Project Description
Fused deposition modeling (FDM) type 3D-printers have become increasingly popular due to their ease of use for manufacturing complex customizable parts and assemblies. Despite their popularity however, they continue to suffer from a series of printing faults that often result in a failed print. At the industry level, commercial 3D-prints that fail can be costly and time consuming. The proposed project will focus on developing a real-time monitoring system that will continuously evaluate the 3D-printing process to detect print faults. The project will include: a) instrumenting a 3D-printer with sensors (acoustic or optical), b) data acquisition using a cloud-computing-capable single-board computer, c) data processing using feature extraction and machine learning algorithms, and d) development of executable corrective actions (through command input and interception) to minimize print failure. The participants of this project will be expected to deliver a 3D-printing automation system that can detect print artefacts and self-adjust to prevent total print failure, with added option for user input through a monitoring interface.
Research Goals
- Instrument a monitoring system that includes sensors to detect print artefacts
- Develop machine learning algorithms to analyze acquired data and differentiate various faults
- Develop a mechanism (downstream workflow) that enables print interception to correct detected artefacts
Learning Goals
- Become proficient in digital design and project conceptualization software, including computer-aided design (CAD)
- Become experienced in fused deposition modeling (FDM) type 3D-printing methodologies (software and hardware)
- Become adept at implementing machine learning algorithms from acquired monitoring data, which can be used to generate actionable corrective print functions
Groups Conducting Research
Theoretical and Applied Mechanics Group (TAMG): http://tamg.mem.drexel.edu/