Título: Treatment of the Uncertainties Involved in the Electric Power Systems Fault Section Estimation with the Aid of Probabilistic Neural Networks
Autores: Cardoso Junior, Ghendy; Rolim, Jacqueline Gisèle; Zürn, Hans Helmut
Resumo: A probabilistic neural net is used as solution methodology during the task of identification of electric faulted sections. Such a net presents a great generalization capacity and it results in the answer with the maximum probability of being correct. It also estimates the likelihood that a new pattern is a member of any of our predefined classes. Models on the operation logic of the protection system used in lines, busbars and transformers are also proposed. A compact neural net, independent of the electric configuration, is associated to each one of these models.
Código DOI: 10.21528/CBRN2001-021
Artigo em pdf: 5cbrn_021.pdf
Arquivo BibTex: 5cbrn_021.bib