ESTIMATION OF THE ELECTRICAL ENERGY CONSUMPTION

C. Volosencu

Abstract

This paper presents a simple example of design, modelling and simulation of a model of the electrical energy consumption based on feedforward neural networks. A database with electrical energy consummated on a year is used to train and to test the neural network. Levenberg-Marquardt method from the neural network toolbox of Matlab was used for training. Good quality criteria were obtained for a prediction for a short term with a simple neural network with only a hidden layer and a reduced number of neurones.

Keywords

measurement of electrical energy, power systems, estimation, feedforward neural networks, Levenberg-Marquardt method, Matlab

Download

IMEKO-WC-2000-TC4-P164.pdf

DOI

-

IMEKO TC

TC4 - Measurement of Electrical Quantities

Back to the proceedings