Load forecasting at Kuantan and Kuala Terengganu in Malaysia

Abstract

Short term regional load forecasting plays an important role in ensuring adequate load supply, preventing shortage and excessive storage of electricity that in turn would add cost to the utility company. Demographic and weather factors do contribute toward the load consumption at the different areas under a particular demographic region. Temperature, the most outstanding weather element has been shown to have great effect on load forecasting. In this study, time series model, ARIMA is applied on the historical load data at two selected meteorological stations, Kuantan and Kuala Terengganu. The selection of the appropriate model for the two data sets were based on the smallest value of AICC statistic. The AICC for Kuantan and Kuala Terengganu stations are 849.253 and 781.335 respectively. The forecasting errors for prediction of the next three hours peak load at the two stations are summarized in terms of root mean square error (RMSE) and mean absolute percentage error (MAPE). The RMSE for Kuantan and Terengganu stations are 0.9 and 0.5 respectively while MAPE are 4.7% and 7.05% respectively. The correlation between temperature and peak load consumption at the chosen stations are analyzed by using simple regression. The results show that there is a positive correlation between peak load and temperature.