1315 Computational Mechanics in Complex Product Development

Markus Zimmerman, Technical University of Munich
Fabian Duddeck, Technical University of Munich
Alexander PoppUniversity of the Bundeswehr Munich
Computational Mechanics is a well-established field in academia and industry. Over the last decades, many research areas have advanced significantly, such as modeling of non-linear material behavior and dynamics, optimization techniques, high-performance computing, etc. At the same time, the complexity of developing products in industrial practice has grown substantially due to diversification of products, ever more numerous and detailed requirements, and development processes that are distributed over many parties involved. The academic field of Product Development and its disciplines such as Systems Engineering provide methodologies for systematic and efficient design of complex systems. Unfortunately, however, methods of Product Development and Computational Mechanics are formulated on different levels of abstraction, and are therefore difficult to connect. 

As an example, consider the early design stages of large technical systems such as passenger vehicles. This development phase is characterized by a significant lack of knowledge about its final state. Product Development methods such as the so-called V-model propose a top-down design process, where design goals are broken down to requirements on system components – however, without specification how this can be accomplished to produce quantitative results. Standard methods of Computational Mechanics are hardly applicable in this scenario, since important data is not yet available. Quantitative treatment can be provided by Computational Mechanics, however, when epistemic uncertainty caused by lack of knowledge is considered appropriately.

The aim of this symposium is to better connect the fields of Computational Mechanics with Product Development and, in doing so, to foster more intense scientific exchange between academia and industry. Areas of interest include, but are not limited to:
- Modelling for early design stages, including physically motivated surrogate models, model reduction, mathematical surrogate modeling, response surface technologies,
- Advanced optimization methods for multi-disciplinary optimization, commonality optimization, solution space optimization,
- Epistemic uncertainty modeling for the treatment of lack of knowledge and model uncertainties in early design stages.