Overview
A microstructure-based model was developed to simulate the mechanical behaviour of polycrystalline Ni-based superalloys containing gamma’ and gamma’’ precipitates and processed by casting and forging.
The model was based on a multiscale approach in which deformation and failure mechanisms as well as microstructural features and defectology are progressively incorporated at three different levels: micron-sized single crystals and small size polycrystals, polycrystalline specimens and components. In this way, the microstructural features which control the mechanical performance (precipitate structure, grain size, texture, porosity, surface condition, etc.) can be taken into account at the appropriate length scale.
The basic tool to predict the mechanical performance of polycrystalline specimens were the finite element simulation of a representative volume element of the microstructure. Crystal plasticity models for Ni-based superalloys were used to simulate the grain behaviour and the model parameters (as well as the grain boundary properties) were obtained from micromechanical tests on single crystals and bicrystals milled from the polycrystalline specimens by focus ion beam in both cast and forged materials. As opposed to purely phenomenological models, relevant microstructural parameters (grain size, texture, etc.), process-specific defects (shrinkage porosity, inclusions, light etching features, etc.), and surface condition can be accounted for in this strategy by modifying the geometric features of the representative volume element.
The proposed model was able to address the effect of temperature (from room temperature up to 700ºC) in the mechanical properties used in the design of components: tensile strength, fatigue, crack propagation and creep. In addition, statistical aspects associated with the scale up from polycrystalline specimens to actual components were incorporated.
Funding
Results
Executive Summary:
A microstructure-based simulation tool has been developed to predict the mechanical behavior of the polycrystalline Ni-based superalloy Inconel 718 under monotonic loading, creep and fatigue. The tool is based in numerical homogenization, linking polycrystalline microstructure (grain size, shape and orientation distributions) and crystal behavior with the macroscopic alloy response. The numerical homogenization is performed by finite element simulations of periodic representative volume elements (RVE) of the microstructure, subjected to a macroscopic strain or stress history: monotonic loading, cyclic loading or creep. The grain behavior is model by crystal plasticity (CP) models developed to represent the monotonic and cyclic response of In718 crystals. The CP models include both physically based and phenomenological assumptions and include the effect of both temperature and grain size. The parameters of the CP model are directly measured from micro-testing on grains (pillar compression) or, if not possible, adjusted by inverse fitting from macroscopic experimental data.
The microstructure based simulation tool is able to predict the mechanical behavior of an In718 wrought alloy as function of its actual microstructure and operation temperature. For monotonic loading, the tool predicts the stress-strain curve for any temperature, strain rate and microstructure, providing also some design parameters as yield stress and ultimate tensile stress. For an applied constant stress (creep), the models can provide the strain versus time response for a given microstructure and temperature. In the case of cyclic loading, the model is able to predict the material response cycle by cycle including all cyclic characteristics of the alloy: hardening, Bauschinger effect, cyclic softening, ratcheting and mean stress relaxation. Finally, fatigue life models are developed, based on the tool to simulate cyclic behavior and that predict the number of cycles for crack initiation or sample fracture for a given cyclic loading conditions. The tools and models have been tested and adjusted for In718 with different microstructures and at different temperatures. The tool was able to accurately predict the material behavior in all the cases.
In addition to this deterministic tool, the effect of random distributions of defects (hard particles, voids and surface roughness) has been accounted in the method to provide a stochastical simulation tool, specially for fatigue life prediction. This tool provides a dispersion of fatigue life results for a given level and type of defects and the inherent experimental scatter of fatigue results can be obtained in addition to the average fatigue life.