Charactering Uncertainty of Frequency Regulation Capability: A Physics-Informed Unit Commitment Model Considering Time-Varying Dynamics

Abstract

As the penetration of low-inertia wind power continues increasing, primary frequency regulation (PFR) capability of power systems faces severe risk of reserve shortage. Considering the complex interaction of wind turbines (WTs) and time-varying frequency dynamics, the PFR capability of wind power is significantly influenced by unit commitment (UC) scheme and uncertain wind speed. Therefore, it is essential to account for the uncertain PFR capability in UC stage to prevent frequency incidents. This paper proposes a physics-informed UC model accounting for the probability distribution of wind power output and droop coefficient. The method employs the Koopman operator theory to establish a high-dimensional linear relationship between wind speed and the maximum droop coefficient, and obtains a probability distribution of PFR capability by predicting wind speed, which is incorporated as a constraint in the UC model. Simulation results demonstrate that this approach ensures sufficient PFR capability in UC stage while exhibiting promising real-time performance.