INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5, pp. 671-677

MAY 2016 / 671

DOI: 10.1007/s12541-016-0082-0

ISSN 2234-7593 (Print) / ISSN 2005-4602 (Online)

A Study on the Mechanical Properties and Springback of 3D Aluminum Sheets Seok-Hwan Oh1, Dae-Cheol Ahn1, and Young-Suk Kim1,# 1 School of Mechanical Engineering, Kyungpook National University, 80, Daehak-ro, Buk-gu, Daegu, 41566, South Korea # Corresponding Author / E-mail: [email protected], TEL: +82-53-950-5580, FAX: +82-53-956-9914 KEYWORDS: 3D sheet, Embossing pattern, Tensile test, Springback

Three-dimensional (3D), cone-shape embossed aluminum sheets are used in automotive exhaust systems to increase their heat dissipation efficiency by increasing the surface area. However, the manufacturing process has various restrictions because wrinkling occurs easily during the press forming process. In this study, A tensile test and a bending test were performed to investigate the mechanical properties and springback characteristics of 3D aluminum sheets. We clarified how the direction in which the specimen is cut affects the tensile properties. The results of the tensile test showed that the characteristics of the parallel and diagonal direction specimens differed from each other and those of the as-received flat sheet. The 3D aluminum sheets had a smaller Young’s modulus and smaller flow stress than the as-received flat sheets in the small plastic range due to the flattening effect of the embossed cone shapes. However, as the plastic strain increased, the flow stress followed the as-received flat specimen’s flow stress curve because the cone-shape was flattened according to increases in the plastic strain. The yield stress increased in the diagonal-direction specimen and decreased in the parallel-direction specimen. The change in Young’s modulus in the 3D sheets affected the amount of springback. Manuscript received: April 15, 2015 / Revised: November 13, 2015 / Accepted: January 11, 2016 This paper was presented at ISGMA 2015

NOMENCLATURE R1 = the radius of the upper punch for the V-bending test R2 = the radius of the lower die for the V-bending test L1 = the distance between the upper die and the lower die for the V-bending test θi = the initial angle of the specimen after the V-bending test θf = the final angle of the specimen after the V-bending test Δθ = the difference in the angle between the initial and final angles of the specimen after V-bending R3 = the radius of the upper punch for the U-draw bending test R4 = the radius of the lower die for the U-draw bending test P1 = the length of the punch for the U-draw bending test L2 = the clearance of the die on the U-draw bending test θ1 = the angle of the specimen on the upper punch shoulder θ2 = the angle of the specimen on the lower die shoulder σY = yield strength σTS = tensile strength E = Young’s modulus

© KSPE and Springer 2016

n = strain hardening exponent K = strength coefficient R = plastic anisotropic coefficient dεwidth, dεthickness = the plastic strain in the direction of the width and thickness, respectively

1. Introduction In the automotive industry, protecting the global environment is an important issue. Aluminum is emphasized as an automotive material because of its low density, and it has been adopted one of the materials for interior and exterior panels in luxury cars.1 As shown in Fig. 1, embossing aluminum, double-layer aluminum, and aluminum-silicone alloy-coated sheets are widely used under car chassis to dissipate and insulate high-temperature heat (about 200~750oC from the exhaust gas in the exhaust system and the muffler).2 A three-dimensional (3D) aluminum sheet is manufactured by rolling and pressing, hydroforming, or electromagnetic forces, and it can be produced in various

672 / MAY 2016

INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5

Fig. 1 Applications of 3D sheets for heat protection

3D shapes, such as a cone, diamond or complex shape.3,4 A widely used size is 2~3 mm-high, and demonstrations show high performance in dissipating heat because of the increased surface area. The sheet shows stiffness reinforcement because of the strain hardening effect while the plastic deforms. Namoco et al. studied the mechanical properties and the strengthening effect of aluminum sheet metals using embossing and restoration. The results of the study demonstrated that by using appropriate embossing and the restoration or combination of both patterns, the deep drawability of sheet metal could be improved.5,6 Fritzsche et al. tested the influence of the structural position on the strength and the deformation behavior of structured sheet metal.7 Cho et al. developed a composite multilayer aluminum sheet and documented its mechanical properties. The tensile strength of the forming direction was 1.44 times higher than that of the transverse direction.8 Although 3D sheet metals have many advantages, the manufacturing process has various restrictions because of the 3D structure; intrinsically, wrinkling occurs during press forming process.9 Finite element analysis using software, such as AutoForm, PamStamp, ABAQUS, LS-Dyna, has been widely adopted in hard tool design and manufacturing industries, and the analysis has reasonable accuracy. However, related studies on hard tool design or simulation techniques for the optimal press forming process for 3D sheet metal are lacking.10,11 In this study, the mechanical properties for each direction and the springback effect in the press forming process were evaluated for a 3D aluminum sheet with 6.4 mm-pitch and 0.7 mm-high cone.

2. Material and Experiments 2.1 Tensile test In this study, 0.5 mm-thick A3004-P aluminum sheets were used. The aluminum sheets were cold rolled to 6.4 mm-pitch and 0.7 mmhigh cone. A tensile specimen that had 12.5 mm-wide and 50 mm-long gauge was cut with a laser from a 0.5 mm-thick as-received flat aluminum sheet and a 3D aluminum sheet with a parallel and diagonal embossed structure, as shown in Figs. 2 and 3 The parallel embossed shape meant the peaks and valleys aligned each other to the tensile direction and the diagonal embossed shape is 45o from the parallel embossed shape, meaning the peaks and valleys aligned alternatively. The difference in tensile properties between the as-received and 3D aluminum sheets was analyzed using the tensile test. The results were averaged for three specimens for each direction, and the average stressstrain curves were plotted for three stress data points for the same strain values.

Fig. 2 3D aluminum sheet and the schematics of the specimens (a) a diagonal embossed shape specimen, (b) a parallel embossed shape specimen

Fig. 3 Tensile test specimen and test results showing an as-received specimens (a) before and (b) after the test and a parallel embossed shape specimens (c) before and (d) after the test

Fig. 4 Die set and springback angle for the V-bending test

Fig. 5 Die set and springback angle for the U-draw bending test

2.2 Bending test The bending test is widely used to understand the springback

INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5

MAY 2016 / 673

Table 1 Mechanical properties of the as-received flat and 3D sheet specimens Item As-received Type Flat 60.5 σY (MPa) σTS (MPa) 183 E (GPa) 70.0 Strain hardening exponent, n 0.30 Strength coefficient, K (MPa) 312 Elongation (%) 0.2 Plastic anisotropic coefficient, R 0.59

3D (Embossed) Parallel Diagonal 57.4 64.0 193 184 47.0 58.3 0.32 0.27 317 296 0.27 0.21 0.03 0.23

phenomenon during plastic deformation. The V-bending and U-draw bending test were conducted with a tool set, as shown in Figs. 4 and 5 for 200 mm-long and 50 mm-wide 3D specimens and the springback was measured after the test.12 The springback values were averaged for three results except the extreme maximum and minimum values among the five test results. Before the bending test, the surface of the specimens was coated with P-DBH lubricant. As shown in Figs. 4 and 5, the increased angle (Dq) was measured in the V-bending test, and the increased angle at the punch (Dq1) and die shoulder (Dq2) were measured in the U-draw bending test as the springback values. The angle was measured in AutoCAD using the scanned section data of the specimens.13

2.3 Tensile test results The tensile test results are summarized in Table 1. Figs. 6 and 7 show the stress-strain curves of the parallel and diagonal embossed sheets compared with the as-received flat sheet. In Fig. 7, Young’s modulus for the 3D sheet, E=47 GPa, is smaller than that of the as-received flat sheet, E=70 GPa. The Young’s modulus of the 3D sheet is the appearance modulus of elasticity, which depends on the 3D shape rather than material properties. The yield stress is defined as the stress of the 0.2% offset strain. The yield stress in the parallel embossed shape was not as clear as that of the as-received flat specimen. Just after the yield occurred, the flow stress was smaller than that of the as-received specimen. We believe that this was caused by the overestimated strain during the test. As shown in Fig. 8(a), the peaks and valleys are aligned in the tensile direction in the parallel embossed specimen. The peaks and valleys are restored by the bending moment induced by the tensile load; they increase the plastic strain. However, as shown in Fig. 7, when the plastic strain is larger than 0.085, the flow stress is almost the same as that of the as-received specimen since the 3D structure is almost completely flattened in the tensile load direction. The total elongation of the parallel embossed sheet is about 27%, which is larger than that of the as-received sheet, which is about 20%. As a result, the press formability should increase because of the low yield stress and large fracture elongation. The Young’s modulus for the diagonal embossed shape specimen, E=58.3 GPa, is lower than that of the as-received specimen, and the flow stress just after yielding is higher than that of the as-received specimen (Fig. 7). In the diagonal embossed specimen, because the peaks and valleys are alternatively aligned in the tensile direction, the embossed shape is retained under plastic deformation and it increases the resistance to the tensile load.

Fig. 6 Stress-strain curves for the as-received and diagonal embossed specimens for the tensile test

Fig. 7 Stress-strain curves for the as-received and parallel embossed specimens for the tensile test

Fig. 8 Deformation patterns of the gauge area for (a) the parallel embossed specimen and (b) the diagonal embossed specimen in the original state and the deformed state with a strain of 0.185

Fig. 8 shows the development of plastic deformation for the parallel and diagonal embossed shape specimens under the tensile load. As explained, the parallel embossed specimen has almost the same depth at the embossed lines, and a large amount of shrinkage occurs in the width direction. For example, when the strain is 0.185, the width of the parallel embossed specimen is reduced to 11.25 mm from 12.01 mm, and the peak height is reduced to 1.22 mm from 1.42 mm. However, in the diagonal embossed specimen, the peak height is dramatically reduced to 0.73 mm from 1.45 mm, because the embossed shape is

674 / MAY 2016

INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5

Fig. 9 Plastic anisotropic coefficient as the strain for the as-received and 3D aluminum sheets

Fig. 10 Deformed shape of specimens: (a) the as-received specimen after the V-bending test, (b) the as-received specimen after the U-draw bending test, (c) the 3D specimen after the V-bending test, and (d) the 3D specimen after the U-draw bending test

Table 2 Measured springback for the as-received and 3D aluminum sheets for each direction Specimen

2θi 93.1 92.2 91.8

As-received Parallel embossed Diagonal embossed

V-bending 2θf 111 115 114

Δθ 17.9 22.8 22.2

θ1 97.6 98.4 99.2

U-draw bending θ2 89.3 90.8 89.1

θ1+θ2 186.9 189.2 188.3

flattened throughout the gauge section area of the specimen under the tensile load; instead, the width reduction is very small, from 12.35 mm to 12.1 mm. Fig. 9 shows the anisotropic coefficient value (R-value) as the strain, and the value is defined as in Eq. (1): dε width R = -------------------dε thickness

(1)

The R-value of sheet metal is an intrinsic value that usually depends on the texture, which is determined during cold rolling. Therefore, the R-value is not significantly different even if the specimen is under plastic deformation unless the texture has not changed. The R-value of the 3D specimen is an appearance value in the macroscopic viewpoint that includes the effect of the embossing rather than the intrinsic value. This macroscopic R-value, the plastic anisotropic coefficient, is useful for explaining the plastic deformation behavior of 3D sheet metal under a multiaxial stress state. Generally, the R-value for the aluminum sheet is measured at the strain 0.1 (or 0.15); however, in this study, the variation of the macroscopic R-value was measured as the strain. At around 0.06 of the strain, the R-value of the as-received sheet was about 0.59, the parallel embossed shape was 0.03, and the diagonal embossed shape was about 0.23. The plastic anisotropic coefficient was dramatically decreased by the applied 3D embossed structure. In particular, the R-value was almost zero at about 0.02 of the strain in the parallel embossed specimen. The width of the specimen was increased by the flattening effect of the embossed structure under the tensile load. Then the behavior was almost same as that of the as-received specimen because the embossing was flattened. Sometimes, the macroscopic R-value was negative. The 3D structured sheet had a negative R-value even when the R-value of the as-received sheet was positive. The deep drawability and the forming limit should be determined in future studies.

Fig. 11 Finite element modeling and mesh of the simulation

2.4 Bending test results The specimens after the V-bending test and the U-draw bending test for the as-received and 3D aluminum sheets are shown in Fig. 10. The results are summarized is in Table 2. Large springback is estimated with the flow curve in Fig. 6, because bending occurred in the small plastic strain range during the V-bending experiment. However, a similar springback to the as-received was estimated because bending occurred in the large plastic strain range in the U-draw bending experiment. This result was reasonable since the flow curve was almost the same between the 3D specimen and the asreceived specimen in the large plastic strain range.

3. Finite Element Simulation of the Tensile Test The tensile test simulations were carried out using ABAQUS Ver. 6.14. Fig. 11 shows the finite element modeling and mesh used for the analysis. The simulation was carried out on the flat specimen and the 3D specimen. The mesh of the specimen on the gauge section was 0.15 mm. The diameter of the embossing of the upper and lower dies was 1.5 mm. The half of the specimen was analyzed because it was

INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5

Fig. 12 The stress distribution of the finite element simulation after the embossing process

MAY 2016 / 675

Fig. 14 The stress-strain curves of the experimental data for 3D structured (solid line) and the finite element simulation under a tensile load for the flat specimen (dotted line), 3D-structured (dashed line), and 3D-structured specimen with anneal the stress after embossing process (dashed-dotted line)

Fig. 13 The shape of the 3D sheet under the tensile load at the total strain (a) 0.019, (b) 0.075, and (c) 0.18 of the specimen

symmetric. Reduced 4-node shell elements (S4R) with five integration points through the thickness direction were used for the specimen. The gauge section of the specimen was 0.5 mm-thick and 12.5 mm-wide. The upper and lower dies were defined rigid bodies. The tensile load was directly applied on the flat specimen on the gripping edge, and the 3D sheet specimen was analyzed using a twostep process. First, the specimen was embossed with the dies, and then the tensile load was applied on the edge. The stress distribution after the embossing process is shown in Fig. 12. The maximum stress is shown on the edge of the embossing die area. Fig. 13 shows the shape of the gauge section during the tensile load for the 3D sheet. The shape of the 3D structure has as similar pattern as the experiment. At the beginning of the tensile load, the embossing pattern started to flatten in the tensile load direction and maintained a wave shape in the transverse direction to the tensile load. As the tensile load increased, the embossed shape was connected to the neighboring embossed shape in the tensile load direction. Fig. 14 shows the stress-strain relation of the flat specimen and the 3D specimen. When the tensile load applied just after the embossing process, the flow stress on the 3D specimen was higher than on the flat specimen. This might have been caused by the hardening effect during the embossing process. When the stress release step added after the embossing process, the flow stress after the yielding showed a similar pattern as in the experiment. The yield point was not clear because of the flattening effect of the embossing. Just after the yielding, the flow stress was lower than that of the flat specimen, and as the plastic

Fig. 15 The height distribution of the 3D sheet as the strain (a) longitudinal and (b) transverse direction to the tensile direction

deformation increased, the flow stress followed the flow stress of the flat specimen. This behavior was due to the flattening and connecting effect of the embossing during the tensile plastic deformation, and the behavior in the range of large plastic strain was similar to that of the flat specimen because the embossing was flattened in the tensile direction. As shown in Fig. 15, the reduction of the height in the longitudinal direction is larger than that of the height in the transverse direction to the tensile load direction. The height in the longitudinal direction is dramatically reduced to 0.10 mm at 0.180 plastic strain from 0.37 mm at the beginning of the tensile test. The height reduction is 73% in longitudinal direction. However, the height in the transverse direction

676 / MAY 2016

INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5

large plastic strain range. In this study, a tensile specimen that had a 12.5 mm-wide gauge section was used. The size of the specimen should be studied in more detail for the 3D sheet because this specimen is optimized compared to flat sheet metal. In addition, the variation in the depth of the embossing will be studied further to obtain more detailed information for industrial applications.

ACKNOWLEDGEMENT Fig. 16 Measured plastic anisotropic coefficient compared with the simulation results for the parallel embossed specimen

is reduced to 0.60 mm at the same plastic strain from 0.75 mm. The height reduction is only 20%, which is a much smaller value than that in the transverse direction. As mentioned earlier, the anisotropic coefficient is defined as the ratio between the strain in the width direction and the strain in the thickness direction. The flattening phenomena of the 3D structure affect the anisotropic coefficient. The reduction of the height is the strain in the thickness direction, and the absolute value of strain is large in this direction while the absolute value of strain in the width direction is small because the reduction in the width direction is compensated due to the flattening of the embossed shapes. Fig. 16 shows the experimental and simulation results of the anisotropic coefficient as the strain for the parallel embossed specimen. The anisotropic coefficient shows a gradually increasing trend, except around 0.01 plastic strain in the experiment. The relatively large value of around 0.01 plastic strain seems to be an experimental error or variation because in this range, the small measurement variation of the length cause relative large variation in the anisotropic coefficient value. For example, a 0.01 mm smaller value in the width direction would result in a 0.1 larger value in the anisotropic coefficient. We believe that decreasing the compensation effect on the width reduction resulted in a larger anisotropic coefficient as the strain increased.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (No. NRF-2014R1A2A2A01005903).

REFERENCES 1. Kim, Y. S., Kim, K. S., and Kwon, N. C., “Press Formabilities of Aluminum Sheets for Autobody Application,” Proc. of the KSAE Conference, pp. 233-246, 1993. 2. Kim, Y. S., “Evaluation Apparatus and Method for Insulation Efficiency of the Heat Protector of the Automobile,” KR Patent, No. 10-1004202, 2010. 3. ArcelorMittal, “Steels Coated with Alusi® Aluminum-Silicon Alloy: Specific Applications,” http://automotive.arcelormittal.com/ europe/products/alusi/alusi_specific_applications/EN (Accessed 23 MAR 2016) 4. Noh, H.-G., An, W.-J., Song, W.-J., Kang, B.-S., and Kim, J., “Experimental and Numerical Study on Patterned Emboss Forming using Electromagnetic Forces,” Int. J. Precis. Eng. Manuf., Vol. 16, No. 7, pp. 1447-1454, 2015. 5. Namoco, Jr, C. S., Iizuka, T., Narita, K., Takakura, N., and Yamaguchi, K., “Effects of Embossing and Restoration Process on the Deep Drawability of Aluminum Alloy Sheets,” Journal of Materials Processing Technology, Vols. 187-188, pp. 202-206, 2007.

4. Conclusions

6. Namoco, Jr, C. S., Iizuka, T., Hatanaka, N., Takakura, N., and Yamaguchi, K., “Influence of Embossing and Restoration on the

In this study, the tensile properties of a 3D aluminum sheet were tested. The Young’s modulus and the appearance anisotropic coefficient, R-value, were reduced because of the flattening of the 3D shapes. The yield stress in the parallel embossed specimen was smaller and that in the diagonal embossed specimen was larger than that of the as-received specimen. In the tensile test, the flow stress of the 3D sheet in the range of the small plastic strain was smaller than that of the flat specimen. This behavior was analyzed using finite element simulation, and the results demonstrated similar behavior. The flow stress of the parallel embossed specimen was smaller than that of as-received flat specimen in the small plastic strain range, and the anisotropic coefficient also smaller than the as-received flat specimen because of the flattening of the 3D shapes. The springback was large within the small plastic strain range and almost the same as that for the as-received specimen in the

Mechanical Properties of Aluminum Alloy Sheets,” Journal of Materials Processing Technology, Vols. 192-193, pp. 18-26, 2007. 7. Fritzsche, S., Ossenbrink, R., and Michailov, V., “Experimental Characterisation of Structured Sheet Metal,” Key Engineering Materials, Vol. 473, pp. 404-411, 2011. 8. Cho, J. H., Choi, W. H., and Park, J. S., “Evaluation on the Mechanical Properties for Two Layer Composite Panel of 1xxx Series Aluminum,” Proc. of KSAE Conference, pp. 528-533, 2014. 9. Kim, M. K., “Mold Structure and Heat Protector for Heat Insulation of Automobile,” KR Patent, No. 10-2005-0042370, 2006. 10. Güler, H. and Özcan, R., “Effects of the Rotary Embossing Process on Mechanical Properties in Aluminum Alloy 1050 Sheet,” Metals

INTERNATIONAL JOURNAL OF PRECISION ENGINEERING AND MANUFACTURING Vol. 17, No. 5

and Materials International, Vol. 18, No. 2, pp. 225-230, 2012. 11. Abe, T., Yasota, T., Nonaka, Y., Saka, S., and Kuwabara, T., “Forming Simulation of Emboss Formation by Roll Forming,” Proc. of Japanese Spring Conference for the Technology of Plasticity, pp. 253-254, 2008. 12. Lee, S. W., “Prediction of Springback of DP590 Steel Sheets using Yoshida-Uemori Kinematic Hardening Model,” M.Sc. Thesis, Kyungpook National University, 2012. 13. Choi, J. G., Choi, S. C., Lee, M. G., and Kim, H. Y., “Measurement of Springback of AZ31B Mg Alloy Sheet in OSU Draw/Bend Test,” Transactions of Materials Processing, Vol. 16, No. 6, pp. 447-451, 2007.

MAY 2016 / 677