Microscale Testing and Characterization Techniques for Benchmarking Crystal. . .
115
the weakest link effect being more prominent in the latter. The cube volume of a
SERVE is more likely to capture single crystal-like behavior and show the full range
of feasible single crystal orientations and strengths. Conversely, the rectangular
microtensile geometry has multiple grains along the sample length and is much
less likely to exhibit strength values that exceed the bulk strength. The small cubes
also contain a higher fraction of surface grains, which must be properly accounted
for in the free boundary condition of the SERVE. These combined factors provide
some understanding for the observed discrepancy and may be used to develop more
detailed physics-based CPFEM models.
The progress reported in this chapter is summarized as follows. An experimental
methodology for machining and testing microtensile samples of Ni-base superalloy
René 88DT across multiple length scales was developed, with the intention of
studying the effect of sample size on yield strength. The expectation that sample
microstructure plays a larger role at smaller sample sizes was confirmed. The
yield strength was observed to decrease with sample size, and increased scatter
in yield strength values was correlated with decreased sample size. Contrary to
initial expectations, none of the tested samples exhibited greater strength than
the bulk value for this material. The trends seen in the experimental data were
studied using simulations with varying levels of complexity. It was shown that the
distribution of maximum Schmid factor for randomly orientated grains in an FCC
alloy is biased towards high Schmid factors, and that geometric averaging of these
grains in ever smaller volumes leads to decreased strength and increased scatter.
The geometry of the sample can also have an effect, as seen by comparison of
the SERVE predictions and the experimental results. These results provide insight
into the influence of polycrystalline microstructure on the mechanical properties
of an alloy and a means of quantifying this behavior to inform selection of RVEs
and the creation of multiscale models. Further work on mesoscale oligocrystals
shows great promise in facilitating explicit comparisons of CPFEM simulations and
experimental results. An end goal for benchmarking the CPFEM models would be
the measurement and modeling of an explicit oligocrystalline specimen where every
grain orientation, size, and shape is known and can be recreated in silico.
5 Orientation and Deformation Maps
Traditional methods for qualifying structural materials using bulk material testing
provide global properties but fail to capture the detailed underlying microstructural
dependencies that can now be included in high-fidelity multiscale models. Measured
global properties such as yield strength, elastic modulus, or strain to failure are
valuable benchmarks for deformation modeling, but recent advances in characterization techniques have made it possible to obtain local microstructural details
and scale-specific benchmarks. Here, we focus on efforts undertaken to perform
detailed microstructural characterization of René 88DT samples in 2D and 3D.
The distribution of local surface strains is a direct output of CPFEM and is also
115
the weakest link effect being more prominent in the latter. The cube volume of a
SERVE is more likely to capture single crystal-like behavior and show the full range
of feasible single crystal orientations and strengths. Conversely, the rectangular
microtensile geometry has multiple grains along the sample length and is much
less likely to exhibit strength values that exceed the bulk strength. The small cubes
also contain a higher fraction of surface grains, which must be properly accounted
for in the free boundary condition of the SERVE. These combined factors provide
some understanding for the observed discrepancy and may be used to develop more
detailed physics-based CPFEM models.
The progress reported in this chapter is summarized as follows. An experimental
methodology for machining and testing microtensile samples of Ni-base superalloy
René 88DT across multiple length scales was developed, with the intention of
studying the effect of sample size on yield strength. The expectation that sample
microstructure plays a larger role at smaller sample sizes was confirmed. The
yield strength was observed to decrease with sample size, and increased scatter
in yield strength values was correlated with decreased sample size. Contrary to
initial expectations, none of the tested samples exhibited greater strength than
the bulk value for this material. The trends seen in the experimental data were
studied using simulations with varying levels of complexity. It was shown that the
distribution of maximum Schmid factor for randomly orientated grains in an FCC
alloy is biased towards high Schmid factors, and that geometric averaging of these
grains in ever smaller volumes leads to decreased strength and increased scatter.
The geometry of the sample can also have an effect, as seen by comparison of
the SERVE predictions and the experimental results. These results provide insight
into the influence of polycrystalline microstructure on the mechanical properties
of an alloy and a means of quantifying this behavior to inform selection of RVEs
and the creation of multiscale models. Further work on mesoscale oligocrystals
shows great promise in facilitating explicit comparisons of CPFEM simulations and
experimental results. An end goal for benchmarking the CPFEM models would be
the measurement and modeling of an explicit oligocrystalline specimen where every
grain orientation, size, and shape is known and can be recreated in silico.
5 Orientation and Deformation Maps
Traditional methods for qualifying structural materials using bulk material testing
provide global properties but fail to capture the detailed underlying microstructural
dependencies that can now be included in high-fidelity multiscale models. Measured
global properties such as yield strength, elastic modulus, or strain to failure are
valuable benchmarks for deformation modeling, but recent advances in characterization techniques have made it possible to obtain local microstructural details
and scale-specific benchmarks. Here, we focus on efforts undertaken to perform
detailed microstructural characterization of René 88DT samples in 2D and 3D.
The distribution of local surface strains is a direct output of CPFEM and is also