A comparative study of severe plastic deformation techniques by finite element analysis.

Comaneci, Radu

1. INTRODUCTION

Severe Plastic Deformation methods were developed as a new way of manufacturing UFG materials. Among various SPD methods, ECAP and MF processes are specially designed to obtain UFG structures in bulk materials (Valiev et al., 2006).

ECAP process involves the extruding of the workpiece through a die consisting of two channel intersecting at a specify angle [phi] (see Fig.1). Because of identical cross section of the two channels, one can re-insert the billet into the entry die channel and impose more severe plastic strain.

[FIGURE 1 OMITTED]

In the vertical channel, the billet moves as a rigid body while all deformation is localized in the small area around the channel's meeting line (the bisect plane). The metal is subjected to a simple shear strain which produces the refinement of the structure. The billet removal involves a new development of ECAP procedure. The theoretical effective strain s according to the die geometry (Fig.1) for each pass is given in Eq. (1):

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

MF process involves multiple forging by changing the axis of the applied strain between passes (see Fig.2) by billet rotation along the axis.

The geometry of specimen and the die channel are in principal designed in that way, that the material flows only in direction of the channel at applied compressive load--thus the width of channel and sample have to be equivalent, Figure 2. Additionally the geometry of the specimen remains about equal before and after a processing step if initial height of the specimen and length of channel are identical. Thus, it is possible to subject the same specimen several times to the MF process by rotating the material by 90[degrees] around the transverse direction and/or normal direction of the channel.

[FIGURE 2 OMITTED]

The theoretical effective strain s of the workpiece in each pass, obtained by compression in the die channel is given by Eq. (2):

[epsilon] = [2/[square root of 3]ln (H/W) (2)

In both procedures, the samples cross section are not changed during the processing which allows to induce large plastic strain into material until the accumulated deformation reaches a desired level.

Comparison of different SPD techniques was very rarely reported previously and only in terms of achieved properties (Cherukuri et al., 2005) or microstructure changes (Cabibbo et al., 2007). The aim of this paper is the comparison of two processes in terms of load and strain distribution by FEA.

2. FINITE ELEMENT ANALYSIS

Annealed Al99,7% specimen with dimensions of 10mmx10mmx60mm was used. Simulations were based of the constitutive curves and material constants from DEFORM 3D material library. The deformations were conducted at RT. The friction factor between the inner surfaces of the die channels and the specimen was assumed to be 0.12. This value was experimentally established. All simulations were carried out at a constant speed of 8.75mm/s, for a 50mm stroke of punch. The isothermal conditions can be fulfilled at that low pressing speed. The workpiece was dicretized in 8000 tetrahedral elements. The tolerance, positioning of both workpiece and top/bottom die, convergence criteria, re-meshing conditions, and boundary conditions must be specified before the execution of the simulation process. In order to estimate strain distribution, a few tracking points ([P.sub.1], [P.sub.2] ...) are defined in longitudinal section of the workpiece.

3. RESULTS AND DISCUSSION

A simulated load-displacement curve for Al99.7 in the mentioned conditions shows that the load level increases dramatically during the last stage of the full-filling of the die channel for MF process. (see Fig.3). That is typically for the die work process. The load level for ECAP process is instead about five times lower.

[FIGURE 3 OMITTED]

Note that theoretical effective strain for a pass and for the sample with dimensions of 10x10x60mm, has the value of 2.06, according to Eq. (2), in frictionless conditions. The strain becomes maximum (2.30) at the contact region of material with the punch ([P.sub.2]) and it is minimum (1.03) at the contact with bottom die (P8). In the steady state, the strain is in the range of 0.73 ... 1.42. That is explained taking into account the friction: because of this, the strain for all tracking points (except [P.sub.2]) is smaller than the theoretical value. As we see, dead zone formation ([P.sub.2], P5, P8 region) is the major factor for non-uniform strain distribution in MF of strain hardening material. Lack of free flow of the sample causes strain heterogeneity.

[FIGURE 4 OMITTED]

Between the inevitable non-uniformly deformed head and tail region, Fig.5 shows clearly a steady-state region in ECAP, where the strain is uniform along the axis. Within this steady-state region, however, deformation inhomogeneity still exists along the transverse direction (e.g. bottom to top along the width) to a degree that varies with material and *P (Li et al., 2004). In ECAP, the channel angle plays a very important role in distribution of strain and material flow. Generally, larger channel angle leads to lesser average strain but better homogeneity (Nagasekhar et al., 2007).

[FIGURE 5 OMITTED]

4. CONCLUSIONS

According to FEA results, there are important differences between ECAP and MF processes in terms of estimated load and strain distribution. In both cases, as we expect, the mean strains are under theoretical values because of strain hardening, friction conditions and gap formation. The estimated strains are in good agreement with real experiments.

The ECAP process has the advantage of smaller forces and more homogenous strain reached in an earlier steady state of plastic deformation.

For the same imposed strain it is preferably to perform ECAP process with extrapasses than MF process even if the last one is more effectively in strain terms.

5. REFERENCES

Cabibbo, M.; El-Mehtedi, M.; Scalabroni, C.; Balloni, L. and Evangelista, E. (2007). A comparative study of microstructural refinement by torsion-compression and equal channel angular pressing, Metallurgia Italiana, Vol 99, Issue 7-8, 2007, pp. 9-14

Cherukuri, B.; Nedkova, T.S. and Srinivasan, R. (2005). A comparison of the properties of SPD-processed AA-6061 by equal channel angular pressing, multi-axial compressions/forgings and accumulative roll bonding, Materials Science and Engineering A, 410-111, 2005, pp. 394-397

Li, S.; Bourke, M.A.M.; Beyerlein, I.J.; Alexander, D.J. and Clausen, B. (2004). Finite element analysis of the plastic deformation zone and working load in equal channel angular extrusion, Materials Science and Engineering A, 382, 2004, pp. 217-236

Nagasekhar, A.V.; Tick-Hon, Y. and Seow, H.P. (2007). Deformation behavior and strain homogeneity in equal channel angular extrusion/pressing, Journal of Materials Processing Technology 192-193, 2007, pp. 449-452

Valiev, R.Z.; Estrin, Y.; Horita, Z.; Langdon, T.G.; Zehetbauer, M.J. and Zhu, Y.T. (2006). Producing Bulk Ultrafine-Grained Materials by Severe Plastic Deformation, JOM, Vol. 58 (4), 2006, pp. 33-39