Wind Turbine Bearing fault detected with IAS combined with Harmonic Product Spectrum

André Hugo (1) , Khelf Ilyes (2) , Quentin Leclere (3, 4)

1 LASPI - Laboratoire d'Analyse des Signaux et des Processus Industriels
2 DCS - Dynamique et Contrôle des Structures
3 CeLyA - centre Lyonnais d'Acoustique
4 LVA - Laboratoire Vibrations Acoustique

A few years ago, Instantaneous Angular Speed (IAS) signal analysis has been proven able to detect natural bearing faults. Moreover, it has been applied on a 2MW wind turbine shaft line and proven able to detect low speed shaft unbalance using an encoder located on the high speed shaft turbine generator. This paper will show that a generator bearing fault can also be monitored using an encoder located on the low speed shaft. This step forward is made difficult by the lack of energy in the bearing fault speed fluctuation amongst the multitude of noisy phenomena, an alternative technique is proposed in this paper to extract dry impacts components from IAS spectrum. Harmonic Product Spectrum has originally been proposed in the sixties to detect fundamental frequency in a noisy signal. Still in use in the domain of speech signal processing, this method eases a precise voice pitch tracking. This paper proposes a revision of HPS to IAS analysis (and more generally to vibration monitoring), and a peculiar adaptation to look for inner ring modulated defects. The efficiency of the technique will be finally shown on real measurements issued from a 2MW wind turbine generator bearing fault.

Mots clés: Bearing Fault; Rotating Machine Monitoring; Harmonic Product Spectrum; Instantaneous Angular Speed

Domaines: Acoustique [physics.class-ph]

André Hugo, Khelf Ilyes, Quentin Leclere. Wind Turbine Bearing fault detected with IAS combined with Harmonic Product Spectrum. COMADEM 2017, Jul 2017, Preston, United Kingdom. ⟨hal-01595841⟩