A computer vision based real-time warpage monitoring and detection system in fused deposition modeling

Additive Manufacturing (AM) manufactures a component by adding layers of material. Fused Deposition Modeling (FDM) is one such AM technique where polymer materials like ABS, and PLA are printed. The FDM method involves considerable time for printing the material. Any disruptions in the bed, power, or nozzle movements during printing will result in defective product. One of the reasons for a defective print arises due to a phenomenon called warping. These disruptions and errors pose challenges for early prevention as they are only visible gradually over time and cause warpage. In the present work, an attempt is made to predict the occurrence of warpage in FDM printed parts. In the present work, an attempt is made to avoid the adverse effect of warping by utilizing a computer vision model for real-time warpage detection. A real-time acquisition system is built to capture print layer-by-layer for a convolutional-based feature extraction process. This integration of a computer vision model aligned with continuous monitoring facilitates a timely shutdown process upon the occurrence of warpage.