Application of data-driven methods in power systems analysis and control

Abstract

The increasing integration of variable renewable energy resources through power electronics has brought about substantial changes in the structure and dynamics of modern power systems. In response to these transformations, there has been a surge in the development of tools and algorithms leveraging real-time computational power to enhance system operation and stability. Data-driven methods have emerged as practical approaches for extracting reliable representations from non-linear system data, enabling the identification of dynamics and system parameters essential for analysing stability and ensuring reliable operation. This study provides a comprehensive review of recent contributions in the literature concerning the application of data-driven identification, analysis, and control methods in various aspects of power system operation. Specifically, the focus is on frequency support, power oscillation detection, and damping, which play crucial roles in maintaining grid stability. By discussing the challenges posed by parametric uncertainties, load and source variability, and reduced system inertia, this review sheds light on the opportunities for future research endeavours.