Improving Wind Power Accommodation Via an Interval Model Predictive Control Strategy

Peng Lu,Qihuan Dong,Ershun Du,Chen Fang,Yun Su,Ning Zhang

DOI: https://doi.org/10.17775/cseejpes.2023.04700

IF: 6.014

2024-01-01

CSEE Journal of Power and Energy Systems

Abstract:The intraday and real-time dispatch of a power system integrated with large-scale wind power is usually implemented via an open-loop strategy (e.g., predict-then-optimize (O-PO)): first, the wind power is predicted; then, the optimized dispatch strategy is developed by using predictions to accommodate more wind power output and pursue the optimal economy. However, accommodation could suffer under the open-loop strategy because the utilization of prediction information may be overly conservative in this optimization strategy. To address this, a closed-loop predict-and-optimize framework based on a model predictive control strategy is developed to improve wind power accommodation, which can transform the intraday dispatch (15 min) and real-time dispatch (5 min) problems into mixed-integer programming problems to coordinate the wind power cluster and upward and downward spinning reserves of the automatic generation control (AGC) units and non-AGC units. Finally, a modified IEEE-39 with 10 wind farms is used to conduct numerical tests. The results show that the proposed strategy exhibits better performance and is suitable for refined allocation schedules for all wind farms. Furthermore, the proposed method's effectiveness in improving wind power accommodation and avoiding power transmission line over-limit is demonstrated.