| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Modeling, Simulation and Machine Learning
|
| Presentation Title |
F-15: An Accurate Machine Learning Approach for Process Optimization in Directed Energy Deposition |
| Author(s) |
Xiao Shang, Ajay Talbot, Hui Lee, Yu Zou |
| On-Site Speaker (Planned) |
Xiao Shang |
| Abstract Scope |
In additive manufacturing, process optimization is a long-standing yet crucial challenge. Traditional design of experiments approaches are costly, have low accuracy, and lack the versatility to address multiple optimization objectives by different production needs. Although efforts have been made using machine learning (ML) methods, most work to date oversimplifies the process and only focus on single-track prints, which hardly translate to multi-track and bulk cases demanded by real-life applications. To address these challenges, we propose a comprehensive ML-aided process optimization approach for the directed energy deposition (DED) process. This approach accurately predicts single-track, multi-tracks and bulk prints geometry from processing parameters. Furthermore, processing maps targeting various objectives including quality, precision, and speed have been established, and the result is validated experimentally. The approach developed here provides an efficient and versatile way for process optimization to meet various needs in industrial productions and is key to developing closed-loop controlled DED systems. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Machine Learning, Other |