Article

RED NEURONAL PARA LA CLASIFICACIÓN DE FALLAS EN LÍNEAS DE TRANSMISIÓN A PARTIR DE REGISTROS DE OSCILOPERTURBOGRAFÍA

Dyna 01/2008;
Source: DOAJ

ABSTRACT El diagnóstico de fallas eléctricas en líneas de transmisión a alto voltaje es una tarea compleja no solo por la cantidad de información involucrada que puede provenir de diversas fuentes como SOE, SCADA y registradores, si no también por la variabilidad misma de las fallas. Dicha complejidad impacta en la oportunidad y certeza del diagnóstico, factores particularmente importantes para el análisis en tiempo real donde rápidamente deben tomarse pautas adecuadas para el restablecimiento del sistema eléctrico de potencia. En este artículo se propone el uso de de una red neuronal con aprendizaje por regularización bayesiana y finalización temprana para la clasificación de fallas a partir de registros de osciloperturbografía provenientes de registradores de falla y se muestra su efectividad para una amplia variedad de casos de entrenamiento y validación, los cuales son obtenidos por medio de un modelo de ATP con el cual se simularon la cantidad de fallas eléctricas requeridas.

0 Bookmarks
 · 
74 Views
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper introduces advanced pattern recognition algorithm for classifying the transmission line faults, based on combined use of neural network and fuzzy logic. The approach utilizes self-organized, supervised Adaptive Resonance Theory (ART) neural network with fuzzy decision rule applied on neural network outputs to improve algorithm selectivity for a variety of real events not necessarily anticipated during training. Tuning of input signal preprocessing steps and enhanced supervised learning are implemented, and their influence on the algorithm classification capability is investigated. Simulation results show improved algorithm recognition capabilities when compared to a previous version of ART algorithm for each of the implemented scenarios.
    IEEE Transactions on Power Delivery 05/2005; · 1.52 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The application of neural networks to power systems has been extensively reported. In the field of protection, neural network based protection techniques have been proposed by a number of investigators including the authors. However, almost all the studies have so far employed the back-propagation neural network structure with supervised learning. It is the purpose of this paper to report some recent studies on different neural network models, particularly those with combined supervised/unsupervised learning applied to fault classification for complex transmission lines. The neural networks concerned here include: (i) back-propagation net; (ii) feature-map net; (iii) radial basis function net; (iv) counter-propagation net and (v) learning vector quantization net. Special emphasis is placed on a comparison of the performance of the five neural networks in terms of size of the neural network, learning process, classification accuracy and robustness. The outcome of the work serves and provides guidelines on how to select a particular neural network from a number of different neural networks for a specific application.
    Electric Power Systems Research 01/1997; · 1.69 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Although Bayesian analysis has been in use since Laplace, the Bayesian method of model--comparison has only recently been developed in depth. In this paper, the Bayesian approach to regularisation and model--comparison is demonstrated by studying the inference problem of interpolating noisy data. The concepts and methods described are quite general and can be applied to many other problems. Regularising constants are set by examining their posterior probability distribution. Alternative regularisers (priors) and alternative basis sets are objectively compared by evaluating the evidence for them. `Occam's razor' is automatically embodied by this framework. The way in which Bayes infers the values of regularising constants and noise levels has an elegant interpretation in terms of the effective number of parameters determined by the data set. This framework is due to Gull and Skilling. 1 Data modelling and Occam's razor In science, a central task is to develop and compare models to a...
    Neural Computation 07/1991; · 1.76 Impact Factor

Full-text (5 Sources)

View
44 Downloads
Available from
May 28, 2014