WIND SPEED FORECASTINGBASED ON TIME SERIES ANALYSIS, WAVELET ANALYSIS AND GARCH ANALYSIS.

Abstract

Now days wind energy is one of the sources of electricity production due to its intermittent nature
there is need for an accurate wind forecasting technique.lot of research has been done to improve the precision
of wind speed prediction. In this paper ARIMA, Wavelet theory and GARCH analysis are used to forecast the
wind speed. In this paper ARIMA approach is used to forecast wind speed using available previous year wind
speed measurements. In this paper wavelet theory is used to decompose highly variable wind speed measurements
in to several approximate stationary time series. ARIMA models are developed for each decomposed approximate
stationary time series and then forecasting results are obtained by this hybrid method. In order to remove the
volatility from the forecasted results of ARIMA model GARCH method is used and then obtaining the forecasting
results. By comparing the MAPE of all the three methods we found GARCH method is the best method for wind
speed forecasting. For this analysis the wind measurements are obtained from Basaveshwar engineering
college, Bagalkot wind site.