About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing: Advanced Characterization With Synchrotron, Neutron, and In Situ Laboratory-scale Techniques IV
|
| Presentation Title |
AI-Driven Spectral Sensing for Distributed Born-Qualified Additive Manufacturing |
| Author(s) |
Steven Storck, Vince Pagan, victor leon, mary daffron, ari lax, Sam Gonzalez, brad bazow, jackson pittman, graham spicer |
| On-Site Speaker (Planned) |
Steven Storck |
| Abstract Scope |
Defect formation in additive manufacturing (AM) creates significant uncertainty, complicating qualification and increasing reliance on post-process inspection. While in-situ monitoring can help to understand the chaotic synthesis and tailored rapid solidification, sensing with AI is needed to maximize the potential. To this end, we developed a custom high-speed spectral sensor integrated with a field-programmable gate array (FPGA), capable of detecting and responding with micron precision. By combining rapid sensing with AI, our system achieves over 99% accuracy for lack-of-fusion defects and 94% for keyhole formation. We will show a method to classify defects within 5 microseconds using an AI assisted technique. We will present a method for visually representing the defects in real-time and the ability to manipulate energy along the laser path with micron accuracy based on sensor inputs. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Solidification, High-Temperature Materials |