Research On Fault Diagnosis Method Of Wind Turbine Gearbox Based On ARMA-MOMEDA And SGMloG

Wind power generators play a vital role in the development of green energy,and the quality of their operation is directly related to the speed of national green energy development.Since the wind turbine is subjected to variable load impacts during operation,it causes great troubles to the efficient operation of its gearbox,so it is imperative to detect the fault state of its gearbox.However,in actual working conditions,the collected gearbox vibration signal contains a lot of environmental noise,which makes it difficult to accurately extract the fault feature information.In addition,in the actual detection process,the faults are often concurrent.If multiple faults cannot be accurately extracted at the same time,the long-term and efficient operation of the wind turbine cannot be guaranteed.In view of the above problems,this paper takes the rolling bearing of wind power gearbox as the research object,and studies the noise reduction of the fault signal of the bearing and the method of multi-fault feature extraction.In this paper,the research status of the fault feature extraction of wind turbines at home and abroad is elaborated,and the research content and ideas of this paper are brought out according to the current problems.Secondly,the main components of the wind turbine are introduced,and the structure of the wind turbine gearbox is described.The common gear failures and bearing failure types in gearboxes are classified,and the causes of failures are elaborated in detail.Considering that the working environment of wind turbines is relatively harsh,most of them are installed in high-pressure areas such as plateaus and Haikou.In complex mechanical environments,the vibration signals of wind turbines are often accompanied by a large amount of white noise,which makes it more difficult to extract fault features.The Multipoint Optimal Minimum Entropy Deconvolution Adjusted can extract periodic shock components,but misdiagnosis or missed diagnosis often occurs in a strong noise environment.Therefore,this paper proposes a method called the multipoint optimal minimum entropy deconvolution adjusted of the autoregressive moving average model(ARMA-MOMEDA),which is used to extract the fault features of wind power gearboxes.First,smooth the noise in the vibration signal by the moving average model to improve the signal-to-noise ratio of the signal;then,perform the second noise reduction on the smoothed signal by the multipoint optimal minimum entropy deconvolution adjusted;Finally,the fault features are extracted through the envelope spectrum,and simulation and experiment verify the effectiveness of the method.The deconvolution theory has shown good advantages in fault diagnosis,but it is necessary to extract the features of each fault when diagnosing a composite fault.In order to efficiently diagnose complex faults of wind power gearboxes,on the basis of Modified Laplacian of Gaussian(MloG),a compound fault diagnosis method for wind power gearboxes based on Savitzky-Golay(SG)smoothing and MloG is proposed.First,the vibration signal is processed by the MloG filter to extract the fault feature information,but the original MloG filter does not have adaptability.When performing fault diagnosis on the vibration signals of different components,the optimal filter parameters cannot be adaptively selected.In order to improve the performance of MloG,a chaotic gray wolf algorithm based on marginal envelope spectral entropy(MBLS)is proposed to adaptively determine the parameters of MloG filter.The MBLS has a good characterization ability for the number of pulses and the intensity change of noise.Finally,the extracted signal is smoothed by SG filter.The simulation and the actual measurement of the bearing vibration signal of the wind turbine are used to verify the effectiveness of the method.