Broken rotor bar detection via four-band wavelet packet decomposition of motor current

Cekic Y., Eren L.

Electrical Engineering, vol.100, no.3, pp.1957-1962, 2018 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 100 Issue: 3
  • Publication Date: 2018
  • Doi Number: 10.1007/s00202-017-0674-4
  • Journal Name: Electrical Engineering
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.1957-1962
  • Keywords: Broken rotor bars, Induction motors, Wavelet packet decomposition, and Multi-band decomposition
  • Istanbul Gelisim University Affiliated: No


© 2017, Springer-Verlag GmbH Germany, part of Springer Nature.The induction motor current is nonstationary by nature, and time-scale analysis techniques such as wavelet packet decomposition (WPD) are more suitable for the analysis of the stator current for broken rotor bar detection. But, WPD is very costly in terms of the computational effort when half-band FIR filter banks are used in analysis. The implementation of four-band FIR filter banks in the analysis of a phase current is proposed here to reduce the computational cost. The use of four-band FIR filter banks with FPGA implementation would also provide higher levels of parallel processing capability resulting in further reduction in computational time required for detection of broken rotor bar faults. In WPD, it is also possible to tailor the frequency band size (resolution) so that one frequency band covers all motor fault induced frequencies due to rotor speed variations. Here, the rms value for broken rotor bar fault-related frequency band is compared with the baseline data to detect any broken rotor bar faults.