The Future Data-Driven Based Voltage and Frequency Control in Renewable Rich Power Systems

The increasing integration of inverter-based resources (IBR) in place of conventional synchronous generation (CSG) poses significant obstacles to the total system frequency stability of renewable energy dominated power systems owing to lack of rotating mass and embedded stored kinetic energy. Renewable energy sources (RESs) contribute a significant amount of generation, and due to the absence of inherent inertia, they experience high rate of change of frequency (RoCoF) and frequency nadir when exposed to a disturbance. Traditional methods for inertia estimation rely on mathematical models and assumptions that may not accurately capture the dynamic behaviour of renewable-dominated microgrids. In this paper, a new data-driven based hybrid approach for frequency nadir and RoCoF improvement is proposed, which will help grid operators to take anticipatory measures like adopting frequency control actions in order to make sure that the grid runs smoothly and reliably. RoCoF is an essential indicator of stability and acts as an input for protection system in power systems. This method aims to capture the real-time frequency and voltage characteristics of the system owing to massive availability of data. Further, machine learning (ML) algorithms are employed to extract meaningful patterns and relationships from the available data, enabling accurate frequency and voltage control of the test system.