Abstract: The spatial distribution of wind turbines and the inherent uncertainty of wind speed led to strong fluctuations in wind farm active power, and such fluctuations deteriorate the grid-friendly operation performance of wind farms. To this end, this paper proposes an optimization method for suppressing wind farm fluctuations based on Model Predictive Control (MPC) coupled with wake interactions. Specifically, this paper leverages the coordinate transformation method to predict the active power of a wind farm for different wind speed magnitudes and directions. Subsequently, based on this prediction model and combined with wind speed forecasting using Least Squares Support Vector Machine (LS-SVM), this study integrates wake effects, wind direction deviations, and unit constraints into a multi-dimensional coupling optimization objective within the MPC framework for the first time. The optimization problem considers the variance of the wind farm's active power output. The case study shows that the average relative error and root mean square error of active power are both reduced about 90% compared with the proportional distribution method, thereby validating the effectiveness of the multi-dimensional coupling model in suppressing fluctuations.

Key words: wind farm, wake control, fluctuation alleviation, model predictive, active power