Title: TWIN Project — Transformation of complex product development processes into knowledge-based services for generative manufacturing

Project duration: October 2019 – April 2023

Research Area: Manufacturing

The focus of the TWIN project is the development process for the production of prototypes using laser powder cladding (LPA) from the field of generative manufacturing (GM). The goal is to increase automation, quality, reliability, and data availability in generative manufacturing, using the LPA process as an example, through automated data processing and machine learning methods. This holistic approach lays the foundation for new intelligent services.

Aims

• Reduce number of erroneous printed parts
• Monitor printing process
• Predictive maintenance of LPA machines

Practical example created during the project

• Working Prototype for analysing a 3D print in laser powder bed fusion (LPBF)
• Predictive maintenance model for LPA
• End-to-End processing pipelines of images and time series data

Technology

• Kubernetes, Docker, Kafka, MQTT
• Python, PyTorch, Typescript, Angular

Publications

• Bauer, M.; Uhrich, B.; Schäfer, M.; Theile, O.; Augenstein, C. and Rahm, E. (2023). Multi-Modal Artificial Intelligence in Additive Manufacturing: Combining Thermal and Camera Images for 3D-Print Quality Monitoring. In Proceedings of the 25th International Conference on Enterprise Information Systems – Volume 1: ICEIS; ISBN 978-989-758-648-4; ISSN 2184-4992, SciTePress, pages 539-546. DOI: 10.5220/0011967500003467
• Uhrich, B., Schäfer, M., Theile, O., Rahm, E. (2023). Using Physics-Informed Machine Learning to Optimize 3D Printing Processes. In: Correia Vasco, J.O., et al. Progress in Digital and Physical Manufacturing. ProDPM 2021. Springer Tracts in Additive Manufacturing. Springer, Cham. https://doi.org/10.1007/978-3-031-33890-8_18
• Markus Bauer, Christoph Augenstein, Martin Schäfer, Oliver Theile, Artificial Intelligence in Laser Powder Bed Fusion Procedures – Neural Networks for Live-Detection and Forecasting of Printing Failures., Procedia CIRP, Volume 107, 2022, Pages 1367-1372, ISSN 2212-8271, https://doi.org/10.1016/j.procir.2022.05.159.

Team

Project Lead

Team Member

Partner

• InfAI e.V.
• Fraunhofer IWS
• LASERVORM GmbH
• Siemens AG
• CPT Präzisionstechnik GmbH
• quapona technologies GmbH
• Leipzig University