DEVELOPMENT OF REAL-TIME DEFECT DETECTION TECHNIQUES USING INFRARED THERMOGRAPHY IN THE FUSED FILAMENT FABRICATION PROCESS

One of the most popular forms of Additive Manufacturing (AM) is Fused Filament Fabrication (FFF), which allows flexibility to work with multiple materials with comparatively low manufacturing costs. However, uncontrollable process factors lead to the underutilization of the full potential of FFF on an industrial scale due to occasional failure in 3D printed parts. These failures can scale from insignificant to critical defects, which can significantly compromise the overall quality of the product, including desired structural integrity, mechanical property, and reliability of the FDM-produced parts. Post-inspection of the printed part is inconvenient as they result in a loss of labor hours, print time, and print materials. To circumvent these issues, an in-situ real-time defect detection approach is proposed in this study utilizing Infrared (IR) thermography. While optical sensors can provide real-time images of in-printing parts, they are unable to measure the thermal deviations essential to evaluate FFF process. Since IR sensors can provide thermal information of in-printing FFF parts with acceptable resolution in the spatiotemporal domain, these sensors can be utilized as an alternative source of information applicable for monitoring the print states and print defects of various types. In this study, an IR sensor-based data acquisition system is designed and integrated into an existing FFF printer to collect in-printing layer-wise real-time thermal data. As the FDM printer starts printing each layer of a part, the thermal data stream produced by the IR sensors from the in-printing part is collected along with the nozzle and the print bed temperature. By establishing a data-driven correlation of the spatiotemporal information between the non-defective and defective parts, a field-deployable approach to detect defects in real time is demonstrated herein.