Multiple signal classification based on automatic order selection method for broken rotor bar detection in induction motors

Gerardo Trejo-Caballero, Horacio Rostro-Gonzalez, Rene de Jesus Romero-Troncoso, Carlos Hugo Garcia-Capulin, Oscar Gerardo Ibarra-Manzano, Juan Gabriel Avina-Cervantes, Arturo Garcia-Perez*

*Autor corresponent d’aquest treball

Producció científica: Article en revista indexadaArticleAvaluat per experts

9 Citacions (Web of Science)

Resum

Multiple signal classification (MUSIC) algorithm has been widely used to obtain high-resolution frequency estimation for an accurate identification of frequency components in low signal-to-noise ratios. One of the main drawbacks associated with the use of the MUSIC algorithm is that its performance is fully deteriorated when a wrong frequency signal dimension order is chosen, producing that some spurious frequencies could appear or some signal frequencies could be missing. In this paper, it is proposed a multi-objective optimization method to address the frequency signal dimension order problem. The proposed approach is based on a novel feature extraction of frequency components, which allows determining an adequate frequency signal dimension order. The methodology has been integrated as part of the MUSIC algorithm, and it can find the optimal order within a predefined frequency bandwidth, where the user is interested to find a frequency component. To evaluate the effectiveness of the proposed methodology, experimental results from several current signals obtained in the detection of broken rotor bar fault in induction motors have been tested.

Idioma originalAnglès
Pàgines (de-a)987-996
Nombre de pàgines10
RevistaElectrical Engineering
Volum99
Número3
DOIs
Estat de la publicacióPublicada - 1 de set. 2017
Publicat externament

Fingerprint

Navegar pels temes de recerca de 'Multiple signal classification based on automatic order selection method for broken rotor bar detection in induction motors'. Junts formen un fingerprint únic.

Com citar-ho