Topology Optimization of Miniaturized Multi-Resonant Systems Accelerated by Model Order Reduction (optiMuM)

• Applicants: Professorin Dr.-Ing. Tamara Bechtold; Professor Dr. Jan Gerrit Korvink
• Subject Area: Microsystems
• Term: since 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 446995632

The main objective of optiMuM is to enable the efficient and rational design of miniaturized multiresonant devices subject to realistic operational and manufacturing constraints. Because of their compactness and manufacturing limitations, miniaturized multiresonant devices, such as needed for energy harvesting, or magnetic resonance microscopy, are hard to design in a way that meets operational requirements such as precise resonances, quality factors, mode isolation, or transfer functions with easily tunable features. In particular, resonances are typically sensitive to slight changes in toplogy, layout, or manufacturing tolerances, and the relationships among these factors arenot easily accessible to the designer. Very often the relationships are governed by multiphysical effects, arising from different energy domains (mechanical, electrical, magnetic, thermal, etc.) so that geometry has a particularly strong influence that is often counter-intuitive. Optimization can accomplish better designs, but suffers from a large computational burden. Automatic approaches to reduce complexity can alleviate this burden, but only if it can be efficiently combined with the optimization procedure. Currently, there is no clear methodology in place to bring together the objective functions of multiresonant design with efficient optimal design. In this proposal two research groups aim to tackle this deficiency, focusing on two key application areas covered by the application expertise of the principle investigators of this proposal, namely micro energy harvesting, and micro nuclear magnetic resonance. Both applications have been chosen from those research areas, in which substantial design and experimental work has already been accomplished. The respective models can, after spatial discretization, both be described by large-scale second order linear differential equation systems. We therefore work towards a general design optimization technique, which addresses the two application areas, and is applicable beyond, for example for resonant inertial sensors, or resonant atomic force sensors.We will base our approach on two methodological areas in which we have significant track record (and a history of successful prior collaboration), namely topology optimization (Jan G. Korvink) and model order reduction (Tamara Bechtold). Our synergy in these fields will lead to an improved optimization technique and will greatly benefit our application fields.

• DFG Programme: Research Grants