Research Group MiP
Research project

KI-ProLaser | KI-ProLaser: AI-based self-configuration and process database generation for laser welding

Within the project KI-ProLaser, a real-time capable and self-configuring laser system for series production including the required knowledge database is being developed. The basis for this is the fusion of different data sources and their analysis with suitable AI methods.

Coordinator: Amina Ziegenbein M. Sc., M. Sc. Enno Lang M. Sc.
Duration: 01.03.2021 – 29.02.2024
Funded by: Federal Ministry of Education and Research (BMBF)
Webseite:www.photonikforschung.de

Motivation

The demands on the manufacturing industry are subject to high availability and the constant increase in effectiveness and quality of production plants. This is accompanied by the minimization of faulty part production and the reduction of drive-in processes. Industry 4.0 ideas offer promising approaches for meeting these requirements. The basis for this is the linking of advanced production technologies with information and communication technologies (ICT), which makes it possible to disclose previously unknown relationships. Modern ICT and innovative analysis algorithms must be used to capture the information generated by highly dynamic processes while at the same time evaluating the sensor and measurement data. Here, artificial intelligence (AI) algorithms offer a way to make the resulting data volumes manageable and transparent and to use them in production. This outlined picture of Industry 4.0 has up to now remained largely untouched in the environment of processing lasers in connection with imaging processes and AI algorithms. This is where KI-ProLaser comes in, using the example of laser beam welding, with intelligent data analysis and the self-configuration of the laser system based on this.

Objectives & Approaches

The overall objective of KI-ProLaser is to use artificial intelligence methods, based on sensor and measurement data from the laser beam welding application, to automatically configure the overall laser system. This results in productivity improvements through increased process transparency, an increase in process quality and the possibility of integrated quality assurance. The innovative aspect is the use of new data analysis methods on several levels. On the one hand, optical coherence tomography (OCT) is used to pre-process image data and to evaluate it for the application. On the other hand, methods of AI as an integrated self-learning system are elaborated, which allow a self-configuration of the photonic system and the creation of a knowledge database for laser beam welding. Benefit-oriented data selection and real-time capability of the overall system as well as the methodical and transferable approach are key points of scientific and economic innovation.

The results will be utilized in different ways within the project consortium. These include the distribution of self-configuring laser systems together with new future-proof communication technologies and knowledge data. In application, significant productivity and quality increases are expected through the use of the developed solutions in series production. Furthermore, the development of image processing solutions in integrated inspection systems forms the basis for a market launch by the project partners. The ICT infrastructure to be developed for AI methods is analogous to this; it provides a comprehensive connection of a technical system in production in a platform solution, which is brought together and visualized in a user-oriented manner. In the future, this development can be extended to other areas of mechanical engineering and may lead to new.

Acknowledgement

This project is funded by the Federal Ministry of Education and Research. We are thankful for the opportunity to work on this project.

Consortium partners

The project is carried out in cooperation with Laserline GmbH, AUDI AG, Automation W+R GmbH and die.interaktiven GmbH & Co. KG.

Funding source

Consortium partner