Infrared thermography–based framework for in situ classification of underextrusions in material extrusion

Material extrusion (ME) is a widely used additive manufacturing (AM) technique, known for its versatility, cost-effectiveness, and ability to produce complex parts on-demand with greater customization and reduced waste. However, the process is impeded by unpredictable factors causing defects such as voids, overextrusions, and underextrusions, which smart manufacturing in Industry 4.0 aims to mitigate. In this study, we report a novel infrared (IR) thermography–based continuous data acquisition and processing framework that can differentiate various levels of in situ underextrusions. While existing underextrusion detection techniques require mid-print interruptions, our framework detects defects without any interruption. The methodology includes integrating an IR camera into a commercially available extrusion-based 3D printer for continuous in-printing data acquisition. The G-code for printing a rectangular block is intentionally modified to induce various levels of known underextrusions. Additionally, a novel signal processing algorithm is developed to automate real-time data processing and analysis, including signal normalization, artifact removal, and feature extraction. Results are obtained by developing a correlation matrix to compare the correlation coefficients of time series thermal data from the printed samples. Time-domain thermal features are also extracted to identify extrusion levels of 25%, 50%, 75%, and 100%. This study demonstrates that by utilizing the proposed framework, thermal data can identify various extrusion levels without mid-print interruption and determine the severity of process deviations within 5 s. This framework paves the way for integrating a thermal data-driven closed-loop monitoring and adjustment system capable of producing first-time-ready parts.