Neural Network Energy Management System for the Optimal Control of Microgrids

The interconnection of large amounts of non-traditional generation may cause problems to networks designed for ‘conventional’ operation. A Microgrid consists of a combination of generation sources, loads and energy storage interfaced through fast acting power electronics. The aim of operating Microgrid sub-systems is to move away from considering DG as badly behaved system components, of which a limited amount can be tolerated in an area, to ‘good citizens’. The paper aims at assessing the economic benefits achievable by a group of industrial and commercial customers aggregated in a Microgrid controlled with a Central Controller that uses a Neural Network to optimise the schedule of generators and responsive loads. The Central Controller receives market signals, load and generation bids, load and weather forecasts and determines hour by hour the correct dispatch of generators to maximise the value of the Microgrid by minimising the energy costs.