Institute of Fundamental Technological Research

This paper describes an application of a machine learning approach for parameter optimization. The method is demonstrated for the elasto-viscoplastic model with both isotropic and kinematic hardening. It is shown that the proposed method based on long short-term memory networks allowed a reasonable agreement of stress–strain curves to be obtained for cyclic deformation in a low-cycle fatigue regime. The main advantage of the proposed approach over traditional optimization schemes lies in the possibility of obtaining parameters for a new material without the necessity of conducting any further optimizations. As the power and robustness of the developed method was demonstrated for very challenging problems (cyclic deformation, crystal plasticity, self-consistent model and isotropic and kinematic hardening), it is directly applicable to other experiments and models.

crystal plasticity, optimization, machine learning, long short-term memory networks, self-consistent modeling, Eshelby solution, cyclic deformation, low cycle fatigue

Recently developed single-phase concentrated solid-solution alloys (CSAs) contain multiple elemental species in high concentrations with different elements randomly arranged on a crystalline lattice. These chemically disordered materials present excellent physical properties, including high-temperature thermal stability and hardness, with promising applications to industries at extreme operating environments. The aim of this paper is to present a continuum plasticity model accounting for the first time for the behaviour of a equiatomic five-element CSA, that forms a face-centred cubic lattice. The inherent disorder associated with the lattice distortions caused by an almost equiatomic distribution of atoms, is captured by a single parameter that quantifies the relative importance of an isotropic plastic contribution to the model. This results in multiple plasticity mechanisms that go beyond crystallographic symmetry-based ones, common in the case of conventional single element metals. We perform molecular dynamics simulations of equiatomic CSAs: NiFe, NiFeCr, NiFeCrCo, and Cantor alloys to validate the proposed continuum model which is implemented in the finite element method and applied to model nanoindentation tests for three different crystallographic orientations. We obtain the representative volume element model by tracking the combined model yield surface.

High entropy alloys, Nanoindentation, Molecular dynamics, Finite element method, Crystal plasticity

Identification of elastic and plastic properties of materials from indentation tests received considerable attention in the open literature. However, unambiguous and automatic determination of parameters in the case of the crystal plasticity (CP) model is still an unsolved problem. In this paper, we investigate the possibility to unambiguously identify the CP parameters from spherical indentation tests using finite element method simulations combined with evolutionary algorithm (EA). To this aim, we check the efficiency and accuracy of EA while fitting either load–penetration curves, surface topographies, or both at the same time. By fitting the results against simulation data with known parameters, we can verify the accuracy of each parameter independently. We conclude that the best option is to fit both load–penetration curve and surface topography at the same time. To understand why a given fitting scheme leads to correct values for some parameters and incorrect values for others, a sensitivity analysis was performed.

crystal plasticity,optimization,evolutionary algorithm,indentation

Fe-based alloys with high chromium and nickel concentrations are very attractive for efficient energy production in extreme operating conditions. We perform molecular dynamics (MD) simulations of nanoindentation on fcc FeNiCr multicomponent materials. Equiatomic FeNi, Fe55Ni19Cr26, and Fe74Ni8Cr18 are tested by using established interatomic potentials and similar conditions, for the elucidation of key dislocation nucleation mechanisms and interactions. Generally, we find that the presence of Cr in these alloys reduces the mobility of prismatic dislocation loops, and increases their area, regardless of crystallographic orientation. Dislocation nucleation and evolution is tracked during mechanical testing as a function of nanoindentation strain and Kocks–Mecking continuum modeling displays good agreement with MD findings. Furthermore, the analysis of geometrically necessary dislocations (GNDs) is consistent with the Ma–Clarke’s model at depths lower than 1.5 nm. The presence of Cr leads to a decrease of the GND density with respect to Cr-less FeNi samples, thus we find that Cr is critically responsible of increasing these alloys’ hardness. Post-indentation impression maps indicate that Ni–Fe–Cr compositions display strain localization and hardening due to high Cr concentration.