Microgrid Energy Management System based ANN of the Two-Step Structure

This paper proposes an energy management system(EMS) based artificial neural network(ANN) for hybrid AC/DC microgrid(MG), which is composed of distributed generation(DG), energy storage system(ESS), variable loads. An ANN is applied to derive the optimal operation mode of the MG and the power reference of the ESS by using the power data of unpredictable DGs and loads. The ANN controller used in EMS was trained by MATLAB deep learning tool box. The effectiveness of the proposed EMS based ANN was verified through the PSIM simulation.