Partitioned cooperative distributed control strategy for fatigue load reduction in waked wind farm

The growing number of wind turbines within wind farms introduces complex wake-induced aerodynamic interactions. These interactions increase the fatigue load on individual turbines and complicate the implementation of real-time control strategies. To address this problem, this paper proposes a partitioned cooperative distributed control strategy for active power regulation, aiming to mitigate wake-induced mechanical load fluctuations and reduce fatigue loads of large-scale waked wind farm. Firstly, the wake sensitivity coefficient is defined, which quantifies the influence of the power output of upstream turbine on the future wake wind speed of downstream turbines. And its size is used to represent the wake interaction strength between turbines. Based on this, the aerodynamic coupling between turbines is modeled as a weighted directed graph. Secondly, Fast Newman network theory, a method of dividing nodes in complex networks into highly cohesive and sparsely interconnected communities, is integrated with wake graph. This integration can divide a large-scale wake wind farm into multiple independent sub-communities. Moreover, a dynamic wake wind speed model is developed using the wake sensitivity coefficient. This model predicts the future wake wind speed changes that result from adjustments in the active power reference. Therefore, the traditional analytical wake model is linearized to make it suitable for real-time control of fatigue load. Then, the predicted wake wind speed is added to the model predictive control as a disturbance term to minimize the wake-induced load fluctuation. Finally, the original centralized optimization problem is decomposed in each sub-communities and solved in parallel to reduce the computational burden. A waked wind farm with 40 wind turbines was used to verify the control performance of the proposed control scheme under different inflow wind directions. Under the wind direction of 180°, the DELs of Mt and Ts are reduced by 37.07 % and 32.22 %, respectively. And the DELs decrease of Mt and Ts are reduced by 13.60 % and 21.59 % under the wind direction of 225°, respectively. The results prove the effectiveness of the proposed method.