Development of tool for oscillation detection and localization

Abstract

For this purpose a tool aiming for near real-time oscillation detection and localization is developed. The tool initially utilizes a variation of the OnLine Algorithm for PMU data processing to tackle missing measurements, without distorting the modal content of the data. The algorithm displays remarkable ability to recover effectively signals with clear temporal behaviors.

The widespread implementation of Phasor Measurement Units (PMU) in the North American grid, revealed that the system is under frequent oscillations, that were unobservable previously. These events cause additional wear and tear of equipment and can even lead to failure. For this purpose, detection of such events and locating the source is of great importance. A tool that can perform this task should operate reliably, while requiring the minimum amount of data possible, in order to leave the operators with enough time for remedial actions.

For the detection three modal estimation methods, enhanced by data driven clustering techniques, are utilized to quickly and reliably detect oscillations in the incoming measurements. Utilizing the detection results, a localizations scheme based on the Dissipating Energy Flow is initiated, utilizing only the measurements from the grid and no model information. The effectiveness of the tool is tested and verified on simulated data and real oscillatory events.