Material extrusion or fused deposition modeling is a popular 3D printing process that currently faces challenges in processing flexible materials such as thermoplastic polyurethane regarding printability and fabrication performance. It has been observed that naturally occurring printing defects, such as stringing and blobs, have a significant impact on the mechanical properties of the prints. To address these challenges, this study presents a system that utilizes a convolutional neural network to control defects in real-time, while simultaneously ensuring the printability of flexible materials through precise parameter control. The proposed system autonomously controls printing parameters such as flow rate and nozzle temperature. Results show that the system can correct stringing and blobs defects within 25 printing layers and reduce printing time by up to 30% while keeping the mechanical strength within an error range of 3.18%. This system has the potential to improve efficiency and reduce waste in advanced 3D printing technologies.
Copyright is held by the author(s).
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Supervisor or Senior Supervisor
Thesis advisor: Soo, Kim, Woo
Member of collection