Electricity Consumption Forecasting in the Western Balkan Countries

Abstract

AbstractThis paper presents results of electricity consumption forecasting in the Western Balkan countries. Electricity consumption forecasting in the Western Balkan countries Bosnia and Herzegovina, Serbia, Montenegro, Albania, Kosovo and North Macedonia are done using conventional techniques based on extrapolation techniques using correlation between gross domestic product and electricity power consumption. Also, for electricity consumption forecasting is used advanced technique artificial neural intelligence. Inputs for training artificial neural network are for the first test case gross domestic product, population and decomposed time series of electricity consumption using Huang’s Empirical Mode Decomposition. The forecasting results using time series of electricity consumption processed with Discrete Wavelet Transform as input for training artificial neural network are also presented. Results of all used techniques are also presented, compared and discussed. Forecasted results of electricity consumption in Western Balkan for 2035 year indicate increase of about 20% of electricity consumption compared to 2017 year.