Exzellenzcluster – Internet of Production

  Nicht verfügbar Copyright: © Dr. Martin Riedel

"The world becomes a Lab" - Cross-domain generation and real-time exchange of information to improve model-based control approaches for production systems

01/01/2019

Key Info

Basic Information

Duration:
01.01.2019 to 31.12.2026
Acronym:
IoP CRD-B2
Group:
Production Systems
Funding:
DFG

Contact

Name

Sebastian Stemmler

Managing Chief Engineer, Head of Industrial Systems and Production Systems

Phone

work
+49 241 80-27479

Email

E-Mail
 

Motivation

Ensuring the sustained competitiveness of manufacturing enterprises in high-wage countries necessitates the cultivation of flexible and efficient production processes while upholding adequat quality standards. In this context, a comprehensive examination and optimization of the production is indispensable, mandating the generation and exchange of knowledge across domains in real-time to enhance existing paradigms in production engineering.

 

Project goals and methods

Sponsored by the German Research Foundation

Aligned with the federal excellence strategy, the excellence cluster "Internet of Production" aspires to explore application-oriented and innovative solutions within the realm of production engineering. As an example, in the field of rough milling, the paramount objective is to optimize geometry and surface quality while simultaneously minimizing production time. Likewise, in the context of plastic injection molding, achieving reproducible adjustment of the specific weight of produced forms, even in the presence of environmental disruptions, holds considerable significance. Achieving these objectives requires the model-based coupling of target and quality variables with machine parameters. Specifically, for the aforementioned examples, the optimization goals in question can be modeled as functions of tool speed and mold internal pressure, respectively. Given the presence of unknown dynamics, non-linearities, and evolving boundary conditions, a perpetual process of optimization and ongoing adaptation of models is required. Thus, the research focuses on harnessing the strengths of data-driven learning techniques, such as neural networks and support vector machines, to enhance existing physical models. Moreover, research aims to explore the real-time operability and compatibility of these adaptation structures with higher-level control processes. By employing these models and pursuing "optimal control," the aim is to operate processes at the pinnacle of productivity without compromising defined quality objectives.

 

Innovations and Perspectives

Interdisciplinary data repositories can be leveraged to enhance models related to critical quality variables in conjunction with physical methodologies. Furthermore, the fusion of physical and data-driven modeling approaches ensures the interpretability of acquired models while expanding knowledge surrounding process operations. Explorations into higher-level control processes in conjunction with these modeling approaches foster both cross-domain interoperability of control methodologies and amplify the flexibility and efficiency within the production process.

 
Project partner