Using Cepstrum Information to Detect Gearbox Faults in Wind Turbine Planetary Stage Gearbox

摘要

Detecting gear related faults in the planetary stage of a wind turbine gearbox is traditionally challenging. This work shows that cepstrum can help identify families of sidebands related to gear faults in the planetary stage of a wind turbine gearbox. Quefrency peaks (representation of those sidebands families) identified can also be trended over time, and early stage fault detection can be achieved. It is imperative to know the kinematical data of a specific gearbox, such as gear mesh frequency and shaft rotational speed, in order to trend the quefrency peaks. This work compiles cepstrum signatures from hundreds of turbines with known and unknown faults present. The results shows that turbines with similar gearbox ratios have common increasing quefrency peaks related to certain faults. Those peaks normally fall within the same limited quefrency ranges. Information regarding these quefrency ranges can then be used for trending the quefrency magnitudes in those ranges for faults detection in many turbines with similar gearbox ratios. In this way, it is not necessary to know the details of gear counting for each turbine.