Machine Learning Database for 3D Printing Optimization
Sara McMains, Professor
Mechanical Engineering, Computer Science
Closed. This professor is continuing with Fall 2023 apprentices on this project; no new apprentices needed for Spring 2024.
Additive Manufacturing (aka 3D Printing) is a set of relatively novel manufacturing techniques that were originally used for prototyping but are increasingly used to fabricate end-use parts, which requires higher quality manufacturing.
The goal of this project is to improve the quality of 3D printed parts using machine learning, as well as, create a benchmark dataset for future machine learning research.
We are working to create practical and industrially relevant tools that provide design and manufacturing feedback for 3D printing. The research apprentice will be working on a team of 3-5 to predict optimal build orientation or part distortion. Unlike existing research efforts in machine learning for this application that create synthetic data that are often not representative of 3D printed parts, we are compiling a database of 3D printed parts using CAD repositories (e.g. FabWave, Thingiverse). The parts we compile are rotated in a series of orientations, labeled by printing orientation quality, and we are working on a machine learning pipeline to predict optimal orientation.
We are working to create practical and industrially relevant tools that provide design and manufacturing feedback for 3D printing. The research apprentice will be working on a team of five to predict optimal build orientation since printing orientation drastically impacts printed part quality. Unlike existing research efforts in machine learning for this application that create synthetic data that are often not representative of 3D printed parts, we are compiling a database of 3D printed parts using CAD repositories (e.g. FabWave, Thingiverse). The parts we compile are rotated in a series of orientations, labeled by printing orientation quality, and we are working on a machine learning pipeline to predict optimal orientation.
Role: The researcher who joins our project will be working with a partially created database to either (a) make the database fully functional, (b) run FEM analysis to be added to the database, (c) research literature search and analysis, and/or (d) build machine learning models to predict optimal build orientation or part distortion.
Qualifications: Desirable skills: Familiarity/experience with databases in JavaScript or finite element method (FEM) Analysis or machine learning models
Day-to-day supervisor for this project: Sara Shonwiler, Graduate Student
Hours: 6-8 hrs
Engineering, Design & Technologies Mathematical and Physical Sciences