Applications of clustering algorithms in long-term load forecasting

Load forecasting is one of the critical activities in electric power system planning. This paper presents clustering algorithms and their usage in load forecasting on a case study in Zagreb, Croatia. Load data acquisition is not always being systematically conducted in distribution networks and some data often has to be extrapolated. For such methods to work additional computation and grouping algorithms have to be used in addition to classical trend forecasting methods. Furthermore, the paper emphasizes on load forecasting in areas with no load history.