One of the challenges of additively manufactured parts is that additive manufacturing processes can lead to anisotropic material properties. For this research, we focus on fused deposition modeling (FDM) which is a common consumer grade process that also is often used for early prototyping due to its low investment and processing cost. In FDM, a semimolten filament is extruded. Within the filament (the intrafilament bonds), the material properties are essentially that of a bulk material that has been injection molded. However, because of the differences in temperature when adjacent filaments are deposited, the material properties between filaments within a layer (the intralayer bonds) are generally less than that of the intrafilament bonds leading to an anisotropic behavior within a layer. Similarly, the temperature difference between layers leads to yet another different material bonding strength (interlayer) that is also less than that of the intralayer or intrafilament bonds.

Our hypothesis is that these anisotropic property differences can be predicted using Discrete Element Models (DEM) to model the process of printing the part filament by filament and layer by layer and the subsequent cooling process. A DEM approach discretizes the filament into discrete elements that are treated as a lumped parameter elements connected to adjacent elements through a set of heat transfer boundary conditions. Elements with external part surfaces are therefore connected to the external environment, or in the case of elements in the base layer of the part, are connected to the print bed which is often heated to encourage bonding between filaments.

The DEM model is validated by comparing the predictions of the model against observed behaviors in FDM printing. For instance, the exposed surfaces of an FDM print will cool faster than elements in the core of the print, or elements that are in contact with the heated printing bed.

This paper describes the process of developing a thermal DEM model in MatLAB, including the assumptions underlying the element level heat transfer model. In addition, discussion of the model results is included to demonstrate the validity of the model as well as the comparisons made to available simulation and experimental data which allows us to validate the underlying behavior of the model. As a result of this research, there are several avenues available for future work including the estimation of bond strength between fibers and layers, the incorporation of viscosity effects, mechanical loading, and the possibilities for process optimization based on intelligent filament path planning, reheating technologies and adaptively controlling the build plate and environmental temperatures.

This content is only available via PDF.
You do not currently have access to this content.