Bearings fault detection in helicopters using frequency readjustment and cyclostationary analysis - Archive ouverte HAL Access content directly
Journal Articles Mechanical Systems and Signal Processing Year : 2013

Bearings fault detection in helicopters using frequency readjustment and cyclostationary analysis


The objective of this paper is to propose a vibration-based automated framework dealing with local faults occurring on bearings in the transmission of a helicopter. The knowledge of the shaft speed and kinematic computation provide theoretical frequencies that reveal deteriorations on the inner and outer races, on the rolling elements or on the cage. In practice, the theoretical frequencies of bearing faults may be shifted. They may also be masked by parasitical frequencies because the numerous noisy vibrations and the complexity of the transmission mechanics make the signal spectrum very profuse. Consequently, detection methods based on the monitoring of the theoretical frequencies may lead to wrong decisions. In order to deal with this drawback, we propose to readjust the fault frequencies from the theoretical frequencies using the redundancy introduced by the harmonics. The proposed method provides the confidence index of the readjusted frequency. Minor variations in shaft speed may induce random jitters. The change of the contact surface or of the transmission path brings also a random component in amplitude and phase. These random components in the signal destroy spectral localization of frequencies and thus hide the fault occurrence in the spectrum. Under the hypothesis that these random signals can be modeled as cyclostationary signals, the envelope spectrum can reveal that hidden patterns. In order to provide an indicator estimating fault severity, statistics are proposed. They make the hypothesis that the harmonics at the readjusted frequency are corrupted with an additive normally distributed noise. In this case, the statistics computed from the spectra are chi-square distributed and a signal-to-noise indicator is proposed. The algorithms are then tested with data from two test benches and from flight conditions. The bearing type and the radial load are the main differences between the experiences on the benches. The fault is mainly visible in the spectrum for the radially constrained bearing and only visible in the envelope spectrum for the "load-free" bearing. Concerning results in flight conditions, frequency readjustment demonstrates good performances when applied on the spectrum, showing that a fully automated bearing decision procedure is applicable for operational helicopter monitoring.
Fichier principal
Vignette du fichier
Publication2011_V24.pdf (444.76 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-00801483 , version 1 (13-05-2013)



Victor Girondin, Komi Midzodzi Pekpe, Jean Philippe Cassar, Hervé Morel. Bearings fault detection in helicopters using frequency readjustment and cyclostationary analysis. Mechanical Systems and Signal Processing, 2013, NA (NA), pp.NA. ⟨10.1016/j.ymssp.2013.03.015⟩. ⟨hal-00801483⟩
257 View
729 Download



Gmail Facebook Twitter LinkedIn More