Wind Farm Cluster Frequency Regulation Control Strategy Based On Prediction Information

Wind power generation is one of the most mature technologies for renewable energy generation,and it is also the priority for the development of energy policies in countries around the world.With the increase of wind power grid-connected scale,the randomness and volatility of wind power output have brought about more and more negative impacts on the dispatching and stable operation of power grids,mainly reflected in the imbalance of power generation and load,which leads to the stability of system frequency.In order to enable the wind power cluster to provide the necessary frequency support for the power grid,this paper takes the centralized wind farm cluster as the research object,and carries out the active power fluctuation characteristics of the wind farm cluster,the ultra-short-term forecast of wind power and the control strategy of the wind farm cluster participating in the frequency modulation.the study.Firstly,the paper analyzes the fluctuation characteristics of wind farm clusters on time scale and space scale through the measured data.The purpose is to explore the law of wind farm cluster output fluctuation,so as to provide solutions and methods for reducing the impact of wind power fluctuations on the grid.The impact of large-scale wind power on system frequency is discussed,and the role of wind power forecasting in solving wind power cluster frequency modulation is discussed.Secondly,a hybrid prediction model based on machine learning algorithm is proposed for the problem of large wind power prediction error.The model uses the improved modal empirical decomposition and wavelet decomposition as the main algorithm in the preprocessing stage to reduce the non-stationarity of the wind power sequence.In the prediction stage,the neural network,the extreme learning machine and the least squares support vector machine are introduced to predict the decomposed sequence.At the same time,the adaptive neural network model based on particle swarm optimization is used to learn and correct the error distribution of three intelligent predictive algorithms,which improves the prediction accuracy and generalization performance of the model.The accuracy and superiority of the proposed hybrid intelligent algorithm are verified by a case study of measured wind farm data.Finally,in order to give full play to the advantages of wind farm clusters in participating in system frequency regulation,a wind farm cluster frequency regulation control strategy based on prediction information is proposed.The strategy manages each wind farm through the establishment of a wind farm cluster control center,and rationally adjusts the frequency of frequency regulation participation of each wind farm based on the predicted information.At the wind farm cluster level,the operation mode control and reference command distribution of the wind farm are realized through the prediction information of each wind farm;at the wind farm level,the adjustment coefficient of the wind farm is adjusted by the prediction information to realize the dynamic adjustment of the frequency participation of the wind farm.In order to prove the effectiveness of the proposed frequency regulation control strategy,the simulation model was established in MATLAB.The results show that the proposed wind farm cluster frequency regulation control strategy based on prediction information can make full use of the wind farm’s prediction information to optimize the power generation and frequency regulation participation depth of each wind farm.