Micromechanics of large deformations

Understanding of the influence of microstructural features on mechanical properties of materials is essential for designing new materials and optimizing manufacturing process like sheet metal forming process. To gain the required knowledge, a large series of experiments are typically employed, which is economically rather inefficient. In this context, micromechanical modeling supports the development of required knowledge, by predicting the mechanical behavior of materials through microstructure-based simulations. This work aims to demonstrate the possibility of using micromechanical modeling to understand the influence of microstructural features on the mechanical behavior of materials at different scales. First, we introduce a software package to generate synthetic microstructure models which can mimic grain geometry and texture. For the latter, we develop a method to reconstruct the orientation distribution function (ODF) and the misorientation distribution function (MDF). With respect to the material model, a non-local crystal plasticity model is incorporated to describe the plastic deformation of the polycrystal. However, a number of material parameters are large which result in great difficulty in defining a unique set of material parameters. We hence develop a robust optimization scheme to parameterize a non-local crystal plasticity model from nanoindentation test. A synthetic microstructure is then treated as a model for finite element analysis with a parameterized non-local crystal plasticity. With this way, we are capable of predicting mechanical properties of the material such as anisotropic plasticity and strain hardening behavior under complex load paths. We can also investigate the influence of key microstructural features such as grain geometry and textures, and defects like pores on corresponding mechanical properties. Furthermore, local deformation fields of the location of interest are captured from the macroscopic simulation of the bending process and imposed as boundary conditions on a synthetic microstructure to study the changes in the microstructure during the manufacturing process. Finally, microstructurebased simulations are used to create training data for machine learning models that are able to predict microstructural properties to a given flow curve