Integration of an artificial neural network to predict electrical energy consumption

The scope of this paper is to present an artificial neural network (ANN) capable to perform a midrange estimation and prediction of the Hellenic electricity load demand. The inputs of the ANN are: (a) temperature, (b) relative humidity, (c) day of the week, (d) month and number of air conditioners. The proposed ANN model implements one hidden layer consisted of eighteen neurons. The learning method used is Levenberg - Marquardt and the transfer function is the Hyperbolic Tangent Sigmoid. Matlab Neural Network Toolbox is used throughout the whole simulation process.