A Comparative Study of the Effectiveness of Adaptive Filter Algorithms, Spectral Kurtosis and Linear Prediction in Detection of a Naturally Degraded Bearing in a Gearbox

Fairs Elasha, C. Ruiz-Carcel, D. Mba, P. Chandra

Research output: Contribution to journalArticle

15 Citations (Scopus)
106 Downloads (Pure)


Diagnosing bearing faults at the earliest stages is critical in avoiding future catastrophic failures. Many techniques have been developed and applied in diagnosing bearings faults; however, these traditional diagnostic techniques are not always successful when the bearing fault occurs in gearboxes where the vibration response is complex; under such circumstances, it may be necessary to separate the bearing signal from the complex signal. In this paper, an adaptive filter has been applied for the purpose of bearing signal separation. Four algorithms were compared to assess their effectiveness in diagnosing a bearing defect in a gearbox, least mean square (LMS), linear prediction, spectral kurtosis and fast block LMS. These algorithms were applied to decompose the measured vibration signal into deterministic and random parts with the latter containing the bearing signal. These techniques were applied to identify a bearing fault in a gearbox employed for an aircraft control system for which endurance tests were performed. The results show that the LMS algorithm is capable of detecting the bearing fault earlier in comparison with the other algorithms.
Original languageEnglish
Pages (from-to)623-636
JournalJournal of Failure Analysis and Prevention
Issue number5
Early online date15 Aug 2014
Publication statusPublished - Oct 2014



  • Vibration
  • Adaptive filter
  • Signal separation
  • Bearing diagnostics
  • Gearbox

Cite this