The Development And Research On Wind Power Short-term Forecasting System Of Wind Farm

With the high-speed consumption of the world energy,as well as the deteriorating of ecological environment,wind energy as a clean,renewable energy has get more and more attention of the whole world.The wind power has a strong randomness and volatility.The unit capacity of wind turbine and the scale of grid wind farms has continuous expansion,so the wind power penetration is also increasing year by year.The attendant problems of wind power have become increasingly prominent.Therefore,the accurate forecasting of wind power is necessary to ensure the power system to be safe,reliable,economical,stable.It is also an problem that to be overcome urgently at this stage of China.Under this circumstance,it is necessary to launch a profound research on the problems of wind power short-term forecasting system.This paper introduced the development of wind power at home and abroad,summarized the current wind power forecasting techniques.Then briefly introduced the classification of wind power forecasting,this paper mainly studies the wind power short-term forecasting of a particular wind farm.This paper analyses the characteristics of wind speed and wind power,and the basic step of realizing wind power forecasting.This paper focuses on the influence factors of Wind power.This paper based on the most widely used neural network principle and using the numerical weather data as the input,this paper put forward an improved wavelet neural network method and summarized the step of wind power forecasting by using the improved wavelet neural network method.Simulations have been done in MATLAB by established the BP neural network forecasting model and the improvement wavelet neural network forecasting model,then using the tow models to predict the wind power in 24 hour advance of one wind farm in Inner Mongolia Autonomous Region.The predicted results were compared with the actual results,it is show that the improvement wavelet neural network forecasting model is better,also verify the feasibility of the improvement wavelet neural network method in wind power forecasting.After the author mastered the algorithm and theories of wind power short-term forecasting,the author used VB to design a wind power short-term forecasting system.The system has made the forecasting process to be simple and convenient.And also made easy to operation for the dispatcher.