Nonplanar Layered Morphologies

Abstract

This thesis investigates the potential of nonplanar robotic 3D printing for architectural applications and develops novel methods to design nonplanar print paths for medium to large-scale robotic FDM 3D printing. Latest developments in additive manufacturing have opened new possibilities for 3D printing objects with unprecedented geometric complexity. However, these advancements are still hindered by the inherent limitations of current printing techniques. Most 3D printing is 2.5D printing, where the accumulation of material happens in planar layers. This results in various drawbacks, such as poor surface quality on high curvature areas and a need for overhang support. In addition, 3D printing produces parts that are inherently anisotropic, with considerably higher strength along the print direction than orthogonally to it. Further research in controlling the orientation of print paths can help alleviate those issues by aligning the path directions as needed to improve surface quality, reduce the need for support, and optimize mechanical properties. With the advent of robotic printing, it is possible to print material along 3D paths with variable orientation and thickness. Consequently, paths can be customized to address the aforementioned issues. However, there is a lack of methodologies for designing robot-fabricable, nonplanar print paths. This research aims to address this gap with a focus on creating intuitive path design tools within the reach of designers and on considering fabrication constraints inherent to nonplanar robotic 3D printing. The proposed computational approach is based on defining a parametrization function over a surface and then tracing the paths as its isolines. The primary challenge then is selecting an appropriate function that fulfills the intended objectives. With that, this research draws upon the wealth of established methods for shape parametrization in computer graphics and repurposes them in the novel context of robotic 3D printing. Two methodologies are developed, namely designing a boundary-controlled and a vector-field-controlled function. The first defines the function as the interpolation of the geodesic distance from boundaries set by the user, resulting in paths always parallel to the boundaries and smoothly interpolating the space between. The second defines a function considering a tangent guiding vector field as its gradient. The user can then control the paths by applying constraints on the guiding vector field, thus having control over both orientation (always orthogonal to the vector field) and spacing (determined by the magnitude of the vector field). To further enable control of the path layouts, an intermediate discrete representation is devised, consisting of two coupled transversal strip networks overlayed into a strip-decomposable quad (SDQ) mesh. Editing operations for altering the strips' connectivity are developed to allow hands-on manipulation of the SDQ mesh structure to suit specific design requirements or aesthetic preferences. The 3D printing method selected as a case study is the robotic FDM extrusion of thermoplastics. In the early research stages, prototypes are carried out using filament extrusion, and later, the paths are adapted for pellet extrusion, which also enables the use of recycled plastics. A series of physical prototypes are produced with both printing setups demonstrating the potential and challenges of the proposed approach. The research's impact lies in design innovation and efficiency. Nonplanar paths unlock a novel design realm, enabling designers to not only shape the exterior but also manipulate the layered configuration of the printed object, fostering artistic and structural innovations and broadening the creative domain of 3D printing. From an efficiency perspective, it minimizes material waste by reducing the need for sacrificial support and enhancing surface quality.