Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
com
Received 1 December 2007; received in revised form 23 December 2007; accepted 9 January 2008
Available online 4 March 2008
Abstract
Elasto-plastic finite element analysis was carried out for analyzing the severe plastic deformation behavior of copper specimens during
groove pressing. Deformation localization was studied in terms of strain variations along the longitudinal direction. Plastic strain is lower
at the local interface between the shear and the flat regions, which receives very little shear during the pressing cycle. Strain localization is
more intensified with the number of groove pressing cycles, although the average strain level increases.
Ó 2008 Elsevier B.V. All rights reserved.
Keywords: Finite element method; Severe plastic deformation; Groove pressing; Strain localization; Copper; Deformation homogeneity
0927-0256/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.commatsci.2008.01.007
642 S.C. Yoon et al. / Computational Materials Science 43 (2008) 641–645
Fig. 1. Schematic of groove pressing process showing 5 stages: (a) first pressing; (b) first flattening; (c) top die shifting towards RHS; (d) second pressing;
and (e) second flattening [5].
300
nificantly from those of the semi-coupling approach, the regions do not deform but the shear regions deform by
latter was taken in this study. the same amount of shear strain in the reverse direction.
The specimen with dimensions of 80 mm in length and The top die was shifted in the simulation towards RHS
5 mm in thickness was considered for simulations (width (or the specimen rotated about 180° in the plane of press-
is unity along the plane normal direction in plane-strain ing) by the width of the flat regions (= that of the shear
condition). The specimen was modeled with CPE4 mesh region 5 mm), between the first flattening and the second
and dies were modeled with analytical rigid lines. The coef- pressing steps, in order to impose deformation in the previ-
ficient of friction l between the die surface and the speci- ously undeformed flat regions after first pressing (unde-
men was selected as 0.1, which is within a typical range formed regions). It can be found that, the deformation
(0.05 0.1) in cold forming of metals. obtained in the first cycle was a little more homogeneous
after second flattening than after the first flattening; the dif-
ferences in strain e between the maximum and minimum
3. Results and discussion
points after the first flattening and the second flattening
are e = 1.23 and e = 1.0, respectively. Even though strain
The specimen was subdivided into three regions accord-
localization was relieved after the second cycle, compared
ing to the deformation modes see Fig. 3: shear region,
to the first cycle, strain is not as homogeneous as expected.
undeformed flat region, and interface region between the
The strain between the maximum and minimum points
shear and flat regions. Plastic deformation occurs mainly
after the second flattening was smaller than that in the first
in the shear region, where the theoretical Von Mises equiv-
flattening due to the differences in the relative strength of
alent strain per pressing is 0.58, assuming the deformation
the flat regions surrounding the shear regions. During the
mode is simple shear (not pure shear!) and there is no inter-
first pressing and flattening steps the relative strength of
action between shear region and flat regions. Fig. 4 shows
the flat regions (symbol d position) surrounding the shear-
the variation of Von Mises equivalent strain distributions
ing region (symbol N position) is weak, and deformation
in GP processed specimen after two cycles. The strain level
diffuses out of the shear region. On the other hand, during
increased with the number of passes. From Fig. 4, it is
the second pressing and flattening steps the relative
observed that the inhomogeneity within the specimen was
strength of the flat regions (symbol N position, which
higher as the number of cycles increased. Looking at the
was the former shear region) surrounding the shear region
first cycle, see Fig. 5, the Von Mises equivalent strain e val-
is high (symbol d position), and deformation is concen-
ues after first pressing and first flattening in the first shear
trated on the shearing regions. Hence, Von Mises equiva-
region are e = 0.86 and e = 1.5, respectively, which are lar-
lent strain in the shearing region after the second
ger than theoretical values e = 0.58 and e = 1.16 of the sim-
flattening (the first maximum value e = 1.81) is higher than
ple shear during the first cycle. This is because of the
that after the first flattening (the second maximum value
interaction between the shear region and flat region: more
e = 1.45). It should be noted that the strain e = 1.45 at
deformation is generated in the shear region in GP than in
the shearing region (symbol N position) does not change
the region of simple shear alone. Because of the interaction
after the second pressing and flattening. It means the first
between the shear region and flat region, the deformation
shear region hardly deforms during the second pressing
mode in GP is not simple shear. The flat regions deform
and flattening, because surrounding region (symbol d posi-
by a negligible amount of deformation during the pressing
tion) is already strong.
and flattening steps. During the flattening steps, the flat
One of the most important results of this simulation is
the existence of the strain minimum regions of symbol w
position. The strain minimum regions are the interfaces
between the shear and flat regions, as can be seen in
Fig. 5. These interface regions receive very little shear dur-
ing the pressing cycles and form the strain minimum points
of symbol w position. These interface regions are the main
cause for strain oscillation after each cycle of groove
pressing.
For the second cycle, the situation of strain localization
does not change. The differences in strain between the max-
imum and minimum points after first flattening and the sec-
ond flattening are 1.65 and 1.6, respectively. It should be
noted that strain distribution after the second flattening
is a little more homogeneous than that after the first flat-
tening, see Fig. 5. This difference in strains between the
maximum and minimum points after the first cycle and
Fig. 3. Schematic of groove pressing showing different deformation the second cycle, shown in Fig. 6, are 1.0 and 1.6, respec-
regions in the specimen. tively. In other words, strain localization is more intensified
644 S.C. Yoon et al. / Computational Materials Science 43 (2008) 641–645
Fig. 4. Von Mises equivalent strain distributions in groove pressing during two cycles.
First pressing
Second pressing
2.2
2.0
1.8
1.6
Effective Strain
1.4
1.2
1.0
0.8
0.6
0.4
0.2 First Flattening
Second Flattening
0.0
0 20 40 60 80
Distance (mm)
Fig. 5. Von Mises equivalent strain path plots along the horizontal line in the first cycle.
with the number of GP cycles, although the average strain same trend, because the stress gradient decreases with
level increases. strain. The nonuniformity of strain distribution shown in
Although the strain localization intensified with increas- Fig. 6 may not indicate the nonuniformity of the strength
ing number of GP cycles, stress need not always follow the (and microstructure) developed. This is shown in Fig. 7,
S.C. Yoon et al. / Computational Materials Science 43 (2008) 641–645 645
4.0 4. Conclusions
3.5
Finite element analysis of the groove pressing process
3.0 was carried out to determine the deformation behavior of
pure copper up to two pressing cycles. The strain distribu-
Effective strain
2.5 tion along the longitudinal length of the specimen was also
analyzed. The strain obtained after the second cycle is more
2.0
inhomogeneous than that after the first cycle. FEM analy-
1.5 sis incorporating a dislocation-based microstructural con-
stitutive model is underway.
1.0
2nd cycle Acknowledgements
0.5
1st cycle
0.0 This work was supported by Korea Science & Engineer-
0 10 20 30 40 50 60 70 80
ing Foundation through the NRL Program (R0A-2007-
Distance (mm)
000-20104-0). A.K acknowledges the financial support of
Fig. 6. Von Mises equivalent strain path plots along the horizontal line the BK21 center through BK postdoctoral fellowship.
after the first and second cycles.
References
400 Experimental 1st cycle Experimental 2nd cycle [1] R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881–981.
FE Analysis 1st cycle FE Analysis 2nd cycle
[2] C. Xu, Z. Horita, T.G. Langdon, Acta Mater. 55 (2007) 203–212.
350
[3] Y.H. Chung, J.W. Park, K.H. Lee, Metal. Mater. Inter. 12 (2006)
289–293.
Effective stress (MPa)
300
[4] C.Y. Lim, S.Z. Han, S.H. Lee, Metal. Mater. Inter. 12 (2006) 225–
250 231.
[5] D.H. Shin, J.J. Park, Y.S. Kim, K.T. Park, Mater. Sci. Eng. A328
200 (2002) 98–103.
[6] J.W. Lee, J.J. Park, J. Mater. Proc. Technol. 130–131 (2002) 208–213.
150
[7] A. Krishnaiah, U. Chakkingal, P. Venugopal, Mater. Sci. Eng. A410–
100 A411 (2005) 337–340.
[8] A. Krishnaiah, U. Chakkingal, P. Venugopal, Scr. Mater. 52 (2005)
50 1229–1233.
[9] J.J. Park, N.J. Park, J. Mater. Proc. Technol. 169 (2005) 299–307.
0 [10] K. Peng, L. Su, L.L. Shaw, K.W. Qian, Scr. Mater. 56 (2007) 987–
0 10 20 30 40 50 60 70 80
990.
Distance (mm) [11] S.C. Baik, R.J. Hellmig, Y. Estrin, H.S. Kim, Z. Metallk. 94 (2003)
754–760.
Fig. 7. Von Mises equivalent stress plots with experimental values [12] S.C. Baik, Y. Estrin, H.S. Kim, Mater. Sci. Forum 408 (2002) 697–
(assumed to be three times the hardness) [7] along the horizontal line 702.
after the first pass and the second pass. [13] H.S. Kim, M.H. Seo, S.I. Hong, J. Mater. Proc. Technol. 113 (2001)
622–626.
[14] H.S. Kim, M.H. Seo, S.I. Hong, J. Mater. Proc. Technol. 130 (2002)
where the calculated stress variations along the specimen 497–503.
length are presented, along with the variations of the mea- [15] H.S. Kim, M.H. Seo, S.I. Hong, Mater. Sci. Eng. A291 (2000) 86–90.
sured strength (assumed to be three times the hardness [16] H.S. Kim, Mater. Sci. Eng. A315 (2001) 122–128.
value) Ref. [7]. It is evident that the stress distribution [17] Abaqus/standard user’s manual version 6.6, Simulia Incorporation,
USA, 2004.
can be considered as relatively uniform, compared to the
[18] Y. Estrin, L.S. Toth, A. Molinari, Y. Brechet, Acta Mater. 46 (1998)
strain distribution. The experimental stress values esti- 5509–5522.
mated from the hardness are in reasonably good agreement
with the simulated results.