Micromechanical simulation approach of dual phase steel artificial microstructure using random field model

Abstract

In this study, a random field (RF) model with a Gaussian kernel was applied to generate an artificial microstructure of dual phase (DP) steels. Micrographs obtained from Scanning Electron Microscopy (SEM) were analyzed using image processing software to extract the grain size and the volume fraction of each phase. Based on watershed (Ws) segmentation and quantitative analysis, the real and artificial microstructures were compared by analyzing grain features related the solidity, grain size and aspect ratio (the proportional relationship between its width and its height). Consequently, this approach allows to simulate the overall stress-strain behavior of the analyzed microstructures. As a result, it was shown that the strain localization starts to develop at the ferrite/martensite interface and that the RF model could replicate the micromechanical behavior of DP steels.