A New Algorithm For Power Systems Controlled Islanding Based on constrained spectral clustering

In this paper a new algorithm is presented for power systems controlled islanding based on constrained spectral clustering. The proposed algorithm minimizes two objective functions of power flow disruption in transmission lines and generation - demand imbalance in islands to ensure transient stability within islands and to minimize necessary actions in reconfiguration of transmission system, generation rescheduling and load shedding, respectively. The algorithm searches all the solution space and investigates feasibility of all islanding candidates. Then the method identifies and selects practical schemes and finally runs the optimum plan. The proposed strategy is not only able to distinguish between important disturbances from the others and prevents unintentional islanding but also creates a minimum number of islands, reduces the time and cost of islanding plan. The proposed strategy is simulated on IEEE 39-bus test case system using MATLAB and PST toolbox. The results verify the efficacy and validity of the proposed islanding algorithm.