In-process thermal imaging to detect internal features and defects in fused filament fabrication

The objective of this work is to identify and measure in situ the embedded features in parts manufactured with a fused filament fabrication (FFF) 3D printer. After implementing the monitoring system consisting of optical and thermal cameras, the efficiency of the system is determined in terms of efficacy for automated defect detection through data analysis. In contrast to our previous work, which involved the detection of a large number of randomly embedded sub-surface defects, this study identifies defects of various sizes, geometries, and depths printed in a rectangular strip. Temperature differences, or ΔT, between certain layers are evaluated to determine their significance to the detection of embedded features and internal voids. ΔT between the final layer of a void within the embedded feature and the subsequent layer was found to increase as void size decreased. ΔT between the formation layer and the subsequent layer decreased as void size decreased. Additionally, embedded feature geometries registered higher ΔT between formation layer and the subsequent layer when they consisted of 3-layer voids, which indicates that larger voids, or multilayer defects, within embedded features led to higher formation layer temperatures. Overall, real-time image acquisition, image processing, and data correlation was demonstrated to effectively detect abnormalities in large datasets.