Research On Fault Diagnosis Methods For Rotor Systems Based On Local Mean Decomposition

The process of rotor systems fault diagnosis includes the acquisition of information and extracting feature and recognizing conditions of which feature extraction is the priority. A novel method of time-frequency analysis, Local Mean Decomposition (LMD) is applied to the rotor system fault diagnosis. Through this method we can acquire a number of product functions (PFs) components whose instantaneous frequencies own physical meaning. Some research on rotor systems diagnosis based on LMD has been done. The main work is done as follows:1. A rotor systems fault diagnosis method based on LMD and neural network is proposed to deal with the non-stationary vibration signal from fault rotor systems. First of all, LMD method is applied to decompose the original signals into a finite number of Product functions (PFs), then several PFs containing main fault information are selected for further analysis; subsequently, energy and time domain statistics feature parameters extracted from PFs could be served as input parameters of neural networks to identify fault patterns of rotor systems. The experimental results show that the rotor systems fault diagnosis method based on LMD and neural network can identify rotor systems fault patterns accurately and effectively.2. A fault diagnosis approach for rotor systems based on LMD and AR model is proposed. Firstly, by using LMD method, the vibration signal of rotor systems is decomposed into a number of PF components whose instantaneous frequencies own physical meaning, and then the AR model of each PF component is established. Furthermore, the model parameters and the variance of remnant are regarded as the fault feature and served as input parameter of neural networks to identify the condition and fault pattern of a rotor system. The study results show that both Empirical Mode Decomposition (EMD) and LMD method could be applied to the rotor systems fault diagnosis effectively. However, the latter has better decomposition results.3. A fault diagnosis approach for rotor systems based on the combination of improved LMD and singular value decomposition is proposed for the frequency confusion of LMD. Firstly, we adopt the wavelet technique to decompose the vibration signal of rotor systems into several wavelet components, then each of wavelet component is decomposed by the LMD method and a number of PF components can be acquired from which the initial feature vector matrix is formed. By applying the singular value decomposition technique to the initial feature vector matrix, take the decomposed singular value as the fault characteristic vector and input it into the network to identify the work condition and the fault patterns of the rotor systems. The experimental results show that the proposed approach can be applied to the rotor systems fault diagnosis effectively.4. A fault diagnosis approach for rotor systems based on the combination of improved LMD and time-frequency entropy is proposed. After the decomposition of wavelet-LMD, the energy of rotor systems vibration signal located in different time scale of the PF components.Under various conditions, the energy of rotor systems in different PF component is inconsistent which can be shown through the different of the time-frequency distribution, the time-frequency entropy based on improved LMD is the quantitative description of the time- frequency distribution. Experimental data analysis shows that the time-frequency entropy based on the improved LMD is very sensitive to the type of rotor fault can be used for fault diagnosis of rotor system.