Potential Applications of Synthetic Anomaly Data in PBF-LB/M

The Laser Powder Bed Fusion (PBF-LB/M) process is prone to a wide array of anomalies that can compromise part quality, mechanical performance, and functional reliability. While in-situ and ex-situ anomaly detection methods continue to evolve, the scarcity of labeled anomaly data and the variability of anomaly appearances pose major challenges for deploying robust AI-based quality assurance. In this work, we propose a comprehensive roadmap highlighting how synthetic data will enhance various stages of the PBF-LB/M process chain, beyond aiming for zero-anomaly manufacturing. In the presented roadmap, synthetic anomaly data enables systematic variation and evaluation of anomaly types and severities, providing a foundation for data-driven imperfection tolerance discovery and validation.