Time series modeling approach for forecasting electricity demand in Sri Lanka.

Abstract

In the midst of the current power crisis, forecasting the electricity demand has turned out to be an important component to develop the energy sector. The demand for electricity is one of the utmost essential data compulsories for assessing the quantity of extra capacity essential to make sure an adequate source of energy. Electricity manufacture and consumption are occupying a substantial role in their governmental economy for a country. Therefore, this study suggests a systematic methodology, constructed on the time series forecasting approach to forecast the effective demand of electricity in Sri Lanka. An ARIMA model is chosen to forecast the future demand of Sri Lanka, one of the fast-developing countries. Thus, public planning necessitates good predictions of forthcoming demand. The study engaged secondary data from the Sustainable Energy Authority, Sri Lanka, spanning from 1970 to 2016 and categorized into sectors utilize electricity as commercial, domestic and industrial. In this analysis, the Box- Jenkins and Autoregressive Integrated Moving Average (ARIMA) based outcomes suggested that ARIMA (1, 1, 1), ARIMA (1, 1, 3) and ARIMA (1, 1, 1) are most appropriate to forecast commercial, domestic and industrial electricity demand, correspondingly. The suggested ARIMA models are used to deliver an eight-year forecast of the electricity demands in the country. Also, the domestic and commercial demands were increasing further rapidly than demand in the industrial sector.