Identificação de sistemas não-lineares multivariáveis usando redes neurais perceptron multicamadas e função de base radial

AUTOR(ES)
DATA DE PUBLICAÇÃO

2005

RESUMO

The identification of multivariable nonlinear dynamic systems is an important area in Engineering. This dissertation presents a methodology based on artificial neural networks for identification of nonlinear system with some inputs and outputs. This study it is mainly motivated by artificial neural networks to present potentialities for identification of nonlinear systems, such as: ability to treat with complex systems, representation of quantitative knowledge, parallel processing, learning, adaptability and generalization. The evaluated artificial neural networks in this study are: (i) radial basis function neural network and (ii) multilayer perceptron neural network. Simulations had been carried through for two case studies of multivariable nonlinear identification. The studies of cases are of nonlinear systems of generation of electric energy. The results gotten for artificial neural networks had been promising, motivating future research in identification of systems based on new configurations of neural networks in applications in power systems.

ASSUNTO(S)

engenharia de producao sistemas de energia elétrica redes neurais (computação) sistemas não-lineares

Documentos Relacionados