Mechanical Engineering
Graduate Seminar
NOTICE OF SEMINAR PRESENTATION
CANDIDATE: Dora Strelkova
DEGREE SOUGHT: PhD
DATE: 9/19/2025
TIME: 11:30am
PLACE: Room 1101 CEI
TITLE: Putting the 3D in 3D Printing: Addressing Manufacturing Variability of Fused Filament Fabrication (FFF) Parts via Non-Planar Toolpaths
Abstract
Material Extrusion (ME), one of the seven ISO/ASTM 52900 Additive Manufacturing (AM) families, covers processes that deposit material extruded through an opening while following a toolpath. Its common subfamily, Fused Filament Fabrication (FFF), builds parts by extruding molten thermoplastic through a nozzle. FFF toolpaths are generated by slicer software, which typically employ planar slicing to convert CAD models into printable layers. Slicers apply user-selected parameters such as layer height, bead width, wall thickness, print orientation, infill pattern and density, raster angle, etc. to precisely define how a part is printed. These process parameters drive the anisotropic behavior of FFF prints, as part performance depends not only on filament type but also on the selected parameters. Notably, raster angle influences inner-layer void formation and build orientation strongly governs strength relative to the applied load direction. Current research leverages non-planar slicing strategies, such as surface-contouring or offset layers, potentially improving surface finish, enhancing part strength, and reducing support structures. However, non-planar toolpaths can still suffer from directional mechanical weakness and inconsistent layer bonding. To address these challenges, this work will investigate woven layer toolpaths as an approach to achieve more isotropic part performance, enhanced load distribution, and improved layer bonding. Current work focuses on analyzing voids generated by planar slicers and evaluating how their morphology influences tensile strength, with the goal of informing a future predictive model of mechanical performance and identifying problem areas that woven toolpath strategies can mitigate. To support these efforts, a universal void-analyzer was developed using K-means clustering, image filtering, and unsupervised learning techniques. The analyzer extracts key metrics such as void volume, shape, and density, which are then directly related to experimental tensile data from Nylon 6-CF and PLA samples across varying build orientations and raster angles.