128
A. Cruzado et al.
a dramatic reduction in the number of experimental tests to characterize new alloys
and will also provide more accurate design allowables because the microstructural
effects can be accounted for.
Computational homogenization of polycrystals is a simulation strategy in which
the effective or macroscopic properties of the polycrystal are obtained by means
of the numerical simulation of a representative volume element (RVE) of the
microstructure [63]. The two key ingredients to make accurate predictions within
this framework are the description of the microstructure (including the grain size
and shape distribution and the texture) and the mechanical properties of the single
crystals. The former can be easily obtained nowadays using different experimental characterization techniques (including X-ray diffraction, optical and scanning
electron microscopy, electron back scattering diffraction, X-ray microtomography),
while crystal plasticity provides a very accurate model of the kinematics of the
plastic deformation of single crystals. In the case of complex alloys, phenomenological crystal plasticity models are normally used to determine the initial value
of the critical resolved shear stress to produce plastic deformation in each slip
system as well as the hardening rate. These model parameters have to be estimated
using different approaches. For instance, they are optimized from a set of estimated
starting values using an inverse optimization strategy by comparison of the predicted
mechanical properties with experimental results on polycrystals [17, 27, 29] or
on single crystals (extracted from the polycrystal) deformed by nanoindentation
[47, 60, 77] or by micropillar compression [14].
In this chapter, a systematic application of computational homogenization of
polycrystals to predict the mechanical behavior of IN718 Ni-based superalloy is
presented. Section 2 describes the microstructure of the alloy, while the experimental techniques to measure the mechanical properties of the single crystals and of the
polycrystals are detailed in Sect. 3. The computational homogenization framework,
including the digital representation of the microstructure, the crystal plasticity
models, and the details of the numerical simulations, is explained in Sect. 5.
Applications to predict the mechanical behavior of IN718 under monotonic tension
and fatigue are presented in Sect. 6, while the main conclusions are summarized in
Sect. 7.
2 Material Description
Polycrystalline IN718 is a Ni-Fe-based superalloy widely used for structural
applications up to 650–700 ◦ C because of its good castability and weldability, high
strength, and corrosion resistance. The microstructure of the wrought IN718 is
made up by a Ni FCC solid solution which contains a dispersion of nm-sized γ
(Ni 3 (Al,Ti)) and γ (Ni 3 Nb) coherent precipitates within the grains together with
μm-sized metal carbides and δ phase (Ni 3 Nb) particles at grain boundaries (Fig. 1)
[55, 68]. The volume fractions of γ and γ phases are in the range 3–5% and 10–
A. Cruzado et al.
a dramatic reduction in the number of experimental tests to characterize new alloys
and will also provide more accurate design allowables because the microstructural
effects can be accounted for.
Computational homogenization of polycrystals is a simulation strategy in which
the effective or macroscopic properties of the polycrystal are obtained by means
of the numerical simulation of a representative volume element (RVE) of the
microstructure [63]. The two key ingredients to make accurate predictions within
this framework are the description of the microstructure (including the grain size
and shape distribution and the texture) and the mechanical properties of the single
crystals. The former can be easily obtained nowadays using different experimental characterization techniques (including X-ray diffraction, optical and scanning
electron microscopy, electron back scattering diffraction, X-ray microtomography),
while crystal plasticity provides a very accurate model of the kinematics of the
plastic deformation of single crystals. In the case of complex alloys, phenomenological crystal plasticity models are normally used to determine the initial value
of the critical resolved shear stress to produce plastic deformation in each slip
system as well as the hardening rate. These model parameters have to be estimated
using different approaches. For instance, they are optimized from a set of estimated
starting values using an inverse optimization strategy by comparison of the predicted
mechanical properties with experimental results on polycrystals [17, 27, 29] or
on single crystals (extracted from the polycrystal) deformed by nanoindentation
[47, 60, 77] or by micropillar compression [14].
In this chapter, a systematic application of computational homogenization of
polycrystals to predict the mechanical behavior of IN718 Ni-based superalloy is
presented. Section 2 describes the microstructure of the alloy, while the experimental techniques to measure the mechanical properties of the single crystals and of the
polycrystals are detailed in Sect. 3. The computational homogenization framework,
including the digital representation of the microstructure, the crystal plasticity
models, and the details of the numerical simulations, is explained in Sect. 5.
Applications to predict the mechanical behavior of IN718 under monotonic tension
and fatigue are presented in Sect. 6, while the main conclusions are summarized in
Sect. 7.
2 Material Description
Polycrystalline IN718 is a Ni-Fe-based superalloy widely used for structural
applications up to 650–700 ◦ C because of its good castability and weldability, high
strength, and corrosion resistance. The microstructure of the wrought IN718 is
made up by a Ni FCC solid solution which contains a dispersion of nm-sized γ
(Ni 3 (Al,Ti)) and γ (Ni 3 Nb) coherent precipitates within the grains together with
μm-sized metal carbides and δ phase (Ni 3 Nb) particles at grain boundaries (Fig. 1)
[55, 68]. The volume fractions of γ and γ phases are in the range 3–5% and 10–