Computational & Technology Resources
an online resource for computational,
engineering & technology publications |
|
Civil-Comp Proceedings
ISSN 1759-3433 CCP: 88
PROCEEDINGS OF THE NINTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY Edited by: B.H.V. Topping and M. Papadrakakis
Paper 251
Crystal Plasticity Finite Element Modelling of Compression of Pure Aluminum Z.Y. Jiang1, H.J. Li1,2, J.T. Han2, D.B. Wei1, H.C. Pi3 and A.K. Tieu1
1School of Mechanical, Materials and Mechatronic Engineering, University of Wollongong, Australia
Z.Y. Jiang, H.J. Li, J.T. Han, D.B. Wei, H.C. Pi, A.K. Tieu, "Crystal Plasticity Finite Element Modelling of Compression of Pure Aluminum", in B.H.V. Topping, M. Papadrakakis, (Editors), "Proceedings of the Ninth International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 251, 2008. doi:10.4203/ccp.88.251
Keywords: uniaxial compression, crystal plasticity finite element method, true strain, true stress.
Summary
The deformation of polycrystalline is determined by the deformation of a single
crystal. The best way of dealing with the polycrystalline deformation behaviour is to
apply the proprieties of single crystal to the polycrystalline by the statistical method.
If the arrangement is disorderly and unsystematic without any obvious textures, the
deformations of polycrystalline can be obtained by averaging the deformation of the
single crystal. If there are some tendencies that the single crystal in the
polycrystalline materials arranges along a certain direction, the properties of
polycrystalline can be obtained by the weighted average. In the polycystic
constitutive model, the plastic deformation of polycrystalline can be obtained from
the average of single crystal plastic deformation. A rate dependent crystal plasticity
constitutive model [1] with respect to latent hardening in the finite element (FE)
analysis is developed to simulate the compression - upsetting of pure aluminium. To
compare the influence of the different finite element models on the simulation
results, the Taylor-Type and finite element polycrystalline models are respectively
employed in the finite element software ABAQUS to simulate the development of
the deformation texture using rate dependent crystal constitutive equations. In order
to obtain the random uniform equiaxed grains, the samples of the electron
backscatter diffraction (EBSD) were prepared with the method of Mao [2] and
Raabe et al. [3], the pure aluminum samples had been forged three times on
the three orthogonal dimensions with the deformations of 25%, 15% and 5%
respectively. The samples were annealed at 500°C. All the initial data from
the experiments were input into the UMAT subroutine of the ABAQUS software to
simulate the uniaxial compression (upsetting) of the aluminium plate [4]. With an
increase of true strain, the crystal can rotate, and form the silk textures easily.
Therefore, the predicted and experimental silk textures tend to be sharper and
stronger accordingly. The results from the two polycrystalline models are both close
to the experimental results. There is only one kind of silk texture (<110> texture)
formed during the uniaxial compression, and the silk axes locate in the centre of the
pole figure. At the same time, it is certain that there is a minority of crystals
distribute along the <110> and <113> axes. However, with the same true strain, the
stress values from the two modelling results are lower than those from the
experimental results. Comparison with the deformation contouring map of the pure
aluminium at the different strains, the finite element polycrystalline model can
predict the drum deformation of pure aluminium accurately during the uniaxial
compression. Whereas the deformation from the Taylor-type model is very even
without any drum deformation. The reason is that the Taylor-type model has a much
lower deformation flow stress than that of the finite element model. The slip systems
can be activated more easily than those of the finite element model.
References
purchase the full-text of this paper (price £20)
go to the previous paper |
|