Wavelet neural network based short term load forecasting of electric power system commercial load

In an electric power system, the system load consists of domestic, commercial, industrial and municipal load sectors. This paper presents an approach for predicting electric power system commercial load using a wavelet neural network. Morlet and Mexican hat wavelets are used to generate the transfer functions of hidden layer nodes of the neural network. A wavelet neural network is trained for a particular power system load. Results show that wavelet neural networks may outperform traditional architectures in approximation and forecasting problems related to electric power system.