Data-enabled predictive control for frequency control in future heterogeneous power systems

To meet the future target of decarbonizing Europe, Transmission system operators (TSOs) in the EU increase the efforts of installing renewable energy sources (RES) to cover the growing demand. Some of these RES, such as wind turbines, work in grid supporting/grid forming modes as required in the recently released grid code and provide frequency support services. The non-homogeneous dynamics, control strategies, and intermittency of the interconnected generating units following load demand and weather conditions result in frequently changing power system topology. This, in turn, directly affects the grid frequency/voltage controllers’ performance and could endanger the system’s overall stability. To address this issue, the recently emerged Data-enabled Predictive Control strategy (DeePC) shows its ability to learn the dynamic of the unknown power system directly and only from the measurements (model-free) and optimize the active power injection, which ends up in a significant improvement of grid frequency support and stability. Simulation results verify the effectiveness of the proposed DeePC strategy compared to the conventional one in improving frequency control.