Monitoring of fused filament fabrication (FFF) : An infrared imaging and machine learning approach

2025/03/14 by Bauriedel, Niklas, Albuquerque, Rodrigo Q., Utz, Julia +2
#620 #IR imaging #additive manufacturing #fused filament fabrication #machine learning #mechanical properties

paper · doi:10.15495/epub_ubt_00008305

Abstract

Additive manufacturing holds great promise for broader future use, but quality assurance and component monitoring present notable challenges. This study tackles monitoring Fused Filament Fabrication (FFF) via infrared imaging to forecast the mechanical traits of 3D-printed items. It highlights how temperature variations, influenced by the infill's alternating orientation, affect printed parts' mechanical properties. Utilizing Machine Learning, notably the Random Forest Regressor, this research validates the capability to accurately predict tensile strength from infrared temperature readings, offering a simple, yet effective, real-time FFF monitoring method without specialized hardware. This approach enhances the quality and dependability of 3D-printed components with IR thermal monitoring and machine learning predictions. Highlights Infrared imaging and machine learning are combined to monitor 3D printing. A cost-effective and accessible non-destructive monitoring method is proposed. Temperature variation patterns of 3D printed layers influence mechanical properties.

Related

• Machine learning for characterizing uncertain elastic properties of fused filament fabricated materials for topology optimization applications
• Evaluation of machine learning models in predicting mechanical properties of additive-manufactured parts
• Data-Driven Process Optimization of Fused Filament Fabrication based on In Situ Measurements
• Infrared Thermography of Complex 3D Printed Components
• An Evaluation of Classification Methods for 3D Printing Time-Series Data