Periodic Reporting for period 1 - DigiCell (BATTERY MATERIAL CHARACTERISATION AND DIGITAL TWINS FOR CELL TO PACK PERFORMANCE IN AGILE MANUFACTURING PILOT LINES AND AUTOMOTIVE FIELD)
Periodo di rendicontazione: 2024-01-01 al 2025-06-30
The overall goal of DigiCell is to improve the manufacturing process of Lithium-ion batteries (LiB) and beyond Lithium batteries using advanced materials, analytical models, and artificial intelligence (AI). This will help to reduce production costs, materials waste, and the CO2 footprint while also improving the battery's electrochemical performance. To achieve this goal, DigiCell employs new measurement tools and multi-scale modelling combined with AI and machine learning (ML), thereby impacting quality control, process optimization, and energy management, as demonstrated in three pilot lines. Both low-power cells and high-power battery packs are manufactured, and metrology graded test methods are developed that are further disseminated via European and international standardization activities. In particular, multi-scale adaptive modelling including materials science, end-of-line characterisation, and nonlinear-electrothermal models will be linked and used together with AI and ML to increase battery safety and state-of-health (SoH) evaluation. Furthermore, digitally integrated process models and calibration methods for pilot lines are developed, which are also applicable to beyond LiB for future pilot lines. Finally, new test methods and quality control systems are developed for integrated automotive module stacks with up to 3000 cells including aging and SoH. Overall, DigiCell aims to make the European battery manufacturing ecosystem more efficient and competitive.
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.
Relevant communication and exploitation activities were performed and significant scientific, economic, and societal impact was generated. For example, a number of open workshops focusing on battery materials and advanced test methods contributed to the scientific impact. Scientific papers were published regarding advanced materials characterization methods and battery tests, as well as battery and pilot-line modeling approaches. Major conferences related to electrochemistry, materials science, and battery technology were attended. Regarding economic impact, various ISO/CEN (International Organization for Standardization; Comite Europeen de Normalization) pre-normative actions were provided supporting standardization and international demonstrations in the field of battery and materials tests. For example, interoperable CHADA (characterization data) and MODA (modeling data) were developed, and DigiCell contributed to the EU CENELEC Workshop Agreement (CWA) revision, which establishes standardised approaches for materials characterisation, data and workflows. A case study was published in Zenodo as an appendix to the updated CWA. Furthermore, hierarchically structured battery test data were provided in the DigiCell Zenodo community allowing for semantic interoperability based on a co-developed battery ontology together with EMMC (European Materials Modeling Council) and EMMO (Elementary Multiperspective Material Ontology).