Agent based model for distributed optimization of microgrids operation

Traditional control schemes for power systems as well as recent research works on microgrid control rely mostly on multiple hierarchical levels, addressing different objectives ranging from optimization (high level) to voltage and frequency control (low level). Assuming the presence of lower-level controllers, this paper proposes an approximate agent-based model of the network, applicable to the computation of high level references for the active and reactive power outputs of a population of distributed storage systems within a microgrid. This model allows the application of well-known techniques to achieve robust and distributed computation of these references. The proposed reference optimization is tested by means of simulation, using data from a real distribution grid. The results indicate the effectiveness of the method.