We work on the computer-aided analysis and prediction of complex systems in engineering including their environment. Therefore, we focus on two key challenges. On the one hand, how to model systems with strong interactions of different physical phenomena such as solid mechanics, fluid mechanics, and electrostatics. To this end, we develop advanced multiphysics computer simulations. On the other hand, in highly complex systems, modeling is often difficult because the governing mechanisms often remain poorly understood or are extremely expensive to model. To address this challenge, we develop advanced, in particular physics-informed, machine learning methods to accelerate the analysis, prediction and design of such systems. We cover the whole range from basic methods development to various application areas including materials science, production technology, biomedical engineering, and biomechanics.
Methods: Smoothed Particle Hydrodynamics (SPH) simulations of fluid-structure interactions (left), coupling the finite element method (FEM) and discrete element method (DEM) (center), Constitutitive Artificial Neural Network (CANN) as a physics-informed machine learning architecture for materials science (right).
Materials Science: High temperature creep in titanium aluminide (TiAl) simulated with Molecular Dynamics (MD) (left), computer simulation of electron microscopy for generating synthetic training data for machine learning in image segmentation (right).
Manufacturing: We simulate complex manufacturing processes such as mixing with FEM and DEM (top left), sintering with phase-field FEM models (top center and right) and different additive manufacturing methods with SPH (center and bottom row).
Biomedical engineering: Biodegradable implants are degraded during healing of a lesion and thus support a particularly comprehensive regeneration. We support their development by simulating degradation processes in vitro (left), mechanical behavior (center) and coupling of degradation and mechanical behavior (right).
Biomechanics: For biomedical engineering it is key to understand not only the properties of implants or biomedical devices but also of their physiological environment. Therefore, we develop also comprehensive models of this environment from the tissue-scale, where cells and fibers of the extra-cellular matrix interact (top left) to the organ scale, for example aneurysms in brain arteries (top right) or peristaltic contractions in the gastro-intestinal tract (bottom row).