As part of this project, an open infrastructure for machine learning is being developed that makes it possible to automatically create artificial intelligence (AI) models for productive use based on existing process data and to exchange them across company boundaries. The open source platform "Open Space for Machine Learning" (Os4ML) is being developed for this purpose by the project partner WOGRA AG (https://wogra.com/). The aim of this platform is to enable non-experts to solve everyday problems with AI. It automates all steps on the way to the finished AI model with the help of an intuitive user interface.
The platform development is supported and checked by DLR with experience and expertise with regard to its usability in a productive environment. For demonstration purposes, categorised data from the thermoplastic tape laying process is supplied, which is selected, classified and annotated for machine learning. The results of the AI are then scientifically checked for plausibility and statistically analysed.
The use case of thermoplastic tape laying is of particular interest to the aviation industry, as this manufacturing process leads to significant weight savings (thinner aircraft skin due to better toughness of the material) in conjunction with considerably simplified component production (no autoclave or oven due to in-situ consolidation) and final assembly (drill-free assembly, weldable material). There are currently no inspection systems available that automatically detect all prescribed defects and general anomalies.
You can find more information and request a free demo on the project page:
The project is open source and can be found freely available on Github:
Supported by:
Logo
Logo
Credit:
DLR
Contact
Dr. Ing. Alfons Schuster
German Aerospace Center (DLR)
Institute of Structures and Design
Center for Lightweight Production Technology Augsburg
Am Technologiezentrum 4, 86159 Augsburg
Lars Larsen
German Aerospace Center (DLR)
Center for Lightweight Production Technology (ZLP)
