Sprecher
Beschreibung
Understanding the micromechanical behavior and degradation mechanisms of duplex stainless steels, especially under hydrogen exposure, is crucial for designing safe and durable materials for the green energy transition. However, investigating local strain partitioning and hydrogen-assisted damage requires realistic three dimensional (3D) representative volume elements (RVEs) that capture both anisotropic, phase-banded spatial morphologies and grain-resolved crystallographic statistics.
This work introduces a novel methodology that combines the reconstruction of grain-resolved two-phase RVEs considering lower and higher-order grain and phase statistics and enables virtual material modeling without relying on expensive, destructive 3D characterization. First, a 3D microstructure containing the spatial phase distribution as a binary phase mask is optimized from 2D EBSD data using directional two-point correlation functions in MCRpy. Second, this phase mask serves as a spatial boundary condition in Kanapy to drive phase-constrained stochastic grain packing and orientation distribution assignment. A distance-transform clearance check and penalized voxelization routine prevent unphysical grain penetration across phase interfaces.
Validated on DIN 1.4462 duplex stainless steel across three heat-treatment states (1050 °C, 1150 °C, and 1250 °C), the workflow accurately reconstructs the phase-banded morphology and polycrystalline aggregate, e.g. grain size, orientation distribution. By serving as input for micromechanical finite-element simulations, these 3D RVEs enable the systematic study of localized phase-boundary interactions and microstructural susceptibility to hydrogen embrittlement. This novel methodology advances virtual materials engineering, enabling targeted microstructural optimization for critical applications in future energy systems.
