| dc.description.abstract | Additive manufacturing (AM), often referred to as three-dimensional (3D) printing, has become one the fastest growing manufacturing technologies across a variety of engineering fields after transitioning from being a prototyping tool to a manufacturing tool for structural end parts. The subsequent introduction of four-dimensional (4D) printing where smart materials are used that can change their shape over time when exposed to an external environmental stimulus has enabled the design of programmable and dynamic structures.
The focus in this thesis is specifically on fused filament fabrication (FFF) which is the most widely available and cheapest AM technology. The quality of components fabricated with FFF is less consistent compared to those manufactured with traditional polymer processing technologies. The layer-by-layer extrusion process notoriously causes residual stresses, reduced bond strength between deposited filaments and reduced dimensional accuracy. The result is a material with a characteristic mesostructure displaying anisotropy, and with mechanical properties differing from those of the raw printing material. Thus, two crucial but separate topics are approached in this thesis: understanding and simulation of the final printed material and the printing process itself.
A multiscale simulation approach is introduced which has the ability to model the complex structural and nonlinear constitutive behaviour exhibited by FFF printed structures. The mesostructure consisting of periodic representative unit cells are numerically homogenized to compute their effective properties. The nonlinear orthotropic behaviour on the macroscale is represented by a B-spline response surface model generated by the mesoscopic data. The macroscopic geometry is modeled by a nonlinear Kirchhoff-Love shell formulation and discretized using isogeometric finite elements, meaning that the concept of isogeometric analysis is applied both for the discretization of the meso- and macroscopic problems. Moreover, various design parameters influencing the local stiffness values and the orientation of the highly orthotropic macroscale material model, are also represented by the isogeometric concept. The framework is applied to 3D functionally graded knitted structures which display similar characteristics as FFF printed components.
Secondly, a simulation framework is introduced which efficiently simulates the heat transfer that takes place during the FFF printing process. The high computational effort required for these simulations is tackled by proposing a combination of efficiency measures whilst mitigating the accuracy loss. A moving heat source model is used as a basis to model the material deposition during the printing process. Modeling simplifications are proposed for efficient discretization of air-filled infill structures and appropriate convective thermal boundary conditions are derived for printed specimens subjected to natural convection. Efficient material deposition strategies are explored and an error-based adaptive coarsening approach is implemented. An appropriate coarsening technique is also presented for geometries with air-filled infill patterns.
Both simulation approaches are numerically and experimentally validated, showing their relevance and potential for real-world applications. | en_US |
