- Email: [email protected]
Mechanical characterization of flow formed FCC alloys
Author’s Accepted Manuscript Mechanical characterization of flow formed FCC alloys M. Haghshenas, R.J. Klassen
www.elsevier.com/locate/msea
PII: DOI: Reference:
S0921-5093(15)30101-5 http://dx.doi.org/10.1016/j.msea.2015.06.046 MSA32484
To appear in: Materials Science & Engineering A Received date: 30 April 2015 Revised date: 15 June 2015 Accepted date: 16 June 2015 Cite this article as: M. Haghshenas and R.J. Klassen, Mechanical characterization of flow formed FCC alloys, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2015.06.046 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Mechanical characterization of flow formed FCC alloys M. Haghshenas1,2,*, R.J. Klassen3 1
Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, Canada
2
Frank Hasenfratz Centre of Excellence in Manufacturing, Linamar Corporation, Guelph, Canada
3
Department of Mechanical and Materials Engineering, Western University, London, Canada
Abstract Flow forming is an old, but still promising, sheet metal forming technique to produce thinwalled, high-precision cylindrical auto-components (i.e. carriers and clutch housings). In the present research, the flow forming process was carried out on some commonly used face centered cubic (FCC) metals to assess the effect of processing parameters on the mechanical responses of materials during flow forming. To this end, flow forming operation with thickness reduction of 70% was performed on 5052 and 6061 aluminum alloys, 70/30 brass, and pure copper. Mechanical responses of starting (as received) materials and flow formed parts, i.e. tensile strength, yield strength, elongation, work hardening rate, and work hardening exponent were discussed for each material separately and the results were compared. Also, flow formed materials were compared in terms of stacking fault energies and twinnability and the effects of these were considered on flow formability of materials and their flow behavior in flow forming operation.
Key words: Flow forming; mechanical characterization; FCC; work hardening rate; twinning.
*
Corresponding author Emails: [email protected] [email protected] Phone: 519 515 000 ext. 32489 1
1. Introduction Flow forming is an incremental sheet forming technique in which the preform, in the form of a disk, is clamped rigidly against the mandrel by means of a tailstock and the shape of the mandrel bears the final profile of the desired product. During the process, both the mandrel and blank are rotated while the spinning tool (i.e. rollers) contacts the blank and progressively induces a change in its shape according to the profile of the mandrel (Fig. 1). In this way, the wall thickness is reduced as material is encouraged to flow mainly in the axial direction, along the mandrel’s wall, increasing the length of the work piece. This technique provides a significant increase in the strength of material, mainly due to strain hardening, and provides excellent dimensional accuracy and surface finish for producing seamless, complex, and high precision metal parts. Flow forming-made parts appear to be an attractive alternative to press formed parts especially with its lower forming load requiring considerable smaller equipment and more flexible tooling as compared to conventional forming processes. Experimental and analytical approaches have been used in the literature to assess mechanics and metallurgy of flow forming in aluminum alloys and copper. Haghshenas et al. [1], Chang et al. [2] Davidson et al. [3], Srinivasulu et al. [4], and Molladavoudi and Djavanroodi [5] studied the flow forming process parameters (i.e. feed rate, mandrel speed, attach angel) and flow formability of different aluminum alloys (i.e. AA2024, AA7075, AA6061, and AA5052) through experimental approach. Rao et al. [6] conducted experiments to study influence of flow forming parameters on surface finish and surface roughness of copper tubes. Pawlicki et al. [7] studied microhardness distribution in the flow formed copper. Zhang et al. [8] and Tang et al. [9] studied ball spin forming of pure copper through experimental and numerical methods and showed that numerical results agree well with the experimental results; the measures to prevent the folding proposed in their paper have been proved efficient in industrial application. Haghshenas et al. [10] studied the hardness distribution in the splined mandrel flow formed 70/30 brass and showed that the hardness magnitude is higher in the mandrel side as compared with the roller side. Despite mentioned (and other) publications on flow forming of aluminum and copper alloys, there are lack of systematic comparisons between post-flow forming mechanical properties of
2
different alloys and the role of structural characterization on flow forming responses. This paper, therefore, tries to analyze and compare the mechanical properties of various FCC alloys, subjected to an identical flow forming operation, to assess the effects of structural characterizations (i.e. staking fault energy, twinning, and work hardening rate) and different deformation mechanisms on the flow forming responses. To this end, metals that are known to deform by different slip twin ratios, and therefore respond differently to plastic deformation were subjected to flow forming operation.
2. Experimental procedure Flat discs, 210 mm diameter and 8.5 mm thickness, were cut from 5052 and 6061 aluminum alloys, 70/30 brass, and pure copper plates. The 6061 and 5052 aluminum alloys were annealed at 415 and 345°C for 2 to 3 hr to get 6061-O and 5052-O alloys. Table 1 shows chemical composition and grain sizes of the four fcc metals studied in this investigation. The dimensions of the preform were chosen according to the final required dimensions of the component (i.e. clutch housing). The discs were each subjected to a multi-pass single-roller flow forming operation over a cylindrical mandrel. Each disc was subjected to three flow forming passes with the work piece and mandrel rotating together at 300 rpm. The flow forming operations were carried out on a VSTR 150/3 Vertical Flow Forming Machine. This machine has been specially developed for the production of high precision cylindrical flow formed parts, i.e. clutch housings transmission components. The work piece was cooled and lubricated with an aqueous-based lubricant during each forming pass. The first two passes were identical for each sample and caused the flat disc to become formed around the cylindrical shape of the mandrel. The final flow forming pass pressed the work piece against the mandrel and endure a net change in thickness. Upon the flow forming process, in the present study, the overall thickness reduction of 70%, from 8.5 mm of initial blank to 2.5 mm of flow formed part, was carried out. In the flow forming, the flow of metal directly beneath the roller consists of two components, axial and circumferential. Tensile coupons, therefore, were machined in the axial and circumferential directions in the flow formed parts (Fig. 2), according to ASTM E8M standard [11]. Coupons were also made from the preform disks. Tensile tests were performed using a servo hydraulic INSTRON 8804 Universal Tensile Testing Machine equipped with a clip-on extensimeter at room temperature. The tensile tests were performed at a constant strain rate of
3
10−3 sec−1. For the sake of accuracy and repeatability, in each condition, three identical tensile tests were carried out. To study and compare the microstructures of starting disks and flow formed parts, routine metallography principles were performed on the as-received and flow formed materials and using an optical microscope (Nikon L150), the microstructural images were recorded.
3. Results and Discussion 3.1. Microstructure Flow forming tends to produce an elongated grain structure. The microstructure of starting disks of studied metals consists of equiaxed grains. Also, some annealing twins are observed in the microstructure of 70/30 brass (to a larger extent) and pure copper (to a lesser extent). However, severe plastic deformation, caused by flow forming, leads to grain refinement, grain orientation and grain flow in the axial direction (along mandrel’s axis). Figure 3 shows the starting and flow formed structures of tested alloys. As seen, grains follow a pancake and elongated pattern in the flow formed parts as compared with starting materials.
Also, some mechanical twins are
observed in the flow formed 70/30 brass (Fig. 3d). 3.2. Engineering stress/strain responses In the flow forming process, the evolution of microstructure eventually changes the mechanical properties which may influence the service performance of the flow formed work piece. It is, therefore, necessary to study the variation of mechanical properties between as received (preform) and flow formed components. Figure 4 shows engineering flow curves of the preform and the flow formed alloys, in the axial and circumferential directions. The tensile curves of flow formed specimens show limited uniform elongation, and peak immediately after yielding. This response is mainly due to severe cold working and strain hardening experienced by flow formed components. This suggests that cross-slip or climb of dislocations responsible for dynamic recovery is suppressed effectively during the deformation process. Work hardening, controlled by dislocation activity, is limited due to high dislocation density caused by flow forming-induced severe plastic deformation. Also, the magnitude of yield stress (YS) and ultimate tensile strength (UTS) in the axial direction of flow formed work pieces are larger than circumferential direction. This is in agreement with the
4
direction of grains’ flow. In a flow forming process, the material is displaced axially, along an axis parallel to the mandrel’s rotational axis; the grains become largely elongated axially in which length-to-width ratio of grains increases. That is, grains are highly deformed, elongated and work hardened in the axial direction when the rollers travel along the mandrel wall to deform the material during flow forming. In the circumferential direction, however, grains are relatively un-deformed (or less-deformed). In the flow curves of as received 5052 aluminum and 70/30 brass alloys, some serration phenomena are observed. Such an irregular behavior in the 5052 aluminum alloy is associated with Portevin-Le Chatelier (PLC) effect [12]. The serrations are attributed to the interaction of moving dislocations with fast diffusing solute atoms (i.e. Mg) and pinning/unpinning of dislocations creating raising/falling in the flow curve [13]. In 70/30 brass, as an alloy which is prone to twinning, the serration in the flow curve is attributed to the competition between twinning and slipping [14]. When the strength of twin boundaries reaches a critical value, they effectively impede the slip of dislocations. significant stress accumulations.
This causes abundant blocked slip bands and
When stress accumulates to a threshold, the blocked
dislocations will proceed along or through twin boundaries, leading to particularly decomposing of twin boundaries and releasing of stress. Summaries of variations in mechanical properties of tested metals are listed in Tables 2 and 3. Among the tested metals, the largest decrease in the elongation belongs to 70/30 brass with 92% decrease in the flow formed condition as compared with as received condition. This can be attributed to large density of mechanical twins in the flow formed 70/30 brass and their effect of elongation.
3.3. Strain hardening rate, θ The material model used in this study is an isotropic power law hardening that behaves during a standard tensile or compression test according to:
K n
(1)
where σ is the stress, K is the strength index or strength coefficient, is the plastic strain and n is the strain hardening exponent. The constant K is structure dependent and is influenced by
5
processing while n is a material property normally lying in the range 0.2–0.5. The strain hardening index can be described by:
n
log log
(2)
This equation can be evaluated from the slope of log vs. log plot. Rearranging allows a determination of the rate of strain hardening, , at a given stress and strain:
d n d
(3)
Strain hardening rate of a material is usually characterized indirectly from the stress-strain curves obtained in tensile tests. Typically, the strain hardening rate is computed numerically from the stress-strain curve and plotted against strain/ stress. In the present study, normalized strain hardening rate ( compare
G
G
), is plotted against strain, , as shown in Fig. 5 (G is shear modulus). To
at different stacking fault energy (SFE) values, 6061 and 5052 aluminum alloys
have also been included in Fig. 5. Accordingly, following equation can be written:
G 70/30brass G Cu
G Al 5052 G Al 6061
(4)
Among tested alloys, 70/30 brass shows a completely different rate hardening behavior (Fig. 5a) as compared with tested pure copper and aluminum alloys. Figure 5a clearly indicates four distinct regimes of strain hardening in the 70/30 brass [15-19]. Stage A is an initial sharp decrease in the work-hardening rate. This region is very similar in form to the starin hardening behaviour exhibited by the high-SFE polycrsytalline FCC metals in the recovery regime (stage III). A constant work-hardening rate, stage B, follows the initial work-hardening decrease. The onset of stage B is correlated with the initiation of deformation twinning. El- Danaf et al. [15, 16] and Asgari [20] reported that a critical dislocation density is required to initiate deformation twinning which is measured by a so called normalized stress parameter, ( 0 )
G
, (where σ is flow stress, σo is initial yield strength, and G is shear modulus). The
normalized stress parameter is believed to be an indicator that reflects the dislocation density
6
built inside materials during plastic deformation [16]. Figure 5a shows that the normalized strain hardening rate in as received 70/30 brass is about 0.02 to 0.03 in the stage B which is in good agreement with El-Danaf et al. [16] findings on the work hardening behavior on 70/30 brass. Stage C is characterized by a renewed decrease in the work-hardening rate. During this stage, the rate of primary twinning decreases. Finally, stage D is correlated with extensive twin intersection, brought about by the activation of a second twinning system non-coplanar with the previous one. As seen in Fig. 5a, normalized strain hardening rate is about 0.01 at the stage D in the tested low-SFE 70/30 brass. Among all test materials, 70/30 brass has the lowest SFE (7 mJ/m2). SFE influences the type of dislocation systems formed during deformation and determines the cross-slip intensity and dislocation climbing rates, which are considered to be the main dynamic recovery processes [21-24].
Note that the stage D strain hardening rate is
significantly larger than the steady state strain hardening rates exhibited by high-SFE aluminum 6061 and 5052 alloys (Fig. 5b). On the microscopic scale, the work hardening behavior of crystalline materials is controlled by the generation, density, motion and interactions between dislocations [25]. Faghihi and Voyiadjis stated that the flow stress of a material, in microscopic scales, can be written as summation of equivalent athermal stress and equivalent thermal stress [26, 27]. In the FCC metals, the thermal activation is strongly dependent on the plastic strain. Kocks and Mecking [28] suggested that the dislocation storage (hardening) and dislocation annihilation (dynamic recovery/softening) are added together to describe the dislocation density evolution. They hypothesized that the hardening component is geometric or statistical in nature and was, thus, athermal. Dynamic recovery (or the softening component), on the other hand, was strongly influenced by thermal activation and, accordingly, temperature and strain rate. Dynamic recovery causes a decrease in the hardening rate by reduction of the long-range internal stresses, either by annihilation of dislocations or by forming energetically lower dislocation systems. A full dislocation tends to split into two partials with a stacking fault ribbon between them in the low SFE alloys. The lower the SFE is, the greater the width of the stacking fault ribbon is due to the smaller repulsive force acting between them (i.e. the separation distance can be 5–10 atomic spacing in the case of pure Cu with SFE=78 mJ/m2 and one inter-atomic spacing in the case of aluminum with SFE=200 mJ/m2). Because of high SFE and activation of dynamic recovery, a much larger decrease in the work hardening rate is anticipated in the tested aluminum alloys as compared to
7
copper and 70/30 brass. In fact, it is difficult for a full dislocation to conduct cross slip and/or climb in the obstruction of stacking fault ribbon when it encounters obstacles, which inhibits the dislocation motions and the dynamic recovery via cross slip or climb. Consequently, a lower SFE usually leads to a higher strain hardening rate. Restricted climb and/or cross-slip results in higher dislocation density accumulated in 70/30 brass and therefore there is a limited capacity for it to store more dislocations during cold work (i.e. flow forming). 3.4. The effect of stacking fault energy (SFE) and work hardening rate (ɵ) A lower SFE makes it easier to deform via twinning which has been reported to cause simultaneous increase in strength and ductility and a higher strain hardening rate. [29-33]. Twining strengthening effect is caused by Hall-Petch effect and Basinski mechanism [32]. Mechanical twins in two ways affect the deformation [34]: the twinning shear makes a relatively small contribution to the deformation and the twin boundaries, which act as barriers to dislocation motion, reduce the dislocation mean free path resulting in a high strain hardening behavior. That is, as the stress increases, the volume fraction of twins increase steadily, continuously dividing grains into smaller units. It can be considered a dynamic Hall-Petch effect as the effective grain size is continuously being decreased. Therefore, the change in deformation with SFE (i.e. cross-slip or climb of dislocations is suppressed effectively [35]) is the primary cause for the observed enhanced strengthening in low-SFE brass compared with medium-SFE copper. Furthermore, it is well established that twin boundaries are effective in strengthening metals as they are counted as grain boundaries [36]. In FCC materials, twinning and slip occur through dislocation processes operating on the same set of crystallographic slip system and are thus in direct competition. The competition can be described by “twinability”, the ease with which a metal deforms by twinning in competing with dislocation mediated slip [37, 38]. The commercial Al alloys can, to a first approximation, be ordered in terms of twinability as follows [39]:
1xxx 4xxx 2xxx 6xxx 7 xxx 5xxx Alloying with Mg is expected to bring about significant increase in the tendency for formation of partial dislocations and for mechanical twinning. Considering the SFE of flow formed alloys in the present study, the twinnability can be ordered as follows:
8
70 / 30brass Copper 5052aluminum 6061aluminum
Therefore, in the high-SFE tested alloys (i.e. aluminum), dislocation cross slip is the dominant mode of plastic deformation. In the low-SFE tested materials (i.e. copper alloys) twinning is the main deformation mechanism. 3.5. The effect of work hardening exponent (n) The response of a metallic material to cold working is given by the n-value. The larger the magnitude of n-value, the greater the strain hardening for a given amount of plastic strain will be. Metals with a low n-value respond poorly to cold working (i.e. flow forming). By using Eq. 2, one can calculate work hardening exponent, n. In this study, the calculated n-values are equal to 0.48, 0.36, 0.29, and 0.22 for 70/30 brass, pure copper, 5052 aluminum and 6061 aluminum, respectively. That is, 70/30 brass ( n 0.48 ) can be given large plastic strain more easily as compared to 6061 aluminum alloy ( n 0.22 ). The uniform elongation can be explained by the necking instability condition. When the work hardening rate is equal to , uniform elongation halts and necking starts. Therefore, high uniform elongation requires high work hardening during tensile deformation leading to a higher instantaneous work hardening (i.e. a slow decrease in work hardening rate) and enhances the uniform elongation by retarding (postponing) local necking.
4. Summary The responses of different FCC materials during a flow forming operation were investigated in this study. The main findings of present study are as follows:
Yield stress and ultimate tensile strength in the axial direction of flow formed work pieces are larger than circumferential direction. This is in agreement with the direction of grains’ flow.
The tensile curves of all flow formed specimens, either axial or circumferential direction, show partial uniform elongation, and UTS immediately after yielding.
The largest drop in the elongation belongs to flow formed 70/30 brass with 92% decrease as compared with as received condition.
9
The rate hardening behavior of tested 70/30 brass was completely different with other tested materials.
PLC effect were observed in 5059 aluminum alloy and 70/30 brass. This is attributed to the Mg solute atoms and to the competition between twinning and slipping in 5059 aluminum alloy and 70/30 brass, respectively.
References 1. M. Haghshenas, J.T. Wood, R.J. Klassen, Materials Science and Engineering A 532 (2012) 287–294. 2. Xu Wen-chen, Shan De-bin, Wang Zhen-long, Yang Guo-ping, LÜ Yan, Kang Da-chang, Trans. Nonferrous Met. Soc. China 17 (2007) 1205-1211. 3. M.J. Davidson, K. Balasubramanian, G.R.N. Tagore, Journal of Materials Processing Technology 203 (2008) 321-325. 4. M. Srinivasulu, M. Komaraiah, C.S. Krishna Prasada Rao, Journal of Mechanical Engineering Research 4(6), 192-198, 2012. 5. Hamid R. Molladavoudi, Faramarz Djavanroodi, Int J Adv Manuf Technol (2011) 52: 949-957. 6. Lakshman Rao M, Rao TVLN, Rao CSKP, J. Manuf. Eng. 5 (2009) 69-71. 7. M. Pawlicki, W.Bbochniak, Archives of Metallurgy and Materials 59 (2014) 537-543. 8. Guang-Liang Zhang, Shi-Hong Zhang, Bing Li, and Hai-Qu Zhang, Journal of Materials Processing Technology, 184, 393–400, (2007) 9. Tang, Y., Chi, Y., Chen, J.Ch., Deng, X.X., Liu, L., Liu, X.K., Wan, Z.P., International Journal of Machine Tools and Manufacture 47 (2007) 1059-1068. 10. M. Haghshenas, J.T. Wood, R.J. Klassen, Materials Science and Engineering A 552 (2012) 376–383. 11. Standard Test Methods for Tension Testing of Metallic Materials, Designation: E8/E8M–09, ASTM international. 12. A. Portevin, Le Chatelier FF., Comptes Rendus de l'Académie des Sciences, Paris 176 (1923) 507. 13. E.L. Huskins, B. Cao, K.T. Ramesh, Materials Science and Engineering A 527 (2010) 1292–1298. 14. S. Zhao, X. Li, L. Rong, Materials Letters 65 (2011) 2388-2390. 15. E.A. El-Danaf, S.R. Kalidinidi, R.G Doherty, Metallurgical and Materials Transaction 30A, (1999), 1223-1233. 16. E.A. El-Danaf, A. Al-Mutlaq, M.S. Soliman, Materials Science and Engineering A 528 (2011) 75797588. 17. Farzad Hamdi, Sirous Asgari, Scripta Materialia 62 (2010) 693–696. 18. R.P. Peng, M. Oden, Y.D. Wang, S. Johansson, Materials Science and Engineering A 334 (2002) 215222. 19. A.A. Salem, S.R. Kalidindi, R.D. Doherty, Acta Mater. 51 (2003) 4225-4237.
10
20. S. Asgari, E. El-Danaf, S.R. Kalidindi, R.D. Doherty, Metallurgical and Materials Transactions A 28 (1997) 1781-1795. 21. V. Subramanya Sarma, J. Wang, W.W. Jian, A. Kauffmann, H. Conrad, J. Freudenberger, Y.T. Zhu, Materials Science and Engineering A 527 (2010) 7624-7630. 22. S. Asgari, E. El-Danaf, S.R. Kalidindi, R.D. Doherty, Metallurgical and Materials Transactions A 28 (1997) 1781-1795. 23. J. Gryziecki, Z. Gdula, Material Science Engineering 93 (1987), 99-105. 24. A.A. Elmustafa, D.S. Stone, Materials Science and Engineering A 358 (2003) 1-8. 25. D. Faghihi, G.Z. Voyiadjis, T. Park, Journal of Engineering Materials and Technology 135, 021003-1021003-17. 26. D. Faghihi, G.Z. Voyiadjis, Mechanics of Materials 44 (2012) 189–211. 27. D. Faghihi, G.Z. Voyiadjis, Proc. IMechE Vol. 225 Part N: J. Nanoengineering and Nanosystems 515. 28. H. Mecking, U.F. Kocks, Acta Metall (1981) 29, 1865-1875. 29. Y.H. Zhao, Y.T. Zhu, X.Z. Liao, Z. Horita, T.G. Langdon, Appl. Phys. Lett. 89, 121906, (2006). 30. Yong-Hao Zhao, John F. Bingert, Xiao-Zhou Liao, Bao-Zhi Cui, Ke Han, Alla V. Sergueeva, Amiya K. Mukherjee, Ruslan Z. Valiev, Terence G. Langdon, Yuntian T. Zhu, Adv. Mater. 18 (2006) 2949– 2953. 31. P. Yang, H. Yang, J. Tao, C. Li, L. Shen, X.K. Zhu, Materials Science Forum 667-669 (2011) 10031008. 32. S.R. Kalidindi, A.A. Salem, R.D. Doherty, Advanced Engineering Materials 5, (2003), 229-232. 33. Y.H. Zhao, Y.T. Zhu, X.Z. Liao, Z. Horita, T.G. Langdon, Applied Physics Letters 89, 121906 (2006). 34. Meyers M.A., Vohringer O., Lubarda V.A. (2001), Acta Mater. 49, 4025 35. H. Bahmanpour, A. Kauffmann, M.S. Khoshkhoo, K.M. Youssef, S. Mula, J. Freudenberger, J. Ecker, R.O. Scattergood, C.C. Koch, Materials Science and Engineering A 529 (2011) 230– 236. 36. V.S. Sarma, J. Wang, W.W. Jian, A. Kauffmann, H. Conrad, J. Freudenberger, Y.T. Zhu, Materials Science and Engineering A 527 (2010) 7624-7630. 37. B.Q. Li, M.L. Sui, S.X. Mao, Journal of Material Science Technology 27 (2011) 97. 38. E.B. Tadmor, N. Bernstein, Journal of the Mechanics and Physics of Solids 52 (2004) 2507. 39. M. Muzyk, Z. Pakiela and K.J. Kurzydlowski, Scripta Materialia 64 (2011) 916. 40. C.-M. Kuo, C.-H. Tso, C.-H. Lin, Mater. Sci. Eng. A 519 (2009) 32.
11
Table 1: Chemical composition and grain size of the four fcc metals studied in this investigation Test material
Composition
70/30 brass
70% Cu, 30%Zn
Average grain size (m) 75
Pure copper
99.9% Cu
25
5052 Aluminum
2.2% Mg, 0.40% Fe, 0.25% Si, 0.17% Cr, 0.1% Mn, 0.1% Cu, 0.1% Zn, Bal. Al
52 [40]
6061 Aluminum
1.0% Mg, 0.6% Si, 0.4% Fe, 0.4% Cu, 0.3% Cr, 0.3% Zn, 0.1% Mn, 0.1% Ti, Bal. Al
100
Table 2: YS and UTS values in tested alloys in the as received and flow formed conditions
Materi al & proper ties
As recei ved
5052 aluminu m
80
6061 aluminu m
80
Pure copper
65
70/30 brass
100
YS (MPa) Flow formed
As recei ved
UTS (MPa) Flow formed
Change in YS As received to flow formed
Change in UTS As received to flow formed
Axi
Circumfer
Axi
Circumfer
Axi
Circumfer
Axi
Circumfer
al
ential
al
ential
al
ential
al
ential
27
240
30
293
70 %
67%
36 %
33%
195
58 %
53%
23 %
15%
305
80 %
76%
37 %
28%
398
79 %
74%
31 %
12%
195
0 19
5 170
165
0 33
5 270
220
0 48
21
35 0
380
0
350
50 5
Table 3: Elongation percent El (%) in the as received and flow formed conditions and the change in the El (%) caused by flow forming
Material &
As received
El (%) Flow formed
12
Change in El (%)
As received to flow formed
properties
Axial
Circumferential
Axial
Circumferential
5052 aluminum
24
3.2
7.4
-86.7%
-69.2%
6061 aluminum
18
6.3
9.6
-65%
-46.7%
Pure copper
43
5.9
7.9
-86.3%
-81.6%
70/30 brass
70
4.3
5.5
-93.8%
-92.1%
Roller
Flow formed work piece
Preform
Forming direction
Mandrel Figure 1: Schematic representation of three roller flow forming process.
13
Figure 2: Schematic diagram of tensile specimens machined from flow formed parts.
14
(a)
a
b
15
(c)
c (d)
d Figure 3: Starting (left side) and flow formed (right side) grain structures in a) 6061 aluminum alloy, b) 5052 aluminum alloy [40], c) pure copper, and d) 70/30 brass. As seen, grains are elongated axially along mandrels axis in the flow formed parts. White arrow shows the direction of forming.
16
250
6061 alumunim_As recieved 6061 alumunim_Flow Formed_Axial 6061 alumunim_Flow Formed_Circumferential
Stress (MPa)
200
150
100
50
0 0
5
10
15
20
Strain a 350
5052 aluminum_As recieved 5052 aluminum_Flow Formed_Axial
300
5052 aluminum_Flow Formed_Circumferential
Stress (MPa)
250
200
150
100
50
0 0
5
10
15
20
Strain b
17
25
30
600
Brass 70/30_As recieved Brass 70/30_Flow Formed_Axial Brass 70/30_Flow Formed_Circumferential
Stress (MPa)
500
400
300
200
100
0 0
20
40
60
80
Strain c 400
Cu_As recieved Cu_Flow Formed_Axial Cu_Flow Formed_Circumferential
Stress (MPa)
300
200
100
0 0
10
20
30
Strain d
18
40
50
Figure 4: Tensile engineering stress/strain curves for as received and flow formed (in axial and circular directions) for the tested materials, a) 6061 aluminum, b) 5052 aluminum, c) 70/30 brass, and d) pure copper.
a
19
b Figure 5: Normalized work hardening rate ( G ) versus strain for 5052 and 6061 aluminum alloys. The presence of Mg solute atoms may lead to the appearance of plastic instabilities resulting in the well-known serrated yielding or Portevin-Le Chatelier (PLC) effect which marks itself as steps in the G curve at strains bigger that 2%.
20
www.elsevier.com/locate/msea
PII: DOI: Reference:
S0921-5093(15)30101-5 http://dx.doi.org/10.1016/j.msea.2015.06.046 MSA32484
To appear in: Materials Science & Engineering A Received date: 30 April 2015 Revised date: 15 June 2015 Accepted date: 16 June 2015 Cite this article as: M. Haghshenas and R.J. Klassen, Mechanical characterization of flow formed FCC alloys, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2015.06.046 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Mechanical characterization of flow formed FCC alloys M. Haghshenas1,2,*, R.J. Klassen3 1
Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, Canada
2
Frank Hasenfratz Centre of Excellence in Manufacturing, Linamar Corporation, Guelph, Canada
3
Department of Mechanical and Materials Engineering, Western University, London, Canada
Abstract Flow forming is an old, but still promising, sheet metal forming technique to produce thinwalled, high-precision cylindrical auto-components (i.e. carriers and clutch housings). In the present research, the flow forming process was carried out on some commonly used face centered cubic (FCC) metals to assess the effect of processing parameters on the mechanical responses of materials during flow forming. To this end, flow forming operation with thickness reduction of 70% was performed on 5052 and 6061 aluminum alloys, 70/30 brass, and pure copper. Mechanical responses of starting (as received) materials and flow formed parts, i.e. tensile strength, yield strength, elongation, work hardening rate, and work hardening exponent were discussed for each material separately and the results were compared. Also, flow formed materials were compared in terms of stacking fault energies and twinnability and the effects of these were considered on flow formability of materials and their flow behavior in flow forming operation.
Key words: Flow forming; mechanical characterization; FCC; work hardening rate; twinning.
*
Corresponding author Emails: [email protected] [email protected] Phone: 519 515 000 ext. 32489 1
1. Introduction Flow forming is an incremental sheet forming technique in which the preform, in the form of a disk, is clamped rigidly against the mandrel by means of a tailstock and the shape of the mandrel bears the final profile of the desired product. During the process, both the mandrel and blank are rotated while the spinning tool (i.e. rollers) contacts the blank and progressively induces a change in its shape according to the profile of the mandrel (Fig. 1). In this way, the wall thickness is reduced as material is encouraged to flow mainly in the axial direction, along the mandrel’s wall, increasing the length of the work piece. This technique provides a significant increase in the strength of material, mainly due to strain hardening, and provides excellent dimensional accuracy and surface finish for producing seamless, complex, and high precision metal parts. Flow forming-made parts appear to be an attractive alternative to press formed parts especially with its lower forming load requiring considerable smaller equipment and more flexible tooling as compared to conventional forming processes. Experimental and analytical approaches have been used in the literature to assess mechanics and metallurgy of flow forming in aluminum alloys and copper. Haghshenas et al. [1], Chang et al. [2] Davidson et al. [3], Srinivasulu et al. [4], and Molladavoudi and Djavanroodi [5] studied the flow forming process parameters (i.e. feed rate, mandrel speed, attach angel) and flow formability of different aluminum alloys (i.e. AA2024, AA7075, AA6061, and AA5052) through experimental approach. Rao et al. [6] conducted experiments to study influence of flow forming parameters on surface finish and surface roughness of copper tubes. Pawlicki et al. [7] studied microhardness distribution in the flow formed copper. Zhang et al. [8] and Tang et al. [9] studied ball spin forming of pure copper through experimental and numerical methods and showed that numerical results agree well with the experimental results; the measures to prevent the folding proposed in their paper have been proved efficient in industrial application. Haghshenas et al. [10] studied the hardness distribution in the splined mandrel flow formed 70/30 brass and showed that the hardness magnitude is higher in the mandrel side as compared with the roller side. Despite mentioned (and other) publications on flow forming of aluminum and copper alloys, there are lack of systematic comparisons between post-flow forming mechanical properties of
2
different alloys and the role of structural characterization on flow forming responses. This paper, therefore, tries to analyze and compare the mechanical properties of various FCC alloys, subjected to an identical flow forming operation, to assess the effects of structural characterizations (i.e. staking fault energy, twinning, and work hardening rate) and different deformation mechanisms on the flow forming responses. To this end, metals that are known to deform by different slip twin ratios, and therefore respond differently to plastic deformation were subjected to flow forming operation.
2. Experimental procedure Flat discs, 210 mm diameter and 8.5 mm thickness, were cut from 5052 and 6061 aluminum alloys, 70/30 brass, and pure copper plates. The 6061 and 5052 aluminum alloys were annealed at 415 and 345°C for 2 to 3 hr to get 6061-O and 5052-O alloys. Table 1 shows chemical composition and grain sizes of the four fcc metals studied in this investigation. The dimensions of the preform were chosen according to the final required dimensions of the component (i.e. clutch housing). The discs were each subjected to a multi-pass single-roller flow forming operation over a cylindrical mandrel. Each disc was subjected to three flow forming passes with the work piece and mandrel rotating together at 300 rpm. The flow forming operations were carried out on a VSTR 150/3 Vertical Flow Forming Machine. This machine has been specially developed for the production of high precision cylindrical flow formed parts, i.e. clutch housings transmission components. The work piece was cooled and lubricated with an aqueous-based lubricant during each forming pass. The first two passes were identical for each sample and caused the flat disc to become formed around the cylindrical shape of the mandrel. The final flow forming pass pressed the work piece against the mandrel and endure a net change in thickness. Upon the flow forming process, in the present study, the overall thickness reduction of 70%, from 8.5 mm of initial blank to 2.5 mm of flow formed part, was carried out. In the flow forming, the flow of metal directly beneath the roller consists of two components, axial and circumferential. Tensile coupons, therefore, were machined in the axial and circumferential directions in the flow formed parts (Fig. 2), according to ASTM E8M standard [11]. Coupons were also made from the preform disks. Tensile tests were performed using a servo hydraulic INSTRON 8804 Universal Tensile Testing Machine equipped with a clip-on extensimeter at room temperature. The tensile tests were performed at a constant strain rate of
3
10−3 sec−1. For the sake of accuracy and repeatability, in each condition, three identical tensile tests were carried out. To study and compare the microstructures of starting disks and flow formed parts, routine metallography principles were performed on the as-received and flow formed materials and using an optical microscope (Nikon L150), the microstructural images were recorded.
3. Results and Discussion 3.1. Microstructure Flow forming tends to produce an elongated grain structure. The microstructure of starting disks of studied metals consists of equiaxed grains. Also, some annealing twins are observed in the microstructure of 70/30 brass (to a larger extent) and pure copper (to a lesser extent). However, severe plastic deformation, caused by flow forming, leads to grain refinement, grain orientation and grain flow in the axial direction (along mandrel’s axis). Figure 3 shows the starting and flow formed structures of tested alloys. As seen, grains follow a pancake and elongated pattern in the flow formed parts as compared with starting materials.
Also, some mechanical twins are
observed in the flow formed 70/30 brass (Fig. 3d). 3.2. Engineering stress/strain responses In the flow forming process, the evolution of microstructure eventually changes the mechanical properties which may influence the service performance of the flow formed work piece. It is, therefore, necessary to study the variation of mechanical properties between as received (preform) and flow formed components. Figure 4 shows engineering flow curves of the preform and the flow formed alloys, in the axial and circumferential directions. The tensile curves of flow formed specimens show limited uniform elongation, and peak immediately after yielding. This response is mainly due to severe cold working and strain hardening experienced by flow formed components. This suggests that cross-slip or climb of dislocations responsible for dynamic recovery is suppressed effectively during the deformation process. Work hardening, controlled by dislocation activity, is limited due to high dislocation density caused by flow forming-induced severe plastic deformation. Also, the magnitude of yield stress (YS) and ultimate tensile strength (UTS) in the axial direction of flow formed work pieces are larger than circumferential direction. This is in agreement with the
4
direction of grains’ flow. In a flow forming process, the material is displaced axially, along an axis parallel to the mandrel’s rotational axis; the grains become largely elongated axially in which length-to-width ratio of grains increases. That is, grains are highly deformed, elongated and work hardened in the axial direction when the rollers travel along the mandrel wall to deform the material during flow forming. In the circumferential direction, however, grains are relatively un-deformed (or less-deformed). In the flow curves of as received 5052 aluminum and 70/30 brass alloys, some serration phenomena are observed. Such an irregular behavior in the 5052 aluminum alloy is associated with Portevin-Le Chatelier (PLC) effect [12]. The serrations are attributed to the interaction of moving dislocations with fast diffusing solute atoms (i.e. Mg) and pinning/unpinning of dislocations creating raising/falling in the flow curve [13]. In 70/30 brass, as an alloy which is prone to twinning, the serration in the flow curve is attributed to the competition between twinning and slipping [14]. When the strength of twin boundaries reaches a critical value, they effectively impede the slip of dislocations. significant stress accumulations.
This causes abundant blocked slip bands and
When stress accumulates to a threshold, the blocked
dislocations will proceed along or through twin boundaries, leading to particularly decomposing of twin boundaries and releasing of stress. Summaries of variations in mechanical properties of tested metals are listed in Tables 2 and 3. Among the tested metals, the largest decrease in the elongation belongs to 70/30 brass with 92% decrease in the flow formed condition as compared with as received condition. This can be attributed to large density of mechanical twins in the flow formed 70/30 brass and their effect of elongation.
3.3. Strain hardening rate, θ The material model used in this study is an isotropic power law hardening that behaves during a standard tensile or compression test according to:
K n
(1)
where σ is the stress, K is the strength index or strength coefficient, is the plastic strain and n is the strain hardening exponent. The constant K is structure dependent and is influenced by
5
processing while n is a material property normally lying in the range 0.2–0.5. The strain hardening index can be described by:
n
log log
(2)
This equation can be evaluated from the slope of log vs. log plot. Rearranging allows a determination of the rate of strain hardening, , at a given stress and strain:
d n d
(3)
Strain hardening rate of a material is usually characterized indirectly from the stress-strain curves obtained in tensile tests. Typically, the strain hardening rate is computed numerically from the stress-strain curve and plotted against strain/ stress. In the present study, normalized strain hardening rate ( compare
G
G
), is plotted against strain, , as shown in Fig. 5 (G is shear modulus). To
at different stacking fault energy (SFE) values, 6061 and 5052 aluminum alloys
have also been included in Fig. 5. Accordingly, following equation can be written:
G 70/30brass G Cu
G Al 5052 G Al 6061
(4)
Among tested alloys, 70/30 brass shows a completely different rate hardening behavior (Fig. 5a) as compared with tested pure copper and aluminum alloys. Figure 5a clearly indicates four distinct regimes of strain hardening in the 70/30 brass [15-19]. Stage A is an initial sharp decrease in the work-hardening rate. This region is very similar in form to the starin hardening behaviour exhibited by the high-SFE polycrsytalline FCC metals in the recovery regime (stage III). A constant work-hardening rate, stage B, follows the initial work-hardening decrease. The onset of stage B is correlated with the initiation of deformation twinning. El- Danaf et al. [15, 16] and Asgari [20] reported that a critical dislocation density is required to initiate deformation twinning which is measured by a so called normalized stress parameter, ( 0 )
G
, (where σ is flow stress, σo is initial yield strength, and G is shear modulus). The
normalized stress parameter is believed to be an indicator that reflects the dislocation density
6
built inside materials during plastic deformation [16]. Figure 5a shows that the normalized strain hardening rate in as received 70/30 brass is about 0.02 to 0.03 in the stage B which is in good agreement with El-Danaf et al. [16] findings on the work hardening behavior on 70/30 brass. Stage C is characterized by a renewed decrease in the work-hardening rate. During this stage, the rate of primary twinning decreases. Finally, stage D is correlated with extensive twin intersection, brought about by the activation of a second twinning system non-coplanar with the previous one. As seen in Fig. 5a, normalized strain hardening rate is about 0.01 at the stage D in the tested low-SFE 70/30 brass. Among all test materials, 70/30 brass has the lowest SFE (7 mJ/m2). SFE influences the type of dislocation systems formed during deformation and determines the cross-slip intensity and dislocation climbing rates, which are considered to be the main dynamic recovery processes [21-24].
Note that the stage D strain hardening rate is
significantly larger than the steady state strain hardening rates exhibited by high-SFE aluminum 6061 and 5052 alloys (Fig. 5b). On the microscopic scale, the work hardening behavior of crystalline materials is controlled by the generation, density, motion and interactions between dislocations [25]. Faghihi and Voyiadjis stated that the flow stress of a material, in microscopic scales, can be written as summation of equivalent athermal stress and equivalent thermal stress [26, 27]. In the FCC metals, the thermal activation is strongly dependent on the plastic strain. Kocks and Mecking [28] suggested that the dislocation storage (hardening) and dislocation annihilation (dynamic recovery/softening) are added together to describe the dislocation density evolution. They hypothesized that the hardening component is geometric or statistical in nature and was, thus, athermal. Dynamic recovery (or the softening component), on the other hand, was strongly influenced by thermal activation and, accordingly, temperature and strain rate. Dynamic recovery causes a decrease in the hardening rate by reduction of the long-range internal stresses, either by annihilation of dislocations or by forming energetically lower dislocation systems. A full dislocation tends to split into two partials with a stacking fault ribbon between them in the low SFE alloys. The lower the SFE is, the greater the width of the stacking fault ribbon is due to the smaller repulsive force acting between them (i.e. the separation distance can be 5–10 atomic spacing in the case of pure Cu with SFE=78 mJ/m2 and one inter-atomic spacing in the case of aluminum with SFE=200 mJ/m2). Because of high SFE and activation of dynamic recovery, a much larger decrease in the work hardening rate is anticipated in the tested aluminum alloys as compared to
7
copper and 70/30 brass. In fact, it is difficult for a full dislocation to conduct cross slip and/or climb in the obstruction of stacking fault ribbon when it encounters obstacles, which inhibits the dislocation motions and the dynamic recovery via cross slip or climb. Consequently, a lower SFE usually leads to a higher strain hardening rate. Restricted climb and/or cross-slip results in higher dislocation density accumulated in 70/30 brass and therefore there is a limited capacity for it to store more dislocations during cold work (i.e. flow forming). 3.4. The effect of stacking fault energy (SFE) and work hardening rate (ɵ) A lower SFE makes it easier to deform via twinning which has been reported to cause simultaneous increase in strength and ductility and a higher strain hardening rate. [29-33]. Twining strengthening effect is caused by Hall-Petch effect and Basinski mechanism [32]. Mechanical twins in two ways affect the deformation [34]: the twinning shear makes a relatively small contribution to the deformation and the twin boundaries, which act as barriers to dislocation motion, reduce the dislocation mean free path resulting in a high strain hardening behavior. That is, as the stress increases, the volume fraction of twins increase steadily, continuously dividing grains into smaller units. It can be considered a dynamic Hall-Petch effect as the effective grain size is continuously being decreased. Therefore, the change in deformation with SFE (i.e. cross-slip or climb of dislocations is suppressed effectively [35]) is the primary cause for the observed enhanced strengthening in low-SFE brass compared with medium-SFE copper. Furthermore, it is well established that twin boundaries are effective in strengthening metals as they are counted as grain boundaries [36]. In FCC materials, twinning and slip occur through dislocation processes operating on the same set of crystallographic slip system and are thus in direct competition. The competition can be described by “twinability”, the ease with which a metal deforms by twinning in competing with dislocation mediated slip [37, 38]. The commercial Al alloys can, to a first approximation, be ordered in terms of twinability as follows [39]:
1xxx 4xxx 2xxx 6xxx 7 xxx 5xxx Alloying with Mg is expected to bring about significant increase in the tendency for formation of partial dislocations and for mechanical twinning. Considering the SFE of flow formed alloys in the present study, the twinnability can be ordered as follows:
8
70 / 30brass Copper 5052aluminum 6061aluminum
Therefore, in the high-SFE tested alloys (i.e. aluminum), dislocation cross slip is the dominant mode of plastic deformation. In the low-SFE tested materials (i.e. copper alloys) twinning is the main deformation mechanism. 3.5. The effect of work hardening exponent (n) The response of a metallic material to cold working is given by the n-value. The larger the magnitude of n-value, the greater the strain hardening for a given amount of plastic strain will be. Metals with a low n-value respond poorly to cold working (i.e. flow forming). By using Eq. 2, one can calculate work hardening exponent, n. In this study, the calculated n-values are equal to 0.48, 0.36, 0.29, and 0.22 for 70/30 brass, pure copper, 5052 aluminum and 6061 aluminum, respectively. That is, 70/30 brass ( n 0.48 ) can be given large plastic strain more easily as compared to 6061 aluminum alloy ( n 0.22 ). The uniform elongation can be explained by the necking instability condition. When the work hardening rate is equal to , uniform elongation halts and necking starts. Therefore, high uniform elongation requires high work hardening during tensile deformation leading to a higher instantaneous work hardening (i.e. a slow decrease in work hardening rate) and enhances the uniform elongation by retarding (postponing) local necking.
4. Summary The responses of different FCC materials during a flow forming operation were investigated in this study. The main findings of present study are as follows:
Yield stress and ultimate tensile strength in the axial direction of flow formed work pieces are larger than circumferential direction. This is in agreement with the direction of grains’ flow.
The tensile curves of all flow formed specimens, either axial or circumferential direction, show partial uniform elongation, and UTS immediately after yielding.
The largest drop in the elongation belongs to flow formed 70/30 brass with 92% decrease as compared with as received condition.
9
The rate hardening behavior of tested 70/30 brass was completely different with other tested materials.
PLC effect were observed in 5059 aluminum alloy and 70/30 brass. This is attributed to the Mg solute atoms and to the competition between twinning and slipping in 5059 aluminum alloy and 70/30 brass, respectively.
References 1. M. Haghshenas, J.T. Wood, R.J. Klassen, Materials Science and Engineering A 532 (2012) 287–294. 2. Xu Wen-chen, Shan De-bin, Wang Zhen-long, Yang Guo-ping, LÜ Yan, Kang Da-chang, Trans. Nonferrous Met. Soc. China 17 (2007) 1205-1211. 3. M.J. Davidson, K. Balasubramanian, G.R.N. Tagore, Journal of Materials Processing Technology 203 (2008) 321-325. 4. M. Srinivasulu, M. Komaraiah, C.S. Krishna Prasada Rao, Journal of Mechanical Engineering Research 4(6), 192-198, 2012. 5. Hamid R. Molladavoudi, Faramarz Djavanroodi, Int J Adv Manuf Technol (2011) 52: 949-957. 6. Lakshman Rao M, Rao TVLN, Rao CSKP, J. Manuf. Eng. 5 (2009) 69-71. 7. M. Pawlicki, W.Bbochniak, Archives of Metallurgy and Materials 59 (2014) 537-543. 8. Guang-Liang Zhang, Shi-Hong Zhang, Bing Li, and Hai-Qu Zhang, Journal of Materials Processing Technology, 184, 393–400, (2007) 9. Tang, Y., Chi, Y., Chen, J.Ch., Deng, X.X., Liu, L., Liu, X.K., Wan, Z.P., International Journal of Machine Tools and Manufacture 47 (2007) 1059-1068. 10. M. Haghshenas, J.T. Wood, R.J. Klassen, Materials Science and Engineering A 552 (2012) 376–383. 11. Standard Test Methods for Tension Testing of Metallic Materials, Designation: E8/E8M–09, ASTM international. 12. A. Portevin, Le Chatelier FF., Comptes Rendus de l'Académie des Sciences, Paris 176 (1923) 507. 13. E.L. Huskins, B. Cao, K.T. Ramesh, Materials Science and Engineering A 527 (2010) 1292–1298. 14. S. Zhao, X. Li, L. Rong, Materials Letters 65 (2011) 2388-2390. 15. E.A. El-Danaf, S.R. Kalidinidi, R.G Doherty, Metallurgical and Materials Transaction 30A, (1999), 1223-1233. 16. E.A. El-Danaf, A. Al-Mutlaq, M.S. Soliman, Materials Science and Engineering A 528 (2011) 75797588. 17. Farzad Hamdi, Sirous Asgari, Scripta Materialia 62 (2010) 693–696. 18. R.P. Peng, M. Oden, Y.D. Wang, S. Johansson, Materials Science and Engineering A 334 (2002) 215222. 19. A.A. Salem, S.R. Kalidindi, R.D. Doherty, Acta Mater. 51 (2003) 4225-4237.
10
20. S. Asgari, E. El-Danaf, S.R. Kalidindi, R.D. Doherty, Metallurgical and Materials Transactions A 28 (1997) 1781-1795. 21. V. Subramanya Sarma, J. Wang, W.W. Jian, A. Kauffmann, H. Conrad, J. Freudenberger, Y.T. Zhu, Materials Science and Engineering A 527 (2010) 7624-7630. 22. S. Asgari, E. El-Danaf, S.R. Kalidindi, R.D. Doherty, Metallurgical and Materials Transactions A 28 (1997) 1781-1795. 23. J. Gryziecki, Z. Gdula, Material Science Engineering 93 (1987), 99-105. 24. A.A. Elmustafa, D.S. Stone, Materials Science and Engineering A 358 (2003) 1-8. 25. D. Faghihi, G.Z. Voyiadjis, T. Park, Journal of Engineering Materials and Technology 135, 021003-1021003-17. 26. D. Faghihi, G.Z. Voyiadjis, Mechanics of Materials 44 (2012) 189–211. 27. D. Faghihi, G.Z. Voyiadjis, Proc. IMechE Vol. 225 Part N: J. Nanoengineering and Nanosystems 515. 28. H. Mecking, U.F. Kocks, Acta Metall (1981) 29, 1865-1875. 29. Y.H. Zhao, Y.T. Zhu, X.Z. Liao, Z. Horita, T.G. Langdon, Appl. Phys. Lett. 89, 121906, (2006). 30. Yong-Hao Zhao, John F. Bingert, Xiao-Zhou Liao, Bao-Zhi Cui, Ke Han, Alla V. Sergueeva, Amiya K. Mukherjee, Ruslan Z. Valiev, Terence G. Langdon, Yuntian T. Zhu, Adv. Mater. 18 (2006) 2949– 2953. 31. P. Yang, H. Yang, J. Tao, C. Li, L. Shen, X.K. Zhu, Materials Science Forum 667-669 (2011) 10031008. 32. S.R. Kalidindi, A.A. Salem, R.D. Doherty, Advanced Engineering Materials 5, (2003), 229-232. 33. Y.H. Zhao, Y.T. Zhu, X.Z. Liao, Z. Horita, T.G. Langdon, Applied Physics Letters 89, 121906 (2006). 34. Meyers M.A., Vohringer O., Lubarda V.A. (2001), Acta Mater. 49, 4025 35. H. Bahmanpour, A. Kauffmann, M.S. Khoshkhoo, K.M. Youssef, S. Mula, J. Freudenberger, J. Ecker, R.O. Scattergood, C.C. Koch, Materials Science and Engineering A 529 (2011) 230– 236. 36. V.S. Sarma, J. Wang, W.W. Jian, A. Kauffmann, H. Conrad, J. Freudenberger, Y.T. Zhu, Materials Science and Engineering A 527 (2010) 7624-7630. 37. B.Q. Li, M.L. Sui, S.X. Mao, Journal of Material Science Technology 27 (2011) 97. 38. E.B. Tadmor, N. Bernstein, Journal of the Mechanics and Physics of Solids 52 (2004) 2507. 39. M. Muzyk, Z. Pakiela and K.J. Kurzydlowski, Scripta Materialia 64 (2011) 916. 40. C.-M. Kuo, C.-H. Tso, C.-H. Lin, Mater. Sci. Eng. A 519 (2009) 32.
11
Table 1: Chemical composition and grain size of the four fcc metals studied in this investigation Test material
Composition
70/30 brass
70% Cu, 30%Zn
Average grain size (m) 75
Pure copper
99.9% Cu
25
5052 Aluminum
2.2% Mg, 0.40% Fe, 0.25% Si, 0.17% Cr, 0.1% Mn, 0.1% Cu, 0.1% Zn, Bal. Al
52 [40]
6061 Aluminum
1.0% Mg, 0.6% Si, 0.4% Fe, 0.4% Cu, 0.3% Cr, 0.3% Zn, 0.1% Mn, 0.1% Ti, Bal. Al
100
Table 2: YS and UTS values in tested alloys in the as received and flow formed conditions
Materi al & proper ties
As recei ved
5052 aluminu m
80
6061 aluminu m
80
Pure copper
65
70/30 brass
100
YS (MPa) Flow formed
As recei ved
UTS (MPa) Flow formed
Change in YS As received to flow formed
Change in UTS As received to flow formed
Axi
Circumfer
Axi
Circumfer
Axi
Circumfer
Axi
Circumfer
al
ential
al
ential
al
ential
al
ential
27
240
30
293
70 %
67%
36 %
33%
195
58 %
53%
23 %
15%
305
80 %
76%
37 %
28%
398
79 %
74%
31 %
12%
195
0 19
5 170
165
0 33
5 270
220
0 48
21
35 0
380
0
350
50 5
Table 3: Elongation percent El (%) in the as received and flow formed conditions and the change in the El (%) caused by flow forming
Material &
As received
El (%) Flow formed
12
Change in El (%)
As received to flow formed
properties
Axial
Circumferential
Axial
Circumferential
5052 aluminum
24
3.2
7.4
-86.7%
-69.2%
6061 aluminum
18
6.3
9.6
-65%
-46.7%
Pure copper
43
5.9
7.9
-86.3%
-81.6%
70/30 brass
70
4.3
5.5
-93.8%
-92.1%
Roller
Flow formed work piece
Preform
Forming direction
Mandrel Figure 1: Schematic representation of three roller flow forming process.
13
Figure 2: Schematic diagram of tensile specimens machined from flow formed parts.
14
(a)
a
b
15
(c)
c (d)
d Figure 3: Starting (left side) and flow formed (right side) grain structures in a) 6061 aluminum alloy, b) 5052 aluminum alloy [40], c) pure copper, and d) 70/30 brass. As seen, grains are elongated axially along mandrels axis in the flow formed parts. White arrow shows the direction of forming.
16
250
6061 alumunim_As recieved 6061 alumunim_Flow Formed_Axial 6061 alumunim_Flow Formed_Circumferential
Stress (MPa)
200
150
100
50
0 0
5
10
15
20
Strain a 350
5052 aluminum_As recieved 5052 aluminum_Flow Formed_Axial
300
5052 aluminum_Flow Formed_Circumferential
Stress (MPa)
250
200
150
100
50
0 0
5
10
15
20
Strain b
17
25
30
600
Brass 70/30_As recieved Brass 70/30_Flow Formed_Axial Brass 70/30_Flow Formed_Circumferential
Stress (MPa)
500
400
300
200
100
0 0
20
40
60
80
Strain c 400
Cu_As recieved Cu_Flow Formed_Axial Cu_Flow Formed_Circumferential
Stress (MPa)
300
200
100
0 0
10
20
30
Strain d
18
40
50
Figure 4: Tensile engineering stress/strain curves for as received and flow formed (in axial and circular directions) for the tested materials, a) 6061 aluminum, b) 5052 aluminum, c) 70/30 brass, and d) pure copper.
a
19
b Figure 5: Normalized work hardening rate ( G ) versus strain for 5052 and 6061 aluminum alloys. The presence of Mg solute atoms may lead to the appearance of plastic instabilities resulting in the well-known serrated yielding or Portevin-Le Chatelier (PLC) effect which marks itself as steps in the G curve at strains bigger that 2%.
20