Optimal Power Generation Control Of Wind Farms With Wake Effect

In order to achieve the ambitious goals of "carbon peaking" by 2030 and "carbon neutrality" by 2060.Clean energy with a low carbon footprint are the inevitable alternative to traditional fossil fuels.Wind power has become one of the energy sources that can be developed on a large scale by virtue of its sustainable and non-pollution characteristics.Following the onshore large-scale wind farms development,large-scale offshore wind farms are the focus of the "14th Five-Year Plan" wind power development.In large-scale wind farms,the wake effect of upstream wind turbines will affect the inlet wind velocity of downstream turbines and generate turbulence,resulting in a reduction in the downstream turbines generation and a negative impact on theirs loads.Studies have shown that although each turbine is working in maximum power point tracking(MPPT)mode,the power generation of the whole wind farm cannot reach the maximum due to the wake effect.Therefore,there has been a great deal of interest in increasing the power generation of wind farm by wake effect control.Yaw-control,pitch-control and speed control of wind turbine can all affect its wake flow.From the perspective of maximizing power generation,yaw-control is the most effective control method.This paper will therefore focus on the optimization of yaw angle of wind turbines to improve the wind farm power generation.The main work of this article is as follows:1.Analyzed the principles and characteristics of various wake models comprehensively,combined with the theory of wake deflection to provide the multi-wake superposition calculation method under the yaw conditions,laying a theoretical foundation for the subsequent optimization calculation.2.Given out a wake correction method for actual wind farms.Firstly,the operational data from the actual wind farm were clustered and analyzed to remove measurement errors,shutdown and power limitations data,as well as the data affected by the dynamic process of wind turbine control are removed.Then,the generator power curve was corrected through the cluster analysis.The parameters of the wake model were identified using a heuristic search strategy,based on a least squares algorithm3.Aiming at the problem that the relative positions of upstream and downstream units change due to the random variation of wind speed and direction in large wind farms,and the coverage area of wake area changes accordingly,a unified two-round screening calculation framework for wind farm wake was proposed,and the optimization strategy of wind turbine collaborative yaw angle based on particle swarm optimization algorithm was presented.The optimal structure of yaw control based on the operating wind farm shows that the generation power of wind farm can be significantly increased and the obtained results can provide reference for the optimization control of wind farm power generation.4.The proposed method of using graph neural networks to construct an optimized output model for wind farms utilizes the good information transfer and integration capability of graph neural networks,effectively combining the operation characteristics and geographical correlation of all turbines in wind farm,solved the time-consuming problem of using analytical model for cyclic calculation,and the time for optimal power output calculation is effectively shortened,which lays a foundation for wind farm control.