Electricity Demand Forecasting Methods Used in Turkey and Their Effects on Investments in Electricity Sector

Abstract

Electrical energy is one of the most critical elements for accessing applications that are related to modern life. Due to the nature of electricity as a commodity that cannot be stored on a large scale, demand management should be handled with particular care. For this reason, realistic demand forecastings must guide investment decisions. This is also an essential element for the balance of payments and investment planning. Thus, electricity investments are also essential to determine accurately associated with unsaturated electricity demand in Turkey. Furthermore, many tools and materials used in the electricity sector are requiring high technology imported goods. Besides, the fuels used in electricity generation and the planning of the power plants that use these fuels affect all economic activity. As a result, planning is the most important factor for power plants. For this purpose, some energy models and forecasting methods have been developed for use in many countries worldwide. Some of these models and methods have also been used in Turkey, but high rate deviations have been observed up to 85% in long-term official forecastings. Even though rapid developments of computer and software technology ensure to enhance forecasting accuracy in many countries, such a deviation ratio is an unacceptable subject in Turkey. In this thesis, as a common and high forecast accuracy rate, the ARIMA Model and Artificial Neural Networks (ANN) have been used for future 10-year demand forecasting in Turkey. These forecasting results have been compared with the actual amounts and official forecasting. Then, the relationship between these results and power plant investments in Turkey has been discussed. Accordingly, it is analyzed that existing investments and future investments can meet the electricity demand. Finally, the proposals for investment in Turkey in the international developments and practical guidance and solutions have been introduced.