Forecasting Model for the Amount of Water Flowing into the Reservoirs of the Electricity Generating Authority of Thailand (EGAT)

Abstract

The objective of this study is to construct the appropriate forecasting model for the amount of water flowing into the 11 reservoirs of the Electricity Generating Authority of Thailand (EGAT), namely, Bhumibol Dam, Srinakarin Dam, Sirikit Dam, Vajiralongkorn Dam, Ratchaprapha Dam, Ubolratana Dam, Sirindhorn Dam, Chulabhorn Dam, Huai Kum Dam, Nampong Dam, and Bang Lang Dam. The forecasting models used in this study were decomposition method, Holt-Winters exponential smoothing method, Box-Jenkins method, and combined forecasting method. The amount of water flowing into the reservoirs was secondary data gathered from Electricity Generating Authority of Thailand during January 2010 to December 2020 (132 values). The data was divided into two sets. The training data set had 120 values, which were the data from January 2010 to December 2019, and were used for constructing the forecasting model. The test data set had 12 values, which were the data collected from January 2020 to December 2020, and were used for checking the accuracy of the forecasting models. The criterion of model evaluation was the lowest root mean square error (RMSE). Minitab 18 and Excel Office 365 were used as the data analysis program. Research results indicated that the combined forecasting method using weights based upon a regression analysis method had given the lowest RMSE.