BattDesign-GAME: Battery Design via GPU-Accelerated Simulation and Machine Learning for Energy Storage Materials

• Applicant: Dr.-Ing. Simon Daubner
• Subject Area: Computer-Aided Design of Materials and Simulation of Materials Behaviour from Atomic to Microscopic Scale
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 578806410

My vision is to develop fast, intelligent microstructure design for batteries validated on lithium-ion cells, with methods that generalize to next-generation systems such as sodium-ion, solid-state, and layered or graded electrodes. By pioneering inverse material design based on differentiable physics (physics-governed simulators that support end-to-end gradients) I will combine the rigor of physical modelling with the speed and flexibility of modern ML. These tools will predict ion transport and degradation directly on 3D voxelized microstructures and enable inverse design: working backwards from target performance to optimal microstructures and manufacturing conditions. By integrating microscopy and electrochemical testing with simulation, this approach reduces reliance on resource-intensive trial-and-error and shortens the path to high-performance cells. This vision is anchored in a fundamental scientific question: How do microstructural properties - such as grain morphology, porosity, and tortuosity - govern performance and the interpretation of electrochemical measurements? To address this, I will integrate high-resolution 3D imaging, classical electrochemical experiments, and differentiable physics to extract material parameters directly and consistently. This integrated approach overcomes limitations of traditional models, which either rely on oversimplified assumptions or are computationally expensive, and it makes large-scale, data-driven microstructure design tractable.

• DFG Programme: WBP Fellowship
• International Connection: United Kingdom
• Participating Person: Professor Samuel Cooper, Ph.D.