A High-Fidelity Thermal Model for a Novel Droplet-Based Additive Manufacturing Process for Polymers

Andreas Schroeffer,Thomas Maciuga,Konstantin Struebig,Tim C. Lueth
DOI: https://doi.org/10.1115/msec2019-2766
2019-06-10
Abstract:Abstract The claim in additive manufacturing (AM) changes from simply producing prototypes as show objects to the fabrication of final parts and products in small volume batches. Thereby the focus is on freedom of material, dimensional accuracy and mechanical component properties. A novel extrusion-based AM technology has been developed focusing on these issues. The working principle is to form spheres from a thermoplastic polymer melt and build parts by single droplets. The material preprocessing is similar to the injection molding technology and enables a wide range of different thermoplastic polymers as build materials. With the droplet-based working principle high mechanical component properties and dimensional accuracy can be reached compared to similar processes. Further improvements to the process need a detailed knowledge of the physical effects during the build process. The temperature distribution during the manufacturing process determines at which temperature material is fused and how solidification takes place and shrinkage can occur or is suppressed. Thus, it has a significant influence on the mechanical properties and warpage effects of produced parts. In this work a thermal model is presented that describes the heat transfer during the build process. The necessary input data are the material properties and a print job description including the part geometry and building strategy. The basic idea is to simulate each single droplet deposition by applying a dynamic Finite Element Method. All relevant heat transfer effects are analyzed and represented in the model. The model was validated with measurements using a thermal imaging camera. Several measurements were performed during the build process and compared to the simulation results. A high accuracy could be reached with an average model error of about 4° Celsius and a maximal error of 10° Celsius.
What problem does this paper attempt to address?