Research On Fault Diagnosis Technology Of Wind Turbine Transmission System Bearings Based On Vibration Signal Analysis

With the goal of "carbon peaking and carbon neutrality",the development of the new energy industry has once again attracted the attention of all sectors of society.As an important part of the new energy industry,wind power generation,the installed capacity of wind turbines will continue to grow steadily.As an important part of the new energy industry,wind power generation,the operating status and output efficiency of wind turbines directly affect the utilization of wind power energy.power quality and output efficiency.Aiming at the problems of strong background noise of vibration signal of wind turbine transmission system,easy mutation of fault signal,difficulty in fault feature extraction,and difficulty in fault identification.A vibration signal noise reduction algorithm based on adaptive white noise complete empirical mode decomposition combined with the correlation coefficient-kurtosis value criterion is proposed,and then the time-shifted multi-scale permutation entropy combined with the kernel principal component analysis method is used for feature extraction.Finally,a fault diagnosis model based on the improved artificial fish swarm algorithm to optimize the support vector machine is proposed to realize the fault diagnosis.The specific research contents are as follows:(1)Aiming at the problem that the background noise of the fault vibration signal is strong and the fault signal is weak,this paper proposes a noise reduction method for the fault signal based on the CEEMDAN algorithm combined with the correlation coefficient and the kurtosis value.Firstly,the original fault signal is decomposed by CEEMDAN,and a series of modal components from high frequency to low frequency are obtained,the selected modal points are reconstructed to achieve signal noise reduction on the premise of retaining fault characteristics.(2)Aiming at the problem that each fault characteristic signal of the bearing is prone to mutation,it is proposed to take TSMPE as the characteristic vector,and then combine the KPCA algorithm to realize the fault characteristic extraction.Firstly,the noise-reduced fault signal is calculated by TSMPE,and the TSMPE value is used as the eigenvalue of each fault state.During the research process,it is found that the eigenvalues of each fault state have cross-aliasing phenomenon with the increase of time scale.To reduce the adverse effect of TSMPE eigenvectors in fault diagnosis and classification,fault features are obtained by fusing TSMPE eigenvalues through KPCA.(3)The SVM fault diagnosis model is optimized by the IAFSA algorithm to realize the fault diagnosis of the transmission system bearing.First,based on the traditional artificial fish swarm algorithm,a decaying exponential function is introduced to replace the fixed field of view and step size parameters,so that it has a faster optimization speed and higher optimization accuracy.Secondly,the IAFSA algorithm is used to optimize the SVM parameters,so that the fault classifier has better learning ability and better diagnosis performance.Finally,the fault diagnosis algorithm proposed in this paper is verified by reference to the measured data of the wind farm,and the data of the main bearing of the wind turbine rotating system and the high-speed bearing of the gearbox are used to verify the data in different fault states,of which the fault diagnosis accuracy of the main bearing is93.3%,and the fault diagnosis accuracy of the gearbox high-speed bearing is 95%.The fault diagnosis results show that the proposed method in this paper has achieved good results in the fault diagnosis of the bearing of the wind turbine transmission system and has certain reference value.Figure 51,table 15,reference article 63.