Intelligent Post-Disturbance Transient Instability Prediction using Wide Area Measurement System

In this paper, the solution of the post-disturbance transient instability prediction problem using the wide area system is considered. In the proposed scheme, at first, the disturbance initiation is detected based on the magnitudes of voltage phasors received from phasor measurement units (PMUs). In the second step, a primary assessment of the transient stability is made according to the post-disturbance magnitudes of voltage phasors, frequency, and differential frequency and by three trained support vector machines (SVMs) classifiers, separately. Finally, the outputs of the classifiers are combined employing the Naive Bayes (NB) algorithm to make the final decision. The proposed algorithm was implemented on the IEEE New England 39-bus system. According to the results obtained, proposed algorithm can predict the transient instability as early as three cycles after the disturbance initiation using noise-free and noisy PMU data.