Applying Artificial Intelligence to Distributed Generation in Local Networks
Abstract:
The paper describes the application of artificial intelligence to the operation and energy management of a local net work consisting of a group of interconnected households operating as a distributed generation system for electric power and a district heating system for thermal energy. The proposed Artificial Intelligence Protocol (AIP) helps a control unit to manage power supply from renewable sources with the aim of a null energy balance for the local network, avoiding grid dependence and optimizing the system energy performance. The AIP selects the most efficient power source for power supply and energy exchange throughout the local network. The AIP achieves the null energy balance by adjusting operational parameters to regulate the output power for the individual household installation and the energy distribution network. The AIP is applied to a group of unbalanced electric and thermal energy households, configuring a local network with an electric distributor and a heating ring to exchange electric and thermal energy between houses. The AIP application to this local network results in an accurate prediction of electric power generation, higher than 99.7%, showing a global deviation of 0.1 kWh/day. Null energy balance prediction is highly accurate, 97.1%, with a maximum daily deviation of 9.61 kWh/day out of 209 kWh/day energy exchanged corresponding to thermal losses.
