New battery materials characterization and battery cell and module modeling techniques were developed. In particular, a cutting-edge GHz (Giga Hertz) electrochemical microscope was developed with high temporal and lateral resolution, operating in a glove box with environmental control. In addition to the broadband frequency measurement of electrochemically active materials using the GHz microscope, a fast 2D nano-mechanical mapping was developed including high-speed nano-indentation protocols for measurement of battery cell interfaces. On the cell and module level, calibrated EIS (electrochemical impedance spectroscopy) and fast time-based pulse test methods were developed for end-of-line quality control. Additional test methodologies were developed on battery modules with 300-500 cells. These battery modules were designed, simulated and manufactured, including electro-thermal aging and corresponding modelling. A test-setup was developed for automated testing of battery modules up to 60 kW, including BMS (battery management system) integration. An interlaboratory comparison of EIS results on cell & module level was done in order to achieve high-quality test-results. For battery modeling, a 1D homogeneous FEM (finite element method) model was developed, delivering high computational efficiency while maintaining robust accuracy. Thereby, the model provides a robust platform for simulating the complex interplay of electrochemical, mechanical, and thermal processes that drive LiB aging. By incorporating multi-scale effects and advanced modelling techniques, this work lays the foundation for predictive simulations of LiB performance and degradation under realistic operating conditions. In addition, multi-fidelity manufacturing modelling tools were developed as well as DEM (discrete element method) contact. We developed an image classification scheme using computer vision for enhanced cell manufacturing, with applications towards SEM (scanning electron microscopy) images of battery anodes. Furthermore, a feature selection method has been developed to identify key battery data characteristics for categorization, and data labelling criteria were defined using parameters such as cycle life, impedance, and charge capacity. Additionally, an integrated SoH prediction model was developed and tested with experimental and simulated datasets. For pilot-line materials optimizations, we studied post-mortem analysis of LiBs, allowing to study electrochemical and calendar aging of cylindrical LiB. Furthermore, to go beyond LiB, Mg-ion batteries (MiB) were developed, including the pilot line assembly of a prototype single-layered pouch cell and its electrical formation. Overall, the developed models and digital twins allow the optimization of cell and pack production at pilot-line scales, reducing development times, waste materials, and battery scrap.
