| About this Abstract |
| Meeting |
MS&T23: Materials Science & Technology
|
| Symposium
|
Additive Manufacturing Modeling, Simulation, and Machine Learning: Microstructure, Mechanics, and Process
|
| Presentation Title |
Analyzing and Predicting Surface Roughness in Laser Powder Bend Fusion |
| Author(s) |
Miguel Correa, Nathan L. Post, Andrew Neils, Jack J. Lesko |
| On-Site Speaker (Planned) |
Nathan L. Post |
| Abstract Scope |
The Laser Powder Bed Fusion (LPBF) technique is becoming widely used due to its ability to print complex metal parts quickly and without extensive tooling. However, due to LPBF’s non-traditional approach of melting together metal particles layer by layer, extra care must be taken in ensuring the initial design matches the resulting part’s intended specifications. Data analytics approaches combined with machine learning (ML) can help in identifying relevant relationships between the printing process and predicting a resulting part’s mechanical and topological properties. The present work evaluates open-source NIST data taken from an LPBF build where the surface roughness was evaluated for different orientation and position within the build. We extend prior efforts by evaluating the performance of several ML methods to predict the surface roughness based on the location and orientation. A successful model will enable optimizing build orientation for complex components with specific topology requirements, reducing post-build machining. |