Akademska digitalna zbirka SLovenije - logo
VSE knjižnice (vzajemna bibliografsko-kataložna baza podatkov COBIB.SI)
  • EEMD-based multiscale ICA method for slewing bearing fault detection and diagnosis
    Žvokelj, Matej, strojnik ; Zupan, Samo ; Prebil, Ivan
    A novel multivariate and multiscale statistical process monitoring method is proposed with the aim of detecting incipient failures in large slewing bearings, where subjective influence plays a minor ... role. The proposed method integrates the strengths of the Independent Component Analysis (ICA) multivariate monitoring approach with the benefits of Ensemble Empirical Mode Decomposition (EEMD), which adaptively decomposes signals into different time scales and can thus cope with multiscale system dynamics. The method, which was named EEMD-based multiscale ICA (EEMD-MSICA), not only enables bearing fault detection but also offers a mechanism of multivariate signal denoising and, in combination with the Envelope Analysis (EA), a diagnostic tool. The multiscale nature of the proposed approach makes the method convenient to cope with data which emanate from bearings in complex real-world rotating machinery and frequently represent the cumulative effect of many underlying phenomena occupying different regions in the time - frequency plane. The efficiency of the proposed method was tested on simulated as well as real vibration and Acoustic Emission (AE) signals obtained through conducting an accelerated run-to-failure lifetime experiment on a purpose-built laboratory slewing bearing test stand. The ability to detect and locate the early-stage rolling - sliding contact fatigue failure of the bearing indicates that AE and vibration signals carry sufficient information on the bearing condition and that the developed EEMD-MSICA method is able to effectively extract it, thereby representing a reliable bearing fault detection and diagnosis strategy.
    Vir: Journal of sound and vibration. - ISSN 0022-460X (Vol. 370, May 2016, str. 394-423)
    Vrsta gradiva - članek, sestavni del
    Leto - 2016
    Jezik - angleški
    COBISS.SI-ID - 15325723
    DOI

Vnos na polico