Short-term Prediction Of Wind Speed And Wind Power Of Large-scale Wind Farm Considering Time Series Random Fluctuation Characteristics

Energy shortage and environmental pollution have become important factors constraining the development of social economy,and have attracted the attention of the whole country and the world.As a renewable energy with the advantages of clean,high efficiency and low cost,wind energy can solve the problem of energy shortage and environmental pollution fundamentally by large-scale development and utilization.It has become the most valuable new energy for commercial development.However,due to the remarkable randomness and uncertainty of wind power,the rapid growth of grid-connected scale will increase the difficulty of system planning and reasonable dispatch,which is not conducive to the safe and stable operation of power systems.Whether wind power can be widely and reasonably consumed by the power grid depends on the ability of the power system to respond to power fluctuations.Accurate prediction of wind speed and wind power can largely smooth the volatility and randomness of wind power,and greatly improve the grid-connected capacity and supply reliability.Therefore,accurate prediction of wind speed and wind power has important practical significance and research value.In view of the rapid growth of wind power grid-connected scale,this paper briefly describes the time-series model of wind power generation technology and its working principle,and discusses the impact of large-scale grid-connected wind power on system network loss,voltage quality and system stability.The characteristics of the random fluctuation of wind turbine output timing and the difference of wind speed and power of each unit in the field are analyzed in depth,and the spatial and temporal distribution characteristics of wind energy resources in large-scale wind farms and the clustering effect of wind turbine output power are analyzed reasonably.It lays a theoretical foundation for the subsequent construction of wind power models with different capacity clusters.The variation of wind speed fluctuation is easily affected by various meteorological factors such as temperature,humidity and pressure,and has great randomness and uncertainty.This paper starts from the two aspects of prediction model input set and key parameters optimization,predictability of wind speed series is quantitatively analyzed according to joint index combined by recurrent rate and determinism.The wind speed series is reconstructed using the parameters optimized with joint index.And the best input sets of prediction model are obtained by embedding dimension and delay time.Finally,combined with the actual wind farm wind speed data,the COA-SVR optimization model trained by the best input set is used to carry out the prediction research.According to the transformation mathematical model of wind speed and wind power,the equivalent wind energy utilization coefficient is introduced to evaluate the interaction between wind turbines.Based on the historical wind speed and wind power measured values,the equivalent wind energy utilization coefficient matrix is obtained,and the output power equivalent model of wind turbine group is constructed according to the maximum predicted wind speed value,and then realize the short-term wind power prediction of the large-scale wind farm.The feasibility and rationality of the proposed method are verified by the measured data of wind farms.