The Battery Revolution: Computers and AI Lead the Future of Energy Storage

Empa researchers hope to speed up the development of much-needed new energy storage systems with the help of Aurora battery robots. The Aurora project is part of the European research program Battery2030+, which recently received more than €150 million in funding from the European Union. Furthermore, the project is part of the ETH Board’s “Open Research Data” initiative, which promotes the digitization and free access to research data. The world is in urgent need of new energy storage. Developing completely new battery concepts and exploring their potential is currently a lengthy process.

Corsin Battaglia, head of the Empa Energy Conversion Materials Laboratory in Diebendorf and professor at ETH Zurich, emphasizes: “Our goal is to accelerate this process. This acceleration is currently reflected in the form of the Aurora robotic platform, The platform will take over the fully automated, future autonomous material selection, assembly and analysis of battery cells in the laboratory." As part of the European Materials Acceleration Platform established within the European Battery2030+ project BIG-MAP, the goal is to achieve an approximately tenfold speedup over current development processes.

For internationally competitive battery research and development, time-consuming and error-prone steps in the innovation process are now being automated using Aurora. The robotic platform is currently being further developed in the Empa laboratory together with Chemspeed Technologies AG.

Empa researcher Svalotto-Ferro is implementing the work steps and "training" Aurora. "When robots weigh, dose and assemble individual battery components with constant accuracy, accurately initiate and complete charging cycles and perform other repetitive steps, researchers can use the data generated to drive the innovation process forward," he said.

However, in the future, Aurora will also learn to work autonomously. Using machine learning, Aurora AI can create mathematical models and decide which experiments should be conducted next, and which materials and components are particularly promising for desired battery applications.

Because Aurora AI can be used independently of materials, battery chemistry, and power generation, it could be used not only to improve lithium-ion batteries, but also to develop replacement sodium-ion batteries or batteries with self-healing mechanisms in the future. With Aurora, researchers can effectively monitor and evaluate the numerous process steps in the development and manufacturing of battery cells, and can trace the data back to the source at any time. They can also make laboratory prototypes (such as saltwater batteries or solid-state batteries) more efficient and more reliable. Rapid introduction to the market will further accelerate the innovation process and provide a comprehensive digital strategy in R&D for Industry 4.0.