Innovative Data Analytics of Low Voltage Frequency Measurements for the Detection of Power System Oscillations

Oscillations. It leverages low voltage frequency measurements acquired from grid-edge devices connected at secondary distribution networks. These single-phase Low Voltage (LV) derived frequency measurements, acquired via the level-crossing method, are shown to capture more effectively the dynamics of demand and distributed generation compared to traditional transmission level devices. However, the LV environment introduces significant noise and distortion, primarily due to harmonics and power quality issues, which further complicate frequency analysis. While conventional filtering techniques can attenuate these distortions, they risk eliminating critical oscillation modes, thereby limiting their suitability for oscillation detection. To address this problem, the paper proposes a novel approach that integrates a notch suppression filtering technique with established analytical tools, such as Fast Fourier Transform analysis, and spectrogram visualisation. The method is validated using data from grid-edge devices across the Great Britain power system, where the system is partitioned into nonoverlapping regions for targeted analysis. Results demonstrate enhanced identification and visualisation of dominant oscillation modes. This facilitates more intuitive and efficient control room decision-making by reducing the unnecessary control actions that can result in over expenditure without compromising the system integrity or neglecting genuine power system oscillations.