Materials Affected Manufacturing: Simulation of the Crystallographic Texture, Microstructure, and Inelastic Mechanical Properties of Additively Manufactured Polycrystalline Materials

Abstract

Additive manufacturing (AM) has attracted huge attention in the past decade due to its capability of producing complex geometries at a much less cost than conventional methods. The additively manufactured parts are being built in a layer-by-layer manner. However, it is difficult to control their mechanical properties due to the intense heat input by laser and the rapid melting/solidification process. There are many numerical and analytical models that can predict the mechanical behavior of materials based on the microstructure of that material. But, these models are very time consuming, especially in the case of AM, the microstructure is very anisotropic with very large grains (on the scale of 100 microns), and therefore, a sufficient size of microstructure can be in the scale of mm which the current finite element models cannot deal with this much data. Meanwhile, producing new samples with AM for a wide range of AM process parameters, then followed by sample preparation and material characterization is also very expensive. In this thesis, to tackle the above challenges, a quantitative model was developed for predicting the crystallographic texture of additively manufactured parts and then combined with a kinetics Monte-Carlo model to simulate the microstructure. This model used sufficient thermodynamics/mechanics terms that made it accurate and time-efficient at the same time. Also, in this work, an effort was made to correlate the inelastic mechanical behavior of additively manufactured parts to the AM process parameters. For simulating the mechanical properties of additively manufactured microstructures, statistical continuum theory was used. This model only uses chemical phase information in the microstructure and outputs 2D inelastic properties response. In the next step, crystallographic texture information was taken into account as well. In this case, the single crystal behavior of each phase was used to predict the inelastic properties of the entire polycrystalline microstructure. The strength of this model is that the input (microstructure) for these models can be either 3D (e.g. CT scan data to represent the phase, grain, and texture in 3D) or 2D (e.g. optical/electron microscopy images for phase information and X-ray diffraction data for texture information representation in 2D) and the output can also be either 3D or 2D depending on the input.