Título: Neural Networks to Transient Stability Analysis of Electrical Power Systems
Autores: Minussi, Carlos R.; Ramos, Lilian Milena; Marchiori, Sandra Cristina; Lopes, Mara Lúcia Martins; Lotufo, Anna Diva P.
Resumo: The objective of this work is the development of a methodology for transient stability analysis of electric power systems (critical time determination for short circuit faults) based on a neural network. Here, it is used back-propagation algorithm with an adaptive process based on fuzzy logic. This methodology results in a fast training, when compared to the conventional formulation of back-propagation algorithm. The input is composed by the nodal active and reactive electrical power vectors and data about the fault, represented in a similar way to the binary code. The output (critical time) is determined by a computational program based on a hybrid methodology – simulation / Liapunov Direct Method. It is a less dimension network when compared to those ones found in the literature, however it needs a low computational cost in the execution training. To illustrate the proposed methodology, an example is presented which considers a multi-machine system composed of 10 synchronous machines, 45 buses, and 73 transmission lines, based on the configuration of a southern Brazilian system.
Código DOI: 10.21528/CBRN2001-006
Artigo em pdf: 5cbrn_006.pdf
Arquivo BibTex: 5cbrn_006.bib