Research On Real-time Forecasting Of Wind Power Based On Chaos

In recent years,the exploitation of wind power in China is rapidly increasing,but because of the fluctuation of wind,wind power has a certain degree of randomness.When wind power is connected to electricity grid,its large-scale fluctuation will have a bad influence for electric power system stability and grid dispatching.The fluctuation characteristics of wind power is useful for the exploitation of wind energy.The accurate forecasting of wind power and the forecasting error analysis is also important for the large-scale exploitation and utilization of wind energy.For an objective understanding of wind power fluctuation characteristics,the measured data of northeast wind farms is used.For an objective understanding of wind power fluctuation characteristics and its chaos characteristics in spatial and temporal distributions,a quantitative indicator to measure wind power chaos,the largest rolling Lyapunov index is presented.And the spatial and temporal distribution characteristics of chaotic characteristics of wind power is analyzed.The existence of periodic components in wind power time series is verified by the autocorrelation coefficient diagram and the periodic chart.The periodic component is extracted by Fourier transform,and the chaotic characteristics of residual component is verified by the recurrence plot.Ensemble empirical mode decomposition is used to denoise the wind power time series in the model,and then the long-range correlation and fractal features of the denoise time series are analyzed.According to wind power forecasting,three forecasting methods based on chaos theory are proposed in this paper.Method-one is ultra-short-term forecasting of wind power based on local one-order weighted method.In the method,the distance is taken as the criterion for determining the adjacent phase points.Method-two is corrected multi-step Lyapunov forecasting of wind power based on Lyapunov forecasting,and the value of rolling forecasting is corrected.Method-three is combination forecasting based on the periodic characteristic of wind power.In the model,the value of the forecasting is the value of forecasting of the periodic component added by the forecasting of the residual,and the new wind power sequence is decomposed and forecasted again.In the paper,the wind power data of a northeast wind farm is used to analyze forecasting error of wind power,and the probability distribution of single-step forecasting error of wind power is analyzed based on chaos.The relationship between the wind power forecasting error and the forecasting step number,the wind power forecasting error and the wind farm output power,the wind power forecasting error and the installed capacity is also analyzed.According to the multi-step forecasting of wind power,a wind power forecasting platform based on VB is established.In the forecasting platform,power information can be directly read from the real-time monitoring system,the modeling is simple,the speed of operation is fast,which can meet the requirement of online application.It is used for ultra-short wind power forecasting,and it is suitable for wind farms in which the meteorological information is insufficient,and the series of wind power is the only available data.