Model based current analysis of electrical machines to detect faults in planetary gearboxes

Vibration based diagnosis to detect damage in gearboxes has been studied to schedule maintenance and reduce possible capital losses from gearbox failures. However, such vibration based techniques are difficult to implement in planetary gearboxes due to the complex nature of measured vibration spectrum. Motor current signal analysis (MCSA) provides an alternative and non-intrusive way to detect mechanical faults through electrical signatures. However, literature still lacks reported investigation to monitor a planetary gearbox in an electro-mechanical drive-train using MCSA. A successful practical implementation of MCSA to planetary gearboxes is challenging because of the harmonics present in the measured current signal that arises from: 1) the structural properties of the electrical machines, and 2) the errors in manufacturing and assembly of large number of gears. This paper proposes a novel fault detection method that extends the resonance demodulation technique from vibration analysis to MCSA for detection of the gear faults where the measured stator current signal may contain high background noise. The capability of this approach is demonstrated through simulation. A nonlinear multi-DOF lumped parameter model is developed for an electro-mechanical drive-train consisting of a driving motor connected through a back-to-back planetary gearbox to a load generator. Afterwards, a seeded gear tooth fault is introduced in this electro-mechanical model to investigate its effect on the stator current. Simulation results result in a higher fault indicator under faulty gear conditions which demonstrate the feasibility of using resonance demodulation based MCSA for monitoring of mechanical faults such as those arising from gearboxes.