This page contains automatically translated content.
Research colloquium: Presentation by Dr. Marc Fina from KIT on design optimization of shell structures taking into account polymorphic fuzziness and structural stability
As part of the research colloquium for students, doctoral and habiliation candidates, we cordially invite you on Tuesday, 09.12.2025, at 16.30 in room 3516 (Mönchebergstr. 7). We are pleased to accounce the lecture by Dr.-Ing. Marc Fina, Institute of Structural Analysis, Karlsruhe Institute of Technology with the title
“Design optimization of shell structures considering polymorphic fuzziness and structural stability”
Abstract
Today, modern load-bearing structures must not only be robust, but also offer high performance and resource efficiency. Shell structures can fulfill these requirements due to their particularly efficient load transfer. However, their load-bearing capacity is often limited by their buckling behavior. In addition, material and geometric imperfections, such as deviations in the shape and thickness of the shell, residual stresses and variations in the boundary conditions and material parameters, have a major influence on the load-bearing behavior. In a design optimization based on a semi-probabilistic safety concept, possible uncertainties of the imperfections are taken into account by very large safety factors.
In reality, data and information are subject to aleatory and epistemic uncertainties. In the concept of polymorphic uncertainty, the basic uncertainty models (random, interval and fuzzy variables) are combined to quantify both types of uncertainty in a structural analysis and optimization. The aim is to adequately capture the uncertainties based on the existing database using suitable uncertainty models. This should reduce the high safety factors in the future. The numerical implementation of polymorphic uncertainty modeling requires a computationally complex, nested loop algorithm (see figure). To reduce the computational effort, efficient surrogate models, such as artificial neural networks, are absolutely necessary.
In this presentation, the concept for the design optimization of shell structures considering polymorphic fuzziness and structural stability is introduced. The procedure is explained in the context of multi-objective optimization. Using the example of plate structures and circular cylindrical shells made of fiber composite materials, it is shown how the robustness and performance of a structure can be evaluated by taking into account blurring.