Permanent magnet synchronous generator stator current AM-FM model and joint signature analysis for planetary gearbox fault diagnosis

Abstract

Wind energy is renewable and wind power generation facilities have been growing rapidly around the world. Permanent magnet synchronous generators (PMSGs) and planetary gearboxes are commonly used in wind turbines (WTs). For planetary gearbox fault diagnosis, PMSG stator current signal analysis provides a potential alternative approach because stator current signals contain fault signatures with easier accessibility comparing to conventional vibration signals. Planetary gearbox faults result in PMSG input torque oscillations, leading to both amplitude modulation and frequency modulation (AM-FM) effects on stator current signals. To better reveal gear fault features in PMSG stator current signals, this paper proposes to diagnose planetary gearbox faults from three complementary analysis results of the PMSG stator current, i.e., Fourier spectrum, amplitude demodulated spectrum and frequency demodulated spectrum. For this purpose, an AM-FM current signal model is derived through mechanical-magnetic-electric interaction analysis. Explicit equation of Fourier spectrum is derived with its sideband characteristics investigated. Both amplitude and frequency demodulation analyses are proposed to compensate for or mitigate the difficulty in the intricate sideband analysis of Fourier spectrum. Due to its capability in signal mono-component decomposition for accurate estimation of instantaneous frequency, adaptive local iterative filtering is utilized in frequency demodulation analysis. Explicit equations of corresponding demodulated spectra are derived, and gear fault signatures in PMSG stator current signals are revealed. The theoretical derivations and proposed methods are validated through lab experiments by diagnosing localized faults on the sun, planet and ring gears.