Wake Fast-Calculation And Power Optimal Dispatch For Wind Farms

Limited by investment in land resources and collection systems, the space between wind turbines that are arranged in a certain way in large wind farm is not too big, which highlights the wake effects. The downstream wind speed is decreased, and the turbulence intensity is increased because of wake effects. And then, the power output of wind turbines in the wake area is reduced, and the fatigue damage is aggravated. Fastly to calculate the wake distribution, and accordingly to optimize load distribution between wind turbines, which can reduce the wind power loss caused by wind farm wake and reduce the fatigue damage of wind turbines. Therefore, It is an important subject for wind farm operation and maintance theory and technoloy to study the wind farm wake distribution fast-calculation method and active power optimal dispatch. Around this topic, the study is completed in four areas.For the influence of wind direction on the flow correlation of wind turbines, a flow correlation calculation method between wind turbines for flat terrain wind farm is proposed, which is used to calculate the flow correlation calculation between wind turbines under varying wind direction. A calculation method of wind turbine wake superposition areaes is proposed, which is used to calculate the superposition areas of wind turbine wakes under different superposition conditions. The flow correlation calculation method and the calculation method of wind turbine wake superposition areaes are integrated, the wind farm wake distribution fast-calculation model is established. The Horns Rev wind farm in Denmark is selected as a case study. The wind farm wake distribution is calculated under the conditions that the wind speeds are 8.5m/s and 12m/s, and the wind directions are is 222 degrees and 270 degrees. The calculation results show that the wind farm wake distribution fast-calculation model can accurately calculate the wind farm wake distribution under different wind speed and direction.A short-term wind speed model of natural wind passing through wind turbine is studied, and the calculation and analysis of the model are presented. The calculation results show that the short-time wind speed model can actually reflect the dynamic characteristics of natural wind passing through wind turbine. The maximum power point tracking method based on extremum seeking control of doubly-fed induction generator is studied. The 6kW doubll-fed wind turbine is selected as object, the extremum seeking control method and hill-climbing method are compared and analyzed, the comparison results show that the extremum seeking control method are better than the hill-climbing method. The coupling relationships between the condition parameters of variable speed wind turbine and the wake distribution of wind farm are studied, and the correspoonding mathematical model is established. The calculation results show that the wake speed decreases with the axial induction factor or thrust coefficient increases, and the wind turbine power output increases with the axial induction factor or thrust coefficient increases.In order to solve the complex problems of wind farm optimal dispatch with high dimension, nonlinear and multi-parameters coupled, the wind farm optimal dispatch model is established based on genetic algorithm(GA) and particle swarm optimization(PSO) algorithm. In shis optimal dispatch model, the wind farm power output is selected as the objective, the wind turbine axial induction factor is selected as optimization variable, genetic algorithm(GA) and particle swarm optimization (PSO) algorithm are chosen as the optimization searching algorithms. Combined with the wake distribution fast-calculation method, the wind farm power output optimal dispatch is realized under varying wind speed and direction. The Horns Rev wind farm in Denmark is selected as a case study. When the wind speed and wind direction is 8.5m/s and 270 degrees, the overall power output of wind farm is improved by 4.88% after optimization with GA algorithm, and the overall power output of wind farm is improved by 9.96% after optimization with PSO GA algorithm. The wind farm optimal dispatch result with PSO algrithm is better than that with GA algorithm.For large-scale wind farm with a large number of wind turbines and strong flow coupling, GA and PSO algorithm are difficult to solve the optimal solution. Therefore, a multi-agents (multi-agents, MA) algorithm is studied for large-scale wind farm optimal dispatch. In MA algorithm, each wind turbine is seen as an agent, and the agent structure having wind turbine characteristics, the objective function of the agent, the grid environment and the learning operator are defined. The MA optimal dispatch algorithm is realized by programming. The Horns Rev wind farm in Denmark is selected as a case study for calculation and analysis. When the wind speed and wind direction is 8.5m/s and 270 degrees, the overall power output of wind farm is improved by 10.74% after optimization with MA algorithm. The superiority of the proposed MA algorithm in solving the wind farm optimal dispatch is proved.