Vector Envelope Spectrum Estimator: A Source Double-Channel Feature Extraction Method for Planetary Gearbox Fault Detection

Traditional feature extraction methods are mostly based on single-channel signals for fault diagnosis. However, in practical engineering, there are differences in source signals emitted by a planetary gearbox in different directions across the cross section, so these methods are employed to easily cause missing information. This article proposes a source double-channel feature extraction method named vector envelope spectrum estimator (VESE) for detecting planetary gearbox faults. First, the source double-channel signals are collected in the same cross section on the gearbox using the on-rotor sensing (ORS) technique. Subsequently, the VESE is developed after the source double-channel model construction to realize double-channel signal feature extraction. Finally, the fault signature information components are recognized in the VESE, which enables enhanced detection of planetary gearbox faults. The effectiveness of the VESE is verified using planetary gearbox experiment data. Compared with the Teager energy operator (TEO) and Hilbert envelope analysis (HEA), the VESE provides more accurate results for detecting planetary gearbox faults.