Study On The Short-Term Wind Power Forecasting Technology Of Large Wind Farm

With the rapid growth of wind power in China, the proportion of wind power in the power grid is increasing gradually, and large scale wind power grid has a serious impact on the safety operation of power system. And effective wind power forecasting can provide reference for the stable operation and dispatch of the power grid. In view of the inaccuracy and unstable of the large wind farm power forecasting, this paper presents a new power forecasting model for large wind farm. The specific research contents are as follows:First, analysis of large wind farm parameters and characteristics.The characteristics of wind speed and wind direction are analyzed and the rules of the whole wind farm are researched, and the power of the wind field is analyzed. Then the physical characteristics of the entire wind farm can be with accurate positioning.Second, for the problem of incomplete and bad data in the large wind farm. The bad data are removed by real power curve and the correlation coefficient matrix method is used to fill the data. The noise and other factors often lead to wind speed data generated burr and spike phenomenon, the paper use the new particle filter to filter the wind speed data, eliminating the burr phenomenon and making the data smoothly. Then the processed data are as the input data of the prediction model.Then, in view of the difficulty in selecting the parameters of the large wind turbine power forecasting model, a IAFSA-SVM wind farm model is established. For support vector machine parameters are difficult to select, the improved fish swarm algorithm was used to optimize the parameters selection. To solve the problem of the limitation of artificial fish in a fixed field of vision and step, by adding adaptive regulator to automatically adjust the fish in foraging and rear end behavior in the visual field and the step. Then the different test function experiments show that the improved fish swarm algorithm has better optimization effect. At last, establishing the single wind turbine power prediction model based on IAFSA-SVM, and the experimental verification is carried out through the wind turbines of two typical wind farms.Last, aiming at the problem of instability and low accuracy of the traditional wind power forecasting method, the paper proposed a new power forecasting strategy based on sample cross correlation function. It combined the strategy with IAFSA-SVM model in predicting wind power, and two typical wind farm case study verified the prediction model has a better effect. In addition, according to the wind power forecasting system of the practice company, a wind farm power forecasting system is designed, which verified the application of the method.