Session: 03-11-01: Future of Smart Manufacturing

Paper Number: 145934

145934 - Towards Resin and Process Invariant Extrusion Additive Manufacturing Driven by Computer Vision and Machine Learning 

Additive manufacturing (AM) is well-known for its ability to place a wide variety of materials at precise locations in 3D space with very few limitations on geometrical complexity.  Recent advances in soft polymer systems have enabled new materials which can be extrusion 3D printed to form intricate geometries or freestanding structures. However, the lack of real-time process control has limited the precision and reproducibility of the printing process, especially for complex materials systems which have diverse printing or post-processing requirements. To overcome these limitations, this study presents an approach for automated, process invariant AM approach using computer vision-based sensing and machine learning (ML)-driven, real-time control of extrusion 3D printing.  Specifically, this study initially focuses on soft polymeric systems such as frontally polymerizing (FP) thermosets and silicone elastomers and extends this application to extrusion-based metal AM approaches. To identify and characterize key printing parameters, an in-situ computer vision system was implemented which included on-printer cameras and computer vision algorithms.  The system developed herein could perform measurements at a rate of 200Hz.  Based on real-time printed object measurements, the printing parameters can be autonomously adjusted to generate repeatable printing geometries, regardless of the material system and printing envorinment.  This auto-regulative approach for extrusion 3D printing offers a promising avenue for the precise, repeatable, and scalable fabrication of complex 3D structures with both high accuracy and resolution.

Presenting Author: Devin Roach Oregon State University

Presenting Author Biography: Devin J. Roach is an Assistant Professor in the Mechanical, Industrial, and Manufacturing Department at Oregon State University (OSU). Prior to starting at OSU, he was a Senior Member of the Technical Staff at Sandia National Laboratories in the Advanced Materials Lab. His research focuses on developing AI-driven additive manufacturing (AM) techniques for the fabrication of smart, multi-functional composites which can find real-world applications. To do this, he develops multi-material multi-method 3D printing approaches and researches new materials, including liquid crystal elastomers (LCE), for the printing of functional devices such as soft robotics. His specialties include soft mechanics, artificial intelligence, materials, and multi-material structural design.

Authors: