Variational Mode Decomposition as Trusted Data Augmentation in ML-based Power System Stability Assessment

Balanced data is required for deep neural networks (DNNs) when learning to perform power system stability assessment. However, power system measurement data contains relatively few events from where power system dynamics can be learnt. To mitigate this imbalance, we propose a novel data augmentation strategy preserving the dynamic characteristics to be learnt. The augmentation is performed using Variational Mode Decomposition. The detrended and the augmented data are tested for distributions similarity using Kernel Maximum Mean Discrepancy test. In addition, the effectiveness of the augmentation methodology is validated via training an Encoder DNN utilizing original data, testing using the augmented data, and evaluating the Encoder's performance employing several metrics.

Identifying Oscillations Injected by Inverter-Based Solar Energy Sources

Inverter-based solar energy sources are becoming widely integrated into modern power systems. However, their impacts on the system in the frequency domain are rarely investigated at a higher frequency range than conventional electromechanical oscillations. This paper presents evidence of the emergence of an oscillation mode injected by inverter-based solar energy sources in Dominion Energy's service territory. This new mode was recognized from the analysis of real-world ambient synchrophasor and point-of-wave data. The analysis was performed by developing customized synchrophasor analysis tools deployed on the PredictiveGrid^{TM} platform implemented at Dominion Energy. Herein, we describe and illustrate the preliminary analysis results acquired from spectrogram observations, power spectral density plots, and mode shape estimation. The emergence and propagation of this new mode in Dominion Energy's footprint is illustrated using a heatmap based on a proposed frequency component energy metric, which helps to assess this oscillation's spread and impact.