Wind speed forecast for wind farms based on ARMA-ARCH model

Short term wind forecasting is a very important work to the operation of wind farms and power systems. In this paper, ARCH (Autoregressive Conditional Heteroscedasticity) effects of wind data series are analyzed with Eviews software. Firstly, an ARMA (Autoregressive Moving Average) model of wind speed time series is built. Secondly, ARCH (Autoregressive Conditional Heteroscedasticity) effect of the residual of ARMA model is tested by Lagrange Multiplier, and the corresponding ARMA-ARCH model is set up. Lastly, forecasting performances of ARMA-ARCH model are compared with ARMA model. Validation of ARMA-ARCH model is proved. And the results show that ARMA-ARCH model possesses higher accuracy.