Title :
A Novel Feature Vector Consisting of AR-MA Model Parameters for Motor Fault Classification
Author :
Ece, Dogan Gokhan ; Gerek, Ömer Nezih
Author_Institution :
Muhendislik Mimarlik Fak., Anadolu Univ., Eskisehir
Abstract :
In this work, AR-MA parameters corresponding to zeros and poles of the z-plane model are extracted from the sampled current -waveform spectra, and considered as elements of a feature vector to classify faulty motors. The classification was observed to be possible due to the differences between these parameters according to the type of the motor faults.
Keywords :
autoregressive moving average processes; fault location; feature extraction; poles and zeros; AR-MA model parameters; feature vector; motor fault classification; zeros and poles; Poles and zeros; Rotors; Stators; Örüntü siniflandirma; AR-MA model; Motor arizalari;
Conference_Titel :
Signal Processing and Communications Applications, 2006 IEEE 14th
Conference_Location :
Antalya
Print_ISBN :
1-4244-0238-7
DOI :
10.1109/SIU.2006.1659844
