Multi-scale Modelling for Materials Design in Additive Manufacturing

Wang, Weiling and Wen, Wei and Rivera Diaz Del Castillo, Pedro and Hou, Xiaonan (2023) Multi-scale Modelling for Materials Design in Additive Manufacturing. PhD thesis, Lancaster University.

Text (2023wangphd)
2023wangphd.pdf - Published Version
Download (9MB)

Abstract

Additively manufactured (AMed) austenitic stainless steels (SSs) possess exceptional properties like high strength and toughness. However, it is unclear how they perform under long-term exposure to high-temperature conditions, such as those found in nuclear reactors. These properties arise due to complex microstructures that develop during additive manufacturing (AM), including nanoscale dislocation cellular structures, microscale sub-grains with a high density of low-angle grain boundaries (LAGBs), and high dislocation density. Although the quasi-static mechanical properties of AM austenitic SSs, such as 316L SS, have been systematically investigated, the creep behaviour of such alloys is still a new area of research, with some experimental studies conducted in recent years. Additionally, the mechanical properties of most AMed alloys are anisotropic due to texture formation, and creep behaviour can be significantly influenced by microstructural differences in various building directions. Furthermore, the presence of AM-characterised microstructure is a notable feature of AM materials, and the size and shape of the pores can greatly influence stress concentration during loading. Thus, it is critical to quantify the effects of AM-characterised microstructure on the mechanical properties of materials. As experimental methods have limitations for studying material properties, it is necessary to use computational modelling to extrapolate existing experimental data, especially for highly time-consuming experiments such as creep testing. This study aims to provide a modelling framework for characterizing the evolution of microstructure and high-temperature creep behaviour in AM and wrought austenitic SSs, considering the impact of the initial microstructure. For AM materials, there are two types of samples: horizontally-built samples (loading direction parallel to the building direction) and vertically-built samples (loading direction vertical to the building direction). The choice of AM materials with different built directions is for studying the effect of the relative loading direction to the building direction on material creep behaviour. The materials strengthening mechanisms, including lattice friction, solid solution strengthening, dislocation hardening, and precipitation hardening, are quantified in detail. In addition to data from literature and experiments used to evaluate each strengthening mechanism, the precipitation evolution during the creep process is simulated through the thermokinetics calculation using Thermo-Calc software. Differently fabricated materials are originally simulated under the visco-plasticity self-consistent (VPSC) framework, using the materials' own characteristics as input. The creep mechanical responses of AM and wrought materials are compared, and the dominant deformation mechanisms are revealed and quantitatively compared. Due to the limitations of the VPSC, only the primary stage and secondary stage of creep behaviour are captured. Based on this, the same physics-based model is employed under the crystal plasticity finite element method (CP-FEM) framework, which is full-field, and combined with the Gurson-Tvergaard-Needleman (GTN) damage model to capture the tertiary stage creep deformation. The original crystal plasticity model is highly microstructure-sensitive, and the detailed local structure can be analyzed through the finite element method. Therefore, the original electron backscatter diffraction (EBSD) information is pictured by MATLAB and used for materials input under the CP-FEM framework. In addition, DREAM3D software is used to extract microstructure information from raw EBSD data. The tertiary creep stages of horizontally-built and vertically-built AMed samples are simulated and compared, revealing that damage tends to accumulate on grain boundaries that are perpendicular to the loading direction. Additionally, the effects of AM-induced pores on creep deformation are evaluated by introducing them into the CP-FEM input. As selecting a specific region on the original EBSD data cannot summarize the overall AM materials characteristics, an artificial input is randomly generated through a Voronoi diagram by MATLAB with assigned grain orientation. The artificial input is characterized by AM-induced elongated grain structure to study the effects of high-angle grain boundaries (HAGBs) on materials creep behaviour, especially the damage evolution.