WIND TURBINE GEARBOX FAULT DETECTION USING MULTIPLE SENSORS WITH FEATURE LEVEL DATA FUSION

Fault detection in complex mechanical systems such as wind turbine gearboxes remains challenging, even with the recently significant advancement of sensing and signal processing technologies. For example, the non-stationary nature of the wind load may require the joint time-frequency domain feature extraction methods for the signals collected from the gearbox. In this paper, a harmonic wavelet based method is adopted, and a speed profile masking technique is developed to account for tachometer readings and gear meshing relationship. In such a way, those features with fault-related physical meanings can be highlighted. While multiple sensors yield redundant features, we fuse them through a statistical weighting approach based on principal component analysis. The fused data are fed to a simple decision making algorithm to verify the effectiveness. Using experimental data collected from a gearbox testbed emulating wind turbine operation, we can detect gear faults statistically for a given confidence level.