flow revealed that the material near the top surface was stretched toward the AS due to the shoulder frictional force and upon increasing the distance from top surface the final position of the material remains almost constant in the transverse direction regardless of rotating around the pin. 2. Formation of tunneling defect occurs at the AS and near the root of the SZ. While powder agglomeration occurs at the AS and near the top surface. 3. Both simulation and experiment indicated that shifting the FSP tool toward the RS can result in a better powder distribution
424
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
Fig. 13 (continued)
Fig. 14. Aggregation of points at the AS in final position.
Fig. 15. Cross section macro-image of FSPed specimen.
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
425
Fig. 16. (a) Initial position of points picked for studying the SZ shape and (b) final position of points.
Fig. 17. Comparison of simulated and experimental SZ shape.
Fig. 19. Cross section macro-image of FSPed specimen with 2 mm tool shifting toward the RS.
Fig. 18. (a) Initial positions of points picked for specification of initial location of points which finally are located in the dashed line, (b) and (c) final position of points.
426
S. Tutunchilar et al. / Materials and Design 40 (2012) 415–426
References [1] Wang Feng, Liu Huimin, Ma Yajun, Jin Yuansheng. Effect of Si content on the dry sliding wear properties of spray-deposited Al–Si alloy. Mater Des 2004;25:163–6. [2] Özyürek Dursun, Aktar Niyazi, Aztekin Hasan. Design and construction of a thixo forming unit and production of Al–Si alloys. Mater Des 2008;29:1070–4. [3] Dwivedi DK, Sharma A, Rajan TV. Influence of silicon morphology and mechanical properties of piston alloys. Mater Manuf Processes 2005;20:777–91. [4] Karthikeyan L, Senthilkumar VS, Padmanabhan KA. On the role of process variables in the friction stir processing of cast aluminum A319 alloy. Mater Des 2010;31:761–71. [5] Ma ZY, Pilchak AL, Juhas MC, Williams JC. Microstructural refinement and property enhancement of cast light alloys via friction stir processing. Scr Mater 2008;58:361–6. [6] Azizieh M, Kokabi AH, Abachi P. Effect of rotational speed and probe profile on microstructure and hardness of AZ31/Al2O3 nanocomposites fabricated by friction stir processing. Mater Des 2011;32:2034–41. [7] Sharifitabar M, Sarani A, Khorshahian S, Shafiee Afarani M. Fabrication of 5052Al/Al2O3 nanoceramic particle reinforced composite via friction stir processing route. Mater Des 2011;32:4164–72. [8] Morisada Y, Fujii H, Kawahito Y, Nakata K, Tanaka M. Three-dimensional visualization of material flow during friction stir welding by two pairs of X-ray transmission systems. Scr Mater 2011;65:1085–8. [9] Colligan K. Material flow behaviour during friction stir welding of aluminum. Welding J 1999;78:229–37. [10] Schmidt HNB, Dickerson TL, Hattel JH. Material flow in butt friction stir welds in AA2024-T3. Acta Mater 2006;54:1199–209. [11] Morisada Y, Fujii H, Nagaoka T, Nogi K, Fukusumi M. Fullerene/A5083 composites fabricated by material flow during friction stir processing. Composites A 2007;38:2097–101. [12] Reynolds AP. Flow visualization and simulation in FSW. Scr Mater 2008;58:338–42. [13] Lorrain O, Favier V, Zahrouni H, Lawrjaniec D. Understanding the material flow path of friction stir welding process using unthreaded tools. J Mater Process Technol 2010;210:603–9. [14] Li Y, Murr LE, McClure JC. Flow visualization and residual microstructures associated with the friction-stir welding of 2024 aluminum–6061 aluminum. Mater Sci Eng, A 1999;271:213–23. [15] Guerra M, Schmidt C, McClure JC, Murr LE, Nunes AC. Flow patterns during friction stir welding. Mater Character 2003;49:95–101. [16] da Silva AAM, Arruti E, Janeiro G, Aldanondo E, Alvarez P, Echeverria A. Material flow and mechanical behaviour of dissimilar AA2024-T3 and AA7075T6 aluminium alloys friction stir welds. Mater Des 2011;32:2021–7. [17] Liechty BC, Webb BW. The use of plasticine as an analog to explore material flow in friction stir welding. J Mater Process Technol 2007;184:240–50.
[18] Zettler R, Donath T, dos Santos JF, Beckman F, Lohwasser D. Validation of marker material flow in 4 mm thick friction stir welded Al 2024-T351 through computer microtomography and dedicated metallographic techniques. Adv Eng Mater 2006;8:487–90. [19] Zhang HW, Zhang Z, Chen’ JT. The finite element simulation of the friction stir welding process. Mater Sci Eng, A 2005;403:340–8. [20] Zhang Z, Zhang HW. A fully coupled thermo-mechanical model of friction stir welding. Int J Adv Manuf Technol 2008;37:279–93. [21] Liechty BC, Webb BW. Modeling the frictional boundary condition in friction stir welding. Int J Mach Tools Manuf 2008;48:1474–85. [22] Asadi P, Mahdavinejad RA, Tutunchilar S. Simulation and experimental investigation of FSP of AZ91 magnesium alloy. Mater Sci Eng, A 2011;528:6469–77. [23] Chen CM, Kovacevic R. Finite element modeling of friction stir welding— thermal and thermomechanical analysis. Int J Mach Tools Manuf 2003;43:1319–26. [24] Zhang Z, Zhang HW. Numerical studies on the effect of transverse speed in friction stir welding. Mater Des 2009;30:900–7. [25] Assidi Mohamed, Fourment Lionel, Guerdoux Simon, Nelson Tracy. Friction model for friction stir welding process simulation: calibrations from welding experiments. Int J Mach Tools Manuf 2010;50:143–55. [26] ASTM standard E 209-00. Standard practice for compression tests of metallic materials at elevated temperatures with conventional or rapid heating rates and strain rates. In: Annual book of ASTM standards, vol. 03.01(July); 2000. [27] Guerra M, Schmidt C, McClure JC, Murr LE, Nunes AC. Flow patterns during friction stir welding. Mater Charact 2002;49:95–101. [28] Schmidt HNB, Dickerson TL, Hattel JH. Material flow in butt friction stir welds in AA2024-T3. Acta Mater 2006;54:1199–209. [29] Gerlich A, Yamamoto M, North TH. Strain rates and grain growth in Al 5754 and Al 6061 friction stir spot welds. Metall Mater Trans A 2007;38A:1291–302. [30] Gerlich A, Avramovic-Cingara G, North TH. Stir zone microstructure and strain rate during Al 7075-T6 friction stir spot welding. Metall Mater Trans A 2006;37A:2773–86. [31] Arbegast William J. A flow-partitioned deformation zone model for defect formation during friction stir welding. Scr Mater 2008;58:372–6. [32] Mishra RS, Ma ZY. Friction stir welding and processing. Mater Sci Eng R 2005;50:1–78. [33] Lorrain Olivier, Favier Véronique, Zahrouni Hamid, Lawrjaniec Didier. Understanding the material flow path of friction stir welding process using unthreaded tools. J Mater Process Technol 2010;210:603–9. [34] Asadi P, Faraji G, Masoumi A, Givi MKB. Experimental investigation of magnesium-base nanocomposite produced by friction stir processing: effects of particle types and number of friction stir processing passes. Metall Mater Trans A 2011;42A:2820–32. [35] Mahmoud ERI, Takahashi M, Shibayanagi T, Ikeuchi K. Effect of friction stir processing tool probe on fabrication of SiC particle reinforced composite on aluminium surface. Sci Technol Weld Joining 2009;14:413–25.