Research On Fault Diagnosis Method Of Wind Turbine Gearbox Bearing Based On Vibration Analysis

Due to the harsh and complicated working conditions,the safety and stability of wind turbines are becoming more prominent as well as the large-scale development of wind turbines.When the gearbox bearing failure causes the failure to stop,the shutdown time is the longest in the various parts.Therefore,the research on the fault diagnosis method of gearbox bearings can provide important theoretical guidance and technical guarantee for detecting bearing failure in time,avoiding serious accidents and reducing maintenance costs,and enhancing the stability of wind turbines.This paper is completed under the support of the national science and technology support program(2015BAA06B03)"Research and demonstration of the key technologies for intelligent operation and maintenance of large wind farms".Based on the vibration monitoring data of the wind turbine,two fault diagnosis algorithms for gearbox bearing based on vibration signals are proposed:(1)The fault diagnosis algorithm is based on variational mode decomposition(VMD)and singular value energy difference spectrum.First,the bearing vibration signal is processed and several intrinsic mode function components(IMF)are obtained by using variational mode decomposition method.Then use the kurtosis-correlation coefficient index to select the most sensitive IMF.Then the singular value energy difference spectrum of the sensitive IMF is calculated,and the effective singular values are selected for signal reconstruction.Finally,the fault characteristics are extracted and the fault diagnosis is carried out by analyzing the envelope spectrum of the reconstructed signal.(2)A new method of high dimensional space fault data processing,tensor decomposition,is proposed and applied to the fault diagnosis of gearbox rolling bearings.First,the IMF-SVD(Singular Value Decomposition,SVD)method is used to estimate the source number of the vibration signal.Then the observation matrix is formed by using vibration signal,and next form a three order tensor.Using tensor decomposition to obtain a source number of low-rank sub tensors,and then use the mode-1 and mode-2 vector space of the corresponding sub tensors to reconstruct source signals.At the end,the source signal was further analyzed by envelope demodulation,the fault characteristic frequencies are extracted,and the purpose of determining fault types is achieved.Under the MATLAB environment,this paper realizes the above algorithms programming.The vibration data are also analyzed.The results show that the proposed VMD-singular differential energy spectrum method and tensor decomposition method can accurately and effectively diagnose wind turbine gearbox bearing faults.Compared the results of traditional analysis methods,envelope spectrum analysis and independent component analysis(ICA),the two proposed methods have obvious advantages.