Prevention of Power Grid Blackouts Using Intentional Islanding Scheme

Intentional islanding is often conducted as the last resort to preserve the electric grid from severe blackouts. The intentional islanding scheme deliberately segregates the power system into a number of self-sustained islands to enhance the transient stability of the power network. In this study, a spectral clustering algorithm is presented to obtain an islanding solution, which results in minimal power flow disruption across boundaries of islands. The constraint in the spectral clustering method is introduced based on generator coherency grouping. The proposed scheme ensures that each island is only comprised of generators, which are synchronized with each other. The proposed clustering scheme utilizes a more computationally efficient and accurate k-medoids algorithm comparing to k-means ones. The proposed intentional islanding method is scrutinized using the IEEE 9-bus and IEEE 118-bus models. The results of simulations demonstrate that the proposed method can effectively prevent blackouts by separating the systems into stable islands.