Power-demand forecasting using a neural network with an adaptive learning algorithm

Abstract

An artificial neural network with an adaptive-Kalman-filter-based learning algorithm is presented for forecasting weather-sensitive loads. The proposed model can differentiate between weekday and weekend loads. This neural-network model has been implemented using real load data. The results reveal the efficiency and accuracy of the proposed approach in terms of short learning time, rapid convergence and the adaptive nature of the learning algorithm.