Optimized Lag Times Of Wind Speed By Monsoon And Short-term Combined Forecast Of Wind Power

Monsoon is a major wind system that prevails in a p lanetary scale and reverses its direction seasonally.The classic monsoon is a peculiar phenomenon in some regions such as Asia and Africa.China is one of the most complex areas of wind speed change in the world.The prediction accuracy of wind power in C hina seems to be lower than those in Europe and the United States,if the observation precision and the distributions of observatories keep the same.To better achieve the smart grid construction in C hina,we need to consider the monsoon phenomenon,and improve the accuracy of wind power prediction.Combined forecast and spatial correlation will become the focus in the wind power prediction,such as support vector machine,artificial neural network,partial least squares regression and so on.Using different prediction methods to set up the combined forecast model,and selecting the best combined forecast model.To analyze the similar degree and the lag times of wind speeds between the different places in monsoon regions by different methods,and to improve the spatial correlation prediction by lag times.Spatial correlation prediction by using characteristics of monsoon is the accurate and reliable prediction and early warning for wind generation ramping.It is the reliable method of ultra short-term wind power prediction.Therefore,the spatial correlation prediction of the future wind speed in the lower reaches using the wind speed recorded in the upper reaches of the monsoon region in C hina,and it is the main research goal of this paper.The main content of this paper is arranged as follows:(1)By using historical data and NWP data of Tianjin,making the single prediction models by AR,ARIMA,RBF and other methods,and compared with NWP.To set up the different combined forecast models by using linear regression,support vector machine,partial least squares regression and other methods.NWP is generally better than the other single prediction models.The combined forecast is generally better than the single prediction models,and the partial least squares regression method is obviously better than other combined forecast methods.Research is focused on spatial correlation prediction of wind speed of the winter monsoon.(2)The wind speed time series between different places are obvious similar after lag times.By wavelet,correlation coefficient,mutual information and other methods,to analyze the correlation of the wind speeds between the different places in monsoon regions,and to have the lag times.To have the autoregressive prediction and the spatial correlation prediction by optimizing the lag times.The spatial correlation prediction is better than the autoregressive prediction in the monsoon periods.It is preliminarily proved that monsoon can improve the effect of spatial correlation,and it is the reliable method for the wind generation ramping and the ultra-short-term and short-term prediction of wind power.