A hybrid physics-based and data-driven approach for characterizing porosity variation and filament bonding in extrusion-based additive manufacturing

Abstract The application of additive manufacturing technology has recently shifted from the fabrication of prototypes to functional end-use products. Consequently, the mechanical strength of products has become of significant importance. To quantify the global and local mechanical strength, it is necessary to characterize the micro-structures and their variation within the product. The extent of bonding between adjacent filaments, both within and between layers, as well as porosity are two of the most important parameters that directly contribute to the mechanical strength of parts in extrusion-based additive manufacturing. However, most of the existing models in the literature either significantly underestimate these parameters or fail to quantify or address their variation along the deposition path and build direction. Hence, in this paper, a hybrid physics-based and data-driven approach is proposed to quantify the extent of filament bonding, porosity, and their distribution within a geometry of interest by characterizing the temperature profile of filaments and their deformation. The proposed models for inter-layer and intra-layer bonding have an average accuracy of 95% and 94%, respectively. In addition, it is observed that the porosity variation model performs better for top layers compared to bottom layers with an average of 51% higher accuracy.