Forecasting model of wind speed along the coast of Tha Sala district, Nakhon Si Thammarat province

Abstract

       This research would like to create the appropriate forecasting model of wind speed at an altitude of 40 meters along the coast of Tha Sala district, Nakhon Si Thammarat province. This time series were collected by Solar and Wind Energy Research Unit, Thaksin University. The first series during February 1 – 28, 2011 for an hour period of 672 observations were utilized to create the forecasting models by six statistical methods: Box-Jenkins method, decomposition method, combined forecasting method using the simple average, combined forecasting method using a weighted by the inverse of root sum squares error, combined forecasting method using a weighted by the inverse of sum absolute error, and combined forecasting method using a weighted by the proportion of the value in the eigenvector from the principal component analysis. The second series during March 1 – 7, 2011 of 168 observations were utilized to compare the performance of forecasting methods by the criterion of minimum mean squared error (MSE). The result showed that the Box-Jenkins method by using SARIMA(1, 1, 1)(0, 1, 1)24 with no constant term was the most suitable for this time series.