Artificial microstructure generator for dual-phase steels

Abstract

The microstructure of low alloyed ferritic-martensitic Dual-Phase (DP) steels consists of hard coarse grained martensite islands embedded in a soft ferrite matrix. Therefore, the macroscopic mechanical properties of DP steels mostly derive from their microstructures, such as volume fractions, morphology of martensite, phase distributions and ferrite grain size. Recently, micromechanical approaches are used to predict ductility and failure mode of DP steels under varying mechanical loading scenarios. In this work, an artificial microstructure generator inspired by topology optimization was developed to construct representative volume element (RVE) with predefined design parameters within a mofidied Voronöi tessellation. Micromechanical modeling of DP steel was performed on the generated artificial RVE. The plastic flow behavior of each single phase in DP steel were calculated by using a dislocation based theory. After numerical simulation, the flow curve on macro scale can be obtained from an asymptotic expansion homogenization (AEH) scheme. This approach allows studying the influence of individual microstructure features on local and global stress-strain response. To improve the robustness of this artificial microstructure generator, a proper orthogonal decomposition (POD) reduction was introduced to identify the optimal design parameters. This approach used a collection of artificial microstructures (snapshots) to find the most representative one of the real microstructure.