The prediction quality and the subsequent transferability of the models to be developed (transfer learning) depend significantly on the selection and placement of individual sensors in forming tools.
Possible areas of work to solve this problem are
- Analysis and requirements for measurement and control functions of interlinked or multi-stage forming sequences with regard to measurement context, time scales, resolution of signals, choice of measurement locations, etc.
- Definition of suitable metrological concepts and derivation of methods for suitable sensor selection and integration in forming tools, taking into account the achievable data quality (sampling rate, resolution rate, repeatability, etc.), the robustness of the measurement (e.g. protection class)
- Transferability concepts of suitable measurement principles to other / different forming processes of the SPP / near-process and remote sensor data preparation
- Definition and specification of data formats / ontologies applicable across SPP
Contact
Univ.-Prof. Dr.-Ing. Dr. h. c. Michael Weyrich
University of Stuttgart
Institute for Automation Technology and Software Systems
Pfaffenwaldring 47
70550 Stuttgart
E-mail: michael.weyrich@ias.uni-stuttgart.de
Website: https://www.ias.uni-stuttgart.de/
Prof. Dr.-Ing. Birgit Vogel-Heuser
Technical University of Munich
Institute for Automation and Information Systems
Boltzmannstr.15
85748 Garching near Munich
E-mail: vogel-heuser@tum.de