Reconstruction methods and analysis of subsurface uncertainty for anisotropic microstructures

Abstract

In the present work, strain localization at the surface of a heavily textured material is modeled, where the subsurface microstructure was statistically reconstructed via a proposed methodology that directly uses the information given by the surface characterization. The elasto-viscoplastic simulations based within a fast Fourier transformation formulation (EVP-FFT) were compared with digital image correlation quantification of strain coupled with electron backscatter diffraction characterization of the microstructure to produce strain maps of the surfaces, and the in-plane maximum shear strain was evaluated. When modeling specimens from rolled plate of AA7050 material, both with and without a 3D subsurface reconstruction, a reasonable improvement was observed when a statistically reconstructed subsurface morphology was included in the simulations (regardless of grain morphology or orientation distribution) since it allowed the surface grains to redistribute and accommodate deformation along the third dimension. Additionally, when performing a case study on fabricated specimens that exhibited a through thickness grain structure, in order to minimize the subsurface uncertainty, the model's ability to capture the strain localization improved, thereby confirming the importance of capturing the subsurface microstructure on the surface response. These results therefore quantify the effect of the subsurface grain structure on the surface strain fields and thereby present a path forward for performing crystal plasticity simulations by presenting a subsurface reconstruction methodology based on surface characterizations.