WP1 established the foundation for BATTwin framework by analyzing pilots' industrial processes, defects and requirements. Upstream and downstream production flows are mapped in IDEF0 diagrams, the root-causes of disturbances and defects across manufacturing stages are identified, and the quantitative KPIs that upcoming digital twins must improve are defined. A catalogue of functional and non-functional requirements for process- and system-level digital twins was created, addressing their interoperability, explainability and deployment. Lastly, a large European industry consultation has been conducted to generalize BATTwin approach beyond the two pilot plants, gather industrial recommendations for future exploitation and adoption across other cell chemistries, geometries and production scales.
In WP2, a docker-based data ingestion and processing system was designed, using Redis, MongoDB, Neo4j and RabbitMQ to integrate both real-time and historical heterogenous data flows to feed process- and system-level digital twins in WP3 and WP4; federated learning strategies for noisy and incomplete data, and uncertainty quantification framework linking measurement noise and model parameter variability are developed; human-operator feedback dashboards were created, enabling qualitative feedbacks about quality; an ontology-based data model was proposed for enhanced traceability, aligned with Digital Battery Passport requirements.
WP3 resulted in several multi-physics and hybrid digital twins for key battery manufacturing processes: from mixing to calendering, a Coarse-Grained Molecular Dynamics method was developed; instead for downstream processes, laser and mechanical notching models, a parametric kinematic digital twin of the stacking robot to analyze positioning errors and dynamics interactions with material deformation models, and a tab welding model are developed. These models target the common defects identified in WP1, such as edge quality, burr formation, deformation, misalignment, adhesion behavior, weld quality. Furthermore, a COMSOL-based DFN model was developed for cell formation and aging to predict SEI formation behavior at various C-rates. One of the key outcomes of WP3 was preparing aforementioned process models for orchestration and integration in a unified workflow to harmonize inputs and cross-link models, paving the path for user-centric workflows in WP5.
WP4 focuses on modelling and analyzing the entire battery production line in terms of system productivity and logistics performance. For that purpose, a continuous–discrete analytical model and a Discrete Event Simulation model are developed, following a hybrid discretization approach to merge continuous upstream processes with discrete downstream processes under a time-synchronization scheme with the law of conservation of mass. An initial analysis of system-level defect management strategies (e.g. downstream compensation of deviations, quality-oriented assembly, offline rework) is performed to verify their impact on system metrics as throughput, yield and Work-In-Progress. Multi-objective optimization platform will integrate both models to analyze realistic scenarios in pilots to support production managers in strategic decision-making.
WP5 integrates individual the individual digital twins from WP3 and WP4 into user-oriented workflows targeting specific goal-driven industrial applications within the design, operation, and control phases, at both process and system level. These workflows are formalized, and under implementation in specific software package, together with dedicated GUIs and cybersecurity protocols and solutions.
WP6 focuses on the customization of technologies developed in WP2 to WP5 for integration and validation within two BATTwin pilots. The achievement of the target KPIs and TRL of the developed solutions will be addressed, highlighting the remaining gap towards upscaling of the solutions, beyond TRL 4-5.
WP7 ensured the visibility, impact, and exploitation of BATTwin outcomes. The project identity was created (logo, templates, website), and regularly updated content—news, documents, partner materials—was published, leading to strong engagement metrics. Communication and dissemination activities exceeded originally set KPIs for social media, media mentions, website updates and partner press releases. The project participated in major European battery and manufacturing events, presenting scientific outputs, and submitted four publications by the first reporting period. WP7 also established the exploitation and business-planning framework, mapping potential applications, market opportunities and industrial benefits. A systematic IP and technology monitoring activity (updated every six months) tracked patent trends and competing technologies relevant to battery manufacturing digital twins. Finally, WP7 reinforced collaboration within the Battery 2030+ initiative, contributing content, participating in roadmap workshops, and ensuring alignment and synergies with the European battery R&I ecosystem.
