Medium-term electricity market price forecasting: A data-driven approach

Medium-term electricity price forecasting is necessary for several applications in electricity markets, such as pricing derivatives, maintenance scheduling for generation companies, and budgeting and fuel contracting. However, this is a complex task because of the inherent dependence of price to other sometimes unpredictable variables, such as variations in availability of different supply resources. This paper presents two regression-based linear forecasting models to predict the monthly average of electricity spot prices in deregulated electricity markets, with specific focus on systems with large penetration of hydro generation units. The forecasting horizon is a full year, i.e., the models are used to generate 12-month-ahead forecasts. Numerical results are provided for Nord Pool market.