A novel high-resolution technique for induction machine broken bar detection

Shuo Chen, Rastko Zivanovic

Research output: Chapter in Book/Report/Conference proceedingsConference contributionpeer-review

9 Citations (Scopus)

Abstract

On-line and non-invasive condition monitoring and fault diagnostics of heavy duty induction machines, for example those used in mining and oil industries, can be accomplished by estimating the frequency components of the stator current. However, due to the inherent drawbacks of the conventional tools, such as Fast Fourier Transform (FFT), and the changing of motor stator current in the frequency domain along with load variations, optimal trade-off between leakage suppression, resolution and tracking is difficult to fulfil. This paper proposes a novel high-resolution technique for detection of rotor broken bars in induction motors. It is able to outperform FFT in the sense of achieving high frequency resolution while using data samples from a very short window. In order to investigate the impact of broken bars and generate representative signals to test the spectral estimation method, a mathematical model that is able to simulate induction machines with broken bars was developed. In the final sections of this paper we present selection of spectral estimation results obtained by analysing simulated signals.

Original languageEnglish
Title of host publication2007 Australasian Universities Power Engineering, AUPEC
DOIs
Publication statusPublished - 2007
Externally publishedYes
Event2007 Australasian Universities Power Engineering, AUPEC - Perth, WA, Australia
Duration: 9 Dec 200712 Dec 2007

Publication series

Name2007 Australasian Universities Power Engineering Conference, AUPEC

Conference

Conference2007 Australasian Universities Power Engineering, AUPEC
CountryAustralia
CityPerth, WA
Period09.12.200712.12.2007

Keywords

  • Broken bar
  • Fault detection
  • Induction machines
  • Prony analysis

Fingerprint Dive into the research topics of 'A novel high-resolution technique for induction machine broken bar detection'. Together they form a unique fingerprint.

Cite this