Distributed Ledger Based Management of Local Energy Markets with a Federated Learning Approach

This paper presents an innovative methodology for managing a local electric market based on artificial intelligence techniques, integrated with a distributed ledger technology platform. The methodology allows an aggregate of users, for example constituting a local energy community, to optimize its energy costs by adopting a local energy market that manages its controllable energy resources in real-time. To achieve this result, the electricity market is managed by means of a distributed ledger platform used for both the certified recording of market operators' bids and for the sharing of a market-solving deep neural network algorithm. This market-solving platform is continuously adapted to the external changes in energy production, consumption and prices. By sharing the state of the system by means of the distributed ledger, the proposed platform allows every operator to locally define its optimal production/consumption and adapting its status according to the community energy needs. The proposed platform has been implemented with a computer-based simulation software and successfully tested for a day-long, 1-minute timestep. The results presented in the paper shown the usefulness of the tool developed in a renewable energy community real case scenario.