A new wind turbine fault diagnosis method based on ensemble intrinsic time-scale decomposition and WPT-fractal dimension

In this paper an ensemble intrinsic time-scale decomposition (EITD) method based on the cubic spline interpolation and linear transformation of intrinsic time-scale decomposition (ITD) was proposed, which can restrain the end effect and avoid the signal distortion. Combining ensemble intrinsic time-scale decomposition (EITD) with wavelet packet transform (WPT) and correlation dimension (CD), a novel method for decomposing nonstationary vibration signal and diagnosing wind turbine faults is presented. In implementation of the method, wavelet packet transform is employed to denoise raw vibration signals. Some important influencing factors relating directly to the computational precision of correlation dimension are discussed. The advantage of combining EITD and fractal dimension is that it does recognize the wind turbine gearbox fault types, and can solve the difficulty of recognizing fault conditions when two or more fractal dimensions are close to each other. To verify the effectiveness of the EITD-WPT-CD in detecting the faults, their induced vibrations are collected from high speed shaft gear under normal and faulty conditions through acceleration measurement. The results show that this method is capable of extracting the signal features and identifying the working conditions. The fault diagnosis application in a wind turbine gearbox indicates that the proposed method improved the accuracy of fault diagnosis. (c) 2015 Elsevier Ltd. All rights reserved.