About this Abstract |
| Meeting |
2025 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2025)
|
| Symposium
|
2025 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2025)
|
| Presentation Title |
Hybrid Sensor in the Loop Approach for Generating Synthetic Event Imager Data of Melt Pool Dynamics |
| Author(s) |
Anthony Starlef, Aidan Gribble, Nicholas Bruns, David Mascareņas, Mahtab Heydari, Bruce Tai |
| On-Site Speaker (Planned) |
Mahtab Heydari |
| Abstract Scope |
High dynamic range neuromorphic event-based imagers show promise for high-speed, memory-efficient in process monitoring of additive manufacturing, as they can observe the high light intensity environment of the melt pools while reducing memory requirements by only detecting change in log light intensity of each pixel. Advancements in machine learning suggest that data augmentation with synthetic data is often desirable in complex dynamic systems with expensive data collection. However, generating synthetic event data presents challenges in accurately modeling noise, high dynamic range phenomena/Fourier optics effects, and multi-time scale dynamics. To make progress on simulating high-fidelity event data, we propose a hybrid sensor in the loop approach that leverages emerging commercially available high speed monitoring technology. In this study, we examine the suitability of using hybrid sensors in the loops approaches to generate synthetic event data representative of multi-time scale dynamics of melt in order to infer 3D melt pool geometry. |
| Proceedings Inclusion? |
Planned: Post-meeting proceedings |