Uni-slice: A unified framework for non-planar 3D printing algorithms

Abstract

3D printing has been a rapidly growing industry since the 1980s. It bridges the gap between design's digital and physical aspects since it has much potential for designers. It has found its place in the designers' workflow with its increased capability, accessibility, and affordability. As a method of rapid prototyping, 3D printing has advantages in design iteration thanks to the fast production times, less material use, and the ability to produce complex geometries. 3D printing in the design process allows designers to decide faster and more potently. Therefore, designers need to use 3D printing tools and methods effectively. There are various methods to create a physical object from a digital model by 3D printing, including Fused Deposition Modeling (FDM), Stereolithography (SLA), and Selective Laser Sintering (SLS). FDM 3D printing is one of the most common and accessible methods, especially after cheap desktop 3D printers in 2009. The working logic of 3D printing systems is very typical in converting a digital model to a machine language. Slicing is the step between the 3D model and the digital data used by the 3D printing tool. The printing process is quite similar in all 3D printing methods, fabricating the 3D model in individual planar layers, known as planar slicing. This process is also referred to as 2,5D printing. This generalized 3D printing process causes problems in the 3D printed model regarding surface quality, structural strength, and optimization in time and material usage. These problems affecting surface quality are stair-stepping effect, overhangs on the 3D model, and support material removal. Besides, the anisotropy problem affects the 3D-printed part's structural quality. These issues are more noticeable in the FDM method than in the other 3D printing methods. There are studies that aim to solve the above-stated issues caused by planar slicing in FDM 3D printing. These studies represent different approaches to tackling the problems, such as bespoke algorithms and tools. These studies obtain non-planar 3D printed models, yet by primarily using custom software built explicitly for their context, and sharing the source code is rare in these studies. Besides, the ones explaining the algorithm with flowcharts and diagrams recreate those algorithms in a way that requires programming language skills. In short, custom 3D printing studies have problems with transparency, modifiability, and accessibility. Designers need more control in the production phase and avoid the adverse effects of planar slicing. Therefore this thesis aims to guide designers to use 3D printing tools more effectively by creating non-planar 3D printing algorithms according to their needs while avoiding the shortcomings of previous studies. A framework for slicing algorithms is presented considering accessibility, modifiability, transparency, and interoperability in this thesis. To demonstrate its usage, Grasshopper is selected as a visual scripting plug-in for Rhinoceros3D. The algorithms created in Grasshopper can be shared in popular forums on the web, such as food4rhino (Url-1) and grasshopper3d (Url-2). In this way, the shared files become accessible so that other users can modify the algorithms and develop them further. The algorithms depend on Grasshopper's components; thus, another user can view them without restrictions. Additionally, the presented framework is explained as a data flow, making it compatible with different visual programming environments. It is expected that a designer following the framework's steps will be able to create a custom slicing code or modify an existing CLDFM algorithm that works better than standard slicing in 3D printing their design. The framework follows a series of steps, from a 3D model to a file format specific for 3D printers called a G-Code file. The 3D model is respectively transformed into sliced surfaces, print curves, toolpaths, and G-Code. A standard slicing algorithm is created within the framework as default. Designers can modify each algorithm step separately to create new and unique slicing algorithms. The last stage of the framework is the toolpath visualization code to see the outcome of the G-Code before 3D printing. The thesis demonstrates multiple non-planar slicing algorithms following a series of design cases as simplified versions of the 3D models derived from the examples found in the literature. The cases are categorized according to their 3D printing tool, slicing approach, and their goal. Qualitative and quantitative evaluations are performed on the design cases. Quantitative testing is based on the purposes of the cases, such as improving surface quality, time, and material optimization. Qualitative testing is based on the accessibility, modifiability, and transparency of the algorithms. Five design case models are sliced with planar and non-planar slicing algorithms. Results are compared digitally for all models. However, only 3-axis 3D printers are used for physical testing due to the lack of available 3D printing tools. The surface quality of the printed models is compared based on visual observation and measurements with a compass. Normally, slicing programs display how much time and material is spent on a 3D print. Although it is possible to calculate this using the framework in Grasshopper, it is excluded from these experiments because of the program's limitations. This limitation is driven by the inconsistency of accurate time and material spent measurement between the Grasshopper's estimate and the actual 3D printing process. Therefore, this thesis only considers the actual data obtained from physical models. In addition, structural testing as a design criterion is omitted in the experiments because of the lack of necessary equipment supply. Instead, experiments consider the existing structural testing records in similar studies in the literature. The results show that it is possible to create non-planar slicing algorithms using a single, unified framework. In the framework, the algorithms have modular steps comprising different basic versions, such as planar and curved slicing. All codes are written using standard Grasshopper components without requiring a plug-in or custom script. Hence, they are accessible and transparent. The algorithms are also explained as dataflows in diagrams. However, they are not tested in other visual programming languages. Based on the presented algorithms in the framework, all 3D models in design cases are printed with a 3-axis FDM 3D printer. Print parameters such as layer height, shell counts, infill amount, print speed, and temperatures are defined as equal or similar for comparing planar and non-planar slicing. In the tests, the algorithms performed similarly to the previous studies in increasing surface quality while decreasing the time and material spent. In conclusion, while previous studies on non-planar slicing algorithms successfully achieved their goal, the framework demonstrated in this thesis creates a guide for various non-planar slicing projects in a unified, designer-friendly, open-access way. It has the potential to be further developed by its users to form a more holistic system. In the future, the non-planar slicing algorithms can be tested on other visual programming interfaces, shared as a standalone slicing tool or a plug-in for another program. Various digital fabrication tools, such as cylindrical and spherical 3D printers (Sencan et al., 2021), robotic arms, and large-scale 3D printers, can be controlled using this framework. Additionally, combinations of these methods are possible, such as large-scale 3D printing on curved surfaces or adaptive non-planar 3D printing on spherical 3D printers.