Edge Computing for Federated Learning–Enabled Energy Management and Monitoring Systems

Abstract: Despite growing digitalization of energy management systems (EMS) in buildings and local networks, integrating data-driven monitoring and control remains challenging. Edge computing offers enhanced local processing that can be embedded in existing EMS infrastructures. Within the European HYSTORE project, two edge-computing use cases were developed: a network-level platform enabling coordinated EMS operation using building thermal mass estimates, and a building-level controller supporting advanced monitoring and control of thermal energy storage. Both architectures are evaluated for design, integration, and potential for federated learning, enabling secure, efficient, and privacy-preserving multi-agent control.