Dynamic Separation of Microgrid System to Maximize Reliability in a Smart Grid

Abstract

The proliferation of renewable generation into distribution networks requires a paradigm shift in designing traditional control and operation strategies regarding system resilience. Distribution networks can be viewed as a group of small microgrids (MGs) when individual MGs are in islanded operation. Therefore, to ensure stability in distribution networks, it is crucial to search for the optimal solutions to define the boundaries of these interconnected small MGs. This paper proposes a dynamic separation algorithm to enable a distribution network to respond to unforeseen disturbances effectively. To achieve this, we consider two constraints: maintaining a dynamic supply-demand balance and securing at least one reserve power source. The proposed algorithm utilizes K-means clustering algorithm based on partition matrix obtained from the $k$-way partitioning scheme for the constraints. The proposed partition matrix can also minimize the total active power flows of distribution lines that connect different small microgrids. The performance of proposed algorithm was tested using a modified IEEE 33-bus radial distribution system and indicates that the proposed algorithm gains more advantages than the original k-way partitioning scheme has.