Research On Early Fault Diagnosis And Degradation State Identification Methods Of Planetary Gear

Large complex electromechanical equipment is an important basis for the development of China's manufacturing industry and national defense equipment,and it is also one of the important cornerstones to support the national comprehensive national strength.With the rapid development of modern industrial technology,the Large-scale electromechanical equipment tends to be complex,precise,efficient and intelligent.However,the transmission system of large electromechanical equipment often fails,the planetary gear is an important part of the mechanical transmission system,and it is one of the important failure sources.Therefore,it is of great theoretical and practical significance to study the early fault diagnosis of planetary gear and the identification of its degraded state.Planetary gear transmission is a typical complex system,and its vibration response is more complicated than the traditional fixed axis gear transmission,which cause the early fault diagnosis and degradation state recognition process has its own characteristics and difficulties.There are many problems in the early fault diagnosis and degradation state recognition,such as the complexity of the frequency components,the weak fault response,the difficulty of feature extraction and so on.Based on this,the planetary gear is taken as the research object in this paper,and the method of early fault diagnosis and degradation state recognition is deeply studied,the main research contents include:(1)Firstly,the basic principle and properties of empirical mode decomposition(EMD)method are introduced.In view of the problem of mode mixing and end effect in EMD,the principle of the improved algorithm ensemble empirical mode decomposition(EEMD)and its parameters selection principle are introduced.To some extent,EEMD can overcome the problem of mode mixing in EMD method,but it has a large reconstruction error.For this reason,the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)method is further introduced.This method inherits the good reconstruction of EMD and overcomes the problem of modal aliasing.By comparing the three methods,it is found that CEEMDAN is more suitable for nonlinear and non-stationary signal processing.(2)Aiming at the problem that the early fault feature of the planetary gear is relatively weak and difficult to find,two new fault feature extraction methods based on permutation entropy and box dimension are proposed.Then,the theory of support vector machine(SVM)is studied,and it is found that SVM is especially suitable for dealing with small sample nonlinear problems.But its parameters have a great influence on the identification results,and are not easy to determine.It is gratifying that genetic algorithm can solve the difficult problem of SVM parameter selection.Therefore,the SVM parameter optimization model based on genetic algorithm is constructed.Finally,based on the advantages of the above algorithms,the early fault diagnosis of planetary gear can be realized.The experimental results show that the proposed methods can effectively identify the early faults of planetary gear.(3)In order to realize the effective identification of different degradation processes of planetary gear,a recognition method based on ICEEMDAN and linear local tangent space alignment(LLTSA)is proposed.Although CEEMDAN can deal with nonlinear and non-stationary signals very well,its decomposition results still has some noise and contain false components.Taking into account the degradation state recognition is different from fault diagnosis,so CEEMDAN is improved further.And according to the decomposition results,the detail features is extracted.Then the original high dimensional feature sets is constructed by the time domain features and the detail features.In order to solve the problem of information redundancy and interference characteristics,LLTSA is used to reduce the dimension and the feature fusion is realized.Further,the effective identification of different degradation states of planetary gear can be realized.(4)It is obvious that the proposed fault identification method has some problems such as inconvenient processing and poor human-computer interaction in the practical application,so a software system for the early fault diagnosis and the identification of degradation states of planetary gear is developed based on LabVIEW and MATLAB.It realizes the functions of real-time acquisition and preservation,basic analysis,early fault diagnosis,degradation state recognition,storage and query of diagnosis results and has friendly interface and convenient operation.At the end of this thesis,the summarizations of the research and expectation of the related technology development are presented.