Data-Driven Distributionally Robust Optimization for Real-Time Economic Dispatch Considering Secondary Frequency Regulation Cost
With the large-scale integration of renewable power generation, frequency regulation resources (FRRs) are required to have larger capacities and faster ramp rates, which increases the cost of the frequency regulation ancillary service. Therefore, it is necessary to consider the frequency regulation cost and constraint along with real-time economic dispatch (RTED). In this paper, a data-driven distributionally robust optimization (DRO) method for RTED considering automatic generation control (AGC) is proposed. First, a Copula-based AGC signal model is developed to reflect the correlations among the AGC signal, load power and renewable generation variations. Secondly, samples of the AGC signal are taken from its conditional probability distribution under the forecasted load power and renewable generation variations. Thirdly, a distributionally robust RTED model considering the frequency regulation cost and constraint is built and transformed into a linear programming problem by leveraging the Wasserstein metric-based DRO technique. Simulation results show that the proposed method can reduce the total cost of power generation and frequency regulation.