An effective power quality classifier using wavelet transform and support vector machines

• Autores: de Yong, David Marcelo; Bhowmik, S.; Magnago, Fernando
• Año de publicación: 2015
• Idioma: inglés
• Tipo de recurso: artículo
• Materia: Complex Disturbance Detection And Classification; Power Quality; Support Vector Machine; Wavelet Transform
• Institución: Consejo Nacional de Investigaciones Científicas y Técnicas

In this paper we propose a method based on a combination of binary classifiers which are optimized for those special cases where the real signals contain a multitude of events within the analyzed temporal window. These type of events are known as complex events. The proposed Power Quality (PQ) classifier is based on Wavelet Transforms (WT) and Support Vector Machines (SVM). The method uses a One vs. One multiclass SVM. We propose a novel method which is simple, easy to train, and can be implemented with low computational cost. The proposed algorithm consists of a set of simple binary SVM classifiers. Each SVM node is trained separately allowing them to be parallelized. The training stage is performed using single events, however due to the structure of the SVM methodology selected, it allows the system to detect complex events. Tests and training were performed using real complex signals and the results show the proposed methodology to be highly efficient.
Fil: de Yong, David Marcelo. Universidad Nacional de Río Cuarto; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina
Fil: Bhowmik, S.. Nexant; Estados Unidos
Fil: Magnago, Fernando. Universidad Nacional de Río Cuarto; Argentina. Nexant; Estados Unidos. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina