Research On Active Power Allocation And Control Of Wind Farm

As a kind of renewable energy,wind energy has been developed and utilized on a large scale in countries all over the world because of its rich resources,high safety and mature utilization technology.With the continuous increase of the grid-connected capacity and unit capacity of wind farms,in order to reduce the error between the output active power of the wind farm and the power command of the grid dispatching center and the impact of output power fluctuations on the grid,the study of the active power distribution of the wind farm has become wind power The research hotspots in the field,this article has carried out the following research around the active power distribution and control of wind farms:(1)Determined to classify units based on unit wind speed,predicted wind speed and output power,and determined the number of unit types.The forecasting method of the combination of gray model prediction,BP neural network and least square support vector machine(LSSVM is used to predict the wind speed of the wind turbine,and the obtained forecast wind speed,real-time wind speed and output power of the wind turbine are smoothed and normalized.,Use the fuzzy C-means clustering algorithm to classify wind turbines,and formulate the order of participating in the active power distribution of wind farms according to the characteristic values of each type of turbines.(2)Use wind farm active power allocation algorithm based on unit classification to allocate wind farm active power.Based on the traditional wind farm power distribution algorithm structure。this algorithm adds a power distribution module for unit classification.According to the unit classification results and power adjustment sequence,the wind farm active power command is assigned to each type of unit,and then each type of unit will receive the active power The instructions are distributed to each unit to complete the distribution of active power of the wind farm.The simulation verification of 14 1.3MW wind turbines has been carried out.The simulation results show that the distribution algorithm proposed in this paper reduces the tracking error of the active power of the wind farm and reduces the fluctuation of the output power of the wind farm.At the same time,it participates in the active power regulation of the wind farm.The number of units has been reduced by 4.(3)Aiming at the shortcomings of traditional PI control strategies in terms of accuracy and response speed of wind farm active power control,model predictive control(MPC)is used to control wind farm active power.Realizes the multi-objective control of wind farm power tracking accuracy,wind farm’s response time to active power commands,and single unit output power.Simulation verification was performed on 10 1.3MW wind turbines.The results show that the use of MPC control strategy improves the accuracy of wind farm active power tracking,shortens the wind farm response time to active power commands,and reduces the output power fluctuation of a single unit.