-
Maísa L. F. Santos
Programa de Pós Graduação de Engenharia Elétrica, Instituto Federal da Paraíba, João Pessoa-PB
-
Suzete E. N. Correia
Programa de Pós Graduação de Engenharia Elétrica, Instituto Federal da Paraíba, João Pessoa-PB
-
Cleumar S. Moreira
Programa de Pós Graduação de Engenharia Elétrica, Instituto Federal da Paraíba, João Pessoa-PB
-
Álvaro de M. Maciel
Programa de Pós Graduação de Engenharia Elétrica, Instituto Federal da Paraíba, João Pessoa-PB
-
Edgard L. L. Fabricio
Programa de Pós Graduação de Engenharia Elétrica, Instituto Federal da Paraíba, João Pessoa-PB
Keywords:
Power quality disturbances, Classification, Variational mode decomposition, Signal processing, Feature extraction
Abstract
This article proposes a computational approach applied to the classification of Electric Power Quality Disturbances (PQD), which are based on the Variational Mode Decomposition (VMD) and the Random Forest (RF) classifier algorithms. Firstly, nine parametric equations- based different signals were generated. Variational mode decomposition (VMD) has been used to decompose the signals into various intrinsic mode functions (IMFs). Seven statistical characteristics were extracted from the instantaneous amplitude (AI) of the decomposed IMFs. The RF model is then developed to classify PQD based on these characteristics. The proposed approach is validated by analytical results, where the classifier accuracy was 99,71%. It confirms that the proposed analysis is capable of providing necessary and accurate information for PQD.