HybridSolver for flexible, resilient and efficient multiscale simulation

In order to simulate and thus understand physical processes ranging from micrometre fibre diameters to rotor blades over a hundred metres long, this project hybridizes classic numerical simulation methods with state-of-the-art machine learning approaches. The use case of fiber-plastic composites promises broad transferability to other material classes, such as filtration media, batteries and rock physics.  Thanks to this long-term and versatile usability, the project promotes the maintenance and expansion of competitiveness in industrial sectors that are considered to be a key area of expertise for the German economy.