Research On Short-term Wind Speed Forecasting Method Based On Time Series Analysis Theory

With the increase of installed capacity of wind power all over the world,The share of wind power in the power grid is increasing.However,due to the intermittent and random of wind speed,large-scale wind power access poses a serious challenge to the safety,stability and power quality of the power grid,which limits the development scale of wind power generation.To solve this problem,the wind speed of the wind farm needs to be predicted accurately.With the prediction results,the power dispatching department can adjust the scheduling plan in time to optimize the system operation and at the same time reduce the standby capacity and operating cost of the power system.Against this background,this paper studies the prediction of short-term wind speed in combination with time series prediction theory.Because of the time series and auto-correlation of wind speed series,the time series model can be established for wind speed series,but the prediction results of time series model have lag.In order to solve this problem,this paper starts with two aspects.(1)A self-excitation threshold auto-regressive model is established based on the nonlinear characteristics of wind speed.The wind speed is segmented and linearly processed,and the accuracy of wind speed prediction is improved to a certain extent.Threshold Auto-Regression(TAR)models employ piece-wise linear auto-regressive models to model and approximate the data.According to the value of the threshold variable,the time series region is divided into several segments,and each segment is established with different linear auto-regressive models.(2)Aiming at the variational unsteady characteristics of time series of wind speed,we need to consider the heteroscedasticity effect of residuals when modeling each part by using time series method.Therefore,ARIMA-GARCH model is established to fully exploit the information in the residuals.(3)In order to further improve the prediction accuracy,this paper introduces the ARIMAX model.Compared with the ARIMA model,the ARIMAX model adds exogenous variables related to wind speed data to obtain more wind speed related information.Exogenous variables are obtained by wavelet decomposition combined with ARIMA-GARCH model.Finally,this paper analyzes and summarizes the prediction errors of each model.