Iran electric energy consumption forcasting using artificial neural network

Nowadays, with the development of industry and technology, electricity energy production has been highly increased. For meeting the increasing demand of energy, the distribution system and transmission facilities should be planned in a way that future priority has been taken into consideration. For accurate using of electricity energy, the produced energy amount has to be based on ordered aggregate. Thus, the electricity energy planning system is aimed to be well designed. Moreover as much as correctness of the estimation of energy production is acceptable, proposed plans will be more successful. In this work, Iran electricity energy consumption estimation is carried out by using of Artificial Neural Network (ANN) with data gathered between the years 1978-2014. The parameters used in this thesis are consumer electricity price, imports, exports, GDP, economic growth and population. In order to evaluate the success of ANN model and performance of estimation, multiple regression and time series models has been applied and comparisons has been made. The obtained MAPE ratio for results of ANN, ANN time series and multiple regression is 1.76%, 1.54% and 2.07% respectively. By using the developed model Iran electricity energy estimation was made up to 2021. Result of the estimation models were compared with future electricity forecasting values obtained from Iran ministry of energy.