Anomaly-driven reinforcement learning for process optimisation in additive manufacturing of hybrid materials (A07)

• Subject Area: Data Management, Data-Intensive Systems, Computer Science Methods in Business Informatics ; Joining and Separation Technology
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 511263698

This project leverages deep learning (DL) for dynamic process parameter adaptation in L-DED when manufacturing HyPo-components. It involves investigating process anomalies and their impact on material properties and identifying optimal process parameter ranges. This will help to establish a comprehensive database for training DL models. Finally, the project targets seminal advances in the field of anomaly detection and reinforcement learning by developing anomaly-driven reinforcement learning for process parameter adaptation.

• DFG Programme: CRC/Transregios
• Subproject of: TRR 375: Multi-functional high performance components made from hybrid porous materials
• Applicant Institution: Rheinland-Pfälzische Technische Universität Kaiserslautern-Landau
• Project Heads: Professor Dr.-Ing. Jan C. Aurich; Professorin Dr. Sophie Fellenz; Professor Dr. Marius Kloft