Short term load forecasting in electricity markets in Turkey

As energy consumption rises, forecasting electricity parameters becomes a significant advantage on efficient power system applications, planning and decision making in deregulated power markets. An accurate short term load and price forecasting model is crucial for efficient decision making, management and utilization to gain economic optimization and avoiding unpro table operations as well as inefficiencies in generation,transmission and distribution from both consumers and producers perspective in competitive electricity markets like Turkish power industry. In this study, time series analysis including lagged variables that have been presented in autoregressive models in combination of speci c categorical variables (hours of day, days of the week, months of the year and special events of Turkey) and environmental indicators as hourly temperature data in terms of heating-cooling degree hours have been applied in short term load and price forecasting, the day ahead hourly forecast of electricity load and market price. With using AR parameters for load and ARIMA parameters for price, 4 different load models in years 2011 and 2012 and a price model for 2015 constructed and with the effect of categorical variables and environmental indicators, new composite models proposed by applying multiple linear regression to forecast future loads and prices with high accuracy. As a result, the comparison of actual and observed data is studied and the power of model is tested with illustrating on various regression tests. Consequently, the results have shown that proposed models gave low percent of errors with extremely accurate day ahead forecasts considering Turkey's electricity load and price profi le.