Long term annual electricity demand forecasting by artificial neural networks including socio-economic indicators and climatic conditions

Abstract

Electricity has become a major form of end use energy in present complex society. The influence of electricity is tremendous and has been recognized as a basic human need. It is an important element of infrastructure on which the socio economic development of the country heavily depends. Electricity demand forecasting is very important and crucial for a utility, in order to make right decisions regarding future power plant and network development. Accurate electricity demand forecasting is one of the challenges and several techniques are used in forecasting demand based on the availability of data in each country. CEB in their long term generation expansion planning studies use three long term demand forecasting methodologies namely econometric approach, time trend approach and end user approach. New application for long term demand forecasting based on Artificial Intelligence has identified as important due to its ability in mapping complex non-linear relationships. Therefore under this study, the use AI method based on Artificial Neural Networks for long term annual electricity demand forecasting in Sri Lanka is discussed and modeled including Socio-Economic Indicators and Climatic Conditions.