Data-driven surrogate modeling of structural instabilities in electroactive polymers

The project will provide a data-integrated multiscale framework for the continuum-mechanical modeling of porous media. A particular focus will be on theory and numerics of variational scale-bridging techniques for diffusion-driven problems in consideration of large deformations. With such a multiscale framework, we aim at the data-based analysis and optimization of porous microstructures with regard to their overall hydro-mechanical properties. One of our major goals is to provide a data-analytics based approach to instability prediction. Instabilities play a crucial role in materials design and arise, for example, in the form of structural instabilities at micro-level. The occurrence of instabilities depends on a number of conditions including material properties, microscopic morphology and overall coupled loading. From a theoretical viewpoint, the investigation of instabilities calls for minimization-type variational principles. These will be numerically implemented in a consistent way into the multiscale framework. Critical instabilities will be revealed by Bloch-Floquet wave analysis. In order to provide efficient and reliable predictions of the associated instability phenomena, we will equip the multiscale formulation with modern tools of machine learning.