DocumentCode :
1381625
Title :
Diagnosis and performance analysis of threephase permanent magnet synchronous motors with static, dynamic and mixed eccentricity
Author :
Ebrahimi, Bashir Mahdi ; Faiz, Jawad
Author_Institution :
Center of Excellence on Appl. Electromagn. Syst., Univ. of Tehran, Tehran, Iran
Volume :
4
Issue :
1
fYear :
2010
fDate :
1/1/2010 12:00:00 AM
Firstpage :
53
Lastpage :
66
Abstract :
Albeit permanent magnet synchronous motors (PMSMs) under static and dynamic eccentricity have been studied in few papers, so far mixed eccentricity has not been investigated. In this study, a novel index is introduced for dynamic and mixed eccentricity fault diagnosis in PMSM. This index is the amplitude of side-band components with a particular frequency pattern that is extracted from spectrum analysis of the motor torque. The index provides precise detection of the eccentricity occurrence, recognition of its type and determination of its degree. To evaluate the ability of the proposed index for the eccentricity diagnosis and estimation of its severity, first the correlation between indices, static and dynamic eccentricity degrees are calculated. The type of eccentricity is determined using k-nearest neighbour (KNN) classifier. At the next step, a three-layer artificial neural network is employed to estimate the eccentricity degree of torque profiles based on their type of eccentricity. After all, a white Gaussian noise is added to the simulated torque and robustness of the proposed index is analysed with respect to the noise variance. Meanwhile, the spectrum analysis of the stator current in faulty PMSM under static, dynamic and mixed eccentricities is studied and the influence of these eccentricities on ohmic, hysteresis and eddy current losses of PMSM are presented. In this examination, the time-domain elemental flux density waveforms and various time-domain waveforms of motor winding currents are computed using time stepping finite element method (TSFEM) for core loss and ohmic loss calculations punctually. The accuracy of obtained simulation results is verified using experimental results.
Keywords :
Gaussian noise; cores; fault diagnosis; finite element analysis; losses; machine theory; machine windings; neural nets; pattern classification; permanent magnet motors; power engineering computing; synchronous motors; time-domain analysis; torque; artificial neural network; core loss; dynamic eccentricity; fault diagnosis; k-nearest neighbour classifier; mixed eccentricity; motor torque spectrum analysis; motor winding current; ohmic loss; static eccentricity; three phase permanent magnet synchronous motors; time domain elemental flux density waveform; time stepping finite element method; time-domain waveform; white Gaussian noise;
fLanguage :
English
Journal_Title :
Electric Power Applications, IET
Publisher :
iet
ISSN :
1751-8660
Type :
jour
DOI :
10.1049/iet-epa.2008.0308
Filename :
5382472
Link To Document :
بازگشت