Increasing the Working Limit of Extruded Aluminum Tubes ... · PDF fileIncreasing the Working...
4
Increasing the Working Limit of Extruded Aluminum Tubes During Draw Bending by Introducing a Wiper Die Yusuke Okude 1,+ , Shuji Sakaki 2 , Shoichiro Yoshihara 3 and Bryan J. MacDonald 4 1 Department of Material and Environmental Technology, Faculty of Engineering, University of Yamanashi, Kofu 400-8511, Japan 2 Emeritus professor of Tokyo Metropolitan Institute of Technology, Tokyo 191-0065, Japan 3 Department of Mechanical System Engineering, Faculty of Engineering, University of Yamanashi, Kofu 400-8511, Japan 4 Mechanical and Manufacturing Engineering, Dublin City University, Dublin 9, Ireland Aluminum alloy sections are widely utilized in many applications such as in general structures and automotive components. Owing to the increasing complexity of these components there is an increasing demand for highly curved thin wall sections. Therefore, reliable bending processes are required to manufacture such components without undesirable deformations such as wrinkling and folding occurring during bending. To prevent undesirable deformations and predict when they might occur, it is necessary to generate much data from various forming configurations in which process parameters are varied. In this paper, the authors concentrate on the problem of bending square-section thin- walled tubes and use finite element analysis to generate the data required to predict working limits. Two different tube materials (A6063S-O and A6063S-T5) were used to investigate the relationship between deformation behavior and material properties. Moreover, the effect of using a wiper die on the working limit of the extruded aluminum sections was investigated. [doi:10.2320/matertrans.MF201113] (Received August 18, 2011; Accepted December 22, 2011; Published February 15, 2012) Keywords: tube bending, draw bending, working limit, wiper die, aluminum extruded section 1. Introduction Square extruded aluminum tubes have advantages in their features as light weight and high bending rigidity. These materials are suitable for use in general structures with curved profiles and in automotive components. In general, secondary bending is necessary to obtain a desired shape/configuration. However, undesirable deformation problems are often encountered in the draw bending of thin extruded tubes. In recent years, various studies have been carried out to prevent such undesirable deformations. Finite element simulation has previously been demonstrated to successfully predict undesirable deformation behavior in draw bending. 1) The positive effects of an additional axial tension and the presence of a center rib are further clarified by finite element analysis (FEA). Li et al. 2) have proposed an FE method of evaluating of wrinkling limit diagrams in pipe bending. The method described by Li et al. presented a calculated stress field in which material anisotropy, a 3D strain field, different loading paths and boosting force were considered. Thus, this method gave a wrinkling criterion that accurately reflects the true wrinkling mode. Kristoffer 3) described their study that examined in detail a multistage forming operation with tube bending followed by tube hydroforming. This study showed the importance of including the effects acquired from the bending and preforming processes in hydroforming simu- lation to obtain reliable simulation results. The effect of material constants in the tube bending requires further clarification. Murata et al. 4) used experimental and FE analyses to show that the work-hardening exponent n has no effect on springback, thickness strain distribution or flatness ratio at the same bending radius of the same tube dimensions. The authors 5) previously reported how a wiper die can be introduced to reduce wrinkling and increase formability. Thus, it is considered that the use of a wiper die is effective in preventing wrinkling; however, the relationship between the wrinkling behavior of extruded aluminum sections and the decrease in the extent of wrinkling when using a wiper die has not yet been determined. In this study, the authors’ aim is to obtain much data on the wrinkling behavior of square section tubes using FE simulation and to thus determine the most effective utilization of a wiper die and axial tension for controlling wrinkling. 2. Finite Element Analysis 2.1 FEA model and condition Representative finite element models of extruded alumi- num alloy (A6063S-O and A6063S-T5) square tubes were constructed. Table 1 shows the properties of the tube materials. The width W 0 of each workpiece is 40 mm and the height H 0 is 40 mm. The wall thicknesses t 0 used were 1.0 and 1.5 mm. The length of each workpiece was varied depending on bending radius. Square section workpieces were modeled using 7552 solid elements and 14224 nodes. An elastoplastic material was assumed, satisfying an n-th power strain hardening law with the constitutive equation, · ¼ Cð¾ n þ ¾ p Þ n ; ð1Þ where C is the plastic modulus, ¾ n is the offset due to elastic strain, ¾ p is the effective plastic strain and n is the work- hardening exponent. Table 1 Material properties of workpieces. 6) Material Density, μ/kg·mm ¹3 Modulus of Elasticity, E/GPa C value, /MPa* n value* A6063S-O 2.7 © 10 ¹6 68.9 189 0.23 A6063S-T5 2.7 © 10 ¹6 68.9 281 0.08 *· ¼ Cð¾ n þ ¾ p Þ n + Graduate Student, University of Yamanashi Materials Transactions, Vol. 53, No. 5 (2012) pp. 875 to 878 Special Issue on Advanced Tube Hydroforming Technology for Lightweight Components © 2012 The Japan Institute of Metals
Transcript of Increasing the Working Limit of Extruded Aluminum Tubes ... · PDF fileIncreasing the Working...
Increasing the Working Limit of Extruded Aluminum TubesDuring Draw Bending by Introducing a Wiper Die
Yusuke Okude1,+, Shuji Sakaki2, Shoichiro Yoshihara3 and Bryan J. MacDonald4
1Department of Material and Environmental Technology, Faculty of Engineering, University of Yamanashi, Kofu 400-8511, Japan2Emeritus professor of Tokyo Metropolitan Institute of Technology, Tokyo 191-0065, Japan3Department of Mechanical System Engineering, Faculty of Engineering, University of Yamanashi, Kofu 400-8511, Japan4Mechanical and Manufacturing Engineering, Dublin City University, Dublin 9, Ireland
Aluminum alloy sections are widely utilized in many applications such as in general structures and automotive components. Owing to theincreasing complexity of these components there is an increasing demand for highly curved thin wall sections. Therefore, reliable bendingprocesses are required to manufacture such components without undesirable deformations such as wrinkling and folding occurring duringbending. To prevent undesirable deformations and predict when they might occur, it is necessary to generate much data from various formingconfigurations in which process parameters are varied. In this paper, the authors concentrate on the problem of bending square-section thin-walled tubes and use finite element analysis to generate the data required to predict working limits. Two different tube materials (A6063S-O andA6063S-T5) were used to investigate the relationship between deformation behavior and material properties. Moreover, the effect of using awiper die on the working limit of the extruded aluminum sections was investigated. [doi:10.2320/matertrans.MF201113]
(Received August 18, 2011; Accepted December 22, 2011; Published February 15, 2012)
Keywords: tube bending, draw bending, working limit, wiper die, aluminum extruded section
1. Introduction
Square extruded aluminum tubes have advantages in theirfeatures as light weight and high bending rigidity. Thesematerials are suitable for use in general structures with curvedprofiles and in automotive components. In general, secondarybending is necessary to obtain a desired shape/configuration.However, undesirable deformation problems are oftenencountered in the draw bending of thin extruded tubes. Inrecent years, various studies have been carried out to preventsuch undesirable deformations. Finite element simulationhas previously been demonstrated to successfully predictundesirable deformation behavior in draw bending.1) Thepositive effects of an additional axial tension and thepresence of a center rib are further clarified by finite elementanalysis (FEA). Li et al.2) have proposed an FE method ofevaluating of wrinkling limit diagrams in pipe bending. Themethod described by Li et al. presented a calculated stressfield in which material anisotropy, a 3D strain field, differentloading paths and boosting force were considered. Thus, thismethod gave a wrinkling criterion that accurately reflects thetrue wrinkling mode. Kristoffer3) described their study thatexamined in detail a multistage forming operation with tubebending followed by tube hydroforming. This study showedthe importance of including the effects acquired from thebending and preforming processes in hydroforming simu-lation to obtain reliable simulation results. The effect ofmaterial constants in the tube bending requires furtherclarification. Murata et al.4) used experimental and FEanalyses to show that the work-hardening exponent n hasno effect on springback, thickness strain distribution orflatness ratio at the same bending radius of the same tubedimensions.
The authors5) previously reported how a wiper die can beintroduced to reduce wrinkling and increase formability.
Thus, it is considered that the use of a wiper die is effective inpreventing wrinkling; however, the relationship between thewrinkling behavior of extruded aluminum sections and thedecrease in the extent of wrinkling when using a wiper diehas not yet been determined.
In this study, the authors’ aim is to obtain much data onthe wrinkling behavior of square section tubes using FEsimulation and to thus determine the most effective utilizationof a wiper die and axial tension for controlling wrinkling.
2. Finite Element Analysis
2.1 FEA model and conditionRepresentative finite element models of extruded alumi-
num alloy (A6063S-O and A6063S-T5) square tubes wereconstructed. Table 1 shows the properties of the tubematerials. The width W0 of each workpiece is 40mm andthe height H0 is 40mm. The wall thicknesses t0 used were 1.0and 1.5mm. The length of each workpiece was varieddepending on bending radius. Square section workpieceswere modeled using 7552 solid elements and 14224 nodes.An elastoplastic material was assumed, satisfying an n-thpower strain hardening law with the constitutive equation,
· ¼ Cð¾n þ �¾pÞn; ð1Þwhere C is the plastic modulus, ¾n is the offset due to elasticstrain, �¾p is the effective plastic strain and n is the work-hardening exponent.
Table 1 Material properties of workpieces.6)
MaterialDensity,
µ/kg·mm¹3Modulus of Elasticity,
E/GPaC value,/MPa*
n value*
A6063S-O 2.7 © 10¹6 68.9 189 0.23
A6063S-T5 2.7 © 10¹6 68.9 281 0.08
*· ¼ Cð¾n þ �¾pÞn+Graduate Student, University of Yamanashi
Materials Transactions, Vol. 53, No. 5 (2012) pp. 875 to 878Special Issue on Advanced Tube Hydroforming Technology for Lightweight Components©2012 The Japan Institute of Metals
Figure 1 shows the FEA model of the draw bendingprocess and the schematic of each workpiece. The exper-imental apparatus consists of a rotary die, a chuck, a bendingroller, a guide roller and brake pads. The draw bendingmachine can bend the workpieces with bending radii R of150, 200, 300 and 400mm. The bending angle used in thisinvestigation was 90°. The friction coefficient between theworkpieces and the dies was defined as 0.5.2.1.1 Laminated elastic mandrel
Figure 2 shows the FEA model of the laminated elasticmandrel, Table 2 indicates the material properties of themandrel. A laminated elastic mandrel consisting of layers ofNylon-66, polyvinyl chloride (PVC) and phosphor bronzeplates was used. The Nylon-66 and PVC plates aresandwiched by the phosphor bronze plates. The mandrelwas inserted into the workpieces to prevent undesirabledeformation during forming.2.1.2 Wiper die
Figure 3 shows a schematic of the wiper die. The wiper dieis located behind the bending point to prevent wrinkling byconstraining both the flange and the web. The square tubepasses inside the wiper die during forming.2.1.3 Axial tension
Axial tension is applied by sandwiching each workpiece
between both brake pads to prevent wrinkling that may occureven though a wiper die was used. Figure 4 shows themethod of applying axial tension. The axial tension ratio Rat
was calculated from the axial tension and tensile strengthusing.
Rat ¼ fT=ð·B � AÞg � 100ð%Þ; ð2Þwhere, T is the axial tension, ·B is the tensile strength and A isthe cross section of each workpiece. Table 3 shows therelationship between axial tension and axial tension ratio.Axial tension was applied on each material at Rat = 0, 5 and10% and a thickness of what. The brake force F for applyingthe axial tension was loaded in the range of 1.16.8 kN.
2.2 Results and discussion2.2.1 Deformation behavior
Figure 5 shows the typical deformation of bent squaretubes. Wrinkling was defined as the deformation thatconsecutively shows a wave shape and folding, characterizedby a combination of a linear portion and local increasing ofwrinkling. The difference in deformation behavior betweenA6063S-O and A6063S-T5 is especially shown at thebending radius of 150mm. Folding occurred in the benttube (t0 = 1.0mm) of A6063S-T5. Wrinkling and foldingwere observed to intensify with decreasing n.
Wiper dieBending point
Rotary die
Chuck
Guide roller
Bending roller Brake pads
W0=
40m
m
H0=40mm
t0=1.0,1.5mm
L0=1000~1400mm
Fig. 1 FEA model of the draw bending process and schematic of workpiece.6)
Cross sectionof mandrel
Phosphorbronze platesPVC plates
Nylon 66plates
36.8mm
36.8mm
277mm
Fig. 2 FEA model of laminated elastic mandrel.
Table 2 Material properties of laminated elastic mandrel.
MaterialDensity,
µ/kg·mm¹3Modulus of
elasticity, E/GPa
Phosphor bronzeplates
10.2 © 10¹6 102
Polyvinylchloride plates
1.7 © 10¹6 3.2
Nylon-66 plates 3.2 © 10¹6 0.8
Fig. 3 Schematic of wiper die.
Brake pads
Chuck
Rotary die
Wiper die
Brake force F
Fig. 4 Method of applying of axial tension.
Y. Okude, S. Sakaki, S. Yoshihara and B. J. MacDonald876
2.2.2 Working limit of square tubesFigure 6 shows the classification of the deformations
generated using the laminated elastic mandrel for variousbend degrees R0/H0, where R0 = R + H0/2 = 10.5 (R =400mm), 8.0 (R = 300mm), 5.5 (R = 200mm) and 4.25(R = 150mm), with thickness ratios t0/H0 = 0.038 (t0 =1.5mm) and 0.025 (t0 = 1.0mm). Undesirable deformationwas enhanced in the case of bending A6063S-T5. Wrinklingat the compression flange and web occurs or folding wasobserved when wrinkling reduces the bending rigidity of thetube. Thus, a wiper die was used in draw bending to preventundesirable deformation. Figure 7 shows the classificationof the deformations generated when the laminated elasticmandrel and wiper die were used. The working limit wasmarkedly improved with the wiper die without axial tension,however, wrinkling was observed on the compression flangeor web in areas A and B. It is confirmed that undesirabledeformation occurred with the use of only restriction diesunder these conditions and without axial tension. Axialtension was thus applied in an attempt to eliminate this. Asthe results show, the square tubes were bent without any
defects when applying axial tension in addition to the use ofrestriction dies.2.2.3 Relationship between wrinkling behavior and
wiper dieFigure 8 shows the effect of the wiper die on the wrinkling
behavior of the bent square tube (R = 150mm, t0 = 1.0mm,A6063S-O, with a mandrel). In the case of using a mandrel,wrinkling occurred at 8° intervals of bending angle. Wrinklesat bending angles between 0 and 15° did not grow along theradial direction in comparison with wrinkling near thebending angle of 50°. In this range, a low growth of wrinkleswas observed, and each workpiece was in contact with therotary die in the initial position during the draw bending.Hence, the growth of wrinkles along the radial direction wasprevented by the filling up of the space between the rotary dieand workpiece. The depth of wrinkling (Wd) at bendingangles between 40 and 70° increased at a 5mm (Wd/H0 =0.13) difference between the convex and concave deforma-tions. As results, wrinkling occurred at 8° intervals despitethe difference in wrinkling depth.
On the other hand, wrinkling near the bending start point(point I) and end point (point II) occurred when a wiper diewas used. The wrinkling near point I increased in comparisonwith the bending in the case with the mandrel only. It wasconsidered that the increase in wrinkling depth was an effectof the wiper die on the decentration of wrinkling. Further-more, wrinkling near point II occurred as the compressivestress and strain exceeded the limit of preventing wrinklingby the wiper die. These compressive stress and strain weremeasured under conditions A and B with high compressivedeformation in comparison with other conditions.2.2.4 Effect of axial tension on the wrinkling
Figure 9 shows the effect of axial tension on wrinkling andcompressive stress (R = 150mm, t0 = 1.0mm, A6063S-O).Compressive stress was markedly decreased by applyingaxial tension as shown in Fig. 9(a). Then, compressive stresswas measured in the element when wrinkling occurred toclarify the effect of axial tension on wrinkling. At Rat = 5%,wrinkling occurred regardless of applying axial tension. Onthe other hand, wrinkling was prevented at Rat = 10% sincecompressive stress decreased by 15% in comparison with thatat Rat = 5%. Therefore, the range of the occurrence ofwrinkling was confirmed. Table 4 shows the minimum axialtension for preventing wrinkling under the conditions in
Table 3 Relationship between axial tension and axial tension ratio.
Material A6063S-O A6063S-T5
Tensile strength, ·B/MPa 96 206
Cross section, A/mm2 156 231 156 231
Axial tension, T/kN
Axial tensionratio Rat (%)
0 0 0 0 0
5 0.7 1.1 1.6 2.4
10 1.5 2.2 3.2 4.8
Wrinkling
(a)Linear portion
Folding
(b)
Fig. 5 Bent square tubes (R = 150mm, t0 = 1.0mm) (a) A6063S-O and(b) A6063S-T5.
(a)
Thickness ratio t0/H0
0.0350.030.0250.23
6
9
12
0.04Thickness ratio t0/H0
0.2 0.025 0.0353
6
9
12
0.04
Ben
d de
gree
R0/H
0
Ben
d de
gree
R0/H
0
(b)
: Folding: Wrinkling at web
: Wrinkling at both compression flange and web : Wrinkling at compression flange
0.03
Fig. 6 Classification of deformations in the draw bending process withoutwiper die (a) A6063S-O and (b) A6063S-T5.
(a)
Thickness ratio t0/H0
0.2 0.025 0.033
6
9
12
0.04Thickness ratio t0/H0
0.0350.030.0250.23
6
9
12
0.04
B
A Ben
d de
gree
R0/H
0
Ben
d de
gree
R0/H
0
(b)
: Wrinkling at web: Wrinkling at compression flange
: Without any defects
0.035
Fig. 7 Classification of deformations in the draw bending process withwiper die (a) A6063S-O and (b) A6063S-T5.
Increasing the Working Limit of Extruded Aluminum Tubes During Draw Bending by Introducing a Wiper Die 877
which wrinkling occurred despite the introduction of a wiperdie. At R = 150mm for each material, wrinkling was com-pletely prevented by applying axial tension (Rat = 10%). Rat
was 5% to prevent wrinkling at R = 200mm. These axialtensions were the upper limit for preventing wrinkling at eachbending radius. From the results, it was confirmed thatwrinkling was prevented in the axial tension ratio rangebetween 10 and 5% at R = 150mm. The range Rat is between5 and 0% at R = 200mm. In addition, it was expected thatthere are suitable axial tensions at each bending radius.
3. Conclusions
(1) The difference in deformation behavior betweenA6063S-O and A6063S-T5 tubes is shown especiallyat a 150mm bending radius, it appears to be an effectof n.
(2) The working limit was improved remarkably utilizing awiper die in the case without axial tension. Introducingwiper die is effective for reducing undesirable defor-mation at every thickness.
(3) Regarding wrinkling behavior in the case with amandrel (A6063S-O, R = 150mm, t0 = 1.0mm), thewavelength of wrinkling was constant at 8° despitethe difference in wrinkling depth. The wrinkling in thevicinity of the bending start point and end pointoccurred when using a wiper die.
(4) It was confirmed that wrinkling is prevented in the axialtension ratio range between 10 and 5% at R = 150mm.The Rat range for preventing the wrinkling is between 5and 0% at R = 200mm. Thus, it is expected that therewill be a suitable axial tension at each bending radius.
REFERENCES
1) N. Utsumi and S. Sakaki: J. Mater. Process. Technol. 123 (2002) 264269.
2) H. Li, H. Yang, M. Zhan and R. J. Gu: J. Mater. Process. Technol. 177(2006) 192196.
3) T. Kristoffer: J. Mater. Process. Technol. 127 (2002) 401408.4) M. Murata, T. Kuboki, K. Takahashi, M. Goodarzi and Y. Jin: J. JSTP
201 (2008) 189192.5) Y. Okude, S. Sakaki and S. Yoshihara: The Proc. 2009 Japanese Spring
Conference for the Technology of Plasticity, (2009) pp. 401402 (inJapanese).
6) Y. Okude, S. Sakaki, S. Yoshihara and B. J. MacDonald: Proc. 5th Int.Conf. on Tube Hydroforming TUBEHYDRO2011, (2011) pp. 133136.
(a) (b)
Fig. 8 Effect of wiper die on wrinkling behavior of bent square tube (R = 150mm, t0 = 1.0mm, A6063S-O, with mandrel) (a) outershape of compression flange and (b) relationship between depth of wrinkling and arc length.
(a) (b)
Fig. 9 Effects of axial tension on wrinkling and compressive stress (R = 150mm, t0 = 1.0mm, A6063S-O) (a) Y-stress distribution ofFEA result without axial tension and with axial tension and (b) relationship between average compression stress and FEA conditions.
Table 4 Minimum axial tensions for preventing wrinkling in the conditionsA and B.
Material A6063S-O A6063S-T5
Condition A(R150) B(R150) B(R150) B(R200) B(R200)
Thickness, t0/mm 1.0 1.0 1.5 1.0 1.5
Axial tension, T/kN 1.5 3.2 4.8 1.6 2.4
Axial tension ratioRat (%)
10 10 10 5 5
Y. Okude, S. Sakaki, S. Yoshihara and B. J. MacDonald878
