Study on real-time clustering method for power system transient stability assessment

This paper addresses the issue in critical machines identification for real-time assessment and control of power system transient stability. To satisfy both accuracy and rapidity, on the basis of clustering study on a 3-unit 9-bus system, a novel real-time critical machines identification method is proposed. First, three candidate decomposition patterns are selected by the top three largest predicted angle variation distance; And then an index is proposed to choose the pattern whose equivalent OMIB system operating state goes farthest from the equilibrium point as the best decomposition pattern. The scheme only needs real-time machine dynamic information, and has the advantage of little computation, rapid identification speed and excellent accuracy. Simulations on IEEE 39-bus 10-unit system validate the efficiency of proposed scheme.