Induction machine fault detection using smartphone recorded audible noise

Toomas Vaimann*, Jan Sobra, Anouar Belahcen, Anton Rassõlkin, Michal Rolak, Ants Kallaste

*Corresponding author for this work

Research output: Contribution to journalArticle

16 Citations (Scopus)
200 Downloads (Pure)

Abstract

This study presents induction machine fault detection possibilities using smartphone recorded audible noise. Acoustic and audible noise analysis for fault detection is a well-established technique; however, specialised equipment for diagnostic purposes is often very expensive and difficult to operate. To overcome this obstacle, a simple pre-diagnostic procedure, using hand-held smartphones is proposed. Different faults of the three-phase squirrel cage induction machine such as various numbers of broken rotor bars and dynamic rotor eccentricity are inflicted to the machine and the resulting audible signals are recorded in laboratory circumstances using two widely available commercial smartphones. The analysis is performed on audible noise and compared with the results of mechanical vibrations measurements, recorded by vibration sensors. Rotational speed frequency and twice-line frequency are used as diagnostic indicators of faults. A simple neural network is composed and probabilities of fault detection using such diagnostic measures are presented. The necessity for further study as well as further implementation and method refinement necessity is pointed out.

Original languageEnglish
Pages (from-to)554-560
Number of pages7
JournalIET Science, Measurement and Technology
Volume12
Issue number4
DOIs
Publication statusPublished - 1 Jul 2018
MoE publication typeA1 Journal article-refereed

Keywords

  • frequency measurement
  • smart phones
  • probability
  • velocity measurement
  • fault diagnosis
  • computerised instrumentation
  • acoustic noise measurement
  • neural nets
  • rotors
  • squirrel cage motors

Fingerprint Dive into the research topics of 'Induction machine fault detection using smartphone recorded audible noise'. Together they form a unique fingerprint.

Cite this