Research On Fault Diagnosis Method Of Helicopter Swashplate Rolling Bearing Based On Information Entropy

Rolling bearing is the key component of helicopter swashplate,once it occurs abnormal situation will affect normal flight of helicopter,therefore,it is of great importance to study the fault diagnosis method of rolling bearing in order to ensure the safety and reliability of helicopter.Based on the aviation science fund,the vibration data collected by a certain type swashplate rolling bearing fault simulation test which carried out by a bear factory is taken as the object of study,the research is conducted on the fault diagnosis method of swashplate rolling bearing by combination of adaptive signal decomposition method and two kinds of common information entropy.The main research contents and results are as follows:(1)The paper introduces the related theoretical basis and preprocessing method.Firstly,aiming at the difference between helicopter bearing and traditional bearing,the principle of two kinds of adaptive signal decomposition methods are introduced.Then,the vibration data are pretreated by morphological filtering,singular value decomposition filtering,wavelet filtering and maximum correlation kurtosis deconvolution respectively.In the end,through the comparison of the root mean square error and signal to noise ratio experiment,results show the singular value decomposition filtering method is the best.(2)The fault diagnosis of helicopter rolling bearing is carried out by VMD-SE and DE-ELM.Firstly,the vibration data of normal state,inner race fault,outer race fault and ball fault are decomposed by VMD.Then,the sample entropy of the first four components are extracted to constitute feature vectors.In the end,the training sets are input into DE-ELM to construct the training model,then the test sets are input into model to carry out fault diagnosis.A large number of fault diagnosis experiments are carried out on the vibration data under different cut and different load,the advantage of DE-ELM is proved under the condition of less hidden layer node,at the same time,the validity and adaptability of this method are verified.(3)Aiming at the shortcomings of CEEMDAN method,an improved CEEMDAN method is proposed for fault diagnosis of swashplate rolling bearing.Firstly,the vibration data of normal state,inner race fault,outer race fault and ball fault are decomposed by improved CEEMDAN.Then,the permutation entropy of the first four components are extracted to constitute feature vectors.In the end,the training sets are input into SVM to construct the training model,then the test sets are input into model soas to achieve the purpose of fault diagnosis.A large number of fault diagnosis experiments with the vibration data collected from different cut and different load are carried out,the experimental results show that the improved method can improved the fault recognition rate effectively,the effect is more obvious in the case of small samples.