Application And Experiment Research Of WVD Based On Improved LCD Method In The Fault Diagnosis

Rotating machinery fault diagnosis is able to provide necessary support for the safety of the equipment. The time-frequency analysis of Fault vibration signal is a very important way of fault diagnosis. The crux of fault diagnosis is the feature extraction of vibration signals. Among the many feature extraction methods, the research of adaptive time-frequency analysis is attracting more attention. Local Characteristic-scale Decomposition(LCD) is a new time-frequency method, and its advantages of computational efficiency and accuracy of the decomposition are obvious. Wigner-Ville distribution(WVD) is widely used in the filed of signal processing due to its excellent characteristics: high energy accumulation, high time frequency resolution and good time varying characteristics. However, the components of the decomposition process does not satisfy additivity and will produce cross terms, which effects the engineering application of WVD. To solve this problem, a cross-terms restraining method based on improved LCD was put forward.Firstly, the fault diagnosis of the rotating equipment and frequency characteristics were researched in this text, and the test equipment was used to simulate it, which laid the foundation for the analysis of the fault singal. Moreover, the LCD method was studied in-depth. In order to prove its superiority, the method of Empirical Mode Decomposition(EMD) and Intrinsic time-scale decomposition(ITD) were compared with LCD separately. Due to the phenomenon of signal distortion and the energy leakage in the process of decomposition when using the method of LCD, a new method was proposed to improve it which based on Cubic B-Spline Interpolation(CBI). By applying the modified LCD method on the singal simulation and the analysis of the vibration singals, the avaliability of it was proved.Additionally, the basic theory and properties of WVD distribution was studied in this paper, and the causes of the cross terms was analysed detailed. Researchers have proved that there was no cross terms of WVD distribution did not exist. Therefore, starting from the idea of inhibition of cross terms, the method of WVD based on improved the LCD was proposed in this paper. This method decomposes a singal into approximation a single component of the signal firstly, and then calculates the WVD distribution of each component, and finally sums to get WVD distribution of the original signal. The analysis results of the simulation signal and the experimental vibration signal showed that the method could successfully suppress WVD cross terms.Lastly, the method of this paper was compared with the traditional WVD method, traditional LCD with WVD distribution method and the EMD with WVD distribution method separately. The results showed that the method in this paper was better in terms of suppress cross terms. The validity of this method was showed in gas turbine generator fault analysis. The project instance analysis also proved that the method could be applied to the engineering practice of the fault diagnosis of rotating equipment.