BATMAN

Overview

Data exploration, artificial intelligence, and digital twins for the next generation of batteries

Prof. Mathieu Salanne (PHENIX, Sorbonne Université / CNRS)
Prof. Alejandro Franco (LRCS, Université de Picardie Jules Verne / CNRS)

Research on electricity storage increasingly relies on numerical simulations, typically based on physical principles. BATMAN aims to develop new models that integrate data science and artificial intelligence approaches. Three main objectives are targeted: finding optimized electrolytes and materials, understanding chemical reactions at battery interfaces, and optimizing manufacturing processes. The project also aims to create a digital twin to predict battery performance based on the properties of the manufactured electrodes.

Keywords

Multi-scale simulations, artificial intelligence, digital twins, high-throughput screening, high-performance computing, electrolytes, materials, manufacturing processes

Tasks

Our research

Screening of new materials for next-generation batteries

The first objective is “high-throughput screening,” which aims to identify optimized electrolytes for next-generation batteries and materials suitable for high-power devices. Well-established techniques, such as electronic structure calculations and molecular dynamics simulations, will be used to explore large material databases, whether these materials have been synthesized or not. This will provide lists of material compositions with improved performance for batteries.

Studying battery interfaces with machine learning approaches

The second objective focuses on understanding the chemical reactions occurring at battery interfaces. New machine learning-based tools will be developed to simulate changes at the electrode/electrolyte interface, incorporating electronic structure effects into molecular dynamics simulations. These methodological advancements will enhance the understanding of phenomena such as dendritic growth in all-solid-state batteries and electrolyte decomposition on Na-ion battery electrodes.

Digital twins for battery manufacturing processes

The third objective is to develop a “digital twin” to optimize battery manufacturing processes. Battery electrodes are composed of blends of different materials, each with a specific role. To improve performance, it is crucial to optimize the textural properties of the composite and the interfaces between materials. A digital twin will be created to predict battery performance based on the properties of the manufactured electrodes.