Analysis and application of load forecasting in power systems using intelligent systems

Load forecasting, the first step of power system planning, is of great importance in economic electric power generation and distribution, improvement of system operating conditions, effective system control and energy pricing. Short-term load forecasting enables the provision of economic operation conditions. In this study, Turkey?s 24-hour-ahead load forecasting without temperature data is aimed. For this purpose, four structures, Artificial Neural Networks (ANN), Wavelet Transform (WT) and ANN, WT and Radial Basis Function Neural Network (RBF NN), Empirical Mode Decomposition (EMD) and RBF NN are constructed. Local holidays? load data is replaced with normal day?s characteristics to remove the disturbing effects of those days on estimation, and estimation results of these days are not included in error computation. To have more accurate forecast, regulated load forecasting is proposed. Unregulated and regulated forecast error percentages are computed as average daily Mean Absolute Percentage Error (MAPE) and maximum MAPE. All MAPE values are compared between the proposed structures. Simulation is performed for years 2009-2010 via the user interface created using MATLAB GUI.Key Words Short-term load forecasting, artificial neural networks, radial basis function neural networks, wavelet transform, empirical mode decomposition