Short-term electricity power demand forecasting for Turkey

In this study, factors that affect electricity demand in Turkey have been determined and econometric model has been developed for forecasting electricity demand. As a result of literature research and correlation tests, it is concluded that factors such as temperature, humidity, sunshine hours and lightening, highest and lowest temperatures, industrial production index, currency basket, household consumption, market coupling price, system marginal price, capacity utilization rate and gross domestic product (GDP) affect electricity consumption. Data set which consists of daily data that are observed from December 2012 to March 2016, were estimated by Least Squares method. While data between 2012 and 2015 were used to develop the model, data of the first quarter of 2016 were used to test the model in analyses. Previous day electricity consumption values and dummy variables were added to the model for the purpose of reflecting autoregressive structure of electricity consumption and seasonal effect. As a result of analyses that were performed through Eviews 8 by Least Square method, the model that consists of temperature, humidity and previous day electricity consumption gave the most successful prediction result. When forecasts of electricity demand for first quarter of 2016 and actual values are compared, deviations ratio between actual and estimated values are seen to be maximum %7,5 and minimum %0,04. The constructed model substantially caught the rising trend and seasonal fluctuations. Keywords: Electricity consumption of Turkey, electricity demand forecast, least squares method.