Distributed optimization for microgrids control and management with virtual voltage source segregation

This paper proposes a distributed optimization framework for islanded microgrids (MGs) control/optimization that achieves a global optimal solution with reduced computational and communication complexity. The proposed method breaks down the global optimization problem into simple sub–optimal problems by dividing the whole network into sub–networks. This is accomplished through virtual segregation of distributed generation (DG) voltage sources. In contrast to the existing distributed schemes, where each agent solves a full–scale optimization problem, taking into account information from the entire network and requires parameter consensus, the proposed approach solves individual sub–optimal problems independently. This substantially reduces the computational complexity and communication requirements at higher speed to promptly exchange information diffusion among the agents. Moreover, this study considers a multi–feeder MG comprised of numerous load feeders and sparsely available DGs, a MG system that has received limited attention in existing literature. The proposed distributed method has been tested for power sharing and load feeder voltage restoration in a radial–type multi–feeder islanded MG network; however, it holds potential for broader applications. A comprehensive analytical formulation, MATLAB numerical simulations, and realistic experimental findings for the proposed distributed method provide a detailed understanding of its capabilities and shortcomings.