Automatic signal segmentation based on abrupt change detection for power systems applications

Abstract

Detection of abrupt changes in the signal parameters can be used to segment the signal. This has been particularly useful in automatic disturbance recognition in power systems, where the signals are first segmented event-specifically and processed subsequently to recognize and analyze the disturbances. This paper describes several abrupt change detection-based segmentation algorithms and their applications in the power systems. The applications include automatic disturbance recognition, synchronization, relay performance analysis etc, where recorded signals from the power network in South Africa have been used.