Research On Wind Power Real-time Prediction Based On Atomic Sparse Decomposition

Wind power is crucial to the low carbon energy,it has the potential to achieve sustainable energy supply.In recent years,wind power has become one of the countries to focus on the development of renewable energy.Wind power is rapidly developing,the installed capacity increase year by year,by 2020,the global wind power installed capacity will reach 1.2 billion kilowatts,able to meet 12% of the world's electricity needs.In recent years,China's wind power installed capacity in multiply,by the end of 2014,China's total wind power installed capacity of 114609 megawatts,has made China become the largest country in the world installed capacity.In 2011,according to the energy bureau released documents "interim measures for the wind power forecasting management",the real-time forecast report refers to the moment next 15 points to four hours of forecasting,time resolution for 15 minutes.So this topic research the real-time wind power prediction is based on the time interval of 15 minutes of the wind power time series as the main research object,and carries on a 16 step rolling forecast the short-term wind power.The prediction results,can serve the wind generator real-time active output adjustment,and has important significance in raising the utilization coefficient of wind energy.This topic starts from the fluctuation characteristics of wind power,first read the literature at home and abroad,find or define indexes of the wind power fluctuation characteristics,analyse the probability distribution of wind power fluctuation characteristics,analyse the reasons of the wind power fluctuation characteristics.Read literature at home and abroad of the fluctuation characteristics of wind power,understand the research progress of wind power prediction,analyzing the cause of the wind power prediction errors,and introduce the characterization of wind power prediction error indicator.Research the document of atomic sparse decomposition theory at home and abroad,the ASD theory was applied to wind power in the early period of the time series decomposition.On the basis of the existing wind power prediction model,the ASD theory combines the existing forecasting model is applied to the short-term real-time prediction of wind power,and analyze the accuracy of the new combined forecasting model for wind power real-time prediction.Analyze the wind power real-time prediction error,to verify the validity of the new combination forecast model.In the last,this picture build a wind power prediction platform based on VB programming language.