Rolling Bearing Fault Diagnosis Based On VMD And Multi-scale Entropy

As an important part of rotating machinery,rolling bearing is also one of the easily damaged mechanical parts because it often works in complex working conditions and harsh environments.Once the fault occurs,it will bring significant safety risks and economic losses.Therefore,it is of great practical significance to study the fault diagnosis method of rolling bearing.This paper on the subject of the study elaborates the background and significance,then introduces the related knowledge,for rolling bearing fault diagnosis includes bearing state of common failure forms,signal acquisition modes and the basic steps of fault diagnosis,then the fault diagnosis of the common extraction methods and future development trend,and some research status at home and abroad.Next,a new time-frequency analysis method VMD is proposed to solve the defects of time-frequency signal decomposition,such as EMD and EEMD,and the difference between the two time-frequency analysis methods(EEMD and VMD)is compared through simulation.In addition,the importance of selecting the value of the modal parameter K in the VMD method is illustrated.In addition,the modal parameter K is determined by decomposing the actual fault bearing signal according to the central frequency similarity criterion.The decomposition of the actual bearing fault signal by VMD better illustrates that the fault feature rate feature can be found more accurately by VMD.The third chapter introduces the basic principle of permutation entropy and multi-scale permutation entropy.It shows that multi-scale permutation entropy is the product of the combination of multi-scale theory and information entropy theory,which can not only dig the essential characteristics of fault signals in a deep level,but also realize the quantitative description of fault characteristics.At the same time,the definition of multi-scale entropy algorithm is better understood by constructing simulation signal to calculate its multi-scale entropy.Some key parameters of the entropy algorithm are determined.Then the process of data acquisition and data processing is introduced in detail.Finally,through the combination of VMD and the multi-scale entropy algorithm,the different types of bearing failure signals collected by the VMD decomposition laboratory are analyzed,and the multi-scale entropy values of each modal component are calculated respectively.Finally,the eigenvector is constructed based on vmd-mpe.The last part is the process of fault identification of rolling bearing.Firstly,the process of fault identification of rolling bearing is introduced more intuitively with flow chart.Secondly,the basic principle of LSSVM is introduced.Finally will feature vector is input into the trained LSSVM,results show that the multiscale permutation entropy is capable of detecting signal randomicity and dynamic mutation behavior,and with the fault diagnosis method,the combination of VMD on rolling bearing signal analysis can show that the method of normal state and failure of inner ring,outer ring and rolling element failure to identify effectively.The feasibility and effectiveness of the method based on vmd-mpe-lssvm are illustrated.