Fast charging, life extension and safety are crucial elements for a wider application of electric vehicles, which is a prerequisite for the transformation of the mobility and transportation sector towards more sustainability and, on a longer time scale, potential CO2 neutrality. A crucial element of these battery systems is the battery management system (BMS), which controls the charging and discharging of a battery consisting of a large number of cells. As the development time of a battery is shorter than its expected lifetime in a car, these BMS are conservatively tuned to ensure safe operation as the main design goal. Increasing performance and longevity while improving safety would be possible if performance degradation or failure of individual cells could be reliably predicted, taking into account new data collected during battery operation.

The Helmholtz Artificial Intelligence Cooperation Unit-funded project i2Batman (“Intelligent, individual battery management using spectroscopy and machine learning”) is investigating the use of advanced AI techniques for an intelligent BMS. In a collaboration between IEK-9, the Steinbuch Centre for Computing (SCC) at KIT and the Theory Department at the Fritz Haber Institute (FHI) of the Max Planck Society in Berlin, a new concept is being developed to dynamically optimize the charging protocol of batteries to increase their lifetime and performance. The project aims at an AI-supported BMS that combines continuous intelligent (AI-optimized, considering design of experiment concepts) data acquisition with a numerical “digital twin” of the battery to develop a better and more detailed charging protocol at the level of individual battery cells. As a result, the model is robustly parameterized and the parameters are continuously and automatically updated as the battery is used.