Short-Term and Mid -Term Load Forecasting Using Multi Layer Perceptron’s and Radial basic Function

Abstract

Load forecasting plays an important role in power systems supply-demand action. For power companies, load forecasting is vital since this
forecasting is a base for planning of future development, the economic dispatch, determining the security and control systems and effective operation
(investment and decisions for electric generating company) in power systems. In this paper, two Neural Networks (NNs); i.e. Multi Layer
Perceptron’s (MLP) and Radial Basic Function (RBF) are proposed for short and mid terms load forecasting. The data of Azarbayjan Electrical
Network in West-North of Iran has employed for training of these NNs. Four statistical indices used to analyze obtained results and compare abilities
of MLP and RBF, these indices are: Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Symmetric Mean Absolute Percentage
Error (SMAPE) and Relative Error Percentage (REP).