Disturbance detection, location and classification in phase space

This study proposes a scheme to detect, locate and classify the disturbances of power system signals in the phase space. The scheme is proposed on the belief that the features of a signal can be more clearly revealed if it is embedded in the phase space through a proper embedding strategy. In this study, because of the fact that a disturbance signal comprises two parts - the disturbance and the normal part, it is aptly embedded to the phase space so that the two parts are decoupled and distinguished from each other. The normal part forms a regular shape in the phase space whereas the disturbance forms a shape that deviates from the regular shape, and the two shapes jointly constitute the embedded signal. Consequently, the detection, location and classification are performed based on the information obtained from the embedded signal. The scheme has been tested on six types of disturbances under various conditions, and the simulation results have demonstrated the accuracy and computational efficiency of the scheme.