Computational micromechanics of polycrystals

The present work deals with the constitutive modeling and development of computational frameworks for the simulation of the micromechanical behavior of polycrystalline materials. The contributions of the current thesis can be broadly classified along two different lines: (a) Modeling of microstructural and micromechanical behavior, and (b) Application to materials. Of particular interest, with reference to (b), are two important materials, magnesium alloy AZ31 and TWIP steels. From a modeling perspective, with reference to (a), both mean-field and full-field frameworks have been used in this work by extending them to model the desired phenomena. An enhanced twinning model has been implemented in the VPSC framework to improve the quality of the predicted textures. Likewise, a finite element model for full-field simulations has been extended to incorporate twinning; both shear due to twinning and reorientation effects form a part of this model. Additionally, a recrystallization model has also been implemented in the VPSC framework thus enabling the texture prediction of high temperature deformation. Both these frameworks are used to study the micromechanical behavior of the two stated materials – AZ31 and TWIP steels. The final part of the thesis deals with a novel full-field algorithm based on fast Fourier transforms, which appears to be computationally a more efficient algorithm – although restricted to periodic microstructures – than its finite element counterpart.