How can we accurately simulate millions of tiny, deformable particles moving through complex fluids? This question lies at the heart of a research retreat hosted at the Speinshart Scientific Center for AI and SuperTech, bringing together experts in computational fluid dynamics, applied mathematics, and artificial intelligence.
Deformable microparticles are found throughout nature and industry—from biological cells and droplets to hydrogel beads, soft catalyst particles, and reactive micro-carriers. As they move through turbulent flows, they continuously change their shape and interact with one another in highly complex ways. Although these processes play a crucial role in medicine, biotechnology, chemical engineering, and environmental science, current simulation methods face a fundamental limitation: they can either model a few particles with high accuracy or many particles with strongly simplified assumptions, but not both at the same time.
The workshop explores a fundamentally new computational approach that combines physics-based modelling with artificial intelligence. Instead of representing every particle through computationally expensive surface meshes, the researchers propose to describe each particle by a compact mechanical state while using neural networks trained on high-fidelity direct numerical simulations to reproduce complex deformation behaviour. This strategy makes it possible to retain the essential physical information while dramatically reducing computational costs.
By integrating machine learning with advanced numerical methods and strict physical conservation principles, the project aims to create a new generation of simulation tools capable of modelling large populations of deformable particles under realistic flow conditions. Such a capability would open entirely new possibilities for understanding complex suspension flows in fields ranging from biomedical engineering and pharmaceutical manufacturing to advanced materials science and industrial process optimisation.
Hosted in the collaborative environment of the Speinshart Scientific Center, the retreat will provide an opportunity to refine this ambitious research vision, strengthen interdisciplinary cooperation, and advance the development of AI-supported scientific computing at the frontiers of computational mechanics.
Please note: This is just an information regarding events taking place at SSC; public attendance is therefore not possible due to the character of the retreat.
