Multi-time Scale Active Power Optimal Dispatch in Wind Power Cluster Based on Model Predictive Control

Multi-time scale optimal dispatch of active power for wind farm cluster is an effective technical way to improve the dispatch of wind farm cluster and promote the accommodation of wind power. To cope with the issues on poor active power control among wind farms and large deviation in tracking the scheduling, a novel multi-scale coordinated dispatch approach for wind farm cluster was proposed based on model predictive control (MPC). In this approach, a combined forecasting model based on the variance- covariance variable weight was adopted to provide forecast for dispatch. A priority set of active power output was established based on the forecasted output characteristics of the cluster, by considering the main varying trends of wind power, including the rising, stability and declining of wind power. At the stage of day-ahead dispatch, a dispatch model aiming at maximizing wind power accommodation was established, by considering the deviation between the forecasting value and the dispatch value of the wind power cluster. At the stage of intra-day dispatch, in order to reduce the adverse impact of day-ahead forecasting error of wind power cluster, the finite time domain rolling optimization and real-time feedback correction of active power in wind power clusters was proposed based on MPC. In the feedback correction procedure, the current output state of the wind power cluster was taken as the initial value of the next round of receding horizon optimization dispatch. A case study was carried out using actual data from a large-scale wind farm cluster. Results show that the proposed approach can improve the accuracy and stability of wind power cluster dispatch. The day-ahead dispatch, the intra-day rolling dispatch and the real-time dispatch can be effectively coordinated, thus the wind power cluster output can be better smoothed. © 2019 Chin. Soc. for Elec. Eng.