IDENTIFICATION OF LOW-FREQUENCY OSCILLATION MODES FROM PMU MEASUREMENTS

Abstract

The implementation of phasor measurement units (PMUs) in power grids has brought about a significant transformation in the monitoring and management of power systems. PMUs enable the acquisition of precise measurements of voltage and current phasors at high sampling rates, thereby facilitating real-time monitoring and control of power systems. However, as power systems become more intricate and interconnected, the emergence of low frequency oscillations has become a critical issue. These oscillations, which are primarily distinguished by their frequency range, can result in system instability and even cascading failures that can have severe economic and societal impacts. The aim of this project is to investigate the use of the Dynamic Mode Decomposition (DMD) algorithm for identifying low frequency oscillation modes from Phasor Measurement Unit (PMU) measurements accurately. Low-frequency oscillations can pose a threat to the stability of power grids, and early identification of these oscillations is crucial for preventing potential blackouts. First, the DMD technique was tested on five case studies from different scientific papers (without and with (0.2%, 0.5%, and 1%) noise), which are used to verify the performance of the proposed approach, then conducted on an IEEE 10-Machine, 39-Bus system, where a five-cycle fault is applied at bus 30 for a duration of one second. Real power flows over lines 20-13 and lines 37-38 are measured for a duration of time t=2 to 12 seconds. The proposed methodology involves several steps, including data collection and processing, mode identification using the DMD algorithm, and validation of results. The DMD algorithm is used to identify the dominant oscillation modes and their associated frequencies and damping ratios. The study demonstrates that the DMD algorithm is a promising technique for identifying low-frequency oscillation modes and provides valuable insights for system analysis and control. The proposed methodology can be applied to other power systems and extended to include application to real-time larger systems and analysis of the impact of measurement noise.