Overview

The research network aims to develop and implement self-adaptive control systems for machines, systems and process chains that can learn from past process steps in order to better adapt to future process requirements. This not only achieves higher component quality (achieving tighter tolerances), but also increased robustness with higher resource efficiency of the process. In terms of robustness, the control concepts should allow the use of recycled materials in particular, which often exhibit greater fluctuations in material properties.  

Where permanently programmed control systems without adaptation to the processes are often unable to deliver reproducible results within the required framework, robust and self-adaptive controlled processes should be created that actively compensate for material fluctuations on the input side in order to ensure constant quality on the output side. With regard to resource efficiency, concepts are to be researched that fully utilize the degrees of freedom in process control in order to process with the lowest possible resource consumption and emissions. For example, adaptive process control anticipates future energy requirements, uses renewable energies depending on their availability and adapts the energy mix and its timing accordingly. 

Both aspects - robustness and resource efficiency - will be used to research and implement approaches for the economic and ecological optimization of production in terms of sustainable process chains using a specific reference process chain. 
For the self-adaptive control systems, basic concepts for process-specific sensor and condition monitoring as well as data-based modelling, control and optimization of manufacturing processes are being researched in the project. Current artificial intelligence (AI) methods are to be used, adapted and further developed in order to make these concepts self-adaptive and adaptive. The investigated processes must also be adapted and networked in such a way that control is possible based on real-time process information. 

In order to ensure the usability of the research beyond the network, the procedure for creating such intelligent manufacturing processes will be analyzed using dedicated individual processes. The concrete process engineering implementation is carried out using the processes of nonwoven production, nonwoven forming, injection molding, an infusion process and a hybrid joining process using ultrasonic welding as well as their interlinking with cross-process control.The processes form a logical process chain that can also be found in a real production line.The complexity of the processes and thus the requirements for sensor technology, modeling and control are to be progressively developed over the course of the project.The starting point will initially be the individual processes involved, which will be enabled to behave in a self-adaptive manner in order to equip them with generic interfaces and ultimately network them across processes and finally develop a general process model. 

To achieve this, it is necessary to bring together interdisciplinary experts from process engineering and materials science with experts from AI and control engineering. Furthermore, cooperation between researchers at research institutions and users in companies is expedient.At the Augsburg site, the infrastructure of the newly created AI production network is to be combined with the innovative strength of a new technical university in Nuremberg.

Funded by

Bayerische Forschungsstiftung

Partners

• Institute for Software & Systems Engineering, University of Augsburg
• Chair of Control Engineering, University of Augsburg
• Chair of Hybrid Composite Materials, University of Augsburg
• Department of Engineering, University of Technology Nuremberg
• Chair for Mechanical Engineering, University of Augsburg
• Chair for Polymer Composites Technology, University of Augsburg
• Fraunhofer Institute for Casting, Composite and Processing Technology IGCV
• Deutsches Zentrum für Luft- und Raumfahrt ZLP Augsburg
• KraussMaffei Technologies GmbH 
• Vallen Systeme GmbH
• Soffico GmbH
• BCMtec GmbH
• SGL Technologies GmbH
• Bolle & Cords Elektrotechnik GmbH
• MAGMA Gießereitechnologie GmbH