Federated Learning for Adaptive Modeling and Edge Forecasting in Energy Communities
FLAME
Privacy-preserving forecasting of demand and renewable production for Energy Communities, trained at the edge.
- Programme
- Horizon Europe, Open Call
- Area
- Energy
- Status
- Ongoing
- Our role
- Technology provider: federated forecasting and FedMaestro orchestration
Overview
FLAME is a decentralised AI platform for intelligent energy forecasting in smart grids and Energy Communities. It combines Federated Learning, edge intelligence and advanced time-series forecasting to predict electricity demand and renewable energy production without centralising sensitive data.
Key features
Household and community-level demand forecasting at the edge · solar and wind production forecasting from local signals · demand response and flexibility management · integration with existing energy management platforms.
Impact
FLAME gives Energy Communities the forecasts they need to share and store energy well, while the data of every household stays where it is produced.
More R&D projects
ARIEL
Federated AI orchestration for edge systems, privacy-preserving forecasting for EV fleets and solar generation.
02FLORA
An ovulation-tracking app that keeps personal data on the device while training prediction models collaboratively.
03FLAMENCO
A federated learning application that helps identify communication disorders in children.