Many phenomena in science and engineering evolve on curved geometries embedded in some higher-dimensional space. This work advances their analysis by deriving mechanical models from first principles and developing a numerical framework for the simultaneous solution of entire families of curved geometries. The central idea is to consider multiple manifolds, implicitly defined by level sets of a scalar function, within one model, rather than considering individual geometries successively. The developed numerical analysis based on finite elements is called Bulk Trace FEM which is a hybrid of classical finite element methods and fictitious domain methods. Numerical studies for the various mechanical models verify optimal higher-order convergence behavior in different error measures. Beyond reporting novel mechanical models and numerical methods, this work lays a solid foundation on the overall topic and presents implementation-ready discretization strategies for the simultaneous analysis of partial differential equations on manifolds. Taken together, these contributions provide both a conceptual foundation and a practical tool set, enabling an efficient, reliable, and simultaneous analysis of multiple curved geometries, setting the stage for future developments.
Issue: Open Access E-Book
ISBN: 978-3-99161-131-8
Language: Englisch
Release date: November 2026
Series: Monographic Series TU Graz / Structural Analysis
Many phenomena in science and engineering evolve on curved geometries embedded in some higher-dimensional space. This work advances their analysis by deriving mechanical models from first principles and developing a numerical framework for the simultaneous solution of entire families of curved geometries. The central idea is to consider multiple manifolds, implicitly defined by level sets of a scalar function, within one model, rather than considering individual geometries successively. The developed numerical analysis based on finite elements is called Bulk Trace FEM which is a hybrid of classical finite element methods and fictitious domain methods. Numerical studies for the various mechanical models verify optimal higher-order convergence behavior in different error measures. Beyond reporting novel mechanical models and numerical methods, this work lays a solid foundation on the overall topic and presents implementation-ready discretization strategies for the simultaneous analysis of partial differential equations on manifolds. Taken together, these contributions provide both a conceptual foundation and a practical tool set, enabling an efficient, reliable, and simultaneous analysis of multiple curved geometries, setting the stage for future developments.






