The 9th International Conference on Multiscale Materials Modeling

Presentation information

Symposium

E. Deformation and Fracture Mechanism of Materials

[SY-E5] Symposium E-5

Wed. Oct 31, 2018 9:45 AM - 11:00 AM Room2

Chairs: Anthony D Rollett(Carnegie Mellon University, United States of America), Evgeniya Kabliman(Austrian Institute of Technology, Austria)

[SY-E5] Use of FFT-based micromechanical modeling for analysis of synchrotron-based diffraction experiments

Invited

Vahid Tari1, Ricardo A Lebensohn2, Reeju Pokharel2, Rachel E Lim1, Darren C Pagan3, Yufeng Shen1, Joel V Bernier4, Robert M Suter1, Anthony D Rollett1 (1.Carnegie Mellon University, United States of America, 2.Los Alamos National Laboratory, United States of America, 3.Cornell High Energy Synchrotron Source, United States of America, 4.Lawrence Livermore National Laboratory, United States of America)

This presentation describes use of image-based modeling to analyze synchrotron-based High Energy Diffraction Microscopy (HEDM) stress-strain experiments. HEDM is both in-situ and non-destructive and was used to measure micromechanical fields such as strain and orientation at the grain scale developed under macroscopic tensile loading of Ti-7Al. Taking the 3D image of the experimentally measured initial microstructure as input, elasto-viscoplastic modeling based on the Micromechanical Analysis of Stress-Strain Inhomogeneities with Fourier transforms (MASSIF) was used to compute the micromechanical fields that develop during loading. To validate the MASSIF calculations, we compared the calculated fields with the ones measured by HEDM. The initial comparisons showed that MASSIF can reproduce the macroscopic stress/strain curve but poor agreement was found between calculated and measured fields at the grain scale. The differences at the grain scale were hypothesized to be caused by the initial residual stress state that was induced during prior material processing, and which was not incorporated in the MASSIF calculation. We used eigenstrain concept to incorporate residual stress in the MASSIF calculation by converting it to an initial eigenstrain field. The results reveal that incorporation of residual stress results in good agreement between calculated and measured fields at the grain scale, thereby validating the computational approach. MASSIF was further used to model an experiment in which a polycrystalline sample of Ti-7Al was cyclically deformed and mapped using boxbeam near field(nf)-HEDM and far field(ff)-HEDM. The ff-HEDM results show a decrease in residual elastic strain over the first cycle followed by a steady increase in (elastic) strain. The initial residual strain in each grain was anti-correlated with the change in that same strain component over the first cycle. The distribution in the von Mises stress, which is a scalar measure of deviatoric stress, broadens as cycles accumulates and develops a long upper tail. Initially, a graph of hydrostatic stress against stress coaxiality angle shows a positive correlation but after about 65 cycles, the trend-line has rotated to a negative correlation. Simulations with MASSIF are used to analyze and understand these results.