Condition Monitoring and Fault Diagnosis for Wind Turbine Gearbox Based on Waterfall

Abstract

Threshold analysis is a common method in condition monitoring of wind turbine gearboxes. It means set alarm thresholds to monitor working state of gearboxes, such as RMS, kurtosis and peak-peak and so on. However, these methods influenced by noise cause false alarms easily. Waterfall has the advance in reflecting the relationship that frequency spectrum changes over time. Before getting frequency spectrum information of each part from the fast Fourier transform, the vibration data of the gearbox is split into several equal parts according to the time. Then frequency spectrum of each part is chronologically stacked into a waterfall. Through the analysis of practical gearbox vibration data, it can obviously differ the waterfall of different state, such as normal state, stop, noise and failure. Waterfall can directly and accurately monitor the working status of the wind turbine gearbox.