Full Vector Spectrum Based Fault Diagnosis Methods For Gear System

With the development of science and technology and the progress of modern industry, all kinds of mechanical equipments become increasingly large-scale, high-speed, and heavy quantitative, the level of automation is also higher and higher. Gears have been widely applied in the running rotating machines because it's big carrying capacity, high driving accuracy and constant power transmission. The security and stability of mechanical equipments depends on directly the gear operation. Therefore, it is important to study the fault mechanism and diagnosis methods of gear system.In most cases, it will have deviations between the perpendiculars two-channeled of the same section due to the characteristics of the eddy whirling of rotor. If the data of single-channel was used for fault diagnosis, the information of fault feature was not complete and this will affect the accuracy and reliability of fault diagnosis. However, the full information technology could effectively integrate information of the two channels, the merged information will accurately reflect the actual state of the rotor whatever the structure or energy.This paper emphasized the full vector spectrum technology of full information technology. The full vector spectrum technology with high resolution could make further energy and three-dimensional analysis; you can also expand the range of analysis methods, such as vector power spectrum, cepstrum vector and short vector spectrum. The main object of the present work is to expand the full vector spectrum analysis and do gear fault diagnosis based on full vector spectrum technology and combined with traditional methods.It is non-stationary gear vibration signals when the gear was found failure. Vector-bispectrum can effectively analysis non-stationary signals and pick up the fault feature of gears. But the vector-bispectrum only remove the Gaussian noise, it is incapable of removing the disturbance of non- Gaussian noise. A new method the wavelet package vector-bispectrum is proposed by combining wavelet package with vector-bispectrum. The original signals are preprocessed by using the noise reduction of wavelet packet and then do vector-bispectrum analysis. Research shows that, this method can effectively eliminate the noise.Wigner higher-order spectrum analyze non-stationary signals from every angle after the suppression of the cross terms. It can truly represent the evolution of signal frequency with time and is a powerful tool for feature extraction and fault diagnosis of gear. Vector Wigner higher-order spectrum is proposed by combining full vector spectrum technology with Wigner higher-order spectrum and applied to gear fault diagnosis. The simulation and experiment results show that the method plays the advantages of Wigner higher-order spectrum and integrates the two-channeled information; the characteristics of gear vibration would be reflected totally through this method.