Yujiong Gu, Lei Song, Ting Xu, Luwei Su, and Guanyu Wu
Order resampling, Angle domain series index, Index-correlation model, Multivariate outlier detection
Under the complex and fluctuant operational condition, the fault feature extraction and fault warning index quantization become the critical technology for the fault warning of wind turbine gearbox. The randomness of wind speed and load always makes the operating staff difficult to estimate the healthy running state, causing the mistakes and delays of fault warning. The paper proposes a novel method based on turbine vibration data analysis to realize the fault warning of wind turbine gearbox and even increase the fault warning accuracy. The paper firstly utilizes the order resampling method to transfer the non-stationary time-domain vibration signal into stationary angle-domain vibration signal, and then extracts dimensionless index of angle domain series to reflect the running trend of wind turbine gearbox qualitatively. Different index-correlation models should be established to recognize different wind turbine gearbox faults. Finally, based on the multivariate outlier detection of angle domain series index, the paper realizes the fault warning of wind turbine gearbox quantitatively. The paper also utilizes the gearbox fault simulation test-rig to verify the method. The result indicates the fault warning method owns validity, avoiding the serious accidents of wind turbine gearbox.
Important Links:
Go Back
