| Along with the progress of science and technology, the development of new type of industrialization, a large number of intelligent mechanical equipment are used in industrial production, and promote the development of economy and society. The rolling bearing as a key component of rotating machinery equipment, its running state is often a direct impact on the performance of. Therefore, to monitor and diagnose the running state has very important significance.When the rolling bearing failure, vibration signal with significant nonlinear, non-stationary contains a large number of fault information. And how to take effective method extracting vibration signal in prominent fault feature component has been a hot research topic in the field of fault diagnosis. Aimed at the problem extracting rolling bearing fault information precision low, the rolling bearing fault type recognition accuracy low, this paper put forward the method based on conditional local mean decomposition (CLMD) and VPMCD bearing fault intelligent diagnosis. The method can extract the fault characteristic component among the vibration signal of the rolling bearing effectively, and construct the fault feature vector. Finally, the fault type of rolling bearing can be recognized effectively.The main contents of this paper are as follows:(1)This paper does much research about the theory of local mean decomposition(LMD) and it application in the nonlinear, non-stationary signal processing. The application of Hilbert transform in signal analysis are studied. The CLMD method is proposed. This method is combined with digital image processing of frequency resolution method and LMD. It can solve the mode confusion phenomenon of LMD in non-continuous and nonlinear, non-stationary signal effectively which affect the recognition efficiency. The effect of this method had been verified by the simulation signal. In addition, the frequency resolution method reduces the computation of LMD in the decomposition of non-continuous nonlinear, non-stationary signal, and also reduces the decomposition time, improves the decomposition efficiency.(2)VPMCD is a kind of multivariate prediction model pattern recognition method, which is a good method to solve the problem of nonlinear classification. The method can be applied to divide the system into different categories, according to the characteristic parameters of different categories, which have a certain intrinsic relationship, and this intrinsic relationship has obvious difference between different categories.In the same category, the corresponding mathematical models can be obtained by training the samples of different classes. That is variable prediction model(VPM). Therefore, these VPM can be used to predict the parameters of the test samples. According to the discriminant function of the square of the prediction error of each characteristic parameter minimum, the type identification of the test samples is carried out. This method can avoid the BP neural network falling into the local optimum easily and the optimization process of support vector machine(SVM) effectively. This method also reduce the amount of computation and training time, improve the recognition efficiency.(3)Because the rolling bearing early weak fault information is difficult to extract, fault type recognition rate is low. This paper propose the fault diagnosis method for rolling bearing based on CLMD and VPMCD. The application result proves the feasibility and effectiveness of the method. This paper compare the method with other methods which are representative method in the fault diagnosis in recent years from signal processing and pattern recognition. The application results show that the method own the advantages of high processing rate and high precision. T has extensive application prospect in the field of rolling bearing fault diagnosis. The method provide an effective new way for the realization of the rotating machinery fault diagnosis. |
PDF Full Text Request