Introduction to Center for Research on Smart Battery and AI to X

As the demand for electrochemical batteries surges,  CROSBAT-AIX aims to be at the forefront of revolutionizing battery management systems (BMS) technology for the future, by incorporating AI accelerators.  While significant progress has been made in energy density, cost, and performance, the challenges of safety and lifetime remain critical and we aim at solving them with the help of AI technology.

In Smart Battery Villum Investigator Project, we have proposed a new BMS technology for large packs featuring cell-level bypass and AI based algorithms for state estimation, thermal balancing and health & safety management including early warnings for accelerated degradation and potential thermal runaways.

Our next goal is to implement the AI-BMS in the current HW platforms of 32bit microcontrollers featured with NPU AI cores and the challenge here is to be able to accelerate the physical models of the electrochemical batteries (e.g. LiB. SSB, LiO2B) using our concept of Multiphysics (thermal, electrochemistry, CFD, mechanics)  physical constrained neural networks in order to be executed in real time and provide a reliable health-scanning function and performance optimization. This new technique will also be applied to electrolysers (SOEC) and fuel-cells (SOFC) which share a common Multiphysics background.

Another goal is to build a foundation model for the power grid using graph neural network and generative AI in order to develop a new AI platform for the control and management in the context of the challenges imposed by the rapid electrification and large scale deployment of AI centers