EDISON

Within EDISON, we develop and demonstrate AI-driven solutions for the efficient management of distributed energy storage systems. We design dedicated protocols and data management systems to enable battery storage to work within Virtual Power Plant (VPP) frameworks. We identify the following key project objectives:

1. Develop an adaptive AI-powered VPP management solution: we build a smart, interoperable platform that coordinates distributed battery storage systems in real time. Edge-based AI algorithms handle monitoring and optimization locally, while secure communication protocols and standardized interfaces keep the system connected. At the local level, deep Reinforcement Learning (RL) agents learn optimal charging and discharging policies directly from interaction with the energy environment, avoiding the need for hand-crafted control rules. For coordination across multiple sites, we employ Multi-Agent RL (MARL) so that each site's agent can learn cooperative strategies while respecting local constraints. To preserve data privacy, we integrate Federated Learning (FL) and Federated Distillation (FD), enabling collective model improvement across sites without transferring sensitive operational data. The platform adapts dynamically to changing grid conditions and environmental factors, with shielding mechanisms ensuring RL agents always remain within safe operating boundaries.
2. Establish AI-based battery lifecycle management: we develop intelligent methods to estimate battery State of Health (SoH) and Remaining Useful Life (RUL) using a self-adaptive approach that combines physics-based battery models with data-driven AI. These hybrid SoX estimators continuously refine their predictions as operational data accumulates, extracting reliable health indicators from field data without requiring disruptive full-discharge cycles. FL is used to improve estimation models across distributed sites, allowing knowledge transfer between heterogeneous battery systems while preserving commercial confidentiality.
3. Enable smart market integration: we develop tools that allow VPPs to participate efficiently in energy markets, including day-ahead, intraday, and reserve markets.
4. System Integration and Validation: we build and test the proposed solutions through both virtual simulations and real laboratory demonstrators. A dedicated VPP simulation environment serves as a testbed for evaluating RL algorithms and FL coordination protocols under realistic grid conditions before physical deployment.

Project Consortium

The EDISON consortium brings together two research institutions and two SMEs with complementary expertise in AI, energy management and battery technology.

• Silicon Austria Labs GmbH (SAL, Lead) is Austria's top research center for electronic-based systems, conducting application-oriented research across microsystems, sensor systems, power electronics, and embedded AI. In EDISON, SAL develops the RL-based energy management algorithms and the federated AI coordination framework for the VPP platform.
• Virtual Vehicle Research GmbH (ViF) is Europe's largest R&D center for virtual vehicle technology, specializing in simulation, digital twins, and model-based development across automotive, rail, energy, and other domains. In EDISON, ViF develops the battery lifecycle management framework with hybrid SoX estimators and contributes cyber resilience strategies through threat risk assessment.
• CISC Semiconductor GmbH is a Klagenfurt-based SME providing hardware and software products for the design, verification, and testing of networked embedded microelectronic systems, with strengths in edge hardware, system integration, and cybersecurity. In EDISON, CISC is responsible for the overall system architecture and solution design and leads the integration of the VPP platform components.
• PIADENO Green Energy Management GmbH is a Klagenfurt-based SME specializing in virtual power plant operation and asset management, enabling customers to participate in energy markets through its all-in-one software platform. In EDISON, PIADENO provides real-world infrastructure for validation and contributes expertise in energy trading and balancing energy markets.