Assessment of fundamental layer features in monolayer parts fabricated via FFF additive manufacturing under varying printing conditions

Abstract The advancement of Additive Manufacturing (AM) technologies, particularly Fused Filament Fabrication (FFF), has significantly broadened the possibilities for precision manufacturing. However, FFF remains susceptible to geometric distortions that can compromise part quality. This study investigates fundamental layer features in monolayer FFF parts produced under three printing conditions: a standard scenario (H) and two defect induced cases (D1 and D2) introduced via G-code modifications. Using optical microscopy and geometric analysis, we measured region specific features, including contour and raster widths, gap distances, and their variability. Under condition H, contour widths in the right region showed high precision, with mean values of 0.6343 mm (CW1), 0.5886 mm (CW2), and 0.6043 mm (CW3), and standard deviations below 0.07 mm. In contrast, defect condition D1 exhibited reduced uniformity, with CW1 and CW2 decreasing to 0.4647 mm and 0.4402 mm, respectively. In the upper region, D1 led to a pronounced degradation: CW1b decreased to 0.3795 mm and gap dimensions became more variable (e.g., G1b = 0.1884 mm, standard deviation = 0.0346 mm). Under condition D2, the central region exhibited the most severe distortion, with the “d” feature mean reaching 20.17 mm and a standard deviation of 2.41 mm. In addition, monolayer features such as RWA and RWP increased markedly under D2 compared to H, reaching 0.6587 mm and 0.5680 mm, respectively. These findings show that extrusion irregularities can critically affect geometric fidelity. Our results quantitatively link process anomalies to feature deviations in FFF printing, reinforcing the need for improved monitoring and correction mechanisms. Beyond quantifying feature changes under induced defects, this study aims to provide actionable first layer acceptance windows to support in situ monitoring. In practice, the signatures reported here (e.g., raster after width ≥ 0.60 mm and inter raster gap ≈ 0.35 mm under D2; contour widths ≤ 0.50 mm under D1) enable simple camera based screening or ML labeling to (i) flag the defect class in real time and (ii) trigger corrective actions, such as pausing the build or adjusting flow and retraction settings. Future work will explore real time defect detection and adaptive G-code compensation to improve dimensional accuracy and repeatability.