Project Details
Projekt
Data- and Response-Surface-Driven Design Assistant for Controlled Flexible Multibody Systems
Subject Area
Mechanics
Term
since 2022
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 501840485
The project aims to revolutionise the design process for controlled flexible multibody systems by shifting from traditional expertise- driven workflows to data- and simulation-driven paradigms, using machine learning techniques for mechanism synthesis and the development of advanced control methods. A key challenge is the exploration of the multimodal solution space, offering diverse, structurally distinct mechanisms with multiple drives that satisfy the same design criteria, thereby ensuring a comprehensive set of proposals. In order to rank the proposed designs, statistically evaluated criteria are developed that include transient performance measures of practical and mechanical relevance as well as control aspects. This will lead to the retrieval and generation of optimal designs that were previously unattainable. Another integral part of the project are surrogate models, which enable efficient simulations of the flexible components and simultaneous optimisation of design and control parameters. Advanced control techniques will be applied to the underactuated multibody systems, using machine learning for real-time, data-driven control without the need for extensive domain expertise. The developed methods will be combined to form a general design assistant capable of designing complex flexible multibody systems and the corresponding control, which will lead to a paradigm shift in the way special purpose machines are devised.
DFG Programme
Priority Programmes
International Connection
Finland
Cooperation Partner
Professor Dr.-Ing. Henrik Ebel
