Data-driven Localization of Forced Oscillations in Power Systems

This paper proposes a data-driven approach to locating the source of forced oscillations, which constitutes an important practical requirement for the normal operation of power systems. The source of forced oscillations is pinpointed by conducting Causality Analysis based on PMU measurements. In order to obtain the portion of PMU data for Causality Analysis in nearly real-time, Sparse Principal Component Analysis is leveraged to determine the starting point of forced oscillations. The effectiveness of the proposed approach is tested in the IEEE 68-bus benchmark system. Extensive simulation results showcase that the proposed method can achieve higher accuracy in comparison with a recent localization algorithm, without assuming any knowledge of system model parameters.