About this Abstract |
| Meeting |
2026 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2026)
|
| Symposium
|
2026 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2026)
|
| Presentation Title |
In-Situ Qualification of Laser Power and Scan Speed via Melt Pool Emissions in Laser Powder Bed Fusion |
| Author(s) |
Brian Johnstone, Nicole Van Handel, Patrick Merighe, Maegan Lenertz, Christopher Saldana, Kyle Saleeby |
| On-Site Speaker (Planned) |
Brian Johnstone |
| Abstract Scope |
Determining optimal process parameters is critical for machine qualification in additive manufacturing. This process requires fabricating numerous parts with varying parameter combinations and ex-situ characterization methods, such as X-ray computed tomography and high-resolution microscopy, which can be time-consuming and costly. However, in-situ sensing technologies have become increasingly prevalent for part qualification and process monitoring. In this work, two qualification models—a convolutional neural network and a k-nearest neighbors model—were developed and compared for classifying melt conditions using photodiode signal metrics acquired during laser track formation in a metal laser powder bed fusion process. The models were trained using classifications derived from ex-situ micrographs and achieved accuracies exceeding 80%. Sensitivity analysis was conducted to identify the signal inputs with the greatest influence on model performance. The proposed models demonstrate the potential to streamline the qualification and requalification of additive manufacturing systems by reducing both cost and evaluation time while improving process efficiency. |
| Proceedings Inclusion? |
Undecided |