AI-Based Acoustic Fault Detection in Fused Filament Fabrication of Polymers

Additive manufacturing using Fused Filament Fabrication (FFF) is increasingly being used for polymer components, yet process faults such as nozzle clogging, over-extrusion, and warping can significantly affect print quality. Traditional monitoring approaches, which rely on visual inspection, typically detect such issues only after the printed part has already developed unacceptable defects. This work investigates sound- and vibration-based process monitoring as an early and low-cost alternative for fault detection. Standard microphones and acceleration sensors capture the process signals near the extrusion zone, which are then analyzed with low latency to enable an early response to process deviations. Changes in these signals occur significantly earlier than those detected through visual monitoring, allowing for faster intervention and thus saving time and material. Machine learning methods are employed to classify characteristic signal patterns associated with each fault type. Preliminary results show that sound and vibration features contain recognizable fault signatures. Therefore, AI-based acoustic monitoring represents an effective and scalable solution for real-time quality assurance in polymer FFF processes.